Three-Dimensional Microcomputed Tomography of Renal Vasculature in Rats
Current microscopic methods to view renal microvasculature reveal only a very limited portion of the total renal volume. Identification of connectivity for postglomerular vessels in the cortex and the medulla during functional states related to changes in sodium excretion will help better to understand the coupling of renal vasculature to tubular function. The purpose of this study was to investigate the possibility of visualizing the continuity of pre- and postglomerular vasculature using three-dimensional micro-computed tomography (micro-CT). Kidneys from normal rats were perfusion fixed in situ at physiological pressure, filled with latex microfil containing lead chromate, and embedded in plastic. The micro-CT scans of the intact kidneys were carried out on a rotating stage illuminated either by a synchrotron x-ray source or a conventional x-ray spectroscopy tube. Images were reconstructed by a filtered backprojection algorithm and volume-rendering techniques were utilized to display the vasculature. The reconstructed images clearly showed the large distribution vessels and the venous drainage of the kidneys, while pre- and postglomerular vessels and their vascular connections throughout the kidney were displayed in great detail. Efferent arterioles showed the characteristics of their peritubular capillary beds in the cortical and medullary regions. The vascular volume of the cortex was 27%, the outer stripe of the outer medulla 18%, the inner stripe of the outer medulla 30%, and the inner medulla 18%. In conclusion, micro-CT is a promising method to evaluate renal vascular architecture relative to physiological and pathological alterations.
- AA = arcuate artery
- AF = afferent arteriole
- AV = arcuate vein
- CCD = charge-coupled device
- Cort = cortex
- EF = efferent arteriole
- G = glomeruli
- MAP = mean arterial pressure
- micro-CT = micro-computed tomography
- IL = interlobar vessels
- IM = inner medulla
- ISOM = inner stripe of outer medulla
- OSOM = outer stripe of outer medulla
- RA = cortical radial artery
- RV = cortical radial vein
- VR = vasa recta
Changes in renal microcirculation often accompany disease conditions,1,2⇓ including hypertension.3,4⇓ However, the complexity of renal microvasculature, its geometric relationship to localized tubular function and its role in the development of disease conditions has defied adequate understanding, partly because of technical limitations on visualization and quantitation of that geometric relationship. Methods such as microangiography, light microscopy of carbon or silicone-based rubber injected specimens have been used extensively to demonstrate in great detail the vascular morphology of the kidney.5–10⇓⇓⇓⇓⇓ With such methods, qualitative and quantitative three dimensional information about the geometric interconnected arrangement of the microvasculature must be inferred from two dimensional tissue sections.11 Scanning electron microscopy of vascular casts has provided an excellent means to envision the three dimensional architecture,12,13⇓ but the method only allows observation of the vessels at prepared surfaces of the specimen.
In the present study, we examine renal vasculature by three-dimensional x-ray micro-tomography (micro-CT). The method provides a multidimensional means to study large representative samples of renal vasculature without limitations related to plane of section or problems related to superimposition. We used two micro-CT systems for this study. The first system was used to determine the feasibility of visualizing renal vasculature; it is installed at the Brookhaven National Laboratory and was developed by Flannery et al14 and subsequently modified prior to our use.15 This micro-CT system uses a synchrotron source that provides an intense collimated beam of monochromatic x-rays. This system provided relatively high resolution images with a high signal-to-noise ratio of a volume 6 mm in transverse diameter and 4 mm high. The second micro-CT that was used in this study was a bench-top version of the synchrotron-based system.16 The system differs from the Brookhaven system in that it has a conventional x-ray spectroscopy tube instead of a synchrotron as the x-ray source. The system is capable of scanning a cubic volume of up to 2 cm on a side, large enough to view a major organ from a small animal, such as the rat; this system was used to perform the whole kidney scans of this study.
Preparation of Kidneys for Micro-CT Scans
In accordance with our Institutional Animal Care and Use Committee guidelines, three male Sprague Dawley rats weighing between 250 and 300 g were anesthetized with a 100 mg/Kg intraperitoneal injection of Inactin (Bykgulden). To prepare the kidneys for micro-CT, a laparotomy was performed, and three loose ligatures were placed so that the first ligature was around the aorta caudal to the renal arteries; the second was around the aorta (between the celiac and cranial mesenteric arteries); the third was around the cranial mesenteric artery. A cannula connected to a perfusion pump (Syringe Infusion Pump 22, Harvard Apparatus) and pressure transducer (Recorder 2000, Gould Inc Instruments Systems Division) was inserted in the aorta at the first ligature caudal to the renal arteries. After the cannula was secured in place by the first ligature, the mean arterial pressure (MAP) was recorded. The aorta was then perfused retrograde with Ringer’s medium containing Lidocaine (0.2 mg/mL), the renal vein was cut open to allow venous drainage, and the second and third ligatures were tightened securely. Perfusion pressure was maintained at the recorded MAP by adjusting the flow rate of the perfusion pump.
When the perfusate draining through the venous vent was essentially free of blood and the kidneys were uniformly blanched, the kidneys were fixed by continuing the perfusion for 5 minutes with a 5% formalin solution. The excess formalin was then removed by continuing the perfusion with Ringers solution for 2 to 3 minutes. Radioopaque Microfil silicone rubber (MV-122, Canton Bio-Medical Products Inc), containing lead chromate, was then perfused through the aortic cannula. When filling was complete, the kidneys had a uniform coloration, and the microfil flowed freely from the renal veins. Throughout the entire procedure, the perfusion pressure was maintained at the recorded aortic pressure by adjusting the flow rate of the perfusion pump.
Following perfusion with microfil, the renal arteries and veins were ligated, and the kidneys were removed. Each kidney was immersed in 70% ethanol and then in aqueous 30%, 50%, 75%, and 100% glycerol solutions, successively, in each solution for 24 hours. The kidney was then agitated in 100% acetone for 1 minute and blotted with a clean dry cloth to remove the excess glycerol from the specimen. Each kidney was embedded in a synthetic resin (Bio-Plastic, Ward’s Natural Science).
All three kidneys were scanned by the Mayo micro-CT scanner.15 This scanner consists of an x-ray source, a rotatable specimen stage, a scintillator, a lens, a charge-coupled device (CCD)-based video camera, and a controlling computer. Scanning was performed by rotating the specimen in specified angular increments in the x-ray beam and acquiring an x-ray transmission image at each angle of view. The x-ray source in this scanner was a Philips spectroscopy x-ray tube (PW 2275/20 molybdenum anode long-line focus) with an effective 0.6×0.4 mm focal spot size. The tube was operated at 35 keV peak and 50 ma, providing molybdenum k-alpha emission at 17.5 keV. The beam was filtered with zirconium foil to suppress unwanted, higher energy components of the beam’s spectrum. The specimen stage was positioned 1 m from the x-ray focal spot, minimizing the angle of the cone beam source and preumbral blurring of the x-ray image. The specimen stage (Newport/Klinger RTN, 120 PP) provided precise rotation to within 0.001 degree and translation to within 0.1 μm (Newport/Klinger UT 100, 50PP). The specimen was rotated at angular increments of 0.499 degrees between views providing 721 views around 360 degrees. The exposure time for each view was up to 85 seconds. An image of each view was acquired on a clear cesium iodide crystal doped with thallium. A CCD camera (Princeton Instruments TE/ccd-1025 TKB/PI-1, Trenton, NJ) that was cooled to −30°C, recorded each view (1024×1024 pixels). The output images from the CCD were digitized to 16 bits and transferred to the controlling computer. A Nikon 50 mm f/2.8 enlarger lens was positioned between the scintillator and the CCD camera to provide variable optical magnification. Scans were made of all three kidneys at ×1 magnification and one kidney was scanned at ×4, thereby reducing the effective CCD pixel size from 21 μm to 6 μm on a side.
Two of the kidneys were also scanned using the National Synchrotron Light source at the Brookhaven National Laboratories. The synchrotron x-ray source provides an intense, collimated beam from which can be selected a narrow band-width (monochromatic) x-rays.14,17⇓ The rotating specimen stage and recording system of the Brookhaven system was similar to that of the Mayo system.
Three-dimensional volume images were reconstructed from the angular views using a modified Feldkamp’s filtered backprojection algorithm.18 The reconstructed images were comprised of cubic voxels 6 to 21 μm on a side, depending on the optical magnification. The radioopacity of each voxel was represented by a 16-bit gray scale value.
Analysis of Images
Image analysis was carried out using the ANALYZEtm19 software package. This package provides methods to compute, display, and analyze orthogonal and oblique sections from the reconstructed volume images. In addition, the software provides volume-rendering methods to envision the three-dimensional architecture of the renal vasculature. Blood vessel diameters were measured by the Pythagorean theorem in the “calipers” application of the software. Glomerular diameters were determined by counting the number of sections through individual glomeruli and multiplying the number of sections by section thickness. To determine the volumes of the kidneys and their regions, a stereologic application of the software was employed that placed an array of points spaced orthogonally 2.1 mm (ie, 50 voxels, each 42 μm) apart throughout the entire scanned volume. The fraction of the volume array that was the tissue of interest was determined by counting the number of points within the boundary of the tissue and dividing by the total number of points. The tissue volume was them determined by multiplying the tissues of interest by the total scanned volume of the image. Vascular volume fractions in the renal tissue were determined following the method of Hillman et al.20 The determinations were made from midtransverse slices that included the cortex, the outer stripe of the outer medulla, the inner stripe of the outer medulla, and the inner medulla. The average opacity was measured from regions of interest within a large interlobar artery (Oartery), in the tissues (Otissue), and in the background matrix outside the kidney (Obg). The vascular volume fraction in a tissue was determined as follows:
The images are represented by three dimensional arrays of cubic voxels having opacities representing the amount of microfil in blood vessels. In sections taken from the arrays, the regions of the kidney were easily identified by their characteristic vascular features (Fig 1). The cortex was characterized by glomeruli and the medulla by parallel bundles of vasa recta. The outer stripe of the outer medulla (OSOM), the inner stripe of the outer medulla (ISOM), and the inner medulla (IM) could be identified by the density of the vasa recta bundles. From these characteristics, the regional volumes of the kidney was determined to be the following: 61.4±1.5% as cortex; 18.4±1.4% as OSOM; 12.4±0.7% as ISOM; and 6.8±1.0% as IM (Table). These values were similar to those reported by Pfaller.21
The arteries and their corresponding veins were opacified by the microfil, indicating that the entire renal circulation had been filled (Figs 2 and 3⇓). The percent of renal mass filled with microfil (ie, the vascular volume fraction) varied with each region of the kidney. The blood vessels occupied 27.6±4.0% of the cortex, 18.6±2.8% of the OSOM; 30.0±1.1% of the ISOM and 18.2±2.1% of the IM. These values are similar to those reported by Hillman, et al.20
In general, with the exception of the capillaries, the morphological features of the blood vessels were distinct. However, the capillaries (ie, the finest details) were clearly evident only in the high magnification images (Fig 4). In the high magnification images, the voxel dimension (6 μm) approached the diameter of a capillary. In this case, a given capillary could be contained entirely within one voxel or shared by several voxels. Consequently, because of this partial volume effect, the brightness of capillaries varied. With low magnification images (ie, larger voxels), the morphological details of the capillaries were blurred and generally not clearly distinguishable.
Using volume rendering techniques, vascular details became quite distinct (Fig 2 and 3⇑). Typically, eight interlobar vessels coursed through the ISOM along outer wall of the renal pelvis (Fig 1). The arcuate vessels branched from the interlobar vessels in the OSOM, and the arcuate vessels continued along the boundary between the cortex and the medulla. The arteries and veins of the interlobar and arcuate vessels were in close apposition. For these vessels, the arteries had a slightly greater opacity and a smaller diameter than the veins (Fig 1 through 3⇑⇑). The vessel diameters were: 536±41 μm (mean±SEM) for the interlobar arteries, 847±10 μm for the interlobar veins, 307±11 μm for the arcuate arteries, and 387±7 μm for the arcuate veins.
The arcuate arteries branched into numerous cortical radial (interlobular) arteries which extended at right angles from the arcuate arteries into the cortex. Glomeruli were arranged cylindrically around the cortical radial arteries, tethered by their afferent arterioles. The glomerular diameter was 186±5 μm. Efferent arterioles of the superficial glomeruli could be seen distributing toward the peritubular capillaries of the cortex. The cortical peritubular capillaries were drained by converging cortical radial (interlobular) veins. The efferent arterioles of the juxtamedullary glomeruli extended into the outer stripe of the outer medulla, forming vasa recta bundles and medullary capillary beds (Fig 4).
The purpose of this study was to determine the extent to which renal vasculature and its three-dimensional interconnectivity and distribution could be envisioned by micro CT. The data reconstructed from the micro-CT scans were computarized volume-images (ie, three dimensional numeric arrays) of the renal vasculature. Such volume-images could be explored and viewed by a number of computer techniques ranging digital sections to volume rendered displays of the entire vasculature. In the volume-images, the architecture of the vasculature was clearly visible and corresponded to observations made in the past using histological techniques.5–10⇓⇓⇓⇓⇓ Furthermore, the fine details of the microvasculature seen with the micro-CT were comparable to observations made by microradiology of histological sections of renal vasculature opacified with x-ray contrast media and to observations of microvascular casts using light microscopy and scanning electron microscopy.1,2,5–10,12,13⇓⇓⇓⇓⇓⇓⇓⇓⇓
The use of vascular casting techniques has been used extensively since the time Trueta’s classic study22 and has provided fundamental understanding of renal microvascular architecture and tubular vascular relationships.23,24⇓ However, the visualization of specific details has required the kidney to be physically cut or sectioned. With micro-CT, the tissue remains intact. Furthermore, micro-CT provides a means for computerized quantitative analysis. The volumetric data of renal tissue composition in the present study was similar to that obtained by Pfaller21 using stereological techniques on histological sections. Likewise, the vascular volume determined in the present study was in agreement with that obtained by Hillman et al.20
The micro-CT may provide a means to explore alterations in peritubular microvasculature in experimental disease conditions. In acute situations where the renal artery is obstructed so that renal perfusion pressure is below the limit of autoregulation, there is a dramatic decline in cortical vascular distribution volume while the medullary distribution volume is less affected.25 In chronic situations of renovascular stenosis, the cortex continues to be underperfused while the medullary circulation is preserved.3,4⇓ Although changes in perfusion have been documented in a stenotic kidney, little is known about the morphological changes that occur in the microvasculature of peritubular circulation. These changes could be important in understanding the recovery process where renovascular circulation is reestablished in a stenotic kidney. Furthermore, processes such as neovascularization26 may be examined by micro CT. Equally intriguing is the state of peritubular circulation in a nonstenotic kidney that is exposed to higher blood pressure. While much is known about glomerular changes,27 little is known about the peritubular vasculature. In other experimental conditions where a portion of the kidney is ablated, the remnant kidney hypertrophies, and the remaining nephrons increase function. In this situation, there is often hypertension accompanied by glomerular enlargement and damage.28 Again, little is known about peritubular circulation in the remnant nephrons or in the zone surrounding the ablation where there is partial ischemia.29 For these ischemic zones, filtration is compromised, but the nephrons may be minimally perfused in combination with neovascularization.
In conclusion, we have shown that microvascular circulation may be seen and quantified with micro-CT. In the past, little attention has been directed toward the peritubular circulation that is integrally involved with nephron function. With micro-CT, anatomical changes in peritubular circulation may be investigated in relation to experimental maneuvers that alter function.
The data of this study, in part, was obtained at the Brookhaven National Synchroton Light Source, Brookhaven National Laboratory and was supported by the United States Department of Energy, Division of Materials Sciences and Division of Chemical Sciences by contract DE-AC02-76CH00016. In addition, this work was supported, in part, by NSF grant BIR9317816 and NIH grant RR11800 and HL16496. Dr. Garcia-Sanz was supported by the National Committee for the Development of Science and Technology of Venezuela (CONICIT), and Elmor Laboratories. Dr. Rodriguez-Barbero was supported by a fellowship of the Spanish Society of Hypertension donated by Merck Sharp & Dohme. The authors are grateful to Steven Jorgensen, Patricia Lund, and Denise Reyes for their technical assistance, to Kristy Zodrow for preparation of this manuscript, and to Julie Patterson for preparation of the illustrations.
- Received September 17, 1997.
- Revision received October 16, 1997.
- Accepted October 31, 1997.
Dworkin LD, Brenner BM. The renal circulation. In: The Kidney. Brenner BM, 5th ed. Philadelphia: WB Saunders Co; 1996; 247 –285.
Lemley KV, Kriz W. Structure and function of the renal vasculature. Tisher CC, Brenner BM, eds: In: Renal Pathology with Clinical and Functional Correlations. 2nd ed. Philadelphia: JB Lippincott Co; 1994; 981 –1026.
Moffat DB, Fourman J. The vascular patterns of the rat kidney. J Anat (London). 1963; 93 : 543 –553.
Plakke RK, Pfeiffer EW. Blood vessels of the renal medulla. Science. 1964; 146 : 1683 –1685.
Flannery BP, Deckman HW, Roberge WG, D’Amico KL. Three-dimensional x-ray microtomography. Science. 1987; 237 : 1439 –1444.
Ritman EL, Jorgensen SM, Lund PE, Thomas PJ, Dunszmuir JH, Romero JC, Turner RT, Bolander ME. Synchrotron-based micro-CT of in situ biological basic functional units and their integration. Proc of SPIE., San Diego, CA, July 27-August 1, 1998 (In press).
Ritman EL, Dunsmuir JH, Faradini A, Finch DV, Smith KT, Thomas PJ. Local reconstruction applied to x-ray microtomography. In: Chavent G, IMA Volumes in Mathematics and its Applications, Inverse Problems in Wave Propagation. Papanicolaou G, Sacks, P, Symes W, eds: New York: Springer-Verlag, 1997; 443 –452.
Margaritondo G. Introduction to synchrotron radiation. New York: Oxford Press; 1988: 280 .
Robb RA, Hanson DP. A software system for interactive and quantitative visualization of multidimensional biomedical images. Aust Phys Eng Sci Med. 1991; 14 : 9 –30.
Trueta J, Barclay AE, Daniel PM, Franklin KJ, Prichard MML. Studies of the Renal Circulation. Springfield, IL: Charles C. Thomas Publisher; 1947; 182 .
Beeuwkes R. Efferent vascular patterns and early vascular-tubular relations in the dog kidney. Am J Physiol. 1971; 221 : 1361 –1374.
Beeuwkes R, Bonventre JV. Tubular organization and vascular tubular relations in the kidney. Am J Physiol. 1975; 229 : 695 –713.
Bentley MD, Lerman LO, Hoffman EA, Fiksen-Olsen MJ, Ritman EL, Romero JC. Measurement of renal perfusion and blood flow with fast computed tomography. Circ Res. 1994; 74 : 945 –951.
Hollenberg NK, Odori T. Centripetal spread of arterial collateral endothelial cell hyperplasia after renal artery stenosis in the rat. Circ Res. 1987; 60 : 398 –401.