Function of the Juxtaglomerular Apparatus: Control of Glomerular Hemodynamics and Renin Secretion

Morphology

The MD cells form an elliptical plaque of epithelial cells located at the distal end of the thick ascending limb, approximately 100 to 200 µm upstream from the transition to the distal convoluted tubule. An MD plaque has been reported to consist of 14 cells in rat and about 25 cells in rabbit. Cellular plasticity is suggested by the finding that this number increased by about 30% following chronic angiotensin II receptor blockade, probably by transdifferentiation of adjacent TAL cells. MD cells are morphologically characterized by a high nucleus-to-cytoplasm ratio, absence of basal infoldings, and numerous mitochondria that are typically not in contact with the basal membrane. The basement membrane is thinner than that found in other areas of the tubule, and shows discontinuities in scanning electron micrographs. The difference in basement membrane appearance is paralleled by a macromolecular composition that differs from that of adjacent TAL cells.

NaCl and Water Movement

Although morphologically distinct, MD cells and neighboring thick ascending limb (TAL) cells share similar NaCl transport mechanisms ( Figure 23.2 ). As in TAL cells, NaCl uptake is mostly through the apical Na,K,2Cl co-transporter (NKCC2 _/ BSC1). Conventional electrophysiology and patch-clamp evidence has established its presence functionally. Presence of the NKCC2 co-transporter has also been shown at mRNA and protein expression levels. Of the three full length isoforms of the co-transporter, both the A and the B types are expressed in the MD. In contrast to TAL cells, in which NKCC2 phosphorylation is strongly enhanced by vasopressin, MD cells have a constitutively high presence of phospho-NKCC2 even in the complete absence of vasopressin. From the rates of MD cell acidification by luminal ammonium it has been concluded that apical NKCC2-mediated flux rates are not markedly different from those in TAL cells. Apical membranes of MD cells are rich in low conductance K channels of the ROMK type that are required for K recycling. Na/H exchange through NHE2 provides a second pathway for a smaller fraction of Na uptake by MD cells. In the rabbit, the apical membrane may also be a site of active Na extrusion, since the luminal presence of ouabain has been found to elevate intracellular Na concentration in both MD and TAL cells; in addition, luminal ouabain prevented the recovery of intracellular Na from the elevated levels resulting from increased luminal NaCl. Apical Na efflux in MD cells may be mediated by a luminal H,K-ATPase, as colonic H,K-ATPase has been shown previously to mediate active Na efflux. Its presence in MD cells is supported by immunocytochemical and functional evidence.

Na,K-ATPase abundance and activity in the basolateral membrane of MD cells in rabbit kidneys appears to be low compared to neighboring TAL cells, a finding that is probably mainly due to the absence of basolateral membrane infoldings. In the rat, basolateral membranes of MD cells identified by nNOS counterstaining clearly express α1 Na,K-ATPase, together with β1- and γ-subunits. Cl exit across the basolateral membrane occurs through abundant Cl channels. Immunocytochemical evidence indicates the presence of the AE2 anion exchanger in basolateral membranes of MD cells in both the rat and the mouse, and Cl/HCO ₃ exchange activity has been observed in isolated rabbit JGA preparations. Together with the apical Na/H exchanger, basolateral AE2 may play a role in the absorption of HCO ₃ or it may act as a pHi-controlling housekeeping gene.

The effect of changes in luminal fluid composition on the volume of macula densa cells has remained a controversial issue. At constant luminal osmolarity of around 300 mOsm, changes in luminal NaCl concentration caused parallel changes in the volume of MD cells. Changes in volume were to some extent transient, indicating some ability of MD cells for volume regulation. Both increases and decreases in MD cell volume have been observed with concomitant increments in Na (25 to 135 mM) and osmolarity (210 to 300 mOsm). When NaCl concentration was kept constant, MD cells behaved like osmometers, swelling with a reduction and shrinking with an increase in osmolarity. Transcellular osmotic water permeability of MD cells, assessed from the initial cell volume change in response to an osmotic step change, was estimated to be similar to that of cortical collecting tubules in the absence of ADH. The main restriction to water movement resides in the apical membrane. Transmembrane channels for water movement have not been identified; apical membranes of MD cells lack aquaporin 1, and the presence of other aquaporins has not been established. Vasopressin receptors of both the V1a and V2 varieties have been found in MD cells, but there is no evidence that MD water permeability is regulated.

Other Cellular Characteristics

Nitric oxide synthase type I (NOS I or nNOS), a constitutive and Ca-dependent NOS isoform, is highly and selectively expressed in MD cells, and has become a useful marker of this cell type. Alternative splicing leads to the generation of several nNOS variants with NO synthase activity. The presence of three of these variants, nNOSα, β, and γ in MD cells and their regulation by salt intake has recently been demonstrated. Catalysis of the conversion of L-arginine to NO and L-citrulline by NOS requires the participation of a number of co-factors. One of these co-factors is NADPH, and it is possible that the relatively high activity of glucose-6-phosphate dehydrogenase (G6PDH) in MD cells is related to the NADPH requirement of NOS. A functional connection between nNOS and G6PDH is suggested by the parallel upregulation of the expression of both enzymes during NaCl restriction. Alternatively, the high pentose shunt activity suggested by the abundance of G6PDH may serve to provide ribose-5-phosphate for nucleic acid synthesis. Avid uptake of labeled uridine has been shown to occur in MD cells, a process inhibited by actinomycin D, and therefore indicative of incorporation of the pyrimidine precursor into the RNA pool.

Cyclooxygenase-2 (COX-2), typically induced by LPS and cytokines in the inflammatory process, is constitutively expressed both in the renal medulla and to a lesser extent in the cortex. Cortical expression of COX-2 in the mature rat and rabbit kidney is restricted to a subgroup of MD and perimacular cells of the TAL. MD expression of COX-2 has also been observed in humans older than 60 years, and in patients with Bartter syndrome. Condensation of COX-2 expression is the remnant of a more intense expression in early postnatal kidneys, where the enzyme can be found in a more contiguous pattern in TAL cells proximal and distal of the MD. At this early stage, COX-2 appears to be specifically excluded from MD cells. Conversion of PGH ₂ into the bioactive PGE ₂ is catalyzed by prostaglandin E2 synthases, of which both a microsomal and cytosolic isoform have been described. Immunocytochemical evidence has demonstrated the presence of a membrane-associated PGE ₂ synthase in MD cells of both rats and rabbits. Co-localization of COX-2 with phospholipase A2 has been described in the MD at the level of single cells. Table 23.1 lists a number of other differences in the expression pattern of MD compared to surrounding TAL cells where the function has remained unclear.

Morphology

The EGM cells (Goormaghtigh or lacis cells are synonyms) are the cells of the JGA which have the most intimate and regular contact with the MD. MD cells and EGM cells are separated by an interstitial cleft of variable width that does not appear to be bridged by gap junctional connections. In three-dimensional reconstructions EGM cells are elongated cells with long cytoplasmic processes, which in general run parallel to the base of the MD cells. Commensurate with extensive gap junctional coupling of EGM cells with each other as well as with mesangial cells and granular cells, connexins 40, 37, and perhaps 43 or 45 are expressed to various degrees in extra- and intraglomerular mesangium. The presence of myofilaments in EM cells suggests that EGM cells, like mesangial cells, have contractile potential.

The extraglomerular mesangial cell field is free of capillaries, lymph terminals or nerve fibers. The absence of blood capillaries may cause a retardation of fluid entry and fluid removal from this compartment. In fact, the interstitial volume density of the EGM cell field increased from 17% during volume depletion to 29% during volume expansion, while no changes were noted in the peritubular interstitium. Nevertheless, recent studies of the flow dynamics across the JGA interstitium using lucifer yellow as a fluorescent marker indicate rapid exchange between afferent arterioles and tubular lumen, presumably mediated by bulk fluid flow.

Biochemical and Functional Aspects

Localization studies using histochemical, autoradiographic or immunological methods usually do not distinguish between intra- and extraglomerular mesangial expression patterns. Nevertheless, in some cases it seems justified to assume parallel expression in both cell types. For example, autoradiographic localization of angiotensin II and atrial natriuretic factor binding suggests the presence of receptors on both intra- and extraglomerular mesangial cells. Relative predominance of AT1 receptor mRNA in EGM cells has been observed by in situ hybridization. While EGM cells normally do not synthesize renin, they can be recruited to form renin with long-standing stimulation such as chronic diuretic abuse.

Differential expression patterns in intra- and extraglomerular mesangium are relatively discrete. EGM cells do not stain with antibodies against Thy-1, while the glomerular mesangium does. Conversely, decay accelerating factor (DAF), a glycoprotein that limits complement activation on cell surfaces, is restricted to the EGM cells, at least in the human kidney. HSP 25 expression has also been reported in extraglomerular, but not intraglomerular, mesangium. Of unknown significance is the observation that two Na,K-ATPase-associated proteins, the FXYD protein phospholemman (FXYD1) and the β2-subunit, are expressed in the extraglomerular mesangium, while being excluded from both MD cells and intraglomerular mesangial cells. Conventional mesangial cell cultures have been used as a model to study JGA-specific issues such as NO and PGE ₂ production, but whether or not this approach permits inferences about the JGA signaling mechanisms in vivo has remained unclear.

Morphology

The granular cells in the arteriolar walls are the main renin-producing cells of the kidney. With a rough endoplasmic reticulum, a well-developed Golgi apparatus, and numerous cytoplasmic granules, they have the fine structure of protein-secreting cells. The renin-containing granules are membrane-bound. Some granules, believed to be the more newly formed, have a crystalline lattice appearance and may mainly contain pro-renin; others, with an amorphous electron-dense content, are believed to represent the mature form. Myofibrils and smooth muscle myosin are sparse, and may be even absent in granular cells at the vascular pole. In the mature rat kidney under control conditions, granular cells are clustered at the vascular pole over a length of about 30 µm or about 20% of the afferent arteriole, but single ring-like renin-positive regions in more proximal locations are sometimes seen. In the developing kidney, as well as during stimulation of renin synthesis, for example with converting enzyme blockade, renin-positive cells can be found all along the afferent arteriole and also in larger vessels.

Coexistence of renin and angiotensin II in granules of rat and human epithelioid cells has been shown by light and electron microscopy. Granular angiotensin II appears to increase in parallel to renin following adrenalectomy and renal artery stenosis. Granular angiotensin II may reflect uptake through either non-specific endocytosis, receptor internalization or intracellular de novo generation. Not unexpectedly, granular cells contain AT1 receptor mRNA, with rats expressing both AT1A and AT1B receptor mRNA and mice expressing only AT1A receptor mRNA. Granular cells express the mRNA for both D1-like and D2-like dopamine receptors mediating stimulation and inhibition of renin secretion. The gap junctional connexins 40 and 37 have been shown to co-localize with renin, indicating functional connections among granular cells. Other proteins found in JG cells include cyclic guanylate kinase II, the ubiquitous basolateral form of the Na,2Cl,K co-transporter NKCC1/BSC2, and GLUT4.

Functional Aspects

Renin release has been found to be episodic or quantal, an observation most consistent with granule exocytosis. Nevertheless, EM images rarely document the classic omega configuration with an open pore to the cell exterior, and no evidence in support of the presence of vesicle or target membrane SNARE proteins in JG cells has been published. On the other hand, isoproterenol and cAMP caused an increase in membrane capacitance in isolated JG cells, an observation usually interpreted to be the result of an exocytotic membrane fusion event. An attempt to observe exocytosis has been made in dissected glomerulus/vessel preparations using optical labeling of renin granules with quinacrine and LysoTracker-Red, fluorophores that are taken up into acidic organelles. When stimulated by isoproterenol or a low arteriolar pressure, labeled granules disappeared at a rate of about 5–10 granules per minute. In renin-releasing As4.1 cells, the extinction of individual granules was followed by the appearance of an extracellular quinacrine cloud, presumably representing the released granule contents.

Studies of the membrane characteristics of JG cells in the hydronephrotic mouse kidney by the whole cell patch-clamp technique have identified an inward rectifying K current whose inhibition was shown to be partly responsible for the depolarizing effect of angiotensin II. In addition, JG cells expressed Ca-activated Cl channels in high density. In contrast, inwardly rectifying K channels were not detected in isolated JG cells from rat kidneys. Instead, Ca-dependent and voltage-gated large conductance K channels (BK _Ca ) were identified that largely determined the resting potential of −32 mV. Presence of BK _Ca channels was verified at the mRNA level by RT-PCR, and at the protein level by immunocytochemistry. The increased outward current caused by cAMP was also due to activation of BK _Ca channels, suggesting that they were of the cAMP-stimulated ZERO splice variant. There is also evidence for the presence of K-ATP channels in JG cells, but their functional role is not clear. While earlier studies failed to obtain functional evidence for the presence of voltage-dependent Ca channels, the presence of L-type Ca channels and their activation by strong depolarizations has been established in isolated JG cells. JG cells express NKCC1, the ubiquitous isoform of the Na,K,2Cl co-transporter, and its inhibition by furosemide stimulates renin exocytosis, as evidenced by increased membrane capacitance.

Effect of Distal Tubule Flow Perturbations on SNGFR

The tubuloglomerular feedback (TGF) response is defined as the change of SNGFR resulting from a change in tubular fluid flow exiting the proximal convoluted tubule, a practical experimental variable to predictably alter tubular fluid composition in the MD segment of the tubule. The average reduction of SNGFR in superficial nephrons of rats caused by a saturating flow increase in 15 independent studies was 13 ±1 nl/min or 40 ±3%. In addition to the rat, TGF responses were found in all mammalian species tested thus far (dog, hamster, mouse, humans), as well as in two non-mammalian species ( Amphiuma means and Necturus maculosus ). Fitting SNGFR measurements at eight different loop perfusion rates to a four parameter logistic equation ( Figures 23.3 and 23.4 ) revealed that TGF responses occur over a defined flow range and show nonlinear saturation kinetics. V _½ , the flow resulting in the half-maximum response, was 17.5 nl/min, a value close to the ambient end-proximal flow rate in the rat ( Figure 23.3 ). The precise location of the TGF operating point has been determined by adding or withdrawing small volumes of fluid from the proximal tubule and determining the resulting changes in proximal flow rate. In these studies, small increases and decreases in loop flow rate were equally and maximally effective in altering SNGFR , an observation that directly demonstrates the position of the operating point at the midpoint of the feedback function curve. Consistent with the conclusion of tonic suppression of GFR by TGF are observations showing that SNGFR based on fluid collections in the proximal tubule where the TGF signal is eliminated is usually higher than SNGFR based on distal collections where the TGF signal is intact ( Figure 23.5 ). Systematically higher values of SNGFR of superficial nephrons determined in proximal compared to distal tubule segments have also been demonstrated in the dog and mouse. TGF responses were also observed in juxtamedullary nephrons of both rats and hamsters, where increased flow past the MD by perfusion of thin ascending limbs produced a reduction in SNGFR by about 25 nl/min or approximately 50%.

The vasoconstriction elicited by TGF at the level of the JGA may be partially offset by a tubular vasodilator effect that appears to be mediated by a tubulovascular contact area at the level of the connecting tubule. This dilator mechanism, called cTGF, is activated by high Na concentrations and inhibited by amiloride or benzamil, suggesting that it is initiated by activation of ENaC-dependent Na transport. Activation of cTGF seems to require relatively high flow rates, and may therefore play an important role only under special circumstances. Nevertheless, the implications of these observations are potentially far-reaching for the interpretation of previous microperfusion studies, and for the understanding of the physiology of TGF regulation of GFR.

TGF Oscillations

In both rats and mice the operation of the TGF system in the closed-loop mode can result in stable oscillations of filtration pressure and filtration rate with a periodicity of 2–3 cycles/min (30–50 mHz). Synchronous pressure oscillations were seen in efferent arteriolar blood flow, with blood flow leading tubular pressure by about a 1 second lag. This phase shift suggests that oscillations in blood flow were the cause of changes in tubular pressure. Oscillations were principally of single nephron origin, since oscillations in random nephrons were not in phase. However, synchronized pressures were observed in adjacent nephrons whose afferent arterioles originate from the same interlobular artery. The simultaneous assessment of the oscillatory pattern of many nephrons on the surface of the kidney using the novel approach of laser speckle contrast imaging has shown that synchronized TGF oscillations can sometimes be observed among nephrons that are not located in the immediate vicinity of each other. Pressure oscillations were abolished by loop diuretics, and they were absent in mice that lack TGF responses, suggesting that they were generated by the TGF system. This contention was further supported by the finding that distal flow rate and Cl concentrations oscillate with the same frequency, but with a fixed phase shift. Mathematical modeling of the TGF system indicates that oscillations are the result of a relatively high feedback gain, in combination with delays in the transmission of the signal across the JGA and along the nephron.

In addition to the slow TGF-dependent oscillations, laser-Doppler velocimetry identified oscillations in star vessel blood flow with a frequency of about 100–200 mHz, probably reflecting myogenic vessel activity. Since TGF and myogenic mechanisms are targeted to identical arteriolar smooth muscle cells, they are expected to interact and become synchronized. As an expression of the interaction, the power of the myogenic oscillations increased during inhibition of TGF, and decreased during TGF saturation. In contrast to the synchronized oscillations of normal animals, irregular fluctuations of proximal tubular pressure have been observed in spontaneously hypertensive as well as Goldblatt hypertensive rats, but not in salt-sensitive Dahl rats with hypertension. Mathematical modeling has shown that desynchronizations like those seen in the hypertensive models can result from parameter variations of the TGF system or from increases in interaction strength, particularly when nephrons are electrotonically coupled. Oxygen tension on the renal surface has also been found to oscillate at the 30 mHz TGF frequency, and it has been speculated that a switch of TAL cell energy production from aerobic to anaerobic metabolism may cause the instability in MD NaCl concentration that activates TGF oscillations.

Effect of Loop of Henle Flow on MD NaCl Concentration

Although flow rate changes are frequently used to activate TGF, their mechanical consequences per se do not appear to be sufficient cause for vasomotor responses in vivo . Flow rate changes do not elicit TGF responses as long as they are not accompanied by changes in NaCl, and conversely, full TGF responses can be induced by low flows, as long as NaCl is supplied at sufficiently high concentrations. Furthermore, widely varying TGF responses can be observed at identical flow rates. However, in a recent study using the isolated perfused JGA preparation, changes in flow even in the absence of NaCl were observed to elicit vasoconstriction and increases of cytosolic calcium in vascular cells at the glomerular pole, an effect attributed to the consequences of mechanical deflection of the central cilium of MD cells. The reasons for this major discrepancy between the in vivo and in vitro effects of flow are not known.

Nevertheless, extensive experimental evidence in vivo favors the notion that the MD cells respond to changes in luminal NaCl concentration, and that the flow dependency of TGF responses reflects flow dependence of luminal [NaCl] in the MD region of the nephron. In situ microperfusion of loops of Henle has revealed a biphasic relationship between flow rate and distal [NaCl] measured 300–600 μm downstream from the MD, the earliest accessible site along the distal convoluted tubule. The increase of distal solute concentrations at subnormal flow rates is the result of modifications of tubular fluid between the MD region and the distal tubule. NaCl influx along the early post-MD epithelium causes [NaCl] to increase over the levels existing at the MD, and the effect of this addition of NaCl is particularly evident at low flow rates.

Effect of MD NaCl Concentration on SNGFR

The precise relationship between MD NaCl concentration and SNGFR was established by perfusing loops of Henle from their distal ends in a retrograde direction. In this approach, the distance between perfusion and sensing sites is greatly shortened, the changes in perfusate composition by tubular transport activities are minimized, and the effects of perfusate composition on SNGFR can be studied at constant flow rate and pressure. At a flow rate of 20 nl/min, SNGFR varied inversely with changes in perfusate NaCl concentration between 15 and 60 mM (or 30 and 120 mOsm), values which extend over the hypotonic range normally occurring at the end of the thick ascending limb. Increments in NaCl concentration above 60 mM did not further suppress filtration rate. Maximum changes of SNGFR caused by saturating flow rates during orthograde perfusion and by saturating NaCl concentrations during retrograde perfusion were identical. Fitting the equation of a hyperbolic tangent to these results ( Figure 23.6 ) indicates that the half-maximum decrease in SNGFR is caused by a NaCl concentration of 33.5 mM, and that the maximum slope is about 0.5 nl/min mM.

Studies designed to discriminate between ionic or osmotic effects of the perfusion fluid indicate that total solute concentration at the MD does not seem to measurably participate in TGF-mediated reductions of SNGFR . Orthograde perfusion with isotonic mannitol solutions in the rat is usually not associated with sustained reductions in SNGFR , even though distal tubular fluid osmolality is greatly increased. TGF responses correlating with alterations in osmolality have been observed during orthograde perfusion with various perfusion solutions, but the variations in distal osmolality were outside the critical osmolality range of 30 to 120 mOsm observed in retrograde perfusion studies. In retrograde perfusion experiments in which fluid osmolality and NaCl concentration were varied independently, TGF responses were exclusively determined by NaCl concentration, and not by osmolality in a range between 130 and 400 mOsm. Finally, the pattern of SNGFR responses during retrograde perfusion with isotonic solutions in which either Na or Cl was replaced, but in which osmolality was kept constant, indicates dependence on the ionic composition and independence of osmolality.

TGF Response and NaCl Transport

Effects of Loop Diuretics

The observation that inhibition of NaCl transport along the loop of Henle is associated with blockade of the TGF mechanism has been of fundamental importance in understanding the initiation of the TGF signaling pathway. TGF inhibition has been rather uniformly observed in the presence of loop diuretics such as furosemide, bumetanide, piretanide, ethacrynic acid, triflocin or l-ozolinone. Concentrations causing half-maximal inhibition of transport and feedback appear to be similar, about 5×10 ⁻⁵ M for furosemide and about 10 ⁻⁶ M for bumetanide (own unpublished data). Furosemide also blocked TGF responses during retrograde perfusion, suggesting that metabolic consequences of TAL inhibition are not transmitted by convective transport to the MD cells. Since distal Na and Cl concentrations are greatly elevated during loop NaCl transport inhibition, TGF responses do not appear to be caused by luminal NaCl concentration changes per se , but by changes in cellular NaCl uptake mediated by the furosemide-inhibitable Na,K,2Cl co-transporter NKCC2. The concentration dependence of feedback responses is probably the result of concentration dependence of NaCl uptake. Studies in mice with selective deletions of the A or B isoform of NKCC2 indicate that NKCC2B mediates TGF in the low NaCl concentration range, while NKCC2A is required for responsiveness to higher NaCl concentrations. Thus, the presence of two isoforms of NKCC2 in the macula densa extends the NaCl range over which TGF operates ( Figure 23.7 ). The concept that TGF responses are generated by the successive activation of NKCC2B and NKCC2A is supported by expression studies in Xenopus oocytes that have shown a higher Cl affinity of NKCC2B than NKCC2A, 9 mM versus 45 mM. Other aspects, such as the dependence of the inhibitory potency of furosemide on Cl concentration, have also been found to hold true for the TGF response. Diuretic agents with primary actions outside the loop of Henle such as acetazolamide, chlorothiazide, and amiloride do not possess TGF-inhibitory properties.

Ion Substitution Studies

That NaCl uptake by NKCC2 is the initial step in the feedback transmission pathway is further supported by parallels in the ionic requirements for both TGF- and NKCC2-mediated NaCl transport across TALH. During retrograde perfusion of the MD segment ( Figure 23.8a ), TGF responses were not seen during perfusion with isotonic or hypotonic solutions of Na salts such as NaHCO ₃ , NaNO ₃ , NaI, NaSCN, Na acetate, Na gluconate or Na isethionate. In contrast, isotonic solutions of Cl salts ( Figure 23.8b ) accompanied by small monovalent cations such as K, Rb, Cs or NH ₄ elicited full TGF responses, as did the bromide salts of Na and K ¹⁴⁸ . It is to be noted that some of these small cations have been found to be substrates for either the Na or the K site on the NKCC2 co-transporter.

The requirement for sizable Cl or Br concentrations, and the apparent lack of dependence on Na concentration, are consistent with an involvement of Na,K,2Cl co-transport, since the apparent overall affinity of NKCC2 for Cl in both TAL and MD cells is much lower than that for Na or K. Thus, the relatively low Cl affinity would predictably create an apparent Cl dependency of transport, while the small amounts of Na or K entering initially Na- or K-free solutions are sufficient to sustain near normal NKCC2 activity. When Na was replaced with large cations such as choline or TMA, TGF responses of normal magnitude were not seen, even though Cl was present in sufficiently high concentrations. Considering the cationic selectivity of the paracellular shunt, replacing luminal Na for choline will result in a sizable lumen positive Na diffusion potential that is predicted to reduce NaCl absorption by increasing Cl backflux. This explanation is consistent with the observation that NaCl transport rates of isolated TAL were found unaltered when studied under symmetric conditions with high choline Cl on both sides of the epithelium. In this case, Na backdiffusion and voltage-dependent inhibition of NaCl absorption is not to be expected as long as low concentrations of Na and K are present.