Three-dimensional functional reconstruction of inner medullary thin limbs of Henle‘s loop

【摘要】  Digital three-dimensional (3-D) functional reconstructions of inner medullary nephrons were performed. Antibodies against aquaporins (AQP)-1 and -2 and the chloride channel ClC-K1 identified descending thin limbs (DTLs), collecting ducts (CDs), and ascending thin limbs (ATLs), respectively, through indirect immunofluorescence. Tubules were labeled in transverse sections and assembled into 3-D arrays, permitting individual tubule or combined surface representations to depths of 3.3 mm to be viewed in an interactive digital model. Surface representations of 75 tubules positioned near the central region of the inner medulla were reconstructed. In most DTL segments that form loops below 1 mm from the inner medullary base, AQP1 expression begins at the base, becomes intermittent for variable lengths, and continues nearly midway to the loop. The terminal DTL segment exhibiting undetectable AQP1 represents nearly 60% of the distance from the medullary base to the tip of the loop. AQP1 expression was entirely undetectable in shorter long-looped DTLs. ClC-K1 is expressed continuously along the terminal portion of all DTLs reconstructed here, beginning with a prebend region 164 µm before the bend in all tubules and continuing through the entire ascent of the ATLs to the base of the inner medulla. CDs express AQP2 continuously and extensive branching patterns are illustrated. 3-D functional reconstruction of inner medullary nephrons is capable of showing axial distribution of membrane proteins in tubules of the inner medulla and can contribute to further development and refinement of models that attempt to elucidate the concentrating mechanism.

【关键词】  threedimensional reconstruction aquaporin ClCK Bcrystallin countercurrent multiplier concentrating mechanism

IN A PREVIOUS study ( 19 ), we described inner medullary mixed-type thin limbs of Henle's loops that consist of segments of descending thin limb (DTL)-type cells interspersed between segments of ascending thin limb (ATL)-type cells in rat, mouse, and rabbit kidneys. Those segments containing cells of DTL-type appearance, as viewed with differential interference-contrast microscopy, express aquaporin-1 (AQP1) and urea transporter 2 (UT-A2) protein or mRNA. Neighboring segments of these same thin limbs that contain cells of ATL-type appearance, as viewed with differential interference-contrast microscopy, do not express AQP1 or UT-A2 protein or mRNA. Furthermore, ATL-type regions express ATL-specific ClC-K1 protein, whereas DTL-type regions do not. These differences in protein or mRNA expression are, of course, also true of the neighboring pure DTLs and ATLs.

The arrangement and functional characteristics of thin limbs of Henle's loop are believed to be critical to the formation of the osmotic gradient within the inner medulla ( 7 ). The functional role of the mixed-type thin limb is presently not known. However, they apparently make up a large percentage of the inner medullary thin limbs ( 35% in the rat kidney) ( 19 ), and their architectural arrangement should be critical to their functional roles.

To understand the functional implications of the three-dimensional (3-D) geometric architectural arrangement of these mixed-type thin limbs within the inner medulla, we found it necessary to develop a method that would enable us to precisely map the physical positions of these segments in 3-D space. We also wanted to map the positions of other inner medullary structures with which the mixed-type thin limbs may carry out functional interactions. In addition to positional mapping, another criterion for this method included the ability to view the structures in an interactive manner, independently and collectively. This would permit us to conveniently obtain quantitative spatial information regarding distances between structures, lengths, and surface areas of structures, volumes occupied by structures, as well as other parameters.

In this paper, we describe a technique for generating digital 3-D reconstructions of identified renal tubules and present examples of the types of information that these reconstructions can provide. By combining immunocytochemistry and semiautomated image-acquisition techniques with graphical, volumetric modeling software, we compiled multiple, serial tissue sections into 3-D surface and volumetric representations of inner medullary thin limbs of Henle's loops and collecting ducts (CDs). Specific antibodies against the well-defined water channels AQP1 and AQP2 and the chloride channel ClC-K enabled us to identify DTLs, CDs, and ATLs, respectively. Similarly, specific antibodies against the heat shock-related protein B-crystallin enabled us to identify portions of tubules not labeled by the other antibodies.

Although our initial partial reconstructions do not show the architectural relationships of the mixed-type thin limbs (which appear to lie outside our initial reconstruction zone), they do demonstrate, for example, the structural organization of long-looped pure thin limbs, as well as the qualitative expression of functional proteins along the entire corticopapillary axis of these individual nephron segments. Present and future structural information obtained with this technique should aid in quantitative analyses of the function of inner medullary tubular and vascular structures.

METHODS

Animals. Young male Munich-Wistar rats (average weight: 90 g) were purchased from Harlan Sprague Dawley (Indianapolis, IN). The animals were anesthetized with pentobarbital sodium (0.2 ml/100 g body wt).

Tissue preparation. Four kidneys from four male Munich-Wistar rats were prepared for immunocytochemistry by retrograde perfusion through the aorta with PBS (pH 7.4) for 5 min, followed by periodatelysine-paraformaldehyde (0.01 M, 0.075 M, 2%) ( 13 ) in PBS (pH 7.4) for 5 min before removal from the animal. The boundary between the inner medulla and outer medulla was identified on the basis of structural criteria ( 9 ). The whole medulla was dissected free and then immersed in fixative for 3 h at 4°C, washed in PBS, and dehydrated through an ethanol series. Each medulla was trimmed with a razor blade such that the dimensions of the outermost medullary face measured 2.0 x 1.4 mm. This tissue was immersed in a solution of Spurr epoxy resin (Ted Pella) and ethanol (1:1) for 16 h (room temperature), then in 100% Spurr for 48 h (4°C), and finally embedded in 100% Spurr (24 h at 60°C). Serial 1-µm transverse sections were cut exhaustively from each medulla, beginning near the base of the inner medulla and continuing in a papillary direction. Every fifth section was placed onto a glass microscope slide for immunocytochemistry (4 sections/slide).

Immunocytochemistry. Immunocytochemistry was conducted using affinity-purified polyclonal antibodies against the COOH-terminal regions of the human water channel AQP1 (diluted 1:200, raised in chicken, provided by John Regan and W. Daniel Stamer, University of Arizona); the rat kidney-specific chloride channel (ClC-K, diluted 1:200, raised in rabbit, Chemicon); the human water channel AQP2 (diluted 1:200, raised in goat, Santa Cruz); and a bovine monoclonal antibody raised against purified B-crystallin (diluted 1:50, raised in mouse, Stressgen). AQP1 is generally expressed in cells of DTLs, and AQP2 is expressed in the principal cells of CDs. Therefore, AQP1 and AQP2 antibodies can serve as markers for DTLs and CDs, respectively. The ClC-K antibody used in these studies binds to epitopes of both ClC-K1 and ClC-K2; however, there is no evidence that ClC-K2 is expressed in the inner medullary thin limbs. Therefore, in the inner medulla this antibody serves to identify ClC-K1. Because ClC-K1 is expressed only in ATLs in the inner medulla ( 23, 24 ), this antibody serves as a marker for these tubules. B-crystallin is expressed in DTLs, ATLs, CDs, and vasa recta (Ref. 5; Fig. 1 ). Therefore, it serves as a common marker for all these tubule and vascular elements.

Fig. 1. Serial section of renal inner medulla showing immunofluorescence localization of aquaporin 1 (AQP1; orange), AQP2 (aqua), ClC-K1 (purple), and B-crystallin (green) in tubule segments from rat inner medulla. Scale bar = 25 µm.

Before antibody application, Spurr resin was etched by applying to each slide 300 µl of a solution of 5 g NaOH, 5 ml 100% ethanol, and 5 ml propylene oxide for 3 min ( 12 ), followed by extensive washing, first with ethanol and then with distilled water. Sections were then treated with 0.2% Triton X-100 (Sigma) in PBS (PBS/Triton) for 2 min and 1% SDS in PBS for 5 min. They then underwent three 5-min PBS/Triton washes. After this, they were treated for 10 min with a blocking solution consisting of 5% BSA, 1% normal donkey serum (Jackson ImmunoResearch), and 0.2% Triton X-100 diluted into PBS. Primary antibodies diluted into the blocking solution were then applied simultaneously for 2 h at room temperature followed by three 5-min PBS/Triton washes. Next, FITC-, TRITC-, CY5-, and DAPI-conjugated donkey immunoglobulins (Jackson ImmunoResearch, diluted 1:200 or 1:100 in PBS/Triton) were applied simultaneously for 60 min at room temperature ending with three 5-min washes with PBS/Triton. Sections were mounted with Dako fluorescent mounting medium (Carpinteria, CA) and were viewed with epifluorescence microscopy.

Image analysis. We created one montage image (3,070 x 4,093 pixels in grayscale) from each of four emission wavelengths from each tissue section, depicting expression of AQP1, AQP2, CLC-K, and B-crystallin. This was accomplished with an Olympus IX70 microscope and x 10 objective, with a cooled CCD camera, fiberoptic coupled mercury arc lamp, high-speed filter changers, electronic shutters, and a high-precision x - y - z positioning system coupled to a Silicon Graphics workstation (DeltaVision, Applied Precision). Image intensity and contrast were adjusted visually on the monitor to optimize the fluorescence signal-to-noise ratio, and images were cropped to 900 x 1,600 pixels with Photoshop 6.0 (Adobe). This area forms the central region of the inner medulla and measures 608 x 1,081 µm and extends to a depth of 3.3 mm from the base of the inner medulla. Serial image alignments, tubule segmentation, and 3-D surface representations were produced with Amira 2.3 visualization and volume modeling software (Indeed-Visual Concepts, Berlin-Dahlem).

Four separate stacks of digitized, serial images were generated by capturing AQP1, AQP2, ClC-K1, and B-crystallin immunofluorescence in serial tissue sections. These four stacks of images enabled us to delineate the entire length of long thin limbs of Henle, beginning from near their entry into the inner medulla and extending toward the tip of the papilla. They also enabled us to delineate the CDs throughout the inner medulla via AQP2. Each of the images obtained from one single channel was positioned manually, using Amira software, into serial alignment (or registration) with the section immediately above it, beginning with the outermost section and continuing in a papillary direction through the entire stack. Rotational coordinates for computing alignments for images of one single channel were then applied to the three remaining sets of images, thereby producing identical alignments for each of the four sets of images.

Image segmentation, a graphics process that separates the exterior surface boundary of the structure of interest, a tubule in this case, from the exterior, nonsurface space, enabled us to depict a continuous surface and volume representation for each tubule. For the 3-D reconstructions illustrated here, we segmented each tubule in cross section by approximating the exterior surface boundary with a circle. A graphical icon was then constructed with its length equivalent to 5 µm in the z -axis. The icons from all sections were constructed and combined by way of batch processing to create images of entire tubules. In the reconstructed 3-D surface representations, the tubule dimensions, positions, and lengths are drawn to scale in the x -, y -, and z -axes. The segmentation procedure produced four sets of digital reconstructions, each set serving as representations of either DTLs, ATLs, CDs, or entire inner medullary thin limbs. These four sets of reconstructions were then superimposed on each other to produce a single image to illustrate expression of AQP1 and ClC-K1 along the entire inner medullary axis of each thin limb and AQP2 along the entire inner medullary axis of each CD. Tubule segments labeled by B-crystallin are depicted with very low opacity, which permits either AQP or ClC-K label to dominate. Tubule segments labeled solely by B-crystallin serve to delineate thin limb segments that do not express either AQP1 or ClC-K1.

RESULTS

Identification of DTL, ATL, and CD segments in transverse sections of the inner medulla with immunocytochemistry. Because AQP1 and ClC-K1 proteins are expressed only in the DTLs and ATLs of inner medullary nephrons, respectively, immunocytochemical detection of these proteins was used initially to establish the identity of each thin limb segment type in serial sections ( Fig. 1 ). These proteins also carry out at least one of the functions characteristic of the segments in which they are expressed. Initially, we believed that we could identify the entire length of the DTLs and ATLs within the inner medulla and connect each DTL and ATL portion of a given loop on the basis of the patterns of immunofluorescence reflecting AQP1 and ClC-K1. However, we found that every DTL stopped expressing AQP1 (and did not express ClC-K1) before its connection with its corresponding ATL. In contrast, ATLs were completely identified through the prebend region by the expression of ClC-K1. Therefore, an additional probe was required to delineate the rest of the DTL and the continuous pathway of each thin limb through the inner medulla. The heat shock-related protein B-crystallin had previously been shown to reside in DTLs, ATLs, and CDs ( 5 ), and, therefore, we chose the antibody to this protein. By observing expression of proteins immunoreactive to the B-crystallin antibody through serial sections of entire inner medullary thin limbs, we found that this protein is expressed in cells along the entire length of both the descending and ascending thin limbs of Henle's loops. Therefore, we used this antibody to delineate segments that do not express either AQP1 or ClC-K1 and to complete the individual loops. Expression patterns for 75 complete inner medullary loops are shown in Fig. 2. Nine tubules that were spatially well separated were selected from these 75 and are shown in Fig. 3 to provide a clearer view of the general architecture of the loops. Visualization of tubule 3-D separation and quantitation of structural features can be effectively enhanced by rotating the tubules in the horizontal or vertical axes.

Fig. 2. Three-dimensional (3-D) representation of 75 rat inner medullary thin limbs of Henle's loops showing protein expression patterns of AQP1 (red) and ClC-K1 (green). White translucent segments express undetectable levels of AQP1 or ClC-K1 and were determined by immunofluorescence of B-crystallin. Thin limbs are oriented in a corticopapillary direction, with the upper edge of the image near the base of the inner medulla. Scale bar = 500 µm.

Fig. 3. Three-D representation of 9 rat inner medullary thin limbs of Henle's loops showing protein expression patterns of AQP1 (red) and ClC-K1 (green). White translucent segments express undetectable levels of AQP1 or ClC-K1 and were determined by immunofluorescence of B-crystallin. Thin limbs are oriented in a corticopapillary direction, with the upper edge of the image near the base of the inner medulla. Scale bar = 500 µm.

Distribution of AQP1, ClC-K1, and AQP2 along the entire axial length of inner medullary nephrons. Our choice of plastic media for embedding the tissue, thereby obtaining laterally rigid sections, was critical. This rigidity was sufficient to enable us to position adjacent sections into good registration with each other. This, in turn, was sufficient to resolve, at the very least, the diameter of a single thin limb, or 15 µm, in the final 3-D reconstruction. Reconstruction used just 20% of the entire inner medullary tissue, as we imaged only every fifth 1-µm serial section. This was more than sufficient to reconstruct the 180° bend of each loop. Each of these bends encompasses about three to four sections in the z -dimension of our reconstructions, or 15 to 20 µm. For the most part, the remainder of each thin limb is a relatively straight segment. Similarly, the branching segments of CDs, some of which may bend in the vertical ( z ) and/or horizontal ( x and y ) dimensions, were fully encompassed by this technique. The general patterns of AQP1, CLC-K1, and AQP2 protein expression that we show here were similar in the inner medullary inner core region of all four kidneys that we examined.

For 31 of the loops shown in Fig. 2, both descending and ascending thin limbs could be reconstructed from the inner levels of the inner medulla to the base of the inner medulla. Because of the curvature of tubules in the outer region of the inner medulla, the thin limbs of the remaining 44 tubules arched beyond the perimeter of the inner core of our initial reconstruction before reaching the base of the inner medulla. The distance from the base of the inner medulla to the hairpin bends of these 31 loops (the z -dimension) ranged from 660 to 3,220 µm.

In long-looped DTLs that form their bend below the upper third of the inner medulla, AQP1 expression begins at the base of the inner medulla and continues toward the papillary tip, becoming intermittent for a variable length and then becoming undetectable near midlength of the loop ( Fig. 4 ). For these tubules, AQP1 expression is undetectable for the lower 60% of the DTL (60.1 ± 12.3, means ± SE, n = 26). The length of the AQP1-expressing segment is proportional to the depth to which the tubule descends below the base of the inner medulla before forming a bend. Longer tubules express AQP1 for a proportionately longer length. Expression of AQP1 was entirely undetectable in some shorter long-looped DTLs, forming a bend as far as 1,350 µm from the base of the medulla. Other tubules, not included in this initial portion of our inner medullary reconstruction, descend below 3.3 mm from the base of the inner medulla. Some of these do express AQP1 at deeper levels of the inner medulla than depicted here; however, these tubules account for a small proportion of the total number of inner medullary thin limbs. It should be noted that AQP1 appears to be present in both the apical and basolateral membranes throughout its range of expression.

Fig. 4. Three-D representation of AQP1 (red) and ClC-K1 (green) distribution in rat inner medullary long-looped thin limbs of Henle's loops from the inner medulla. AQP1 is expressed continuously in upper regions, intermittently in middle regions, and is undetectable in the lowest regions (white). Right : AQP1 immunofluorescence in tissue sections at 845 ( A ); 1,295 ( B ); 1,335 ( C ); and 1,410 µm ( D ) below the base of the inner medulla. Left : these sections correspond to the levels identified by black arrows. Right, A, B, D : white arrowheads identify AQP1 immunofluorescence associated with the specific descending thin limb (DTL; shown in red) that is positioned immediately to the right of the black arrows at left. Right, C : white arrowhead identifies the portion of this same DTL that does not express AQP1 (positioned immediately to the right of the arrow at left ). The portion of the DTL not expressing AQP1 has been identified by the antibody to B-crystallin (fluorescence not shown at right ). Thin limbs are oriented in a corticopapillary direction, with the upper edge of the image near the base of the inner medulla. Left : scale bar = 200 µm; right : scale bars = 100 µm.

ClC-K1 is expressed in all ATLs, beginning 164 µm before the bend. The length of this prebend region was similar for all tubules (163.71 ± 7.82, means ± SE, range 85 to 250 µm, n = 31), regardless of the distance from the base of the medulla to the hairpin bend of each tubule. This pattern of ClC-K1 expression was also seen in shorter long-looped DTLs not expressing AQP1. ClC-K1 expression continues in all inner medullary tubules examined so far, through and beyond the hairpin bend to the base of the inner medulla ( Fig. 5 ). The immunocytochemical fluorescence signal-to-background ratio appears to remain qualitatively unchanged along the entire length of the ATL. Like AQP1, ClC-K1 appears to be present in both the apical and basolateral membranes throughout its range of expression.

Fig. 5. Three-D representation of ClC-K1 distribution in rat inner medullary long-looped thin limbs of Henle's loops. ClC-K1 (green) is expressed in a prebend region (beginning at white arrows) before looping 180° and continuing as the ascending limb toward the base of the inner medulla. Right : ClC-K1 immunofluorescence in tissue sections at 2,615 ( A ) and 2,720 ( B ) µm below the base of the inner medulla. Left : these sections correspond to the levels identified by black arrows. Right, B : white arrowhead identifies ClC-K1 immunofluorescence associated with the DTL (prebend region) that is positioned immediately to the right of the black arrow at left. Right, A : white arrowhead identifies section of the DTL positioned immediately to the right of the black arrow at left that does not express ClC-K1. This section of the DTL was identified by the antibody to B-crystallin (fluorescence not shown at right ). Thin limbs are oriented in a corticopapillary direction, with the base of the inner medulla above the upper edge of the image. Left : scale bar = 300 µm; right : scale bars = 50 µm.

AQP2 is expressed in the apical membrane of all principal cells of CDs ( Fig. 1 ) throughout the regions we examined. Therefore, the entire inner medullary collecting duct network can be followed with this antibody ( Fig. 6 ). The 26 CDs that are positioned near the base of the inner medulla, shown in Fig. 6, coalesce into three CDs at a level of 3.3 mm below the base of the inner medulla.

Fig. 6. Three-D representation of inner medullary collecting ducts (CDs). Twenty-six CD branches coalesce into 3 CDs at 3.3 mm below the base of the inner medulla. AQP2 is expressed in principal cells along the entire length of the tubules. Branches are clearly delineated throughout. CDs are oriented in a corticopapillary direction, with the upper edge of the image near the base of the inner medulla. Scale bar = 500 µm.

DISCUSSION

We developed a technique that enables us to produce high-resolution, 3-D surface representations that permit positional mapping of thin limbs of Henle's loops and CDs in the inner medulla. These structures may extend to 4 or 5 mm in length below the base of the inner medulla, although the reconstructions shown here extend to only 3.3 mm. Although we reconstructed a random sample of tubules within the central region of the inner medulla to this depth, it seems reasonable to suggest that the patterns of AQP and ClC-K1 expression that we describe here are representative of all the DTLs, ATLs, and CDs that reach this depth. Preliminary observations suggest that AQP1 expression is increasingly reduced at lower depths, however.

Tubule outer diameters at depths to 3.3 mm range between 15 and 35 µm. Structural features such as the 180° bend of the thin limb of Henle's loop and more complex lateral bends of the CD are clearly resolved with this technique. All structures are reconstructed to scale, and volume-modeling software permits measurements of distances between any two points within the 3-D volume. Individual or selected groups of tubules can be viewed interactively, zooming in or out, with full rotational capabilities. The number of classes of structures incorporated into the model is limited primarily by the adequacy of immunocytochemistry to label different structures on biological tissue sections. For the models shown here, we used 20% of the tissue, leaving additional sections unused that can be labeled with antibodies identifying other structures, such as vasa recta. This can be done at a later time via appropriate immunocytochemistry and, after carrying out appropriate alignments, we can readily position structural elements into the existing model. Moreover, the immunolabeling of additional proteins of various functions will permit the exact 3-D positioning of these other functional elements within the already reconstructed tubular and vascular structures.

Our reconstructions illustrate functional aspects of each nephron qualitatively, indicating either presence or absence of AQP1 or ClC-K, although visibly detectable and significant differences in labeling intensity appear at different depths. Most notable, and shown here, is the gradual disappearance of AQP1 label from DTLs near midlength and its complete absence from terminal segments. At the outset of these studies, we observed a greater abundance of ClC-K1-expressing thin limbs compared with AQP1-expressing thin limbs at most levels of the rat inner medulla, and others had reported this as well ( 14 ). Because thin limbs do not form branches and for each DTL there is a single ATL, this could only be explained by the absence of AQP1 expression in some DTLs and/or presence of ClC-K1 in some DTLs. As shown here, the absence of AQP1 in some DTL segments accounts for at least part of this disparity in the inner medulla.

The rat and mouse thin limb epithelium exhibits structural heterogeneity, with three segment types having been distinguished for the DTL and one type for the ATL ( 4 ). These segments differ regarding cell junction complexes, organelle abundance, and other structural features. These observations imply that functional differences exist for individual DTLs at different levels along the corticopapillary axis. Although it has generally been assumed that AQP1 protein is continuously expressed along the length of the DTL ( 14 ), our recent study indicates that discontinuous expression occurs in some segments of DTLs in rat, rabbit, and mouse kidneys ( 19 ). Several earlier papers reported evidence for lower water permeability in segments from deeper levels of the inner medulla. Chou and Knepper ( 2 ) found that in the chinchilla, the degree of water permeability declined significantly in segments near the papilla tip compared with segments from the outer region of the inner medulla, and a range of permeabilities was noted at all levels of the chinchilla inner medulla, including the outer region. In the rat, at least, low water permeability in segments from lower levels of the inner medulla apparently is due to the absence of expression of AQP1.

Ultrastructural studies previously suggested that a significant reduction of water channel abundance may occur at deeper levels of the inner medullary thin limbs. Freeze-fracture electron microscopy studies showed that intramembrane particles (IMPs) in inner medullary long-looped DTLs consist predominantly of AQP1 water channels ( 27 ). The density of IMPs in DTLs extending through the outer medulla and into the initial third of the inner medulla is quantitatively greater than the density of IMPs in DTLs from the lower two-thirds of the inner medulla ( 21 ). The density of IMPs in DTLs from the lower two-thirds of inner medullary DTLs is quantitatively similar to the density of inner medullary ATL IMPs. In the ATL, IMPs likely reflect expression of proteins unrelated to AQP1 ( 27 ). In view of our initial results, it seems likely that many of the remaining IMPs in DTLs from the lower two-thirds of the inner medulla ( 21 ) also reflect proteins unrelated to AQP1.

A prebend segment has been noted in the rat inner medulla and has been described as making an abrupt transition from the lower DTL epithelium type to the ATL epithelium ( 21 ). Lengths of this prebend segment ranging from 50 to 140 µm (or longer) were reported. Mejia and Wade ( 14 ) reported immunolabeling by ClC-K1 antibody in "turning" thin limbs, suggesting that ClC-K1 is expressed along this prebend region. On the basis of the nephrons that we examined, it appears likely that thin limbs always express ClC-K1 along this prebend segment. The abrupt transition in epithelial structure therefore corresponds to an abrupt onset of ClC-K1 expression.

The existence of a large number of inner medullary DTL segments that express little or no AQP1 would likely have significant impact on models describing the concentrating mechanism ( 8, 22 ). At least seven AQP isoforms are known to be present in the kidney, and only AQP1 is localized to the thin limbs ( 17 ). AQP1 is believed to be the principal transepithelial fluid pathway in DTLs, permitting osmotic equilibration of tubular fluid with the interstitium as the DTL descends the medullary axis ( 17 ). The existence of other, as yet uncharacterized, transporters capable of providing a pathway for water flow across the epithelial cell plasma membrane within the DTL cannot be ruled out. The existence of minimal water permeability at terminal ends of inner medullary DTLs would imply that within these regions luminal fluid in DTLs does not equilibrate with interstitial fluid through water reabsorption. The discontinuous expression of AQP1 underscores the importance of determining permeability properties along the AQP1-null segments of individual inner medullary thin limbs.

Transfer of fluid and solutes between different segments of the nephron requires a distinctive architecture for optimal function of urinary concentration and dilution. Current models propose that neighboring structures, expressing different or similar water and solute permeabilities, are critical to formation of the medullary gradient ( 8, 22 ). To understand the interactions of multiple membrane transport proteins and their contributions to development and maintenance of the concentrating and diluting mechanisms in the renal inner medulla, it is important to clearly delineate the expression of relevant proteins along the various nephron segments. Conceptual and mathematical models of the mammalian concentrating mechanism and fluid and solute transport and transport-mediated compartmentation, and their dynamics, require quantitative details of transport function ( 7, 11, 28 ). As the complexity of the transport functions associated with the concentrating mechanism becomes better understood, the 3-D geometric architecture of thin limbs of Henle's loops, CDs, and vasa recta and their membrane permeabilities and transport properties will likely become increasingly important in modeling these mechanisms. Three-D reconstructions of the inner medullary tubular and vascular elements using techniques such as those described here provide an effective method for providing these quantitative details.

ACKNOWLEDGMENTS

We thank Drs. J. Regan and D. Stamer of the University of Arizona for providing AQP1 antibodies and M. Landis of the University of Arizona Center for Computing and Information Technology for assistance with image analyses.

GRANTS

This study was supported in part by National Institutes of Health Grants DK-16294 and ES-06694 (Southwest Environmental Health Sciences Center).

【参考文献】
  Brokl OH and Dantzler WH. Amino acid fluxes in rat thin limb segments of Henle's loop during in vitro microperfusion. Am J Physiol Renal Physiol 277: F204-F210, 1999.

Chou CL and Knepper MA. In vitro perfusion of chinchilla thin limb segments: segmentation and osmotic water permeability. Am J Physiol Renal Fluid Electrolyte Physiol 263: F417-F426, 1992.

Chou CL, Nielsen S, and Knepper MA. Structural-functional correlation in chinchilla long loop of Henle thin limbs: a novel papillary subsegment. Am J Physiol Renal Fluid Electrolyte Physiol 265: F863-F874, 1993.

Dieterich HJ, Barrett JM, Kriz W, and Bulhoff JP. The ultrastructure of the thin loop limbs of the mouse kidney. Anat Embryol (Berl) 147: 1-18, 1975.

Iwaki T, Kume-Iwaki A, and Goldman JE. Cellular distribution of B-crystallin in non-lenticular tissues. J Histochem Cytochem 38: 31-39, 1990.

Kaissling B and Kriz W. Morphology of the loop of Henle, distal tubule, and collecting duct. In: Handbook of Physiology. Renal Physiology. Bethesda, MD: Am Physiol Soc, 1992, sect. 8, vol. I, chapt. 3, p. 109-168.

Knepper MA, Saidel GM, Hascall VC, and Dwyer T. Concentration of solutes in the renal inner medulla: interstitial hyaluronan as a mechanoosmotic transducer. Am J Physiol Renal Physiol 284: F433-F446, 2003.

Kokko JP and Rector FC. Countercurrent multiplication system without active transport in inner medulla. Kidney Int 2: 214-223, 1972.

Kriz W and Bankir L. A standard nomenclature for structures of the kidney. Kidney Int 33: 1-7, 1988.

Kriz W and Kaissling B. Structural organization of the mammalian kidney. In: The Kidney: Physiology and Pathophysiology, edited by Seldin DW and Giebisch G. New York: Raven, 1992.

Layton HE, Knepper MA, and Chou CL. Permeability criteria for effective function of passive countercurrent multiplier. Am J Physiol Renal Fluid Electrolyte Physiol 270: F9-F20, 1996.

Maxwell MH. Two rapid and simple methods used for the removal of resins from 1.0 µm thick epoxy sections. J Microsc 112: 253-255, 1977.

McLean IW and Nakane PK. Periodate-lysine-paraformaldehyde fixative a new fixative for immunoelectron microscopy. J Histochem Cytochem 22: 1077-1083, 1974.

Mejia R and Wade J. Immunomorphometric study of rat renal inner medulla. Am J Physiol Renal Physiol 282: F553-F557, 2002.

Moon C, Preston GM, Griffin CA, Jabs EW, and Agre P. The human aquaporin CHIP gene: structure, organization, and chromosomal localization. J Biol Chem 268: 15772-15778, 1993.

Nielsen S and Agre P. The aquaporin family of water channels in kidney. Kidney Int 48: 1057-1068, 1995.

Nielsen S, Frokiaer J, Marples D, Kwon TH, Agre P, and Knepper MA. Aquaporins in the kidney: from molecules to medicine. Physiol Rev 82: 205-244, 2001.

Nielsen S, Pallone T, Smith BL, Christensen EI, Agre P, and Maunsbach AB. Aquaporin-1 water channels in short and long loop descending thin limbs and in descending vasa recta in rat kidney. Am J Physiol Renal Fluid Electrolyte Physiol 268: F1023-F1037, 1995.

Pannabecker TL, Dahlmann A, Brokl OH, and Dantzler WH. Mixed descending- and ascending-type thin limbs of Henle's loop in mammalian renal inner medulla. Am J Physiol Renal Physiol 278: F202-F208, 2000.

Sands JM and Kokko JP. Current concepts of the countercurrent multiplication system. Kidney Int 50, Suppl 57: S93-S99, 1996.

Schwartz MM and Venkatachalam MA. Structural differences in thin limbs of Henle: physiological implications. Kidney Int 6: 193-208, 1974.

Stephenson JL. Concentration of urine in a central core model of the renal counterflow system. Kidney Int 2: 85-94, 1972.

Uchida S. In vivo role of CLC chloride channels in the kidney. Am J Physiol Renal Physiol 279: F802-F808, 2000.

Uchida S, Sasaki S, Furukawa T, Hiraoka M, Imai T, Hirata Y, and Marumo F. Molecular cloning of a chloride channel that is regulated by dehydration and expressed predominantly in kidney medulla. J Biol Chem 268: 3821-3824, 1999.

Uchida S, Sasaki S, Nitta K, Uchida K, Horita S, Nihei H, and Marumo F. Localization and functional characterization of rat kidney-specific chloride channel, ClC-K1. J Clin Invest 95: 104-113, 1995.

Vandewalle A, Cluzeaud F, Bens M, Kieferle S, Steinmeyer K, and Jentsch TJ. Localization and induction by dehydration of ClC-K chloride channels in the rat kidney. Am J Physiol Renal Physiol 272: F678-F688, 1997.

Verbavatz JM, Brown D, Sabolic I, Valenti G, Ausiello DA, Van Hoek A, Ma T, and Verkman AS. Tetrameric assembly of CHIP28 water channels in liposomes and cell membranes: a freeze-fracture study. J Cell Biol 123: 605-618, 1993.

Wexler AS, Kalaba RE, and Marsh DJ. Three-dimensional anatomy and renal concentrating mechanism. I. Modeling results. Am J Physiol Renal Fluid Electrolyte Physiol 260: F368-F383, 1991.

作者单位：Department of Physiology, College of Medicine, University of Arizona, Tucson, Arizona 85724-5051