Am J Physiol Renal Physiol 294: F1306-F1314, 2008.
First published April 16, 2008; doi:10.1152/ajprenal.00068.2008
0363-6127/08 $8.00
Quantitative analysis of functional reconstructions reveals lateral and axial zonation in the renal inner medulla
Thomas L. Pannabecker,
Cory S. Henderson, and
William H. Dantzler
Department of Physiology, University of Arizona Health Sciences Center, Tucson, Arizona
Submitted 7 February 2008
; accepted in final form 11 April 2008
 |
ABSTRACT
|
|---|
Three-dimensional functional reconstructions of descending thin limbs (DTLs) and ascending thin limbs (ATLs) of loops of Henle, descending vasa recta (DVR), ascending vasa recta (AVR), and collecting ducts (CDs) permit quantitative definition of lateral and axial zones of probable functional significance in rat inner medulla (IM). CD clusters form the organizing motif for loops of Henle and vasa recta in the initial 3.0–3.5 mm of the IM. Using Euclidean distance mapping, we defined the lateral boundary of each cluster by pixels lying maximally distant from any CD. DTLs and DVR lie almost precisely on this independently defined boundary, placing them in the intercluster interstitium maximally distant from any CD. ATLs and AVR lie in a nearly uniform pattern throughout intercluster and intracluster regions, which we further differentiated by a polygon around CDs in each cluster. Loops associated with individual CD clusters show a similar axial exponential decrease as all loops together in the IM. Because 
3.0–3.5 mm below the IM base CD clusters cease to form the organizing motif, all DTLs lack aquaporin 1 (AQP1), and all vasa recta are fenestrated, we have designated the first 3.0–3.5 mm of the IM the "outer zone" (OZ) and the final 1.5–2.0 mm the "inner zone" (IZ). We further subdivided these into OZ-1, OZ-2, IZ-1, and IZ-2 on the basis of the presence of completely AQP1-null DTLs only in the first 1 mm and on broad transverse loop bends only in the final 0.5 mm. These transverse segments expand surface area for probable NaCl efflux around loop bends from 40% to 140% of CD surface area in the final 100 µm of the papilla.
countercurrent systems; concentrating mechanism; aquaporin; kidney-specific chloride channel; urea transporter B; three-dimensional reconstruction
THE HIGHLY ORGANIZED three-dimensional architecture of the mammalian renal inner medulla (IM) underscores the important role structure likely plays in enabling the urinary concentrating mechanism (UCM) to proceed (13, 15–18). Mathematical models of the UCM provide insights into the functional contributions of spatial orientations of structural elements in the outer medulla (OM) and IM (4, 13, 22–25). These models have met with variable success in explaining the observed urine-concentrating ability, especially in the IM. Additional information about the details of renal architecture may permit modelers to suggest testable mechanisms that come closer to explaining the experimental observations.
Discrete lateral orientation of nephrons, blood vessels, and collecting ducts (CDs) occurs at successive levels along the inner medullary axis, and these arrangements change in clearly defined ways with increasing depth below the IM-OM border (15–18). To better define these inner medullary lateral and axial arrangements, we have performed quantitative analyses of 1) CD cluster borders and 2) loop of Henle and vasa recta distributions and their positional relationships with regard to CD clusters. The results show, first, that loops of Henle and blood vessels intermingle with CD clusters (16, 17) along the inner medullary axis in such a well-organized fashion that two major functional compartments appear to exist laterally across the IM. These two compartments comprise 1) the intracluster interstitium and 2) the intercluster interstitium, defined largely by CD cluster arrangements.
Second, the axial arrangements of nephron segments and blood vessels occur in such a clearly repeating manner that the IM can be readily divided into four distinct axial zones. In contrast to the inner and outer stripes of the OM, the axial zones of the IM are not readily observable with standard histological techniques. These four zones include two outer zones and two inner zones. Zonation drawn from architectural features can be correlated to recognized functional features of nephron and vessel segments for the outermost three of these four zones. To provide a functional corollary for the innermost zone, which exists along the terminal 500 µm of the papilla tip, we propose, on the basis of analyses of three-dimensional reconstructions, that wide-bend loops of Henle markedly impact total delivery of NaCl to this region, thereby performing a critical role in sustaining the solute gradient of the terminal papilla.
|
METHODS
|
|---|
Animals.
Young male Munich-Wistar rats (average wt 90 g; Harlan, Indianapolis, IN) were anesthetized with pentobarbital sodium (0.2 ml/100 g body wt) or with CO2. All experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (Washington, DC: National Academy Press, 1996) and approved by the Institutional Animal Care and Use Committee.
Tissue preparation for immunohistochemistry.
Kidneys were prepared for immunohistochemistry by retrograde perfusion through the aorta with PBS (pH 7.4) for 5 min and then with 0.01 M periodate-0.075 M lysine-2% paraformaldehyde in PBS (pH 7.4) for 5 min before removal from the animal. The whole medulla was dissected free, the OM was discarded, and the IM was immersed in fixative for 3 h at 4°C and washed in PBS. Tissue was dehydrated through an ethanol series and embedded in Spurr's epoxy resin (Pella). Serial transverse sections were cut at a thickness of 1 µm either exhaustively or partially from medullas beginning near the OM-IM border and continuing in a papillary direction or from medullas beginning near the papillary tip and continuing in a cortical direction. To reconstruct uninterrupted segments of nephrons or vessels from the complete kidney, two to five sets of 1-µm serial sections were prepared from epoxy-embedded tissue; in each set, sections were 5 µm apart. Image overlays showing blood vessels and nephrons together were obtained from two sections offset from each other by 2.5 µm. Consecutive sections were placed onto glass microscope slides for immunohistochemistry (4 sections/slide). The boundary between the IM and the OM was identified on the basis of structural criteria (9).
Immunohistochemistry.
Generally, two sets of serial sections (see above) were prepared for each kidney: one set was labeled for nephrons and CDs, and one set was labeled for vasa recta and CDs. Nephron segments were labeled by indirect immunohistochemistry as described previously (15–18) using affinity-purified polyclonal antibodies against the COOH-terminal regions of the human water channel aquaporin 1 (AQP1, diluted 1:200, raised in mouse; Serotec), the rat kidney-specific chloride channel (ClC-K, diluted 1:200, raised in rabbit; Chemicon), the human water channel aquaporin 2 (AQP2, diluted 1:200, raised in goat; Santa Cruz Biotechnology), and a bovine monoclonal antibody raised against purified 
B-crystallin (diluted 1:50, raised in mouse; Stressgen). AQP1 and AQP2 antibodies serve as markers for descending thin limbs (DTLs) and CDs, respectively, the ClC-K antibody serves as a marker for ascending thin limbs (ATLs), and B-crystallin serves as a common marker for all tubules (15–18).
In the second set of sections, descending vasa recta (DVR) were labeled with a polyclonal antibody raised in rabbits against rat urea transporter B (UT-B, diluted 1:200; provided by Jeff Sands, Emory University). Ascending vasa recta (AVR) and capillaries were labeled with a polyclonal antibody raised in chicken against rat PV-1, a plasmalemmal vesicle protein formerly known as gp68 (21) (diluted 1:500; provided by Radu Stan, Dartmouth College). PV-1 is a component of the fenestral diaphragm, but its physiological function is poorly understood. In the rat IM, the AVR and capillaries are fenestrated and believed to have diaphragms. In addition, the terminal portions of UT-B-positive DVR are fenestrated, as is the entire subsequent descending UT-B-negative portion (17, 18), and all are believed to have diaphragms. CDs were labeled for AQP2 as described above.
Secondary antibodies, including Alexa 499-, Alexa 568-, FITC-, tetramethylrhodamine isothiocyanate-, Cy5-, 7-amino-4-methylcoumarin-3-acetic acid-, and 4,6-diamidino-2-phenylindole-conjugated donkey Igs (Invitrogen/Molecular Probes or Jackson ImmunoResearch; diluted 1:200 or 1:100 in PBS-Triton), were applied simultaneously for 60 min at room temperature, and the sections were washed three times for 5 min each with PBS-Triton. Sections were mounted with Dako fluorescent mounting medium (Carpentaria, CA) and viewed with epifluorescence microscopy.
Image analysis.
Separate stacks of digitized, serial images were generated by capture of AQP1, AQP2, ClC-K1, B-crystallin, UT-B, or PV-1 immunofluorescence from each tissue section. Continuous surface and volume representations for each vessel and tubule were constructed as described previously from serial sections 
5 µm apart (15–18) with Amira visualization and volume modeling software (Mercury, Chelmsford, MA). Quantitative analyses were carried out in two- or three-dimensional images using PhotoShop (Adobe) and the Image Processing Toolkit (Reindeer Graphics) or Amira, respectively.
Determining CD cluster boundaries with the Euclidean distance map.
Each pixel in the background is assigned a luminosity value equal to its distance from the nearest edge of any CD; in this case, pixels at a greater distance are assigned a darker value (19). This value encodes the straight-line distance to the nearest point on any CD edge. The Euclidean distance map (EDM) is then skeletonized to produce a boundary around each CD. The process of skeletonization sets pixels to white if they have neighbors that are darker. The boundaries are then linked to encompass individual CD clusters (Fig. 1). Boundaries are represented in images as contour lines that correspond to the linear array of pixels that are most distant from any CD.
View larger version (86K):
[in this window]
[in a new window]
|
Fig. 1. Five primary clusters making up a single secondary collecting duct (CD) cluster that consists of 31 CDs [aquaporin 2 (AQP2), blue] at the base of the inner medulla (IM). Image shows section 400 µm below IM base. CD cluster boundaries (white) were determined by Euclildean distance map (EDM). Polygons (red) delimit intracluster interstitium for each of the 5 primary CD clusters. Scale bar, 100 µm.
|
|
Determining loops of Henle associated with each cluster and measuring distances from loop or vessel to CD cluster boundary.
For quantitation of loops of Henle within individual CD clusters at successive levels along the papillary axis, the following criteria were applied. At the IM base, those ATLs with cross-sectional area within the EDM boundaries of a single CD cluster were assigned to that cluster. Those ATLs lying on EDM boundaries were assigned to the cluster in which the majority of the ATL cross-sectional area lay. Each cluster includes only those DTLs that are contiguous with ATLs that had been assigned to the cluster, unless otherwise stated (see RESULTS). AQP1-positive and AQP1-negative DTL segments are included. In some cases, DTLs will physically reside in a cluster different from that determined by the location of their ATL with respect to the CD cluster boundary (EDM boundary). For quantitation of DVR and AVR within individual CD clusters along the papillary axis, those vessels located within the EDM boundaries of each cluster, at each successive level, were included. Those DVR and AVR within multiple subclusters were assigned to the cluster in which the majority of the cross-sectional area lay.
The minimum, transverse distance of each loop or vessel to the CD cluster boundary was determined for all loops and vessels associated with single clusters from five kidneys. Since CD clusters from four of these kidneys had not been completely reconstructed, the clusters were identified on the basis of individual CD proximity to one another as well as the arrangement of DTLs and DVR that encircled them. We previously showed this to be a satisfactory method for estimation of CD cluster boundaries (16). A single section within 500 µm from the IM base was examined for each kidney. The mean distance for each segment type was determined for each cluster.
The volume shrinkage factor for ethanol dehydration of rat medullary tissue has been reported to be 20% (2), and the linear shrinkage factor has been reported to be 20–25% (3). Thus the linear distances that we report would underestimate the distances measured for fresh tissue by a maximum of 20–25%.
Statistical analyses.
Data combined from three or more samples are reported as means and SD (n is number of replicates). The statistical significance of differences between means for each category of multiple data sets was determined with ANOVA and Tukey's post hoc test (P < 0.05).
|
RESULTS
|
|---|
Definition of CD clusters, their lateral boundaries, and the quantitative lateral relationship of loops of Henle and vasa recta to them.
As we reported previously (16, 17), clusters of CDs form the lateral organizing motif for loops of Henle and vasa recta throughout the first 3.0–3.5 mm of the IM (outer zone of the IM). This section first describes the lateral relationships in more detail and is followed by a description of the axial zonation. DTLs and DVR are arrayed outside these clusters (intercluster interstitium), whereas ATLs and AVR are nearly uniformly distributed outside, as well as inside (intracluster interstitium), the clusters (16, 17). At the OM-IM border, each of these primary CD clusters consists of 6–12 CDs that coalesce into a single CD deep in the outer zone of the IM. About five to six of the terminal CDs from these primary CD clusters also coalesce to form a single CD >3.0–3.5 mm below the base of the IM (in the inner zone of the IM). Therefore, about five to six of the primary clusters seen at the IM base make up a larger secondary CD cluster. The terminal CD from each secondary cluster apparently combines with the terminal CD from one other secondary cluster in the final 1.5–2.0 mm (inner zone of the IM) to form one of the 118 CDs found to be remaining 0.5 mm above the papillary tip in one reconstructed kidney (18). These 118 CDs combine to form 13 ducts of Bellini at the papillary tip (18). The number of CDs at the IM base making up each of the larger secondary clusters varies, but if we assume that number is 31 for all of them (as determined for one; see below), then the total number of CDs at the IM base is 7,316. This estimate in these Munich-Wistar rats is almost identical to the 7,300 estimated for Sprague-Dawley rats by Han et al. (3). This total will vary somewhat, depending on the actual average number of CDs in each secondary cluster at the IM base, but it clearly fits with other types of estimates.
To examine more exactly how loops of Henle and vasa recta relate laterally to each of the primary clusters, we used the EDM technique (19) (see METHODS) to determine boundaries for the clusters. This approach leads to a boundary around each of these clusters, which, by definition, is the maximum distance from any CD in this or any neighboring cluster (Fig. 1). We find that most of the AQP1-positive DTLs and UT-B-positive DVR lie directly on the boundaries that demarcate these primary CD clusters (Fig. 2). The DVR tend to be arranged in their own clusters, the members of which are spaced in nearly equal numbers on either side of the CD cluster boundaries (Fig. 3). These patterns of DTL and DVR distribution occur transversely across the entire IM (Fig. 4).
View larger version (145K):
[in this window]
[in a new window]
|
Fig. 2. Five primary clusters making up a single secondary CD cluster that consists of 31 CDs at the IM base. Image shows section 400 µm below IM base. CD cluster boundaries (white) were determined by EDM. Magenta, descending thin limb (DTL)/aquaporin 1 (AQP1); blue, CD/AQP2; green, ascending thin limb (ATL)/kidney-specific Cl– channel (ClC-K1). Scale bar, 100 µm.
|
|
View larger version (141K):
[in this window]
[in a new window]
|
Fig. 3. Five primary clusters making up a single secondary CD cluster that consists of 31 CDs at the IM base. Image shows section 400 µm below IM base. CD cluster boundaries (white) were determined by EDM. Yellow, descending vasa recta (DVR)/urea transporter B (UT-B); blue, CD/AQP2; green, ATL/ClC-K1. Scale bar, 100 µm.
|
|
View larger version (138K):
[in this window]
[in a new window]
|
Fig. 4. Inner medullary section showing 23 primary CD clusters with boundaries determined by EDM (white outlines). EDM boundaries around each CD are shown in gray. Image is an overlay of 2 images from 2 sections 400 µm below the IM base. Red, DTL/AQP1; blue, CD/AQP2; green, DVR/UT-B. Scale bar, 500 µm.
|
|
We further quantified the relationship of loops of Henle and vasa recta to the CD clusters in transverse sections from five kidneys (see METHODS; Fig. 5). For these sections, the average distance from the center of each DTL and DVR to the cluster boundary is 6.8 and 7.8 µm, respectively, whereas the mean distance from the center of each ATL and AVR to the cluster boundary in these same transverse sections is 14.9 and 17.8 µm, respectively (Fig. 5). Thus the DTLs and DVR lie within the intercluster interstitium at the farthest possible distance from the CDs. These lateral relationships of loops of Henle and vasa recta to CD cluster boundaries continue axially from the OM-IM border through the first 3.0–3.5 mm of the IM (outer zone). Below this level (in the inner zone of the IM), the clusters consist of so few CDs that clusters can be defined only by their origin from a duct of Bellini.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 5. Distance between center point of nephron or vessel and nearest CD cluster boundary. Values are means and SD for nephrons and vessels associated with single CD clusters from 5 kidneys. Significantly different means for each of the data sets are indicated by a common symbol (P < 0.05 by ANOVA with Tukey's post hoc test).
|
|
Axial zones of IM.
As discussed above, the arrangement of CD clusters as a clear organizing motif for the arrangement of loops of Henle and vasa recta continues only through the outer 3.0–3.5 mm of the IM. Also, our previous studies have shown that all DTLs extending into the final 2.0–2.5 mm of the IM lack AQP1 and that all vasa recta in the final 2.0 mm are fenestrated, with DVR differentiated from AVR only by the direction of blood flow (18). Because of these major changes at 3.0–3.5 mm below the IM base, we have designated the first 3.0–3.5 mm of the IM the "outer zone" and the final 1.5–2.0 mm the "inner zone" (Table 1). In addition, our previous work indicated that the DTLs of all loops of Henle with bends within the first 1 mm of the IM lack AQP1 for their entire inner medullary length, whereas the DTLs of loops with bends below this level express AQP1 for the first 40% of their length and lack AQP1 for the remaining 60% of their length (16). Because of this difference, we have designated the first 1 mm of the IM outer zone 1 and the next 2.0–2.5 mm outer zone 2. Finally, about half of the loops of Henle extending into the terminal 0.5 mm of IM, instead of having narrow hairpin bends with only a small transverse segment, are bent so that they have long transverse segments. Because of this distinct difference in loop bends in this terminal 0.5 mm, we have designated the first 1.0–1.5 mm of the inner zone inner zone 1 and the terminal 0.5 mm inner zone 2.
Numerical assessment of axial distribution of loops of Henle in CD clusters.
Two complete secondary CD clusters were reconstructed in their entirety, beginning within 100 µm below the IM base. Each of these two clusters coalesced into a single CD 4.5 mm below the IM base. At the IM base, these secondary clusters consisted of 31 and 41 separate CDs, respectively.
For this analysis, we first identified those ATLs or prebend segments within the spatial domain of each secondary CD cluster as defined by EDM in the first section near the IM base. This yielded a discrete population of identified nephrons that was then examined at descending levels. The numbers of these ATLs, as well as their contiguous AQP1-positive or AQP1-negative DTLs, were counted at descending intervals. Since prebend segments were not included in the DTL count, the sum of AQP1-positive and AQP1-negative DTLs at each interval is always less than the ATL count. Loops were counted at each 5-µm level from near the IM base to the level where the CD segments had coalesced into a single, terminal segment (Fig. 6). Both limbs of these loops were closely associated with a single CD cluster along the entire papillary axis, although some drifted outside the cluster border (as defined by EDM), entering into spatial domains of neighboring clusters intermittently. The data for the secondary cluster with 31 CDs are shown in Fig. 6. A similar loop distribution was seen for the other secondary CD cluster (data not shown).
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 6. Loop of Henle decay rates along the IM axis. All ATLs that are associated with a single CD cluster near the IM base and their contiguous AQP1-positive and AQP1-negative DTLs (which lie near but not necessarily within the same CD cluster as defined by EDM) were counted in each section at 5-µm intervals from the base to the CD cluster terminus. DTLs that lack AQP1 upon entering the IM are called entire-null DTLs. Prebend segments were not included in the DTL count. Consequently, the sum of AQP1-positive and AQP1-negative DTLs at each interval is always less than the number of ATLs. Inset: DTL and ATL segments for 2 inner medullary loops of Henle. Entire-null DTL is shown on right. Loop population for this single secondary CD cluster [as reflected by the ATLs (ATL Exp Decay)] was fitted to the same exponential that was used previously for all loops in the entire IM (3, 10, 13).
|
|
The number of ATLs declines along the corticopapillary axis in a shallow exponential (Fig. 6A). The decay rate for all loops that lie in this single secondary CD cluster (as reflected by the ATL distribution) appears to be well represented by the exponential used previously for all loops in the entire IM (3, 10, 13). As shown in Fig. 6A, the published exponential and the rates of decay for the DTLs and ATLs of this secondary CD cluster appear to be essentially the same. The AQP1-positive DTLs decline in number in a near-linear fashion, with none present below 1,700 µm from the IM base (Fig. 6A). For this initial 1,700 µm below the IM base, the number of AQP1-negative DTLs declines at a low rate, producing the pattern of a quasi-plateau (Fig. 6A). This plateau in the first 1,700 µm results from a combination of two factors (Fig. 6B): 1) AQP1-negative DTLs that are AQP1-negative before entering the IM ("entire-null DTLs", 50% of DTLs) form bends by 1 mm below the IM base, thereby reducing the number of AQP1-negative DTLs through attrition; and 2) AQP1-positive DTLs gradually become AQP1-negative, and these loops add to the existing AQP1-negative DTL population through a one-for-one replacement of AQP1-positive DTLs, thereby tending to offset the AQP1-negative segments lost by attrition. After the point where AQP1-positive DTLs disappear from the cluster entirely (1,700 µm below the IM base), the number of AQP1-negative DTLs declines at a shallow exponential rate comparable to that of the ATLs as loops form bends.
ATLs associated with a primary cluster (as delimited by EDM) can be further described by their relationship to the CDs in that cluster. To do this more precisely than we have in the past (16), we drew polygons around the outer border of the CDs making up that cluster. We drew a straight line from the outermost convex point on the surface of each outer CD to a similar point on its neighboring CD within that cluster (Fig. 1). By this process, we subdivided the total area of each primary cluster within the EDM boundary into the polygon area within the polygon area boundary (intracluster interstitium) and the area between the EDM boundary and the polygon area boundary (intercluster interstitium; Fig. 1). With this determined, we evaluated the position of each ATL relative to the CDs in each of five primary clusters that form a single secondary cluster, first at the IM base and then at 1,000-µm intervals axially until all loops associated with this secondary cluster had turned back (Fig. 7). As in our previous work (16), we found that we could divide the ATLs within each primary CD cluster into three groups. Group 1 includes ATLs that are bounded by at least two CDs or lie entirely within the polygon area. The mean tubule length for group 1 was 465.5 µm (SD 553.3, n = 21). Group 2 includes ATLs that are bounded by just one CD. These ATLs lie primarily on the border of the polygon area, with most of their cross-sectional area outside the polygon area. The mean tubule length for group 2 was 1,156.6 µm (SD 664.8, n = 40). Group 3 includes ATLs that are not bounded by any CDs. These ATLs lie primarily in the area outside the border of the polygon area of the cluster but within the EDM boundary of the cluster. The mean tubule length for group 3 was 2,352.7 µm (SD 1,099.6, n = 33). The frequency of ATLs in each group is nearly equal along the axis of the IM (Fig. 7).
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 7. ATLs associated with a primary cluster (as delimited by EDM) can be further described by their relationship to the CDs in that cluster. ATL groups 1–3 are shown for 4 different levels below the IM base: those that touch at least 2 CDs or lie entirely within the polygon area (group 1), those that touch just 1 CD (group 2), and those that do not touch any CDs (group 3).
|
|
Surface area of wide bends of loops of Henle.
Loops of Henle form two types of bends in the IM. One is the narrow hairpin bend that occurs throughout the IM from the base to the tip. The other, found only within the terminal 500 µm of the IM (inner zone 2), has a bend with a very broad transverse segment (18). These wide-bend loops, which include part of the ClC-K1-positive prebend region of the DTL and part of the ATL, tend to be close to and curved laterally around the CDs in this region (18).
Because the prebend segments and the loop bends, whether narrow hairpin or broad transverse bends, are ClC-K1-positive (15, 18), we infer that, similar to the ATLs (5), which are also ClCK-1 positive (15), they are highly permeable to NaCl. Significant efflux of NaCl from the loops of Henle presumably occurs only from the ClC-K1-positive regions (13). Moreover, model studies indicate that significant net NaCl efflux from the loops occurs only for a short axial distance around the loop bend, probably about the length of the prebend and a comparable length of the ATL (11). In the case of the broad transverse bends, the region for net NaCl efflux appears to be increased transversely so that large amounts of NaCl can be delivered over a very narrow axial distance near the tip of the IM. The surface area of the portions of the loops involved in NaCl efflux should be important in determining the magnitude of NaCl efflux. Therefore, to determine more quantitatively the importance of these broad transverse bends, we examined their effect on the axial distribution of the surface area most important for NaCl efflux in the terminal 500 µm of the IM (inner zone 2). To do this, we determined the basolateral surface area of all loop bends in this entire terminal region from the start of the prebend segments through an equivalent axial distance on the ATLs. First, we determined the surface area at 5-µm intervals along the corticopapillary axis for all native narrow-bend and wide-bend loops. Next, we repeated this measurement at 5-µm intervals for the same loop population after we had hypothetically converted the wide-bend loops to narrow-bend loops.
For this conversion, we considered loops with >50-µm bends in the transverse direction to be wide-bend loops. We then replaced the prebend and transverse extension of each wide-bend loop with two axial segments totaling that length: one segment on the descending side and one on the ascending side of the loop (Fig. 8). Finally, we repositioned the loop bend so that it lay at the same axial level as the original transverse segment (Fig. 8). This hypothetical conversion has the effect of moving nearly all the transverse extension of each wide-bend loop to a higher axial level closer to the base of inner zone 2 and away from the tip of the papilla.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 8. Architecture of native wide-bend loop of Henle (left) and wide-bend loop converted to narrow-bend loop of Henle (right). Prebend and equivalent-length ATL postbend segments are shown as dotted lines.
|
|
As shown in Fig. 9, when all loops are converted to narrow hairpin bends, the cumulative surface area-to-papilla volume ratio for all loops (including, for each loop, the area from the start of the prebend to an equal axial distance on the ATL) decreases steadily over the final 300 µm of the papilla, even though it is corrected for the decreasing papillary volume. In contrast, the same surface area of the native population (wide-bend and narrow-bend loops) remains more nearly constant until the last loops turn back in the terminal 25 µm. Indeed, over most of the terminal 100 µm, the surface area of the native population is about twice as great as that of the converted loops. Thus the presence of the wide bends maintains a large surface area for NaCl efflux in a very small and decreasing papillary volume.
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 9. Cumulative loop of Henle bend surface area-to-papillary volume ratio for all loops (including for each loop the area from the start of the prebend to an equal axial distance on the ATL). Ratios were determined at 5-µm intervals throughout the final 500 µm of the IM (inner zone 2). This was first done for all native loops (wide-bend + narrow-bend). Wide-bend loops were then converted to equivalent narrow-bend loops (Fig. 8), and their 3-dimensional surface area was determined. Cumulative loop surface area-to-papilla volume ratio was then computed for all converted wide-bend loops + all native narrow-bend loops.
|
|
The reabsorption of NaCl from the loops of Henle relative to the reabsorption of urea, water, and NaCl from the CDs is important for most proposed models of the concentrating process in the IM (13). It appeared likely that the wide transverse bends might be important in determining the relationship between loop and CD reabsorption. Therefore, we examined the ratio of the surface area of all loop bends (from the start of the prebend through an equal axial distance on the ATL) to the surface area of all CDs at 5-µm intervals along the terminal 500 µm of the papilla. We did this for the native population of loop bends (wide-bend and narrow-bend loops) and for the converted population of loop bends (all wide bends converted to narrow bends). The results are shown in Fig. 10. For the converted loop bends, the surface area ratio averages 0.40 until the terminal 25 µm, where the final loops turn back. For the native population of loop bends, this ratio also averages 0.40 until the final 100 µm. In this region, the ratio increases dramatically until the surface area of the loop bends equals or exceeds the surface area of the CDs. Indeed, it reaches 1.40 times the CD surface area 50 µm above the tip. Below this point, the ratio falls as the last loops turn back. The absolute magnitude of this ratio in the terminal 100 µm reflects some decrease in CD surface area (data not shown), as well as an increase in loop bend area resulting from the broad transverse bends. However, the rate of decrease in CD surface area approximately matches the rate of decrease in loop-bend surface area when all loop bends are converted to narrow ones (Fig. 9), resulting in the nearly constant CD surface area-to-loop-bend surface area ratio (Fig. 10). Therefore, it is clear that it is the presence of the broad transverse segments of loop bends in this terminal region of the IM (18) that alters the relationship of the loop-bend surface area to CD surface area.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 10. Cumulative loop of Henle bend surface area-to-CD surface area ratio for all loops (including for each loop the area from the start of the prebend to an equal axial distance on the ATL). Ratios were determined at 5-µm intervals throughout the final 500 µm of the IM (inner zone 2). This was first done for all native loops (wide-bend + narrow-bend). Wide-bend loops were then converted to equivalent narrow-bend loops (Fig. 8), and their 3-dimensional surface area was determined. Cumulative loop surface area-to-CD surface area ratio was then computed for all converted wide-bend loops + all native narrow-bend loops.
|
|
|
DISCUSSION
|
|---|
CD clusters form by iterative branching and extension of CD segments beginning in early embryonic development (1, 20), leading to a variety of repeating CD branching patterns and other distinct architecture clearly discerned in the adult (15, 16). As described in our previous studies (16, 17), these CD clusters form the lateral organizing motif for loops of Henle and vasa recta in the first 3.0–3.5 mm of the rat IM. To quantitate loops and vessels associated with each CD cluster, we chose to use the EDM technique to empirically define the lateral, interstitial, and spatial domains of each primary and secondary CD cluster. The procedure creates boundaries that lie in the intercluster interstitium separating adjacent CD clusters. Each boundary was created for the full axial length of each cluster (although continuous axial boundaries are not shown), encompassing, in addition to each CD cluster, those nephron segments and blood vessels that are closest to that cluster. Compartments defined by EDM boundaries thereby include all components that exist in the intercluster interstitial space as well as in the intracluster space.
The intracluster interstitium is populated largely with the microdomains of interstitial nodal spaces created by the arrangement of two AVRs abutting a single CD and at least one ATL (17, 18). Although ATLs and AVR also reside throughout the intercluster regions, nodal spaces are absent, except for the region immediately adjacent to the peripheral edges of the CDs in the clusters where ATLs and CDs with abutting AVR coexist. Thus, in a functional sense, the intracluster interstitium may in fact extend laterally to include those nodal spaces that abut the outer edges of the CDs in the cluster.
Nearly all the AQP1-positive DTLs and UT-B-positive DVR lie almost directly on EDM-defined cluster boundaries, as far as possible from any CD; moreover, they lie entirely in the intercluster region. The contiguous, deeper, AQP1-negative DTL segments tend to lie closer to the edge of CD clusters, at the interface between the intercluster and intracluster regions, and infrequently lie within the cluster (16). Fenestrated blood vessels carrying fluid in a descending direction (PV-1-positive, UT-B-negative vessels that can be defined as DVR), as well as UT-B-positive DVR, appear to be restricted to the intercluster region (unpublished observations). Therefore, all inner medullary countercurrent exchange between DVR and AVR is restricted to this region.
ATLs, more than DTLs, tend to be unambiguously associated with a particular CD cluster. This close association with an individual cluster begins with the prebend segment. However, the number and identity of loops associated with a given CD cluster vary to some degree along the IM axis, as seen by variably increasing and decreasing numbers of ATLs and DTLs and by the variations in the sum of DTLs and ATLs (Fig. 6). This instability underscores the fact that, for a given loop, its association with a single cluster may not be entirely constant along the IM axis.
In the present study, we have proposed a nomenclature to facilitate reference to apparently specific axial regions of the IM. These regions are defined by differences in 1) the lateral arrangements of CDs, loops of Henle, and vasa recta discussed above, 2) protein expression patterns (Table 1), and 3) structure. We have suggested that the IM be divided into an outer zone and an inner zone. We have further proposed that the outer zone be subdivided into zones 1 and 2 on the basis of the distribution of AQP1 in DTLs (Table 1) and that the inner zone be subdivided into zones 1 and 2 on the basis of the presence of loops with wide bends (Table 1).
The lateral and axial compartmentation of loops of Henle and vasa recta of specific structural type, direction of flow, and expression of transporters and channels seems likely to play a significant role in the inner medullary concentrating process. For example, lateral solute gradients may be involved with developing and sustaining the concentrating mechanism, but technical limitations prevent satisfactory analysis of NaCl and urea concentrations in transverse compartments of the IM, and it is not known if lateral solute gradients exist. Restricted transverse diffusion of fluid and solutes could result from high interstitial fluid viscosity (e.g., from hyaluronan in the inner zone) (7, 14) and/or architectural barriers and axial components that are spatially isolated in the lateral dimension, such as those described here (Table 1). In any case, there must be some form of lateral separation of function, notably between the intercluster and intracluster regions, at least through the outer 3.0–3.5 mm of the IM, since AQP1-positive DTLs and UT-B-positive DVR are entirely absent from intracluster nodal spaces.
The CD cluster, as a model, may be useful for understanding similar repeating units representing whole medullary function, thereby simplifying models of physiological processes. The geometric shape of the CD cluster roughly parallels the conical shape of the papilla, although it is much smaller. Importantly, the exponential decay rate of loop number along the axis of the CD cluster appears to be nearly identical to the exponential decay rate previously reported for the sum total of all loops in the IM of Sprague-Dawley rats (6, 10, 13) (Fig. 6). From this information, the membrane surface area available for fluid and solute transport by tubules and vessels within each functional unit can be determined from tubule population and diameter.
ATL positional analysis confirms previous work (16) showing mean loop length proportional to position relative to CD cluster. Shorter loops, on average, lie within the intracluster region, whereas the longest loops lie in the intercluster region, outside CD clusters. Although not shown in this study, the contiguous DTLs tend to show comparable positional relationships to CD clusters (16). Because of the probable importance of NaCl reabsorption by prebend and equivalent-length postbend segments (11, 13), detailed quantitative data on prebend positions relative to interstitial nodal spaces abutting CDs, as well as loop length, and loop origin and destination in the OM would provide significant information for modeling the concentrating mechanism. Similarly, detailed knowledge of positions of AQP1-negative DTLs and transitions from AQP1-positive segments would provide important information for modeling. The zonation as designated here should be useful in mapping and referencing these elements throughout the IM.
Significant net NaCl efflux from the loops arguably occurs only for a short axial distance around the loop bend, above which the driving force rapidly diminishes (11, 13). Wide-bend loops of Henle, recently described in the terminal 500 µm of the papilla (18), suggest that significant efflux of NaCl may occur over a very small axial distance near the papilla tip. Mathematical models indicate that such delivery of NaCl at a single point would be most effective for producing highly concentrated urine (12). To gain insight into the possible significance of the wide-bend loops for this type of NaCl delivery, we examined their effect on the axial distribution of the loop surface area that is apparently most important for NaCl efflux. The results indicate that the presence of the wide bends in a limited axial region prevents the decrease in this surface area, especially over the final 250 µm of the papilla tip, which would occur if all loops had narrow bends (Fig. 9). Of perhaps even greater significance, the occurrence of wide-bend loops leads to a striking increase in this apparently critical loop surface area relative to the CD surface area over only 50 µm in axial distance (from 25 to 75 µm above the papilla tip; Fig. 10). These quantitative analyses suggest that the wide-bend architecture may well underlie the very high osmolality of the deep papilla.
Anatomic proximity between different structures likely does not result in rigid interactions but, rather, a continuum of functional relationships; fluid and solute diffusion between structures lying in neighboring CD clusters certainly occurs to some extent. The extent to which that occurs would be influenced by distance, diffusivity of interstitial fluid, and physical barriers such as we describe here. Another barrier likely includes interstitial cells, which have not been studied sufficiently for inclusion in architectural models or in UCM equations. Nevertheless, the localization of structures and transporter and channel proteins to the lateral and axial regions delimited in the present study may be significant for the function of the UCM in the IM.
|
GRANTS
|
|---|
This study was supported in part by National Institutes of Health Research Grant DK-16294, Grant ES-06694 for the Southwest Environmental Health Sciences Center, and Training Grants HL-07249 and GM-08400.
|
ACKNOWLEDGMENTS
|
|---|
We gratefully acknowledge contributions and collaborations with Harold Layton, who has greatly advanced our concepts of functional compartmentation in the renal inner medulla. We also thank the following students for assistance with three-dimensional reconstructions: Brandi Hoopes, Nick Pyle, and Joni Saquilayan.
|
FOOTNOTES
|
|---|
Address for reprint requests and other correspondence: T. L. Pannabecker, Univ. of Arizona Health Sciences Center, Dept. of Physiology, AHSC 4130, 1501 N. Campbell Ave., Tucson, AZ 85724-5051 (e-mail: pannabec{at}u.arizona.edu)
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES
|
|---|
- Al-Awqati Q, Goldberg MR. Architectural patterns in branching morphogenesis in the kidney. Kidney Int 54: 1832–1842, 1998.[CrossRef][Web of Science][Medline]
- Bankir L, Fischer C, Fischer S, Jukkala K, Specht HC, Kriz W. Adaptation of the rat kidney to altered water intake and urine concentration. Pflügers Arch 412: 42–53, 1988.[CrossRef][Web of Science][Medline]
- Han JS, Thompson KA, Chou CL, Knepper MA. Experimental tests of three-dimensional model of urinary concentrating mechanism. J Am Soc Nephrol 2: 1677–1688, 1992.[Abstract]
- Hervy S, Thomas SR. Inner medullary lactate production and urine-concentrating mechanism: a flat medullary model. Am J Physiol Renal Physiol 284: F65–F81, 2003.[Abstract/Free Full Text]
- Imai M. Function of the thin ascending limb of Henle of rats and hamster perfused in vitro. Am J Physiol Renal Fluid Electrolyte Physiol 232: F201–F209, 1977.[Abstract/Free Full Text]
- Knepper MA, Danielson RA, Saidel GM, Post RS. Quantitative analysis of renal medullary anatomy in rats and rabbits. Kidney Int 12: 313–323, 1977.[Web of Science][Medline]
- Knepper MA, Saidel GM, Hascall VC, Dwyer T. Concentration of solutes in the renal inner medulla: interstitial hyaluronan as a mechano-osmotic transducer. Am J Physiol Renal Physiol 284: F433–F446, 2003.[Abstract/Free Full Text]
- Koepsell A, Nicholson WAP, Kriz W, Höhling HJ. Measurements of exponential gradients of sodium and chlorine in the rat kidney medulla using the electron microprobe. Pflügers Arch 350: 167–184, 1974.[CrossRef][Web of Science][Medline]
- Kriz W, Bankir L. A standard nomenclature for structures of the kidney. Kidney Int 33: 1–7, 1988.[Web of Science][Medline]
- Layton HE. Distribution of Henle's loops may enhance urine concentrating capability. Biophys J 49: 1033–1040, 1986.[Web of Science][Medline]
- Layton HE. Concentrating urine in the inner medulla of the kidney. Comments Theoret Biol 1: 179–196, 1989.
- Layton HE, Davies JM. Distributed solute and water reabsorption in a central core model of the renal medulla. Math Biosci 116: 169–196, 1993.[CrossRef][Web of Science][Medline]
- Layton AT, Pannabecker TL, Dantzler WH, Layton HE. Two modes for concentrating urine in rat inner medulla. Am J Physiol Renal Physiol 287: F816–F839, 2004.[Abstract/Free Full Text]
- MacPhee PJ. Estimating rat renal medullary interstitial oncotic pressures and the driving force for fluid uptake into ascending vasa recta. J Physiol 505: 529–538, 1998.[Web of Science]
- Pannabecker TL, Abbott DE, Dantzler WH. Three-dimensional functional reconstruction of inner medullary thin limbs of Henle's loop. Am J Physiol Renal Physiol 286: F38–F45, 2004.[Abstract/Free Full Text]
- Pannabecker TL, Dantzler WH. Three-dimensional lateral and vertical relationships of inner medullary loops of Henle and collecting ducts. Am J Physiol Renal Physiol 287: F767–F774, 2004.[Abstract/Free Full Text]
- Pannabecker TL, Dantzler WH. Three-dimensional architecture of inner medullary vasa recta. Am J Physiol Renal Physiol 290: F1355–F1366, 2006.[Abstract/Free Full Text]
- Pannabecker TL, Dantzler WH. Three-dimensional architecture of collecting ducts, loops of Henle, and blood vessels in the renal papilla. Am J Physiol Renal Physiol 293: F696–F704, 2007.[Abstract/Free Full Text]
- Russ JC. The Image Processing Handbook (3rd ed.) Boca Raton, FL: CRC, 1999.
- Saxen L. Organogenesis of the Kidney. Cambridge, UK: Cambridge University Press, 1987.
- Stan RV, Ghitescu L, Jacobson BS, Palade GE. Isolation, cloning, and localization of rat PV-1, a novel endothelial caveolar protein. J Cell Biol 145: 1189–1198, 1999.[Abstract/Free Full Text]
- Wang X, Wexler AS. The effects of collecting duct active NaCl reabsorption and inner medulla anatomy on renal concentrating mechanism. Am J Physiol Renal Fluid Electrolyte Physiol 270: F900–F911,1996.[Abstract/Free Full Text]
- Wang X, Thomas SR, Wexler AS. Outer medullary anatomy and the urine concentrating mechanism. Am J Physiol Renal Physiol 274: F413–F424, 1998.[Abstract/Free Full Text]
- Wexler AS, Kalaba RE, Marsh DJ. Passive, one-dimensional countercurrent models do not simulate hypertonic urine formulation. Am J Physiol Renal Fluid Electrolyte Physiol 252: F358–F378, 1987.
- Wexler AS, Kalaba RE, Marsh DJ. Three-dimensional anatomy and renal concentrating mechanism. I. Modeling results. Am J Physiol Renal Fluid Electrolyte Physiol 260: F368–F383, 1991.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
A. T. Layton, H. E. Layton, W. H. Dantzler, and T. L. Pannabecker
The Mammalian Urine Concentrating Mechanism: Hypotheses and Uncertainties
Physiology,
August 1, 2009;
24(4):
250 - 256.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. L. Pannabecker
Loop of Henle interaction with interstitial nodal spaces in the renal inner medulla
Am J Physiol Renal Physiol,
December 1, 2008;
295(6):
F1744 - F1751.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
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]
|
|
|
Copyright © 2008 by the American Physiological Society.
TOP
ABSTRACT
METHODS
