Search

As a consequence of its ability to absorb salt in excess of water, the thick ascending limb of the mammalian kidney dilutes the urine and supplies the energy for counter current multiplication. This latter effect follows directly from the ability of the medullary thick ascending limb (mTALH) to enrich medullary osmolality. In this review, we consider certain selected aspects of mTALH function that determine its ability to dilute the urine and to contribute to overall renal concentrating power. Specifically, we shall review the mechanisms for salt absorption in the mTALH and the modulation of salt absorption in the mTALH and hence urinary concentrating power, by antidiuretic hormone (ADH), prostaglandin E2 (PGE2) and peritubular hypertonicity. Furthermore, we shall advance an explanation of how these latter three agents modulate mTALH function without affecting external salt balance.

These experiments in isolated mouse cortical thick ascending limbs (cTALH) provide information about: the relative contributions of cellular and paracellular pathways to the transepithelial electrical conductance Ge (mS cm-2); the effects of (CO2 + HCO3-) on Ge; the ratio of net K+ secretion (JKnet) to net Cl- absorption (JClnet); and the K+ requirement for apical membrane furosemide-sensitive NaCl entry. The combination of luminal Ba++, zero K+ reduced Ge; at 5 mM luminal Ba++, the residual conductance was about 75% of the control Ge. The Ba++-insensitive Ge of 73.1 +/- 6.4 mS cm-2 was only slightly greater than the shunt conductance of 57 mS cm-2 computed from the sum of the dissipative bath to lumen fluxes of 22Na+ and 36Cl-. Moreover, luminal 5 mM Ba++, zero K+ had no effect on the Na+/Cl- permselectivity ratio of the paracellular pathway. Thus, Ba++ blockaded transcellular conductance by blocking apical membrane K+ channels. The combination of (CO2 + HCO3-) in external solutions increased Ge solely by augmenting the Ba++-sensitive, that is, transcellular, component of Ge. Finally, in paired experiments, the ratio JKnet/JClnet was 0.27 +/- 0.05; and both Ve, the spontaneous transepithelial voltage (mV), and the equivalent short circuit current Je (pEq sec-1 cm-2) were reduced dramatically by luminal K+ omission. Thus, apical membranes of the cTALH appear to contain a pathway for net K+ secretion, and apical membrane NaCl entry may involve co-transport with K+. Transcellular conductance contributes at least 25% to the total Ge, and the combination (CO2 + HCO3-) augments transcellular conductance.

We have explored evidence that suggests that lateral intercellular spaces is the mammalian proximal nephron do not serve as a hypertonic "central compartment" driving volume absorption. A primary consideration is the very low transepithelial resistance of this tissue as demonstrated by several laboratories. By making the reasonable assumption that passive ion permeation occurs via a paracellular route, we have concluded that the diffusion resistance of the spaces in insufficient to allow the development of a significant compositional difference between the spaces and the peritubular medium. This conclusion led us to look for potential osmotic gradients existing between the luminal and peritubular solutions. From the perfusion rate dependence of osmotic volume flow in the absence of active transport in isolated convoluted and straight proximal tubules, we calculated that both segments have very high hydraulic conductances, on the order of 3,000-5,000 micron/sec. Consequently, slight differences in the effective osmolality of the external solutions are sufficient to explain net volume absorption both in vivo and in vitro. We have provided evidence for two such driving forces. First, the development of asymmetrical anion concentration differences along the length of the proximal nephron due to preferential reabsorption of HCO-3 provides a driving force if the reflection coefficient for HCO-3 exceeds that for Cl-. Second, slight luminal hypotonicity may develop as a consequence of active solute absorption. Although both mechanisms probably occur simultaneously in vivo, we consider the former to be quantitatively the most important.

The present experiments were designed to evaluate the effects of varying the osmolality of luminal solutions on the antidiuretic hormone (ADH)-independent water and solute permeability properties of isolated rabbit cortical collecting tubules. In the absence of ADH, the osmotic water permeability coefficient (cm s(-1)) P(f) (l-->b), computed from volume flows from hypotonic lumen to isotonic bath, was 20 +/- 4 x 10(-4) (SEM); the value of P(f) (b-->l) in the absence of ADH, computed from volume flows from isotonic bath to hypertonic lumen, was 88 +/- 15 x 10(-4) cm s(-1). We also measured apparent urea permeability coefficients (cm s(-1)) from (14)C-urea fluxes from lumen to bath (P(DDurea) (l-->b)) and from bath to lumen (P(DDurea) (b-->l)). For hypotonic luminal solutions and isotonic bathing solutions, P(DDurea) (l-->b) was 0.045 +/- 0.004 x 10(-4) and was unaffected by ADH. The ADH-independent values of P(DDurea) (l-->b) and P(urea) (b-->l) were, respectively, 0.216 +/- 0.022 x 10(-4) cm s(-1) and 0.033 +/- 0.002 x 10(-4) cm s(-1) for isotonic bathing solutions and luminal solutions made hypertonic with urea, i.e., there was an absolute increase in urea permeability and asymmetry of urea fluxes. Significantly, P(DDurea) (l-->b) did not rise when luminal hypertonicity was produced by sucrose; and, bathing fluid hypertonicity did not alter tubular permeability to water or to urea. We interpret these data to indicate that luminal hypertonicity increased the leakiness of tight junctions to water and urea but not sucrose. Since the value of P(f) (b-->l) in the absence of ADH, when tight junctions were open to urea, was approximately half of the value of P(f) (l-->b) in the presence of ADH, when tight junctions were closed to urea, we conclude that tight junctions are negligible paracellular shunts for lumen to bath osmosis with ADH. These findings, together with those in the preceding paper, are discussed in terms of a solubility-diffusion model for water permeation in which ADH increases water solubility in luminal plasma membranes.

We conclude that the incidence of IHD in men undergoing hemodialysis for end-stage renal disease was not greater than that seem in a nondialysis population of men with similar risk factors. In contrast, we found the rate of IHD in women with end-stage renal disease to be accelerated when compared to similar nondialysis subjects. Our data also suggest that during the 6 yrs of hemodialysis, ischemic heart disease did not represent a major mortality risk in patients who had not had symptoms of IHD prior to therapy although it dose pose a significant mortality risk in patients with evidence of established coronary artery disease prior to the onset of dialysis.

This paper reports a theoretical analysis of osmotic transients and an experimental evaluation both of rapid time resolution of lumen to bath osmosis and of bidirectional steady-state osmosis in isolated rabbit cortical collecting tubules exposed to antidiuretic hormone (ADH). For the case of a membrane in series with unstirred layers, there may be considerable differences between initial and steady-state osmotic flows (i.e., the osmotic transient phenomenon), because the solute concentrations at the interfaces between membrane and unstirred layers may vary with time. A numerical solution of the equation of continuity provided a means for computing these time-dependent values, and, accordingly, the variation of osmotic flow with time for a given set of parameters including: P(f) (cm s(-1)), the osmotic water permeability coefficient, the bulk phase solute concentrations, the unstirred layer thickness on either side of the membrane, and the fractional areas available for volume flow in the unstirred layers. The analyses provide a quantitative frame of reference for evaluating osmotic transients observed in epithelia in series with asymmetrical unstirred layers and indicate that, for such epithelia, P(f) determinations from steady-state osmotic flows may result in gross underestimates of osmotic water permeability. In earlier studies, we suggested that the discrepancy between the ADH-dependent values of P(f) and P(DDw) (cm s(-1), diffusional water permeability coefficient) was the consequence of cellular constraints to diffusion. In the present experiments, no transients were detectable 20-30 s after initiating ADH-dependent lumen to bath osmosis; and steady-state ADH-dependent osmotic flows from bath to lumen and lumen to bath were linear and symmetrical. An evaluation of these data in terms of the analytical model indicates: First, cellular constraints to diffusion in cortical collecting tubules could be rationalized in terms of a 25-fold reduction in the area of the cell layer available for water transport, possibly due in part to transcellular shunting of osmotic flow; and second, such cellular constraints resulted in relatively small, approximately 15%, underestimates of P(f).

The present studies tested the hypothesis that osmotic disequilibrium between luminal and peritubular fluids is the driving force for net volume absorption in the isolated proximal straight tubule. Isolated tubule segments from superficial rabbit renal cortex were perfused at varying rates with a high chloride and bicarbonate-free solution as they were bathed with a normal bicarbonate-Krebs-Ringer buffer solution at 38 degrees C. Increasing the perfusion rate from congruent to 4 to congruent to 30 nl/min produced a monotonic increase in net volume absorption (Jv) from 0.18 +/- (sem) 0.03 to 0.62 +/- 0.08 nl . min-1. The chloride concentration in collected fluid samples rose from congruent to 137 to congruent to 147 mEq/liter over the same perfusion rate range. Ouabain (10(-4) m) added to the bathing solution inhibited Jv by a rate which varied from 0.20 to 0.28 nl . min-1 . min-1, depending on the perfusion rate. A mathematical model of the axial flows and transepithelial transport processes was developed. This model, and the experimental data, is consistent with the view that the driving force for isotonic fluid absorption in these tubules depends on the axial maintenance of osmotic disequilibrium between the perfusate and the bathing solution. Increasing the perfusion rate opposes osmotic equilibration by minimizing the extent to which dissipative fluxes of chloride and bicarbonate ions change the transepithelial chloride and bicarbonate concentration gradients, and by minimizing the tendency of the luminal cryoscopic osmolality to increase as volume absorption occurs.

The present experiments were designed to evaluate the effective thickness of the unstirred layers in series with native and porous (i.e., in the presence of amphotericin B) lipid bilayer membranes and, concomitantly, the respective contributions of membranes and unstirred layers to the observed resistances to the diffusion of water and nonelectrolytes between aqueous phases. The method depended on measuring the tracer permeability coefficients for the diffusion of water and nonelectrolytes (P(DDi), cm sec(-1)) when the aqueous phase viscosity (eta) was increased with solutes having a unity reflection coefficient, such as sucrose or dextran. The effective thickness of the unstirred layers (alpha(t), cm) and the true, or membrane, permeability coefficients for diffusion of water and nonelectrolytes (P(mmi), cm sec(-1)) were computed from, respectively, the slope and intercept of the linear regression of 1/P(DDi) on eta. In both the native and porous membranes, alpha(t) was approximately 110 x 10(-4) cm. The ratio of P(f), the osmotic water permeability coefficient (cm sec(-1)) to P(mmH2O) was 1.22 in the native membranes and 3.75 in the porous membranes. For the latter, the effective pore radius, computed from Poiseuille's law, was approximately 5.6 A. A comparison of P(mmi) and P(DDi), indicated that the porous membranes accounted for 16, 25, and 66% of the total resistance to the diffusion of, respectively, H(2)O, urea, and glycerol, while the remainder was referable to the unstirred layers.

The present experiments were designed to evaluate coupling of water and nonelectrolyte flows in porous lipid bilayer membranes (i.e., in the presence of amphotericin B) in series with unstirred layers. Alterations in solute flux during osmosis, with respect to the flux in the absence of net water flow, could be related to two factors: first, changes in the diffusional component of solute flux referable to variations in solute concentrations at the membrane interfaces produced by osmotic flow through the unstirred layers; and second, coupling of solute and solvent flows within the membrane phase. Osmotic water flow in the same direction as solute flow increased substantially the net fluxes of glycerol and erythritol through the membranes, while osmotic flow in the opposite direction to glycerol flow reduced the net flux of that solute. The observed effects of osmotic water flow on the fluxes of these solutes were in reasonable agreement with predictions based on a model for coupling of solute and solvent flows within the membrane phase, and considerably in excess of the prediction for a diffusion process alone.

These experiments were intended to evaluate the antidiuretic hormone (ADH)-dependent reflection coefficients of urea, sucrose, and NaCl in cortical and outer medullary collecting tubules isolated from mammalian kidney. In one group of experiments, the ADH-dependent osmotic water flows, when the perfusing solutions contained hypotonic NaCl solutions, were indistinguishable from control observations when either urea or sucrose replaced, in part, NaCl in isotonic bathing solutions (cortical collecting tubules). Similarly, both in cortical and outer medullary collecting tubules exposed to ADH, there was zero net osmotic volume flow when a portion of the NaCl in the bathing and/or perfusing solutions was replaced by either sucrose or urea, so long as the perfusing and bathing solutions were isosmolal. Taken together, these observations suggest that the ADH-dependent reflection coefficients of NaCl, urea, and sucrose, in these tubules, were identical. Since the effective hydrodynamic radii of urea and sucrose are, respectively, 1.8 and 5.2 A, it is likely that sigma(i), for urea, sucrose, and NaCl, was unity. In support of this, the diffusion permeability coefficient (P(Di) cm sec(-1)) of urea was indistinguishable from zero. Since the limiting sites for urea penetration were the luminal interfaces of the tubules, these data are consistent with the view that ADH increases diffusional water flow across such interfaces.

Optically black, thin lipid membranes prepared from sheep erythrocyte lipids have a high dc resistance (R(m) congruent with 10(8) ohm-cm(2)) when the bathing solutions contain NaCl or KCl. The ionic transference numbers (T(i)) indicate that these membranes are cation-selective (T(Na) congruent with 0.85; T(Cl) congruent with 0.15). These electrical properties are independent of the cholesterol content of the lipid solutions from which the membranes are formed. Nystatin, and probably amphotericin B, are cyclic polyene antibiotics containing approximately 36 ring atoms and a free amino and carboxyl group. When the lipid solutions used to form membranes contained equimolar amounts of cholesterol and phospholipid, these antibiotics reduced R(m) to approximately 10(2) ohm-cm(2); concomitantly, T(Cl) became congruent with0.92. The slope of the line relating log R(m) and log antibiotic concentration was congruent with4.5. Neither nystatin (2 x 10(-5)M) nor amphotericin B (2 x 10(-7)M) had any effect on membrane stability. The antibiotics had no effect on R(m) or membrane permselectivity when the lipids used to form membranes were cholesterol-depleted. Filipin (10(-5)M), an uncharged polyene with 28 ring atoms, produced striking membrane instability, but did not affect R(m) or membrane ionic selectivity. These data suggest that amphotericin B or nystatin may interact with membrane-bound sterols to produce multimolecular complexes which greatly enhance the permeability of such membranes for anions (Cl(-), acetate), and, to a lesser degree, cations (Na(+), K(+), Li(+)).

These experiments were intended to evaluate the effects of antidiuretic hormone (ADH) on dissipative water transport in cortical collecting tubules isolated from rabbit kidney. In the absence of ADH, the osmotic (P(f), cm sec(-1)) and diffusional (P(DW) cm sec(-1)) water permeability coefficients were, respectively, 6+/-6 and 4.7+/-1.3 (SD). When ADH was added to the bathing solutions, P(f) and P(DW) rose to, respectively, 186+/-38 and 14.2+/-1.6 (SD). In the absence of ADH, the tubular cells were flat and the lateral intercellular spaces were closed when the perfusing and bathing solutions were, respectively, hypotonic and isotonic; in the presence of ADH, the cells swelled and the intercellular spaces dilated. These data suggest that ADH increased the water permeability of the luminal membranes of the tubules. It was possible that the ADH-dependent P(f)/P(DW) ratio was referable to the resistance of the epithelial cell layer (exclusive of luminal membranes) to water diffusion (R(DW), sec cm(-1)). Such a possibility required that R(DW) be approximately 650, i.e., approximately 25-fold greater than in an equivalent thickness of water. To test this view, it was assumed that R(Di) values for lipophilic solutes in lipid bilayer membranes and in luminal membranes were comparable. In lipid bilayer membranes, R(Di) was substantially less than 90 sec cm(-1) for pyridine, n-butanol, and 5-hydroxyindole. In renal tubules, R(Di) for these solutes ranged from 795 to 2480 with and without ADH. It was assumed that, in the tubules, R(Di) was referable to cellular constraints to diffusion; for these solutes, the latter were 12-25 times greater than in water. Accordingly, it is possible that the ADH-dependent P(f)/P(DW) ratio was also due to cellular constraints to diffusion.

Optically black membranes prepared from sheep red cell lipids have a high electrical resistance (1-3 x 10(8) ohm-cm(2)). The ionic transference numbers (T(i)) for cations (Na(+) or K(+)) are equal to each other but at least four to five times greater than for Cl(-). The cyclic depsipeptide valinomycin produces a striking decrease in the membrane resistance when K(+), but not when Na(+) is in the solutions bathing the membrane. The ratio T(Na)/T(K), estimated from membrane voltages in the presence of ionic concentration gradients, approaches zero. The order of membrane monovalent cation selectivity, in the presence of valinomycin, is H(+) > Rb(+) > K(+) > Cs(+) > Na(+). Addition of the antibiotic to one side of a membrane which separates identical solutions of NaCl produces a substantial (up to 80 mV) membrane voltage (side opposite valinomycin negative). These data are consistent with the hypothesis that valinomycin can interact with appropriately sized cations (hydrated diameter ??? 6 A) to increase their membrane permeability, perhaps by forming hydrogen bonds between the solvation shell of the cations and carbonyl oxygens in the valinomycin molecule which are directed toward the aperture of the ring.

Amphotericin B modifies the permeability properties of thin lipid membranes formed from solutions containing sheep red cell phospholipids and cholesterol. At 10(-6)M amphotericin B, the DC membrane resistance fell from approximately 10(8) to approximately 10(2) ohm-cm(2), and the membranes became Cl(-)-, rather than Na(+)-selective; the permeability coefficients for hydrophilic nonelectrolytes increased in inverse relationship to solute size, and the rate of water flow during osmosis increased 30-fold. These changes may be rationalized by assuming that the interaction of amphotericin B with membrane-bound sterol resulted in the formation of aqueous pores. N-acetylamphotericin B and the methyl ester of N-acetylamphotericin B, but not the smaller ring compounds, filipin, rimocidin, and PA-166, produced comparable permeability changes in identical membranes, and amphotericin B and its derivatives produced similar changes in the properties of membranes formed from phospholipid-free sterol solutions. However, amphotericin B did not affect ionic selectivity or water and nonelectrolyte permeability in membranes formed from solutions containing phospholipids and no added cholesterol, or when cholesterol was replaced by either cholesterol palmitate, dihydrotachysterol, epicholesterol, or Delta5-cholesten-3-one. Phospholipid-free sterol membranes exposed to amphotericin B or its derivatives were anion-selective, but the degree of Cl(-) selectivity varied among the compounds, and with the aqueous pH. The data are discussed with regard to, first, the nature of the polyene-sterol interactions which result in pore formation, and second, the functional groups on amphotericin B responsible for membrane anion selectivity.

This paper reports the effects of amphotericin B, a polyene antibiotic, on the water and nonelectrolyte permeability of optically black, thin lipid membranes formed from sheep red blood cell lipids dissolved in decane. The permeability coefficients for the diffusion of water and nonelectrolytes (P(DDi)) were estimated from unidirectional tracer fluxes when net water flow (J(w)) was zero. Alternatively, an osmotic water permeability coefficient (P(f)) was computed from J(w) when the two aqueous phases contained unequal solute concentrations. In the absence of amphotericin B, when the membrane solutions contained equimolar amounts of cholesterol and phospholipid, P(f) was 22.9 +/- 4.6 microsec(-1) and P(DDHDH2O) was 10.8 +/- 2.4 microsec(-1). Furthermore, P(DDi) was < 0.05 microsec(-1) for urea, glycerol, ribose, arabinose, glucose, and sucrose, and sigma(i), the reflection coefficient of each of these solutes was one. When amphotericin B (10(-6)M) was present in the aqueous phases and the membrane solutions contained equimolar amounts of cholesterol and phospholipid, P(DDHDH2O) was 18.1 +/- 2.4 microsec(-1); P(f) was 549 +/- 143 microsec(-1) when glucose, sucrose, and raffinose were the aqueous solutes. Concomitantly, P(DDi) varied inversely, and sigma(i) directly, with the effective hydrodynamic radii of the solutes tested. These polyene-dependent phenomena required the presence of cholesterol in the membrane solutions. These data were analyzed in terms of restricted diffusion and filtration through uniform right circular cylinders, and were compatible with the hypothesis that the interactions of amphotericin B with membrane-bound cholesterol result in the formation of pores whose equivalent radii are in the range 7 to 10.5 A.

A cyclic depsipeptide antibiotic, valinomycin, was found to produce increased selective permeability of the plasma membranes of HK and LK sheep red blood cells to potassium but not to sodium ions. The compound had relatively little effect on the active extrusion of sodium from HK sheep red blood cells or on the Na + K-stimulated ATPase activity of membranes derived from these cells. It is proposed that the selective cation permeability produced by this compound depends primarily on steric factors, particularly the relationship between the diameter of the ring and the effective diameter of the ion. The significance of these results for the problem of the mechanism of ionic selectivity in natural membranes is discussed.

This paper describes dissipative Cl(-) transport in "porous" lipid bilayer membranes, i.e., cholesterol-containing membranes exposed to 1-3 x 10(-7) M amphotericin B. P(DCl) (cm.s(-1)), the diffusional permeability coefficient for Cl(-), estimated from unidirectional (36)Cl(-) fluxes at zero volume flow, varied linearly with the membrane conductance (Gm, ohm(-1).cm(-2)) when the contributions of unstirred layers to the resistance to tracer diffusion were relatively small with respect to the membranes; in 0.05 M NaCl, P(DCl) was 1.36 x 10(-4) cm.s(-1) when Gm was 0.02 ohm(-1).cm(-2). Net chloride fluxes were measured either in the presence of imposed concentration gradients or electrical potential differences. Under both sets of conditions: the values of P(DCl) computed from zero volume flow experiments described net chloride fluxes; the net chloride fluxes accounted for approximately 90-95% of the membrane current density; and, the chloride flux ratio conformed to the Ussing independence relationship. Thus, it is likely that Cl(-) traversed aqueous pores in these anion-permselective membranes via a simple diffusion process. The zero current membrane potentials measured when the aqueous phases contained asymmetrical NaCl solutions could be expressed in terms of the Goldman-Hodgkin-Katz constant field equation, assuming that the P(DNa)/P(DCl) ratio was 0.05. In symmetrical salt solutions, the current-voltage properties of these membranes were linear; in asymmetrical NaCl solutions, the membranes exhibited electrical rectification consistent with constant-field theory. It seems likely that the space charge density in these porous membranes is sufficiently low that the potential gradient within the membranes is approximately linear; and, that the pores are not electrically neutral, presumably because the Debye length within the membrane phase approximates the membrane thickness.

This paper reports the electrical properties of thick lipid membranes in the absence and presence of valinomycin. The thick lipid membranes were formed by placing a solution of sheep red cell lipids in decane between two cellophane partitions which formed the interfaces between the membrane and the two aqueous bathing solutions. The DC electrical resistance of these structures was found to be directly proportional to the reciprocal of the concentration of lipids in the decane (C(L)). The limiting resistance, as (C(L) (-1)) approached zero, was 3 x 10(8) ohm-cm(2). Resistance was also found to be linearly related to membrane thickness. The limiting resistance at zero thickness was again 1-3 x 10(8) ohm-cm(2). These data are interpreted to indicate that the DC resistance of thick lipid membranes comprises two surface resistances (R(S)) at each interface with the aqueous bathing solutions, and a bulk resistance (R(B)) of the lipid-decane solution, arranged in series. Measurements of the effect of variations of area on resistance were consistent with this interpretation. Valinomycin reduced R(S) but had no effect on R(B). Under certain conditions, thick lipid membranes containing valinomycin behaved like highly selective K(+) electrodes.

Lipids were obtained from high potassium (HK) and low potassium (LK) sheep red cells by sequential extraction of the erythrocytes with isopropanol-chloroform, chloroform-methanol-0.1 M KCl, and chloroform. The extract contained cholesterol and phospholipid in a molar ratio of 0.8:1.0, and less than 1% protein contaminant. Stable thin lipid membranes separating two aqueous compartments were formed from an erythrocyte lipid-hydrocarbon solution, and had an electrical resistance of approximately 10(8) ohm-cm(2) and a capacitance of 0.38-0.4 microf/cm(2). From the capacitance values, membrane thickness was estimated to be 46-132 A, depending on the assumed value for the dielectric constant (2.0-4.5). Membrane voltage was recorded in the presence of ionic (NaCl and/or KCl) concentration gradients in the solutions bathing the membrane. The permeability of the membrane to Na(+), K(+), and Cl(-) (expressed as the transference number, T(ion)) was computed from the steady-state membrane voltage and the activity ratio of the ions in the compartments bathing the membrane. T(Na) and T(K) were approximately equal ( approximately 0.8) and considerably greater than T(Cl) ( approximately 0.2). The ionic transference numbers were independent of temperature, the hydrocarbon solvent, the osmolarity of the solutions bathing the membranes, and the cholesterol content of the membranes, over the range 21-38 degrees C. The high degree of membrane cation selectivity was tentatively attributed to the negatively charged phospholipids (phosphatidylethanolamine and phosphatidylserine) present in the lipid extract.