Three-Dimensional Reconstruction of the Mouse Nephron
Xiao-Yue Zhai*,
Jesper S. Thomsen,
Henrik Birn*,
Inger B. Kristoffersen*,
Arne Andreasen and
Erik I. Christensen*
Departments of * Cell Biology, Connective Tissue Biology, and Neurobiology, Institute of Anatomy, University of Aarhus, Aarhus, Denmark
Address correspondence to: Dr. Erik Ilsø Christensen, Department of Cell Biology, Institute of Anatomy, University of Aarhus, DK-8000 Århus C, Denmark. Phone: +45-89-42-30-57; Fax: +45-86-19-86-64; E-mail: eic{at}ana.au.dk
Received for publication August 1, 2005.
Accepted for publication October 1, 2005.
Renal function is crucially dependent on renal microstructurewhich provides the basis for the regulatory mechanisms thatcontrol the transport of water and solutes between filtrateand plasma and the urinary concentration. This study providesnew, detailed information on mouse renal architecture, includingthe spatial course of the tubules, lengths of different segmentsof nephrons, histotopography of tubules and vascular bundles,and epithelial ultrastructure at well-defined positions alongHenle’s loop and the distal convolution of nephrons. Three-dimensionalreconstruction of 200 nephrons and collecting ducts was performedon aligned digital images, obtained from 2.5-µm-thickserial sections of mouse kidneys. Important new findings werehighlighted: (1) A tortuous course of the descending thin limbsof long-looped nephrons and a winding course of the thick ascendinglimbs of short-looped nephrons contributed to a 27% averageincrease in the lengths of the corresponding segments, (2) thethick-walled tubules incorporated in the central part of thevascular bundles in the inner stripe of the outer medulla wereidentified as thick ascending limbs of long-looped nephrons,and (3) three types of short-looped nephron bends were identifiedto relate to the length and the position of the nephron andits corresponding glomerulus. The ultrastructure of the tubulesegments was identified and suggests important implicationsfor renal transport mechanisms that should be considered whenevaluating the segmental distribution of water and solute transporterswithin the normal and diseased kidney.
The renal architecture is arranged elaborately to fulfill thephysiologic demands for reabsorption of filtered substancesand for urine concentration. The architecture includes the formationof the renal zones, the population of short-looped nephrons(SLN) and long-looped nephrons (LLN), the distribution of thetubule segments of nephrons, the tubular-vascular histotopographyin medulla, and the epithelial configurations of the differenttubule segments.
From the 1960s to the 1980s, Kriz and others studied the renalmicrostructure, including the distribution of nephrons, thetubular-vascular relations in the medullary zones, and the ultrastructureof the epithelia along Henle’s loop. Simultaneously, thediversities in renal structure were demonstrated between differentspecies, including rat, mouse, hamster, and rabbit (1–3).In the late 1980s, a standard nomenclature for the kidney structure,which has been widely adopted, was published (4). With the developmentof digital techniques, three-dimensional (3D) representationsof the tubule segments of nephrons were performed to a limitedextent at the proximal tubules (PT) (5), distal tubule (6),and at the thin limbs (TL) of the LLN in the inner medulla (IM)(7) of rats. Two mathematical regional-based models combiningmorphologic and immunohistochemical findings have been set upto simulate the urine concentration mechanism in rat medulla(8).
At present, a large body of basic nephrology research focuseson mapping the segmental distributions of a variety of receptors(9), water channels (10), tight junction integral proteins (11–13),and ion transporters (14) in mouse kidneys, as mice are consideredto be the most convenient animal model for physiologic studiesand gene manipulation (15, 16). However, for obtaining detailedand systematic information on renal architecture in mice, furtherinvestigations are imperative. Our study provides a detaileddescription of the 3D structure of the mouse nephrons, includingthe course and the lengths of individual nephrons, histotopographyin outer medulla, and ultrastructure in well-defined segmentsof Henle’s loop, and distal convolution.
Preparation of Renal Tissue
The tissue preparation has previously been described in detail(17). Briefly, the kidneys from three 8-wk-old male C57/BL/6Jmice, 25 g of body weight, were fixed by perfusion through theabdominal aorta with 1% glutaraldehyde in 0.06 M sodium cacodylatebuffer and 4% hydroxyethyl starch. Tissue blocks cut perpendicularto the longitudinal axis of the kidney were postfixed for 1h in 1% OsO4, and embedded in Epon 812. From each kidney, atotal of 2500, 2.5-µm-thick consecutive sections wereobtained from the surface to the papillary tip and stained withtoluidine blue. The average section thickness was determinedby counting the number of sections that contained a sphericalstructure of measured diameter (renal corpuscle). The analysisof five corpuscles from each of the three kidneys gave a meansection thickness of 2.5 µm (2.56 µm ± 0.11;n = 3.). All animal experiments were carried out in accordancewith provisions for the animal care license provided by theDanish National Animal Experiments Inspectorate.
Image Recordings and Alignment
Using an Olympus AX 70 microscope with a x2 objective and equippedwith a digital camera (Olympus DP 50; Olympus, Tokyo, Japan)attached to a standard PC, every second section was digitized.Four overlapping digital recordings from each section were combinedinto one 24-bit color image by use of "analySIS" (Soft ImagingSystem, Version 3.2; Soft Imaging System, Munster, Germany).The final image size was 2596 x 1889 pixels, whereby each pixelcorresponded to 1.16 µm x 1.16 µm. The image alignmentwas based on a previously established alignment algorithm (18,19). Each color image was split into three grayscale images(-r, -g, and -b). The relative transformation values were determinedbetween two consecutive images and then summarized into a setof absolute transformation values. The absolute transformationvalues underwent a high-pass filtration to avoid potential distortionsof the image stacks as a result of small but accumulating "trends."Then, the grayscale images were "moved" according to these absolutetransformation values. Finally the moved -r, -g, and -b grayscaleimages were merged to form a set of transformed color images.The quality of the alignment was checked in two different ways:In the first test, the aligned images were arranged as consecutiveimages similar to a movie film. A copy of this elongated imagewas made beginning with image 2 instead of image 1. Then thecopy was superimposed onto the original so that the consecutiveimages could be compared two by two. In the second test, theconsecutive images were turned into an animated film by theprogram "gifsicle" and inspected by use of "gifview." In casesin which the alignment was not optimal, the alignment was repeatedwith a new set of initial conditions until the alignment convertedinto a global optimum. The image recording and alignment procedurewas carried out on a Windows-based PC (Microsoft Windows 98;Intel, Santa Clara, CA), whereas the control procedures involving"gifsicle" and "gifview" were made on the Linux platform (MandrakeLinux 10.1; AMD Athlon, Sunnyvale, CA).
Digital Tracing and 3D Presentation
The tracing of the nephron paths was performed on the Linux-basedPC. A series of custom-made computer programs were written inC on the Linux system for tracing and for calculation of lengths,distances, etc.
The tracing was conducted among collecting duct (CD) "families."Such a "family" is defined as a CD and the nephrons that draininto the CD by the connecting tubules (CNT) at the superficialcortical level. Nephron tracing started at the urinary poleof a randomly chosen glomerulus and ended at the meeting ofthe CNT. The CD was traced from the cortex toward the IM.
A marker was placed manually in the tubule lumen along the courseof the tubule. For each marker, the x-, y-, and z-coordinateswere recorded in a data file, where the z-coordinate correspondedto the image number. The data file thus documented the pathsof all traced nephrons in three dimensions. All subsequent dataanalyses were based on this data file. In addition, the transitionsbetween each segment of the nephrons were recorded manuallyin a separate data file. The tubular profiles and the transitionsbetween different segments were normally identified from thestructure in the digital images. In only a few cases were theoriginal sections inspected in a light microscope to identifythe tubules or the transitions. Furthermore, the surface ofthe renal capsule and the papillary epithelia was traced toobtain the distance from the glomerulus center to the renalcapsule and to describe the spatial relationship between thetraced nephrons and the kidney boundary.
The CD and the deeper bends of the LLN exited the tissue blocksin the lower IM. Therefore, these LLN were traced both fromthe urinary pole of the glomerulus and from the macula densauntil they exited the block.
Nephron Length Measurements
Each nephron was subdivided into four segments: PT, TL, thickascending limbs (TAL), and distal convoluted tubules (DCT).The lengths of the TL of the completely traced LLN were subdividedinto two continuous parts: Descending thin limb (DTL) and ascendingthin limb (ATL). The length of each segment was calculated asthe sum of the Euclidian distance between two subsequent markersin the course of the nephrons. For reducing measurement noisearising from the alignment procedures and the manual placementof the markers, the path was smoothed using a triangular movingaverage window in such a way that the previous point in thepath was weighted with one fourth, the current point with onehalf, and the next point in the path with one fourth.
Electron Microscopy
On the basis of the tracing, sections that represented the differentsegments of the nephrons and the transitions in distal convolutionswere selected for ultrastructural analysis. These sections werere-embedded in Epon, sectioned, and observed in a Philips CM100 electron microscope (FEI Company, Hillsboro, OR).
General Description of the Spatial Course of Nephrons
In total, 30 CD, 127 SLN, and 41 LLN were traced. Of these,24 LLN were complete, whereas 17 LLN were incomplete as a resultof their deep bends exiting the tissue blocks. The nephronswere grouped into two types according to the location of Henle’sloop bends (4): SLN, which form their loop bends within theouter medulla, and LLN, which have their loop bends in the IM.A Henle’s loop consists of a pars recta of the PT, a DTL,an ATL, and a TAL (pars recta of the distal tubule). The majorfindings in the spatial course of the various types of SLN andLLN are illustrated in Figure 1.
Figure 1. Schematic representation of the typical nephrons and collecting duct (CD) organization: Tortuous descending thin limbs (DTL) of the long-looped nephrons (LLN; large arrow); winding course of thick ascending limbs (TAL) of short-looped nephrons (SLN) and CD (arrowheads); a piece of TAL inserted in the DTL of LLN (LLNt; small arrow), which forms its bend just beneath the "transitional zone" within the inner medulla (IM); and three different types of SLN bends (SLN1, SLN2, and SLN3). On the basis of the distribution of the nephron segments, the renal zones were defined, including cortex, outer stripe of outer medulla (OSOM), inner stripe of outer medulla (ISOM), and IM. The postmacula densa segments and the directions of the proximal tubules (PT) arising from their glomeruli are also illustrated.
Renal Zones.
In Figure 2A, the boundary between the outer medulla and theIM was defined clearly by the transition of the LLN from ATLto TAL. In contrast, the boundary between the outer stripe ofthe outer medulla (OSOM) and the inner stripe of the outer medulla(ISOM) was less sharply defined by the transition from the parsrecta of the PT to the DTL. The thicknesses of the OSOM andthe ISOM were approximately 0.5 and 1.2 mm, respectively. Thecortical-medullary boundary was determined by the location ofthe arcuate vessels. The thickness of the cortex was approximately1.3 mm, which was subdivided arbitrarily into superficial, middle,and juxtamedullary cortex, corresponding to outer 30%, middle30%, and inner 40% of the cortex, respectively. The thicknessof the IM was not determined as a result of the curved papilla,which was not included in the tissue blocks.
Figure 2. All SLN, LLN, and CD traced within one kidney (A) and selected cross-sections showing tubular-vascular histotopography (B through D). (A) The red lines represent the renal capsule and the papilla. The white dots represent glomeruli (not actual size). Tubule segments are represented in different colors: PT, blue; thin limbs (TL), green; TAL, red; DCT, purple; connecting tubules (CNT), orange; and CD, brown. (B) Level of the OSOM (approximately 1570 µm from the renal surface). (C) Level of the initial part of the ISOM (approximately 1925 µm). (D) Level of the middle to lower ISOM (approximately 2585 µm). White circles highlight the same vascular bundles. Each red number denotes a specific nephron or CD. In the vascular bundle, tubules 84, 88, 92, 96, 102, 103, and 105 are the TAL of the LLN. (C) The DTL of SLN 4, 5, 47, 48, 49, 50, 79, and 93 are integrated into the vascular bundle. In the interbundle region, the elongated tubules 0, 8, 31, and 91 reflect a tortuous course of the DTL of LLN, whereas the elongated tubules 4 and 47 reflect a winding course of the TAL of SLN, likewise the tubule 37, a winding course of the CD. In B and D, the same vascular bundle became smaller (white circles). In D, tubules 25, 45, 50, 75, 82, 93, 100, and 104 are longer SLN with a type 3 bend, e.g., tubule 93 and 104 reveal a prebend TAL segment. Scale = 50 µm.
Proximal Tubules.
Notably, most (>90%) of the PT that initiated at the superficialcortex left their glomeruli ascending toward the renal surface,whereas >90% of the PT that initiated at juxtamedullary cortexleft their glomeruli in a short descent and then turned to ascend.The PT that originated within the middle or early juxtamedullarycortex left their glomeruli taking a horizontal course but turnedto ascend after a short distance (Figure 1).
The proximal convoluted tubules made coils around their ownglomeruli and occupied either small, tightly packed domainsin the superficial and middle cortex or larger, loosely packeddomains in the juxtamedullary cortex. The corresponding DCTand cortical TAL were located within or nearby the domains.The individual domains contained only one nephron, and onlya few peripherally situated tubules were found to interminglewith the neighboring nephron tubules from the same "family"from time to time.
In this study, none of the LLN was found to originate from thesuperficial cortex, and none of the SLN was found to originatefrom the juxtamedullary cortex (Figure 3, A and C). Moreover,no nephrons formed loop bends within the cortex, as opposedto findings in humans, pigs, and occasionally in rabbits (20).
Figure 3. All LLN (A) and SLN (C) traced in one kidney, together with selected LLN (B) and SLN (D) at higher magnification. (A and B) The glomeruli are placed at different levels within the middle to juxtamedullary cortex. The PT have no straight part, and the tortuous course continues into the initial part of the DTL (arrows). All loop bends are located in the IM. The TAL start at the boundary of ISOM and IM and continue toward the cortex. The DCT and their corresponding proximal convoluted tubules intermingle and coil to form the structure domain. The CNT go upward into the superficial cortex. (C) The glomeruli are located within the superficial and middle cortex. The loop bends are located at different levels within the ISOM. The lower bends have TAL epithelia (red) lining the bends and the prebend descending limbs (prebend TAL segment; one LLN as reference). (D) Three bend types of SLN: "1" with DTL epithelium covering and continuing a short distance in ascending limb; "2" with a transitional epithelium, DTL to TAL; and "3" with a prebend TAL segment of 288 µm. "3i" has a type 3 bend (446-µm prebend TAL segment) that is deeper in the innermost stripe and the terminal part of the loop is winding. The bend types correspond to the position of their glomeruli. The TAL of SLN run windingly through the upper part of the ISOM. Those originating from the middle cortex reveal a tortuous course of the pars recta of the PT and initial part of the DTL.
Loops of Henle and Tubular-Vascular Histotopography.
The most intriguing observation was that, through the outerpart of the ISOM, the tortuous course of the "straight" partof the PT continued into the upper part of the DTL in all LLNand the initial part of the DTL of the SLN, which originatedfrom middle to juxtamedullary cortex. The tortuous course ofa DTL increased the total length of the DTL of the LLN by approximately27%. The epithelium that covered the tortuous course of theDTL was characterized at the electron microscope level (seebelow). The TAL of the SLN and their CD also ran windingly throughthe ISOM, where the tortuous course of the DTL of the LLN wasseen (Figure 3, B and D). The winding course of the TAL increasedthe total length of the TAL of the SLN by approximately 27%as well.
The corresponding descending and ascending limbs of SLN andLLN that originated from the superficial and middle cortex ranclosely together through the medullary rays. The loop limbsfrom one "family" were found to run together with their commonCD. However, at different levels through the OSOM close to theISOM, the corresponding descending and ascending limbs startedto separate, running in parallel. The distance between the correspondinglimbs varied up to approximately 100 µm. The tubules occasionallyshifted positions relative to neighboring tubules from the same"family." The thick descending limbs of LLN from the juxtamedullarycortex occupied a large tissue volume because of the tortuous"straight" part of the PT, and the corresponding TAL ran nearby(Figure 2B, e.g., tubule 0).
In the ISOM, the tubule segments were located in a distincthistotopographic position. In the outer part of the ISOM, enclosinglarge vascular bundles, single thick-walled tubule profileswere constantly found in the central part of the vascular bundles(Figure 2C) and identified as TAL of LLN with deeper loop bends.However, most TAL of LLN were located peripherally. The DTLof SLN were constantly integrated in the outer or middle layerof the vascular bundles, and the tortuous DTL of LLN were locatedin the interbundle regions, together with the winding TAL ofSLN and their CD (Figures 1 and 2C). In the inner part of theISOM, SLN successively looped back, and the vascular bundlesplit into smaller bundles with both DTL and TAL of LLN andremaining SLN running in the vicinity of the bundles (Figure 2D).
In the IM, the limbs of Henle’s loop were only DTL andATL from LLN, and they ran close to each other and next to theirCD and vasa recta, which were identified at the electron microscopelevel. The ATL were constantly closer to their CD than the DTL.
Bend of Henle’s Loop.
We identified three types of bends of the SLN according to theepithelium lining the bends: Type 1 bends (21% of the totalSLN) with DTL epithelium covering the bends and continuing fora short distance into the ascending limbs (<60 µm),type 2 bends (33%) with a transforming epithelium from DTL toTAL within the bend, and type 3 bends (47%) with TAL epitheliumcovering the bend and constituting 50 to 450 µm of thedescending limb (prebend TAL segment). Generally, the deeperthe bends, the longer the prebend TAL segments of type 3 bends(Figure 3C).
Types 1 and 2 bends were located at the middle level of theISOM, and their corresponding glomeruli were located at theaverage outer 17 and 24% of the cortex, respectively (200 and300 µm from the renal surface). Type 3 bends were locatedat the various levels of the inner half of the ISOM, a few reachingthe so-called innermost stripe (1), and their correspondingglomeruli were located at the average outer 40% of the cortex(500 µm from the renal surface; Figure 3D). The SLN withtype 1 and type 2 bends were comparable in tubule course, lengthsof segments, and position of the bends. The lengths of the DTLand TAL of the SLN with a type 3 bend were 8 and 9% longer,respectively, than the lengths of the same segments of the SLNwith type 1 and type 2 bends. Most SLN bends had limbs separatedby approximately 20 µm. Broad and flat bends with limbsthat were separated by a distance of 40 to 50 µm wereoccasionally identified among all bend types. The bends of allfully traced LLN were located at successive levels in the IM.
A transitional zone was identified around the boundary betweenISOM and IM, in which almost no loop bends were found, eitherfrom SLN or from LLN. It comprised the innermost stripe andthe initial part (approximately 100 µm thick) of the IM.Very few bends from SLN reached the innermost stripe, and theseSLN were characterized by a winding terminal loop and a type3 bend (Figure 3D, "3i"). Only a few LLN looped back just beneaththis zone. These LLN were structurally similar to the otherLLN, except for an approximately 100-µm-long and thick-walledsegment inserted in the DTL at the level of the transitionalzone (Figure 1, LLNt). The inserted segment was ultrastructurallycharacterized by TAL epithelium.
Distal Convolution and CD.
The distal convolution consists of DCT, connecting tubule (CNT),and initial collecting tubule, which is the top and branchedpart of the cortical CD (4). Similar to the rat (21), the DCTcommenced downstream from the macula densa at an average distanceof 85 µm. The DCT of the SLN were more coiled comparedwith the DCT of the LLN. The DCT that were situated in the outermostof the cortex touched the renal capsule one to two times, comparedwith five to seven times for the proximal convoluted tubules.The transition from late DCT to CNT was defined by the decreasein epithelium height of the CNT. The CNT in the superficialcortex occasionally ran just beneath the capsule for some distance.The CNT that originated from LLN ascended toward the superficialcortex within medullary rays to form an "arcade" (22) (Figure 3A).The CNT of the SLN drained into their own CD either byan arcade or directly into the initial collecting tubule (Figure 4).
Figure 4. Ten CD with the CNT deriving from 10 "families" (A through F). A, B, D, and E consist of two families. The nephrons are omitted to show the pattern that a CD (brown) connects with five to seven nephrons by CNT (different colors) at superficial cortex. G is a higher magnification of cortical part of the C: At the top of the CD (brown) is the branched cortical CD, initial collecting tubule (ICT; pale yellow); yellow is the CNT of an LLN. It ascends in the cortex to form the arcade, and the CNT from SLN drain into the CD by either joining directly into ICT or joining the arcade first and then draining into the ICT.
In total, 30 "families" were traced. According to the completelytraced "families," a CD collected six to seven nephrons, ofwhich either one or two were LLN, reflecting an 82:18 ratio(n = 25 "families") of SLN to LLN. The cortical CD ran withinmedullary rays inward to the medulla without receiving morenephrons in the middle and the juxtamedullary cortex. The firstjoining of two CD often occurred around the boundary betweenISOM and the IM. Subsequent joinings took place at successivelevels in the IM.
Segment Lengths
The lengths of the four well-defined segments of the nephronwere obtained for SLN and LLN. They were in sequence PT, TL,TAL, and DCT. The average length for each segment of the SLNwas determined (Table 1).
Table 1. The lengths of segments of SLN in mouse kidneysa
The lengths of the different segments for the entirely tracedLLN are shown in Figure 5A. The lengths of the PT and the TLthat originated from deeper than 70% of the cortex increasedwith increasing depth of their glomeruli in cortex. The lengthsof the PT and TL that originated from 40 to 69% of the cortexvaried slightly. The lengths of the TAL and DCT were relativelyconstant. Therefore, the increase in total length paralleledthe increase in PT and TL lengths. The TL of LLN was subdividedinto DTL and ATL, which on average constituted 80 and 20% ofthe total length of the TL, respectively.
Figure 5. (A) Bar chart showing the lengths of the four segments of the completely traced LLN. They are ordered according to the distance (in percentage) of their glomeruli from the renal surface. The lengths of the TAL and DCT are constant. The increase in total length corresponds to an increase in length of the PT and the TL. (B) Bar chart showing the relative lengths of all completely traced nephrons. The nephrons are ordered according to absolute length within each of the four nephron types: SLN1 (n = 27), SLN2 (n = 41), SLN3 (n = 59), and LLN (n = 24).
The normalized length for all completely traced nephrons isshown in Figure 5B. The relative lengths of the four segmentsare constant for all SLN, independent of bend type and independentof the total nephron length. However, for LLN the relative lengthsof the TAL decreased corresponding to an increase of TL withincreasing total nephron length. The relative lengths of theDCT remained constant and smaller as compared with the relativelengths of the SLN.
Electron Microscopy
We have previously quantitatively described the ultrastructurealong the PT (17). In this study, we therefore focused on theTL of Henle’s loops, as well as the distal convolutions.On the basis of multiple micrographs at successive levels throughthe renal cortex and medulla, our analyses largely agreed withand expanded previous findings.
Four ultrastructurally distinct segment epithelia along theTL described in different species (3, 23) were observed in miceas well (Figure 6). The DTL of most SLN were lined by a uniformlythin and simple epithelium (type 1). The DTL of LLN were linedby three distinctly differentiated epithelia: The epitheliumof the upper tortuous DTL of LLN and the SLN with the deepestbends within the ISOM was characterized by complex and highlydifferentiated apical and basolateral membranes and radiallyoriented mitochondria (type 2). The epithelium of the DTL ofLLN located at the boundary between ISOM and IM was characterizedby many thin luminal microvilli and numerous finger-like infoldingsof the basal membrane, fewer mitochondria and other organelles,and fewer and less shallow tight junctions (type 3). The epitheliumof the lower part of the DTL, the loop bends, and the ATL wascharacterized by many lateral processes with tight junctionsand dense apical membrane with few luminal microprojections(type 4). The part of the DTL covered by type 4 epithelium wasapproximately 700 µm long, i.e., a prebend ATL segment.The limbs in the IM of those LLN with the bends located justbeneath the transitional zone were covered by type 4 epitheliumonly. The epithelia along DTL of the LLN transformed graduallyfrom one type to another.
Figure 6. Four distinct epithelia along the TL of both SLN and LLN. The figure comprises two columns of images representing lower and higher magnifications (right). Type 1 is a cross-section of the DTL of an SLN (440 µm from the transition of the PT to the DCT); higher magnification of type 1 shows a simple and thin epithelium with few microprojections, abundant ribosomes, and sparse mitochondria throughout the cytoplasm. Type 2 is a cross-section of the DTL of an LLN (370 µm from the same transition as above); higher magnifications of type 2 show a highly specialized epithelium: Abundant short and plump microvilli, basolateral infoldings, and the highly interdigitated cellular processes with the shallow tight junctions (small arrows) and the radially oriented mitochondria. Type 3 is a cross-section of a DTL of LLN around the boundary between the ISOM and the IM (2790 µm from the transition of PT to DTL); higher magnifications of type 3 show the highly differentiated membranes: Numerous relatively long microvilli, abundant finger-like basolateral infoldings throughout the cytoplasm (large arrows), and the less shallow tight junctions (small arrow) and few mitochondria. Type 4 is a cross-section of the ATL of an LLN (700 µm from the transition of the ATL to the TAL); higher magnifications of type 4 show the bold and dense apical membrane, basolateral infoldings, cellular processes with tight junctions (small arrows), and various organelles with sparse ribosomes throughout the cytoplasm. Scales = 1 µm.
The transition from the TAL to the DCT was observed at a distanceof an average of 80 µm from the macula densa and definedby an abrupt increase in cell height. Ultrastructurally, theepithelia of the TAL and the DCT are characterized by extensiveinvaginations of basolateral plasma membrane and numerous elongatedmitochondria, which were placed basolaterally and perpendicularto the basement membrane (Figure 7). The transition from theDCT to the CNT was defined by the appearance of CNT cells andan abrupt decrease in epithelial height. Along this segment,four distinct cell types were identified by their specific morphologiccharacters (24): DCT cells, intercalated cells, CNT cells, andprincipal cells. However, principal cells were rarely seen inthe early CNT. In addition, three subtypes of intercalated cellswere identified: Type A, type B, and non-A–non-B intercalatedcells, based on their ultrastructural and immunochemical characters(Figure 8) (25).
Figure 7. The transition from TAL to DCT. Electron micrographs show the transition from TAL (40 µm beyond the macula densa) to the DCT. (A) The lower lumen is the TAL and the upper is DCT, corresponding to B and D, respectively. (C) One of the TAL cells showing a centrally placed nucleus and less mitochondria in the basal compartment. A cilium is visible. (E) One of the DCT cells showing an apically placed nucleus and tightly packed mitochondria and membrane infoldings in the basal compartment. Scales for A, B, and D = 5 µm; scales for C and E = 1 µm.
Figure 8. The transition from DCT to CNT of an LLN. (A) The tubular transformation from the late DCT into the CNT (upper lumen). (B through E) Different cell types constituting the wall of the CNT: CNT cell (B), type A intercalated cells (C and D), part of DCT cell in C (the short arrow), and non-A–non-B intercalated cell (E). The wall of the late DCT is composed mainly of DCT cells; however, a few intercalated cells are interspersed. Compared with DCT cells, the segment-specific cells, CNT cells, were obviously lower and characterized by a centrally placed nucleus, randomly oriented mitochondria, basolateral invaginations of the plasma membrane, and a relatively smooth lateral margin. Type A and non-A–non-B intercalated cells were commonly characterized by numerous microvilli covering the apical surface, abundant vacuoles, and mitochondria. Compared with type A, the mitochondria in non-A–non-B intercalated cells were mainly located in the bulging apical area above the nucleus, and smaller vesicles were distributed throughout the cytoplasm. Scale for A = 5 µm; scales for B, C, D, and E = 1 µm.
This study presents the results of computer-assisted, 3D reconstructionof 200 nephrons and CD from mouse kidneys. In contrast to earlierstudies, the digital tracing allowed a detailed characterizationof the spatial structure and interrelation of a large numberof nephrons and enabled for the first time an ultrastructuralcharacterization of the tubule epithelium at well-defined distancesfrom the glomerulus and relevant tubule segment transitions.The study reveals several new, important features of mouse kidneymicrostructure with substantial influence on renal function.However, it should be emphasized that the findings of this studyare based on C57/BL/6J mice only.
A correlation has been established between the position of theglomeruli within the renal cortex and the point of origin andthe direction of the PT from the glomeruli. The course takenby the PT is similar to observations of efferent arteriolesfrom superficial nephrons in dog and rat kidneys (26, 27). TheDCT was surrounded by proximal convolutions of the same nephron,forming a relatively separate structural domain. This structuraldomain in the cortex may be related morphologically to the regulationof ion reabsorption in the proximal-distal tubules by tubuloglomerularfeedback (28), as observed in renal physiologic studies andcomputer models (29–31).
The ratio of the number of SLN to LLN was 82:18, which may becompared with previous observations—70:30 in rats (32),34:66 in rabbits (22), 0:100 in dogs (33), and 85:15 in humans(34)—reflecting species differences in urine-concentratingmechanisms.
Our findings of the spatial and histotopographic arrangementof nephron tubules within the medulla provide crucial structuralbasis for elucidating the formation of countercurrent multiplierin mouse kidney and for evaluating species differences in theability of concentrating urine. (1) A tortuous course of theupper part of the DTL of LLN and a few longer SLN, togetherwith a winding course of TAL of SLN and CD were located in theinterbundle regions through the ISOM. These windings increasedthe corresponding segment lengths by approximately 27%. (2)Single thick-walled tubules constantly found in the centralpart of the larger vascular bundles were identified as TAL ofLLN, whereas most of the TAL of LLN were located in the peripheryof the vascular bundles through the outer medulla; the DTL ofSLN were always integrated into the outer or middle layers ofthe vascular bundles. (3) A classification was introduced todescribe three different SLN bends relating to the lengths andthe positions of the tubule segments and the corresponding glomeruli.
The epithelium lining the tortuous part of the tubules of theDTL of LLN was ultrastructurally characterized by a highly specializedapical and basolateral membrane, described as type 2 epithelium(35). This highly differentiated membrane has been confirmedto have a high permeability to sodium, chloride, and water bystudies in rabbits, rats, and hamsters (36). The lateral cellinterdigitations with a high number of single-strand tight junctionsare consistent with the distribution of a subtype of a tight-junctionintegral protein, claudin 2, which is believed to be a componentor channel for the solute paracellular transport (13). The epithelialining the DTL and TAL contribute to different solute compositionsin the interbundle regions exchanging solutes with the vascularbundles because of different permeability properties. Therefore,it is fair to assume that the increase in length as a resultof the tortuous or winding course of the tubules and increasein membrane surface of DTL owing to the specialized membranearrangement correspond to a similarly increased functional capacityin countercurrent multiplier mechanisms.
The TAL of LLN in relation to the vascular bundles through theouter medulla may result from the layer-arranged tubule development(37). However, the finding of the TAL of LLN within the centerof the vascular bundles probably reflects local physiologicdemands in the active transport by TAL. The tubules that integratedinto the vascular bundles, e.g., the DTL of SLN, have been proposedto be involved in generating the osmotic gradient in both thetubule lumen and the medullary interstitium by exchanging importantsolutes with vasa recta (38, 39). Moreover, the identificationof three types of bends gives us a better understanding of theformation of the osmotic gradients in the ISOM. In rats andrabbits, the bends of the SLN are located roughly at the samelevel of the ISOM near the junction to the IM, and the TAL epitheliumcontinues for a short distance in the descending limb (20).This study on mice revealed that the bends of the SLN were situatedat successive levels within the middle and inner half of theISOM. In the inner part of the ISOM, where the vascular bundlesdecreased in size, the length of the prebend TAL segment oftype 3 bends increased with increasing depth of the bends (Figure 2D).This type 3 bend may be assumed to play an important rolein the formation of higher osmolality in the surrounding interstitium,because TAL epithelium is impermeable to water, but transportsdifferent electrolytes (40). Our study also identified a transitionalzone. The representation of the deepest type 3 bends that reachthe innermost stripe and the LLNt (Figure 1) revealed an extracontribution to the higher osmolality by the exclusive TAL epitheliumcovering the loop limbs around the transitional zone.
We believe that these findings in the outer medulla may havefunctional implications in the formation of the countercurrentmultiplier. Therefore, it is important to take the spatial coursesand the epithelial characteristics of the tubules into considerationwhen modeling the urine concentrating mechanism in mice.
In the IM, the formation of the gradient osmolality is alsostructurally associated with the epithelia covering the limbsand bends of the LLN. The type 3 epithelium covering the DTLwas gradually replaced with type 4 epithelium at a distanceof approximately 700 µm before the bends. With increasingdepth of the bends in the IM, more tubules are covered withtype 4 epithelium than with type 3 epithelium. The ATL is assumedto contribute to the higher osmolality in the middle to deepIM by passive NaCl absorption, urea permeability, and waterimpermeability (41, 42). This assumption is corroborated bythe findings from a study in 3D functional reconstruction ofthe IM in rats (43).
The structural-functional relationships along the distal convolutionand CD have been studied intensively morphologically and physiologicallyin different species (14, 24, 44). The DCT started at an averagedistance of 80 µm from the macula densa, compared witha distance of 0 to 500 µm for the nephrons of a varietyof mammalian species (22). The transition from DCT to CNT wasdetermined by the appearance of CNT cells during tracing. However,an agreement on the definition of the transition has not beenreached between the morphologic and physiologic criteria (45).The multiple cell types in the CNT suggest the variety of itsphysiologic functions in fine-tuning transcellular electrolytetransport, as well as hormone-regulated water transport (14).
In conclusion, the presented 3D representation of the mousekidney has provided a detailed characterization of the spatialarrangement and interrelationship of a large number of nephrons,of the tubular-vascular histotopography in the renal medulla,and of the ultrastructure of the tubule epithelia in well-definedsegments. New, important features of mouse kidney microstructurewith potential functional implications have been revealed inaddition to findings that are consistent with our basic knowledgeof the human and animal kidneys. This information is essentialfor the development of accurate mathematical models simulatingurine concentration mechanisms and tubular permeability properties.
Acknowledgments
This work was supported by grants from the Danish Medical ResearchCouncil, the Novo Nordic Foundation, the Danish BiotechnologyProgram, Daloon Foundation, the European Commission (EU FrameworkProgram 6, EuroGene, contract no. 05085), and the Universityof Aarhus Research Foundation.
This work was presented in part at the American Society of Nephrology37th Annual Meeting; St. Louis, MO; October 27 to November 1,2004 (J Am Soc Nephrol 15: A542, 2004).
Merete Fischer is thanked for secretarial assistance.
Footnotes
Published online ahead of print. Publication date availableat www.jasn.org.
Kriz W, Koepsell H: The structural organization of the mouse kidney.
Z Anat Entwicklungsgesch144
:137
–163,1974[CrossRef][Medline]
Kriz W, Schnermann J, Koepsell H: The position of short and long loops of Henle in the rat kidney.
Z Anat Entwicklungsgesch138
:301
–319,1972[CrossRef][Medline]
Bachmann S, Kriz W: Histotopography and ultrastructure of the thin limbs of the loop of Henle in the hamster.
Cell Tissue Res225
:111
–127,1982[CrossRef][Medline]
Kriz W, Bankir L: A standard nomenclature for structure of the kidney. The Renal Commission of the International Union of Physiological Sciences (IUPS).
Anat Embryol (Berl)178
:N1
–N8,1988[CrossRef][Medline]
Dorup J, Maunsbach AB: Three-dimensional organization and segmental ultrastructure of rat proximal tubules.
Exp Nephrol5
:305
–317,1997[Medline]
Dorup J, Morsing P, Rasch R: Tubule-tubule and tubule-arteriole contacts in rat kidney distal nephrons. A morphologic study based on computer-assisted three-dimensional reconstructions.
Lab Invest67
:761
–769,1992
Pannabecker TL, Dahlmann A, Brokl OH, Dantzler WH: Mixed descending- and ascending-type thin limbs of Henle’s loop in mammalian renal inner medulla.
Am J Physiol Renal Physiol278
:F202
–F208,2000[Abstract/Free Full Text]
Layton AT, Pannabecker TL, Dantzler WH, Layton HE: Two modes for concentrating urine in rat inner medulla.
Am J Physiol Renal Physiol287
:F816
–F839,2004[Abstract/Free Full Text]
Christensen EI, Birn H: Megalin and cubilin: Multifunctional endocytic receptors.
Nat Rev Mol Cell Biol3
:256
–266,2002[Medline]
Chou CL, Knepper MA, Hoek AN, Brown D, Yang B, Ma T, Verkman AS: Reduced water permeability and altered ultrastructure in thin descending limb of Henle in aquaporin-1 null mice.
J Clin Invest103
:491
–496,1999[Medline]
Li WY, Huey CL, Yu AS: Expression of claudin-7 and -8 along the mouse nephron.
Am J Physiol Renal Physiol286
:F1063
–F1071,2004[Abstract/Free Full Text]
Kiuchi-Saishin Y, Gotoh S, Furuse M, Takasuga A, Tano Y, Tsukita S: Differential expression patterns of claudins, tight junction membrane proteins, in mouse nephron segments.
J Am Soc Nephrol13
:875
–886,2002[Abstract/Free Full Text]
Enck AH, Berger UV, Yu AS: Claudin-2 is selectively expressed in proximal nephron in mouse kidney.
Am J Physiol Renal Physiol281
:F966
–F974,2001[Abstract/Free Full Text]
Loffing J, Loffing-Cueni D, Valderrabano V, Klausli L, Hebert SC, Rossier BC, Hoenderop JG, Bindels RJ, Kaissling B: Distribution of transcellular calcium and sodium transport pathways along mouse distal nephron.
Am J Physiol Renal Physiol281
:F1021
–F1027,2001[Abstract/Free Full Text]
Meneton P, Ichikawa I, Inagami T, Schnermann J: Renal physiology of the mouse.
Am J Physiol Renal Physiol278
:F339
–F351,2000[Abstract/Free Full Text]
Birn H, Fyfe JC, Jacobsen C, Mounier F, Verroust PJ, Orskov H, Willnow TE, Moestrup SK, Christensen EI: Cubilin is an albumin binding protein important for renal tubular albumin reabsorption.
J Clin Invest105
:1353
–1361,2000[Medline]
Zhai XY, Birn H, Jensen KB, Thomsen JS, Andreasen A, Christensen EI: Digital three-dimensional reconstruction and ultrastructure of the mouse proximal tubule.
J Am Soc Nephrol14
:611
–619,2003[Abstract/Free Full Text]
Andreasen A, Drewes AM, Assentoft JE, Larsen NE: Computer-assisted alignment of standard serial sections without use of artificial landmarks. A practical approach to the utilization of incomplete information in 3-D reconstruction of the hippocampal region.
J Neurosci Methods45
:199
–207,1992[CrossRef][Medline]
Andreasen A, Ren H: Extending the resolution of light microscopy and electron microscopy digitized images with reference to cellular changes after in vivo low oxygen exposure.
J Neurosci Methods122
:157
–170,2003[CrossRef][Medline]
Kaissling B, Kriz W: Morphology of the loop of Henle, distal tubule, and collecting duct. In:
Handbook of Physiology. Renal Physiology, edited by Windhager EE, New York, Oxford University Press,1992
, pp109
–168
Kaissling B, Peter S, Kriz W: The transition of the thick ascending limb of Henle’s loop into the distal convoluted tubule in the nephron of the rat kidney.
Cell Tissue Res182
:111
–118,1977[Medline]
Kaissling B, Kriz W: Structural analysis of the rabbit kidney.
Adv Anat Embryol Cell Biol56
:1
–123,1979[Medline]
Takahashi-Iwanaga H, Iwata Y, Adachi K, Fujita T: The histotopography and ultrastructure of the thin limb of the Henle’s loop: A scanning electron microscopic study of the rat kidney.
Arch Histol Cytol52
:395
–405,1989[Medline]
Madsen KM, Tisher CC: Structural-functional relationships along the distal nephron.
Am J Physiol250
:F1
–F15,1986
Kim J, Kim YH, Cha JH, Tisher CC, Madsen KM: Intercalated cell subtypes in connecting tubule and cortical collecting duct of rat and mouse.
J Am Soc Nephrol10
:1
–12,1999[Free Full Text]
Beeuwkes R 3rd: Efferent vascular patterns and early vascular-tubular relations in the dog kidney.
Am J Physiol221
:1361
–1374,1971[Free Full Text]
Moffat DB, Fourman J: A vascular pattern of the rat kidney. 1963.
J Am Soc Nephrol12
:624
–632,2001[Free Full Text]
Thomson SC, Blantz RC: Ions and signal transduction in the macula densa.
J Clin Invest106
:633
–635,2000[Medline]
Vallon V, Richter K, Huang DY, Rieg T, Schnermann J: Functional consequences at the single-nephron level of the lack of adenosine A1 receptors and tubuloglomerular feedback in mice.
Pflugers Arch448
:214
–221,2004[CrossRef][Medline]
Lorenz JN, Schultheis PJ, Traynor T, Shull GE, Schnermann J: Micropuncture analysis of single-nephron function in NHE3-deficient mice.
Am J Physiol277
:F447
–F453,1999[Medline]
Layton HE, Pitman EB, Moore LC: Spectral properties of the tubuloglomerular feedback system.
Am J Physiol273
:F635
–F649,1997[Medline]
Kriz W: [The architectonic and functional structure of the rat kidney].
Z Zellforsch Mikrosk Anat82
:495
–535,1967[CrossRef][Medline]
Bulger RE, Cronin RE, Dobyan DC: Survey of the morphology of the dog kidney.
Anat Rec194
:41
–65,1979[Medline]
Peter K:
Untersuchungen über Bau und Entwicklung der Niere, Jena, Germany, Fischer,1909
Dieterich HJ, Barrett JM, Kriz W, Bulhoff JP: The ultrastructure of the thin loop limbs of the mouse kidney.
Anat Embryol (Berl)147
:1
–18,1975[Medline]
Imai M: Functional heterogeneity of the descending limbs of Henle’s loop. II. Interspecies differences among rabbits, rats, and hamsters.
Pflugers Arch402
:393
–401,1984[CrossRef][Medline]
Speller AM, Moffat DB: Tubulo-vascular relationships in the developing kidney.
J Anat123
:487
–500,1977[Medline]
Bankir L, Bouby N, Trinh-Trang-Tan MM: Heterogeneity of nephron anatomy.
Kidney Int Suppl20
:S25
–S39,1987[Medline]
Obermuller N, Kunchaparty S, Ellison DH, Bachmann S: Expression of the Na-K-2Cl cotransporter by macula densa and thick ascending limb cells of rat and rabbit nephron.
J Clin Invest98
:635
–640,1996[Medline]
Yang B, Bankir L: Urea and urine concentrating ability: New insights from studies in mice.
Am J Physiol Renal Physiol288
:F881
–F896,2005[Abstract/Free Full Text]
Imai M, Kokko JP: Sodium chloride, urea, and water transport in the thin ascending limb of Henle. Generation of osmotic gradients by passive diffusion of solutes.
J Clin Invest53
:393
–402,1974[Medline]
Pannabecker TL, Abbott DE, Dantzler WH: Three-dimensional functional reconstruction of inner medullary thin limbs of Henle’s loop.
Am J Physiol Renal Physiol286
:F38
–F45,2004[Abstract/Free Full Text]
Loffing J, Kaissling B: Sodium and calcium transport pathways along the mammalian distal nephron: From rabbit to human.
Am J Physiol Renal Physiol284
:F628
–F643,2003[Abstract/Free Full Text]
Kaissling B: Ultrastructural organization of the transition from the distal nephron to the collecting duct in the desert rodent Psammomys obesus.
Cell Tissue Res212
:475
–495,1980[Medline]
This article has been cited by other articles:
T. L. Pannabecker, W. H. Dantzler, H. E. Layton, and A. T. Layton Role of three-dimensional architecture in the urine concentrating mechanism of the rat renal inner medulla
Am J Physiol Renal Physiol,
November 1, 2008;
295(5):
F1271 - F1285.
[Abstract][Full Text][PDF]
E. Riveira-Munoz, O. Devuyst, H. Belge, N. Jeck, L. Strompf, R. Vargas-Poussou, X. Jeunemaitre, A. Blanchard, N. V. Knoers, M. Konrad, et al. Evaluating PVALB as a candidate gene for SLC12A3-negative cases of Gitelman's syndrome
Nephrol. Dial. Transplant.,
October 1, 2008;
23(10):
3120 - 3125.
[Abstract][Full Text][PDF]
X.-Y. Zhai, R. A. Fenton, A. Andreasen, J. S. Thomsen, and E. I. Christensen Aquaporin-1 Is not Expressed in Descending Thin Limbs of Short-Loop Nephrons
J. Am. Soc. Nephrol.,
November 1, 2007;
18(11):
2937 - 2944.
[Abstract][Full Text][PDF]
T. L. Pallone Aquaporin 1, Urea Transporters, and Renal Vascular Bundles
J. Am. Soc. Nephrol.,
November 1, 2007;
18(11):
2798 - 2800.
[Full Text][PDF]
R. A. Fenton and M. A. Knepper Urea and Renal Function in the 21st Century: Insights from Knockout Mice
J. Am. Soc. Nephrol.,
March 1, 2007;
18(3):
679 - 688.
[Abstract][Full Text][PDF]
Top
Abstract
Introduction
