Potassium recycling EDITORIAL REVIEW

Transcription

1 Kidney International, Vol. 31 (/987), pp EDITORIAL REVIEW Potassium recycling The concept of potassium recycling It has been amply demonstrated by micropuncture in nephrons accessible on the surface of the renal cortex that potassium is filtered and then nearly completely reabsorbed by the beginning of the distal tubule [1 4], confirming a hypothesis based on earlier clearance studies [5 8]. In fractional terms, potassium excreted in the urine is completely accounted for by secretion in the connecting tubule, initial cortical collecting tubule, and cortical collecting tubule segments. With the aid of various methods, including perfusion of segments of individual renal tubules perfused in vitro, as well as micropuncture, we have learned how secretion of potassium is influenced by hormones and metabolic changes [9 18]. Under conditions of potassium conservation, secretion of potassium is completely abolished, revealing a reabsorptive mechanism underneath [1, 2, 4, 14]. In circumstances of chronic high potassium intake, the morphologic adaptations in the distal renal tubule provide a particularly convincing case for the role of secretion in potassium balance [19]. Other evidence, however, indicates the classic theme of proximal reabsorption and distal secretion of potassium is not the whole story. Studies in several laboratories, including our own[20 26], strongly suggest that some of the potassium leaving the collecting tubule in the cortex is reabsorbed in the medullary collecting tubule, trapped in the medullary interstitium by countercurrent exchange, and secreted into the pars recta of the superficial nephron and pars recta and descending limb of the juxtamedullary nephron, that is, that potassium undergoes recycling in the renal medulla [23]. This hypothesis is illustrated in Figure 1. These findings reopen the question are the transport processes in the proximal tubule and loop of Henle excluded from playing a role in the urinary excretion of potassium? The purpose of this editorial is to reconsider this question and to suggest that in the acute (day to day) regulation of potassium balance, the nephron proximal to the distal tubule does play a role by means of potassium recycling. In rodents, particularly the young, the tip of the renal papilla protrudes into the pelvic ureter and can be exposed by excision of the ureter, a procedure which provides access to the end descending limb of the long Henle's loops of the juxtamedullary nephron, the papillary collecting duct, and the vasa recta [22]. Combining this approach and the micropuncture technique, we demonstrated in rats that fractional delivery of potassium to the end of the juxtamedullary descending limb is normally equal to the filtered load of potassium [22]. Administration of benzolamide, a carbonic anhydrase inhibitor that decreases reabsorption in the proximal convoluted tubule, Received for publication August 20, 1985 and in revised form February 17, by the International Society of Nephrology 695 caused potassium delivery to the end-descending limb to increase to a value equivalent to 177% of the filtered load of potassium, unequivocally establishing that potassium is secreted into the juxtamedullary nephron upstream to the bend of Henle's loop [22]. Potassium deliveries to the end-descending limb clearly exceeding the filtered load were also demonstrated in animals chronically [23] or acutely loaded [25] with potassium. From the high potassium concentrations in vasa recta plasma [23] and in the medullary interstitium [27], we concluded that the potassium secreted upstream to the juxtamedullary end-descending limb comes from the medullary interstitium. Conversely, if the potassium concentration of the medullary interstitium is diminished by furosemide [22], chronic water diuresis [23] or dietary potassium deprivation [24], potassium delivery to the end-descending limb was shown to be reduced. During the course of our studies, an unexpected relationship was observed [23], one that was not easily reconciled to the view that urinary potassium excretion is supplied by and controlled through influences operating exclusively on potassium secretion in the distal nephron. That relationship is depicted in Figure 2, which is taken from an article recently published elsewhere [28] and summarizes the findings from several laboratories obtained by micropuncture of the long loops. Fractional delivery of potassium to the end-descending limb is plotted as a function of fractional excretion of potassium. With two exceptions furosemide administration [22] (point labeled as number "2" in Fig. 2) and metabolic alkalosis [28] (number 1 1) fractional delivery of potassium to the end descending limb is equal to or greater than 100% when urinary potassium excretion is 30% or more. The relationship expressed in Figure 2 could be explained by two mutually exclusive hypotheses. One is that potassium delivered to the end descending limb is reabsorbed from the medullary ascending limb, trapped in the interstitium, and secreted by the medullary collecting duct into the urine. The alternative is that potassium is reabsorbed from urine in the medullary collecting duct, trapped in the interstitium, and re-enters the juxtamedullary nephron upstream to the end descending limb. The latter explanation implicitly assumes that urinary excretion of potassium is proportional to the delivery of potassium to the beginning of the medullary collecting duct. One way to distinguish between these alternatives is to consider the effect of reducing potassium secretion in the connecting tubule and cortical collecting tubule. If the first hypothesis is correct, delivery of potassium to the end descending limb of the juxtamedullary nephron should not be affected, while urinary potassium excretion should be reduced. The net effect would be to shift the relationship depicted in Figure 2 to the left without necessarily changing the slope of the regression line. If the second hypothesis is correct, then delivery of potassium to the end-descending limb should decline in proportion to the fall in delivery of potassium to the begin-

2 696 Jamison PCI CNT Cortex PCI DCI Medullary ray -a E 0) OC >.-D 0 CC Ii!! Inner medulla tdl tal Fig. I. Schematic representation of superficial and deep nephrons and a collecting duct with their principal segments to indicate the net transepithelial movement of potassium in various segments of the renal tubule. After Wright and Giebisch [45] but modified to indicate the correct position of the nephrons relative to the collecting ducts (after Bankir and de Rouffignac [571). Not depicted in the figure is the fact that in the rat and other rodents the thin descending limb of the short loop (superficial) nephron swings to the left in the figure across Henle's loop of the deep nephron to lie adjacent to or within the vascular bundle. The arrows indicate the net transepithelial movement of potassium under normal conditions in which excess potassium is being excreted. Abbreviations are: PCT, proximal convoluted tubule; P3, third or straight segment of the proximal tubule; tdl, thin descending limb; tal, thin ascending limb; TAL, thick ascending limb; DCT, distal convoluted tubule; CNT, connecting tubule (stippled regions); CCD, cortical collecting duct; OMCD05, outer medullary collecting duct, outer stripe; OMCD15, outer medullary collecting duct, inner stripe; and IMCD, inner medullary collecting duct. fling of the medullary collecting duct. To the extent that excretion of potassium in the final urine reflects the flow of potassium to the beginning of the medullary collecting duct', the points should decline in the direction indicated by the regression line in the illustration. Accordingly, in the initial test to distinguish between these hypotheses, amiloride was given, Potassium excretion in the urine would, according to the hypothesis of potassium recycling, actually be somewhat lower than, but still ought to change proportionately to, potassium delivery to the beginning of the medullary collecting duct Fractional excretion of potassium, % Fig. 2. Summary of micropuncture studies of potassium recycling. Fractional delivery of potassium to the end descending limb of the juxtamedullary nephron (in percent) is plotted on the ordinate as a function of fractional excretion of potassium (in percent) on the abscissa. The numbers beside each symbol correspond to the following experimental conditions and investigators: I. normal, 2. furosemide, and 3. benzolamide, Jamison et al [22]; 4. chronic K load, and 5. chronic K load + amiloride, Battilana et al [23]; 6. low K diet, Dobyan et al, [24]; 7. acute K load, Arrascue, et al, [26]; 8. normal (different diet); 9. acute metabolic acidosis; 10. acute respiratory acidosis; 11. acute metabolic alkalosis, and 12. sulfate infusion, Roy et al [23]; 13. adrenalectomy + dexamethasone and 14. Adx + Dex + aldosterone, Higashihara and Kokko [31]; 15. "hormone deprived rat" and 16. "hormone deprived rat" + ddavp, Elalouf et al [33]. The regression line was calculated from all conditions except nos. 2 and II, as discussed in the text, and no. 3, since benzolamide alters K delivery to the beginning of the decending limb. The horizontal line indicates 100% fractional delivery to the end descending limb. since the drug reduces potassium excretion by inhibiting secre tion in the distal segments of the renal tubule, and does not have a significant effect on potassium transport in the proximal tubule or loop of Henle [30], and in particular does not enhance potassium reabsorption by the proximal nephron. If the first hypothesis were correct, potassium delivery to the descending limb should not change despite the expected decline in urinary potassium; if the second hypothesis were correct, both potassium delivery and potassium excretion should decline. As depicted in the figure, potassium delivery to the descending limb in the chronic potassium loaded rat fell strikingly from 180% (point number 4 in Fig. I), to less than 65% (point number 5), and urinary potassium excretion fell from 40% to 20%, respectively. These findings are consistent with the second hypothesis that potassium recycles in the renal medulla [23]. Since then other maneuvers designed to reduce potassium excretion have also yielded findings consistent with the recycling hypothesis. Potassium delivery to the medullary collecting duct was reduced by a brief feeding of a potassium free diet (number 6 in Fig. 2) [24] and aldosterone deficiency (number 13 in Fig. 2)131], both of which inhibit potassium secretion in the connecting tubule cortical collecting tubule segment but have

3 no known effect on potassium movement in the proximal tubule or pars recta (and if aldosterone deficiency did have an effect on these segments, it would have to be to enhance potassium reabsorption to account for the decline in potassium delivery to the end descending limb). In animals deprived of potassium, fractional potassium delivery was diminished to a value essentially the same as the fractional delivery of sodium to the end descending limb [24]. Since fractional delivery of potassium from the proximal tubule is thought to be approximately the same as that of sodium (actually it is somewhat higher) and little or no net transepithelial addition of sodium is believed to occur in the descending limb of the rat [32], it was as if potassium secretion upstream to the end of the descending limb (and therefore potassium recycling) were completely abolished by a potassium free diet. Higashihara and Kokko [311 showed that correction of the aldosterone deficiency, which would enhance potassium flow to the medullary collecting duct, increased delivery to the end-descending limb (number 14 in Fig. 2) [31]. In the "hormonally deprived" rat, which lacks calcitonin, parathyroid hormone, insulin, and antidiuretic hormone [33], potassium delivery to the juxtamedullary end-descending limb was low (number 15 in Fig. 2) and was then increased by the administration of ddavp, a synthetic antidiuretic agonist of antidiuretic hormone (number 16 in Fig. 2) [33]. Potassium secretion in the distal nephron of the hormonally deprived rat was shown to be stimulated by ddavp, which would be analogous to correction of aldosterone deficiency. These and other findings illustrated in Figure 2 are consistent with the second hypothesis, the potassium recycling hypothesis formulated by Battilana and his colleagues in 1978 [23]. The recycling hypothesis is also entirely consistent with previous clearance and stop flow studies [5 7], since potassium reabsorption still precedes secretion in the nephron, although in a more complex pattern: reabsorption in the proximal tubule, secretion in the pars recta and descending limb, reabsorption in the ascending limb, secretion at a site beyond the ascending limb, and finally reabsorption. Micropuncture experiments in the Psammomys by de Rouffignac and his coworkers [20, 21] strongly suggested medullary recycling of potassium and recycling of other cations as well in this species. In most studies employing micropuncture of surface nephrons which demonstrate that potassium is secreted along the superficial distal nephron, there is considerable variability between the fraction of potassium reaching the beginning of the collecting tubule in the cortex and that excreted in the urine [1, 4, 10, II, 14]. Data obtained by micropuncture of the medullary collecting duct accessible in the exposed papilla also yields considerable variability between base and tip potassium mass flows [15, 23]. Generally, there appears to be reversal of net transepithelial potassium transport as one descends from the cortical to medullary collecting tubule secretion in the cortical segment, reabsorption somewhere along the medullary collecting duct. The key may be the fluid flow rate. Under circumstances of high urinary potassium excretion and low urinary flow, the potassium concentration in the collecting tubule would reach very high levels unless there were some potassium reabsorption in the medullary collecting duct, as first noted by Reineck, Osgood and Stein [15]. Potassium recycling 697 Recent tests of the recycling hypothesis Three predictions of the medullary recycling hypothesis have been tested in several laboratories by a variety of methods. Potassium is reabsorbed in the collecting duct The site and extent of potassium reabsorption by the medullary collecting duct under various conditions of potassium balance are unknown. In vitro studies of the isolated perfused cortical collecting duct have demonstrated active potassium secretion [34]. In contrast, evidence of potassium reabsorption by the cortical collecting duct is lacking. Compared to the cortical collecting duct, however, the outer medullary collecting duct behaves quite differently [35]. Its epithelium acts as a passive membrane to sodium and potassium, that is, the transepithelial movement of either cation is determined by the corresponding epithelial permeability and transepithelial electrochemical gradient for each cation, The potassium concentration in collecting tubule fluid entering the beginning of the medullary collecting duct under most circumstances is likely to be much greater than the potassium concentration in the outer medullary interstitium. Thus, the outer medullary collecting duct, in particular the inner stripe segment [35], is a prime candidate for the site of potassium reabsorption in the collecting duct (Fig. 1). While it seems incongruous to have one segment in the collecting duct that permits passive transepithelial movement of potassium located just downstream to another segment in which energy has been expended to secrete potassium into the lumen to a high concentration, the juxtaposition of the two segments is entirely consistent with the recycling hypothesis. In the inner medullary collecting duct, microcatheterization and micropuncture studies have not shown consistent net potassium reabsorption in animals fed a normal potassium diet [36]. In potassium loaded animals there is some evidence of potassium secretion [36]. Potassium is reabsorbed in the inner medullary collecting duct in animals fed a low potassium diet. These findings, while they do not exclude the inner medullary collecting duct as a potential site for potassium reabsorption under normal circumstances, also suggest that the outer medullary collecting duct is the more likely site in which potassium recycling is initiated. Potassium is secreted in the proximal convoluted tubule or pars recta It is generally accepted that potassium reabsorption in the proximal convoluted tubule in vivo is proportional to that in sodium and water, that is, that the concentration of potassium in the proximal fluid remains equal to that of the filtered plasma. A recent study of a desert rodent, however, has shown that the tubular fluid to plasma (TFIP) potassium ratios exceed unity towards the end of the accessible proximal convoluted tubule. At the end of the accessible proximal tubule of the Perognathus, the TF/P K ratio increases to 1.4 [37]. In the rabbit proximal convoluted tubule perfused in vitro, Kaufman and Hamburger [38] observed that although half of the segments exhibited net potassium reabsorption, the other half evidenced net potassium secretion. Ouabain diminished both potassium reabsorption and potassium secretion. These authors later concluded, however, that the mechanism of potassium reab-

4 698 Jamison sorption is passive and dependent on the concentration difference across the epithelium [39]. Grantham and his colleagues [40] were the first to observe potassium secretion in the pars recta of the proximal tubule. Recently, Work, Troutman and Schaffer [41], and Wasserstein and Agus [42] independently demonstrated potassium secretion in the pars recta of superficial and juxtamedullary nephrons (Fig. I). The secretory component was inhibitable by ouabain in Wasserstein's study, but not in Work's. When a modest favorable bath to lumen concentration gradient of 5 m of potassium was established, potassium secretion was markedly enhanced. The proximal convoluted tubule and pars recta of the juxtamedullary nephron both have a higher potassium permeability than the corresponding segments of the superficial nephron [38, 39, 41, 42]. If the outer medullary interstitium had even a slightly elevated potassium concentration from potassium recycling, potassium secretion would be especially facilitated in the juxtamedullary nephron [41, 42]. of urinary potassium concentrations. Not only is potassium concentrated in the cortical collecting duct where the ion is secreted, but in hydropenic conditions, water reabsorption along the collecting duct will raise the concentration of potassium in urine more than two orders of magnitude higher than that in the blood. Failure to sustain the high potassium concentrations within the lumen of the collecting duct under these conditions (owing to back leak of potassium along the collecting duct) would impair the ability to excrete potassium. That is, regulation of potassium balance would be subjected to limits imposed by regulation of water balance, a potentially dangerous linkage. Wright and Giebisch suggest that reabsorption of potassium in modest amounts by the collecting duct and trapping of potassium by countercurrent exchange would account for the increased interstitial concentration of potassium in the medulla. The higher level could serve to diminish transepithelial potassium reabsorption from the collecting duct, and preserve urinary excretion. Potassium is secreted in the descending limb Even under the most favorable conditions, the increase in potassium mass flow owing to secretion in the pars recta entering the descending limb could account for less than half of the mass flow at the end of the juxtamedullary descending limb that occurs in a normally fed animal [43]. Unless potassium secretion regularly occurs in the juxtamedullary proximal convoluted tubule, potassium secretion must also occur in the descending limb (Fig. 1). We devised a theoretical model which demonstrated that potassium secretion in the thin descending limb could account for the mass flow of potassium at the end of the descending limb [22]. It would not have accounted for the potassium mass flow observed after chronic [23] or acute [26] potassium loading, however. Recently Imai [44] reported that the proximal portion of the descending limb of the juxtamedullary nephron of the rat has a higher potassium permeability than does the thin descending limb of the rabbit, which may suffice to adequately account for potassium entry even under potassium loading conditions. Role of potassium recycling in renal function Since the evidence is convincing that potassium undergoes medullary recycling, it is reasonable to ask why. Is the increase in potassium recycling attendant upon the rise in urinary potassium excretion simply the unavoidable consequence of the very large transepithelial difference in potassium concentration across the medullary collecting tubule and without any functional significance? Or does potassium recycling have a physiological role in renal function, and if so what role? The answers to these questions remain largely unknown, but some intriguing possibilities are worth considering. Regulation of potassium movement across the collecting duct We initially suggested that the increased potassium concentration in the medullary interstitium might affect potassium reabsorption in the medullary collecting duct [22]. This idea has been expanded in a thoughtful way by Wright and Giebisch [451. Since urinary flow rate depends on water balance, a constant rate of potassium excretion will be associated with a wide range Regulation of potassium delivery to the collecting duct Other suggestions have in common the idea that in the non-steady state, such as in the response to an acute load of potassium, medullary recycling facilitates the transition from a low to a high urinary excretion rate of potassium. By analogy, urea, which also undergoes recycling in the renal medulla, is excreted at a greater rate during the transition from antidiuresis to water diuresis and at a lower rate during the reverse transition. Battilana and his colleagues [23] proposed that medullary recycling of potassium, which occurs to a much greater extent in the juxtamedullary loop of Henle than in the superficial loop, would selectively enhance the delivery of potassium to the juxtamedullary distal tubule and thence to the collecting tubule. A major drawback to this proposal is that it cannot be tested directly. Another possibility was proposed by Stokes [46]. He studied the effect of transepithelial potassium concentration gradients on the rabbit medullary thick ascending limb perfused in vitro. If the perfusate concentration was 25 m and the bath concentration was 5 m, potassium reabsorption was markedly enhanced, although the mass flow of potassium leaving the end of the thick ascending limb segment was still elevated as a consequence of the increased potassium flow to the beginning of thick ascending limb. The most striking finding, however, was the virtual abolition of sodium reabsorption, from 102 to 13 picoequivalents/mm tubule Iength/min. Chloride, however, was decreased only by 20%; potassium substituted for sodium as the cation accompanying chloride reabsorption across the medullary thick ascending limb. Conversely, if the bath concentration of potassium was 25 mm, and the perfusate potassium concentration was 5 mm, potassium reabsorption was reversed to potassium secretion, which also increased potassium mass flow from the distal end of the ascending limb. In this case it was chloride reabsorption that was essentially abolished from 114 to 5 peq/mm tubule length/mm, while sodium reabsorption was reduced to a lesser extent from 92 to 57. There was an apparent ion for ion exchange between potassium secretion and sodium reabsorption. These findings suggest that the medullary thick ascending limb converts the effect of medullary potassium

5 Potassium recycling 699 Percent of filtered K delivered (4 A 01 0) o o 'J 0 F Percent of filtered K delivered N) 6) 01 0).J o F r _ I I I I I I Fig. 3. Fraction of filtered potassium delivered (in percent) to end of accessible proximal tubule (Prox), beginning of accessible distal tubule (Dist), and final urine in Periods I and II. Control animals (LI) were infused with isotonic saline throughout experiment. KCI rats (U) were also infused with KCI after the end of Period I. Statistical comparisons are between control and KCI groups. From Sufit and Jamison [47]. Reproduced by permission. recycling into an inhibitory effect on sodium chloride reabsorption by the thick ascending limb. In this manner potassium recycling could augment potassium mass flow in the connecting tubule and cortical collecting tubule in three ways: 1.) Increased potassium delivery from the end of the medullary thick ascending limb. 2.) Increased fluid flow to the connecting tubule which stimulates potassium secretion. The increased flow of fluid results from the decline in medullary interstitial osmolality secondary to the inhibition of NaCI reabsorption in the thick ascending limb. As a consequence, water extraction from the descending limb would decrease and fluid flow entering the thick ascending limb and distal tubule would be enhanced. 3.) Increased sodium chloride delivery to the cortical collecting duct, which stimulates potassium secretion. Several experiments have been performed to test Stokes' hypothesis. Normal rats were studied to determine whether potassium recycling enhances the delivery of potassium, sodium and water to the distal tubule of the superficial nephron [47]. Samples of fluid were obtained by recollection micropuncture from the end accessible proximal tubule and the beginning of the accessible distal tubule before and after acute potassium loading to estimate reabsorption by the intervening segment (the loop of Henle) and delivery to the distal tubule. Potassium reabsorption by the superficial loop of Henle fell from 75 to 58% after KC1 infusion. As illustrated in Figure 3, fractional delivery of potassium to the beginning of the distal tubule increased from 12 to 26%. If these findings in accessible nephrons are representative of all nephrons, the potassium delivery to the distal tubule was equivalent to half of the urinary excretion of potassium. Reabsorption of sodium and water by the loop of Henle did not change significantly, however, although the sodium concentration in the early distal tubule fluid rose. Fig. 4. Fraction of filtered potassium delivered (in percent) to end of accessible proximal tubule (Prox), beginning of accessible distal tubule (Dist), and final urine in Periods I and II. NaCI control animals (a) received NaCI at a rate of 72 mol/min/kg body wt added to the saline infusion in Period II. Time control animals (LI) had no solute added to the saline infusion in Period II. Combined control () illustrates the results of the NaCI and Time control groups combined. The KCI group (U) received KCI at a rate of 72 tmol/min/kg body wt in the second period. The vertical bars represent 1 SE. P < 0.05 compared to KC1. Statistical comparisons are between each control group and the KCI group. From Milanes and Jamison [48]. Reproduced by permission. Therefore we turned to a model of chronic renal failure on the assumption that any inhibition of sodium chloride reabsorption in the thick ascending limb might be more readily detectable [48]. In rats the right kidney was removed and branches of the left renal artery were ligated to create a functioning remnant kidney. A week later the animals were studied before and after an acute load of potassium which increased potassium excretion from 5 to 50%. Potassium reabsorption by the superficial loop of Henle fell from 64 to 48%. As depicted in Figure 4, fractional delivery of potassium to the beginning of the distal tubule increased from 17 to 35%, which, if representative of all nephrons, was equivalent to three fourths of the urinary potassium excretion. Although there was an increase in fractional delivery of sodium and water to the beginning of the distal tubule after KC1 loading, the changes did not differ significantly from those in the control animals. Next we used the technique of microperfusion of the loop of Henle in vivo [49]. The end accessible proximal tubule was punctured and tubule fluid flow upstream was stopped by the injection of an immobile wax block. The loop was perfused by an artificial fluid whose composition and flow resembled that of fluid entering the pars recta. The beginning of the accessible distal tubule was punctured and the perfusate was collected and analyzed. Rats were studied in this manner before and after an intravenous potassium load. Preliminary findings suggested that fractional reabsorption by the loop of sodium chloride as well as that of potassium was reduced by the potassium load. By controlling the composition and flow of fluid entering the loop of Henle, an apparent inhibitory effect of acute potassium loading on sodium chloride reabsorption in the thick ascending limb

6 700 Jamison was uncovered. Further investigation is necessary, however, to verify these findings.2 Inhibition of NaCl reabsorption in the ascending limb by potassium loading may explain the well known diuretic and natriuretic effects of potassium salts [47]. In rats with normal renal function [47] or with chronic renal insufficiency, KC1 loading increased urinary flow and sodium excretion. Evidence con sistant with the view that the natriuretic effect of potassium is mediated by inhibition of salt readsorption in the loop of Henle was recently reported by Sonnenberg et al [51]. Acute KCI loading increased sodium delivery to the medullary collecting duct, but did not affect sodium reabsorption in the duct itself. How might elevating the potassium concentration in the interstitium by potassium recycling inhibit NaCI transport in the thick ascending limb? The effect may be mediated by the luminal transport carrier which, at least in the cortical thick ascending limb, requires one sodium, one potassium, and two chloride ions to move from lumen to cell interior [52]. An increase in interstitial potassium concentration might enhance cell potassium and thus oppose the inward movement of the carrier. Whether the NaCI transport carrier in the medullary thick ascending limb requires potassium, however, is controversial. Koenig, Ricapito and Kinne [53] found a K dependency of Na and Cl uptake into microsomes from rabbit medullary thick ascending limb cells, whereas in the isolated intact medullary thick ascending limb, Alvo and her colleagues [54]could not find any evidence for K dependency. An inhibitory effect of potassium recycling could be mediated in other ways, for example by stimulation of medullary production of prostaglandins which inhibit NaCl reabsorption in the thick ascending limb [55]. Does potassium recycling play a role in the adaptation to a chronic potassium load? Rabinowitz and his colleagues [56] have shown that adaptation to a change in potassium intake occurs in sheep and rats in a matter of a few days, in contrast to the generally accepted but experimentally unverified view that adaptation is a chronic process requiring a week or more to occur. Moreover the rapid potassium adaptation often occurred without a detectable change in systemic plasma potassium concentration. We now realize that, at least in the rat, increased potassium intake stimulates medullary recycling of potassium which may play a role in initiating the adaptation to a high potassium intake. In Figure 2 there are two instances in which potassium delivery to the end descending limb did not exceed 100% when potassium excretion rose above 30% furosemide administration [22] (number 2 in Fig. 2) and acute metabolic alkalosis [27] (number 11 in Fig. 2). If potassium recycling inhibits sodium chloride reabsorption by the thick ascending limb, then conditions which themselves decrease sodium chloride reabsorption in the thick ascending limb might disrupt the relationship between potassium recycling and urinary excretion of potassium depicted in Figure 2. Furosemide administration, which profoundly inhibits sodium chloride reabsorption in the thick 2 Work and colleagues [50], however, reported preliminary findings that acute KCI loading enhanced chloride reabsorption by the superficial loop of Henle. ascending limb, would obscure an inhibitory effect of potassium on sodium chloride reabsorption. Similarly, in acute metabolic alkalosis, salt depletion and the reduced chloride delivery would reduce NaCI reabsorption by the thick ascending limb. In either case potassium reabsorption by the thick ascending limb or by the outer medullary collecting tubule might be reduced, in turn decreasing potassium entry into the pars recta or descending limb of the long loops. Present status of potassium recycling hypothesis The potassium recycling hypothesis does not challenge or supplant the classic view that potassium secretion by the connecting and collecting tubules is the key transport mechanism which sets and regulates potassium excretion (Fig. 5). Indeed, it is potassium secretion by those segments that initiates and determines the extent of potassium recycling. Rather, the potassium recycling hypothesis is intended to indicate that the process of renal potassium transport is more complex, involving functional heterogeneity between superficial and juxtamedullary nephrons and among successive segments of individual nephrons. We have proposed (Fig. 5B) that potassium is recycled in the medulla, that is, potassium is reabsorbed from the collecting duct, probably in the outer medullary segment (and normally also from the thick ascending limb) and secreted into the pars recta of superficial and juxtamedullary nephron and the thin descending limb of Henle's loop juxtamedullary nephron. Potassium delivery from the thick ascending limb added to the potassium secretion in the connecting tubule and cortical collecting tubule transiently augments total potassium mass flow. In circumstances in which the acute load of potassium is increased, the recycling pathway is enhanced (Fig. SC); the mass flow of potassium. to the outer medullary collecting duct is increased by the enhanced potassium secretion in the connecting tubule and cortical collecting duct. Potassium reabsorption in the outer medullary collecting duct is thereby increased although the mass flow of potassium beyond the OMCD is still much greater than normal. Potassium reentry by secretion in the pars recta and descending limb rises sharply. Conversely, in states of acute reduction of potassium intake (Fig. 5A), potassium secretion by the connecting tubule and cortical collecting duct is abolished and superseded by potassium reabsorption. The mass flow of potassium reaching the OMCD is so low that potassium reabsorption is abolished, which in turn erases potassium secretion in the pars recta (where potassium reabsorption now predominates) and the descending limb. Recycling of potassium also may inhibit NaC1 reabsorption in the medullary thick ascending limb to enable the increased sodium chloride delivery from that segment (primarily in the juxtamedullary nephron) to stimulate secretion of potassium in the cortical collecting tubule. Fluid flow reaching the distal tubule may also be enhanced, but since it is secondary to a reduced axial osmotic gradient primarily in the inner renal medulla, the increased fluid flow is likely to occur primarily from the long descending limb of the juxtamedullary nephron rather than in the short descending limb of superficial nephron. In effect, medullary recycling of potassium could initiate a positive feedback which accelerates the excretion of an acute potassium load, initiated by potassium secretion in the connect-

7 Potassium recycling 701 A CNT B PCT CCD PCT CNT ccd C CNT CCD OMCD tdl tal IMCD tal MCD MCD Fig. 5. Schematic representation of potassium mass flow throughout the renal tubule under conditions of an acute reduction in potassium intake (A); normal potassium intake (B); and acute load of potassium (C). The arrows depict the net transepithelial movement of potassium in tubule segments. The path narrows after a segment in which net potassium reabsorption occurs and widens after a segment in which potassium secretion occurs. Abbreviations are in legend to Figure 1. ing and collecting tubules. More work is required before this can be considered established, however. Speculation about clinical implications of potassium recycling If medullary recycling of potassium normally plays a role in the excretion of potassium, it is tempting to speculate, at the risk of being premature, about the consequences of diseases of the kidney that are confined primarily or exclusively to the renal medulla. In certain tubular interstitial diseases the ability to excrete acute potassium loads is impaired despite a more than adequate glomerular filtration rate. Examples of kidney disease in which this phenomenon has been observed are cited elsewhere [48]. While a reduced secretion of renin and aldosterone may explain the impaired potassium excretion in several of these diseases, other findings suggest that a disordered renin aldosterone system does not adequately account for the reduced potassium excretory ability in these diseases; they point instead to a primary tubule defect. It is noteworthy that in this family of tubular interstitial diseases, the medullary architecture is characteristically disarranged. The countercurrent exchange of potassium among adjacent tubules and vasa recta is very likely to be disrupted which would impair the efficient medullary trapping and recycling of potassium and thereby decouple the loop contribution to the excretion of an acute potassium load. Moreover, some of the potassium secreted in the distal tubule and leaking out of the medullary collecting tubule would be carried out of the medulla by ascending vasa recta to the cortex and thus escape excretion. REX L. JAMISON Stanford University School of Medicine Stanford, California, USA Acknowledgments Research cited in this article performed in the author's laboratory was supported by funds from the National Institutes of Health Grant AM and by the American Heart Association Grant Reprint requests to Rex L. Jamison, M.D., Division of Nephrology, Department of Medicine, Stanford University School of Medicine, Stanford, California 94305, USA. References 1. MALNIC G, KLOSE RM, GIEBI5cH G: Micropuncture study of renal potassium excretion in the rat. Am J Physiol 206: , MALNIC G, KL05E RM, GIEBISCH G: Micropuncture study of distal tubular potassium and sodium transport in the rat nephron. Am J Physiol 211: , WRIGHT FS, GIEBI5cH G: Renal potassium transport: contributions of individual nephron segments and populations. Am J Physiol 235:F515 F527, Giisci-i G, STANTON B: Potassium transport in the nephron. Ann Rev Physiol 41: , BERLINER RW: Renal mechanisms for potassium excretion. Harvey Lect 55: , MUDGE GH, FOULKS J, GILMAN A: The renal excretion of potassium. Proc Soc Exp Biol Med 67: , 1948

8 702 Jamison 7. MALVIN RL, WILDE WS, SULLIVAN LP: Localization of nephron transport by stop flow analysis. Am J Physiol 194: , MOREL F: Les modalites de l'excretion du potassium par Ia rein: etude experimentale a l'aide du radio potassium chez Ic lapin. He/v Physiol Pharmaco! Ada 13: , MALNIC G, KLOSE RM, GIESISCH G: Microperfusion study of distal tubular potassium and sodium transfer in rat kidney. Am J Physiol 211: , MALNIC G, DE MELLO AIRES M, GIEBISCH G: Potassium transport across renal distal tubules during acid base disturbances. Am J Physiol 221: , 1971 Ii. WRIGHT FS, STRIEDER N, FOWLER NB, GIEBIscH G: Potassium secretion by distal tubule after potassium adaptation. Am J Physiol 221: , GROSS JB, IMAI M, KOKKO JP: A functional comparison of the cortical collecting tubule and the distal convoluted tubule. J C/in invest 55: , PETERSON LN, WRIGHT FS: Effect of sodium intake on renal potassium excretion. Am J Physio/ 233:F225 F234, WRIGHT FS: Sites and mechanisms of potassium transport along the renal tubule. Kidney in! 11: , REINECK Hi, Osaooo RW, STEIN JH: Net potassium addition beyond the superficial distal tubule of the rat. Am J Physio/ 235:Fl04 FllO, GooD DW, WRIGHT FS: Luminal influences on potassium secretion: Sodium concentration and fluid flow rate. Am J Physio/ 236:Fl92 F205, Gooo DW, VELAZQUEZ H, WRIGHT FS: Luminal influences on potassium secretion: Low sodium concentration. Am J Physiol 246: F609 F619, FIELD MJ, STANTON BA, GIEBISCH GH: Differential acute effects of aldosterone dexamethasone and hyperkalemia on distal tubular potassium secretion in the rat kidney. J C/in Invest 74: , STANTON BA, BIEME5DERFER D, WADE JB, GIEBISCH G: Structural and functional study of the rat distal nephron: effects of potassium adaptation and depletion. Kidney 1w' 19:36 48, DE ROUFFIGNAC C, MOREL F: Micropuncture study of water, electrolyte and urea movements along the loops of Henle in Psammomys. J C/in Invest 48: , DE ROUFFIGNAC C, MOREL F, Moss N, ROINEL N: Micropuncture study of water and electrolyte movements along the loop of Henle in Psammomys with special reference to magnesium, calcium and phosphorous. Pflugers Arch 344: , JAMI50N RL, LACY FB, PENNELL JP, SANJANA VM: Potassium secretion by the descending limb of pars recta or juxtamedullary nephron in vivo. Kidney in! 9: , BATTILANA CA, DOBYAN DC, LACY FB, BHATTACHARYA J, JOHNSTON RA, JAMISON RL: The effect of chronic potassium loading on potassium secretion by the pars recta or descending limb of the juxtamedullary nephron in the rat. J C/in Invest 62: 1093 I 103, DOBYAN DC, LACY FB, JAMI5ON RL: Suppression of potassium recycling in the renal medulla by short term potassium deprivation. Kidney mt 16: , DOBYAN DC, ARRASCUE if, JAMISON RL: Terminal papillary collecting duct reabsorption of water, sodium, and potassium in Psammomys obesus. Am J Physio/ 239:F539 F544, ARRASCUE if, DOBYAN DC, JAMISON RL: Potassium recycling in the renal medulla: Effects of acute KCI administration to rats fed a potassium free diet. Kidney in! 20: , BULGER RE, BEEUWKES R 111, SAUBERMANN AJ: Application of scanning electron microscopy to analysis of frozen hydrated sections. III. Elemental content of cells in the rat renal papillary tip. J Ce// Bio/ 88: , JAMIsON RL, MULLER SUUR R: Recycling of potassium in the renal medulla, in Potassiu,n Transpor!: Physiology and Pat hophysio/ogy, edited by GIEBISCH G. New York, Academic Press (in press) 29. Ro DR, BLOUCH KL, JAMIs0N RL: Effects of acute acid base disturbances on K delivery to the juxtamedullary end descending limb. Am J Physio/ 243:F188 Fl96, DUARTE CG, CHOMETY F, GIEBISCH G: Effect of amiloride, ouabain, and furosemide on distal tubular function in the rat. Am J Physiol 221: , HIGASHIHARA E, KOKKO JP: Effects of aldosterone on potassium recycling in the kidney of adrenalectomized rats. Am J Physiol 248:F219 F227, JAMISON RL, BUERKERT J, LACY F: A micropuncture study of Henle's thin loop in Brattleboro rats. Am J Physiol 224: , ELALOUF JM, ROINEL N, DE ROUFFIGNAC C: Effects of ADH on electrolyte movements in rat juxtamedullary nephrons: Stimulation of medullary K recycling. Am J Physio/ (in press) 34. GRANTHAM ii, BuRG MB, ORLOFF J: The nature of transtubular Na and K transport in isolated rabbit renal collecting tubules. J C/in Invest 49: , Si'ous JB: Na and K transport across the cortical and outer medullary collecting tubule of the rabbit: Evidence for diffusion across the outer medullary portion. Am J Physio/ 242:F514 F520, HAYSLETT JP, MYKETEY N, BINDER HJ, ARONSON PS: Mechanism of increased potassium secretion in potassium loading and sodium deprivation. Am J Physiol 239:F378 F382, BRAUN EJ, Ro DR, JAMISON RL: Micropuncture study of the superficial nephron of Perognathus pendi//afus. Am J Physiol 231:F612 F617, KAUFMAN is, HAMBURGER Ri: Potassium transport in the isolated proximal convoluted tubule. Am J Physio/ 244:F297 F3 12, KAUFMAN JS, HAMBURGER RJ: Passive potassium transport in the proximal convoluted tubule. Am J Physio/ 248:F228 F232, GRANTHAM JJ, QUALIZZA PB, IRWIN RL: Net fluid secretion in proximal straight renal tubules in vitro: role of PAH. Am J Physiol 226: , WORK J, TROUTMAN SL, SCHAFER JA: Transport of potassium in the rabbit pars recta. Am J Physiol 242:F226 F237, WASSERSTEIN AG, AGUS ZS: Potassium secretion in the rabbit proximal straight tubule. Am J Physio/ 245:F167 F174, Jusor. RL, WORK J, SCHAFER JA: New pathways for potassium transport in the kidney. Am J Physio/ 242:F297 F312, IMAI M: Functional heterogeneity of the descending limb of Henle's loop. II. Interspecies differences among rabbits, rats and hamsters. Pflugers Arch 402: , WRIGHT FS, GIEBISCH G: Regulation of potassium excretion, in The Kidney: Physio/ogy and Pathophysiology, edited by SELDIN D, GIEBLSCH G. New York, Raven Press, 1985, pp STOKES JB: Consequences of potassium recycling in the renal medulla. Effects of ion transport by the medullary thick ascending limb. J C/in Invest 70: , SUFIT CR, JAMISON RL: Effect of acute potassium load on reabsorption of Henle's loop in the rat. Am J Physiol 245:F569 F576, MILANES CL, JAMISON RL: Effect of acute potassium load on reabsorption of Henle's loop in chronic renal failure in the rat. Kidney mt 27: , MULLER SUUR R, JAMISON RL: Effects of acute potassium loading on Henle's loop electrolyte transport: A microperfusion study in vivo. (abstract) Kidney mt 27:3 17, WORK J, BOOKER BB, GALLA JH, LUKE RG: Acute KCI loading increases superficial loop of Henle chloride reabsorption. (abstract) Kidney Int 29:411, SONNENBERG H, HORVATH C, CHONG K, WILSON DR: Atrial natriuretic factor inhibits sodium transport in the collecting duct. Am J Physio/ 250:F963 F966, GREGER R, SCHLATTER E: Presence of luminal K, a prerequisite for active NaCI transport in the cortical thick ascending limb of Henle's loop of rabbit kidney. Pfiugers Arch 392:92 94, KOENIG B, RICAPITO S, KINNE R: Chloride transport in the thick ascending limb of Henle's loop: Potassium dependence and stoichiometry of the NaCI cotransport system in plasma membrane vesicles. Pflugers Arch 399: , ALVO M, CALAMIA J, EVELOFF i: Lack of potassium effect on Na-Cl cotransport in the medullary thick ascending limb. Am J

9 Potassium recycling 703 Physiol 249: F34.-F39, 1985 TZENDZALIAN PA: Time course of adaptation to altered K intake in 55. STOKES JB: Effect of prostaglandin E2 on chloride transport across rats and sheep. Am J Physiol 247:F607 F617, 1984 the rabbit thick ascending limb of Henle. J Clin Invest 64: , 57. BANKIR L, OK ROUFFIGNAC C: Urinary concentrating ability: 1979 Insights from comparative anatomy. Am J Physiol 249:R643 R666, 56. RABIN0wITz L, SARASON RL, YAMUCHI H, YAMANAKA KK, 1985