Maintaining K⁺ balance on the low-Na⁺, high-K⁺ diet

Abstract

A low-Na⁺, high-K⁺ diet (LNaHK) is considered a healthier alternative to the “Western” high-Na⁺ diet. Because the mechanism for K⁺ secretion involves Na⁺ reabsorptive exchange for secreted K⁺ in the distal nephron, it is not understood how K⁺ is eliminated with such low Na⁺ intake. Animals on a LNaHK diet produce an alkaline load, high urinary flows, and markedly elevated plasma ANG II and aldosterone levels to maintain their K⁺ balance. Recent studies have revealed a potential mechanism involving the actions of alkalosis, urinary flow, elevated ANG II, and aldosterone on two types of K⁺ channels, renal outer medullary K⁺ and large-conductance K⁺ channels, located in principal and intercalated cells. Here, we review these recent advances.

the majority of the United States population consumes a “Western” diet of high Na⁺ and low K⁺. The Western diet produces an acid load (29, 71) and is associated with hypertension, especially in genetically predisposed, Na⁺-retaining individuals (4, 38, 43, 166, 167). The “ancient” or “Paleo” diet consists of fruits, vegetables, and nuts and produces an alkaline load (29). The Yanomami (no-salt culture) tribe of South America, when studied in the 1970s, gave us a window into the ionic content and renal handling of the ancient diet. The urinary K⁺ concentration was very high and the Na⁺ and Cl⁻ concentrations were near zero, indicating the consumption of an exaggerated form of the modern-day equivalent of the ancient diet (95). However, it has been questioned how K⁺ is eliminated on such a diet of minimal Na⁺ when Na⁺ reabsorption in the distal nephron is the driving force for K⁺ secretion. Moreover, a low Na⁺ intake may be detrimental to health (93). Nevertheless, a longitudinal study (5) has shown that a diet of fruits and vegetables, known as the dietary approaches to stop hypertension diet, reduced blood pressure in hypertensive subjects, and most studies now agree that the Paleo diet is healthier, antihypertensive, and associated with stroke reduction (4, 13, 23, 23, 51, 84, 95). To understand the beneficial effects of a low-Na⁺, high-K⁺ (LNaHK) diet, studies are needed to determine how such a long-term diet is handled by the mammalian kidney.

Very few studies have explored renal handling of the LNaHK diet. However, recent progress has been made in the field of distal renal transport physiology with respect to the regulation of Na⁺ and K⁺ homeostasis after consumption of either a low-Na⁺ or high-K⁺ diet. With low volume or low Na⁺ intake, Na⁺ is reabsorbed by the Na⁺-Cl⁻ cotransporter (NCC) in the early distal tubule (168). During high K⁺ intake, Na⁺ bypasses the NCC and is reabsorbed in the connecting tubules (CNTs) and cortical collecting ducts (CCDs) via the epithelial Na⁺ channel (ENaC), which is localized in principal cells (PCs) (19). Na⁺ reabsorbed by amiloride-sensitive ENaC causes an electronegative driving force for the secretion of K⁺ into the lumen (24, 121, 122). Na⁺ reabsorbed with Cl⁻ via NCC will restore Na⁺ and volume (110). Na⁺ reabsorbed via ENaC will restore Na⁺ but reduce blood K⁺ concentration (121, 169). A set of key regulators and signaling networks are involved in distinguishing between these Na⁺ transporters. The LNaHK diet reveals several interesting findings that can help distinguish these regulatory pathways.

This review will emphasize three major simultaneously occurring events associated with the LNaHK diet that are responsible for maintaining K⁺ balance and may explain why the LNaHK diet is beneficial. The first is the dietary anionic content, which is responsible for creating an acidic or alkaline load for the kidneys. The second is the role of elevated plasma ANG II concentration (P[ANG II]) and plasma aldosterone (Aldo) concentration (P[Aldo]). The third is the generation and effects of the high distal flow.

The ability of the LNaHK diet to produce an alkaline load appears to be a critical factor for maintaining K⁺ homeostasis. The urine of the Yanomami people contained near 0 mM Cl⁻, indicating the production of an alkaline load, as expected from a diet of fruits, vegetables, and nuts. Like the Yanomami, mice maintain their K⁺ balance with a high dietary K⁺ of 5%, so long as it is accompanied with alkaline counteranions (30, 169).

Because of the combination of low Na⁺ and high K⁺, ANG II, P[Aldo], and plasma K⁺ concentration (P[K⁺]) are all elevated in the LNaHK diet. Thus, we will discuss the importance and interactions of K⁺, ANG II, and Aldo and their importance in influencing K⁺ secretion to maintain K⁺ balance on the LNaHK diet. Two studies in particular, one study with mice (31) and the other study with rats (150), have reported renal handling on LNaHK diets of similar Na⁺ and K⁺ composition, and one or the other has addressed the roles of ANG II and Aldo in maintaining K⁺ homeostasis.

The final critical event for animals on the LNaHK diet is the generation of a high urinary flow. The beneficial blood pressure-lowering effects of high K⁺ may be related to this natural diuretic effect (146, 169). The mechanism for generating high flow will be discussed. High flow in the distal nephron stimulates K⁺ secretion by activating large-conductance Ca²⁺-activated K⁺ (BK) channels (75, 178). BK channels are composed of two subunits: pore-forming BK-α and one of four BK-β proteins (β₁–β₄), which confer differing properties on BK channels (150). CNT cells contain BK-α–BK-β₁ (103, 104), and intercalated cells (ICs) contain BK-α–BK-β₄ (52). Mice with knockout (KO) of the BK β₁-subunit (β₁-KO mice) or BK β₄-subunit (β₄-KO) exhibit attenuated K⁺ excretion compared with wild-type (WT) mice when given a high-K⁺ diet.

A variety of K⁺ channels may be involved in the secretion of K⁺, depending on dietary intake and distal flow. BK channels and renal outer medullary K⁺ (ROMK) channels have been the most studied channels that secrete K⁺ in the CNT/CCD of the Aldo-sensitive distal nephron (ASDN). The ROMK channel mediates K⁺ secretion in the Aldo-sensitive distal nephron when flow is normal or low (47, 119, 153, 180). Both ROMK (113, 162, 180) and BK (31, 36, 171) channels are regulated by Aldo. However, it has been difficult to discern the functional activity of ROMK channels in the CCD in vivo using ROMK KO mice because the thick ascending limb (TAL) also contains ROMK channels (32, 33, 66, 161). In the TAL, the ROMK channel recycles K⁺ absorbed via Na⁺-K⁺-2Cl⁻ cotransporter (NKCC)2 in the apical membrane. ROMK KO mice exhibit Bartter's syndrome, with flow-stimulated, BK channel-mediated K⁺ secretion in the distal nephron (9). The small-conductance Ca²⁺-activated K⁺ channel (16) also has been ascribed a K⁺ secretory role in the Aldo-sensitive distal nephron. The small-conductance Ca²⁺-activated K⁺ channel is very sensitive to Ca²⁺ (10) and may also be involved in flow-induced K⁺ secretion.

This review will not discuss the “with no lysine kinase” (WNK) and SP-related proline/alanine-rich kinase (SPAK) pathways because this topic has been expanding rapidly and would add considerably to the length. These pathways, which will have large roles in sorting the downstream responses to changes in P[K⁺], P[Aldo], and P[ANG II] in animals on the LNaHK diet, have been extensively studied and reviewed by other groups (2, 53, 54, 57, 140).

Unless otherwise noted in the review, the Na⁺ and K⁺ contents of the diets are in percent composition by weight as follows: regular diet = 0.3% Na⁺ and 0.6% K⁺; low Na⁺ = 0.01% Na⁺ and 0.6% K⁺; high K⁺ = 0.3% Na⁺ and 5% K⁺; and LNaHK = 0.01% Na⁺ and 5% K⁺. The high-K⁺ and LNaHK diets contain 5% equal parts citrate-carbonate-Cl⁻ as the counteranions to K⁺. The high-K⁺-Cl⁻ and LNaHK-Cl⁻ diets contain 5% Cl⁻ as the counteranion to K⁺. K⁺ imbalance is defined by P[K⁺] > 5.5 mM, which is clinically considered “hyperkalemia” (67).

The LNaHK Diet and the Role of Acid/Base

The reciprocal relation between K⁺ and H⁺ secretion is well established; however, the ability of high-K⁺ diets to generate an acid or alkaline load has been seldom considered. The Yanomami study revealed nearly zero urinary Cl⁻, indicating that their LNaHK diet was very alkaline. The acid load is considerably greater (hyperchloremic acidosis) when Cl⁻ (5%) is associated with 5% K⁺ in the diet, meaning that renal mechanisms will be required to eliminate the high H⁺ concentration as well as K⁺ concentration. The dramatic differences between the effects of an alkaline-producing and acid-producing LNaHK diet are shown in Table 1, which shows different values relating the K⁺ handling of mice given either a LNaHK diet (alkaline producing), a LNaHK diet plus acidic drinking water (280 mM NH₄Cl + 2% sucrose), or a LNaHK-Cl⁻ diet (5% K⁺ + 5% Cl⁻). As shown in Table 1, the P[K⁺] of 4.71 for mice on the LNaHK diet is within the normal range; however, mice become very hyperkalemic (6.8 mM) when the LNaHK diet is accompanied by acidic drinking water. When the diet was an acid-producing LNaHK-Cl⁻ diet, P[K⁺] increased to 9.3 mM. P[Aldo] of mice on the LNaHK-Cl⁻ diet was very high [12,674 pg/ml (unpublished observations)], showing that the primary defect was the failure to secrete K⁺ rather than a failure to produce Aldo. In the LNaHK mice, the transtubular K⁺ gradient, a measure of the passive driving force for K⁺ across the CNT/CCD, is approximately twofold the value of mice on the LNaHK-Cl⁻ diet. Therefore, these mouse experiments indicate that the ability of the Yanomami diet to produce an alkaline load was key to maintaining K⁺ balance when consuming a LNaHK diet.

Table 1.

Differences in K⁺ handling with alkaline versus acidic LNaHK diets

	Diet	Anion	Urine pH	Plasma K+ Concentration, mM	Transtubular K+ Gradient
Ref. 169	LNaHK	Carb/Citr/Cl−	8.59	4.71	22.9
Ref. 30	LNaHK-AW	Carb/Citr/Cl−	6.63	6.8	15.1
Ref. 169	LNaHK	Cl−	5.89	9.3	10.2

All diets were given for 7–10 days. LNaHK-AW = low-Na⁺, high-K⁺ diet (LNaHK) with acid (280 mM NH₄Cl plus 2% sucrose) drinking water. Carb/Citr/Cl, equal parts carbonate, citrate, and Cl⁻ anions with K⁺.

Considerable in vivo and in vitro work has illustrated the competition between the transport pathways for eliminating K⁺ versus H⁺ in the distal nephron (6, 21, 69). However, the recently described link between ENaC-mediated Na⁺ reabsorption and pendrin-mediated HCO₃⁻ secretion (64, 100, 101, 158) is an explanation for why an alkaline-producing diet facilitates K⁺ secretion. HCO₃⁻ increases the activity of ENaC in cultured cells (100), and pendrin KO mice exhibit diminished apical expression of ENaC (101). The converse, inhibition of ENaC-mediated Na⁺ absorption promotes H⁺ secretion, has also been demonstrated (91). Therefore, K⁺ secretion could be linked to pendrin activity based on ties between K⁺ secretion and ENaC activity. Such a mechanism could account for the fact that an alkaline LNaHK diet, which generates systemic HCO₃⁻, promotes a high rate of K⁺ secretion, and the K⁺ balance is maintained.

The driving force for K⁺ secretion not only depends on Na⁺ reabsorption but also on the resistance for luminal anions to traverse the paracellular pathway. As shown in Fig. 1, IC-β are predominant in the CNT (89, 147), particularly in alkaline conditions. With predominant IC-β, Cl⁻ reabsorbs via pendrin in exchange with secreted HCO₃⁻. Secreted luminal HCO₃⁻ is a larger anion than Cl⁻. Similar to the large anions gluconate and sulfate, HCO₃⁻ is more resistant to paracellular reabsorption than Cl⁻, thereby enhancing the electronegative transepithelial voltage (28, 44, 109, 118). This principal for HCO₃⁻ was first demonstrated with microperfused distal tubules of rats on low-Na⁺ diets (28). The transepithelial voltage was significantly more negative by 6 mV when the distal tubule was perfused with NaHCO₃ compared with NaCl. Therefore, with a LNaHK diet that produces an alkaline load, reabsorbing NaCl is tightly exchanged for secreted KHCO₃.

Fig. 1.

Illustration of the effects of acidic or alkaline loading on coordinated transport between prinicpal cells (PCs) and intercalated cells (ICs) of mice on a low-Na⁺, high-K⁺ (LNaHK) diet. A: mechanism of HCO₃⁻-promoted K⁺ secretion in predominant IC-β in the connecting tubule (CNT) of mice on a LNaHK diet. With only IC-β, the reabsorbed Cl⁻ is exchanged for secreted HCO₃⁻ via pendrin and the lumen negative potential from Na⁺ reabsorbed via epithelial Na⁺ channels (ENaC) drives the secretion of K⁺ from PCs and ICs. B: mechanism of K⁺ secretion with the LNaHK-Cl⁻ diet. As the load becomes more acidic, IC-β are converted to IC-α (108, 126) in the CNT and initial cortical collecting duct (CCD). Cl⁻ is reabsorbed via pendrin in exchange for secreted HCO₃⁻, which binds to H⁺, which is pumped into the lumen from IC-α. CO₂ recycles back through IC-β. H₂O would recycle back via aquaporin 5, which is localized on IC-β (105), and its apical membrane insertion is linked with K⁺ intake (106). CO₂ and H₂O would combine to generate HCO₃⁻, which would be secreted, and H⁺, which would be actively transported into the plasma via basolateral H⁺-ATPase. The result is Na⁺ reabsorption with Cl⁻ rather than exchange with K⁺. The reduced driving force for K⁺ secretion is reflected by the lower transtubular K⁺ gradient and subsequent rise in plasma K⁺ concentration (P[K⁺])for mice on an acidic LNaHK diet (Table 1). The urine is acidified in the inner medullary collecting ducts, where the predominating cells are acid secreting IC-α. BK, large-conductance Ca²⁺-activated K⁺ channels.

As the LNaHK diet becomes more acidic, IC-β are converted to IC-α (108, 126). In the CNT/CCD, IC-β and IC-α will coexist. In this state, secreted H⁺ from IC-α, which contain apical carbonic anhydrase IV (107), will react with HCO₃⁻ to form H₂O and CO₂. CO₂ will be recycled back to IC-α. HCO₃⁻ recycling results in net Na⁺ reabsorption with Cl⁻ rather than exchanging for K⁺. With a preponderance of Cl⁻ in the lumen, some Cl⁻ may be reabsorbed through the paracellular pathway (118). The consequence of decreased resistance to anion reabsorption via the paracellular pathway and H⁺ secretion will be less driving force for K⁺ secretion, thereby explaining the reduced transtubular K⁺ gradient and elevated P[K⁺] in the LNaHK-Cl⁻ diet. The urine can be additionally acidified in medullary collecting ducts, where there is a predominance of IC-α (155).

In both models (Fig. 1, A and B), Cl⁻ reabsorption with Na⁺ is dependent on pendrin, with luminal HCO₃⁻ activating ENaC (100). It is not understood at the molecular level how HCO₃⁻ activates ENaC from the luminal side. However, ENaC-mediated Na⁺ reabsorptive exchange for K⁺, along with HCO₃⁻ excretion, was enhanced in acid-loaded electrogenic Na⁺-HCO₃⁻ cotransporter (NBCe2) KO mice. Therefore, NBCe2, which resides in distal nephron cells, prevents ENaC-mediated Na⁺ for K⁺ exchange during acidosis (172–174). The role of NBCe2 and other transporters and signaling mechanisms linking ENaC and pendrin need to be determined.

Although it is clear that the LNaHK diet requires alkaline anions to prevent hyperkalemia, the results are mixed with respect to K⁺ homeostasis when mice or rats are given a normal-Na⁺, high-K⁺ diet with Cl⁻¹ as the counteranion. Most studies have not reported the anion content associated with high K⁺. Some studies have shown that P[K⁺] is within the normal range when a 10% KCl (5% K⁺ and 5% Cl⁻) diet is given (40, 97); however, one of those studies administered the high-K⁺-Cl⁻ for only 2 days (97). Other studies have shown that adapting mice (30) or rats (79, 156) with a high-K⁺-Cl⁻ diet produces hyperkalemia with a P[K⁺] of 5.8–6.8 mM. Nevertheless, the degree of hyperkalemia in animals on a high K⁺-Cl⁻ diet is not nearly as severe as animals on a LNaHK-Cl⁻ diet. The reason is not clear, but increased Na⁺ availability to ENaC in the high K⁺-Cl⁻ diet may reduce the dependency on pendrin-mediated HCO₃⁻ secretion to enhance ENaC activity.

The LNaHK Diet and Distinct Roles of P[Aldo], P[ANG II], and P[K⁺]

Aldo, ANG II, and K⁺ should all be considered critical regulators of K⁺ homeostasis in animals on a LNaHK diet. The independent importance of each for maintaining K⁺ balance will be discussed in this section. In Table 2, we compare data from three previously published studies (7, 31, 150) and some unpublished data from mice and rats on a variety of diets with differing Na⁺ and K⁺ contents to better understand the independent and interdependent roles of P[Aldo], P[K⁺], and P[ANG II] in the regulation of K⁺ excretion with the LNaHK diet. Values of P[Aldo] were measured by either radioimmunoassay or ELISA. Also shown are P[K⁺] and renal ANG II concentrations, when determined. Although the absolute values for P[Aldo] are different, probably because of the method of analysis, the relative changes in P[Aldo] are similar with the four diets. These values will be referenced below, which discuss Aldo versus high P[K⁺] and the ANG II regulation of K⁺ excretion.

Table 2.

Effects of varying dietary Na⁺ and K⁺ on plasma K⁺ concentration, plasma aldosterone concentration, and renal ANG II

	Na+, %	K+, %	Plasma K+ Concentration, mM	Plasma Aldosterone Concentration, pg/ml	Method	ANG II, fmol/g
Rat (150)
Regular	0.5	0.8	4.8	128	RIA	120
Low Na+	0.001	0.8	4.7	428	RIA	230
High K+*	0.5	5	5.1	420	RIA	120
LNaHK*	0.001	5	5.9	1355	RIA	220
Mouse (31)
Regular	0.3	0.6	3.8	239	ELISA	ND
Regular ADX	0.3	0.6	4.5	56	ELISA	ND
High K+	0.3	5	4.6	753	ELISA	ND
LNaHK	0.01	5	4.3	3630	ELISA	ND
LNaHK-ADX-low Na+	0.01	5	7.4	607	ELISA	ND
LNaHK-ADX-high Na+	0.01	5	4.2	4293	ELISA	ND
Mouse (7)
Regular	0.3	0.9	4.5	202	RIA	ND
Low Na+	0.01	0.9	ND	440	RIA	ND
Low K+	0.3	0.05	3	64	RIA	ND
High K+*	0.3	3	4.9	354	RIA	ND
Mouse (unpublished observations)
Low Na+	0.01	0.6	4.1	700	ELISA	ND
LNaHK-captopril	0.01	5	9.2	9278	ELISA	ND

ADX, adrenalectomy; RIA, radioimmunoassay; ND, not determined. Unpublished observations were from the Sansom laboratory. High-K⁺ and LNaHK diets contained 5% citrate/carbonate/Cl⁻.

^*Anionic content unknown.

In response to high K⁺ intake, Aldo places K⁺ into extrarenal cells (182) as well as stimulating K⁺ excretion in the distal nephron. Aldo has two major roles in the regulation of Na⁺ and K⁺ transport in the distal nephron. With low volume or low Na⁺ intake, the renin-angiotensin system stimulates aldosterone production to retain Na⁺, Cl⁻, and volume. The major site of action of low volume-stimulated Aldo is NCC of the early distal tubule. The second responsibility of Aldo is to respond to elevated P[K⁺]. With elevating P[K⁺], Aldo eliminates K⁺ in the CNT/CCD via a process involving enhancement of the activity of ENaC, which reabsorbs Na⁺ in exchange for K⁺ secretion. The regulatory mechanisms involving the differential regulation of Na⁺ reabsorption via NCC versus ENaC are very complex and involve the WNK-SPAK system, as mentioned above in the Introduction. However, evidence is clear that high K⁺ intake, probably through an increase in P[K⁺], will inhibit Na⁺ reabsorption via NCC, resulting in increased delivery of Na⁺ to ENaC to enhance the elimination of K⁺.

As shown in Table 2, the P[K⁺] of 5.8 mM for rats on a LNaHK diet was greater compared with the value of 4.6 mM in mice. The best explanation for the hyperkalemia in rats on the LNaHK diet is that the diet is more acid producing, with Cl⁻ as the counteranion of K⁺ (150). However, the anionic content of the diet was not revealed.

Aldo Versus High P[K⁺]

When a high-K diet is consumed, elevated plasma levels of K⁺ increase the production of Aldo in the adrenal glomerulosa. Aldo regulates K⁺ secretion by increasing the activity of ENaC, which enhances Na⁺ reabsorption and, consequently, the driving force for K⁺ secretion (14, 27, 39, 85, 145). A long-standing question has been whether an increase in K⁺ secretion can be stimulated by high P[K⁺], independent of an increase in P[Aldo].

The question of Aldo-independent K⁺ secretion was addressed for LNaHK diet (0.01% Na⁺ and 5% K⁺)-fed mice, which have a P[Aldo] of >3,500 pg/ml (31). When adrenalectomized (no corticosteroid replacement), LNaHK diet-fed mice do not survive more than 2 days without Aldo replacement (31). As shown in Table 2, when given low Aldo replacement to a reach P[Aldo] of 607 pg/ml, the mice had a P[K⁺] of 7.4 mM and survived only 5 days. The K⁺ balance was maintained only with a high replacement dose of Aldo, which yielded a P[Aldo] of 4,293 pg/ml. Therefore, a very high P[Aldo] was necessary to maintain K⁺ balance in mice on a LNaHK diet.

Unlike animals on a LNaHK diet, when animals are given a normal-Na⁺, high-K⁺ diet, the K⁺ balance is maintained fairly well in the absence of Aldo. Early studies showed that adrenalectomized animals on a high-K⁺ diet maintained K⁺ balance due to high P[K⁺]-stimulated K⁺ excretion (177, 181). However, Arrighi et al. (7) found that P[Aldo] increased by only 1.5-fold with a high-K⁺ diet of normal Na⁺ and 3% K⁺, indicating that K⁺ is secreted without the need of Aldo. However, when K⁺ intake increases, the demand for Aldo increases. For example, the K⁺ balance was maintained in aldosterone synthase KO (AS^−/−) mice with normal Na⁺ and 2% dietary K⁺. When AS^−/− mice were given a 5% high-K⁺ diet, the K⁺ balance was not maintained as P[K⁺] increased (146).

The mechanism for Aldo-independent K⁺ secretion is not clearly understood but may rely on high P[K⁺] directly stimulating Na⁺-K⁺-ATPase activity in the CNT/CCD (88) and inhibiting Na⁺ reabsorption by NCC (112), thereby enhancing Na⁺-reabsorptive exchange for secreted K⁺. ENaC activity decreases as luminal Na⁺ concentration increases (98). This feedback effect is the result of increased cell Na⁺ entry because ENaC activity returns with amiloride treatment (42). However, large increases in luminal Na⁺ concentration can enhance the intracellular Na⁺ concentration of PCs (102). Moreover, if Na⁺-K⁺-ATPase is stimulated with high P[K⁺], then the intracellular Na⁺ concentration is maintained at a low level (131).. Therefore, increased ENaC and Na⁺-K⁺-ATPase activity will stimulate K⁺ secretion, causing only a mild increase in P[K⁺] that only increases Aldo by 50%.

It is not understood why P[Aldo] (and Aldo dependence) is more than fivefold in LNaHK diet-fed mice compared with high-K⁺ diet-fed mice. Based on transtubular Na⁺ gradients, estimated from urine osmolalities, the luminal Na⁺ concentration at the terminal CCD is between 30 and 40 mM in high-K⁺ diet-fed mice and only 1 mM in LNaHK diet-fed mice (169). Palmer et al. (98) found that ENaC activity was much greater in rats on a low-Na⁺ diet. Therefore, Aldo may be required for the increase in ENaC when Na⁺ intake is low. However, the Aldo demand to enhance ENaC activity may be reduced when the distal Na⁺ concentration is high.

In conclusion, high levels of P[Aldo] are required to maintain K⁺ homeostasis in LNaHK diet-fed mice. However, the Aldo-independent, P[K⁺]-stimulated K⁺ excretion of the high-K⁺ diet may be the result of an abundance of distal Na⁺ that is inhibited from reabsorbing via NCC and delivered to ENaC to stimulate K⁺ secretion. At the same time, Na⁺-K⁺-ATPase activity is stimulated by the high P[K⁺] from the basolateral side.

ANG II

Although very high Aldo is the most striking feature of the LNaHK diet, ANG II also has an important role to maintain K⁺ homeostasis. ANG II is increased in the plasma and locally in the kidney with the LNaHK diet in rats (150), probably because of low Na⁺ delivery to the macula densa. The relevance of ANG II was indicated when captopril (100 mg·kg⁻¹·day⁻¹ in drinking water) was given for the duration of the LNaHK diet (unpublished observation). Captopril treatment (n = 4) caused P[Aldo] to increase to very high levels of 9,278 ± 360 pg/ml, and P[K⁺] reached 9.23 ± 0.62 mM. Takaya et al. (141) showed that losartan, an ANG II type 1 receptor blocker given to low-Na⁺ diet-fed mice, resulted in increased P[K⁺] and a twofold increase in P[Aldo] (141). Therefore, ANG II may be necessary to increase P[Aldo] to 3,500 pg/ml; however, a very high P[Aldo] cannot compensate for the lack of ANG II, which is also required to maintain a normal P[K⁺].

The mechanism by which ANG II stimulates K⁺ secretion in LNaHK diet-fed mice can be complicated by the fact that inhibition of ANG II by captopril and losartan will decrease blood pressure with systemic effects. However, the mechanism may be connected to its direct effects on Na⁺ and K⁺ transport properties of the CNT/CCD. ANG II increases ENaC or Na⁺ reabsorption in isolated collecting ducts, independent of Aldo or any systemic effects (102). Mamenko et al. (82) showed that ANG II increased the open probability and membrane insertion of ENaC, as an independent and additive effect to that of Aldo. In isolated CCDs, ANG II application to either the luminal or basolateral side increased benzamil-sensitive Na⁺ uptake (102). In Aldo-sensitive distal nephron segments of mice, immunohistochemical staining revealed an increase in ENaC subunits (17), and patch-clamp analysis indicated an Aldo-independent increase in ENaC activity due to both low dietary Na⁺ and chronic ANG II infusion (83).

The role of ANG II in LNaHK diet-fed animals may be to insure a substantial ENaC-mediated Na⁺ reabsorptive driving force for K⁺ secretion with the minimally available luminal Na⁺ in the terminal CCD (102). A role for ANG II to maintain K⁺ homeostasis seemingly contradicts studies that have indicated that low volume-stimulated ANG II enhances NCC-mediated Na⁺ reabsorption. Since a high-K⁺ diet does not increase ANG II, its effect on NCC could explain the Aldo paradox (114, 127, 149, 151). Although ANG II may enhance NCC with low or normal P[K⁺], recent studies have shown that the high P[K⁺] of animals on a LNaHK diet should override the effects of ANG II and maximally inhibit NCC (150).

ANG II regulation of pendrin is also important for maintaining K⁺ homeostasis in animals on a LNaHK diet because it is necessary to secrete high HCO₃⁻ with K⁺. Pech et al. (99) showed with isolated perfused tubules that ANG II increased Cl⁻ absorption via pendrin. Therefore, ANG II may be another common activator that links ENaC with pendrin activity (50).

Aldo and ANG II Regulation of K⁺ Secretory Channels

Aldo enhances not only the driving force for K⁺ secretion by increasing Na⁺ reabsorption (137) but also the K⁺ conductive pathways for K⁺ secretion (65, 116) in distal nephron segments. Initial studies have revealed that rabbits treated with mineralocorticoids or Aldo exhibited enhanced K⁺ conductance of the apical membrane of isolated rabbit CCDs (65, 116). These early studies measured the apical membrane conductance for K⁺. Subsequent patch-clamp studies made it possible to understand how the individual channel types might be regulated.

ROMK and BK channels have distinct and redundant roles to secrete K⁺ in the CCD (119). Although Aldo can increase the activity of both ROMK and BK channels, there may be separate signaling mechanisms for these channels. With a high-K⁺ diet, Aldo signaling increases insertion of ROMK channels in the apical membrane (156, 163, 180) and ANG II increases ROMK activity (164). A high-K⁺ diet increases ROMK mRNA (156) and protein expression (153) in the CNT/CCD. However, mice with KO of the Aldo-induced serum and glucokinase pathway (SGK−/− mice) exhibit elevated P[K⁺], a sixfold increase in P[Aldo], and downregulation of ENaC and Na⁺-K⁺-ATPase, as expected, but ROMK expression is enhanced (58). This finding indicates that ROMK channels may be regulated by elevated P[K⁺] rather than the Aldo-induced SGK pathway. In support of this notion, Todkar et al. (146) described an increase in ROMK expression in AS−/− mice, a model of elevated P[K⁺] in the absence of Aldo.

Patch-clamp analysis revealed that ANG II inhibited ROMK activity via ANG II type 1 receptors in mice on a low-K⁺ diet (165) but increased ROMK activity via ANG II type 2 receptors in mice on a high-K⁺ diet (164). Considering that ANG II is not elevated in rats on a high-K⁺ diet (150), the direct effects of ANG II on ROMK channels may be more relevant when animals are placed on a LNaHK diet. Van der Lubbe et al. (150) did not find an increase in ROMK (protein) expression in mice on a LNaHK diet. However, the open probability of ROMK channels or amount of ROMK channels in the membrane versus cytoplasm might be increased by ANG II (164). The ANG II-induced increase in ROMK activity, along with the increased ENaC activity, supports the notion that ANG II has a role to maintain K⁺ homeostasis in mice on a LNaHK diet.

Many studies found that BK channels in a variety of epithelial cells are regulated by Aldo (52, 60, 77, 133, 171). BK channels are contained in PCs as BK-α–BK-β₁ (104) and in ICs as BK-α–BK-β₄ (96). For the kidney, considerable evidence supports the notion that Aldo regulates BK-α–BK-β₄ in ICs. A high-K⁺ diet enhanced the renal mRNA expression of BK-α and BK-β₄ (90) and expression of BK-α protein in the cytoplasm of ICs (150, 171). However, BK-α remains in the cytoplasm if the high-K⁺ diet is acidic (171). BK-α does not express in the apical membrane of ICs of β₄-KO mice on a LNaHK diet (171), which explains the considerably attenuated K⁺ excretion of β₄-KO mice.

The role of BK-β₄ in ICs is to prevent the lysosomal breakdown and promote the incorporation of BK-α into the plasma membrane (171). This trafficking function of BK-β₄ is not supported by past studies, which have shown that BK-β₄, along with BK-β₁, prevents BK-α incorporation in the plasma membrane of cultured hair (8) and neuronal (130) cells. It is uncertain whether BK-α is associated with more than one subunit (159). The β₂-subunit has been described in the CCD (90) and may also associate with BK-α in ICs. Moreover, it is now understood that palmitoylation of the β₄-subunit controls the insertion of BK-α to the membrane by masking a trafficking motif on the COOH-terminus of BK-α (26). Thus, it remains to be determined how palmitoylation of BK-β₄ is regulated in ICs. The roles and properties of β-subunits and their interactions with BK-α appear much more complex in different native cell systems than indicated in the original studies with cultured cells.

CNT cells and PCs of the CCD have been considered the primary targets of Aldo. Aldo-induced SGK (152) enhances ENaC (78) and Na⁺-K⁺-ATPase (138). However, IC-β also contains mineralocorticoid receptors as well as 11-β-steroid dehydrogenase (92), an enzyme that renders an Aldo-selective effect by inactivating glucorticoids (1, 134). These findings support the notion that Aldo regulates pendrin and BK channels in IC-β.

Although Aldo regulates IC-β of mice on a LNaHK diet, a recent study by Shibata et al. (128) reported that mineralocorticoids do not regulate IC transporters when mice are given a high-K⁺ (normal Na⁺) diet (128). Both IC-α and IC-β contain mineralocorticoid receptors with a specific site (MR^S843) that prevents Aldo receptor binding when phosphorylated. High K⁺ enhances the phosphorylation of MR^S843 via WNK4, thereby preventing an increase in Cl⁻ reabsorption via pendrin in IC-β. Low extracellular volume, which increases ANG II in addition to Aldo, decreases the phosphorylation of MR^S843, rendering mineralocorticoid receptor binding with enhanced H⁺-ATPase activity in IC-α and enhanced pendrin activity in IC-β. The model of Shibata et al. requires that IC-α secrete H⁺ that reacts with HCO₃⁻, which is secreted from nearby IC-β (128, 158). Alternatively, H⁺-ATPase secretes H⁺ and pendrin secretes HCO₃⁻ from the same non-A-non-B ICs (63). Therefore, the study of Shibata et al. reports another explanation for the low volume or high K⁺ Aldo paradox.

The Shibata et al. model explains the paradox of either high K⁺ or low volume (high ANG II)-induced Aldo secretion by a mechanism that uses either high K⁺-induced PCs or low volume-induced IC transport. However, LNaHK diet-fed mice have increased plasma levels of Aldo, P[K⁺], and ANG II. Mice on a LNaHK diet require an effect of Aldo on both PCs and ICs, which are cooperating to maximize the ratio of K⁺ secreted to Na⁺ reabsorbed (169). Aldo enhances HCO₃⁻ secretion from a predominant population of IC-β and enhances ENaC-mediated Na⁺ reabsorption, which drives K⁺ secretion from both PCs and ICs (see Fig. 2). A modification of the Shibata et al. model has been proposed in which the high ANG II and Aldo of the LNaHK condition causes the dephosphorylation of MR^S843 in only the predominating population of IC-β, allowing Aldo to enhance both pendrin and BK-α–BK-β₄ activity as well as enhance the ENaC-mediated Na⁺ reabsorptive exchange for secreted K⁺ in PCs. This effect will enhance the exchange rate of secreted K⁺ per reabsorbed Na⁺ in animals on a LNaHK diet. This model explains why mice that consume the acid-loading LNaHK-Cl⁻ diet are unable to maintain K⁺ balance. With a predominant population of IC-α with few IC-β (or non-A-non-B-cells), there will be increased Cl⁻ reabsorption with Na⁺ rather than complete exchange of K⁺ for Na⁺, as observed with a predominant population of IC-β.

Fig. 2.

Cell models illustrating the effects of the LNaHK diet on Na⁺ and K⁺ transport properties of the distal nephron. A and B: illustrations of the Na⁺ and K⁺ transport pathways of the early distal convoluted tubule (DCT1) and CNT/CCD of animals on a regular diet (A) or LNaHK diet (B). A: for mice on a regular diet, ENaC reabsorbs 50% more Na⁺ than the Na⁺-Cl⁻ cotransporter (NCC) in the distal nephron. Approximately 75% of the K⁺ transported on Na⁺-K⁺-ATPase is secreted into the lumen, and 25% is recycled across the basolateral membrane to yield a Na⁺ reabsorbed-to-K⁺ secreted ratio of 0.5. Cl⁻ is reabsorbed via pendrin to account for the charge balance with Na⁺ reabsorption. B: the high K⁺ intake or elevating P[K⁺] turns off NCC of the DCT1, increasing the delivery of Na⁺ to ENaC. Renal and plasma ANG II levels elevate in response to the low Na⁺ intake with an increase in ENaC open probability and activation of renal outer medullary K⁺ (ROMK) channels. Aldosterone (Aldo) has several effects to enhance K⁺ secretion, including increasing BK channels in ICs and Na⁺-K⁺-ATPase, ENaC, and ROMK channels in PCs. The elevated flow stimulates BK channels in PCs and ICs. To enhance the secreted K⁺-to-Na⁺ reabsorbed ratio greater than 2, it is hypothesized that Na⁺ is recycled via the Na⁺-dependent Cl⁻/HCO₃⁻ exchanger.

Although a modification of the Shibata et al. model, with a predominance of IC-β, can explain the Aldo effects on both PCs and IC-β, it cannot explain the results of studies that showed that a high-K⁺ (normal Na⁺) diet causes Aldo-dependent BK-α–BK-β₄-mediated K⁺ secretion in IC-β (171). In the high K⁺ condition, P[Aldo], but not ANG II, is elevated, and only PCs would be influenced by Aldo. However, the effects of Aldo on BK-α–BK-β₄ were only observed with the alkaline-producing high-K⁺ diet (171). It remains to be determined whether an alkaline load causes dephosphorylation of MR^S843 specifically in IC-β.

LNaHK and High Flow

Young et al. (183) originally reported that high K⁺ intake caused a profound natriuresis with hypotension. Although natriuresis is mild with low Na⁺ intake, high urinary flow is one of the most remarkable events associated with animals on a LNaHK diet (18, 30, 40, 146, 150, 169). In mice, flow increases from 1–1.5 ml/day with a regular diet to 4–6 ml/day when treated with a LNaHK diet (31). However, the mechanism underlying the high flow is uncertain.

High distal flow is necessary to eliminate a high K⁺ load. When a 12-fold normal K⁺ load is consumed, the rate of K⁺ excretion matches the input by increasing the rate of K⁺ secretion in the CNT and CCD. The rate of K⁺ secretion is composed of two components: the elevation of luminal K⁺ concentration and the rate of flow; however, the capacity to elevate luminal K⁺ concentration, which is ∼40 mM in mice on a regular diet, is limited, with a maximal transtubular K⁺ gradient of 20–25 (169). With a P[K⁺] near 4.5 mM, the luminal K⁺ concentration saturates at a value of 100–120 mM in mice on a LNaHK deit. Therefore, with an increase in luminal K⁺ concentration of only threefold (169), distal volume needs to increase by at least threefold to match the increase in K⁺ intake. High flow continually dilutes the volume and maintains the luminal K⁺ concentration near 110 mM.

While the high flow is necessary to prevent high K⁺ concentration in the CNT/CCD, conversely, the rate of K⁺ secretion must increase to match the rate of flow to maintain the high luminal K⁺ concentration. It has been shown that the continued reabsorption of Na⁺ is necessary to maintain a constant high transepithelial voltage driving force for K⁺ secretion during high flow (79). To increase Na⁺ reabsorption and transepithelial voltage with flow, it is necessary to continually activate ENaC. Isolated perfused tubule studies have shown that ENaC-mediated Na⁺ reabsorption is enhanced by flows from 0.5–3 (120) or 5 nl/min, even when ENaC membrane insertion is inhibited (87). Expression studies have shown that shear stress can increase the open probability of ENaC (3, 20, 120). Therefore, mechanical, flow-induced activation of ENaC may account for the increased Na⁺ reabsorption at a distal flow of 0.5–5 nl/min, which is considered a normal physiological range in rat distal tubules (56).

Malnic et al. (87) showed that the flow rate of the perfused distal tubule of high K⁺-treated rats stimulated K⁺ secretion to a steep level from 2 to10 nl/min, consistent with the flow range for mechanically activating ENaC and Na⁺ reabsorption. However, the rates of Na⁺ reabsorption and K⁺ secretion saturated and the luminal K⁺ concentration decreased with perfusion rates from 10 to 30 nl/min (79). The case was different with free flow micropuncture, in which high distal flow was increased by saline volume expansion. In these conditions, both Na⁺ reabsorption and K⁺ secretion continued with flows from 6 to 30 nl/min (61, 62, 68). These data indicate that the flow-induced shear stress of microperfusion increased the open probability of ENaC from 2 to 10 nl/min, as shown by the isolated tubule perfusion study by Morimoto et al. (87). However, as revealed by the free-flow experiments of Khuri et al. (61), another mechanism with more ENaC may be required for flows of >10 nl/min to increase Na⁺ reabsorption.

One explanation for the difference in the response of ENaC to the high flows of tubule perfusion and free flow is the presence of an unknown paracrine regulator of Na⁺ reabsorption in the endogenous tubular fluid. The precedence for this notion is the finding that flow-induced fluid reabsorption in the proximal tubule is exhibited with free-flow micropuncture but inconsistently, or at all, with microperfusion (49). It has been shown that if the proximal tubule is microperfused for 10 min with a solution containing the plasma ultrafiltrate, as opposed to an artificial solution, then flow-induced fluid reabsorption can be demonstrated (11). However, we have not found reports of similar studies that determined the relation between flow and ENaC-mediated Na⁺ reabsorption using plasma ultrafiltrate. A paracrine effect would be more evident with high flow in vivo as it is observed to act over the course of several minutes to increase the number of ENaC in the membrane. In contrast, the shear stress effect to increase ENaC open probability with flows from 1 to 5 nl/min could be regulated in the order of seconds.

High flow not only maintains a chemical driving force for K⁺ secretion, but also the luminal K⁺ conductive pathways are enhanced. This is accomplished by flow stimulation of the activity of BK channels in the distal nephron (9, 73, 142, 178, 178). However, very few in vivo studies have supplied evidence showing that BK channels are activated by flow. Mice with KO of the BK β₁-subunit, which is localized in CNT cells, exhibit a deficient kaliuretic response to high flow induced by NaCl volume expansion (104). In ROMK KO mice, micropuncture revealed that K⁺ secretion in the distal nephron was inhibited by the BK channel blocker iberiotoxin (9). It is not known, however, whether the BK channels were responding to high flow from reduced Na⁺ reabsorption in the TAL or compensating for the lack of ROMK-mediated K⁺ secretion in the CNT. Another study revealed that BK-α–BK-β₄ in the medullary collecting ducts were activated to reduce the volume of ICs and increase the luminal diameter of the tubules to accommodate the flow (52). However, with a LNaHK diet, BK-α–BK-β₄ secretes K⁺ as a resident of the CNT, where there is a preponderance of HCO₃⁻ secreting IC-β (89), especially under alkaline conditions. Urinary flow is decreased only slightly, but K⁺ excretion is considerably reduced in β₄-KO mice on a high-K⁺ or LNaHK diet (52). Therefore, BK-α–BK-β₄ in the medullary collecting ducts activate to reduce cell volume in response to high flow, and BK channels of the CNT/CCD mediate net K⁺ secretion with high flow.

That flow-induced K⁺ secretion is blocked by iberiotoxin, a BK channel blocker, is seemingly inconsistent with the notion that BK-α–BK-β₄ of ICs respond to flow. A previous expression study (86) with Xenopus oocytes indicated that the β₄-subunit renders BK-α insensitive to iberiotoxin. However, other more recent reports have indicated that the BK-α/BK-β₄ resistance to Iberiotoxin is cell specific (159) and that BK-α–BK-β₄ is sensitive to iberiotoxin in native cells (179). Patch-clamp experiments are needed to determine whether BK-α–BK-β₄ in IC-β is resistant to iberiotoxin.

Flow may activate BK channels by stimulating Ca²⁺ cell entry (73, 76) or releasing ATP, which increases cell Ca²⁺ through purinergic receptors (46, 55). Ca²⁺ enters the cell via flow-stimulated transient receptor potential vanilloid (TRPV)4 channels (15, 16) or ATP-activated TRPV3 channels (46). Mice with KO of TRPV4 do not exhibit flow-mediated K⁺ secretion (81, 143).

We do not completely understand how high flow is generated in mice on a LNaHK diet. High flow can be initiated in the kidney by either inhibiting Na⁺ reabsorbing transporters or increasing the osmolality of the distal tubule. Most of these Na⁺ reabsorbing transporters can be inhibited by diuretic pharmaceutical agents. Accordingly, these diuretics can be used to diagnose the origin of the site of Na⁺ inhibition, with the notion that a diuretic will be relatively ineffective if the Na⁺ transporter is already inhibited in animals on a LNaHK diet.

NKCC2 and NCC Pathways

High K⁺ consumption can enhance flow by acting as a loop diuretic inhibiting NKCC2 or inhibit the NCC like a thiazide diuretic. Previous studies have suggested that K⁺ recycling/reabsorption from the medullary region inhibits Na⁺ and Cl⁻ absorption in the TAL (12, 25, 136), thereby acting as a loop diuretic. Furosemide was given in vivo to determine whether the TAL was inhibited in mice on a LNaHK diet. Furosemide was found to inhibit the TAL in mice on a LNaHK diet, with the osmolality decreasing from 2,555 ± 522 to 717 ± 15 mosmol/L and the Na⁺ output increasing from 9.0 ± 1.4 to 98.8 ± 20.4 (160). This magnitude of change of Na⁺ and osmolality was the same as found for furosemide-treated mice on a regular diet (160). It is therefore unlikely that the transport of NKCC2 in the TAL is affected by the LNaHK diet.

NCC localized in the early distal convoluted tubule (DCT) normally reabsorbs 5% of filtered Na⁺. With a LNaHK diet, the high P[K⁺] inhibits NCC of the early DCT, acting as a thiazide diuretic (41, 112, 112, 132, 148, 150). A recent study (59) indicated that acute inhibition of NCC does not result in enhanced Na⁺ reabsorption via ENaC. Rather, long-term inhibition of NCC causes morphological adaptations of the distal nephron with enhanced ENaC-mediated reabsorption for secreted K⁺. That high K⁺-induced natriuresis was prevented in NCC KO mice is evidence that high P[K⁺] inhibits NCC (132). High P[K⁺] probably causes a slight DCT depolarization that increases intracellular Cl⁻, resulting in the dephosphorylation of NCC (144).

The question is whether the natriuresis due to inhibition of NCC can result in fivefold diuresis. An early study (80) has shown that substantial Na⁺ and Cl⁻ are delivered to NCC with a Na⁺ concentration of ∼50 mM, even with a low-Na⁺ diet. Therefore, ∼100 mosM will be delivered to the CNT/CCD. On a LNaHK diet, Na⁺ is reabsorbed by ENaC in the CNT/CCD, as indicated by <5 mM Na⁺ concentration in the urine of LNaHK diet-fed mice (169). Before water is extracted in the medullary regions, the Na⁺ concentration at the end of the CCD is <2 mM. Moreover, hydrochlorothiazide (HCTZ), an inhibitor of NCC, only decreased tubular fluid/plasma inulin from 3.79 to 2.78 in the late distal nephron of dogs (34), indicating a mild increase in flow of only 1.4-fold in the late DCT. HCTZ caused a natriuresis in normal diet-fded mice, but the urinary flow was not significantly increased (169). Thus, the natriuresis from inhibition of NCC might carry enough volume to stimulate BK channels but would probably not account for a four- to fivefold increase in flow through the CNT/CCD.

Kaliuretic Osmotic Diuresis and the IC-β Pathway

Luminal Na⁺ in the CNT/CCD of mice on a LNaHK diet is avidly reabsorbed via ENaC; however, the luminal Na⁺ and Cl⁻ are replaced by K⁺ and HCO₃⁻, such that there is no net loss of solutes. It was hypothesized that the ENaC-mediated Na⁺-dependent K⁺ secretion could be greater than the ENaC-dependent Na⁺ reabsorption. This hypothesis was tested with amiloride, an inhibitor of ENaC. It was important to establish with HPLC that the amiloride concentration of the tubular fluid totally inhibited ENaC. At a luminal concentration of >22 μM (169) and a K_i of ∼100 nM (123), ENaC activity was inhibited by >99%. When mice were placed on a normal diet and then given amiloride for 12 h, the amiloride-induced natriuresis was more than double the amount of amiloride-sensitive K⁺ excretion, consistent with a ratio for K⁺ secreted to Na⁺ reabsorbed of 0.5. However, when mice on a LNaHK diet for 7 days were given amiloride for 12 h, the amiloride-sensitive Na⁺ diuresis was ∼30% of the amount of amiloride-sensitive kaliuresis (169). This experiment indicated that more ENaC-dependent K⁺ is secreted than ENaC-dependent Na⁺ is reabsorbed in the CNT/CCD.

With the alkaline-producing LNaHK diet, the secreted K⁺ is essentially a nonreabsorbable solute because the major reabsorbing pathway in the CCD and medullary collecting ducts, H⁺-K⁺-ATPase, is inhibited (176). Thus, there is no net loss of solutes in the distal lumen when inhibiting NCC and Na⁺ is delivered to ENaC. The net addition of solute in the CNT/CCD could account for a large osmotic diuresis.

A kaliuretic osmotic diuresis could also result from Na⁺-independent K⁺ secretion, with the addition of luminal K⁺ without reabsorbing Na⁺, in the CNT/CCD. Rats exhibited Na⁺-independent K⁺ excretion when fed a high-K⁺-Cl⁻ diet (5% K⁺ and 5% Cl⁻) for 7–9 days, and amiloride treatment significantly inhibited kaliuresis (40). This diet with Cl⁻ as the counteranion produced an acid load, as indicated by the high urine Cl⁻ output. However, amiloride treatment only increased P[K⁺] to 5.5 mM, which is considered only slightly hyperkalemic. In contrast, when mice were given the high-K⁺-Cl⁻ diet for 7–10 days, they exhibited hyperkalemia (P[K⁺] = 6.85 mM), and P[K⁺] increased to lethal levels when mice were given amiloride (169). These results suggest that rats, unlike mice, may have a very efficient way to adapt to the high-K⁺-Cl⁻ diet.

The question arises regarding how more K⁺ secretion is inhibited than Na⁺ reabsorbed when inhibiting ENaC when the Na⁺-K⁺-ATPase ratio of PCs is only 0.67 (2 K⁺ for 3 Na⁺). It has been suggested that K⁺ is secreted from a parallel pathway from ICs when mice are on a LNaHK diet (169). IC-β normally secrete HCO₃⁻ via pendrin, do not contain ENaC, and contain minimal, if any, Na⁺-K⁺-ATPase. It was therefore very surprising when amiloride caused a much greater reduction in K⁺ secretion in WT mice compared with β₄-KO mice. These data supported the notion that BK-α–BK-β₄ of ICs mediate ENaC-dependent K⁺ secretion in mice on a LNaHK diet.

A hypothetical mechanism, as shown in Fig. 2, was previously presented that describes a parallel path that is dependent on ENaC-mediated Na⁺ reabsorption yet results in a ratio of K⁺ secretion per Na⁺ reabsorbed of much greater than 2:3 and could be as much as 3:1 (169). The model was partly based on recent findings by Chambrey et al. (22) showing that transepithelial transport is powered by H⁺-ATPase, which helps establish an electronegative cell potential and sets up the driving force for HCO₃⁻ secretion via pendrin or one of the recently discovered members of the solute carrier (SLC)4 family across the apical cell membrane (35, 72) and the exit of Na⁺ via Na⁺-HCO₃⁻ cotransporter 1 (NBCn1) in the basolateral cell membrane. This model was established with chemical gradients that supported reabsorption of Na⁺ and Cl⁻. However, the transporters would reverse when faced with the chemical driving forces for animals on a LNaHK diet.

With a Na⁺-deficient diet, the urine Na⁺ concentration is reduced to <5 mM in rats (129) and mice (169). Accounting for water extraction, the Na⁺ concentration in the terminal CCD is <2 mM. The low luminal Na⁺ concentration establishes a chemical gradient for Na⁺ to passively recycle back through a paracellular pathway and be pumped back through PCs. Na⁺ can enter IC-β via NBCn1 in the basolateral membrane (22) or NKCC1. Although NBCn1 has been shown to extrude Na⁺ across the basolateral membrane of ICs, NBCn1 carries Na⁺ into the TAL (70). NKCC1 has been described functionally with the isolated perfused tubule (74) and was localized with immunohistochemistry in the basolateral membrane of ICs (45). That NKCC1 is involved in K⁺ excretion was indicated by Wall et al. (157), who showed that NKCC1 KO mice, compared with WT mice, on a regular K⁺ diet exhibited decreased K⁺ excretion and a significant increase in P[K⁺] (157). However, the increased P[K⁺] (to 4.6–4.8 mM) was still in the normal range. It would be interesting to determine the K⁺ handling of NKCC1 KO mice when treated with a high-K⁺ diet. A role for basolateral NKCC1 in K⁺ secretion has also been demonstrated for the intestine (48) and a variety of secretory epithelial cells (37).

K⁺ and Cl⁻ can enter ICs via NKCC1 and exit via apical K⁺ and basolateral Cl⁻ channels, respectively. However, it is not understood how Na⁺ exits ICs with levels of Na⁺-K⁺-ATPase too low to support transport. The best candidate for Na⁺ exit is the luminal Na⁺-dependent Cl⁻/HCO₃⁻ exchanger (45, 74), with chemical forces driving the HCO₃⁻ and Na⁺ back to the lumen. In the intestine, Na⁺ is also recycled from the blood to the lumen compartment (175). Moreover, a substantial Na⁺ permeability for Na⁺ backflux has been shown with isolated perfused CCDs (124, 125, 135). With Na⁺ secreting/recycling, more net K⁺ is secreted than net Na⁺ reabsorbed, thereby causing an addition of K⁺ solute that is nonreabsorbable. However, this mechanism is only hypothetically based on previously identified transporters and needs to be verified with in vivo and in vitro studies.

It is difficult to compare recent in vivo studies using LNaHK diets with earlier isolated perfused tubule experiments because most of these early studies focused on direct effects of Aldo on transport properties of the CCD by treating animals with Aldo or the synthetic mineralocorticoid DOCA (65, 94, 115, 116, 124, 124, 135). Using electrophysiological analysis with microelectrodes, DOCA treatment elicited a threefold increase in the amounts of secreted K⁺ and reabsorbed Na⁺ in the isolated perfused rabbit CCD (117). The ratio of K⁺ secreted to Na⁺ reabsorbed was ∼0.5 in both control and mineralocorticoid-treated rabbits. A ratio of K⁺ secreted to Na⁺ reabsorbed of 0.5 is much less than the ratio of >1 found for LNaHK diet-fed mice. However, with exogenous mineralocorticoid treatment, the plasma K⁺ concentration is reduced and the volume expansion reduces ANG II. It remains to be determined with isolated perfused tubules whether the high ratio of K⁺ secreted to Na⁺ reabsorbed indicated in vivo with the LNaHK diet can be demonstrated with the isolated CNT or CCD.

In summary, three different mechanisms, including natriuresis from inhibition of Na⁺ reabsorption in the TAL, natriuresis from inhibition of NCC in the DCT, and a kaliuretic osmotic kaliuresis, can contribute to the high flows associated with a high-K⁺ diet. It is possible that the natriuretic effects increase flow to reach a threshold value that is enough to activate BK channels, which contribute to the kaliuresis.

LNaHK Diet and Diuretic Therapy

With the LNaHK diet, Na⁺ reabsorption along the nephron is altered to maximize the use of the available Na⁺ that is necessary for exchange with secreted K⁺. It should not be surprising, therefore, that diuretics will have a different effect on electrolyte homeostasis in patients on LNaHK than in patients on a Western diet. Diet may explain hyperkalemia, instead of the expected hypokalemia, that often results from antihypertensive diuretics meant to reduce extracellular fluid volume (111, 139). When disregarding the patient's diet, administration of loop, thiazide, and K⁺-sparing diuretics could give unexpected changes in P[K⁺] and can lead to increased morbidity, expense, and time wasted in an attempt to obtain the correct response. We found several differences in the electrolyte response to three diuretics (furosemide, HCTZ, and amiloride) in mice on a LNaHK diet that were not observed in mice on a regular diet.

The loop diuretic furosemide, which inhibits NKCC2 of the TAL, caused a natriuresis and small reduction in P[K⁺] in mice on a regular diet. Decreased P[K⁺] is a prevalent complication of loop diuretics as high flow presumably causes BK channel-mediated K⁺ secretion. In mice on a LNaHK diet, furosemide still caused a small natriuresis, despite low Na⁺ intake, and decreased urine osmolality. Unexpectedly, however, furosemide inhibited K⁺ excretion in mice on a LNaHK diet (160). The mechanism for the inhibition of K⁺ excretion by furosemide is not understood and is under investigation. We conclude, however, that furosemide, normally considered a K⁺-wasting diuretic, may actually prevent K⁺ excretion in subjects on an ancient diet.

A striking feature of the LNaHK diet is the total ineffectiveness of thiazide diuretics, such as HCTZ. HCTZ normally inhibits NCC-mediated Na⁺ and Cl⁻ reabsorption in the DCT (169). In the normal diet condition, NCC is phosphorylated and traffics to the apical membrane so that Na⁺ and Cl⁻ are reabsorbed. It is now well documented that high K⁺ intake results in a dephosphorylation and reduced abundance of NCC (112, 132, 150) and low external K⁺ increases the phosphorylation of NCC (144, 154). Therefore, the high K⁺ intake and elevated P[K⁺] have already effectively turned off the NCC, rendering HCTZ ineffective as a diuretic.

The ENaC inhibitor amiloride also has a different effect in mice on a LNaHK diet. Amiloride is a more effective diuretic than HCTZ and reduces K⁺ excretion and causes a natriuresis in mice on a regular diet (170). However, in mice on a LNaHK diet, amiloride decreases K⁺ excretion more than it increases Na⁺ excretion. Because ENaC-dependent K⁺ secretion per Na⁺ reabsorption is close to 3, creating an osmotic diuresis, amiloride treatment tended to reduce urinary volume of mice on a LNaHK diet (169). Even with a very large increase in Na⁺ excretion, amiloride decreased flow and urine osmolality due to the large decrease in K⁺ excretion.

Conclusions

The LNaHK diet is an exaggerated form of present day “ancient” diets, but in vivo studies have revealed underlying physiological mechanisms for the regulation of K⁺ balance that have not been considered previously. Maintenance of K⁺ balance with a LNaHK diet requires that the diet is alkaline because HCO₃⁻ excretion plays a major role to assist in BK-α–BK-β₄-mediated K⁺ secretion from IC-β. In contrast to animals on a high-K⁺ diet, the LNaHK diet requires a very high P[Aldo] of >10-fold normal to maintain K⁺ balance. However, with normal, high K⁺, the demand for Aldo is minimal as the elevating P[K⁺] stimulates Na⁺-K⁺-ATPase in the CCD and inhibits NCC, causing an increased delivery of Na⁺ to ENaC. Both ANG II and Aldo are required to maximize the ENaC-mediated Na⁺ reabsorptive driving force to maintain K⁺ balance in mice on the LNaHK diet. While ANG II may increase Cl⁻ reabsorption from IC-β and IC-α in animals on a LNaHK-Cl⁻ diet, ANG II may dephosphorylate mineralocorticoid receptors and enhance Aldo effects on K⁺ secretion from IC-β in animals on a LNaHK diet. The high urinary flow of mice on a LNaHK diet may be attributed to a combination of inhibiting NCC-mediated Na⁺ reabsorption and the high excretion rate of K⁺ in the CNT/CCD, causing an osmotic diuretic flow that activates BK channels.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: R.J.C. and S.C.S. conception and design of research; R.J.C. performed experiments; R.J.C., B.W., and J.W.-F. analyzed data; R.J.C., B.W., and J.W.-F. interpreted results of experiments; R.J.C., B.W., and S.C.S. prepared figures; R.J.C. drafted manuscript; R.J.C., B.W., J.W.-F., and S.C.S. edited and revised manuscript; R.J.C., B.W., J.W.-F., and S.C.S. approved final version of manuscript.