Colon-Specific Deletion of Epithelial Sodium Channel Causes Sodium Loss and Aldosterone Resistance

Abstract

Aldosterone promotes electrogenic sodium reabsorption through the amiloride-sensitive epithelial sodium channel (ENaC). Here, we investigated the importance of ENaC and its positive regulator channel-activating protease 1 (CAP1/Prss8) in colon. Mice lacking the αENaC subunit in colonic superficial cells (Scnn1a^KO) were viable, without fetal or perinatal lethality. Control mice fed a regular or low-salt diet had a significantly higher amiloride-sensitive rectal potential difference ([INCREMENT]PD_amil) than control mice fed a high-salt diet. In Scnn1a^KO mice, however, this salt restriction-induced increase in [INCREMENT]PD_amil did not occur, and the circadian rhythm of [INCREMENT]PD_amil was blunted. Plasma and urinary sodium and potassium did not change with regular or high-salt diets or potassium loading in control or Scnn1a^KO mice. However, Scnn1a^KO mice fed a low-salt diet lost significant amounts of sodium in their feces and exhibited high plasma aldosterone and increased urinary sodium retention. Mice lacking the CAP1/Prss8 in colonic superficial cells (Prss8^KO) were viable, without fetal or perinatal lethality. Compared with controls, Prss8^KO mice fed regular or low-salt diets exhibited significantly reduced [INCREMENT]PD_amil in the afternoon, but the circadian rhythm was maintained. Prss8^KO mice fed a low-salt diet also exhibited sodium loss through feces and higher plasma aldosterone levels. Thus, we identified CAP1/Prss8 as an in vivo regulator of ENaC in colon. We conclude that, under salt restriction, activation of the renin-angiotensin-aldosterone system in the kidney compensated for the absence of ENaC in colonic surface epithelium, leading to colon-specific pseudohypoaldosteronism type 1 with mineralocorticoid resistance without evidence of impaired potassium balance.

Sodium and potassium transport across tight epithelia (kidney and colon) is important to keep the body in a constant balance, despite large dietary variations. Aldosterone promotes sodium reabsorption as an electrogenic sodium transport through the amiloride-sensitive epithelial sodium channel (ENaC).¹ Systemic autosomal recessive pseudohypoaldosteronism type 1 (systemic PHA-1) is caused by ENaC mutations and characterized by a severe salt-losing syndrome paralleled with hypotension, hyperkalemia, metabolic acidosis, and high plasma aldosterone levels.² Liddle’s syndrome is caused by mutations within the PPxY motif (PY) domain of the β- or γENaC subunits and results in severe salt-sensitive hypertension with renal salt retention, alkalosis, and low plasma aldosterone levels.³

ENaC was originally identified in rat colon from animals challenged with low salt diet and is made of three homologous subunits: α, β, and γ.^4,5 In the mouse, the constitutive knockout of each subunit is postnatally lethal.^6–8 In the absence of αENaC, the β- and γ-subunits are not transported to the membrane, and no amiloride-sensitive sodium current is measured in vitro or ex vivo.^6,9

Along the intestine, sodium absorption occurs through electroneutral sodium transport through the sodium/hydrogen exchanger rather than electrogenic absorption through ENaC that is limited to surface epithelial cells of the distal colon and rectum.^10,11 After proctocolectomy, ENaC starts to be expressed in the distal part of the small intestine (i.e., the ileum), thereby unveiling the importance of an electrogenic amiloride-sensitive transport for the reabsorption of salt and water in the intestine.¹² Thereby, aldosterone stimulates β- and γENaC mRNA transcript expression in rat distal colon.^13–15 If dietary sodium intake is low and plasma aldosterone levels are high, the distal colon can efficiently absorb dietary sodium against a large concentration gradient.^11,16 Enhanced ENaC expression in colon, thus, contributes to sodium retention observed in mice with Liddle’s syndrome^17,18 along with increased responsiveness to aldosterone.¹⁹ On the other side, downregulation of ENaC with reduction in sodium reabsorption in colon may contribute to diarrhea associated with inflammatory bowel disease.^20,21

The membrane-bound serine protease CAP1/Prss8, also known as prostasin, activates ENaC by rapidly increasing the open probability.^22–25 CAP1/Prss8 is coexpressed with ENaC in many salt-absorbing tight epithelia, such as distal colon, urinary bladder, and airways.^23,24In vivo evidence that CAP1/Prss8 is an important and physiologically relevant activator of ENaC came from the study of mice lacking CAP1/Prss8 in the alveolar epithelium, unveiling a crucial role for lung fluid balance.²⁶ In the colon, however, the physiologic role of this membrane-bound serine protease was hitherto unknown, and it was unclear whether CAP1/Prss8 was implicated in regulating colonic ENaC activity.

In the present study, we, thus, addressed whether suppression of colonic ENaC activity affected sodium and/or potassium balance and what are the compensatory mechanisms that lead to increased renal sodium reabsorption. Finally, we unveiled the role of the positive ENaC activator CAP1/Prss8 in colon. We specifically deleted either αENaC/Scnn1a or CAP1/Prss8 in the colonic surface epithelium and determined in vivo the electrogenic sodium transport to correlate plasma electrolytes with fecal sodium loss and plasma aldosterone concentrations.

Results

Intestine-Specific αENaC-Deficient Mice Are Viable and Exhibit Normal Colon Histology

To ablate αENaC expression in colonic superficial cells, we mated Scnn1a^+/−; villin::Cre^tg/0 mice with mice harboring two floxed αENaC alleles (Scnn1a^loxlox) (Figure 1A). Analysis of a total of 252 offspring at weaning showed no deviation from the expected Mendelian distribution (Scnn1a^Lox, n=60; Scnn1a^Het, n=68; Scnn1a^Hetc, n=70; Scnn1a^KO, n=54). Adult Scnn1a^KO mice were viable, showed no postnatal mortality, and were indistinguishable in appearance, growth, and body weight (Table 1). In the Scnn1a^KO mice, colonic superficial cells lack near 99% of Scnn1a mRNA transcript expression, whereas heterozygotes (Scnn1a^Het) exhibit intermediate (71%) expression levels compared with Scnn1a^Lox (P<0.05) (Figure 1B). The expression of β- and γENaC mRNA transcripts was not significantly higher in Scnn1a^KO mice (Figure 1B). The successful deletion of Scnn1a in scraped colonic superficial cells was further confirmed on the protein expression level (Figure 1, C and D). Heterozygotes for the Scnn1a allele (Scnn1a^Het and Scnn1a^Hetc) showed intermediate expression (70% and 50% of Scnn1a^Lox, respectively).

Figure 1: Loss of αENaC mRNA transcript and protein expression in colonic superficial cell-specific Scnn1a-deficient mice. DNA, mRNA, and protein samples were analyzed from (A) tissues or (B–D) isolated scraped distal colonic superficial cells. (A) PCR analysis on ear biopsies with primers distinguishes between the lox (580 bp) and the KO (360 bp) allele of the Scnn1a gene locus (row 1). In row 2, the primers indicated distinguish between wild type (220 bp) and lox (280 bp) allele in experimental Scnn1a ^lox/−; villin::Cre^tg/+ (Scnn1a ^KO) and controls (lanes 1 and 2), Scnn1a^lox/− (Scnn1a ^Het); lanes 3 and 4, Scnn1a^lox/+; villin::Cre^tg/+ (Scnn1a ^Hetc); lanes 5 and 6, Scnn1a ^lox/+ (Scnn1a ^Lox); lanes 7 and 8, littermates. Detection of the villin::Cre transgene (upper band) and myogenin (internal control, lower band) (row 3). (B) Quantification of α-, β-, and γENaC mRNA transcripts by quantitative RT-PCR in cells from Scnn1a ^Lox (n=5; white), Scnn1a ^Het (n=4; light gray), and Scnn1a ^KO mice (n=4; black column). Results are expressed as the ratio of mRNA/β-actin mRNA (*P<0.05). (C) Representative immunoblot showing the expression of αENaC (row 1) and β-actin (row 2) protein in cells from Scnn1a ^Lox, Scnn1a ^Het, Scnn1a ^Hetc, and Scnn1a ^KO mice. (D) Quantification of αENaC protein expression levels in cells from Scnn1a ^Lox (white), Scnn1a ^Het (light gray), Scnn1a ^Hetc (dark gray), and Scnn1a ^KO (black) mice after analysis with ImageJ software (n=3 mice per group; *P<0.05). Results are expressed as the ratio of αENaC protein/β-actin protein. Values are mean±SEM.

Table 1: Physiologic parameters of Scnn1a ^KO mice

Macroscopically, the morphology of the adult distal colon was not different (Supplemental Figure 1). The colon epithelium and mucin-secreting goblet cells appeared normal in knockout mice, without any effect on the number of crypt cells (not shown). The intestine length-to-body weight ratio was not different between the Scnn1a^Lox (1.97±0.05; n=5), Scnn1a^Het (1.89±0.05; n=6), and Scnn1a^KO (1.83±0.06; n=6) groups.

Implication of ENaC in Intestinal Electrogenic Sodium Transport and Sodium Balance

ENaC-mediated sodium transport is electrogenic and generates an amiloride-sensitive transepithelial potential difference ([INCREMENT]PD_amil) that varies on different salt diets and follows a circadian rhythm.²⁷ We measured ΔPD_amil, and the switch from high salt (HS) (Figure 2A) to regular salt (RS) (Figure 2B) and low salt (LS) (Figure 2C) diets induced a progressive increase in plasma aldosterone (Figure 3). On HS diet, plasma aldosterone (0.1–0.2 nmol/L) (Figure 3) and baseline ΔPD_amil (Figure 2A) were equally low (−5 to −6 mV) between groups (Figure 2A). On RS diet, plasma aldosterone increased from 0.7 nmol/L in Scnn1a^Lox to 1.2 nmol/L in Scnn1a^Het and 1 nmol/L in Scnn1a^Hetc mice (Figure 3). All mice showed a significant (−10 to −15 mV) increase in ΔPD_amil compared with the HS diet. The circadian rhythm expressed as (a.m./p.m.) cyclicity was readily observed (Figure 2, A and B). The highest plasma aldosterone level was observed in the Scnn1a^KO group (2.4 nmol/L), contrasting with the ΔPD_amil that remained low (−6 to −8 mV) and without cyclicity (Figures 2B and 3). On LS diet, plasma aldosterone increased in all groups to reach high values in the Scnn1a^KO group (8.5 nmol/L) (Figure 3). Despite this drastic increase in plasma aldosterone level, ΔPD_amil remained low (−5 to −6 mV) with blunted cyclicity (Figures 2C and 3). In all conditions, a residual amiloride-insensitive negative PD was observed (between −6 and −8 mV; data not shown). The observed hyperaldosteronism suggested that loss of sodium in the feces could have caused a significant hypovolemia and triggered the activation of the renin-angiotensin-aldosterone system (RAAS). We, therefore, analyzed total sodium and potassium in the feces and found that, with RS and LS diets, Scnn1a^KO mice lost significantly more sodium (RS, P<0.05; LS, P<0.001). This difference was not observed with HS diet (Figure 4A). Fecal potassium was not significantly different among the groups (Figure 4B). Moreover, wet/dry ratio of feces was similar in all groups (Scnn1a^KO: 0.32±0.02; Scnn1a^Lox: 0.30±0.02; Scnn1a^Het: 0.34±0.02).

Figure 2: Colonic sodium transport is impaired in Scnn1a ^KO mice. Morning and afternoon measurements of amiloride-sensitive rectal PD (ΔPDamil) on 2 consecutive days in Scnn1a ^Lox mice (n=7; line), Scnn1a ^Het (n=7; dashed line), Scnn1a ^Hetc (n=7; dotted line), and Scnn1a ^KO (n=8; dashed/dotted line) mice treated with (A) HS, (B) RS, or (C) LS diet. ***P<0.001. Values are mean±SEM.

Figure 3: Scnn1a ^KO mice show elevated plasma aldosterone levels. Plasma aldosterone (nanomoles per liter) concentrations in Scnn1a ^Lox (n=6; white), Scnn1a ^Het (n=7; light gray), Scnn1a ^Hetc (n=6; dark gray), and Scnn1a ^KO (n=7; black) mice were analyzed on various sodium diets. *P<0.05; **P<0.01. Values are mean±SEM.

Figure 4: Increased sodium loss through feces in Scnn1a ^KO mice. Measurements of (A) sodium and (B) potassium electrolytes levels in feces from Scnn1a ^Lox (n=6; white), Scnn1a ^Het (n=7; light gray), and Scnn1a ^KO (n=7; black) mice on various sodium diets. Values are mean±SEM. *P<0.05; ***P<0.001, Scnn1a ^KO versus Scnn1a ^Lox and Scnn1a ^Het mice.

Lack of Colonic ENaC-Mediated Sodium Absorption Compensated by the Kidney

Scnn1a^KO mice should be able to compensate for the fecal sodium loss by an aldosterone-dependent sodium absorption by the distal nephron. Hence, mice were followed in metabolic cages, and on HS, RS, and LS diets, food and water intake, feces output, urinary volume, and plasma and urinary sodium and potassium were measured (Table 1). On LS diet, cumulative sodium excretion in the Scnn1a^KO group was significantly diminished compared with all groups (P<0.05) (Figure 5, A–C). Cumulative potassium loss was not different, even when challenged with high potassium (5%) (Figure 5, D–F, Table 1).

Figure 5: Diet-dependent reduced sodium loss in urine of Scnn1a ^KO mice. Measurement of cumulative urinary (A–C) sodium and (D–F) potassium electrolyte levels in Scnn1a ^Lox (n=8; white), Scnn1a ^Het (n=7; light gray), Scnn1a ^Hetc (n=8; dark gray), and Scnn1a ^KO (n=8; black) mice on (A and D) HS, (B and E) RS, and (C and F) LS diets; *P<0.05. Values are mean±SEM.

CAP1/Prss8 Identified as an In Vivo Regulator of ENaC in Distal Colon

To test the role of CAP1/Prss8 on ENaC in distal colon in vivo, intestine-specific CAP1/Prss8-deficient mice (Prss8^KO, Prss8^{[INCREMENT]/lox}; villin::Cre^tg/0) were generated (Figure 6A). At weaning, analysis of a total of 219 offspring showed no deviation from the Mendelian distribution (Prss8^Lox, n=55; Prss8^Het, n=55; Prss8^Hetc, n=56; Prss8^KO, n=53). In Prss8^KO mice, colonic superficial cells lacked CAP1/Prss8 mRNA transcript expression (<1%), whereas heterozygotes (Prss8^Het) exhibited intermediate expression levels compared with Prss8^Lox cells (70%) (Figure 6B). The mRNA transcript expression of CAP2/Tmprss4 and CAP3/Prss14 was not altered (Figure 6B). The successful deletion of CAP1/Prss8 in scraped colonic superficial cells was further confirmed on the protein level (Figure 6, C and D).

Figure 6: Loss of Prss8 mRNA transcript and protein expression in colonic superficial cell-specific Prss8 ^KO mice. DNA, mRNA, and proteins samples were analyzed from isolated scraped intestinal superficial cells as indicated. (A) DNA-based PCR analysis on ear biopsies using primers distinguishing wild type (+; 379 bp), lox (413 bp), and [INCREMENT] (473 bp) alleles in Prss8^lox/+ (Prss8 ^Lox; lanes 1–4), Prss8^{lox/[INCREMENT]} (Prss8 ^Het; lanes 5–8), and Prss8^{[INCREMENT]/lox}; villin::Cre^tg/+ (Prss8 ^KO; lanes 9–12) littermates. The villin::Cre transgene (400 bp) and myogenin (internal control) are detected using specific primers. (B) Quantification of CAP1/Prss8, CAP2/Tmprss4, and CAP3/SP14 mRNA transcripts by quantitative RT-PCR in cells from Prss8 ^Lox (n=4; white), Prss8 ^Het (n=5; gray), and Prss8 ^KO (n=8; black) mice. Data are expressed as the ratio of mRNA/β-actin mRNA. *P<0.05. (C) Representative immunoblot showing the expression of CAP1/Prss8 and β-actin protein in cells from Prss8 ^Lox, Prss8 ^Het, and Prss8 ^KO. (D) Quantification of CAP1/Prss8 signals in Prss8 ^Lox (n=4; white), Prss8 ^Het (n=5; gray), and Prss8 ^KO (n=6; black) cells analyzed with ImageJ software. ***P<0.001. Results are expressed as the ratio of CAP1/Prss8 protein/β-actin protein. Values are mean±SEM.

Prss8^KO mice did not differ in body weight, food and water intake, urine or feces output, and plasma and urinary sodium and potassium levels (Table 2). Colon histology was normal (Supplemental Figure 2A) without any apparent effect on the number of crypt cells (data not shown). The intestine length-to-body weight ratio was not different between the control (Prss8^Lox: 2.04±0.14; n=6), heterozygotes (Prss8^Het: 2.16±0.13; n=6), and knockout (Prss8^KO: 1.91±0.1; n=7). When we monitored the intestinal permeability after fluorescein isothiocyanate dextran supply in blood plasma, we found no difference amongst the groups, indicating a normal intestinal barrier function in the knockouts (P=0.09 to Prss8^Het and P=0.39 to Prss8^Lox) (Supplemental Figure 2B). When mRNA expression levels of ENaC subunits were quantified in distal colon and the kidney, there was no difference among the groups, with the exception of βENaC mRNA transcripts (KO versus Lox and Het; P<0.05) (Supplemental Figure 3, A and B). Western blot analysis using the anti-αENaC antibody revealed the full-length 93 kDa form and its cleaved 30 kDa form (Supplemental Figure 3C). The 95 kDa full-length β- and γENaC, including the cleaved 75 kDa γENaC proteins, are equally present in all groups (Supplemental Figure 3, C–F). We finally measured ΔPD_amil after HS, RS, and LS diets that induced a progressive increase in plasma aldosterone levels in all groups (Figure 7, A–D). On HS diet, baseline ΔPD_amil and plasma aldosterone levels (0.1–0.2 nmol/L) were equally low (−8 to −10 mV), and cyclicity was maintained, although blunted (Figure 7, A and D). On RS diet, ΔPD_amil of Prss8^Lox and Prss8^Het mice increased markedly (−15 to −25 mV) with respect to HS diet, and (a.m./p.m.) cyclicity was readily observed (Figure 7, A and B). Despite increased (0.5 nmol/L) plasma aldosterone levels, the cyclicity of the Prss8^KO group was blunted, mainly because of a significant decrease of ΔPD_amil in the afternoon. On LS diet, ΔPD_amil in Prss8^KO remained significantly lower with blunted cyclicity, although plasma aldosterone levels reached comparable high and even significant values (P<0.05 to Prss8^Lox and Prss8^Het) (Figure 7, C and D). Interestingly, however, the feces wet/dry ratio was not altered in the knockout (Prss8^KO: 0.33±0.02 versus controls; Prss8^Lox: 0.31±0.02 and Prss8^Het: 0.37±0.02), and sodium, but not potassium, was significantly lost in feces from the knockouts (P<0.05) (Figure 7, E and F).

Table 2: Physiologic parameters of Prss8 ^KO mice

Figure 7: CAP1/Prss8 is as a regulator of ENaC in the colon. (A–C) Morning and afternoon measurements of amiloride-sensitive rectal PD (ΔPD_amil) on 2 consecutive days in control Prss8 ^Lox (n=8; line), Prss8 ^Het (n=7; long dashed line), and Prss8 ^KO (n=8; dashed line) mice treated with (A) HS, (B) RS, and (C) LS diet. *P<0.05. (D) Plasma aldosterone concentrations after various sodium diets. (E) Fecal sodium and (F) fecal potassium concentrations in Prss8 ^Lox (n=5; white), Prss8 ^Het (n=6; gray), and Prss8 ^KO (n=6; black) mice on HS, RS, and LS diets. *P<0.05. Values are mean±SEM.

In summary, our data clearly show that in vivo stimulation of the amiloride-sensitive ENaC-mediated sodium transport is dependent on the expression of the membrane-bound serine protease CAP1/Prss8 and more striking in the afternoon, when the RAAS is maximally activated.

Discussion

ENaC-Mediated Electrogenic Sodium Transport Is Limiting for the Final Absorption of Sodium in Distal Colon and Rectum: Evidence for Colon-Specific Haploinsufficiency

In the present study, we studied mice with an efficient deletion of αENaC along the colon and found a strict gene dosage effect at the mRNA transcript and protein expression levels (Figure 1). Electrogenic sodium transport in distal colon was mainly mediated by ENaC, even if a low but significant electrogenic transport was measured after amiloride application (Figure 2). We cannot exclude some residual ENaC activity caused by incomplete recombination, although on a HS diet, mRNA expression of ENaC subunits should be rather repressed. Sodium/hydrogen exchanger 3 that is sensitive to amiloride is electroneutral and thus, undetectable by our PD measurements (Figure 2). On RS diet, despite increased plasma aldosterone levels, the ENaC KO mice remained at a low [INCREMENT]PD_amil. Under LS diet, a significant dissociation between the heterozygotes and the floxed (Scnn1a^Lox) group was observed, indicating haploinsufficiency possibly caused by upregulation of AT1 receptors, although those mice showed an intact capacity to maintain BP and sodium balance.²⁸

Differential Activation of RAAS on Lowering Salt Intake: Evidence for Colon-Specific Mineralocorticoid Resistance

In our study, we varied salt intake from HS to RS (19-fold) and from RS to LS, with an additional 17-fold decrease in salt intake (Supplemental Figure 4). The Scnn1a^KO mice showed 14-fold (versus 7-fold in control groups, P<0.05; HS to RS) increased aldosterone response that declined on switch from RS to LS to a 4-fold (P<0.01; versus 2-fold) induction (Supplemental Figure 4). Absence of ENaC in the colon and consequently, failure of the colon to absorb sodium against an electrochemical gradient might lead to a colon-specific salt-losing syndrome accompanied by high aldosterone response, which was shown by the clear correlation between plasma aldosterone (P_aldo) levels and [INCREMENT]PD_amil response; the KO mice remained unresponsive, whereas the Scnn1a^Lox mice stayed sensitive to increased P_aldo. The response of the heterozygous mice was intermediate (P<0.05) (Supplemental Figures 4 and 5). We interpreted these data as indicating a colon-specific mineralocorticoid resistance (or decreased aldosterone responsiveness) that led to a colon-specific PHA-1 phenotype. Interestingly, a mirrored image of this phenotype was observed in the colon of Liddle mice that harbor a point mutation within the βENaC subunit, leading constitutively to hyperactivity of ENaC and an increased aldosterone responsiveness of the sodium transport in colon.^19,29

Differential Effect of Colon-Specific αENaC Knockouts on Sodium and Potassium Balance

As summarized in Figure 8, on HS diet, Scnn1a^KO mice exhibit a sodium balance, and the total recovery of urinary and fecal sodium accounts for approximately 85% of sodium intake. From HS to LS diet, we found a progressive fecal sodium loss in Scnn1a^Het, Scnn1a^Hetc and Scnn1a^KO mice. Under LS, the fecal loss of sodium is compensated for by a maximal retention of sodium in the kidney because of high P_aldo (Figures 4 and 8). The missing sodium might be caused by loss into the transcellular fluid compartment, which may account for about 6% along the entire intestine and/or into the skin compartment.³⁰ Although systemic PHA-1 is normally also characterized by hyperkalemia, we did not find a shift in the potassium balance in the Scnn1a^KO mice (Figures 4 and 5), which may be explained by differentially regulated and spatially separated electrogenic sodium absorption and potassium secretion.³¹

Figure 8: Diet-dependent shift of sodium balance in Scnn1a ^KO mice. Sodium balance is considered as the ratio between the quantity of sodium output (in urine or feces) at day 1 normalized by the quantity of sodium intake at day 1. Data were taken from the experiments summarized in Table 1 (food intake), Figure 4 (fecal sodium), and Figure 5 (urinary sodium). For each of the genotypes (Scnn1a ^Lox, Scnn1a ^Het, and Scnn1a ^KO), the average sodium intake through food (gray column) is compared with urinary sodium output (white column) and fecal sodium output (black column) on HS, RS, or LS diet.

CAP1 Regulates Colon ENaC Activity by Blunting Its Circadian Cyclicity

Previous studies have emphasized the importance of CAP1/Prss8 in vivo^32–35 and its implication in ENaC regulation in alveolar fluid clearance and lung fluid balance.²⁶ In colon, we clearly identify CAP1/Prss8 as a protease activating ENaC in vivo, because on RS and LS diets, ENaC-mediated transport becomes limiting in Prss8^KO mice (Figure 7). These data are in the same line as recent findings in hairless (fr^CR) rats and frizzy (fr/fr) mice harboring spontaneous mutations of CAP1/Prss8.³² We do not see an implication of CAP1/Prss8 in epithelial barrier formation and permeability in colon (Supplemental Figure 2), which is contrary to mice that specifically lack CAP1/Prss8 in the epidermis and exhibit a severely impaired epidermal barrier caused by defective function of tight junctions.³⁴ Interestingly, lack of the serine protease in colon superficial cells is not consistent with a failure to cleave ENaC, because the cleaved 75 kDa ENaC fragment is present in Prss8^KO mice (Supplemental Figure 3). These data are consistent with previous findings, where the 80 kDa and the cleaved 70 kDa γENaC protein forms were detected when CAP1/Prss8 was absent in lung.²⁶ This lack of difference in γ-cleavage is maybe not too surprising in view of the relative small difference in ΔPD_amil between the KO and the controls.

In conclusion, we showed that, in the colon of mice lacking ENaC and/or CAP1/Prss8, amiloride-sensitive sodium transport is drastically diminished. This result leads to increased fecal sodium loss, which is accompanied by mineralocorticoid resistance in ENaC-deficient mice. In patients with PHA-1 mutations, it might become pathophysiologically relevant and aggravate sodium loss, particularly on low dietary salt intake. Because the amount of sodium in the body is the main determinant of extracellular volume, disturbances in sodium balance will lead to clinical situations of volume depletion or overload; the latter will lead to arterial hypertension and heart failure. In CKD, when the ability of the kidneys to excrete sodium decreases, pharmacological inhibition of colonic ENaC may lead to increased intestinal excretion of sodium, which may help to maintain sodium homeostasis in CKD, where diuretics have only limited success.

Concise Methods

Intestine-Specific CAP1/Prss8 and αENaC-Deficient Mice

Intestine-specific αENaC (Scnn1a) or CAP1/Prss8 knockout mice were generated by interbreeding Villin::Cre transgenic mice, which were heterozygous mutant for the αENaC⁶ or CAP1/Prss8³⁴ knockout allele, with mice homozygous for the respective conditional alleles Scnn1a^loxlox36 or CAP1/Prss8^lox/lox.³⁷ To generate an intestine-specific αENaC KO, we mated Scnn1a^+/−; villin::Cre^tg/0 mice with mice harboring two floxed αENaC alleles (Scnn1a^loxlox). Age-matched wild type-like Scnn1a^lox/+(Scnn1a^Lox), heterozygous mutant Scnn1a^lox/− (Scnn1a^Het), intestine-specific heterozygous mutant Scnn1a^lox/+; villin::Cre^tg/0 (Scnn1a^Hetc), and intestine-specific αENaC knockout Scnn1a^lox/−, villin::Cre^tg/0 (Scnn1a^KO) mice were obtained. To generate intestine-specific CAP1/Prss8 KO, we mated Prss8^{[INCREMENT]/+}; villin::Cre^tg/0 mice with mice harboring two floxed CAP1/Prss8 (Prss8^loxlox). Age-matched wild type-like CAP1/Prss8^lox/+ (Prss8^Lox), heterozygous mutant CAP1/Prss8^{lox/[INCREMENT]} (Prss8^Het), intestine-specific heterozygous mutant CAP1/Prss8^lox/+; villin::Cre^tg/0 (Prss8^Hetc), and intestine-specific CAP1/Prss8 knockout CAP1/Prss8^{lox/[INCREMENT]}; villin::Cre^tg/0 (Prss8^KO) mice were obtained.

All animal work was conducted according to Swiss federal guidelines. All mice were kept in the animal facility under animal care regulations of the University of Lausanne. They were housed in individual ventilated cages at 23±1°C with a 12-hour light/dark cycle. All animals were supplied with food and water ad libitum. This study has been reviewed and approved by the “Service de la consommation et des affaires vétérinaires” of the Canton of Vaud, Switzerland. If not otherwise indicated, 6- to 12-week-old age-matched male and female αENaC and CAP1/Prss8 control and experimental (knockout) mice (homozygous for Ren-1^c) were fed for at least 3 weeks on an RS (0.17% Na⁺), HS (3.2% Na⁺), or LS (0.01% Na⁺) diet. All diets were obtained from ssniff Spezialdiäten GmbH (Soest, Germany).

Genotyping

Genotyping by PCR was performed using the following primers: CAP1/Prss8^{+/lox/[INCREMENT]}:Prss8–1 sense (5′-GCAGTTGTAAGCTGTCATGTG-3′); Prss8–2 sense (5′-CAGCAGCTGAGGTACCACT-3′); Prss8-3 antisense (5′-CCAGGAAGCATAGGTAGAAG-3′); αENaC ^+/−:αENaC^+/−-1 antisense (5′-TTAAGGGTGCACACAGTGACGGC-3′); αENaC^+/−-2 antisense (5′-TTTGTCACGTCCTGCACGACGCG-3′); αENaC^+/−-3 sense (5′-AACTCCAGAAGGTCAGCTGGCTC-3′); αENaC^{+/lox/[INCREMENT]}:αENaC^lox/+-1 sense (5′-CTCAATCAGAAGGACCCTGG-3′); αENaC^lox/+-2 sense (5′-GTCACTGTGTGCACCCTTAA-3′); αENaC^lox/+-3 antisense (5′-GCAAAAGATCTTATCCACC-3′).

If not otherwise stated, 35 cycles were run, and each run consisted of 1 minute each at 94°C, 56°C (58°C for ENaC), and 72°C. The Villin::Cre transgene was detected by PCR using the following primers: Villin-Cre sense (5′-CCTGGAAAATGCTTCTGTCCG-3′) and Villin-Cre antisense (5′-CAGGGTGTTATAAGCAATCCC-3′). Myogenin-specific primers (sense, 5′-TTACGTCCTCGTGGACAGC-3′) and (antisense, 5′-TGGGCTGGGTGTTAGTCTTA-3′) were used to control the DNA integrity of each sample.

Quantitative RT-PCR Analysis on Distal Colon and Kidney Samples

Total RNA was prepared from freshly isolated mouse colon superficial cells and whole kidney using the RNeasy Extraction Kit (Qiagen, Hilden, Germany). The RNA (1 µg/sample) was reverse-transcribed at 37°C for 1 hour using superscript II RNAse H reverse transcriptase (Invitrogen, Basel, Switzerland) and oligo-dT(20) primers (Invitrogen). The products were then diluted 10 times before proceeding with the real-time PCR reaction. Real-time PCRs were performed by Taqman PCR with the Applied Biosystems 7500 (Foster City, CA). The primer and probe mix (2×) (Mm00504792 m1 for mCAP1 and 4352341E for β-actin) was purchased with the Universal Taqman Mix (2×) and used according to the manufacturer’s instructions (Applied Bio Systems, Foster City, CA). Quantification of fluorescence was performed with the [INCREMENT][INCREMENT]C_T normalized to β-actin. Each measurement was performed in duplicate. Additional primers have been used: Scnn1a: FOR, 5′-GCACCCTTAATCCTTACAGATACACTG-3′ and REV, 5′-CAAAAAGCGTCTGTTCCGTG-3′, Probe 5′-FAM-AGAGGATCTGGAAGAGCTGGACCGCA-BHQ1–3′; Scnn1b: FOR, 5′-GGGTGCTGGTGGACAAGC-3′, REV, 5′-ATGTGGTCTTGGAAACAGGAATG-3′, Probe, 5′-FAM-CAGTCCCTGCACCATGAACGGCT-BHQ1–3′; Scnn1g: FOR, 5′-AACCTTACAGCCAGTGCACAGA-3′, REV, 5′-TTGGAAGCATGAGTAAAGGCAG-3′, Probe, 5′-FAM-AGCGATGTGCCCGTCACAAA>CATCT-BHQ1–3′; Prss8: FOR, 5′-CCCATCTGCCTCCCTGC-3′, REV, 5′-CCATCCCGTGACAGTACAGTGA-3′, Probe, 5′-FAM CCAATGCCTCCTTTCCCAACGGC-BHQ1–3′; Prss14: FOR, 5′-GAAGCTTTGATGTCGCTCCC-3′, REV, 5′-GGAGGGTGAGAAGGTGCCA-3′, Probe, 5′-FAM- CCACGCTGTGGTGCGGCTG-BHQ-1–3′; Tmprss4: FOR, 5′-AGTAGGCATCGTGAGCTGGG-3′, REV, 5′-GGACGGCAGCGTTACATCTC-3′, Probe, 5′-FAM-ATGGATGCGGCGGCCCAA-BHQ1–3′.

Western Blot Analysis

Animals (3–4 months) were kept under an RS or LS diet for 2 weeks. Colon and kidney were freshly isolated and snap frozen in liquid nitrogen. Proteins were extracted by homogenization using polytron and sonication with an IKA sonicator in 8 M urea buffer; then, they were incubated for 30 minutes on ice and centrifuged for 30 minutes at 4°C at 14,000 rpm. The supernatant was taken and centrifuged again for 10 minutes at 4°C at 14,000 rpm. The supernatant was used to detect the protein concentration with a BCA protein kit (PIERCE, Rockford, IL). Samples of protein extracts were separated by SDS-PAGE on 10% acrylamide gels, electrically transferred to polyscreen polyvinylidene difluoride transfer membrane (Perkin Elmer, Boston, MA), and subsequently probed for CAP1/Prss8, Scnn1a (αENaC), Scnn1b (βENaC), Scnn1g (γENaC), and β-actin using primary rabbit antibodies Scnn1a (1:500),³⁸ Scnn1b and Scnn1g (1:1000),³⁹ CAP1 (1:1000),⁴⁰β-actin (1:1000; Sigma-Aldrich), and anti-rabbit IgG secondary antibody (1:10,000; Amersham, Burkinghampshire, UK). The signal was developed with the ECL+ system (Hyperfilm ECL; Amersham). Quantification of protein level was obtained using National Institutes of Health image software.

Histologic Analysis of Proximal and Distal Colon

Colon was fixed in 4% paraformaldehyde overnight and subjected to paraffin embedding and sectioning (4-µm sections). Sections were stained with hematoxylin and eosin and examined by light microscopy using an Axioplan microscope (Carl Zeiss Microimaging, Inc., Oberkochen/Jena, Germany), and images were acquired with a high-sensibility digital color camera (Carl Zeiss Microimaging, Inc.).

Determination of Intestine Structural and Functional Parameters

Determination of Length-to-Body Weight.

Length of intestine (centimeters) was measured and normalized to the body weight in 3-month-old mice. Results were determined as mean±SEM.

Feces Wet-to-Dry Weight and Electrolyte Measurements.

Feces samples were collected from age-matched 3-month-old control (n=6), heterozygote mutant (n=6), and knockout (n=7) mice that were kept under RS diet in metabolic cages for 4 consecutive days. Wet-to-dry weight was determined by determining the wet weight feces samples collected within 24 hours, drying the feces at 80°C for another 24 hours, and weighing again to calculate the wet-to-dry feces ratio as described.³² Sodium and potassium fecal electrolytes were determined from samples as described.⁴¹ Briefly, the feces were collected over 2 consecutive days, weighed, and resuspended overnight into 0.75 N nitric acid at 4°C. After centrifugation, an aliquot of supernatant was measured for Na⁺ and K⁺ content with a flame photometer (943 Electrolyte Analyzer; Instrumentation Laboratory, UK).

Intestinal Permeability Assay.

In vivo intestinal permeability was determined as described previously.⁴² Briefly, mice were kept under RS diet and gavaged with 10 ml/kg solution of 22 mg/ml fluorescein isothiocyanate–dextran (4 kDa; Sigma-Aldrich, St. Louis, MO) in PBS (pH 7.4). Three hours after gavage, plasma was collected at the end of the experiment and centrifuged at 3000 rpm for 20 minutes at 4°C. After a 1:1 dilution in PBS, the concentration of fluorescein was determined using a 96-plate reader with an excitation wavelength at 485 nm and an emission wavelength at 535 nm using serially diluted samples of the tracer as a standard.

Metabolic Cage Studies.

Six- to twelve-week-old age-matched control and knockout mice were individually placed in metabolic cages (Tecniplast, Buguggiate, Italy) for 5 consecutive days to measure urine and feces output. Food and water intake were daily measured. For the entire experiment, mice had free access to food and water. During experimental days, urine and feces were collected. Sodium intake was measured as sodium (millimoles) intake per day in percentage of total food intake. Sodium output was measured as urinary sodium (millimoles) and fecal sodium (millimoles) excretion per day in percentage of total food intake.

High Potassium Diet

Experimental mice and control mice were placed in individual metabolic cages and fed a standard diet for 2 consecutive days (0.95% potassium), which was followed by 2 days on 5% potassium in drinking water (the potassium was added as KCl). During the experiment, the animals had free access to food and water. During experimental days, urine was collected. Blood was collected 2 days after the experiment.

Analysis of Urinary and Plasma Electrolytes

Urine samples (24 hours) were collected in metabolic cages. Blood samples were collected at the end of the experiment. Urine and plasma electrolytes were analyzed using an Instrumentation Laboratory 943 Electrolyte Analyzer, UK.

Blood Collection for Aldosterone Measurements

Control and knockout mice (8–12 weeks old) were kept in standard cages with free access to food and water and fed with RS, LS, or HS diets for 12 consecutive days. At the end of the experiment, blood samples were collected. Plasma aldosterone levels were measured according to standard procedures using a radioimmunoassay (Coat-A-Count RIA Kit; Siemens Medical Solutions Diagnostics, Ballerup, Denmark).⁴³ Samples with values >1200 pg/ml were further diluted using a serum pool with a low aldosterone concentration (<50 pg/ml). Aldosterone concentration is indicated as nanomoles per liter.

Amiloride-Sensitive Rectal Transepithelial PD Measurements

Mice were fed a LS or HS diet for 3 weeks. Amiloride-sensitive transepithelial rectal PD measurements were performed as previously described.^27,32 Briefly, rectal PD and amiloride-sensitive rectal PD were measured in the morning (10 a.m. to 12 p.m.) and the afternoon (4 p.m. to 6 p.m.) on 2 days of the same week. The rectal PD was monitored continuously by a VCC600 electrometer (Physiologic Instruments, San Diego, CA) connected to a chart recorder. After stabilization of rectal PD (approximately 1 minute), 0.05 ml saline solution was injected through the first barrel as a control maneuver, and the PD was recorded for another 30 seconds. A similar volume of saline solution containing 25 µmol/L amiloride was injected through the second barrel of the pipette, and the PD was recorded for 1 minute. The PD was recorded before and after the addition of amiloride to determine the amiloride-sensitive PD.

Statistical Analyses

Results are presented as mean±SEM. Throughout the study (if not otherwise stated), data were analyzed by one-way ANOVA. Unpaired t test was used for the comparison between two groups (Figure 7D). P<0.05 was considered statistically significant.

Disclosures

None.

We thank Anne-Marie Mérillat for excellent photographic work and Friedrich Beermann for critically reading the manuscript. We are grateful to Jean-Christophe Stehle and Samuel Rotman from the mouse histology platform of the University of Lausanne.

S.M. was a recipient of the Faculty of Biology and Medicine Fellowship Program of the University of Lausanne. This work was supported by grants from the Swiss National Science Foundation and the National Center of Competence in Research: Kidney.CH: Control of Homeostasis and the Leducq Foundation (to E.H.).

Part of this work has been published in abstract form (Malsure et al., J Am Soc Nephrol (Suppl): 14A, 2009 and Malsure et al., J Am Soc Nephrol (Suppl): 496A, 2012).