Potassium Homeostasis: The Knowns, the Unknowns, and the Health Benefits

Abstract

Potassium homeostasis has a very high priority because of its importance for membrane potential. Although extracellular K⁺ is only 2% of total body K⁺, our physiology was evolutionarily tuned for a high-K⁺, low-Na⁺ diet. We review how multiple systems interface to accomplish fine K⁺ balance and the consequences for health and disease.

“There are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns–the ones we don't know we don't know.” Donald Rumsfeld, U.S. Secretary of Defense, 2001–2006

Organisms are believed to have evolved in a geothermal field environment with a high potassium (K)-to-sodium (Na) ratio, setting the stage for evolution of the internal milieu of modern cells (65). Eukaryotes have high intracellular (ICF) K⁺, critical for protein synthesis and cell volume regulation maintained by the active uphill pumping of K⁺ into the cell by the ubiquitous plasma membrane Na-K-ATPase. Na-K-ATPase activity also contributes to low extracellular fluid (ECF) K⁺, creating a steep transmembrane potassium gradient that is a key determinant of the membrane potential and a source of stored potential energy that is used to drive action potentials, control muscle contractility, and power ion transporters. When ECF [K⁺] falls or rises, cell membranes hyperpolarize or hypopolarize, respectively, which disrupts normal electrical excitability and can lead to life-threatening cardiac arrhythmias. Thus the ECF [K⁺] is regulated, i.e., “K⁺ homeostasis,” within narrow limits by multiple renal and extrarenal mechanisms.

Importantly, the ECF pool of potassium is very small (70 meq) relative to the recommended daily dietary potassium intake (120 meq/day), leading to daily K⁺ homeostasis challenges (2, 48). That is, ECF K⁺ must be maintained during absorption of a K⁺-rich meal as well as during fasting. Normally, daily K⁺ output equals K⁺ intake; however, efficient K⁺ homeostasis requires rapid adaptation. For example, during periods of fasting, ECF K⁺ is maintained by reducing renal K⁺ excretion and redistributing muscle ICF K⁺ to ECF. When the fast is broken by a high-K⁺ meal (e.g., the hungry fasting lion catches the gazelle), the ingested and absorbed K⁺ must be rapidly taken up by muscle ICF and/or excreted by the kidney to prevent detrimental fluctuations in ECF [K⁺]. In contrast, during pregnancy, dietary K⁺ intake chronically exceeds K⁺ excretion to facilitate the growth of the fetus; positive K⁺ balance continues after birth in the growing child until steady-state muscle mass is attained (92).

Both feedback and feedforward mechanisms exist to control this vital variable in the face of acute challenges (FIGURE 1). First, ECF [K⁺] is controlled by feedback mechanisms: increased ECF [K⁺] stimulates renal K⁺ excretion by direct effects on renal tubule cells as well as indirect effects mediated by stimulating aldosterone production, both of which stimulate K⁺ excretion (32, 72, 75). Increased ECF [K⁺] also promotes cellular K⁺ uptake, especially by muscle (14). Second, ECF [K⁺] is controlled by feedforward mechanisms (81, 117): ingested K⁺ appears to be sensed in the gut, provoking renal K⁺ excretion and cellular K⁺ uptake before or without increases in ECF [K⁺]. Feedforward control, proposed decades ago by Rabinowitz and colleagues (81), was recently shown to operate in humans (79). Third, long-recognized circadian rhythms in renal K⁺ excretion provide evidence that ECF [K⁺] may also be modulated by “predictive” control of K⁺ homeostasis, which enhances K⁺ excretion during the times when K⁺ intake is anticipated; recent studies demonstrate that clock genes control these circadian rhythms in K⁺ excretion (41, 42).

FIGURE 1.

A schematic diagram illustrating three different control mechanisms for ECF K⁺ homeostasis during dietary K⁺ intake

Feedback control is driven by a rise of plasma [K⁺], i.e., perturbation of the system, and feedforward control is driven by the sensing of dietary K⁺ intake in the gastrointestinal tract, independent of plasma [K⁺]. Predictive or adaptive control is driven by circadian rhythms.

Feedforward control operates independently of the variable-controlled, activating corrective responses that anticipate changes in the controlled variable (FIGURE 1). An undisputed mechanism of feedforward control is insulin's action to promote K⁺ shift from ECF to ICF (20, 24). Insulin is secreted during a meal when dietary glucose and K⁺ are absorbed, and, since a meal can routinely contain as much potassium as is found in the entire ECF, insulin's action to stimulate K⁺ transport into ICF is critical to buffer the rise in ECF potassium. Insulin secretion from pancreatic β-cells is controlled primarily by plasma glucose and incretin hormones secreted from intestinal cells during food absorption. Thus control of K⁺ redistribution to ICF by insulin is largely independent of ECF [K⁺] and thus is considered to be a feedforward control. The feedforward response is also independent of gut sensing of dietary K⁺ (or dietary K⁺ intake), since ECF [K⁺] falls after a K⁺-deficient meal as a result of insulin action (20, 79).

A number of excellent recent reviews have covered the topic of K⁺ homeostasis quite comprehensively (31, 32, 42, 72, 75). The goal of this brief analysis is to review new findings (“known knowns”) as well as to identify remaining gaps in our understanding of the mechanisms responsible for potassium homeostasis (“known unknowns”). In addition, we discuss recent findings linking the cardiovascular benefits of a potassium-rich diet to the molecular mechanisms engaged in excreting the potassium.

The Cardiovascular Benefit of a K⁺-Rich Diet Is a Gift From Our Past

Early terrestrial animals, like hunter-gather groups, consumed diets with high K⁺ and very low Na⁺ (71). To maintain extracellular fluid, animals evolved salt-retaining genes, including genes for salty food preference; once salt was commercialized by humans, salt-seeking behaviors drove very high NaCl consumption (10, 86). Consequently, although our human physiology has been fine-tuned to handle high-K⁺, low-Na⁺ diets, typical Western diets contain far less K⁺ than Na⁺. For example, as people transition from non-industrialized to industrialized societies, their dietary K⁺-to-Na⁺ ratio falls from >1 (3, 60) to <0.5 (63). Epidemiological studies repeatedly establish significant inverse relationships between dietary K⁺-to-Na⁺ ratio and blood pressure, whether estimated from urinary K⁺-to-Na⁺ ratio (45, 50, 85) or diet recall (116). Progression of kidney disease is reported to be blunted with higher K⁺/Na⁺ intakes in some (6, 97) but not all (43) studies and warrants careful examination. A study in adolescent girls followed for 10 years suggests it is important to consume a K⁺-rich diet while young in life: a lower rate of rise in BP was observed in girls consuming a K⁺-rich diet, yet there was no association with dietary Na⁺ (13). A recent study of half a million participants in China queried the relationship between fruit consumption (a rich source of diet K⁺) and cardiovascular disease and discovered lower BP, blood glucose, and rate of cardiovascular deaths in participants that ate fruit frequently vs. rarely (28). Although these studies establish correlations, studies involving dietary K⁺ recommendations (95) or supplementation also detect similar benefits (1, 12, 114). The highly regarded Dietary Approaches to Stop Hypertension (DASH) investigated the impact of consuming a diet rich in vegetables, fruits, and low-fat dairy products at various levels of Na⁺ intake and concluded that, although BP was reduced when Na⁺ was lowered, the effect was amplified by the K⁺-rich DASH diet, especially at higher Na⁺ levels, in the elderly, African Americans, and hypertensives; elevating dietary K⁺ alone also had an effect but not as great as the DASH diet (5, 37, 88). Arguably the most deliberate interventional study was conducted by Chang et al. (17) in a population of elderly male veterans living in retirement communities in Taiwan. K⁺-enriched (49%) cooking salt was provided to three of five kitchens (condiments like soy sauce not limited); K⁺ intake increased 70%, and Na intake decreased 25%. Impressively, relative risk of cardiovascular disease was reduced 40% over 3 years in these veterans with a mean age of 75. Likewise, hypertensive Tibetans given 25% KCl table salt exhibit reduced BP after 3 mo (119). Taken together, these studies illustrate the cardiovascular benefits of diets with higher K⁺-to-Na⁺ ratios independent of dietary Na⁺. As summarized in subsequent sections, recent findings demonstrate that the blood pressure-lowering benefit arises from physiological mechanisms recruited to excrete K⁺ consumed in a K⁺-rich diet to maintain plasma [K⁺] in a narrow range.

Overview of Organ System Regulation of K⁺ Homeostasis: Adaptations to an Acute K⁺ Load

Multiple organ systems and signaling cascades work in concert to regulate the small pool of ECF [K⁺]. FIGURE 2 provides an overview of the pathways of K⁺ compartmental redistribution (solid arrows) and routes of regulatory cross talk (dotted arrows) between organs, both known and potential but unknown, that control the small pool of ECF [K⁺]. Daily K⁺ intake is taken up rapidly into the extracellular fluid across the small intestine. Evidence suggests absorption is passive, that is, driven by electrochemical gradients and solvent drag, not unlike reabsorption along the kidney proximal tubule (4). K⁺ is usually ingested with dietary nutrients that raise plasma glucose, which stimulates insulin secretion. Insulin activates cellular Na-K-ATPase activity, which pumps ECF K⁺ into the ICF, especially muscle (dotted arrow, a) (59); additionally, a rise in [K⁺] in muscle T-tubules, evident during exercise, is sufficient to stimulate Na-K-ATPase-driven K⁺ uptake (25). Exercise training, by increasing muscle mass and the number of sodium pumps in muscle, improves potassium adaptation by increasing the rate of K⁺ uptake and the size of the buffer pool (21).

FIGURE 2.

An overview of K⁺ fluxes (solid arrows) and routes of regulatory cross talk (dotted arrows) between organs during dietary K⁺ intake

See text for explanations of individual K⁺ fluxes and regulatory cross talk.

ECF K⁺, via plasma, is continuously filtered into the kidneys, which, ideally, match K⁺ output to K⁺ intake to achieve whole body K⁺ homeostasis. Thus renal K⁺ excretion is increased or decreased in response to K⁺-rich or -poor meals, respectively. The molecular mechanisms are beginning to be unraveled. The kidneys regulate the fraction of K⁺ reabsorbed vs. excreted (bidirectional solid arrows) depending on plasma [K⁺] (dotted arrow, b), ICF K⁺ stores (dotted arrow, c), and plasma aldosterone level (dotted arrow, d), which is itself regulated by plasma [K⁺] (dotted arrow, b). Evidence suggests signals emanating from gut-sensing of [K⁺] likely stimulate renal K⁺ excretion (dotted arrow, e). Finally, CNS is implicated from studies of circadian variability in K⁺ homeostasis (41), but there is a paucity of knowledge supporting specific cross talk between brain and gut, kidney, or muscle (“unknown” dotted arrows, f, g, h, respectively).

These layers of regulation that drive cellular uptake and renal excretion efficiently blunt the rise in ECF during K⁺ intake. In a recent study (84), we acutely raised plasma [K⁺] to 5.5 mM in conscious rats via tail vein infusion and calculated its compartmental disposition: at 3 h, only 10% of the infused K⁺ remained in the ECF, 40% was excreted into the urine, and, by subtraction, 50% of the infused [K⁺] was redistributed into the ICF. Subsequently, during muscle contraction, K⁺ redistributes from muscle ICF and reenters ECF pool for excretion (FIGURE 2, solid arrow leaving muscle). Overall, K⁺ balance studies in people indicate that ∼90% of the ingested K⁺ is excreted in the urine, and the remaining 10% is excreted in feces (4); this fecal excretion is also regulated by plasma [K⁺] indirectly via mineralocorticoid regulation of colonic sodium and potassium channels (FIGURE 2, dotted arrow, d) (105) and colonic H-K-ATPase (62, 91).

The Kidneys Match Output to Input: Regions Involved

Feedback control, although very effective, requires an error signal, that is, a deviation from the baseline or set point (46), e.g., a rise in ECF [K⁺] that drives the compensatory responses such as renal K⁺ excretion and/or cellular K⁺ uptake. Studies have verified a rise in ECF [K⁺] in response to K⁺ ingestion in rodents (84, 98) and man (79, 80); the K⁺-rich Paleolithic diet, containing up 400 meq/day, would certainly have provided such an error signal (71, 94). Recent studies have provided some clarity in understanding the molecular mechanisms connecting the increase in plasma K⁺ to increase in K⁺ excretion. It has long been appreciated that >80% of the filtered load of K⁺ is reabsorbed in the proximal tubule and loop of Henle, independent of dietary intake, whereas regulated K⁺ transport that matches K⁺ output to intake occurs in the late distal convoluted tubule (DCT) through connecting tubule (CNT) and collecting duct (CD) (72). In fact, more than 50 years ago, Malnic et al. (56) demonstrated, by micropuncture, that in normal and high-K⁺ diet-fed rats, tubular fluid K⁺ increases from DCT through cortical CD, evidence for K⁺ secretion, and, in K⁺-deficient rats, tubular K⁺ decreases along the same region, evidence for K⁺ uptake. The K⁺ secretion channels were eventually identified as the renal outer medulla K⁺ channel (ROMK) (44) and tubular flow-sensitive, large-conductance Ca²⁺-activated K⁺ (BK) channels (15, 78) and the active K⁺ reabsorption route as an H-K-ATPase (26).

Molecular Mechanisms Regulating Renal K⁺ Excretion

FIGURE 3 (adapted from Ref. 99) illustrates how key cellular and molecular mechanisms, including sensors and effectors, work in concert to regulate K⁺ output along the distal nephron in the face of variable intake. Remarkably, control of K⁺ secretion vs. reabsorption is strongly influenced by Na⁺ handling upstream of these CNT/CD K⁺ transporters, specifically, by the DCT Na⁺-Cl⁻ cotransporter (NCC). It was long known that thiazide diuretics, which target and inhibit DCT NCC activity, can provoke urinary K⁺ loss (83), but there was not a mechanistic connection between NCC and K⁺ homeostasis until a 2009 study by Vallon et al. (106), which demonstrated that dietary K⁺ itself regulates NCC phosphorylation (NCC-P) and activation: low K⁺ increases and high K⁺ decreases NCC-P abundance. The physiological consequence of DCT NCC activation during low K⁺ intake is an increase in fractional reabsorption of NaCl in the DCT and, thus, reduced delivery of Na⁺ and volume to the CNT/CD where less Na⁺ movement through the apical epithelial Na⁺ channel (ENaC) reduces the electrochemical driving force for K⁺ secretion into the tubular fluid via ROMK and flow-sensitive BK channels (7), thus conserving K⁺ by reducing its excretion; likewise, the consequence of inhibiting NCC during high K⁺ intake is reduced fractional Na⁺ reabsorption by NCC and increased Na⁺ delivery and flow to ENaC, where increased fractional Na⁺ reabsorption drives K⁺ secretion and increased flow increases K⁺ secretion by BK channels, resulting in increased K⁺ output to match the increased input. The molecular mechanisms involved in the transporters' regulation have begun to be revealed and include regulation of abundance in the membrane (by synthesis and/or trafficking), phosphorylation state (by kinases and phosphatases), and other modifications such as proteolytic cleavage by cellular and tubular proteases.

FIGURE 3.

Schematic diagram and flow charts illustrating molecular events regulating Na⁺, K⁺, Cl⁻, and H⁺ transport and their transporters

A schematic diagram illustrating molecular events regulating Na⁺, K⁺, Cl⁻, and H⁺ transport and their transporters in different regions of the distal nephron (top). Also shown are flow charts showing how these individual molecular events work in concert to regulate K⁺ and Na⁺ excretion in response to low (bottom left) or high (bottom right) plasma [K⁺]. BK, high-conductance Ca²⁺-activated K⁺ channel; CD, collecting duct; CNT, connecting tubule; DCT, distal convoluted tubule; DCT1, the first portion of the DCT; DCT2, the second portion of the DCT; ENaC, epithelial Na⁺ channel; IC, intercalated cells; Kir, inwardly rectifying K⁺ channel; KCC, K⁺-Cl⁻ cotransporter; MR, mineralocorticoid receptor; NCC, Na⁺-Cl⁻ cotransporter; NCC-P, NCC phosphorylation; PC, principal cells; ROMK, renal outer medulla K channel; SGK1, serum- and glucocorticoid-inducible kinase 1; SGK1-P, SGK1 phosphorylation; SPAK, Ste20p-related proline and alanine-rich kinase; SPAK-P, SPAK phosphorylation; TK, tissue kallikrein; V, voltage or membrane potential; WNK, with-no-lysine kinase.

The first portion of the DCT (DCT1) expresses high levels of apical NCC and basolateral Na-K-ATPase, which together effect transepithelial NaCl reabsorption in this region. The basolateral heteromeric K⁺ channel (Kir 4.1/5.1) not only recycles K⁺ pumped in by the Na-K-ATPase across this membrane but its activity is responsive to changes in extracellular [K⁺] in a manner that can set off a signaling cascade that regulates apical NCC activity (118). Said another way, Kir 4.1/5.1 is a plasma [K⁺] sensor that plays a role in feedback control to normalize plasma K⁺ (99, 102). Although some details remain to be confirmed in vivo, when ECF [K⁺] is reduced, as during K⁺-deficient intake, K⁺ exit through Kir 4.1/5.1 increases, which hyperpolarizes the membrane potential and increases basolateral Cl⁻ conductance and exit, which lowers cell [Cl⁻] in cultured cells (101) and in native DCT (76). Lowering cell [Cl⁻] relieves its inhibition of the with-no-lysine kinase (WNK) and Ste20p-related proline and alanine-rich kinase (SPAK) cascade, and promotes SPAK phosphorylation and its activation of NCC (11). This feedback system works in a physiological range in vivo: pooled and modeled results suggest that decreasing plasma [K⁺] from 4.5 to 3 mM lowers cell [Cl⁻] by 5 mM and raises NCC-P three- to fourfold (102).

Reciprocally, an acute rise in plasma [K⁺] dephosphorylates NCC within minutes (98) and in a physiological range typical of that measured after a meal (84). Whether raising plasma K⁺ reduces the gradient for K⁺ exit via the Kir4.1/5.1 enough to raise cell [Cl⁻] and inhibit the WNK-SPAK cascade has been hard to determine in native kidney tissue (76). Our group measured a >50% decrease in NCC-P associated with raising plasma [K⁺] to 5.5 mM in rats, and although there was an accompanying 35% decrease in SPAK-P, it did not reach significance, suggesting that the decrease in NCC-P may not require a decrease in SPAK-P. A recent study using kidney tissue studied ex vivo confirmed the cascade from low plasma [K⁺] to increased WNK-SPAK activity and higher NCC-P, but this same study failed to demonstrate a connection between decreased cell [Cl⁻] conductance and NCC dephosphorylation (76). These authors discussed evidence that NCC-P is controlled by not only kinases but also by protein phosphatases (PP) (39, 51, 77) yet showed that the decrease in NCC-P during exposure to high extracellular [K⁺] persists during PP1, PP2A, and PP3 inhibition (76). They concluded that potassium regulation of NCC activity involves both Cl⁻-dependent and -independent pathways, including WNK-SPAK and unidentified signaling mechanisms. Since trafficking of NCC out of the apical membrane can play a role in lowering NCC activity (35, 89, 90), this mechanism should also be added to the “known unknown” category for further investigation. The bigger question of how high plasma K⁺ dephosphorylates NCC is another unknown that warrants further investigation given the prevalence of hyperkalemia in hospitalized populations and in chronic kidney diseases.

The second portion of the DCT (DCT2) expresses all the components expressed in the DCT1 (Kir 4.1/5.1, NCC, Na-K-ATPase, WNK-SPAK cascade) as well as components expressed in the CNT and CD principal cells (CD-PC) including ENaC, ROMK, mineralocorticoid receptor (MR), and 11β-hydroxysteroid dehydrogenase type 2 (36), which degrades glucocorticoids, thus conferring aldosterone specificity to the MR. A more complete cataloging of proteins expressed along the DCT is reviewed elsewhere (7, 58), and aldosterone regulation of K⁺ homeostasis is discussed in the next section. The DCT2 is notable for co-expression of NCC and the ENaC. As discussed, the inhibition of NCC in DCT1 during hyperkalemia increases delivery of both salt and water to the CNT/CD-PC, where it increases ENaC-mediated fractional reabsorption of Na⁺. But what is the effect of NCC inhibition on ENaC located in the same cell? We established, using native gels, that apical membrane NCC is localized to 700-kDa multimers in rat kidneys (52). Using this strategy, the Hoover lab recently showed that ENaC subunits are also localized to the 700-kDa multimer in a cultured DCT cell line and validated the NCC-ENaC association by co-immunoprecipitation, electronmicroscopy, and FRET (64); the study also demonstrated that thiazide diuretic inhibition of NCC reduced ENaC open probability 50% in these cultured cells (64). Although tantalizing, it remains to be established whether indirect regulation of ENaC via NCC-ENaC association and thiazide inhibition is also evident and physiologically relevant in the native DCT2 (a known unknown).

The next regions downstream are the CNT, the last portion of the nephron's distal tubule, followed by the CD, composed of principal cells (CD-PC) and intercalated cells (CD-IC). Relevant to K⁺ handling along this region, the CNT shares transport characteristics with the CD-PC, and these will be discussed together as CNT/CD-PC; it is worth noting that DCT2 and CNT reabsorb and secrete a much higher fraction of the filtered loads of Na⁺ and K⁺ than CD-PC, which probably plays a relatively minor role unless a very high K⁺/low Na⁺ diet is consumed (61, 73). Neither CNT nor CD-PC express NCC or its kinase SPAK. Kir 4.1 is expressed in this region but plays only a minor role in regulating membrane potential, probably because there are at least two other unidentified K⁺ channels expressed in the basolateral membranes of CNT/CD (99). If and how this region senses plasma [K⁺], independent of aldosterone, are not known. The transporters in the CNT/CD-PC include the basolateral Na-K-ATPase, which pumps K⁺ from the blood side into the cell and the apical ROMK and BK channels that can secrete the cellular K⁺ into the tubule lumen if a favorable driving force is provided. Na-K-ATPase couples K⁺ uptake and Na⁺ extrusion from the cell into the blood, which maintains low cellular [Na⁺], and this low Na⁺ is the driving force for Na⁺ to move from tubular fluid through apical ENaC into the cell, creating the electrochemical gradient for K⁺ to be secreted into the tubular fluid via ROMK. There is tight synchronizing of Na-K-ATPase and ENaC transport activities, mediated by hormones and kinases, to prevent fluctuations in cell volume when Na⁺ reabsorption is increased or decreased in this region (33).

In response to altered dietary K⁺, both ENaC and ROMK traffic between apical membranes and intracellular compartments and are modified by protease and kinase cascades, thus impacting the number of active channels in the membrane (35, 53, 61, 112, 113). In general, high K⁺ intake rapidly increases apical expression of both ENaC subunits and ROMK (35) to drive K⁺ secretion, and the increased tubular flow that results from inhibition of DCT Na⁺ reabsorption stimulates activity of BK channels. [Signaling mechanisms for BK discussed in a recent review (15).] In parallel, elevated plasma [K⁺] stimulates aldosterone production by adrenals, which, via MR receptors, stimulate serum- and glucocorticoid-inducible kinase 1 (SGK1) transcription and activation, which both activate Na-K-ATPase and promote ENaC activity. By inactivating the ubiquitin ligase Nedd 4-2, which promotes ENaC degradation, SGK1 indirectly promotes apical accumulation of ENaC by reducing endocytosis (57). Interestingly, the serine protease tissue kallikrein (TK), synthesized in renal CNT cells, amplifies the excretion of a K⁺ load by both proteolytic activation of ENaC (74) and inhibition of CD-IC H-K-ATPase (16), identifying TK functionally as a kaliuretic factor. TK acts rapidly in an aldosterone- and kinin-independent manner to facilitate excretion of a K⁺ load; how CNT TK is regulated by dietary K⁺ intake or plasma [K⁺] and how it inhibits H-K-ATPase activity are unknowns remaining to be determined (29). In contrast, K⁺-deficient intake reduces ENaC (30, 66), even during Na⁺ restriction (34), and provokes ROMK retraction from the apical membrane mediated by the effects of cSrc family protein tyrosine kinase on ROMK (53) and on SGK1-WNK4 inhibition of ROMK (54), effected via clathrin-mediated endocytic machinery (112, 113). Since low-K⁺ diets stimulate DCT Na⁺ reabsorption, flow to the CNT/CD is reduced, minimizing BK-mediated K⁺ secretion. CD-IC apical colonic H-K-ATPase activity is increased during K⁺ depletion, which can further reduce K⁺ excretion (22, 29).

Update on Role of Aldosterone in K⁺ Homeostasis

Aldosterone synthesis is stimulated by elevated plasma [K⁺] and acts on the kidney to increase urinary K⁺ excretion, a classic feedback mechanism that acts to normalize plasma [K⁺]. The DCT NCC has been viewed as an aldosterone target, yet this is not consistent with the suppression of NCC by elevated plasma K⁺, which is viewed as important for shifting Na⁺ reabsorption downstream from NCC to ENaC to drive K⁺ secretion and kaliuresis. However, a direct test of this notion, by treating mice with a thiazide to inhibit NCC, measured natriuresis but not kaliuresis, indicating that simply increasing Na⁺ and volume delivery to the CD is not sufficient to activate ENaC and ROMK channel activity (47). Elevated aldosterone production, stimulated directly by hyperkalemia, likely contributes to K⁺ secretion by triggering SGK1 transcription, which, via well described signaling cascades, initiates redistribution and/or activation of ENaC and ROMK channels (35, 38, 40, 115). Recent studies in mice lacking aldosterone synthase (AS^−/−) (105) or lacking renal tubular MR^−/− (23, 100) provide more answers and new questions. Complete aldosterone deficiency (AS^−/−) lowers abundance of renal NCC and NCC-P relative to that in AS^+/+ mice; 2% K⁺ diet further depresses NCC-P in both genotypes, which raises flow and Na⁺ delivery downstream. Abundance of both activated α-ENaC and ROMK is higher in AS^−/− mice, likely driven by compensatory elevated angiotensin II levels, consistent with aldosterone-independent K⁺ adaptation. In contrast, in the colon of AS^−/− mice, the responses of Na⁺ and K⁺ channels to K⁺ loading are absent, indicating dependency of colonic K⁺ adaptation on aldosterone (105). In a study to directly visualize whether aldosterone regulates NCC as well as ENaC, the Loffing lab generated mice with MR selectively knocked out of ∼20% of renal tubule cells to view cells with and without MR in the same image (23). NCC and NCC-P abundance and their responses to low-salt diet (which elevates plasma aldosterone) were unchanged in MR^−/− vs. MR^+/+ cells; further down the nephron in the collecting system, α-ENaC and Na-K-ATPase were significantly reduced in MR^−/− cells and did not respond to low-salt diet. In a parallel study, the Ellison group generated renal tubule-specific MR^−/− mice that developed salt wasting and hyperkalemia, both likely due to deficient ENaC expression; NCC and NCC-P abundance were also very low in these mice but could be elevated by K⁺ restriction, indicating that the suppression of NCC in the MR^−/− was secondary to hyperkalemia rather than the lack of MR. Together, these studies support the idea that aldosterone and MR directly regulate ENaC yet indirectly regulate NCC abundance and phosphorylation secondary to altering ENaC, K⁺ secretion, and ECF K⁺ disruptions.

Molecular Mechanisms of Feedforward Control

Is there anticipatory K⁺ excretion independent of a rise in plasma K⁺ at the DCT? When the long-fasting seal (108) that has been conserving K⁺ by maximizing renal reabsorption and shifting K⁺ from ICF to ECF returns to the sea and eats a tuna, it would be very helpful to preemptively convey the message from the K⁺-rich gut to the kidneys to prepare to secrete K⁺ and to the muscles to take it up. Gut-mediated feedforward control of plasma K⁺ homeostasis was proposed almost three decades ago by Rabinowitz and colleagues (81) based on their finding that plasma [K⁺] did not measurably increase in rats during their active period when K⁺ intake and K⁺ excretion were both elevated. They proposed that dietary K⁺ was sensed in the splanchnic bed and signaled to the kidney to increase K⁺ excretion to prevent a rise in plasma [K⁺]. Supporting this idea, Oh and Youn (69) demonstrated that, after ingesting K⁺ with a meal, feedforward control plays a major role in stimulating renal K⁺ excretion (FIGURE 4). A recent study by Preston et al. (79) showed that feedforward control of K⁺ excretion also operates in humans: when oral K⁺ is provided with a K⁺-deficient meal, serum [K⁺] did not increase but K⁺ excretion increased sharply. The molecular mechanisms underlying the feedforward control of K⁺ excretion (or muscle uptake) are not known and remain to be unraveled. There are numerous possibilities for sensing of dietary K⁺ in the gut. Dietary K⁺ causes huge variations in intestinal [K⁺]. A simple idea is that intestinal cells sense the rise in luminal [K⁺] and secrete peptide hormones to regulate renal K⁺ excretion; we have tried unsuccessfully to identify gut peptides whose secretion is affected by dietary K⁺ intake (69) or pituitary peptides that might play a role in K⁺ homeostasis (69, 70). Another possibility that remains to be explored is a role of the gut microbiota. Recent studies have provided ample evidence that the gut microbiota play a crucial role in host physiology and pathology (87). K⁺ transport plays a critical role in bacterial growth and survival and regulates protein (or toxin) secretion or interaction with the host (55, 68). It is conceivable that, in response to dietary K⁺, gut bacteria secrete substances that act as signals to affect K⁺ homeostasis of the host. Although the mechanisms by which dietary K⁺ is sensed in the gut and the signals transmitted are unknown, we postulate that kidneys possess sensors to respond to the signals (independent of plasma [K⁺]) to alter K⁺ excretion. A key unknown is whether the renal mechanism affecting feedforward regulation are the same as feedback control mechanisms via the DCT effectors (Kir 4.1/5.1, cell Cl⁻, WNK-SPAK and NCC-phosphorylation state).

FIGURE 4.

A demonstration of the operation of a feedforward control during normal dietary K⁺ intake in rats

Top: a normal, K⁺-containing meal may increase renal K⁺ excretion by increasing plasma [K⁺] and/or activating a feedforward control in response to gut sensing of dietary K⁺ intake. Bottom: when the same but K⁺-deficient meal was given to rats (no gut sensing of dietary K⁺ and thus feedforward control) and plasma [K⁺] was matched by intravenous K⁺ infusion, renal excretion was significantly smaller (69), indicating that a predominant portion of the increase in renal K⁺ excretion during normal dietary K⁺ intake is due to a feedforward control.

Chronic K⁺ Adaptation Independent of Plasma [K⁺]

Choi et al. (19) found that high fat feeding in rats decreased insulin stimulation of both glucose and K⁺ uptake. Whereas decreased glucose uptake arose from high dietary fat, decreased K⁺ uptake (and renal K⁺ excretion) occurred because rats ate less of the calorie-dense, high-fat diet and, consequently, less K⁺, since the reduced K⁺ uptake was normalized by K⁺ supplementation. More recently, Nguyen and Moe confirmed that insulin-stimulated cellular K⁺ uptake was preserved in humans with Type 2 diabetes, whereas cellular glucose uptake was reduced (67). Both studies illustrate that molecular mechanisms of insulin-stimulated K⁺ uptake are distinct from insulin-stimulated glucose uptake. Thus the reduced cellular K⁺ uptake and renal K⁺ excretion observed in high-fat-fed rats represent adaptive responses to reduced K⁺ intake (reduced to one-third of controls fed normal diet) (19). Interestingly, plasma [K⁺] was not altered in the high-fat-feeding study, indicating that renal or extrarenal K⁺ adaptation can occur without changes in plasma [K⁺]. To directly support this, a subsequent study by Chen et al. (18) confirmed (in rats fed normal fat chow) that chronically reducing K⁺ intake to one-third of normal resulted in robust downregulation of renal K⁺ excretion as well as cellular K⁺ uptake without changes in plasma [K⁺] or aldosterone concentration. Oh et al. (70) further determined that these K⁺ adaptations occur very rapidly, following overnight feeding with reduced K⁺ intake. Taken together, these studies demonstrate that both renal and extrarenal K⁺ adaptation can occur very rapidly (overnight) and without measurable changes in plasma [K⁺]. This K⁺ adaptation without changes in plasma [K⁺] identifies an intriguing gap: we hypothesize there are signals generated from altered K⁺ intake to which renal and extrarenal tissues respond, even in the absence of changes in aldosterone and plasma K⁺ concentrations, such as brief transients in plasma [K⁺], or gut factors. Additionally, there is evidence that the brain is involved in the regulation of electrolyte balance (82). Impaired potassium homeostasis is evident in hypophysectomized rats (70); in particular, both K⁺ excretion and K⁺ tolerance are reduced, despite normal K⁺ intake, similar to rats maintained on low-potassium diets, suggesting that hypophysectomized rats do not sense the dietary K⁺ intake and are, rather, geared to K⁺ conservation. One idea is that the pituitary releases a humoral factor subsequent to gut sensing of K⁺ intake that participates in both acute control of K⁺ homeostasis during absorptive periods and also exerts postabsorptive K⁺ handling independent of plasma [K⁺] (117). This unknown remains to be identified. Likewise, the possibility that NCC phosphorylation is also regulated by a mechanism independent of plasma [K⁺] warrants consideration, i.e., that NCC phosphorylation is increased when K⁺ intake is modestly reduced and renal K⁺ excretion is decreased without changes in plasma [K⁺] as in the Chen et al. study (18).

Wang and colleagues (9, 110) have shown that K⁺ depletion increases protein tyrosine kinase (i.e., cSrc)-mediated phosphorylation of ROMK, which reduces K⁺ secretion and excretion by translocation of the K⁺ channels out of the apical membrane, effects mediated by NADPH oxidase activation and superoxide anion production. A more recent study by this group further reported that these changes occurred in the absence of changes in plasma [K⁺] in rodents maintained on a low (0.1%) vs. normal (1%) K⁺ diet (111). Chen et al., with the collaboration of Wang and colleagues (18), also showed that modest K⁺ deprivation increased cSrc expression and ROMK phosphorylation without changes in plasma K⁺ or aldosterone levels. Taken together, these findings suggest that NADPH oxidase in the kidney may be involved in plasma [K⁺]-independent regulation of renal K⁺ excretion, an unknown meriting further exploration.

Is there cross talk between muscle and kidney to maintain ECF [K⁺]? This question is in the category “unknown unknowns” but should be discussed. In the postabsorptive state (when no dietary K⁺ is absorbed), K⁺ removed from the ECF by excretion must be replenished by K⁺ donated by ICF stores, mainly muscles, to maintain constant plasma [K⁺]. This takes on importance during fasting and disease states when K⁺ output is chronically greater than K⁺ intake. How do muscles sense that renal K⁺ excretion is altered or how much K⁺ to donate in the absence of changes in ECF [K⁺]? More than 10 years ago, we determined that, during dietary K⁺ deprivation and falling plasma [K⁺], muscle Na-K-ATPase α2 catalytic isoform and active K⁺ uptake are reduced, favoring redistribution of K⁺ from ICF to ECF to buffer the fall in ECF [K⁺] (20, 103, 104). That finding suggests that the link between renal K⁺ excretion and muscle K⁺ redistribution could simply be plasma [K⁺]. Subsequently, discussed above, we showed that when dietary K⁺ was reduced to one-third of control, both renal K⁺ excretion and active K⁺ uptake were reduced before any change in plasma [K⁺] or muscle Na-K-ATPase α2 (18), suggesting some cross talk. Also supporting cross talk, a recent study by Veiras et al. (107) demonstrates that angiotensin II provokes a mild hypokalemia in rats, attributed to ENaC activation driving K⁺ secretion and excretion. When angiotensin II-infused rats were fed a single K⁺-rich meal, plasma [K⁺] rose from 3.6 to 4.7 mM, but K⁺ excretion did not increase, suggesting that the ingested K⁺ is preferentially routed to replenish intracellular K⁺ stores when whole body K⁺ is depleted. This would be a useful adaptation when a long-fasting lioness catches a gazelle. Exercise is another scenario that may illuminate whether muscle-kidney cross talk exists: during intense exercise, the rate of passive K⁺ leakage from working muscles exceeds the rate of active K⁺ reuptake, resulting in a rise in plasma [K⁺], e.g., to 6 mM [K⁺] during intense rowing (8). Whether the signaling cascade reducing NCC dephosphorylation is activated (leading to negative K⁺ balance) or suppressed via cross talk is a discoverable unknown that could identify novel mechanisms of K⁺ adaptation.

Elucidation of the Cardiovascular Benefits of a Potassium-Rich Diet

A discussion of the physiology of K⁺ homeostasis would be incomplete without considering the negative impact of the rise in dietary Na-to-K ratio as humans transitioned from a natural diet to a Western diet. The homeostatic mechanisms developed to maintain K⁺ homeostasis, discussed above, evolved in animals consuming a diet (plants and meat) very rich in K⁺ and low in Na⁺; the homeostatic responses are arguably unnecessary when the high-Na⁺/low-K⁺ “Western diet” is consumed. Significantly, the recent literature reveals (FIGURE 3) that the mechanisms evolved to excrete a K⁺-rich diet require the excretion of Na⁺ (reduced NCC activity) and thereby exert a cardiovascular benefit, analogous to a diuretic (without the diuretic's negative side effects). In fact, a number of investigations have come to the conclusion that the body places a higher priority on maintaining plasma [K⁺] than on maintaining the effective circulating volume. As a consequence, responses to maintain K⁺ homeostasis while consuming a low-K⁺/high-Na⁺ Western diet can actually raise blood pressure: Vitzhum et al. (109) reported that dietary K⁺ depletion drives higher SPAK and NCC activity and provokes renal Na⁺ retention and higher blood pressure (even when ENaC or MR were antagonized) to minimizing Na⁺ delivered to the CNT/CD-PC and K⁺ secretion; these findings were confirmed by Terker et al. (102). Jensen et al. (49) report that high K⁺ intake provokes rapid NCC dephosphorylation and urinary Na⁺ loss even under Na⁺-depleted conditions. Frindt and Palmer measured ENaC channel activity during Na⁺ depletion and discovered that ENaC activity was undetectable during low K⁺ intake, despite Na⁺ depletion, presumably to minimize ENaC-driven K⁺ secretion and urinary K⁺ loss (34). These studies and others support the conclusion that K⁺ homeostasis is prioritized over maintenance of extracellular Na⁺ and volume, a reality that likely contributes to elevated blood pressure in populations with low K⁺ intake.

Conclusions

Recent studies reviewed herein have changed unknowns to knowns and revealed new unknowns. Regarding feedback regulation, the functional identification of the Kir 4.1/5.1 channel as a K⁺ sensor in the DCT unmasked a feedback loop between low plasma [K⁺], NCC phosphorylation in DCT, and reduced K⁺ secretion downstream in the CCD. Important unknowns remain concerning feedback during high plasma [K⁺], including mechanisms of rapid NCC dephosphorylation and H-K-ATPase activation, and how/whether plasma [K⁺] is sensed in the CNT/CCD. These warrant further investigation given the prevalence of hyperkalemia in the hospitalized population and in patients taking renin angiotensin system inhibitors. On a related point, the effects of gastric K⁺-binding resins (93) on K⁺ homeostasis mechanisms in hyperkalemic patients are largely unknown.

Feedforward control of K⁺ homeostasis is likely unnecessary in populations consuming high-Na⁺/low-K⁺ diets, yet takes on prominence in the setting of high-K⁺/low-Na⁺ diet given the small pool of ECF potassium. For example, a K⁺-rich meal (lion catches the gazelle) is generally well tolerated because of our evolutionary adaptation, but intravenous K⁺ infusion of a similar load can be deadly; a lack of feedforward control (via insulin or gut factor) driving K⁺ clearance from the ECF by cell uptake and renal excretion distinguishes intravenous K⁺ infusion from oral K⁺ intake. This example underscores the importance of feedforward control in K⁺ homeostasis. Significant unknowns remain, including cross talk between gut (intake) and muscle (ICF storage pools) and brain (central control). For example, when the long-fasting seal eats a tuna, are there signals between these organs that “decide” to replenish muscle stores rather than excrete the ingested [K⁺]? Do feedback and feedforward regulation use the same intrarenal mechanisms? When plasma [K⁺] is elevated by exercise, is renal K⁺ excretion blunted to facilitate muscle reuptake?

The recently elucidated molecular mechanisms of K⁺ homeostasis indicate that elevating dietary K⁺ should contribute to a lowering of the global burden of hypertension. Silver and Farley (96) editorialize that manufacturers should reduce Na⁺ added during food processing and process food in a way that maintains natural cation composition. They recommend public policies that promote increased K⁺ intake, lower the cost of fruits and vegetables, a known barrier to consumption (27), and display K⁺ content on nutritional facts panels. The recently described “known knowns” should drive high-priority strategies to accomplish these global public health goals.