*Article* **MST3 Involvement in Na<sup>+</sup> and K<sup>+</sup> Homeostasis with Increasing Dietary Potassium Intake**

**Chee-Hong Chan <sup>1</sup> , Sheng-Nan Wu <sup>2</sup> , Bo-Ying Bao 3,4 , Houng-Wei Li <sup>5</sup> and Te-Ling Lu 3,\***


**Abstract:** K + loading inhibits NKCC2 (Na-K-Cl cotransporter) and NCC (Na-Cl cotransporter) in the early distal tubules, resulting in Na<sup>+</sup> delivery to the late distal convoluted tubules (DCTs). In the DCTs, Na<sup>+</sup> entry through ENaC (epithelial Na channel) drives K<sup>+</sup> secretion through ROMK (renal outer medullary potassium channel). WNK4 (with-no-lysine 4) regulates the NCC/NKCC2 through SAPK (Ste20-related proline-alanine-rich kinase)/OSR1 (oxidative stress responsive). K<sup>+</sup> loading increases intracellular Cl−, which binds to the WNK4, thereby inhibiting autophosphorylation and downstream signals. Acute K<sup>+</sup> loading-deactivated NCC was not observed in Cl−-insensitive WNK4 mice, indicating that WNK4 was involved in K<sup>+</sup> loading-inhibited NCC activity. However, chronic K + loading deactivated NCC in Cl−-insensitive WNK4 mice, indicating that other mechanisms may be involved. We previously reported that mammalian Ste20-like protein kinase 3 (MST3/STK24) was expressed mainly in the medullary TAL (thick ascending tubule) and at lower levels in the DCTs. MST3−*/*<sup>−</sup> mice exhibited higher ENaC activity, causing hypernatremia and hypertension. To investigate MST3 function in maintaining Na+/K<sup>+</sup> homeostasis in kidneys, mice were fed diets containing various concentrations of Na<sup>+</sup> and K<sup>+</sup> . The 2% KCl diets induced less MST3 expression in MST3−*/*<sup>−</sup> mice than that in wild-type (WT) mice. The MST3−*/*<sup>−</sup> mice had higher WNK4, NKCC2- S130 phosphorylation, and ENaC expression, resulting in lower urinary Na<sup>+</sup> and K<sup>+</sup> excretion than those of WT mice. Lower urinary Na<sup>+</sup> excretion was associated with elevated plasma [Na<sup>+</sup> ] and hypertension. These results suggest that MST3 maintains Na+/K<sup>+</sup> homeostasis in response to K<sup>+</sup> loading by regulation of WNK4 expression and NKCC2 and ENaC activity.

**Keywords:** MST3; STK24; high potassium; ENaC; NKCC2; SPAK; OSR1; WNK4

## **1. Introduction**

An increase in dietary K<sup>+</sup> intake stimulates aldosterone release, which stimulates renal K+ secretion, and does not influence Na<sup>+</sup> retention. Several Na<sup>+</sup> and K<sup>+</sup> channels coordinate to maintain K<sup>+</sup> secretion without Na<sup>+</sup> retention. The model of the process suggests that K<sup>+</sup> loading inhibits NKCC2 in the loop of Henle [1,2] and NCC activation [3,4], thus, the Na<sup>+</sup> is delivered to the distal nephron. The increased Na<sup>+</sup> in the distal nephron stimulates K<sup>+</sup> secretion through ROMK due to an electrochemical gradient generated by reabsorption of Na<sup>+</sup> through ENaC.

These channels are regulated by a group of serine/threonine kinases, WNKs. Mutations in WNK1 and WNK4 genes cause a hereditary disease known as pseudohypoaldosteronism type II (PHAII) characterized with hyperkalemic hypertension [5]. WNK4 is a physiological Cl<sup>−</sup> sensor that manipulates dietary K<sup>+</sup> intake [6] and regulates NCC

**Citation:** Chan, C.-H.; Wu, S.-N.; Bao, B.-Y.; Li, H.-W.; Lu, T.-L. MST3 Involvement in Na<sup>+</sup> and K<sup>+</sup> Homeostasis with Increasing Dietary Potassium Intake. *Int. J. Mol. Sci.* **2021**, *22*, 999. https://doi.org/ 10.3390/ijms22030999

Academic Editor: Pasquale Strazzullo Received: 15 December 2020 Accepted: 15 January 2021 Published: 20 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

activation through downstream kinases SAPK and OSR1 [7]. Chloride efflux from the cells occurs in K<sup>+</sup> deficiency, resulting in low intracellular Cl<sup>−</sup> ([Cl−]<sup>i</sup> ) that stimulates WNK4 kinase, which phosphorylates SPAK and thus induces NCC phosphorylation [8,9]. The acute K<sup>+</sup> loading by oral gavage dephosphorylates NCC in wild-type (WT) mice after 30 min oral gavage of K<sup>+</sup> . The decrease in phospho-NCC is not observed in the WNK4-Cl− insensitive knock-in mice. These results indicated that high extracellular K<sup>+</sup> by increasing [Cl−]<sup>i</sup> inhibits WNK4 and thus inactivates NCC [6]. Interestingly, long-term K<sup>+</sup> loading still dephosphorylates NCC in the WNK4-Cl− insensitive knock-in mice, indicating that other molecules may be involved in HK-inhibited WNK4 and its downstream signaling [6].

WNK4 also modulates NKCC2 activity. NKCC2 abundance and NKCC2 activity are lower in WNK4−*/*<sup>−</sup> mice than that in controls [10]. Phosphorylation of NKCC2 is regulated by Ste-20 family kinases, including SPAK [11] and OSR1 [12]. SPAK mutant mice have a SPAK-activation deficiency, manifest reduced NKCC2 phosphorylation at T96, and are substantially hypotensive [13]. In addition to T96, NKCC2 overexpressed in the cells is phosphorylated at S91, T100, T105, and S130 by SPAK/OSR1 activation under hypotonic low-chloride conditions. Mutation of T105 or S130 reduces NKCC2 activity by 30–40% [14]. NKCC2 is known to account for approximately 20–25% of Na<sup>+</sup> reabsorption in the kidney [15], and phosphorylation of T105 and S130 plays the most important role in stimulation of NKCC2 activity. However, S130 phosphorylation has not been detected in the mouse kidney and the mechanism of regulation of phosphorylation of NKCC2 at S130 in vivo is unclear [14].

Na<sup>+</sup> delivered from K<sup>+</sup> loading-inhibited NCC and NKCC2 is reabsorbed through ENaC. Hence, K<sup>+</sup> loading induces ENaC expression and increases the channel activity to prevent Na<sup>+</sup> loss. ENaC is composed of the α, β, and γ subunits, which are delivered to the apical surface after the synthesis. The activity of ENaC at the apical surface is regulated by proteases, which cleave the α- and γ-ENaC subunits to increase open probability of the channel [16]. An increase in dietary K<sup>+</sup> intake significantly increased both ENaC and ROMK currents; however, K<sup>+</sup> loading-induced stimulation of Na<sup>+</sup> and K<sup>+</sup> currents was smaller in mice carrying PHAII-mimicking mutations [17]. These results indicate that molecules downstream of WNK4 may be involved in K<sup>+</sup> loading-regulated ENaC and ROMK activity.

We found that MST3 expression is higher in the medullary thick ascending limb (TAL) than that in the distal convoluted tubules (DCTs) in mice [18], and MST3 protects MDCK cells from physiological hypertonic stress in vitro [19]. To investigate whether MST3 was involved in ion homeostasis, we generated MST3-targeted mutant mice. Since complete knockout of MST3 was not achievable, we reported the phenotype of MST3 hypomorphic mice (referred to as MST3−*/*<sup>−</sup> mice) that manifested enhanced ENaC activity and hypertension [18]. These results indicated that MST3, similar to other Ste20 family members, played an important role in the maintenance of Na<sup>+</sup> homeostasis. In the present study, we investigated whether MST3 was involved in the regulation of Na<sup>+</sup> and K<sup>+</sup> homeostasis in response to K<sup>+</sup> loading. The expression levels of WNK4, ROMK, BK, ENaC, NCC, and NKCC2 were determined in mice fed the control and HK diets. Plasma Na<sup>+</sup> and urinary Na+/K<sup>+</sup> excretion were also assayed.

#### **2. Results**

#### *2.1. An Increase in MST3 Levels in Mouse Kidneys with Increasing K<sup>+</sup> Intake*

Since we previously reported that MST3−*/*<sup>−</sup> mice have higher ENaC activity [18], we hypothesized that MST3−*/*<sup>−</sup> mice have higher ability to reabsorb Na<sup>+</sup> with a low-Na (LNa, 0.04% Na) diet challenge. To preserve Na<sup>+</sup> in Na<sup>+</sup> deficiency, WT (MST3*+/+*) mice reduced urinary Na<sup>+</sup> excretion from 782.62 ± 152.57 to 90.89 ± 61.78 µmol/day (Figure 1A,B) and reduced urine volume from 5345.00 <sup>±</sup> 861.55 to 3748.67 <sup>±</sup> 789.03 <sup>µ</sup>L/day (Figure 1C,D). MST3−*/*<sup>−</sup> mice also preserved Na<sup>+</sup> through a reduction in urinary Na<sup>+</sup> excretion from 589.13 ± 90.61 to 61.78 ± 22.26 µmol/day (Figure 1A,B) and urine volume from 3958.00 ± 1047.19 to 2729.31 ± 861.14 µL/day (Figure 1C,D). The urinary Na<sup>+</sup> excretion ratio of low Na (LNa) to control diets was 12.11 ± 3.8% in WT mice and

10.42 <sup>±</sup> 2.7% in MST3−*/*<sup>−</sup> mice. The urine volume ratio of LNa to control diets was 70.58 <sup>±</sup> 13.19% in WT mice and 67.55 <sup>±</sup> 12.91% in MST3−*/*<sup>−</sup> mice. There were no differences between WT and MST3−*/*<sup>−</sup> mice in Na<sup>+</sup> reabsorption in Na<sup>+</sup> deficiency. However, we previously reported that MST3 protein level is upregulated in WT mice after high-salt (HS) intake (8% Na, 1.1% K) [19]. The HS diet-fed mice intake two-fold higher levels of water and chow than those animals fed the control diets (Table 1), indicating that the animals intake higher levels of both Na<sup>+</sup> and K<sup>+</sup> . To investigate the effects of Na<sup>+</sup> or K<sup>+</sup> separately, we fed mice with increasing Na<sup>+</sup> and increasing K<sup>+</sup> by adding 1% NaCl and 1% KCl in drinking water to determine the effects of Na<sup>+</sup> and K<sup>+</sup> on MST3. The MST3 expression was similar in kidney to that of control, LNa or 1.43% Na diet (0.43% Na in chow with additional 1% NaCl in drinking water) (Figure 2A). Interestingly, 2% KCl (1% K in chow with additional 1% KCl in drinking water) stimulated MST3 expression (Figure 2B). To determine the effect of K<sup>+</sup> -loading on MST3 function in the kidney, we fed WT and MST3−*/*<sup>−</sup> mice 2% KCl diets. The 2% KCl diets stimulated an approximately 1.6-fold increase in MST3 expression in WT mice (Figure 2C, lanes 4–6). MST3−*/*<sup>−</sup> mice consistently expressed a lower level of MST3 than that in WT mice (Figure 2C, lanes 7–9); however, we observed only a 1.1-fold increase in MST3 expression in 2% KCl diet-fed MST3−*/*<sup>−</sup> mice (Figure 2C, lanes 10–12). MST3*−/−* mice also preserved Na+ through a reduction in urinary Na+ excretion from 589.13 ± 90.61 to 61.78 ± 22.26 μmol/day (Figure 1A,B) and urine volume from 3958.00 ± 1047.19 to 2729.31 ± 861.14 μL/day (Figure 1C,D). The urinary Na+ excretion ratio of low Na (LNa) to control diets was 12.11 ± 3.8% in WT mice and 10.42 ± 2.7% in MST3*−/−* mice. The urine volume ratio of LNa to control diets was 70.58 ± 13.19% in WT mice and 67.55 ± 12.91% in MST3*−/−* mice. There were no differences between WT and MST3*−/−* mice in Na+ reabsorption in Na+ deficiency. However, we previously reported that MST3 protein level is upregulated in WT mice after high-salt (HS) intake (8% Na, 1.1% K) [19]. The HS diet-fed mice intake two-fold higher levels of water and chow than those animals fed the control diets (Table 1), indicating that the animals intake higher levels of both Na+ and K+. To investigate the effects of Na+ or K+ separately, we fed mice with increasing Na+ and increasing K+ by adding 1% NaCl and 1% KCl in drinking water to determine the effects of Na+ and K+ on MST3. The MST3 expression was similar in kidney to that of control, LNa or 1.43% Na diet (0.43% Na in chow with additional 1% NaCl in drinking water) (Figure 2A). Interestingly, 2% KCl (1% K in chow with additional 1% KCl in drinking water) stimulated MST3 expression (Figure 2B). To determine the effect of K+-loading on MST3 function in the kidney, we fed WT and MST3*−/−* mice 2% KCl diets. The 2% KCl diets stimulated an approximately 1.6-fold increase in MST3 expression in WT mice (Figure 2C, lanes 4–6). MST3*−/−* mice consistently expressed a lower level of MST3 than that in WT mice (Figure 2C, lanes 7–9); however, we observed only a 1.1-fold increase in MST3 expression in 2% KCl diet-fed MST3*−/−* mice (Figure 2C, lanes 10–12).

**Figure 1.** Urinary Na+ excretion and urine volume on a low Na (LNa) diet. Wild-type (WT) (MST*+/+*) (●) and MST3*<sup>−</sup>/<sup>−</sup>* (○) mice were fed the control diet for 3 days; then, the diet was changed to the LNa diet for an additional 16 days. The average group values were used to generate the graphs, and the error bars correspond to SE. Urinary Na+ excretion (**A**) and urine volume (**C**) were recorded over 9 days. The bar graphs (**B**,**D**) show the average of the data in (**A**,**C**), respectively. **Figure 1.** Urinary Na<sup>+</sup> excretion and urine volume on a low Na (LNa) diet. Wild-type (WT) (MST*+/+*) (•) and MST3−*/*<sup>−</sup> (#) mice were fed the control diet for 3 days; then, the diet was changed to the LNa diet for an additional 16 days. The average group values were used to generate the graphs, and the error bars correspond to SE. Urinary Na<sup>+</sup> excretion (**A**) and urine volume (**C**) were recorded over 9 days. The bar graphs (**B**,**D**) show the average of the data in (**A**,**C**), respectively.

(**A**)

(**B**)

(**C**)

(**D**)

**Figure 2.** A 2% KCl diet stimulates MST3 expression. MST3 protein expression was detected in WT mice fed the (**A**) control (n = 3 or 4 animals), low Na (LNa), 1.43% Na or (**B**) 2% KCl diets for 16 days (n = 3 animals/group)*.* The sodium, potassium, chloride and water contents in the diet are indi-


**Table 1.** The values of weight, water intake, and food intake in 8-week-old C57Bl/6 male mice fed control or high salt diets.

*Int. J. Mol. Sci.* **2021**, *22*, x FOR PEER REVIEW 4 of 15

(**D**)

**Figure 2.** A 2% KCl diet stimulates MST3 expression. MST3 protein expression was detected in WT mice fed the (**A**) control (n = 3 or 4 animals), low Na (LNa), 1.43% Na or (**B**) 2% KCl diets for 16 days (n = 3 animals/group)*.* The sodium, potassium, chloride and water contents in the diet are indi-**Figure 2.** A 2% KCl diet stimulates MST3 expression. MST3 protein expression was detected in WT mice fed the (**A**) control (n = 3 or 4 animals), low Na (LNa), 1.43% Na or (**B**) 2% KCl diets for 16 days (n = 3 animals/group). The sodium, potassium, chloride and water contents in the diet are indicated. (**C**) MST3 protein expression was detected in WT and MST3−*/*<sup>−</sup> mice fed the control and 2% KCl diets (n = 3 animals/group). (**D**) The bar graph shows quantification of MST3 expression in (**C**). \* *p* < 0.05 vs. the control group.

#### *2.2. Reduction of Diuresis, Kaliuresis, and Natriuresis in MST3*−/<sup>−</sup> *Mice Fed 2% KCl Diets*

Consistent with the notion that K<sup>+</sup> loading causes diuresis and kaliuresis [20], K + loading induced a rapid diuresis in WT mice (urine volume from 4812.22 ± 695.51 to 5700.80 ± 792.40 µL) on day 1 of K<sup>+</sup> loading and remained at 6813.83 <sup>±</sup> 2229.64 <sup>µ</sup>L during days 2–6 of K<sup>+</sup> loading. However, the urine volume of MST3−*/*<sup>−</sup> mice was increased only from 3884.58 <sup>±</sup> 695.51 to 5875.00 <sup>±</sup> 993.79 <sup>µ</sup>L on day 1 of K<sup>+</sup> loading and was reduced to 3652.50 ± 893.94 µL on subsequent days of K<sup>+</sup> loading (Figure 3A,B). A kaliuresis was also observed from 1082.92 <sup>±</sup> 136.58 to 1692.90 <sup>±</sup> 136.58 <sup>µ</sup>mol on day 1 of K<sup>+</sup> loading and remained at 1903.83 ± 435.69 µmol/d in WT mice (Figure 3C,D). These results indicated that in response to increasing K<sup>+</sup> intake, the kidneys normally excreted approximately 1.75-fold of K<sup>+</sup> , which was between 90 to 95% of the daily 2-fold increase of K<sup>+</sup> intake (Figure 3D); however, urinary K<sup>+</sup> excretion in MST3−*/*<sup>−</sup> mice was substantially increased from948.49 <sup>±</sup> 105.69 to 1873.12 <sup>±</sup> 370.63 <sup>µ</sup>L on day 1 of K<sup>+</sup> loading and was then slightly increased to 1230.35 ± 205.69 µL (Figure 3C,D). Comparison of the fold change of the urine volume on K<sup>+</sup> loading to the control diets indicated an approximately 1.4-fold increase after K<sup>+</sup> loading in WT mice. The urine volume in K<sup>+</sup> loading-treated MST3−*/*<sup>−</sup> mice was only 0.8-fold of that in control diet-fed mice, significantly less than the 1.4-fold increase in WT mice (*p =* 1.8 <sup>×</sup> <sup>10</sup>-5) (Figure 3). When comparing the fold change of K<sup>+</sup> secretion on K<sup>+</sup> loading, only a 1.29-fold increase was detected in MST3−*/*<sup>−</sup> mice, which was significantly less than a 1.75-fold increase in WT mice (*p =* 4.0 <sup>×</sup> <sup>10</sup>-6) (Figure 3D), indicating that MST3−*/*<sup>−</sup> mice exhibited reduced diuresis and kaliuresis than that in WT mice on K<sup>+</sup> loading.

The urinary Na<sup>+</sup> excretion in WT mice was slightly reduced from 763.10 ± 87.68 to 680.59 <sup>±</sup> 44.45 <sup>µ</sup>mol on day 1 of K<sup>+</sup> loading and returned back to 712.87 ± 102.34 µmol/d on subsequent days of K<sup>+</sup> loading, indicating that Na<sup>+</sup> was maintained at homeostasis. In contrast, urinary Na<sup>+</sup> excretion was significantly decreased in MST3−*/*<sup>−</sup> mice from 663.74 <sup>±</sup> 79.60 to 532.68 <sup>±</sup> 71.95 <sup>µ</sup>mol after K<sup>+</sup> loading (Figure 3E,F) and reduced to approximately 0.8-fold of that in mice fed control diet; these levels were significantly less than those in WT mice (0.93-fold change, *p =* 0.0006) (Figure 3F). These results indicated that MST3−*/*<sup>−</sup> mice reabsorbed higher amounts of Na<sup>+</sup> and water than those in WT mice on 2% KCl diets. This increase in Na<sup>+</sup> reabsorption was associated with an increase in the plasma [Na<sup>+</sup> ] (in mM, 154.67 in MST3−*/*<sup>−</sup> mice vs. 152.5 in MST3*+/+* mice) (Table 2). Overall, the Na<sup>+</sup> - and flow-dependent K<sup>+</sup> secretion was inhibited in MST3−*/*<sup>−</sup> mice. After K<sup>+</sup> loading, systolic blood pressure (SBP) of WT mice was 119 ± 10 mm Hg, which was similar to the SBP in mice fed the control diet; however, SBP of MST3−*/*<sup>−</sup> mice was <sup>131</sup> <sup>±</sup> 9 mm Hg (Table 2). These results suggested that only a 1.1-fold increase in MST3 in MST3−*/*<sup>−</sup> mice fed the 2% KCl diets was insufficient to excrete Na<sup>+</sup> and K<sup>+</sup> , causing elevated SBP in MST3−*/*<sup>−</sup> mice.


**Table 2.** Blood pressure (BP) in WT and MST3−*/*<sup>−</sup> mice on control diets for 3 days and challenged with 2% KCl diets. \* *p* < 0.05 vs. the WT group.

**Figure 3.** Reduction of diuresis, kaliuresis, and natriuresis in MST3*<sup>−</sup>/<sup>−</sup>* mice fed 2% KCl diets. MST*+/+* (●) and MST3*<sup>−</sup>/<sup>−</sup>* (○) mice were fed a control diet for 3 days; then, the diet was changed to the 2% KCl diet for an additional 16 days. The urine volume (**A**), urinary K+ (**C**), and urinary Na+ (**E**) excretion were recorded over 9 days. The average group values were used to generate the graphs, and the error bars correspond to SE. The bar graphs (**B**,**D**,**F**) show the average of the data in (**A**,**C**,**E**) respectively. **Figure 3.** Reduction of diuresis, kaliuresis, and natriuresis in MST3−*/*<sup>−</sup> mice fed 2% KCl diets. MST*+/+* (•) and MST3−*/*<sup>−</sup> (#) mice were fed a control diet for 3 days; then, the diet was changed to the 2% KCl diet for an additional 16 days. The urine volume (**A**), urinary K<sup>+</sup> (**C**), and urinary Na<sup>+</sup> (**E**) excretion were recorded over 9 days. The average group values were used to generate the graphs, and the error bars correspond to SE. The bar graphs (**B**,**D**,**F**) show the average of the data in (**A**,**C**,**E**) respectively.

duced kaliuresis.

mice on 2% KCl diets (Figure 4D, h).

with 2% KCl diets. \* *p* < 0.05 vs. the WT group.

**Table 2.** Blood pressure (BP) in WT and MST3*<sup>−</sup>/<sup>−</sup>* mice on control diets for 3 days and challenged

Plasma [Na+] (mmol/L) control diet 152.33 ± 0.52 153.4 ± 0.89 \*

Blood pressure (mmHg) control diet 117 ± 9 130 ± 13 \*

2% KCl diets 152.5 ± 0.84 154.67 ± 1.53 \*

2% KCl diets 118 ± 10 131 ± 9 \*

2.3, WNK4 and WNK4-regulated channels in 2% KCl diet-fed mice. WNK4 plays an important role in modulating renal K+ secretion and Na+ absorption. We found that 2% KCl diets induced increased WNK4 expression in MST3−/− mice (Figure 4A), indicating that MST3 might be involved in WNK4 regulation. WNK4 have been shown to inhibit ROMK activity by stimulating clathrin-mediated endocytosis [21] and inhibiting maxi-K by a kinase-dependent mechanism [22]. Figure 4B showed that BK and ROMK were increased after K+ loading in WT mice. In MST3*−/<sup>−</sup>* mice, there was no obvious difference in BK expression; however, ROMK was not induced after K+ loading, which may cause re-

Consistent with the notion that K+ loading induced ENaC γ-cleavage [6], 2% KCl diets induced increased cleaved γ-ENaC expression in WT mice, indicating that 2% KCl diets increased ENaC activity. Compared with WT mice, MST3*−/−* mice had a higher cleaved γ-ENaC on control diets and higher full-length ENaC on 2% KCl diets (Figure 4C). IHC results showed that feeding the 2% KCl diets slightly increased γ-ENaC expression at the apical plasma membrane of the DCT2/CNT in WT mice compared with that in mice fed the control diets (Figure 4D, c vs. a). Consistent with our previous report, MST3*−/−* mice exhibited higher ENaC expression at the apical plasma membrane of the DCT2/CNT. A higher intensity of ENaC staining was observed on 2% KCl diets (Figure 4D, g vs. e). The MST3 expression showed that higher levels of MST3 protein were present in the cytosol of the DCT2/CNT in WT mice on 2% KCl diets compared to that in WT mice on control diets (Figure 4D, d vs. b). Lower levels of MST3 were observed in MST3*−/<sup>−</sup>*

Isoforms A, B, and F of NKCC2 are estimated to account for 20–25% of all renal Na+ reabsorption. The NKCC2-F isoform mainly located in the inner medullary TAL accounts for 70% of NKCC2 expression [23]. Since MST3 is primarily localized in the inner me-

**Value 16 Days MST3***+/+* **(WT) MST3 −/<sup>−</sup>**

#### *2.3. WNK4 and WNK4-Regulated Channels in 2% KCl Diet-Fed Mice*

WNK4 plays an important role in modulating renal K<sup>+</sup> secretion and Na<sup>+</sup> absorption. We found that 2% KCl diets induced increased WNK4 expression in MST3−*/*<sup>−</sup> mice (Figure 4A), indicating that MST3 might be involved in WNK4 regulation. WNK4 have been shown to inhibit ROMK activity by stimulating clathrin-mediated endocytosis [21] and inhibiting maxi-K by a kinase-dependent mechanism [22]. Figure 4B showed that BK and ROMK were increased after K<sup>+</sup> loading in WT mice. In MST3−*/*<sup>−</sup> mice, there was no obvious difference in BK expression; however, ROMK was not induced after K<sup>+</sup> loading, which may cause reduced kaliuresis.

Consistent with the notion that K<sup>+</sup> loading induced ENaC γ-cleavage [6], 2% KCl diets induced increased cleaved γ-ENaC expression in WT mice, indicating that 2% KCl diets increased ENaC activity. Compared with WT mice, MST3−*/*<sup>−</sup> mice had a higher cleaved γ-ENaC on control diets and higher full-length ENaC on 2% KCl diets (Figure 4C). IHC results showed that feeding the 2% KCl diets slightly increased γ-ENaC expression at the apical plasma membrane of the DCT2/CNT in WT mice compared with that in mice fed the control diets (Figure 4D, c vs. a). Consistent with our previous report, MST3−*/*<sup>−</sup> mice exhibited higher ENaC expression at the apical plasma membrane of the DCT2/CNT. A higher intensity of ENaC staining was observed on 2% KCl diets (Figure 4D, g vs. e). The MST3 expression showed that higher levels of MST3 protein were present in the cytosol of the DCT2/CNT in WT mice on 2% KCl diets compared to that in WT mice on control diets (Figure 4D, d vs. b). Lower levels of MST3 were observed in MST3−*/*<sup>−</sup> mice on 2% KCl diets (Figure 4D, h).

Isoforms A, B, and F of NKCC2 are estimated to account for 20–25% of all renal Na<sup>+</sup> reabsorption. The NKCC2-F isoform mainly located in the inner medullary TAL accounts for 70% of NKCC2 expression [23]. Since MST3 is primarily localized in the inner medullary TAL, we determined whether MST3 is involved in K<sup>+</sup> loading-mediated NKCC2 phosphorylation. K<sup>+</sup> loading inhibited nonglycosylated NKCC2 phosphorylation at S130 in WT mice; however, the level of nonglycosylated and glycosylated phospho-S130-NKCC2 was increased in MST3−*/*<sup>−</sup> mice on 2% KCl diets (Figure 4E). These results indicated that MST3 inhibited NKCC2F phosphorylation at S130. IHC results showed that both MST3 (Figure 4F, G, b) and NKCC2 (Figure 4F, G, a) are mainly expressed at the apical membrane of the inner medullary TAL in control diet-fed WT mice. A 2% KCl diet induced MST3 expression ((Figure 4F, G, d) in the cytosol of the inner medullary TAL. NKCC2 was present at the subapical membrane of the inner medullary TAL (Figure 4F, G, c vs. a). This pattern is more clearly observed in an enlarged image (Figure 4G). However, in 2% KCl diet-fed MST3−*/*<sup>−</sup> mice, low levels of MST3 and higher levels of NKCC2 were still present at the apical membrane of the inner medullary TAL ((Figure 4F, 4G, g and h). These results indicated that MST3 inhibited medullary NKCC2 expression at the apical membrane of the medullary TAL in mice fed the 2% KCl diets.

The 2% KCl diets reduced the level of NCC in MST3*+/+* and MST3−*/*<sup>−</sup> mice (Figure 4H, lower panel), indicating that K<sup>+</sup> loading-inhibited NCC expression promoted K<sup>+</sup> secretion. However, there were no differences in phospho-NCC levels in MST3*+/+* and MST3−*/*<sup>−</sup> mice fed the 2% KCl diets (Figure 4H, upper panel). Additionally, there were no differences in the NCC distribution in the DCT1 in mice fed the control and 2% KCl diets (Figure 4I). These results indicated that MST3 may have a small or no effect on K<sup>+</sup> loading-mediated inhibition of NCC activity.

dullary TAL, we determined whether MST3 is involved in K+ loading-mediated NKCC2 phosphorylation. K+ loading inhibited nonglycosylated NKCC2 phosphorylation at S130 in WT mice; however, the level of nonglycosylated and glycosylated phospho-S130-NKCC2 was increased in MST3*−/−* mice on 2% KCl diets (Figure 4E). These results indicated that MST3 inhibited NKCC2F phosphorylation at S130. IHC results showed that both MST3 (Figure 4F, G, b) and NKCC2 (Figure 4F, G, a) are mainly expressed at the apical membrane of the inner medullary TAL in control diet-fed WT mice. A 2% KCl diet induced MST3 expression ((Figure 4F, G, d) in the cytosol of the inner medullary TAL. NKCC2 was present at the subapical membrane of the inner medullary TAL (Figure 4F, G, c vs. a). This pattern is more clearly observed in an enlarged image (Figure 4G). However, in 2% KCl diet-fed MST3*−/−* mice, low levels of MST3 and higher levels of NKCC2 were still present at the apical membrane of the inner medullary TAL ((Figure 4F, 4G, g and h). These results indicated that MST3 inhibited medullary NKCC2 expression at the apical membrane of the medullary TAL in mice fed the 2% KCl diets. The 2% KCl diets reduced the level of NCC in MST3*+/+* and MST3*−/−*mice (Figure 4H, lower panel), indicating that K+ loading-inhibited NCC expression promoted K+ secretion. However, there were no differences in phospho-NCC levels in MST3*+/+* and MST3*−/−* mice fed the 2% KCl diets (Figure 4H, upper panel). Additionally, there were no differences in the NCC distribution in the DCT1 in mice fed the control and 2% KCl diets (Figure 4I). These results indicated that MST3 may have a small or no effect on K+ loading-mediated

inhibition of NCC activity.

**Figure 4.** *Cont*.

(**E**)

(**I**)

**Figure 4.** WNK4 and WNK4-regulated channels in mice fed 2% KCl diets. Western blot analysis of (**A**) WNK4, (**B**) BK, and ROMK, (**C**) γ-ENaC, (**E**) NKCC2, and p-NKCC2, and (**H**) NCC and p-NCC in the kidney of MST*+/+* and MST3*<sup>−</sup>/<sup>−</sup>*mice 8 weeks after the treatment with the control and 2% KCl diets for 16 days (*n* = 3 animals/group). The bar graph shows **Figure 4.** WNK4 and WNK4-regulated channels in mice fed 2% KCl diets. Western blot analysis of (**A**) WNK4, (**B**) BK, and ROMK, (**C**) γ-ENaC, (**E**) NKCC2, and p-NKCC2, and (**H**) NCC and p-NCC in the kidney of MST*+/+* and MST3−*/*<sup>−</sup> mice 8 weeks after the treatment with the control and 2% KCl diets for 16 days (*n* = 3 animals/group). The bar graph shows quantification of Western blot relative to the levels in the control group of WT mice. \* *p* < 0.05 vs. the control group. Serial sections of the kidney of WT and MST3−*/*<sup>−</sup> mice fed the control and 2% KCl diets were stained for γ-ENaC (**D**), NKCC2 (**F**,**G**, outlined images of **F** were enlarged using a 100× objective.), NCC (**I**), and MST3. G, glomerular; D1, early distal convoluted tubule; D2, late distal convoluted tubule; CN, connecting tubule.

#### **3. Discussion** delivered to DCT2/CNT is due to K+ loading-induced NCC and NKCC2 inhibition; thus,

convoluted tubule; D2, late distal convoluted tubule; CN, connecting tubule.

**3. Discussion** 

K <sup>+</sup> preferentially leaves the cells through K<sup>+</sup> channels, such as ROMK and BK, at the apical membrane of the DCT2/CNT. This process is driven by an electrochemical gradient generated by reabsorption of Na<sup>+</sup> through ENaC to induce a K<sup>+</sup> -secreting state. Na<sup>+</sup> delivered to DCT2/CNT is due to K<sup>+</sup> loading-induced NCC and NKCC2 inhibition; thus, K<sup>+</sup> loading-inhibited NCC and NKCC2 and K<sup>+</sup> loading-induced ENaC activation needs to be strictly regulated to maintain Na<sup>+</sup> homeostasis (Figure 5). We found that the 2% KCl diets induced higher MST3 expression in WT mice than that in MST3−*/*<sup>−</sup> mice. MST3−*/*<sup>−</sup> mice with reduced MST3 expression had higher WNK4 expression, which might be involved in ENaC activity and NKCC2 phosphorylation at S130. These results indicated that MST3−*/*<sup>−</sup> mice reabsorbed more Na<sup>+</sup> at TAL, thus reducing K<sup>+</sup> secretion. In DCT2/CNT, MST3−*/*<sup>−</sup> mice had higher ENaC activity than that in WT mice, indicating that MST3−*/*<sup>−</sup> mice could not inhibit ENaC activation to prevent ENaC overactivation. Overall, MST3−*/*<sup>−</sup> mice reabsorbed more Na<sup>+</sup> and K<sup>+</sup> than did WT mice on HK diets. Our results indicate that MST3 functions to maintain Na<sup>+</sup> and K<sup>+</sup> homeostasis in mice on 2% KCl diets in vivo. K+ loading-inhibited NCC and NKCC2 and K+ loading-induced ENaC activation needs to be strictly regulated to maintain Na+ homeostasis (Figure 5). We found that the 2% KCl diets induced higher MST3 expression in WT mice than that in MST3*−/−* mice. MST3*−/<sup>−</sup>* mice with reduced MST3 expression had higher WNK4 expression, which might be involved in ENaC activity and NKCC2 phosphorylation at S130. These results indicated that MST3*−/−* mice reabsorbed more Na+ at TAL, thus reducing K+ secretion. In DCT2/CNT, MST3*−/−* mice had higher ENaC activity than that in WT mice, indicating that MST3*−/−* mice could not inhibit ENaC activation to prevent ENaC overactivation. Overall, MST3*−/−* mice reabsorbed more Na+ and K+ than did WT mice on HK diets. Our results indicate that MST3 functions to maintain Na+ and K+ homeostasis in mice on 2% KCl diets in vivo.

K+ preferentially leaves the cells through K+ channels, such as ROMK and BK, at the

apical membrane of the DCT2/CNT. This process is driven by an electrochemical gradient generated by reabsorption of Na+ through ENaC to induce a K+-secreting state. Na+

*Int. J. Mol. Sci.* **2021**, *22*, x FOR PEER REVIEW 11 of 15

quantification of Western blot relative to the levels in the control group of WT mice. \* *p* < 0.05 vs. the control group. Serial sections of the kidney of WT and MST3*<sup>−</sup>/<sup>−</sup>* mice fed the control and 2% KCl diets were stained for γ-ENaC (**D**), NKCC2 (**F**,**G**, outlined images of **F** were enlarged using a 100× objective.), NCC (**I**), and MST3. G, glomerular; D1, early distal

**Figure 5.** Hypothetical model of MST3 function on 2% KCl diets. NKCC2 and NCC in TAL and DCT1, respectively, are inhibited upon an increase in K+ intake, thus, delivering Na+ and water to the DCT2/CNT. Na+ in the DCT2/CNT is reabsorbed through ENaC, resulting in K+ secretion due to an electrochemical gradient. Reabsorption of Na+ through ENaC at the DCT2/CNT prevents Na+ loss. An increase in K+ intake stimulates MST3 expression, which inhibits NKCC2 and ENaC expression at the apical membrane of the nephron, thus preventing excessive absorption of Na+ and **Figure 5.** Hypothetical model of MST3 function on 2% KCl diets. NKCC2 and NCC in TAL and DCT1, respectively, are inhibited upon an increase in K<sup>+</sup> intake, thus, delivering Na<sup>+</sup> and water to the DCT2/CNT. Na<sup>+</sup> in the DCT2/CNT is reabsorbed through ENaC, resulting in K<sup>+</sup> secretion due to an electrochemical gradient. Reabsorption of Na<sup>+</sup> through ENaC at the DCT2/CNT prevents Na<sup>+</sup> loss. An increase in K<sup>+</sup> intake stimulates MST3 expression, which inhibits NKCC2 and ENaC expression at the apical membrane of the nephron, thus preventing excessive absorption of Na<sup>+</sup> and maintaining Na<sup>+</sup> homeostasis.

maintaining Na+ homeostasis. K+ loading induces an increase in the circulating plasma levels of aldosterone to stimulate K+ secretion, and then aldosterone has a smaller increase for longer periods [24]. In addition, HK-induced K+ secretion is dependent on Na+ delivery and flow, resulting from inhibition of Na+ reabsorption in the TAL and DCT1. The resultant Na+ delivery and flow along with increased aldosterone facilitate renal K+ excretion through K<sup>+</sup> loading induces an increase in the circulating plasma levels of aldosterone to stimulate K<sup>+</sup> secretion, and then aldosterone has a smaller increase for longer periods [24]. In addition, HK-induced K<sup>+</sup> secretion is dependent on Na<sup>+</sup> delivery and flow, resulting from inhibition of Na<sup>+</sup> reabsorption in the TAL and DCT1. The resultant Na<sup>+</sup> delivery and flow along with increased aldosterone facilitate renal K<sup>+</sup> excretion through ROMK and BK channels [25–27]. Our results indicated that both WT and MST3−*/*<sup>−</sup> mice exhibited apid diuresis and kaliuresis on the 1st day K<sup>+</sup> loading; however, MST3−*/*<sup>−</sup> mice could not continually increase K<sup>+</sup> secretion on the subsequent days of HK challenge (Figure 3). We

suggested that aldosterone might be involved in K<sup>+</sup> secretion at the beginning of the K<sup>+</sup> challenge, and then MST3 plays a role in Na<sup>+</sup> -dependent and flow-dependent K<sup>+</sup> secretion.

Most studies used 5% K<sup>+</sup> in an HK diet; feeding this diet decreased the abundance of both NCC and phospho-NCC [6,8]. However, 5% K<sup>+</sup> is unphysiological. Modest changes in dietary K<sup>+</sup> affect plasma [K<sup>+</sup> ] and NCC in a graded manner [8]. We fed mice with modest changes of K<sup>+</sup> by increasing K<sup>+</sup> from 1% (1% in chow) to 2% K<sup>+</sup> (1% in chow and 1% in drinking water). An approximately 2-fold increase of urinary K<sup>+</sup> was excreted in response to increased K<sup>+</sup> intake (from 1% to 2%). The abundance of NCC was reduced in both MST3+/+ and MST3−*/*<sup>−</sup> mice, indicating that NCC was inhibited, thus promoting K<sup>+</sup> secretion (Figure 4H, lower panel); however, phospho-NCC was not obviously inhibited (Figure 4H, upper panel), which may be due to the modest K<sup>+</sup> challenge in mice in the present study. The results of IHC analysis showed a lack of differences in NCC distribution in MST3*+/+* and MST3−*/*<sup>−</sup> mice (Figure 4I). These results suggested that the lowest MST3 expression in the DCT1 was not involved in 2% KCl loading-inhibited NCC activation in vivo.

WNK4 kinase activity is regulated by different mechanisms to transduce signals to downstream molecules. The KLHL3/CUL3 ubiquitin ligase complex degrades WNK4. In PHAII, the loss of interaction between KLHL3 and WNK4 increases levels of WNK4 [28]. Phosphorylation of WNK4 by PKC and PKA regulate the WNK40 s activity and downstream signaling [29]. Protein phosphatase 1 binds to WNK4 and modulates the inhibitory effect of WNK4 on ROMK [30], and activation of protein phosphatases (PPs) may mediate NCC dephosphorylation in response to high extracellular K<sup>+</sup> [31]. In addition, WNK4 is a Cl− sensor. WNK4 regulates WNK4 activity by binding to Cl−. A WNK Cl−-sensing mechanism explains WNK-mediated regulation of NCC/NKCC2 by diets with various levels of K<sup>+</sup> . At high intracellular chloride concentrations ([Cl−]<sup>i</sup> ), chloride ions binds to WNK4, thus inhibiting WNK4 activity. The acute K<sup>+</sup> loading-dephosphorylated NCC was not observed in the WNK4-Cl− insensitive knock-in mice, indicating that high extracellular K<sup>+</sup> by increasing [Cl−]<sup>i</sup> inhibits WNK4 and thus inactivates NCC [6]. However, the long-term K<sup>+</sup> loading still dephosphorylates NCC in the WNK4-Cl− insensitive knockin mice, indicating that another mechanism was involved in HK-inhibited WNK4 and its downstream signaling. MST3−*/*<sup>−</sup> mice exhibited higher WNK4 expression (Figure 4A), higher ENaC activity (Figure 4C,D), and higher p-NKCC2 (Figure 4E,G) on 2% KCl diets for 16 days, indicating that MST3 was involved in WNK4 and its downstream signals. We have previously reported that MST3 was phosphorylated at the tyrosine residues. Tyrosine phosphorylation of MST3 may create a docking site for molecules involved in diverse signaling pathways [32]. We demonstrated that MST3 inhibited protein tyrosine phosphatase activity to inhibit cell migration through paxillin regulation [33]. Involvement of MST3 in phosphatase activity, ubiquitination, or WNK4 phosphorylation, which regulates ENaC activity and NKCC2 phosphorylation in the case of long-term K<sup>+</sup> loading, requires additional investigation.

Dietary potassium inhibits NCC- and NKCC2-mediated Na<sup>+</sup> reabsorption and shifts Na<sup>+</sup> downstream for reabsorption by ENaC, which can drive K<sup>+</sup> secretion and prevent Na<sup>+</sup> loss. This study reports that increased K<sup>+</sup> intake stimulates MST3 expression to inhibit Na<sup>+</sup> reabsorption. This effect is mediated by inhibition of NKCC2 and ENaC; inhibition of NKCC2 inhibits Na<sup>+</sup> reabsorption and promotes K<sup>+</sup> secretion; inhibition of ENaC does not increase Na<sup>+</sup> reabsorption, thus maintaining Na<sup>+</sup> homeostasis (Figure 5).

#### **4. Methods**

#### *4.1. Animals*

This study was approved by the Committee on the Ethics of Animal Experiments and was performed according to the Guidelines for Animal Experiments of the China Medical University (#CMUIACUC-2019-284-1; approved July 2019). C57BL/6 male mice in Table 1 were 12-weeks-old and housed in metabolic cages and allowed ad libitum access to food and water for 4 weeks [19]; the animals were divided into two groups: the control diet

group (diets: 0.43% Na and 1.1% K (*w/w*)) and high-salt (HS) diet group (diets: 8% Na and 1.1% K (*w/w*); TestDiet, St. Louis, MO, USA). MST3*+/+* (WT) and MST3 hypomorphic mutant (designated MST3−*/*−) mice were obtained as reported previously [18]. To investigate the effect of an increase in Na<sup>+</sup> or K<sup>+</sup> separately, we added additional 1% NaCl and 1% KCl into drinking water to make sure that the mice took in additional Na<sup>+</sup> and K<sup>+</sup> . The mice were fed control, low Na (LNa), 1.43% Na (0.43% Na in chow with additional 1% NaCl in drinking water), and 2% KCl (1% K in chow with additional 1% KCl in drinking water) diet. Male WT and MST3−*/*<sup>−</sup> mice (8–12-weeks-old) were allowed ad libitum access to food and water. The mice were kept in metabolic cages, fed the control diet for the first 3 days and then challenged with LNa and HK diets for the following 6 days. Urine was collected during this period. Then, the animals were moved to the mouse cages and exposed to the corresponding challenge diet for the next 10 days.

#### *4.2. Immunohistochemistry*

The procedures have been described previously in detail. Briefly, serial sections of mouse kidneys were deparaffinized in xylene before rehydration in a graded series of ethanol and then the sample was incubated in a buffer (1 mM Tris in PBS, pH 8.2) at 100 ◦C for 20 min for antigens retrieval. Serial sections were incubated with antibodies against MST3 (1:500, a gift from Dr. Ming-Derg Lai, Taiwan), γ-ENaC (1:200, cat. no. 13943-1-AP; Proteintech, IL, USA), NCC (1:8000, cat. no. ab3553; Millipore, MA, USA), and NKCC2 (1:200, cat. no. AF2850; R&D Systems, MN, USA). Specificity of the anti-MST3 antibody was confirmed as previously report [18,33,34]. The secondary antibodies (1:1000, cat. no. 111-035-144; Jackson ImmunoResearch, PA, USA) were incubated at room temperature for 1 h with subsequent 3,30 -diaminobenzidine (DAB) (DAKO, Denmark, Hilden, Germany) staining.

#### *4.3. Immunoblotting*

Kidneys harvested from mice were homogenized using a PT 2100 Polytron homogenizer in ice-cold lysis solution containing 50 mM HEPES, pH 7.2, 150 mM NaCl, 1% Triton X-100, and protease inhibitors. The lysate from the kidney was first centrifuged at 100,000× *g* for 30 min at 4 ◦C. Samples from the supernatant were resolved by SDS-PAGE and transferred to NC paper. Western blot analysis was performed as reported previously [18] using primary antibodies against MST3 (1:1000), WNK4 (1:1000, cat. no. 22326-1-AP; Proteintech" IL, USA), BKα (1:500, cat. no. APC-151; Alomone labs, Jerusalem, IL), ROMK (1:500, cat. no. APC-001; Alomone labs, Jerusalem, IL), γ-ENaC (1:1000), NCC (1:500), NKCC2 (1:1000), phospho-T53-NCC (1:500, cat. no. ab254039; Abcam, Cambridge, UK), and phospho-S130-NKCC2 (1:500, a gift from Dr. Dario Alessi, UK).

#### *4.4. Measurement of Blood Pressure, Serum Na<sup>+</sup> , and Urinary Concentrations of Na<sup>+</sup> and K<sup>+</sup> and Statistical Analysis*

The steady-state SBP (systolic blood pressure) of restrained conscious mice was measured by a programmable tail-cuff sphygmomanometer (MK-2000ST, Muromachi, Tokyo, JP). SBP was initially estimated by inflating the cuff at approximately 25 mm Hg/sec. SBP was accurately determined during cuff deflation at approximately 4–5 mm Hg/sec. SBP was defined as blood pressure (BP) corresponding to the reappearance of the pulse. Blood samples were obtained via cheek pouch bleeding. Urine was collected from the urine collection tubes of the metabolic cages every 24 h for 9 days. The concentrations of Na<sup>+</sup> and K<sup>+</sup> were measured using an Advia 1800 chemistry system (Siemens). The levels of Na<sup>+</sup> and K<sup>+</sup> in urine and urine volume are shown as the mean ± SD. The statistical analysis was performed using Microsoft Excel 2013 by one-way analysis of variance (ANOVA) followed by Bonferroni post hoc tests. *p*-values \*<0.05, \*\*<0.005, and \*\*\*<0.0005 were considered significant. The mean values of the animals fed the control diet were averaged for the first 3 days of the treatment.

**Author Contributions:** C.-H.C. was responsible for experiment design, data collection and discussion. S.-N.W. contributed to experiments and discussion. B.-Y.B. was responsible for data collection and statistical analysis. H.-W.L. contributed to experiments and material preparation. T.-L.L. was responsible for writing, experimental design and materials. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by the research grant MOST 108-2635-B-039-001 from the National Science Council of Taiwan, in part by grant CMU109-MF-89 from the China Medical University, Taiwan, and grant RD108006 from Chang Bing Show-Chwan Memorial Hospital in Taiwan.

**Institutional Review Board Statement:** This study was approved by the Committee on the Ethics of Animal Experiments and was performed according to the Guidelines for Animal Experiments of the China Medical University (#CMUIACUC-2019-284-1; approved July 2019). We have described it in the "Animals".

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** Stk*24tm1a(GEMMS)Narl* (RMPC13241, MST3-1a/+) was used in this study as previously report [18].

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Review* **Sodium Intake and Heart Failure**

## **Yash Patel <sup>1</sup> and Jacob Joseph 2,3,\***


Received: 3 November 2020; Accepted: 10 December 2020; Published: 13 December 2020

**Abstract:** Sodium is an essential mineral and nutrient used in dietary practices across the world and is important to maintain proper blood volume and blood pressure. A high sodium diet is associated with increased expression of β—myosin heavy chain, decreased expression of α/β—myosin heavy chain, increased myocyte enhancer factor 2/nuclear factor of activated T cell transcriptional activity, and increased salt-inducible kinase 1 expression, which leads to alteration in myocardial mechanical performance. A high sodium diet is also associated with alterations in various proteins responsible for calcium homeostasis and myocardial contractility. Excessive sodium intake is associated with the development of a variety of comorbidities including hypertension, chronic kidney disease, stroke, and cardiovascular diseases. While the American College of Cardiology/American Heart Association/Heart Failure Society of America guidelines recommend limiting sodium intake to both prevent and manage heart failure, the evidence behind such recommendations is unclear. Our review article highlights evidence and underlying mechanisms favoring and contradicting limiting sodium intake in heart failure.

**Keywords:** sodium; salt; heart failure; ambulatory heart failure; epidemiological studies

#### **1. Salt and Sodium**

Salt is an ionic compound made up of cation and anion. Edible salt consists of 40% sodium and 60% chloride by weight. Salt was historically used as a preservative since bacteria cannot flourish in the presence of high salt concentrations. Human cells require approximately 0.5 g/day of sodium to maintain vital functions. Most food preservatives have high sodium content and are major causes of increased dietary intake of sodium. The average sodium intake in most Americans is 3.4 g/day or 1.5 teaspoons of salt, which is greater than the physiological requirement for the human body. High sodium or salt intake can lead to chronic comorbidities including hypertension, heart failure (HF), chronic kidney disease, stroke, cardiovascular diseases, and increase mortality. Hence, current guidelines recommend restricting sodium consumption to 2–3 g/day [1].

HF is a major burden of morbidity and mortality on the health care system and is classified into two major groups, heart failure with reduced ejection fraction (HFrEF) and heart failure with preserved ejection fraction (HFpEF). Treatment of HFrEF involves both pharmacologic and non-pharmacologic strategies, while mainly heart rate and blood pressure control strategies are used in HFpEF since multiple clinical trials have not shown significant benefits of pharmacologic therapy [2]. Sodium restriction has historically been taught in textbooks as a cornerstone of the management of HF patients. However, data on this management strategy are controversial. In addition, the adherence to following a low sodium diet is challenging, especially after a recent hospitalization, as shown by Riegel et al. [3]. Before we vigorously start educating HF patients to limit sodium intake in their diet, we need to

understand the evidence behind such recommendations. In this paper, we review evidence relating sodium to HF, pathophysiological mechanisms of increased sodium intake, and the relation of sodium intake to HF outcomes.

#### **2. Guideline Recommendations for Sodium Intake**

A low sodium diet is recommended in most national and international guidelines, as described in Table 1, with the intent of promoting health and preventing and managing comorbidities including HF.


**Table 1.** Guideline recommendations for sodium restriction in the general population.

#### **3. Low Sodium Intake and Prevention or Management of HF**

#### *3.1. Evidence in Favor of Low Sodium Intake in Prevention or Management of HF*

Systemic hypertension is one of the main risk factors for the development of HF. The lifetime risk of HF decreases with adequate treatment of blood pressure. Data from meta-analysis suggest a dose–response relationship between salt intake and increased blood pressure [12]. In a pooled analysis from four large prospective studies involving 133,118 patients, higher sodium intake was associated with increased risk of cardiovascular events and death compared with moderate sodium intake in hypertensive populations over a median of 4.2 years [13]. Systemic hypertension, if untreated, is a major risk factor for development of left ventricular hypertrophy. In the hypertensive patient population, diastolic dysfunction, left ventricular hypertrophy, and arterial stiffness are associated with urinary sodium excretion, and limiting sodium intake is associated with regression of left ventricular hypertrophy [14–17]. The proposed mechanism of regression of left ventricular hypertrophy with sodium restriction is improved large-arterial stiffness and microvascular endothelial dysfunction [18,19]. Sodium restriction is appropriate in patients with stage A (at risk for HF) and B (asymptomatic) HF due to its effect on lowering blood pressure, the incidence of hypertension, left ventricular hypertrophy, cardiovascular disease, and even incidence of HF [17,20–24]. However, there is insufficient evidence for such recommendation for stage C (with prior or current symptoms) and D (refractory) HF [25]. The Dietary Approaches to Stop Hypertension (DASH) diet, which emphasizes limiting sodium intake, has been shown to be associated with a lower incidence of HF in a prospective observational study of 36,019 participants in the Swedish Mammography Cohort over a course of seven years [26].

#### *3.2. Pathogenic Mechanisms for Beneficial E*ff*ect of Low Sodium Intake in Management of HF*

Figure 1 shows potential mechanisms of benefit with low sodium intake in patients with HF. A low sodium diet is shown to be associated with decreased pulmonary artery and capillary wedge pressures in patients with New York Heart Association (NYHA) Class III to IV heart failure [27]. Previous studies have shown that HF patients have systemic inflammation characterized by increased levels of tumor

necrosis factor *(TNF)-alpha*, interleukin *(IL)-1B* and *IL-6*, chemokine (monocytes chemoattractant protein-1 and IL-8), as well as enhanced expression of adhesion molecules. Moderate sodium restriction (up to 2.8 g/d) was associated with reduced values of neurohormonal (B-type natriuretic peptide (BNP), aldosterone, plasma renin activity) and cytokine levels (*TNF-alpha, IL-6*) and increased levels of anti-inflammatory cytokine (*IL-10*) over 12 months of follow up compared to low sodium restriction (up to 1.8 g/d) [28]. A recent review of the effects of low dietary sodium intake in patients with HF revealed that 2.6–3 g/d of dietary sodium restriction is effective for decreased BNP, renin, and aldosterone plasma levels [29]. Similarly, low sodium intake in the DASH diet is associated with low systolic and diastolic blood pressure, arterial stiffness, and markers of oxidative stress including urinary F2-isiprostane levels in HFpEF patients [30]. Adherence to the DASH diet was shown to be associated with improvement in arterial compliance, improved exercise capacity, and quality of life in patients with stage C HF [31]. increased levels of tumor necrosis factor *(TNF)-alpha*, interleukin *(IL)-1B* and *IL-6*, chemokine (monocytes chemoattractant protein-1 and IL-8), as well as enhanced expression of adhesion molecules. Moderate sodium restriction (up to 2.8 g/d) was associated with reduced values of neurohormonal (B-type natriuretic peptide (BNP), aldosterone, plasma renin activity) and cytokine levels (*TNF-alpha, IL-6*) and increased levels of anti-inflammatory cytokine (*IL-10*) over 12 months of follow up compared to low sodium restriction (up to 1.8 g/d) [28]. A recent review of the effects of low dietary sodium intake in patients with HF revealed that 2.6–3 g/d of dietary sodium restriction is effective for decreased BNP, renin, and aldosterone plasma levels [29]. Similarly, low sodium intake in the DASH diet is associated with low systolic and diastolic blood pressure, arterial stiffness, and markers of oxidative stress including urinary F2-isiprostane levels in HFpEF patients [30]. Adherence to the DASH diet was shown to be associated with improvement in arterial compliance, improved exercise capacity, and quality of life in patients with stage C HF [31].

*Int. J. Mol. Sci.* **2020**, *21*, x FOR PEER REVIEW 3 of 13

Previous studies have shown that HF patients have systemic inflammation characterized by

Figure 1 shows potential mechanisms of benefit with low sodium intake in patients with HF. A low sodium diet is shown to be associated with decreased pulmonary artery and capillary wedge

*3.2. Pathogenic Mechanisms for Beneficial Effect of Low Sodium Intake in Management of HF* 

**Figure 1.** Potential mechanisms linking dietary sodium restriction to better heart failure outcomes DBP—diastolic blood pressure, PCWP—pulmonary capillary wedge pressure, SBP—systolic blood pressure. Abbreviations: DBP, diastolic blood pressure; SBP, systolic blood pressure; PCWP, pulmonary capillary wedge pressure. **Figure 1.** Potential mechanisms linking dietary sodium restriction to better heart failure outcomes DBP—diastolic blood pressure, PCWP—pulmonary capillary wedge pressure, SBP—systolic blood pressure. Abbreviations: DBP, diastolic blood pressure; SBP, systolic blood pressure; PCWP, pulmonary capillary wedge pressure.

#### **4. Low Sodium Intake and Worsening of HF**

#### **4. Low Sodium Intake and worsening of HF**  *4.1. Evidence Against Low Sodium Intake in HF*

*4.1. Evidence Against Low Sodium Intake in HF*  In a randomized clinical trial, Aliti et al. studied the effect of the intervention of <2 g/d of salt intake in patients admitted with acute decompensated HFrEF with EF ≤ 45% on HF clinical congestion score compared to a control group with >2 g/d of salt intake [32]. On 30 days follow-up, there were no differences between the groups in the number of hospital readmissions and length of stay, though the patients in the intervention group had significantly more congestion than the control group (*p* = 0.02) [32]. Similarly, Velloso et al. did not see any significant difference in time needed for resolution of HF symptoms in adult patients admitted to the hospital with acute illness due to underlying chronic HF between the intervention group with <2 g/d salt intake and the control group with more than 2 g/d salt intake [33]. In a large Italian study in patients admitted with HF, patients assigned to low sodium intake (1.84 g/d) compared to moderate sodium intake (2.76 g/d), had reduced diuresis, more HF readmissions, poorer renal function, and a trend towards increased mortality [34]. Subjects in this study did not receive optimal neurohormonal blockade and received strict fluid restriction of 1 L/d and had high diuretic doses (up to 100 to 1000 mg of furosemide) without adjustment of clinical status. A recent pilot study done to see the effects of three-months of In a randomized clinical trial, Aliti et al. studied the effect of the intervention of <2 g/d of salt intake in patients admitted with acute decompensated HFrEF with EF ≤ 45% on HF clinical congestion score compared to a control group with >2 g/d of salt intake [32]. On 30 days follow-up, there were no differences between the groups in the number of hospital readmissions and length of stay, though the patients in the intervention group had significantly more congestion than the control group (*p* = 0.02) [32]. Similarly, Velloso et al. did not see any significant difference in time needed for resolution of HF symptoms in adult patients admitted to the hospital with acute illness due to underlying chronic HF between the intervention group with <2 g/d salt intake and the control group with more than 2 g/d salt intake [33]. In a large Italian study in patients admitted with HF, patients assigned to low sodium intake (1.84 g/d) compared to moderate sodium intake (2.76 g/d), had reduced diuresis, more HF readmissions, poorer renal function, and a trend towards increased mortality [34]. Subjects in this study did not receive optimal neurohormonal blockade and received strict fluid restriction of 1 L/d and had high diuretic doses (up to 100 to 1000 mg of furosemide) without adjustment of clinical status. A recent pilot study done to see the effects of three-months of 1.5 g versus 3.0 g daily sodium intake in patients with HFrEF showed that both dietary interventions reduced urinary sodium without adverse quality of life improvements [35].

In animal models, sodium restriction in early stages of HF was seen to be associated with early aldosterone activation compared to normal or excess sodium intake [36]. These findings suggest that sodium restriction in early stages of HF should be avoided to prevent neuroendocrine disease

progression. The data on sodium and fluid restriction in HFpEF patients are limited. A randomized clinical trial to see the effect of a diet with sodium and fluid restriction compared to an unrestricted diet in patients admitted with acute decompensated HFpEF showed that aggressive sodium and fluid restriction does not decrease readmission and mortality rate, and that it impairs the patient's food intake without any significant neurohormonal effect [37]. A recent systematic review by Mahtani et al. in 2018 including nine randomized control trials that enrolled a total of 479 patients from a total of 2655 retrieved references, revealed no robust high-quality evidence of the effects of sodium restriction in patients with HF [38]. There was a trend in improvement of HF functional class symptoms in outpatient studies with reduced sodium intake, but no effects were observed on all-cause mortality, hospitalization, or length of stay [38]. Similarly, a recent randomized trial of 44 patients hospitalized for acute decompensated HF showed that a normal sodium diet (7 g/d), when compared to a low sodium diet (3 g/d) is associated with similar degrees of decongestion with lower neurohormonal activation during acute HF treatment [39].

#### *4.2. Potential Mechanism for Adverse Impact of Low Sodium Intake in HF*

Figure 2 shows the potential mechanism for decompensated HF with low sodium intake. In short, HF is characterized by activation of the sympathetic system and renin–angiotensin–aldosterone system (RAAS) activation due to decreased renal perfusion leading to sodium and water reabsorption from renal tubules [40,41]. A sodium-restricted diet in HF patients has been shown to be associated with activation of antidiuretic and anti-natriuretic systems [42]. A recent Cochrane review of 185 clinical studies randomizing persons to low- vs. high-sodium diet revealed that in plasma or serum, there was a statistically significant increase in renin, aldosterone, noradrenaline, adrenaline, cholesterol, and triglyceride levels in groups with low sodium intake as compared to groups with high sodium intake [43]. These increases in hormones can lead to further development of congestive symptoms. Vascular congestion in HF activates pro-oxidant and pro-inflammatory genes in endothelial cells, which contributes to cardiorenal dysfunction [44–46]. Reduced sodium intake can lower blood pressure, which in turn can increase the heart rate and thereby negate the effects of beta-blockers. This was shown in a recent meta-analysis of 63 studies, although the effect was marginal with a heart rate increase of as little as 2.4% [47]. Reverse causation could also explain the observed association of lower sodium intake and outcomes. Higher-risk individuals with HF might consume less sodium due to their underlying illness but still have higher risks of adverse events. *Int. J. Mol. Sci.* **2020**, *21*, x FOR PEER REVIEW 5 of 13

**Figure 2.** Potential mechanisms whereby dietary sodium restriction may worsen heart failure. Abbreviations: Na, sodium; RAAS, renin–angiotensin–aldosterone system. **Figure 2.** Potential mechanisms whereby dietary sodium restriction may worsen heart failure. Abbreviations: Na, sodium; RAAS, renin–angiotensin–aldosterone system.

sodium diet is associated with an increase in cardiac expression of *β-MHC* and a decrease in the *α/β-MHC* ratio [48]. A low sodium diet was seen to be associated with increased α*/β-MHC* ratio, which in turn improves myocardial mechanical performance [48]. Similar effects of a high sodium diet were seen to be associated with myocyte enhancer factor (*MEF*) 2/nuclear factor of activated T cell (*NFAT*) transcriptional activity, and thereby increasing the expression of *MHC* genes [49]. Systemic hypertension can lead to a shift in the isoform distribution towards overexpression of the *β-MHC* gene with simultaneous downregulation of the *α-MHC* gene [50–52]. A high salt diet is associated with an increase in salt-inducible kinase 1 expression, which mediates the activation of *MEF2/NFAT*

There are five main proteins that are involved with calcium homeostasis and myocardial contractility—L-type Ca2+ channel (*LTCC*), phospholamban (*PLB*), *SERCA2a,* Na+/Ca2+ exchanger (*NCX*), and ryanodine receptors (*RYR*). A high sodium diet is also associated with reduced expression of both *PLB* and *NCX*. Ca2+ handling is important to maintain myocardial performance. The *LTCC* plays an important role in action potential during systole. Once Ca2+ enters the myocardial cell, it activates *RYR*, which in turn triggers Ca2+ release from the sarcoplasmic reticulum. This increase in Ca2+ release is responsible for the activation of myocardial contraction during systole. During diastole, the opposite mechanism happens; Ca2+ is pumped back from the cytosol to the sarcoplasmic reticulum by *SERCA2a* and sarcolemmal *NCX-1*, which mediates regulation of Ca2+ and Na+ exchange and thereby maintains excitation–contraction coupling. Altered Ca2+ handing is an important pathophysiological mechanism by which preclinical HF develops. Salt restriction has been shown to be associated with decreased *LTCC* protein levels in the left ventricle, increased *PLB* expression, and reduced *NCX* levels. Combined, these mechanisms together decreases sarcoplasmic reticulum Ca2+ overload by having an inhibitory effect on *SRCA2a* activity, and thereby is associated with a decrease

Low-sodium diet recommendations not only apply to hospitalized patients but also to ambulatory patients to prevent acute worsening of symptoms. However, the evidence behind these

**5. Potential Molecular Mechanism of Salt Diet and Heart Failure** 

and genes associated with left ventricular hypertrophy [49].

in the contractility index [53–55].

**6. Sodium Intake and Ambulatory Heart Failure** 

#### **5. Potential Molecular Mechanism of Salt Diet and Heart Failure**

The myosin heavy chain (*MHC*) protein is formed of α and β filaments. Changes in the proportion of these protein filaments are associated with cardiac mechanical performance. A high sodium diet is associated with an increase in cardiac expression of β*-MHC* and a decrease in the α/β*-MHC* ratio [48]. A low sodium diet was seen to be associated with increased α/β*-MHC* ratio, which in turn improves myocardial mechanical performance [48]. Similar effects of a high sodium diet were seen to be associated with myocyte enhancer factor (*MEF*) 2/nuclear factor of activated T cell (*NFAT*) transcriptional activity, and thereby increasing the expression of *MHC* genes [49]. Systemic hypertension can lead to a shift in the isoform distribution towards overexpression of the β*-MHC* gene with simultaneous downregulation of the α*-MHC* gene [50–52]. A high salt diet is associated with an increase in salt-inducible kinase 1 expression, which mediates the activation of *MEF2*/*NFAT* and genes associated with left ventricular hypertrophy [49].

There are five main proteins that are involved with calcium homeostasis and myocardial contractility—l-type Ca2<sup>+</sup> channel (*LTCC*), phospholamban (*PLB*), *SERCA2a,* Na+/Ca2<sup>+</sup> exchanger (*NCX*), and ryanodine receptors (*RYR*). A high sodium diet is also associated with reduced expression of both *PLB* and *NCX*. Ca2<sup>+</sup> handling is important to maintain myocardial performance. The *LTCC* plays an important role in action potential during systole. Once Ca2<sup>+</sup> enters the myocardial cell, it activates *RYR*, which in turn triggers Ca2<sup>+</sup> release from the sarcoplasmic reticulum. This increase in Ca2<sup>+</sup> release is responsible for the activation of myocardial contraction during systole. During diastole, the opposite mechanism happens; Ca2<sup>+</sup> is pumped back from the cytosol to the sarcoplasmic reticulum by *SERCA2a* and sarcolemmal *NCX-1*, which mediates regulation of Ca2<sup>+</sup> and Na<sup>+</sup> exchange and thereby maintains excitation–contraction coupling. Altered Ca2<sup>+</sup> handing is an important pathophysiological mechanism by which preclinical HF develops. Salt restriction has been shown to be associated with decreased *LTCC* protein levels in the left ventricle, increased *PLB* expression, and reduced *NCX* levels. Combined, these mechanisms together decreases sarcoplasmic reticulum Ca2<sup>+</sup> overload by having an inhibitory effect on *SRCA2a* activity, and thereby is associated with a decrease in the contractility index [53–55].

#### **6. Sodium Intake and Ambulatory Heart Failure**

Low-sodium diet recommendations not only apply to hospitalized patients but also to ambulatory patients to prevent acute worsening of symptoms. However, the evidence behind these recommendations is not conclusive. Alvelos et al. reported that in patients with chronic HFrEF with Ejection Fraction (EF) ≤40%, sodium restriction was not associated with improvement in NYHA functional class during 15-day follow-up [42]. Colin-Ramirez et al. in 2004 showed that in patients with HFrEF or HFpEF, 2.0–2.4 g/d of sodium restriction was associated with an improvement in NYHA functional class and less reported signs of HF on 6-months follow up [56]. However, Colin-Ramirez et al. in 2015 showed no significant difference in NYHA functional class between the intervention group with sodium restriction of 1.5 g/d in patients and the control group of moderate sodium intake of 2.4 g/d in patients with HFrEF and HFpEF who are on optimal medical therapy during 6-months follow up [57]. In a study by Philipson et al., sodium and fluid restriction of 2.3 g/d and 1500 mL/d respectively were associated with lower NYHA functional class and symptoms of edema in patients with a history of HF in NYHA classes II and IV over a 12-week follow-up [58]. Hummel et al. reported that 30-day readmissions were lower in the group with sodium restriction of 1.5 g/d in patients with a history of hypertension and recent admission or acute decompensated HF who are followed by discharge into the community [59]. However, they reported that the Kansas City Cardiomyopathy Questionnaire clinical summary score was not different between the two groups over 12 weeks of follow-up [59]. Amongst 123 ambulatory HFrEF patients from two outpatient HF clinics over a median follow-up of three years, higher sodium tertile was associated with a 39% increased risk for all-cause hospitalization and a 3.5-fold increase in risk for mortality [60]. A recent propensity-matched analysis from the HF Adherence and Retention Trial showed that sodium restriction to <2.5 g/d in NYHA class II/III HF patients is associated with a 72% higher risk of death or HF hospitalization compared to a higher sodium

intake of >2.5 g/d, especially in patients not receiving therapy with renin–angiotensin antagonists with a hazard ratio of 5.23 [61]. However, sodium intake was determined from a food-frequency questionnaire, which is subject to recall bias.

#### **7. Sodium Intake in Selected Patient Populations**

Recent meta-analyses of randomized control trials of treatment of hypertension reveal that the older population, non-white population, and only study groups with blood pressure in the highest 25th percentile show a clinically significant drop in blood pressure with a low sodium diet [62,63]. The Prospective Urban Rural Epidemiology study data showed that an increase in dietary sodium intake is associated with worse cardiovascular morbidity and mortality in a population with high basal sodium intake [64]. A dietary sodium restriction in such a population should be efficacious. Moreover, dietary sodium restriction was not efficacious in a population with low basal serum intake [64]. Amongst The National Health and Nutrition Examination Survey I participants over an average of 19 years of follow-up, a higher intake of dietary sodium was shown to be a strong independent risk factor for HF in overweight men and women with a body mass index of <sup>≥</sup>25 kg/m<sup>2</sup> [24]. Such effect was not seen amongst adult U.S. men and women with a body mass index <25 kg/m<sup>2</sup> . It also appears that sodium restriction is more beneficial for patients with advanced heart failure symptoms. Amongst 302 patients with HF, greater than 3 g/d dietary sodium intake was found to be associated with a hazard ratio of 2.54 (95% CI 1.10–5.84) for cardiac event-free survival in patients with NYHA III/IV HF symptoms compared to a hazard ratio of 0.44 (95% CI 0.20–0.97) in patients with NYHA I/II HF symptoms [65]. These data suggest that sodium restriction should be applied in only such a targeted population to obtain a substantial benefit. A study by Dolanski et al. examined the association of cognitive decline and low-sodium dietary adherence in 339 HF patients [66]. Interestingly, cognitive decline was not associated with low sodium intake; higher socioeconomic status and higher body mass index was associated with higher sodium intake. Similarly, Creber et al. studied the predictors of high sodium excretion in patients with previously or currently symptomatic HF amongst 280 community-dwelling adults [67]. They found that concomitant obesity and diabetes, and intact instead of deprived cognitive function, were associated with higher odds of sodium excretion. Similarly, sodium consumption was evaluated in 305 outpatients with HFrEF after receiving education to follow a <2 g sodium diet [68]. The authors found that sodium consumption exceeded recommended amounts in men and those with higher body mass indexes. These findings narrate the importance of addressing such demographic discrepancies to target in clinical trials to evaluate clinical outcomes with sodium restriction.

#### **8. Serum Sodium Values and HF**

Research has shown that low serum sodium value (hyponatremia) is seen in about 20% of hospitalized patients with acute HF [69]. Serum sodium concentration is closely regulated by water homeostasis, which in turn is regulated by thirst, arginine vasopressin, and kidney function [70]. Hyponatremia can be caused by excessive water retention from neurohormonal activation as well as by negative sodium balance from loop diuretics and with a low sodium intake diet [39]. Serum sodium values can be used to prognosticate outcomes in both HFrEF and HFpEF. Low serum sodium is a risk factor for poor long-term outcomes in acute HF, regardless of ejection fraction [71]. The Organize Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure registry (OPTIMIZE-HF) involving 48,612 patients recruited from 259 hospitals revealed that each 3 mmol/L drop in serum sodium values below 140 mmol/L in hospitalized patients is associated with a 19.5% increased risk of in-hospital mortality, 10% increased risk of mortality on follow-up, and 8% increase risk of death or rehospitalization on follow-up [69]. A meta-analysis of HF patients showed that low serum sodium values are associated with an increased risk of mortality [72]. We have previously shown in a national Veterans Affairs database study of 25,540 HFpEF patients that a J-shaped relationship is observed between serum sodium levels and a higher risk of number of days of HF hospitalizations and all-cause hospitalizations per year [73]. Such a relationship exists with

baseline measurements of serum sodium levels at the time of diagnosis of HF as well as during longitudinal follow-up. Among 50,932 HFpEF patients with a median follow-up of 2.9 years, a J-shaped relationship was seen between serum sodium values and all-cause mortality, HF hospitalizations, and all-cause hospitalizations [74]. These data are further supported by the fact that the improvement of hyponatremia in HF patients is associated with long-term clinical outcomes [75].

#### **9. Future Directions**

There are multiple clinical trials that aim to examine if sodium restriction in HF patients is associated with improved clinical outcomes. The Study of Dietary Intervention under 100 MMOL in Heart Failure (SODIUM-HF) is an open-label, multicenter, international, randomized controlled trial in ambulatory patients with chronic HF and aims to assess the effects of dietary sodium restriction on clinical outcomes [76]. The Geriatric out of Hospital Randomized Meal Trial in Heart Failure (GOURMET-HF) is a multicenter, randomized, single-blind, controlled trial of 3-months duration to see the effect of sodium restriction/DASH diet in older patients after discharge from acute decompensated HF admission [59].

#### **10. Our Recommendations**

There are likely many potential reasons for conflicting evidence regarding the benefit/harm of sodium restriction. These include heterogeneity of HF patient population studied, lack of uniformity in limiting the amount of sodium restriction per day, unclear data on associated use of fluid restriction, and simultaneous usage of diuretics and neurohormonal blockade agents. Given there is clear evidence of the benefit of limiting sodium intake to prevent various comorbidities leading to HF, we recommend limiting sodium intake in those who are at risk to develop comorbidities to prevent the onset of heart failure. In patients with HF, we recommend to continue limiting sodium intake to prevent morbidity associated with HF. We also recommend avoiding too much limitation in sodium intake as this has been associated with worse outcomes in HF patients.

#### **11. Conclusions**

The data supporting the restriction of dietary sodium intake in heart failure patients are unclear. While there appears to be a trend in reducing HF symptoms amongst patients using dietary sodium restriction, there appears to be no effect or slightly higher risk in mortality compared to no sodium restriction. A randomized control trial is hence needed to address this important clinical question.

**Author Contributions:** Y.P., data extraction and writing original draft; J.J., editing the draft; Y.P. and J.J., responsible for final edits of the draft. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


## **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **Sodium Intake and Chronic Kidney Disease**

**Silvio Borrelli 1,\* ,**† **, Michele Provenzano 2,**† **, Ida Gagliardi <sup>2</sup> , Michael Ashour <sup>2</sup> , Maria Elena Liberti <sup>1</sup> , Luca De Nicola <sup>1</sup> , Giuseppe Conte <sup>1</sup> , Carlo Garofalo <sup>1</sup> and Michele Andreucci <sup>2</sup>**


Received: 30 May 2020; Accepted: 1 July 2020; Published: 3 July 2020

**Abstract:** In Chronic Kidney Disease (CKD) patients, elevated blood pressure (BP) is a frequent finding and is traditionally considered a direct consequence of their sodium sensitivity. Indeed, sodium and fluid retention, causing hypervolemia, leads to the development of hypertension in CKD. On the other hand, in non-dialysis CKD patients, salt restriction reduces BP levels and enhances anti-proteinuric effect of renin–angiotensin–aldosterone system inhibitors in non-dialysis CKD patients. However, studies on the long-term effect of low salt diet (LSD) on cardio-renal prognosis showed controversial findings. The negative results might be the consequence of measurement bias (spot urine and/or single measurement), reverse epidemiology, as well as poor adherence to diet. In end-stage kidney disease (ESKD), dialysis remains the only effective means to remove dietary sodium intake. The mismatch between intake and removal of sodium leads to fluid overload, hypertension and left ventricular hypertrophy, therefore worsening the prognosis of ESKD patients. This imposes the implementation of a LSD in these patients, irrespective of the lack of trials proving the efficacy of this measure in these patients. LSD is, therefore, a rational and basic tool to correct fluid overload and hypertension in all CKD stages. The implementation of LSD should be personalized, similarly to diuretic treatment, keeping into account the volume status and true burden of hypertension evaluated by ambulatory BP monitoring.

**Keywords:** salt intake; sodium; hypertension; cardiovascular risk; mortality; prognosis

#### **1. Introduction**

Sodium is not only considered an important mineral in maintaining the balance of body fluid, but it has also played an important role in the history of the world for its economical, religious, and symbolic importance. The concept that salt is a beneficial substance was so ingrained that although the earlier studies on the relationship between low sodium intake and reduction of blood pressure (BP) date back to 1948 [1], it was only after nearly forty years that the international community has recognized the role of salt intake in the pathophysiology of hypertension [2]. According to the World Health Organization, the restriction of sodium intake to less than 2.3 g/day of sodium corresponding to 5.8 g of salt (or 100 mmol) is one of the most cost-effective measures to improve public health [3]. Cumulating evidence highlights that higher sodium consumption contributes to higher BP [4], thus increasing the risk of cardiovascular disease (CVD) [5,6]. However, recent studies have raised some concerns about

the real benefit of a low salt diet in the healthy general population [7–10]. In particular, in a large cohort study in over 100,000 patients from 18 countries the role of higher salt consumption was associated with increased BP levels [7], and poor CV outcomes [8]. At the same time, it emerged that sodium intake of <3 g forecasted a higher CV risk, drawing a U-shaped mortality curve [8]. Although these findings have also been obtained by other investigators [8–10], these studies are methodologically flawed by reverse causality (e.g., patients eat less salt because they are sicker and/or more malnourished), collinearity (e.g., a low sodium intake may be associated with a low protein-energy intake) and biased methods used to assess individual salt intake (e.g., single spot urine sample) [11].

In Chronic Kidney Disease (CKD) patients, high BP is a frequent finding, which is traditionally considered as a direct consequence of sodium sensitivity. Hence, a low salt diet (LSD) is widely considered a cornerstone in the treatment of hypertension in CKD.

In this review, we address the importance of the kidney in sodium regulation, the relationship of sodium intake with hypertension from earlier CKD stages to end-stage kidney disease (ESKD), and the available evidence on the benefit of salt restriction in non-dialysis CKD and in the ESKD population.

#### **2. Adherence to Low-Salt Diet: Definition and Assessment in CKD Patients**

The terms sodium and salt (generally sodium chloride) are used interchangeably, generating confusion about sodium intake. Table 1 illustrates formulas to convert sodium in salt (sodium chloride) and *vice versa*, according to the units of measurement.

**Table 1.** Formulas to convert sodium in salt (sodium chloride) and vice versa, according to units of measurement.


The first concern in the evaluation of adherence to LSD is the method used for evaluation of sodium intake. Accordingly, the measurement of sodium excretion by 24 h urine sample collection (UNaV) is considered as the gold standard. However, UNaV may be cumbersome for the patient, and, therefore, estimation from spot urine samples using the Nerbass, Kawasaki, Tanaka, and INTERSALT formulas have been proposed to evaluate sodium intake. The rationale arises from the assumption that spot urine excretion would be proportionate to UNaV corrected for creatinine excretion. However, a cross-sectional study in CKD patients showed that these formulas might provide an inaccurate estimate of sodium intake, irrespective of severity of CKD and use of diuretics [12].

Another issue is the evaluation of salt intake by a single measurement, which is not generally considered sufficient to evaluate an individual's usual salt intake because of the wide day-to-day variability in salt consumption and urinary excretion [13].

Among CKD cohorts, the Chronic Renal Insufficiency Cohort (CRIC) study reported that only about one out of four patients had a sodium intake <100 mmol/24 h, evaluated by three measurements [14]. As reported in Table 2, these findings are consistent with the prevalence of LSD reported in secondary analyses of trials [15–19], and also when CKD patients were regularly followed in nephrology clinics (<25% had a salt intake below 6 g/day) [20].


Abbreviations: n.a.: not applicable. ARBs: Angiotensin Receptor Blockers.

**Table 2.** Studies evaluating the effect of urinary sodium excretion (UNaV) on end-stage kidney disease (ESKD) and cardiovascular (CV) outcomes in patients with or

*Int. J. Mol. Sci.* **2020**, *21*, 4744

Novel methods for self-monitoring of salt intake have been developed to improve the adherence to the LSD; these are based on urine chloride strips, given that urinary chloride excretion is very tightly correlated with urinary sodium excretion. The potential benefit of self-monitoring is the ability to immediately achieve an adequate estimate of sodium intake immediately (75.5% sensitivity and 82.6% specificity to correctly classify patients with UNaV >100 mmol/24 h) in order to make proper dietary adjustments aimed at achieving recommended intake [23]. However, any benefit on the achievement of the BP goal with use of chloride strips has still to be proved with use of chloride strips. To answer this question, one randomized clinical trial, the SALUTE-CKD (SALt lowering by Urine sodium self-measurement Trial in Chronic Kidney DiseasE) has advanced to the final stage of development and results are expected in the next few months.

Furthermore, a recent trial performed in 99 patients has evaluated the efficacy of a web-based self-management program for dietary sodium restriction compared with routine care. After 3 months in intervention group a significant reduction of sodium intake (−40 mmol/day) and systolic BP (−8 mmHg) was registered in the intervention group, whereas no significant difference was found in control group. Surprisingly, in the following maintenance phase, no difference in sodium intake was detected between the two groups, due to the inadvertent adoption of the intervention by the control group. Notably, the largest effect was reported in the first 3 months, when participants actively used the web-based self-management program [24].

In ESKD patients, sodium intake can be estimated with a dietary questionnaire, though several factors, such as high dialysate sodium concentration and sodium plasma concentration, can affect thirsty and water intake in these patients, irrespective of their sodium intake [25].

Finally, in ESKD patients, residual kidney function must be carefully evaluated: in this subgroup of patients, dialysis is started with an incremental approach, corresponding to a low dose of dialysis (peritoneal or hemodialysis) integrated into the conservative management [26,27]. In these patients, the assessment of sodium intake by UNaV may be misleading, because of the aliquot of sodium intake removed by dialysis.

#### **3. Hypertension and Salt in CKD**

Hypertension and CKD are common chronic noncommunicable diseases strictly inter-related with each other; indeed, elevated BP is not only a frequent complication of CKD [28], but it can also act as the cause of CKD [29]. A recent meta-analysis showed that hypertensive patients have a 75% greater risk than normotensive individuals of development of *de novo* CKD (GFR <60 mL/min/1.73 m<sup>2</sup> ), estimating a 10% increase of CKD onset for each increase of 10 mmHg of either BP component. Notably, even pre-hypertension (Systolic BP of 120–139 mm Hg and/or Diastolic BP of 80–89 mm Hg) was associated with a 25% higher risk of developing low GFR [29].

Furthermore, the prognostic role of lowering BP assumes greater importance in CKD patients if we bear in mind at least three basic points: (1) higher prevalence of hypertension in CKD than in the general population, which increases progressively from 65% to 95% as GFR falls from 85 to 15 mL/min/1.73 m<sup>2</sup> [29]; (2) hypertension is the main known risk factor for CKD progression and for CV mortality [30]; (3) Hypertension is often resistant to the treatment in CKD patients, resulting in worsening CV prognosis [31,32].

Salt and water retention play a key role for development of hypertension in CKD. In fact, according to the classical model, under normal conditions, high salt intake temporarily increases plasma sodium level, which is soon buffered by movement of water from the intracellular to the extracellular compartment. Thus, increased plasma sodium concentration also stimulates the thirst center, leading to an increase in water intake and secretion of antidiuretic hormone, which restores plasma sodium concentration to a normal level while increasing and maintaining extracellular fluid volume. On the other hand, high salt intake suppresses the renin-angiotensin-aldosterone system (RAAS), which consequently reduces sodium tubular reabsorption, thus contributing to re-establishing sodium and water homeostasis [33].

In CKD patients, external sodium balance is preserved by expansion of the extracellular volume (ECV), which however causes the persistence of high BP levels. Therefore, hypertension in CKD is an early manifestation of ECV expansion and, at the same time, a maladaptive mechanism aimed at limiting ECV expansion that corresponds to approximately 5% to 10% of body weight, generally without peripheral edema, when cardiac and hepatic function is normal and the transcapillary Starling forces are not disrupted [34]. In spite of ECV expansion, RAAS is inappropriately activated in CKD, leading to vasoconstriction and sodium retention, which contribute significantly to the raising of BP levels [35].

As reported in a classic experiment [36], the BP response to sodium load is amplified in CKD patients. In particular, increasing sodium intake was increased from 20 to 120 mmol/day in patients with advanced renal failure, this caused a significant acute increase of BP (+12.2 ± 1.4 mmHg). On the other hand, the same increase in sodium intake in healthy people was not associated with any BP change and, even greater elevation of sodium intake up to 1120 mmol/day, did not produce any effect on BP values. This experiment is the proof of concept of the sodium sensitivity of BP in CKD. Notably, sodium sensitivity may be already detectable in the earlier CKD stages, as reported in a study comparing patients with glomerular disease vs healthy controls, which showed a significant BP reduction in response to lowering salt intake, whereas BP did not change in controls [37].

Moreover, experimental studies showed that high salt intake induces intrarenal production of Angiotensin-II [38], stimulates the synthesis of pro-inflammatory cytokines [39] and increases oxidative stress [40], as well as triggering sympathetic activity [41], whose activation is already increased in CKD, as a result of increased arterial stiffness and/or endothelial dysfunction [42].

#### **4. Alternative Mechanism of Sodium Toxicity**

Recent experimental findings suggest that skin could work as a reservoir of sodium, escaping from renal control [43]. In particular, high salt intake might cause sodium accumulation in the skin, which is detected by cells of the Monocytes Phagocytes System (MPS) located in the skin interstitium, which act as osmoreceptors by expression of the tonicity enhancer-binding protein (Ton-EBP). This transcription factor leads to Vascular Endothelial Growth Factor (VEGF) production that increases sodium clearance by the lymphatic network [44,45]. Moreover, high sodium levels in the CKD condition would promote the expression of pro-inflammatory factors, such as Interleukin-6, VEGF, and Monocyte Chemoattractant Protein-1 (MCP-1), via Ton-EBP pathway, leading to local inflammation and vascular proliferation in peritoneal, heart, and vascular tissue [46].

Figure 1 summarizes the potential mechanisms underlying the increase of BP levels and dependent CV risk associated with high salt intake in CKD.

The concept that sodium balance is regulated by additional extra-renal mechanisms was first reported by Herr et al., whose study showed that high salt intake increased total sodium content, whereas total body water and body weight did not change [47]. More recently, a space flight simulation study has reported that in healthy subjects under controlled sodium intake, UNaV changes periodically (every 6 days), independently from BP levels and total body water [48].

The recent availability of <sup>23</sup>Na Magnetic Resonance Imaging (MRI) in humans has allowed detection and quantification of sodium storage in the skin [49]. In particular, a higher tissue sodium content was detected in patients affected by hyperaldosteronism. Interestingly, surgical and/or medical correction of hyperaldosteronism was associated with a significant reduction in tissue sodium content; whereas body weight did not change [50]. Recently, in a cross-sectional analysis of 99 CKD patients, skin sodium content was strongly associated with left ventricular mass independently from BP levels and volume status [51]. Finally, sodium stored into the skin is modifiable in CKD patients, as reported by a recent study showing a significant reduction of skin sodium content, after a single hemodialysis session, though the mechanism by which sodium is removed from skin remains still unclear [52].

**4. Alternative Mechanism of Sodium Toxicity** 

raising of BP levels [33].

early manifestation of ECV expansion and, at the same time, a maladaptive mechanism aimed at limiting ECV expansion that corresponds to approximately 5% to 10% of body weight, generally without peripheral edema, when cardiac and hepatic function is normal and the transcapillary Starling forces are not disrupted [32]. In spite of ECV expansion, RAAS is inappropriately activated in CKD, leading to vasoconstriction and sodium retention, which contribute significantly to the

As reported in a classic experiment [34], the BP response to sodium load is amplified in CKD patients. In particular, increasing sodium intake was increased from 20 to 120 mmol/day in patients with advanced renal failure, this caused a significant acute increase of BP (+12.2 ± 1.4 mmHg). On the other hand, the same increase in sodium intake in healthy people was not associated with any BP change and, even greater elevation of sodium intake up to 1120 mmol/day, did not produce any effect on BP values. This experiment is the proof of concept of the sodium sensitivity of BP in CKD. Notably, sodium sensitivity may be already detectable in the earlier CKD stages, as reported in a study comparing patients with glomerular disease vs healthy controls, which showed a significant BP

Moreover, experimental studies showed that high salt intake induces intrarenal production of Angiotensin-II [36], stimulates the synthesis of pro-inflammatory cytokines [37] and increases oxidative stress [38], as well as triggering sympathetic activity [39], whose activation is already increased in CKD, as a result of increased arterial stiffness and/or endothelial dysfunction [40].

Recent experimental findings suggest that skin could work as a reservoir of sodium, escaping from renal control [41]. In particular, high salt intake might cause sodium accumulation in the skin, which is detected by cells of the Monocytes Phagocytes System (MPS) located in the skin interstitium, which act as osmoreceptors by expression of the tonicity enhancer-binding protein (Ton-EBP). This transcription factor leads to Vascular Endothelial Growth Factor (VEGF) production that increases sodium clearance by the lymphatic network [42,43]. Moreover, high sodium levels in the CKD condition would promote the expression of pro-inflammatory factors, such as Interleukin-6, VEGF, and Monocyte Chemoattractant Protein-1 (MCP-1), via Ton-EBP pathway, leading to local

reduction in response to lowering salt intake, whereas BP did not change in controls [35].

inflammation and vascular proliferation in peritoneal, heart, and vascular tissue [44].

**Figure 1.** Potential pathogenic mechanisms of hypertension in CKD due to high salt intake. Abbreviations: CKD: Chronic Kidney Disease; AT-II: Angiotensin-II; CNS; Central Nervous System; CV: cardiovascular.

#### **5. Clinical E**ff**ects of Low Salt Diet in Non-Dialysis CKD**

We have recently completed a metanalysis comparing low versus high salt diet in 738 CKD patients [53]. Analysis included nine trials [54–62]. This meta-analysis showed that a moderate salt restriction of 4.4 g/day (from 179 mEq/day to 104 mEq/day) was associated with a significant lowering of 4.9 mmHg [95% C.I.: 6.8/3.1 mmHg; *p* < 0.001] in systolic BP and of 2.3 mmHg [95% C.I.: 6.8/3.1 mmHg; *p* < 0.001] in diastolic BP measured by traditional method [53]. A similar effect was found in the five out of eleven studies [57,60,61,63,64] evaluating the effect of LSD on Ambulatory BP (ABP). In particular, we found that salt restriction reduces systolic and diastolic ABP of 5.9 mmHg (95% C.I.: 2.3/9.5 mmHg; *p* < 0.001) and 3.0 mmHg (95% C.I.: 1.7/4.7 mmHg; *p* < 0.001), respectively [53].

As regards ABP studies, it is worth mentioning that in CKD cohorts, sodium sensitivity has been associated with a higher prevalence of altered circadian rhythm and nocturnal hypertension [65,66], which are predictors of poor cardio-renal prognosis [67].

Moreover, in seven out of eleven studies [54–59] reporting the effect of salt restriction on proteinuria, pooled analysis showed a significant improvement of 0.4 g/day (95% C.I.: 0.2–0.6 g/day) associated with lower salt intake [53]. These findings are in agreement with a previous meta-analysis reporting that in patients following a lower salt diet, there was an augmented antiproteinuric effect of RAAS blockers [68]. The synergic effect of LSD and RAAS inhibition may be correlated to the finding that high salt intake enhances angiotensin-converting enzyme (ACE) activity in renal tissues, in spite of decreased plasma renin and angiotensinogen concentrations, which could reduce the effect of RAAS blockers in tissues [38].

Although these effects of LSD on BP and proteinuria suggest an improvement of prognosis in CKD patients, few studies [14–19] have evaluated the long-term effect of salt restriction on the cardio-renal outcomes (Table 2).

In the CRIC study, a large observational study carried out in 3757 CKD patients followed for almost seven years, the group of patients with a UnaV of >195 mmol/day was associated with a higher risk of CKD progression [14]. Among participants of this study, 804 composite CV events (575 heart failure, 305 myocardial infarctions, and 148 strokes) occurred during a median of 6.8 years of follow-up, drawing a linear relation between higher sodium intake and higher CV risk [22]. Similarly, a post-hoc

analysis of the Reduction of Endpoints in NIDDM with the Angiotensin II Antagonist Losartan (RENAAL) and Irbesartan Diabetic Nephropathy Trial (IDNT) trials in a subgroup of 1177 patients with available 24 h urinary sodium measurements, showed that the beneficial effects of RAAS blockers on renal and cardiovascular outcomes were greater in patients with lower sodium intake [18]. Furthermore, in Autosomal Dominant Polycystic Kidney Disease (ADPKD) patients, fast progressors irrespective of intensive CKD management [69,70], a recent post-hoc analysis of the HALT-PKD trial has shown that a moderate salt restriction reduces CKD progression [21].

On the other hand, other studies have not confirmed these results, finding no association between low salt intake and improvement of the renal prognosis, in CKD patients [15–17]. In particular, secondary analysis of the first and second Ramipril Efficacy in Nephropathy (REIN) trials showed that low salt intake was associated with a lower risk of ESKD, but this association disappeared after adjustment for basal proteinuria [15]. In the longitudinal follow up of the Modification of Diet in Renal Disease (MDRD) Study, no association of single baseline 24 h urinary sodium excretion with kidney failure and a composite outcome of kidney failure or all-cause mortality was found [16]. Similarly, post-hoc analysis of the ongoing telmisartan alone and in combination with ramipril global endpoint trial (ONTARGET) and telmisartan randomized assessment study in ACE intolerant subjects with cardiovascular disease (TRANSCEND) studies trials showed no association between UNaV (though estimated by morning spot urine) and renal endpoints (30% decline of eGFR or ESKD) in patients with or without CKD at baseline [17]. Surprisingly, in diabetic non-CKD patients, UNaV was inversely associated with a cumulative incidence of ESKD, and in fact, patients with the lowest sodium excretion had the highest cumulative incidence of ESKD [19].

Of note, the negative studies are post hoc analyses of clinical trials designed to test the efficacy of RAAS inhibitors rather than of low-sodium intervention, confounding thus a possible association[15–18]. Furthermore, in some of these studies, UNaV was measured by a single 24 h urine [19] or spot urine sample [17]. On the other hand, we cannot exclude that other factors might play a role: a renal hemodynamic response to an acute reduction of sodium intake was impaired by aging, especially when atherosclerotic damage coexists [71]. This may expose patients to acute kidney injury and hypotension [72]. Furthermore, patients with CKD have a higher prevalence of white coat effect [73,74], exposing CKD patients to "inappropriate" antihypertensive treatment, which may potentially cause renal hypoperfusion [75]. Therefore, particular attention must be paid in the management of CKD patients, personalizing salt intake on the basis of "true" hypertensive status measured by ABPM and volemic status, and monitoring the adherence and anti-hypertensive effect LSD over time.

#### **6. Sodium Intake in End-Stage Kidney Disease**

In ESKD patients, similarly to early CKD stages (Figure 1), the deleterious effects of high salt intake are mainly related to the fluid overload, resulting in high BP levels, left ventricular hypertrophy, and increased CV mortality [76–80]. Therefore, sodium restriction is a major therapeutic goal in these patients. Indeed, it has been estimated that, in ESKD patients with no residual diuresis, a salt intake of <6 g should cause patients to gain no more than 0.8 kg/day in interdialytic weight.

A recent metanalysis of four trials (3 in HD/1 in PD) showed that ESKD patients with lower salt intake (N = 67) had a significant improvement of both systolic [−8.4 (−12.0; −4,8) mmHg] and diastolic BP [−4.4 (−6.6; −4.2) mmHg] levels compared with the higher salt intake group (N = 64) [81]. Moreover, a post-hoc analysis of the HEMO study revealed that low sodium intake, evaluated by a 24 h food questionnaire, allowed to decrease the need for ultrafiltration, even if it was not associated with pre-dialysis systolic BP levels [82].

Similarly, hypervolemia is prevalent in Peritoneal Dialysis (PD) patients, because of the common mismatch between intake and removal of sodium and fluid [83–85]. In a recent study performed in a cohort of 1054 incident PD patients, overhydration was evident in over 50% of patients starting PD [83]. This finding is relevant because persistence of volume overload heralds a 60% higher mortality risk [84]. Interestingly, recent experimental findings have reported that high sodium intake is related to direct toxicity on the peritoneal membrane, leading to chronic inflammation, fibrosis, and hypervascularization, increasing, in turn, peritoneal permeability [86].

Surprisingly, few studies have addressed the relationship between sodium intake and mortality in ESKD patients (Table 3). In hemodialysis, secondary analysis of HEMO study showed that higher dietary sodium correlated with mortality rate independently from patients' nutritional status [82]. A retrospective study on 305 Chinese PD patients has reported that sodium intake, assessed by a 3-day diet questionnaire, was inversely associated with all-cause and cardiovascular mortality. It is noteworthy that patients with lower sodium intake also had lower serum albumin levels and reduced lean body mass, as well as lower energy and protein intake, when compared with patients with higher sodium intake, suggesting that patients with a lower dietary sodium intake were more malnourished and with reduced appetite. Moreover, the group of patients with highest salt intake had mean sodium intake of 2.5 g/day, which is lower than the mean intake reported in the general population, suggesting a possible measurement bias. This study, moreover, was flawed by methodological issues (small sample size, monocentric, few events, and overfitted cox models), further reducing the generalizability of the results [87].


**Table 3.** Studies evaluating the effect of urinary sodium excretion (UNaV) on mortality in ESKD patients.

Therefore, salt restriction remains the basic approach to achieve volume control in PD patients, keeping in mind that sodium removal is lower in PD patients treated with cycler (Automated PD, APD), because of greater sodium sieving as compared with Continuous Ambulatory PD (CAPD). In these patients, high salt intake may not be counterbalanced by sodium removal, consequently leading to hypervolemia and hypertension [88].

#### **7. Conclusions**

The negative effects of sodium on BP values are amplified in CKD patients, as a result of fluid overload and of direct toxicity on the heart, the vascular system, and kidney. In non-dialysis CKD patients, LSD is beneficial for hypertension control, irrespective of BP levels, to lower proteinuria by enhancing the antiproteinuric effect of RAAS inhibition. Whether these effects can improve cardio-renal prognosis still remains unclear. Nonetheless, salt restriction assumes a greater importance in ESKD because of the common mismatch between intake and removal of sodium, which leads to hypertension, LVH, and higher CV risk. Therefore, reducing salt intake is crucial for hypertensive CKD patients from earlier stages to ESKD. However, it remains insufficiently and/or inadequately applied. More studies are therefore needed to improve adherence to LSD in the long term.

**Author Contributions:** S.B.: Data curation, Writing - original draft; M.P.: Data curation, Funding acquisition, Writing - original draft; I.G.: Funding acquisition Writing - review & editing; M.A. (Michael Ashour) and M.A. (Michele Andreucci): Funding acquisition Writing - review & editing; M.E.L., L.D.N., C.G. and G.C.: Supervision Writing - review & editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
