Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Concentrated urine is produced through the generation of an osmotic gradient in the renal medulla. The gradient in the outer medulla is widely believed to be generated by means of a mechanism in which an osmotic pressure difference is amplified by countercurrent multiplication. This difference arises from active NaCl reabsorption from thick ascending limbs, which dilutes ascending limb flow relative to flow in vessels and other tubules. In the inner medulla, the mechanisms responsible for generating the osmotic gradient have not been definitively identified. Important advances in understanding the urine concentrating mechanism include: identification and localization of key transport proteins for water, sodium, and urea; better resolution of the anatomical relationships in the medulla; and improvements in mathematical modeling of the urine concentrating mechanism. Continued experimental investigation of transepithelial transport and its regulation, both in normal animals and in knock-out mice, and incorporation of the resulting information into mathematical simulations, may help to more fully elucidate the inner medullary urine concentrating mechanism.
Keywords
vasopressin; urine osmolality; aquaporins; urea transporter; mathematical modeling
The ability to vary water excretion is essential for mammals, which generally do not have continuous access to water, but must maintain a nearly constant blood plasma osmolality. Mammals, therefore, need a mechanism that allows them to regulate water loss to closely match water intake. In addition, because sodium and its anions are the principal osmotic constituents of blood plasma, and plasma sodium concentration must be kept nearly constant, water loss must be regulated by a mechanism that decouples water and sodium. These critical regulatory capabilities are provided by the kidney’s urine concentrating mechanism: when water intake is large enough to dilute blood plasma, urine more dilute than plasma is produced to concentrate the plasma; when water intake is so small that blood plasma is concentrated, urine more concentrated than plasma is produced to dilute the plasma. In both cases, the rate of sodium excretion is small and varies little; indeed, the total solute excretion rate varies little ( Figure 43.1 ).
Urine osmolality varies widely in response to changes in water intake. Following a prolonged period without water intake, such as occurs when an individual sleeps, human urine osmolality may rise to ~1200 mOsm/kg H 2 O, about four times plasma osmolality (~290 mOsm/kg H 2 O). However, following the ingestion of large quantities of water, such as commonly occurs at breakfast, urine osmolality may decrease rapidly. Humans (and other mammalian species) are able to dilute their urine to ~50 mOsm/kg H 2 O. Such large and rapid changes in osmolality require that the cells of the inner medulla have adaptive mechanisms (e.g., osmolytes) to protect them from osmotic damage.
Maximum urine osmolality varies widely among mammalian species. The long-nosed bat Leptonycteris sanborni can concentrate only to about 350 mOsm/kg H 2 O, while the Australian hopping mouse Notomys alexis can concentrate to nearly 9400 mOsm/kg H 2 O. Primates can typically concentrate their urines from ~1000 to 2000 mOsm/kg H 2 O. Beluga whales and bottle-nosed dolphins, which have access only to hypertonic ocean water (~1000 mOsm/kg H 2 O), can concentrate urine up to ~1800 mOsm/kg H 2 O. Most laboratory data relevant to the urine concentrating mechanism have been obtained from rabbits or rodents that can achieve higher maximum urine osmolalities than humans: the European rabbit can concentrate to ~1400 mOsm/kg H 2 O, whereas the eastern cotton tail can concentrate to ~3300 mOsm/kg H 2 O; rats to ~3000 mOsm/kg H 2 O; mice and hamsters to ~4000 mOsm/kg H 2 O; and chinchillas to ~7600 mOsm/kg H 2 O.
Regardless of maximum concentrating ability, the kidneys of all mammals maintain an osmotic gradient that increases from the cortico-medullary boundary to the tip of the medulla (papillary tip). This osmotic gradient is sustained even in diuresis, although it is diminished in magnitude relative to antidiuresis. The major constituent of the osmotic gradient in the outer medulla is NaCl; in the inner medulla, the major constituents are NaCl and urea. The cortex is nearly isotonic to plasma, while the papillary tip is hypertonic to plasma and, in antidiuresis, has osmolality similar to urine. The major urinary solutes are sodium and potassium accompanied by univalent anions and by urea; urea is the predominant solute in urine during antidiuresis. The sodium, potassium, and urea concentrations in rat plasma, papillary tissue, and urine, during both diuresis and antidiuresis, are given in Table 43.1 .
A. Diuresis (urine flow/animal=192:l/min) | |||
---|---|---|---|
Component | Plasma | Papilla | Urine |
Na + (mEq/l) | 138 | 159 | 5.4 |
K + (mEq/l) | 6.0 | 66.0 | 5.9 |
Urea (mM) | 4.5 | 34.1 | 22.6 |
Osmolality (mOsm/kg H 2 O) | 304 | 572 | 59 |
B. Antidiuresis (urine flow/animal=~5:l/min) | |||
---|---|---|---|
Component | Plasma | Papilla | Urine |
Na + (mEq/l) | 145 | 417 | 148 |
K + (mEq/l) | 6.7 | 102 | 140 |
Urea (mM) | 4.4 | 605 | 946 |
Osmolality (mOsm/kg H 2 O) | 314 | 1832 | 1805 |
The mechanisms responsible for the separate control of water and sodium excretion are largely located in the renal medulla, where the nephron segments and vasa recta are arranged in complex but specific anatomic relationships, both in terms of which segments connect to which segments, and in terms of three-dimensional configuration. The production of concentrated urine involves complex interactions among the nephron segments and vasculature. In the outer medulla, thick ascending limbs of the loop of Henle actively absorb NaCl, diluting the luminal fluid and providing NaCl to increase the osmolality of the medullary interstitium, pars recta, descending limbs, collecting ducts, and vasculature. The countercurrent configuration of nephron segments and vessels allows the generation of a medullary osmolality gradient along the cortico-medullary axis. In the inner medulla, osmolality continues to increase, but the source of the concentrating effect remains controversial. However, the most widely accepted mechanism remains passive absorption of NaCl, in excess of solute secretion, from thin ascending limbs of the loops of Henle.
The structural organization of the mammalian kidney is discussed in detail elsewhere in this book. This section, based in large measure on key studies (e.g., ), summarizes features that are pertinent to the urine concentrating mechanism.
In most mammals studied, the kidney contains short looped and long-looped nephrons; both have loops of Henle that are arranged in a hairpin configuration ( Figure 43.2 ). They differ in two important aspects: the loops of short-looped nephrons turn near the inner-outer medullary border and lack a thin ascending limb, whereas the loops of long-looped nephrons extend into the inner medulla and contain a thin ascending limb. Thin ascending limbs are found only in the inner medulla, and their transition to thick ascending limbs defines the inner–outer medullary border. Thus, only thick ascending limbs are found in the outer medulla, regardless of the type of loop. Some mammalian kidneys, e.g., human kidneys, have nephrons whose loops of Henle do not reach into the medulla; these nephrons are called cortical nephrons. Tubular fluid flows from thick ascending limbs of both short and long looped nephrons to distal convoluted tubules. Several distal tubules merge to form cortical collecting ducts that descend through the cortex and then become medullary collecting ducts that pass through the outer medulla. The collecting ducts merge along the entire length of the inner medulla, ultimately forming the ducts of Bellini, which open into the renal pelvis at the papillary tip.
Small mammals, such as rodents, have unipapillate kidneys. In these mammals, the papilla is an inverted pyramid-shaped portion of the innermost inner medulla; the papilla descends into the renal pelvis. Larger mammals (including humans) have multipapillate kidneys in which each papilla descends into a renal calyx. The renal pelvis is formed from the merging of these calyces. In all mammals, urine exits through the ducts of Bellini into the renal pelvis. The pelvis, which connects to the ureter, is bounded by two epithelia: the papillary surface epithelium lining the surface of the papilla, and a ureteral-type epithelium extending from the ureter up into the renal pelvic fornices.
The descending and ascending vasa recta, which provide the blood supply for the medulla, are arranged roughly in parallel. Although their configuration is similar to the hair-pin configuration of the loops of Henle, there is an important anatomic difference: the tubular segments that make up the loops of Henle are contiguous, whereas the descending and ascending vasa recta are separated by capillary plexuses. Blood enters the medulla through descending vasa recta, passes through capillary plexuses located at various depths within the medulla, and then enters ascending vasa recta. Vascular bundles, which are aggregations of both descending and ascending vasa recta, form in the outer stripe, but become much more prominent in the inner stripe. Lemley and Kriz have proposed using the vascular bundle (see detail, Figure 43.3 ) as the histotopographical core around which the various outer medullary tubule structures are arranged.
Studies of inner medullary structure by Kriz and co-workers, and recent studies by Pannabecker and Dantzler, found that the inner medullary collecting ducts (IMCDs) in the inner medullary base form clusters that coalesce along the cortico–medullary axis. Using immunohistochemical labeling and computer-assisted reconstruction, Pannabecker, Dantzler, and colleagues have elucidated new detail of the functional architecture of the rat inner medulla (see recent review ). A computerized reconstruction of the inner medullary portion of several long-looped nephrons from rats is shown in Figure 43.4 , in which antibodies to the water channel aquaporin-1 (AQP1, shown in red) and the chloride channel ClC-K1 (shown in green) are used to label the thin descending and ascending limbs of the loops of Henle, respectively (reviewed in ). In the base of the inner medulla, thin descending limbs are predominantly present at the periphery of these clusters, and appear to form an asymmetric ring around each collecting duct cluster. In thin descending limbs of loops of Henle that turn within the upper first millimeter of the inner medulla, no AQP1 was detected. In contrast, there are three discernible functional subsegments in thin descending limbs of loops of Henle that turn below the first millimeter: the upper 40% expresses AQP1, whereas the lower 60% do not. ClC-K1, a marker of the thin ascending limb-type epithelium, is first detected in the final ~165 micrometers of the thin descending limb, as well as in the contiguous thin ascending limb. Thus, ClC-K1 is detected before the bend of the loops of Henle. This finding is consistent with previous morphological studies demonstrating that the descending limb to ascending limb transition occurs before the loop bend. In addition, a substantial portion of the inner medullary thin descending limbs of long-looped nephrons did not express either AQP1 or ClC-K1, as indicated in gray in Figure 43.4a .
In contrast, thin ascending limbs are distributed relatively uniformly among collecting ducts and thin descending limbs. In Munich-Wistar rats, Pannabecker and Dantzler identified three population groups of loops of Henle, distinguished by thin ascending limb position at the base of the inner medulla and by differing loop length. Group 1 loops, having thin ascending limbs that are interposed between collecting ducts, reach a mean length of 700 μm into the inner medulla; Group 2 loops, having thin ascending limbs that are adjacent to just one collecting duct, reach 1500 μm and Group 3 loops, having thin ascending limbs that lie more than a half-tubule diameter from a collecting duct, reach 2200 μm. As collecting ducts coalesce and shorter loops disappear, the originating portions of longer thin ascending limbs run alongside the collecting ducts for substantial distances (a more detailed description of inner medullary loop subgroups was given in ).
Moreover, Pannabecker et al. proposed that the inner medulla has at least four distinguishable zones (axial subsections) that can be differentiated by the variable characteristics of vasa recta and loops of Henle. The four zones are illustrated in Figure 43.4b . Three countercurrent systems were hypothesized to exist within these zones: (1) an intercluster system within the CD clusters where most of the work of concentrating is conducted in and around nodal spaces that are surrounded by IMCDs, thin ascending limbs, and AVR; (2) an extracluster system in the outerzones (outer 3 to 3.5 mm) that serves to carry water absorbed from water-permeable portions of descending limbs back to the OM via AVR; and (3) a papillary system in the inner zones where the highest concentrations are attained with the aid of direct interactions between transversely-running segments of loops of Henle and collecting ducts.
Distinct types of interstitium are found in the vascular bundle in the outer medulla, in the interbundle region of the outer medulla, and in the inner medulla. These interstitia may play an important role in medullary solute and water transfer, especially in the inner medulla, where interstitial cells are interspersed in a gelatinous matrix of acid mucopolysaccharides, which is largely devoid of any capillary plexuses, laterally flowing capillaries or lymphatics. Thus, the inner medullary interstitium should greatly slow lateral bulk flow of solutes and water.
The number of nephrons found in mammals, and thus the number of loops of Henle, varies over many orders of magnitude, increasing sub-linearly with increasing body mass. The mouse has about 12,400 nephrons per kidney ; rat, 30,000–40,000 ; rabbit, 230,000 ; human, 0.3–1.4 million ; elephant, 7.5 million ; and fin whale, 192 million. In contrast, medullary thickness in mammals varies from 3 to 25 mm, thus indicating that maximum loop of Henle length varies over about an order of magnitude ; proximal tubule diameter changes little from rat to fin whale, increasing by a factor of about 1.3.
Although loops of Henle of variable length are found in all mammals, most mammals are thought to have both short- and long-looped nephrons. Exceptions include the dog, with all long loops, and the mountain beaver Aplodontia rufa , which has thick ascending limbs only, and has a renal medulla that corresponds to the outer medulla of other mammals. Generally, however, there are more short-looped than long looped nephrons, and the long-looped nephrons tend to exhibit substantial variation in the depth reached within the inner medulla. Measurements in the rabbit and rat indicate that the decrease in loop of Henle and collecting duct population in the inner medulla is approximately exponential, with most loops of Henle turning back in the outer portion of the inner medulla, and with collecting ducts converging to a few ducts of Bellini. A similar pattern is seen in the medullary cones of the avian kidney which, like the mammalian kidney, is able to produce a concentrated urine, although only to osmolalities of about twice that of blood plasma.
The pattern of decrease in the tubule populations of the rat renal medulla is portrayed in Figure 43.5 , which gives curves approximating loop and collecting duct population as a function of normalized medullary depth. About 38,000 loops and 7300 collecting ducts extend through most of the outer medulla. About 28–33% of the loops of Henle in rat have thin ascending limbs and reach into the inner medulla. The populations of loops of Henle and collecting ducts both decline rapidly, but the loop population decreases more rapidly, so that the loop and collecting duct populations are more nearly equal in the papilla. In human, about
of the loops of Henle reach into the inner medulla.
Figure 43.5 also portrays the concentration of urea, the sum of the concentrations of sodium and its anions, and the osmolality, as a function of medullary depth, as determined in tissue slices harvested from vasopressin-treated Wistar rats. The experimental data points indicated in Figure 43.5 are connected by natural cubic splines which generate smooth curves; these curves have shapes supported by other studies in rat. Osmolality increases by a factor of about 2 in the outer medulla, and by an additional factor of 3 in the inner medulla, where urea makes a substantive contribution. As can be inferred from the values for urine (U) in Figure 43.5 , sodium is largely carried by flow in the loops of Henle and vasculature, while urea makes up a large portion of the solute in collecting duct flow. Potassium has a tissue concentration of about 80–100 mM along the medulla, but it makes a larger contribution (~150 mM) to urine.
The osmolality increase along the outer medulla arises from the vigorous transepithelial transport of NaCl from thick ascending limbs into the surrounding interstitium. This effect is believed to be augmented by a process of countercurrent multiplication, described in a subsequent section ( vide infra ). However, as shown in Figure 43.5 , the osmolality gradient is largest in the papilla, even though only a small fraction of the loops, tubules, and vasa recta reach into the papilla, and even though the population of tubules and vessels is rapidly decreasing there. The remarkable capacity for generating high osmolalities in so small a volume (~0.5% of total kidney volume ) has thus far resisted a generally satisfactory explanation.
This section will review the water, urea, and sodium permeability values measured in isolated perfused tubules in nephron segments involved in producing concentrated or dilute urine. Thin limb segments are difficult to perfuse, and most measurements involving different species have been made by different laboratories. Thus, some caution must be used in comparing these values. Tables 43.2–43.5 contain values obtained from animals receiving food and water ad libitum ; representative values were chosen since space does not permit citing every original manuscript.
Species | ||||
---|---|---|---|---|
Chinchilla | Rat | Rabbit | Hamster | |
THIN DESCENDING LIMB TYPE I | ||||
Na b | 0–2 c | 4.8 | ||
Urea b | 13 c | 1–2 c | 8 | |
Water d | 2295 c | 2420 c | 3257 | |
Na + /K + -ATPase e | 2–5 c | 2–4 c | ||
THIN DESCENDING LIMB TYPE II | ||||
Na b | 1 c | 23–66 | ||
Urea b | 3 | 0 | 1 c | 3 |
Water d | 2600 | 2295 c | 2315 c | 5378 |
Na + /K + -ATPase e | 2–5 c | 2–4 c | ||
THIN DESCENDING LIMB TYPE III | ||||
Na b | 29 | 4 | ||
Urea b | 17–29 | 13 c | 13 | |
Water d | 1550 | 2295 c | 1693 | |
Na + /K + -ATPase e | 3 | |||
THIN DESCENDING LIMB TYPE III DISTAL (PAPILLARY SUBSEGMENT) | ||||
Na b | 74 | |||
Urea b | 48 | |||
Water d | 60 |
a References are in supercripted.
c Subsegment not specified, thus cannot differentiate between thin descending limb subtypes.
In the past two decades, many of the proteins which mediate water, urea, and sodium transport in nephron segments important for urinary concentration and dilution have been cloned ( Figure 43.6 ). The water channels (called aquaporins) and sodium transporters are discussed in detail elsewhere in this book. The urea transport proteins and their role in the long-term regulation of the urine concentrating mechanism are discussed later in this chapter.
In general, the water, urea, and sodium transport proteins are highly specific. Reflection coefficients are not included in Tables 43.2–43.5 , since the specificity of these transport proteins appears to eliminate a molecular basis for solvent drag and suggests that the reflection coefficients should be 1.
Thin descending limbs are conventionally divided into types I, II, and III: types I and II are located in the outer medulla in short and long looped nephrons, respectively, while type III limbs are located in the inner medulla. The osmotic water permeability of thin descending limb subtypes that express aquaporin 1 (AQP1) water channels is extremely high in all species studied ( Table 43.2 ). AQP1, a constitutively active water channel, is present in both the apical and basolateral plasma membranes in sufficient abundance to account for the measured rates of transepithelial water transport. Transgenic mice lacking the AQP1 channel (which is also found in proximal tubule and descending vasa recta), were found to have greatly impaired urine concentrating capability, which was attributed in large measure to defective water absorption from the proximal tubules and descending limbs, which may lead to an overloading of available concentrating capacity.
The chinchilla has an additional inner medullary subsegment (type II distal) in the deepest 20% of the longest loops of Henle. This subsegment has low osmotic water permeability and lacks AQP1. The Munich-Wistar rat has a similar subsegment in the deepest 60% of the longest loops and also a prebend segment of ~164 μm in length that labels for the ClC-K1 chloride transporter, but not for AQP1. Avian loops of Henle appear to have similar pre-bend segments. The pre-bend segments may be functionally important as a site of NaCl absorption.
Urea permeability varies in different portions of the thin descending limb ( Table 43.2 ). Urea permeability is relatively low in types I and II thin descending limbs. Urea permeability is higher in type III thin descending limbs, and is quite high in the chinchilla type III distal thin descending limb. The reflection coefficient for urea is close to 1 in thin descending limbs.
Sodium permeability is relatively low in rabbit types I and II and hamster types I and III thin descending limbs, but is relatively high in hamster type II and chinchilla types III and III distal thin descending limbs. Na + /K + -ATPase activity is very low in all thin descending limb segments in which it has been measured. Rabbit types I and II thin descending limbs have a NaCl reflection coefficient that is close to 1. However, the measured NaCl reflection coefficient is heterogeneous in hamster: 0.83 in type II and 0.99 in type III thin descending limbs.
The perfused tubule studies reviewed above provide important information about the transport properties of the individual nephron segments comprising the descending limb of the loop of Henle. In contrast, micropuncture studies provide in vivo information about the concentrations of solute within the portions of the descending limb that are accessible to micropuncture: comparisons can be made between the composition of tubular fluid near the ends of those proximal convoluted tubules that are accessible on the cortical surface and the composition of fluid in the bends of the longest loops of Henle near the papillary tip. Since proximal tubules on the surface of the kidney originate from superficial glomeruli, and the loops of Henle which reach the papillary tip generally originate from juxtamedullary nephrons, it is not possible, currently, to compare fluid samples taken from cortical and medullary sites in a single nephron. Thus, the validity of this comparison depends upon the assumption that the composition and delivery of solute and water to the beginning of the superficial and juxtamedullary descending limbs are similar. In addition, papillary micropuncture requires the removal of the ureter, which also reduces maximum urinary concentrating ability by a mechanism that is not completely understood. Thus, micropuncture studies are limited to studies performed during moderate, not maximal, antidiuretic conditions.
In the rat, osmolality increases along the length of the descending limb. Water removal accounts for 90% of this increase in osmolality in Brattleboro rats that are not treated with vasopressin. When Brattleboro rats are treated with vasopressin, there is an increase in the osmotic pressure of the descending limb fluid and in the volume of water absorbed from the descending limb. This rise in descending limb fluid osmolality results from water extraction (60%), and from urea addition (40%). The delivery of urea to the end of the thin descending limb averages 550% of the filtered load of urea. Thus, urea is either secreted into the descending limb fluid or there is a major difference between the filtered load of urea in superficial versus juxtamedullary glomeruli. When urea is infused into rats fed a low-protein diet, both water extraction from the descending limb and urinary concentrating ability are significantly increased.
In hamster, ~65% of the osmotically active solute in fluid obtained near the bend of the loop of Henle is due to sodium (plus a univalent anion), while ~20% is due to urea. Since only ~10% of the filtered load of water reaches the bend of the loop of Henle, the high luminal fluid sodium concentration results primarily from water extraction from the descending limb. Both sodium and inulin concentrations increase along the length of the hamster descending limb, showing that water is extracted from the descending limb fluid. As in the rat, significant amounts of urea are added to the descending limb fluid.
Psammomys obesus , a desert rodent, feeds on halophilic plants which provide water along with large quantities of NaCl. In these animals, tubular fluid flow rate decreases by 1.7-fold along the descending limb, while osmotic pressure increases four-fold. Water removal accounts for 40% of the increase and solute addition accounts for 60% under moderate NaCl-loading conditions. Unlike the rat, NaCl is the principal solute added to descending limb fluid ; urea is added but is much less important than in the rat. In Psammomys which are producing more highly concentrated urine (although still less concentrated than can be achieved by the intact animal), NaCl addition accounts for nearly 80% of the rise in osmolality.
The thin ascending limb ( Table 43.3 ) has an extremely low osmotic water permeability in all species studied, and no aquaporin proteins have been detected (reviewed in ). Although the thin ascending limb has a urea permeability that is lower than its NaCl permeability, it is significantly higher than the value that mathematical models indicate is required for the effective operation of the hypothesized passive mechanism ( vide infra ). While this is true in all species, it is especially true in chinchilla.
Species | ||||
---|---|---|---|---|
Chinchilla | Rat | Rabbit | Mouse | |
THIN ASCENDING LIMB | ||||
Na b | 238 | 80 | 26 | 55–88 |
Urea b | 171 | 14–23 | 7 | 19 |
Water c | 0–8 | 25 | 13 | 29 |
Na + /K + -ATPase d | 2–4 | 3 | ||
MEDULLARY THICK ASCENDING LIMB | ||||
Na b | 6 | 2 | ||
Cl b | 1 | 1 | ||
Urea b | 1.4 (outer stripe) | 1 | ||
0.6–0.9 (inner stripe) | ||||
Water c | 0 | 23 | ||
K b | 1 | |||
PD e | 2–3 | 3–7 | ||
Na + /K + -ATPase d | 41–139 (outer stripe) | 41–124 | 62 | |
260 (inner stripe) | ||||
CORTICAL THICK ASCENDING LIMB | ||||
Na b | 1 | 3 | ||
Cl b | 1 | 1 | ||
Urea b | 1.5 | 2 | ||
Water c | 0 | 23 | ||
PD e | 3–7 | |||
Na + /K + -ATPase d | 83–133 | 16–31 | 61 |
The thin ascending limb has a very low level of Na + /K + -ATPase activity that would not support a significant rate of active sodium transport. However, some in vivo studies have found evidence for active sodium transport in thin ascending limbs. The thin ascending limb has a high passive NaCl permeability. Chloride transport occurs transcellularly via the ClC-K1 chloride channel, which is present in both the apical and basolateral plasma membranes. Vasopressin increases chloride transport in thin ascending limbs, and water deprivation increases the mRNA abundance of ClC-K1. Transgenic mice lacking the ClC-K1 transporter were found to have greatly reduced urine concentrating capability, which was attributed to defective chloride transport in the thin ascending limb. Sodium transport is thought to occur paracellularly, since no apical plasma membrane sodium transport pathway has been demonstrated.
When rabbit thin ascending limbs are perfused in vitro with concentration gradients of NaCl and urea that simulate in vivo conditions (NaCl gradient from lumen-to-bath and a urea gradient from bath-to-lumen), they are able to dilute their luminal fluid by purely passive means. Perfusing rabbit thin ascending limbs in vitro with solutions whose osmolality is increased from 290 to 600 mOsm/kg H 2 O by adding NaCl to the perfusate and urea to the bath (to mimic the higher concentration of NaCl in the tubule lumen and the higher urea concentration in the medullary interstitium) reduces collected fluid osmolality to 70% of perfusate osmolality, suggesting that it may be possible to dilute the luminal fluid within thin ascending limbs without active transport in vivo .
Pannabecker et al. investigated inner medullary functional structure in Munich-Wistar rats by means of computer-assisted three-dimensional reconstructions of cross-sections in which tubules were identified and labeled by direct immunofluorescence of antibodies raised against specific transport proteins. The reconstructions indicate that thin descending limbs of Henle’s loops that have bends within the first millimeter below the outer–inner medullary boundary lack the water transporter AQP1. Thin descending limbs of loops that have bends beyond the first millimeter express AQP1 for about the first 40% of their length below the outer–inner medullary boundary, but beyond that point lack AQP1 expression. Expression of ClC-K1 chloride channels begins abruptly with a prebend segment of length ~165 μm, and ClC-K1 expression continues uniformly along the entire length of thin ascending limbs. Co-localization of AQP1 and ClC-K1 was not found in any loop of Henle segment. Preliminary sections show no evidence of expression of the urea transporters UT-A1, UT-A2 or UT-A4 in thin limbs below the first millimeter of the inner medulla. These observations are generally consistent with expression patterns indicated in other immunocytochemical studies in rat. However, Mejia and Wade found in Sprague-Dawley rats that ~30% of thin descending limbs that reached deep into the papilla labeled for AQP1 (in Brattleboro rats, ~11%); and Wade et al. found co-labeling of a UT-A urea transport protein and AQP1 in thin descending limbs in the base of the inner medulla of Brattleboro rats (these limbs may correspond to the longer population identified by Pannabecker et al. ).
Both the medullary and cortical portions of the thick ascending limb ( Table 43.3 ) have osmotic water permeabilities that are essentially zero, and neither subsegment expresses aquaporin proteins (reviewed in ). Thus, the primary mechanism for diluting the luminal fluid in thick ascending limbs is net absorption of solute, particularly NaCl. NaCl is actively absorbed by the Na + -K + -2Cl − co-transporter (NKCC2, BSC1) in the apical plasma membrane, and the sodium pump (Na + /K + -ATPase) in the basolateral plasma membrane. The thick ascending limb from short looped nephrons can lower the concentration of NaCl in the luminal fluid at loop bend from ~300 to ~117–40 mM at the cortico–medullary border, while the cortical thick ascending limb can lower the concentration of NaCl to ~32 mM. However, the medullary portion has the capacity to absorb more NaCl than the cortical portion, as evidenced by the higher Na + /K + -ATPase activity in the medullary thick ascending limb. The regulation of NaCl absorption in the thick ascending limb is discussed in detail in Chapter 34.
Vasopressin increases NaCl absorption in medullary and cortical thick ascending limbs in mouse. This response is consistent with vasopressin’s role in urinary concentration, and suggests that vasopressin can increase or maintain concentrating ability by increasing NaCl absorption across thick ascending limbs. However, vasopressin does not increase NaCl absorption in human and canine, and only weakly stimulates absorption in rabbit medullary thick ascending limbs.
Urea permeability in the medullary thick ascending limb is lower than in the cortical thick ascending limb. In rat, the transition to a higher urea permeability occurs between the inner and outer stripe portions of the medullary thick ascending limb, while in rabbit it occurs between outer medulla and cortex. Urea permeability in the thick ascending limb could permit dilution of tubular fluid by passive urea absorption or increase urea concentration in the thick ascending limb by secretion.
The cortical collecting duct has an extremely low osmotic water permeability (Table 4) in the absence of vasopressin. Vasopressin significantly increases the osmotic water permeability by a factor of 10 to 100 in both rat and rabbit. Arachidonic acid metabolites, produced by cytochrome P450, inhibit vasopressin-stimulated osmotic water permeability by a post-cyclic AMP (cAMP) mechanism ; the mechanism by which vasopressin increases osmotic water permeability is discussed in detail in Chapter 41 on Water Channels.
The cortical collecting duct has a low urea permeability that is unaffected by vasopressin. Thus, vasopressin-induced water absorption will increase the urea concentration within the lumen of the cortical collecting duct, and also the osmolality, provided that there is no significant net absorption of solutes.
The cortical collecting duct is the major site for aldosterone-mediated sodium absorption and potassium secretion. Vasopressin also stimulates sodium absorption in the cortical collecting duct. Sodium is actively absorbed via the epithelial sodium channel (ENaC) in the apical plasma membrane of principal cells, and sodium absorption is responsible for the generation of a lumen-negative voltage. Sodium exits the principal cell via Na + /K + -ATPase in the basolateral plasma membrane. Chloride is transported by both paracellular and transcellular pathways. Chloride absorption is primarily passive in rabbit, although some evidence for chloride absorption against an electrochemical gradient exists. Active chloride absorption occurs in rat, and is stimulated by vasopressin and inhibited by bradykinin.
Few permeability measurements exist for the rat outer medullary collecting duct ( Table 43.4 ). In rabbit, the outer medullary collecting duct has a low osmotic water permeability which is increased 20- or 30-fold by vasopressin. The urea permeability is low in the outer medullary collecting duct in both rat and rabbit.
Species | ||
---|---|---|
Rat | Rabbit | |
CORTICAL COLLECTING DUCT | ||
Na b | 0.1 | |
K b | 1–2 | |
Cl b | 2–5 | |
Urea b ± c AVP | 1 | 0–1 |
Water d −AVP | 17–43 | 4–13 |
+AVP | 389–994 | 166–280 |
Na + /K + -ATPase e | 13–81 | 12–23 |
OUTER MEDULLARY COLLECTING DUCT | ||
Na b | 0.39 | |
K b | 0.59 | |
Cl b | 0.5 | |
Urea b | 3.5 | 0.3 |
Water d –AVP | 14 | |
+AVP | 445 | |
Na + /K + -ATPase e | 11–41 | 8–19 |
a References are superscripted;
c ±AVP: value unchanged by AVP; −AVP: no vasopressin; +AVP: with vasopressin.
The inner medullary collecting duct (IMCD) was originally divided into three subsegments: IMCD 1 ; IMCD 2 ; and IMCD 3 . Subsequent studies showed that the inner medullary collecting duct could generally be viewed as consisting of two morphologically and functionally distinct subsegments: the initial IMCD (corresponding to the IMCD 1 ) and the terminal IMCD (corresponding to the IMCD 2 and IMCD 3 ). However, some recent studies have found functional differences between the IMCD 2 and IMCD 3 . Histologically, the rat initial IMCD (or IMCD 1 ) contains 90% principal cells and 10% intercalated cells ; the rat terminal IMCD (or IMCD 2 and IMCD 3 ) contains a unique cell type, the IMCD cell. Most of the permeability values available for IMCD subsegments are from the rat ( Table 43.5 ).
Species | |||
---|---|---|---|
Rat | Rabbit | Hamster | |
INITIAL INNER MEDULLARY COLLECTING DUCT – IMCD 1 | |||
Urea b ± c AVP | 2–5 | 1 | 8–9 |
Sodium-Urea d | 0 | ||
Water e −AVP | 16–81 | ||
+AVP | 148–460 | 534 | |
Na + /K + -ATPase f | 18–42 | ||
TERMINAL INNER MEDULLARY COLLECTING DUCT – IMCD 2 | |||
Na b | 1 | 2 | |
K b | 4 | ||
Cl b | 1–2 | ||
Urea b –AVP | 15–46 | 12 | 12 |
+AVP | 69–93 | 32 | |
+Hypertonic bath: | 120–143 | ||
+Hypertonic bath and AVP: | 163–190 | ||
Sodium-Urea g | 0–1 | ||
Water e −AVP | 70–333 | ||
+AVP | 208–749 | 646 | |
Na + /K + -ATPase f | 12–40 | ||
TERMINAL INNER MEDULLARY COLLECTING DUCT – IMCD 3 | |||
Na b | 1 | ||
Urea b −AVP | 39–49 | 13 | |
+AVP | 110 | ||
Sodium-Urea g | −9 | ||
Water e −AVP | 43–145 | ||
+AVP | 389–749 | ||
Na + /K + -ATPase f | 8–17 | ||
PAPILLARY SURFACE EPITHELIUM | |||
Chloride b ±AVP | 2–3 | ||
Urea b ±AVP | 1 | ||
Water d ±AVP | 14 |
a References are superscripted.
C ±AVP: value unchanged by AVP; −AVP: no vasopressin; +AVP: with vasopressin.
d sodium-urea co-transport, units: pmol/mm/min.
g sodium-urea counter-transport, units: pmol/mm/min; +: urea absorption; −: urea secretion.
In the absence of vasopressin, the initial IMCD has a low osmotic water permeability which is increased 10- to 30-fold by vasopressin. Urea permeability is low in the initial IMCD, and is unaffected by vasopressin. The initial IMCD from normal rats does not show any active urea transport.
The terminal IMCD has a higher basal (no vasopressin) osmotic water permeability than other portions of the collecting duct. Vasopressin can rapidly increase osmotic water permeability by a factor of 10. The terminal IMCD also has a higher basal urea permeability than other portions of the collecting duct. Vasopressin and hypertonicity can each increase urea permeability by a factor of 4–6, and together they can increase urea permeability by a factor of 10. Although early studies suggested a urea reflection coefficient of less than 1, more recent studies which re-measured the urea reflection coefficient and explicitly measured the dissipation of the imposed urea gradient, showed that the urea reflection coefficient equals 1. The IMCD 2 subsegment from normal rats does not show any active urea transport. However, active urea secretion, which is completely dependent upon luminal sodium, is present in the IMCD 3 from normal rats, suggesting that sodium absorption may be coupled to urea secretion.
Sodium and chloride permeabilities are low in the terminal IMCD. Micropuncture studies indicate substantial rates of NaCl absorption from the IMCD, but perfused tubule studies have been unable to detect it.
Only a few permeability coefficients have been measured across the papillary surface epithelium, and these have been measured only in rabbit ( Table 43.5 ). The urea and osmotic water permeabilities are low and unaffected by vasopressin. The basal chloride permeability is higher than that of the terminal IMCD, and is inhibited by vasopressin. The apical membrane of the papillary surface epithelial cell, which faces the urinary space, expresses a Na + –K + –Cl − co-transporter that is stimulated by vasopressin and inhibited by bumetanide. The basolateral membrane contains a potassium conductive pathway in rat and rabbit.
Since the late 1950s, the countercurrent multiplication hypothesis has been the generally accepted explanation for the generation of the osmolality gradient along the cortico–medulary axis in both the outer and inner medullas. This hypothesis holds that, at each level of the medulla, a small osmolality difference between tubular fluid flows in ascending and descending limbs is multiplied by the countercurrent flow configuration to establish a large axial osmolality difference. The principle of countercurrent multiplication is illustrated in Figure 43.7 . The loop shown in the figure panels may be identified with a short loop of Henle: the left channel is analogous to the descending limb, whereas the right channel is analogous to the thick ascending limb. The channels are separated by a water-impermeable barrier. Vertical arrows indicate flow down the left channel and flow up the right channel. Left-directed horizontal arrows indicate active transport of solute from the right channel to the left channel. The numbers within channels indicate local fluid osmolality. Successive panels represent the time course of the multiplication process.
Panel (a) of Figure 43.7 illustrates a loop with isosmolar fluid throughout. In panel (b), an active transport mechanism has pumped enough solute to establish a 20 mOsm/kg H 2 O osmolality difference between the ascending and descending flows at each level. This small difference, transverse to the flow, is called the “single effect.” Panel (c) illustrates the osmolality values after the fluid has advected (or carried) the solute half-way down the left channel and half-way up the right channel. In panel (d), the active transport mechanism has re-established a 20 mOsm/kg H 2 O osmolality difference, and the luminal fluid near the bend of the loop has attained a higher osmolality than in panel (a). By successive iterations of this process, a progressively higher osmolality is attained at the loop bend, and a large osmolality difference is generated along the flow direction. This is illustrated in panel (e), where the osmolality at the loop bend is nearly 300 mOsm/kg H 2 O above the osmolality of the fluid entering the loop. Thus, the “single effect” of a 20 mOsm/kg H 2 O difference has been multiplied axially down the length of the loop by the process of countercurrent multiplication.
In the outer medulla, countercurrent multiplication is believed to occur in the short loops of Henle by a process that is similar to that shown in Figure 43.7 . The tubular fluid of the proximal tubule that enters the outer medulla is isotonic to plasma (about 290 mOsm/kg H 2 O). That fluid is concentrated, as it passes through the pars recta and the thin desending limb, by osmotically driven water absorption; the absorption is driven by vigorous active transport of NaCl from the thick ascending limbs. At the bend of the loop of Henle, the tubular fluid osmolality attains an osmolality about twice that of blood plasma. Because the thick ascending limbs are nearly impermeable to water, its tubular fluid is diluted by NaCl absorption as it flows toward the cortex, so that the fluid emerging from this segment is hypoosmotic to blood plasma.
Countercurrent multiplication in the outer medulla, however, differs in important ways from the process illustrated in Figure 43.7 . In some of the most completely studied mammals, the descending and ascending limbs do not abut one another ; therefore, solute is not directly transported from ascending limbs to descending limbs. Rather, NaCl is pumped from thick ascending limbs to the interstitium, raising the osmolality of the interstitial fluid and the blood flowing through the vasa recta and capillaries. The increased interstitial osmolality withdraws water from thin descending limbs, and some NaCl may diffuse into thin descending limbs, thus raising the osmolality of descending limb fluid. The NaCl absorbed from ascending limbs and the water absorbed from descending limbs is carried to the cortex by the vasa recta, which, somewhat like the loops of Henle, are arranged in a countercurrent configuration. Thus, a large axial osmolality difference, from the cortico–medullary boundary to the boundary of the inner and outer medulla, is established in the loops of Henle, the vasculature, and the interstitium.
In addition, Figure 43.7 does not represent the flow in the collecting ducts. Some of the water and solute in thick ascending limb tubular fluid delivered to the cortex re-enters the outer medulla in the collecting ducts, and in the presence of vasopressin, sufficient water is absorbed from the collecting ducts, as a consequence of the hyperosmotic medullary interstitium, to bring collecting duct flow to near osmotic equilibrium with the surrounding interstitium. Thus, a large axial osmolality difference, similar to that in the thin descending limb, is established in collecting duct fluid.
Finally, the discrete, sequential process represented in Figure 43.7 does not arise under normal physiological conditions. Rather, the axial osmolality difference, or gradient, is sustained in near steady-state, much as indicated in panel (e), with a bend osmolality that is limited primarily by the rate of active transport, the diffusive back-leak of NaCl into the thick ascending limb, the length of the loop, the rate of water absorption from collecting duct flow, and the dissipative effects of the vasculature.
In recent years, some reasons have emerged for skepticism of the countercurrent multiplication hypothesis as the explanation for the axial osmolality gradient in the outer medulla. Several laboratories have reported evidence for the absence of AQP1 in significant portions of the terminal thin desending limbs of short loops of Henle. Moreover, two modeling studies have suggested that the osmotic load that is put on the concentrating mechanism in the outer medulla may be increased by water-permeable descending limbs, relative to water-impermeable limbs, and thus may reduce or eliminate the hypothesized concentration advantage of water absorption from descending limbs of short loops.
The axial osmolality gradient in the inner medulla has also been generally believed to be generated by the countercurrent multiplication of a small transverse osmolality difference, presumably between thin ascending and thin descending limbs. However, evidence for significant active transport from thin ascending limbs is lacking, and experiments indicate that the thin ascending limbs are highly permeable to both NaCl and urea. Thus, the inner medullary single effect must arise from a mechanism different from that found in the outer medulla. The roles of the vasculature and collecting duct are considered below; a more detailed treatment of countercurrent multiplication, and in particular, the concentrating mechanism of the inner medulla, is given in a subsequent section ( vide infra ).
The descending and ascending vasa recta are arranged in a counter-flow configuration connected by a capillary plexus. Vasa recta are freely permeable to water, sodium, and urea, and achieve osmotic equilibration through a combination of water absorption and solute secretion. Descending vasa recta gain solute and lose water, while ascending vasa recta lose solute and gain water. The exchange of solute and water between the descending and ascending vasa recta and the surrounding interstitium is called “countercurrent exchange.”
Efficient countercurrent exchange is essential for producing concentrated urine, because hypotonic fluid carried into the medulla and hypertonic fluid carried away from the medulla both tend to dissipate the cortico–medullary gradient of countercurrent multiplication. Thus, to minimize wasted work, fluid flowing through the vasa recta must achieve near osmotic equilibrium with the surrounding interstitium at each medullary level, and fluid entering the cortex from the ascending vasa recta must have an osmolality close to that of blood plasma. Conditions which decrease medullary blood flow, such as volume depletion, improve the efficiency of countercurrent exchange and urine concentrating ability by allowing more time for blood in the ascending vasa recta to lose solute and achieve osmotic equilibration. Conversely, conditions which increase medullary blood flow, such as osmotic diuresis, impair the efficiency of countercurrent exchange and decrease urine concentrating ability. For a detailed treatment of countercurrent exchange, see Chapter 24.
The collecting duct, under the control of vasopressin and other factors, is the nephron segment responsible for final control of water excretion. Whereas the osmolality gradient along the cortico–medullary axis, in both the outer and inner medulla, presumably arises from mechanisms that principally involve participation of the loops of Henle, and countercurrent exchange in the vasa recta minimizes the dissipative effect of vascular flow, the excretion of water requires another structural component, the collecting duct system, which starts in the cortex and ends at the papillary tip. In the absence of vasopressin, the cortical, outer medullary, and initial inner medullary portions of the collecting duct are nearly water-impermeable. (The terminal IMCD has a moderate water-permeability even in the absence of vasopressin ( vide supra ).) Since the fluid that leaves the thick ascending limb and enters the cortical collecting duct is dilute relative to plasma, excretion of dilute urine only requires that not much water be absorbed nor much solute be secreted along the collecting duct.
In the presence of vasopressin, the entire collecting duct becomes highly water-permeable. This process takes place in the following way. Plasma osmolality increases when a person or an animal becomes water depleted. Osmoreceptors in the hypothalamus, which can sense an increase of only 2 mOsm/kg H 2 O, stimulate vasopressin secretion from the posterior pituitary. Vasopressin binds to V 2 -receptors in the basolateral plasma membrane of principal and IMCD cells in the collecting duct, stimulates adenylyl cyclase to produce cAMP, activates protein kinase A (PKA), phosphorylates aquaporin-2 (AQP2) at serines 256, 261, 264, and 269, inserts AQP2 water channels into the apical plasma membrane, and increases water absorption across the collecting duct ( and reviewed in ). This regulated trafficking of AQP2 between subapical vesicles and the apical plasma membrane is the major mechanism for acute regulation of water absorption by vasopressin (reviewed in ). Wade and colleagues originally proposed the “membrane shuttle hypothesis,” which proposes that water channels are stored in vesicles and inserted exocytically into the apical plasma membrane in response to vasopressin. Since the cloning of AQP2, the shuttle hypothesis has been proven experimentally in rat inner medulla (reviewed in ). Subsequent studies have elucidated several signal transduction pathways that are involved in regulating AQP2 trafficking (insertion and retrieval of AQP2), the role of vesicle targeting proteins (SNAP/SNARE system), and the cytoskeleton (reviewed in ); these processes are discussed in more detail in Chapter 41.
Vasopressin-induced water permeability allows water to be absorbed across the collecting ducts at a sufficiently high rate for collecting duct tubular fluid to attain near osmotic equilibration with the hyperosmotic medullary interstitium; the absorbed water is returned to the systemic circulation via the ascending vasa recta. The majority of water is absorbed from collecting ducts in the cortex and outer medulla. Although the inner medulla has a higher osmolality than the outer medulla, its role in absorbing water from the collecting duct is important only when maximal water conservation is required. More water is actually absorbed across the IMCD during diuresis than antidiuresis, owing to the large transepithelial osmolality difference.
The conceptual history of the urine concentrating mechanism may be divided into three periods. The first period, extending from 1942 through 1971, was inaugurated by the publication of a study by Kuhn and Ryffel, who proposed that concentrated urine is produced by the countercurrent multiplication of a “single effect,” and who constructed a working apparatus that exemplified the principles of countercurrent multiplication. During this first period, the theory of the countercurrent multiplication hypothesis was developed further, and experimental evidence accumulated that supported the hypothesis as the explanation for the concentrating mechanism of the outer medulla. In particular, active transport of NaCl from thick ascending limbs was identified as the source of the outer medullary single effect.
The second period of conceptual history, extending from 1972 through 1992, was inaugurated by the simultaneous publication, by Kokko and Rector and by Stephenson, of papers proposing that a “passive mechanism” provides the single effect for countercurrent multiplication in the inner medulla. According to the passive mechanism hypothesis, a net solute efflux from thin ascending limbs results from favorable transepithelial NaCl and urea gradients; these gradients arise from the separation of NaCl and urea, which is largely driven by the outer medullary concentrating mechanism. Although initially much experimental evidence appeared to support the passive mechanism, findings from many subsequent studies are difficult to reconcile with this hypothesis. Moreover, mathematical models incorporating measured transepithelial permeabilities failed to predict a significant inner medullary concentrating effect. The discrepancy between the consistently negative results from mathematical modeling studies and the very effective inner medullary concentrating effect has persisted through more than three decades. The discrepancy has helped to stimulate research on the transport properties of the renal tubules of the inner medulla and the formulation of several highly sophisticated mathematical models (notably, ), but no model study has resolved the discrepancy to the general satisfaction of experimentalists and modelers.
In the early 1990s, new hypotheses for the inner medullary concentrating mechanism began to receive serious consideration, and a third period of conceptual thought may be considered to have begun in 1993: in that year, Knepper and colleagues proposed a key role for the peristalsis of the papilla, and in 1994 Jen and Stephenson examined the principle of “externally driven” countercurrent multiplication, arising, e.g., by the net production of osmotically active particles in the interstitium. At about the same time, perfused tubule studies in chinchilla, which can produce very highly concentrated urine, provided evidence that the passive mechanism, as originally proposed, cannot explain the inner medullary concentrating mechanism. Recent studies have sought to further develop hypotheses involving peristalsis, the potential generation of osmotically active particles, especially lactate, and the role of complex inner medullary anatomy and detailed transporter localization. In 2004, evidence suggesting an absence of significant urea transport proteins in loops of Henle reaching deep into the medulla led to a reconsideration of hypotheses related to the passive mechanism.
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here