Research ArticleSKIN DISEASE

Sodium channel Nax is a regulator in epithelial sodium homeostasis

See allHide authors and affiliations

Science Translational Medicine  04 Nov 2015:
Vol. 7, Issue 312, pp. 312ra177
DOI: 10.1126/scitranslmed.aad0286

Rubbing sodium in a wound

The skin serves as a critical barrier to the outside world; however, little is known about how this barrier returns to homeostasis after it is disturbed. Water loss occurs during many skin disorders, resulting in an increase in extracellular sodium concentration. Now, Xu et al. report that the sodium channel Nax functions as a sodium sensor that contributes to epithelial homeostasis. Nax, which is present in multiple epithelial tissues and up-regulated in scars, increases sodium flux and induces the downstream production of mediators of epithelial cell proliferation and inflammation that may lead to scar formation. Indeed, blocking Nax in animal models decreases scarring and atopic dermatitis–like symptoms, suggesting that Nax may contribute to epithelial homeostasis.


The mechanisms by which the epidermis responds to disturbances in barrier function and restores homeostasis are unknown. With a perturbation of the epidermal barrier, water is lost, resulting in an increase in extracellular sodium concentration. We demonstrate that the sodium channel Nax functions as a sodium sensor. With increased extracellular sodium, Nax up-regulates prostasin, which results in activation of the sodium channel ENaC, resulting in increased sodium flux and increased downstream mRNA synthesis of inflammatory mediators. Nax is present in multiple epithelial tissues, and up-regulation of its downstream genes is found in hypertrophic scars. In animal models, blocking Nax expression results in improvement in scarring and atopic dermatitis–like symptoms, both of which are pathological conditions characterized by perturbations in barrier function. These findings support an important role for Nax in maintaining epithelial homeostasis.


Sodium homeostasis is critical to cellular function. In animals, sodium homeostasis is tightly regulated centrally by the kidney. However, in the skin and other epithelia exposed to the external environment, cells potentially need to respond to a different set of cues to maintain sodium homeostasis (1). Many lines of evidence support the importance of maintaining extracellular and intracellular sodium homeostasis, but the upstream mechanism by which sodium levels are sensed and controlled remains unknown. We report here that the atypical sodium channel Nax (scn7a) acts as a sodium sensor in the skin and other epithelia, which activates the principal sodium channel ENaC as well as inflammatory signals, which have the downstream effects of restoring epithelial homeostasis. However, in many instances, epithelial barrier function is chronically disturbed, resulting in sodium dysregulation and chronic inflammation characteristic of many pathological conditions of the skin including hypertrophic scarring and atopic dermatitis.

We have recently shown that ENaC (scnn1), the major sodium channel in epithelial cells, initiates an inflammatory pathway by inducing sodium flux in the epidermis in response to skin barrier disturbance (2). Blocking the inflammatory pathway induced by ENaC resulted in an improvement in scarring and reduction in collagen synthesis. This indicates the importance of these inflammatory mediators in the pathogenesis of excessive scarring. In the lung, sodium dysregulation is an important component of many inflammatory diseases, with ENaC dysregulation specifically implicated in cystic fibrosis (3, 4). In the kidney, ENaC also plays an important role in sodium balance, with dysregulation contributing to hypertension (5, 6). Activation of ENaC heavily depends on proteolytic activities in the cell membrane; therefore, there must be a sensor of sodium concentration that controls these proteolytic activities and thus the activation of ENaC.

An atypical sodium channel, Nax has been demonstrated to act as a sodium-sensing molecule in the central nervous system (CNS), critical to maintaining sodium homeostasis in mammals (7, 8). Here, we report for the first time that Nax in epithelial cells responds to perturbations in sodium homeostasis. We show that this Nax-mediated inflammatory pathway in response to sodium perturbations is present in the epithelium of the skin with potential implications for pathological conditions of the skin, such as human hypertrophic scarring. In animal models of hypertrophic scarring and atopic dermatitis, in vivo knockdown of Nax resulted in a drastic reduction of scarring and improvement of atopic dermatitis, respectively. We hypothesize that Nax is critical in maintaining epithelial barrier function homeostasis by responding to changes in sodium concentration. We also propose that blocking Nax or some of its downstream mediators may have therapeutic value in reducing conditions such as excessive scar and inflammatory dermatitis with barrier dysfunction.


Nax regulates the production of proinflammatory factors in response to perturbation of sodium homeostasis in the skin

Previously, our laboratory showed that reduced hydration status in the skin led to an increase in sodium flux and up-regulation of downstream molecules that are involved in cutaneous wound repair and scarring (2). A reduced hydration condition was simulated by increasing extracellular sodium concentration by 10%, which induced sodium flux and up-regulation of downstream molecules (Fig. 1A). A 10% increase in sodium is a physiological level reached in the extracellular fluid in a water-deprived thirsty animal (911). We previously confirmed that sodium flux in keratinocytes is mediated by ENaC, a constitutively active transmembrane channel (2). Its α-subunit is essential for channel activity (1214). ENaC is activated by proteases, and though its function in sodium flux is well described, the upstream regulation of ENaC activation has not been clearly elucidated. ENaC is a voltage-independent channel with very long opening and closing times, which is consistent with protease activation and inconsistent with a direct sensing mechanism that would have implied rapid opening and closing times (1517).

Fig. 1. Nax is an upstream regulator of inflammatory gene expression in keratinocytes.

(A) Hydration status of skin wounds can be controlled by using occlusive dressings on the top of the wounding area in vivo (left), which is simulated in human ex vivo skin culture (HESC) by hydration (middle) and in cell culture by sodium concentration (right). (B) Summary of expression patterns of ENaC and Nax in mouse tissues (n = 4). (C) Expression changes of ENaC-α, COX-2, IL-1β, and IL-8 under the high-sodium condition in HaCaT cells (n = 4). (D) Effect of the knockdown of Nax or ENaC-α on the expression of COX-2, IL-1β, and IL-8 in HaCaT with high sodium concentration. Gene expression level was compared to mRNA expression in HaCaT cultured with the control sodium concentration, which was set as 1 (n = 4). (E) Expression of COX-2, IL-1β, and IL-8 in HaCaT by the conditional overexpression of Nax with the control sodium concentration. Gene expression was compared to mRNA expression in wild-type HaCaT, which was set as 1 (n = 4). (F and G) Microarray analysis. Global gene (F) and inflammatory gene (G) expressions analysis in wild-type and Nax knockdown HaCaT cultured under the high-sodium condition (165 mM) for 4 and 16 hours (n = 5). **P < 0.01. CSC, control sodium concentration; HSC, high sodium concentration; KD, knockdown.

Nax functions as a sodium-sensing molecule controlling the sodium homeostasis in the CNS (7, 8), but it has not previously been described outside of the CNS. Initially, Nax in the skin was identified in our laboratory when looking for sodium channels that might be responding to an increase in extracellular sodium. Nax was found to be expressed not only in human skin keratinocytes (fig. S1) but also in many different mouse epithelial cells such as those found in the palate, lung, small intestine, stomach, kidney, and colon (Fig. 1B and fig. S2). The distribution of Nax and ENaC in these epithelia was highly overlapped. Critical cytokines and enzymes were identified, including cyclooxygenase-2 (COX-2), interleukin-1β (IL-1β), and IL-8, which are up-regulated in the epidermis when the barrier function of the skin is disrupted. Such disruptions occur in vivo in the maturation phase of wound healing after epithelialization is complete, or they can be simulated in vitro under conditions of reduced hydration or increased extracellular sodium concentration (18). Here, COX-2, IL-1β, and IL-8 were confirmed to all be up-regulated in the human keratinocyte cell line HaCaT, using a 10% increase in cell culture sodium concentration (165 mM compared to 150 mM sodium found in standard culture medium) (Fig. 1C). Up-regulation of these genes was also found in adult human primary keratinocyte (HK) cultures by the 10% increased sodium concentration (fig. S3A). However, ENaC-α gene expression was not up-regulated under conditions of high sodium concentration.

Loss-of-function (LOF) studies were performed to delineate the relation between Nax and ENaC-α on the expression of downstream cytokines. Three small interfering RNA (siRNA) sequences were designed to knock down Nax from HaCaT cells, namely, siNx-1, siNx-2, and siNx-3 (table S1). Knockdown of Nax with each siRNA or all three siRNAs (siNx-mix) led to down-regulation of COX-2, IL-1β, and IL-8 under high-sodium conditions (Fig. 1D and fig. S3B). In contrast, ENaC-α knockdown showed down-regulation of COX-2 but not of either IL-1β or IL-8 (Fig. 1D). Our laboratory has previously characterized the ENaC–COX-2 signaling pathway in the skin (2). For gain-of-function (GOF) studies, Nax was conditionally overexpressed using a tetracycline repressor (TetR) system (fig. S3, C and D). The efficiency of this overexpression system was estimated by overexpressing green fluorescent protein (GFP) in HaCaT cells (fig. S3E). With the presence of doxycycline, an inhibitor of the repressor (TetR), the expression level of Nax was significantly elevated (fig. S3, F and G). Overexpressing Nax in HaCaT cells enhanced the expression of COX-2, IL-1β, and IL-8 without stimulation with high sodium concentration, which further confirmed that Nax is upstream of these three molecules (Fig. 1E). These findings suggest that Nax controls downstream cytokine expression using multiple pathways including ENaC–COX-2. Microarray analysis of Nax knockdown HaCaT cells revealed down-regulation of many genes (Fig. 1F and table S2), especially inflammation-related genes (Fig. 1G), which were up-regulated in response to high sodium treatment. Notably, many genes with the greatest change in Nax knockdown were also identified in our previous microarray study on rabbit wounds with reduced hydration (18). These results suggest that Nax plays a critical role in the response to changes in epidermal barrier function and sodium homeostasis.

Nax manipulates sodium flux of keratinocytes through activation of ENaC

Given that ENaC is the major sodium channel for sodium flux in the skin, we wanted to investigate the potential role of Nax in regulating sodium flux through this channel. The scanning ion-selective electrode technique (SIET) (19, 20) was used to measure sodium flux in a stratified keratinocyte culture model (Fig. 2, A and B) that mimics the human epithelium in vivo (2). To mimic the reduced hydration status of wounded human skin, the cells were kept in a flow-air interface that resulted in water loss (compared to the standard static interface at 100% humidity) (Fig. 2A). Stratified keratinocytes maintained underneath the medium were used as control hydration status, without water loss (Fig. 2A). Sodium flux was measured on the surface of samples after placing them in bathing buffer. A significantly increased sodium flux was observed in the samples pretreated with reduced hydration compared to control HaCaT cells maintained fully submerged in medium (Fig. 2C). Notably, increased sodium flux was abolished in the samples where Nax was knocked down with lentivirus-mediated RNA interference (RNAi) (Nax-KD, fig. S1B) in the reduced hydration condition (Fig. 2C). Increased sodium flux was also detected in stratified HaCaT cells cultured with 10% increased sodium concentration relative to the control, and this increased flux was similarly blocked in the Nax knockdown cells (Fig. 2D).

Fig. 2. Nax controls sodium transport in the skin.

(A) Creation of control and reduced hydration status on stratified keratinocytes. (B) Measurement of sodium flux with SIET on the stratified keratinocytes. (C and D) Wild-type and Nax knockdown HaCaT cells were stratified and placed under decreased humidity (C) and increased extracellular sodium (D) conditions, and sodium flux was measured by SIET. Comparisons were made between wild-type HaCaT control condition and other conditions. (E to G) Intracellular sodium was measured in wild-type (E), Nax knockdown (F), and ENaC-α knockdown (G) HaCaT cells cultured under high sodium (165 mM) and control (150 mM) conditions, using a sodium indicator, SBFI/AM. **P < 0.01. Scale bar, 10 μm. Time course measurements of intracellular SBFI/AM fluorescence signal for all the images were performed. Arrow indicates the time point when cells were treated with high-sodium medium.

Next, using a sodium binding fluorescent dye, SBFI/AM, changes in intracellular sodium concentration as a function of sodium flux were analyzed in epithelial cells in response to increased extracellular sodium concentrations. An increase in intracellular sodium was found in cells pretreated with high sodium, but in cells not pretreated, changing the extracellular sodium from 150 to 165 mM had no effect. This observation would be consistent with a requirement for ENaC activation, which takes time, to see an increase in sodium flux (Fig. 2E). Increased intracellular sodium found in wild-type HaCaT was blocked in the Nax (Fig. 2F) or ENaC-α (Fig. 2G) knockdown HaCaT in response to high extracellular sodium.

Nax regulates ENaC activity through prostasin

We recently demonstrated that sodium flux was mainly mediated through ENaC in epidermal keratinocytes (2). However, the regulation and control of ENaC could be either through increased expression of ENaC or through activation of existing ENaC. To test this, we quantified the expression levels of ENaC-α mRNA in HESC by quantitative polymerase chain reaction (qPCR) in response to reduced hydration. We found that ENaC-α mRNA expression was unchanged compared to the control hydration status in the epidermis (fig. S4B). The activity of ENaC primarily depends on serine proteases in the cell membrane (21). Our microarray studies showed that prostasin, a membrane-bound serine protease, also known as CAP1 or Prss8, was up-regulated in wild-type HaCaT cells in response to elevated sodium concentration but was not changed in Nax knockdown cells (table S2). Prostasin is known to play a critical role in the activity of ENaC (21, 22). Immunostaining demonstrated that prostasin colocalized with Nax or ENaC-α in human skin and was found primarily in the suprabasal epidermal layer (Fig. 3, A and B, and fig. S5). Knockdown of prostasin resulted in the block of sodium flux, demonstrated by no change in the intracellular sodium concentration in response to increased extracellular sodium concentration in keratinocytes (fig. S6A). In addition, the expression of the downstream inflammatory genes, COX-2, IL-1β, and IL-8, was suppressed under the high-sodium condition when prostasin was knocked down (Fig. 3C). This evidence indicates that prostasin is upstream of ENaC and that the channel function of ENaC is principally regulated by activation rather than an increase in channel quantity.

Fig. 3. Regulation of Nax, ENaC, and prostasin expression in epidermal keratinocytes.

(A) Colocalization of Nax and prostasin in human skin was analyzed using their specific antibodies. (B) Colocalization of ENaC and prostasin in human skin. (C) Expression of COX-2, IL-1β, and IL-8 in HaCaT cells with knockdown of prostasin under high sodium concentration (n = 4). (D) Expression of Nax, ENaC-α, and prostasin in HaCaT cells by conditional overexpression of Nax with doxycycline (n = 4). (E)The expression of prostasin was analyzed in Nax or ENaC-α knockdown HaCaT under the high-sodium condition (n = 4).The epidermal basement membrane was indicated by white dashed lines. The epidermis was indicated by a double-headed arrow. *P < 0.05; **P < 0.01. Scale bar, 50 μm. Dox, doxycycline.

The experimental results directly measuring sodium flux using SIET (Fig. 2D) and indirectly measuring intracellular sodium by sodium imaging with SBFI/AM (Fig. 2E) together demonstrate that an hour of incubation in high sodium concentration was necessary to see changes. We hypothesize that a 1-hour period is required for the activation of ENaC, which is essential for the sodium flux. Trypsin, a serine protease, has been used to mimic the function of prostasin in vitro (23). Supporting a protease-based mechanism, pretreatment with trypsin (10 μg/ml) in wild-type HaCaT cells resulted in an immediate increase in sodium flux without the need for pretreatment with high sodium (fig. S6B). Treatment with trypsin (10 μg/ml) also resulted in increased sodium influx into keratinocytes in Nax knockdown cells (fig. S6C). However, trypsin treatment does not induce sodium flux in ENaC-α knockdown cells (fig. S6D). It is shown that the increase of the proteolytic cleavage of ENaC indicates higher activity of ENaC protein (23). Compared to wild-type HaCaT cells, high sodium treatment resulted in less cleaved ENaC-α in Nax knockdown HaCaT cells, when the ratio of cleaved and noncleaved forms of ENaC-α was measured (fig. S6E). Nax knockdown HaCaT cells showed a similar amount of cleaved ENaC-α compared to wild-type HaCaT cells when they were pretreated with trypsin (10 μg/ml) (fig. S6E). These data suggest that Nax regulates ENaC activation through a protease-based mechanism. Together, these results suggest that sodium flux in HaCaT is mediated by ENaC, which is downstream of Nax and prostasin. Consistent with this hypothesis, we found that conditional overexpression of Nax up-regulated the expression of prostasin, but not ENaC, in HaCaT (Fig. 3D and fig. S7). In addition, LOF studies showed that the expression of prostasin in HaCaT upon stimulation with high sodium concentration was regulated by Nax and not by ENaC (Fig. 3E). Both GOF and LOF studies suggest that prostasin is downstream of Nax and mediates the activation of ENaC, and that COX-2 is downstream of both Nax and ENaC.

Intermediate steps in the signaling pathway between sodium sensing by Nax and sodium transport by ENaC were then addressed, focusing on mitogen-activated protein kinase (MAPK) pathways that are activated by environmental stress and show fast kinetics and reversibility (24). A human phospho-kinase antibody array containing 45 phospho-kinase antibodies was used for the identification of the downstream phosphorylation cascade of Nax. Phosphorylation of 14 proteins was increased in keratinocytes within 60 min of stimulation with high sodium concentration (fig. S8A). Among them, phosphorylation of WNK1, p38α, and PYK2 was highly altered by the knockdown of Nax but not ENaC. These kinases are most likely to be involved in signal transduction from Nax to prostasin and ENaC. Knockdown studies of WNK1 (upstream of extracellular signal–regulated kinase 5, fig. S8B) and p38α (fig. S8C), both of which belong to MAPK families, were performed to further validate this signaling pathway. Knockdown of either WNK1 (fig. S8D) or p38α (fig. S8E) in HaCaT cells (fig. S9A) caused a considerable decrease in the expression of COX-2, IL-1β, and IL-8 compared to wild-type cells under the stimulus of high sodium concentration. Overexpression of Nax increased the phosphorylation levels of p38α, but not WNK1, with the control sodium concentration compared to that with high sodium concentration (fig. S8F). Because LOF studies showed that WNK1 and p38α are upstream of ENaC, we investigated whether activation of ENaC rescues their deficiency in downstream expression. Trypsin treatment of WNK1 or p38α knockdown keratinocytes reestablished the up-regulation of COX-2, IL-1β, and IL-8 driven by high sodium concentration (fig. S9, B to D). These data show that increases in sodium concentration are sensed by Nax, which in turn activates MAPK pathway and downstream genes.

Expression of Nax is associated with skin disorders

Given the importance of Nax in regulating both prostasin and ENaC, we investigated whether abnormal expression of Nax is observed in many skin disorders. Notably, serine proteases in keratinocytes (25) and ENaC (26) dysregulation are both known to be involved in skin disorders. Among the most important of these disorders is hypertrophic cutaneous scarring, the ropey stiff and disfiguring scars that occur after burns, injuries, or surgical incisions, and for which the treatment options are currently limited. Impaired barrier function with increased transepidermal water loss (TEWL) plays an important role in the pathogenesis. We analyzed human skin tissues that were clinically diagnosed as hypertrophic scars. The expression of Nax and prostasin in hypertrophic scars is significantly higher than that in normal skin as evidenced by immunofluorescence staining (Fig. 4, A, B, and D). However, the expression of ENaC-α did not show notable difference between normal skin and hypertrophic scar (Fig. 4, C and D).

Fig. 4. Expression of Nax and ENaC-α in human hypertrophic scars and their effect on dermal fibroblasts.

(A to C) Expression of Nax (A), prostasin (B), and ENaC-α (C) in normal (n = 5) and hypertrophic scar skin tissue (n = 9). (D) Quantification of the immunofluorescence (IF) staining using ImageJ. (E and F) Immunofluorescence analyses in the keratinocyte-fibroblast coculture model. The expression of α-SMA (E) and procollagen I (F) in dermal fibroblasts cocultured with wild-type, Nax knockdown, and ENaC-α knockdown HaCaT cells was quantified using ImageJ (n = 4). The epidermal basement membrane was indicated by white dashed lines. The epidermis was indicated by a double-headed arrow. **P < 0.01. Scale bar, 100 μm. CHS, control hydration status; RHS, reduced hydration status.

Dermal fibroblasts play a critical role in scar formation; activated dermal fibroblasts vigorously produce collagen, which results in tissue fibrosis. A keratinocyte-fibroblast coculture model was used to determine whether inhibition of Nax production in keratinocytes could affect the behavior of cocultured dermal fibroblasts (fig. S10A). An increase in the expression of α–smooth muscle actin (SMA) and procollagen I was found when stratified wild-type HaCaT were stimulated with reduced hydration (fig. S10B and Fig. 4, E and F). The genes α-SMA and procollagen I are also up-regulated in activated fibroblasts (myofibroblasts), which are found during the scarring phase of wound healing. After Nax knockdown in the HaCaT, the activation of dermal fibroblast cells was minimized (fig. S10C and Fig. 4, E and F). Similarly, the activation of dermal fibroblasts was also inhibited when the cells were cocultured with ENaC-α knockdown HaCaT in response to reduced hydration status (fig. S10D and Fig. 4, E and F).

In vivo inhibition of Nax expression leads to a reduction in scar formation

Targeting Nax as a strategy in skin conditions characterized by excess inflammation and disturbed barrier function with increased TEWL was assessed in a scar formation model on rabbit ears. Dicer-substrate siRNAs against rabbit SCN7a (Nax-DsiRNAs) were designed and tested for efficiency in in vivo knockdown of Nax, using a well-validated rabbit ear incisional wound model that recapitulates the behavior of human clinical hypertrophic scars (2, 27). Two 2 × 2 incisional grids were created on each rabbit ear to generate multiple superficial wounds (Fig. 5A). Nax-DsiRNA or sham-DsiRNA was conjugated with an alginate-chitosan nanoparticle (ALG/CS NP) as a delivery vehicle and applied directly onto the surface of the incisional grids. A significant reduction of Nax (by 60%) was observed in the epidermis 3 days after treatment with Nax-DsiRNA compared to sham-DsiRNA (Fig. 5, B and C). The expression level of prostasin in the epidermis was also decreased by 50% (fig. S11). The expression of Nax downstream genes, COX-2, IL-1β, and IL-8, was also reduced as predicted (Fig. 5C). We then investigated the effects of Nax knockdown on scarring using our well-validated rabbit ear hypertrophic scar model (2, 18, 27), which recapitulates the behavior of human clinical hypertrophic scars in multiple parameters—both aggravating factors that increase scarring and response to therapy are similar to clinical responses. A 7-mm dermal punch was used to create full-thickness excisional wounds on rabbit ears (Fig. 5D). Nax-DsiRNA or sham-DsiRNA in ALG/CS NP was mixed with 2% methylcellulose and topically applied on wounds of one ear or contralateral ear of rabbits, respectively, after wounds were fully reepithelialized. The scar elevation index (SEI) was calculated by histological analysis to evaluate the hypertrophy of each wound (Fig. 5E). SEI was reduced by 25% in the Nax-DsiRNA–treated wounds compared to sham DsiRNA–treated wounds (Fig. 5F).

Fig. 5. In vivo inhibition of Nax reduced hypertrophic scar formation.

DsiRNAs were applied on rabbit ear incisional wounds to suppress the expression of Nax in the skin. (A) Rabbit ear incisional wound model. (B) Nax levels were examined by Western blotting within the rabbit ear incisional model. Control and treatment ears were administered with sham-DsiRNA and Nax-DsiRNA, respectively, 2 days after wounding. RNA samples were isolated from both groups 3 days after treatment. (n = 4). (C) Quantification of the expression of Nax, prostasin, COX-2, IL-1β, and IL-8 by qPCR (n = 6). Comparisons were made between sham-DsiRNA– and Nax-DsiRNA–treated samples. (D) Rabbit ear excisional wounds at postoperation day 0 (POD 0, left) and POD 28 (right). (E) Calculation of SEI to evaluate the hypertrophy of the scar. (F) Hematoxylin and eosin (H&E) staining of the rabbit ear hypertrophic scars treated with sham-DsiRNA or Nax-DsiRNA in 2% methylcellulose at POD 28 (n = 12). **P < 0.01. Scale bar, 500 μm.

In vivo inhibition of Nax expression improved the quality of atopic dermatitis–like skin

Atopic dermatitis, which currently lacks effective treatment, is an increasingly prevalent skin disorder characterized by intense itch and increased proinflammatory gene expression along with barrier dysfunction and increased epithelial water loss (25). Nax was targeted in a well-validated atopic dermatitis–like model induced by topical application of oxazolone (2830) on the back of hairless mice (SKH1-Elite), which produces the characteristic signs of atopic dermatitis including dry and scaly skin (fig. S12, A to C). The expression of Nax and its downstream proteins was elevated in the epidermis of atopic dermatitis–like skin compared to healthy skin (fig. S12D). To inhibit the expression of Nax in the epidermis, DsiRNA was synthesized against mouse Nax (table S1), and its efficacy was tested in vitro by confirming the down-regulation of Nax and its downstream proteins (fig. S12E). Nax-DsiRNA or sham-DsiRNA in ALG/CS NP was applied to either side of the dorsal skin of oxazolone-induced dermatitis mice (Fig. 6A). Nax-DsiRNA–treated skin showed markedly less signs of irritation and rash compared to the sham-DsiRNA–treated side (Fig. 6B). The thickness of skin epidermis and dermis (hyperkeratosis), another parameter to evaluate the skin quality, was significantly decreased in the Nax-DsiRNA–treated side (Fig. 6, B and C). Molecular analysis showed that the expression level of Nax in the epidermis of Nax-DsiRNA–treated skin was decreased by 40% compared to the sham control side (Fig. 6D). In addition, downstream genes of Nax, prostasin, COX-2, and IL-1β, showed high reduction by Nax-DsiRNA treatment (Fig. 6D). This result indicates that the proinflammatory expression pathway is regulated by Nax in atopic dermatitis.

Fig. 6. In vivo inhibition of Nax relieved atopic dermatitis–like symptoms in hairless mice.

(A) Schematic of the atopic dermatitis–like model and application of treatments. (B) Effectiveness of the treatment of atopic dermatitis–like mouse skin with DsiRNA was estimated by the measurement of the thickness of the epidermis and dermis. (C) Blue represents the thickness of the dermis, whereas orange represents the thickness of the epidermis in the bar graph. (D) Expression levels of Nax, prostasin, COX-2, and IL-1β in the epidermis with Nax-DsiRNA or sham-DsiRNA treatment were measured by qPCR (n = 6). Comparisons were made between sham-DsiRNA– and Nax-DsiRNA–treated samples. **P < 0.01. Scale bar, 50 μm. E, epidermis; D, dermis.


Impaired skin barrier function with increased water loss is characteristic of many skin disorders such as scarring, atopic dermatitis, ichthyosis, and psoriasis. During the remodeling phase of wound healing after epithelialization is complete, the return of epidermal barrier function to normal takes weeks to months. Whereas it is known that there is a close association between perturbed barrier function, increased water loss, and excess inflammation, the upstream signals resulting in inflammation have not been characterized. Our hypothesis is that increased water loss raises extracellular sodium concentrations, which induces a Nax response. In turn, ENaC is activated through the membrane serine protease prostasin, resulting in increased sodium flux. A signal transduction pathway mediated directly through Nax and secondarily through ENaC results in production of secretory inflammatory mediators. These mediators result in epithelial proliferation and restoration of epidermal homeostasis, but they can also have negative effects including excess inflammation that ultimately lead to activation of fibroblasts.

Nax controls a proinflammatory signaling pathway

Simulation of reduced hydration or increased water loss in the skin was performed in vitro by increasing extracellular sodium concentration by 10%, which mimics the level reached in the extracellular fluid in a water-deprived thirsty animal (2). Here, Nax was identified as a sodium sensor in the skin, and the pathway that directs the signaling transduction was characterized (Fig. 7). Sodium flux, which is conducted by ENaC in epithelia, was detected in a short period (within 1 hour) in response to reduced hydration or increased sodium conditions. Supporting this notion, MAPKs were found to be phosphorylated immediately in response to this condition. On the basis of gene knockdown studies with both Nax and prostasin, as well as overexpression studies using Nax, we propose that the signal transduction pathway in response to increased sodium concentration is initiated by Nax and mediated by prostasin through MAPKs. Our previous work (2) demonstrated that sodium flux through ENaC results in rapid production of prostaglandin E2, an important proinflammatory mediator produced by COX-2. However, in this study, IL-1β and IL-8 were found to be downstream of Nax but not affected by ENaC in ENaC-α knockdown studies. Another pathway, which is not fully investigated in this study, may control the production of IL-1β and IL-8. One of the candidates is protease-activated receptor 2 (PAR-2), a G protein–coupled receptors (GPCRs) in epithelial cells, which is an important target of prostasin with potential downstream cytokines including IL-1β and IL-8 (3133). Knockdown of PAR-2 eliminated the up-regulation of IL-1β and IL-8 under the stimulation with high sodium concentration in HaCaT cells (fig. S13A). Moreover, trypsin (10 μg/ml) treatment up-regulated the expression of COX-2, but not IL-1β or IL-8, in PAR-2 knockdown HaCaT cells (fig. S13B). These results suggest that IL-1β and IL-8 are downstream of PAR-2. Activation of the downstream pathways of Nax through ENaC or unidentified proteins, such as PAR-2, results in the rapid production and release of COX-2 and IL-1β/IL-8 (Fig. 7). These proinflammatory factors are commonly seen in numerous inflammatory diseases and damages in the lung, kidney, and skin, and often result in the fibrosis of connective tissues. Nevertheless, these pathways function in the skin and possibly all other epithelia, such as the respiratory epithelium, which express Nax. Targeting Nax in vivo in two different animal models led to both reductions in scarring and amelioration of atopic dermatitis–like symptoms.

Fig. 7. The Nax pathway in keratinocytes.

The putative pathway of Nax in the regulation of downstream proinflammation cytokine expression. The downstream pathways of Nax were activated through ENaC or unidentified proteins (such as GPCR) to up-regulate the expression of COX-2/prostaglandin E2 and IL-1β/IL-8, respectively.

Nax (SCN7A) is classified as a member of voltage-gated sodium channels (VGSCs). It is called as a typical VGSC because the important sequences for voltage sensing and inactivation are different from other VGSCs. When Nax complementary RNA was injected into Xenopus oocytes and cell line, no functional channel activity was detected (34, 35). Analysis showed that Nax functions as a sodium concentration–sensitive channel, not a voltage-dependent channel, which responds to increased extracellular sodium concentration in the excitable neuronal cells (7, 8). Others suggest that Nax is a sodium leak channel that allows sodium to pass the membrane in the nervous system (36). Our results indicated that Nax mediates the sodium influx through activating ENaC because activation of ENaC with trypsin treatment in Nax knockdown HaCaT cells recovered sodium intake (fig. S6). Although our data showed that Nax senses extracellular sodium concentration changes, the mechanisms of the sodium sensing and channel function of Nax are yet unknown.

Nax regulates prostasin at both the mRNA and protein levels

The activities of prostasin (prss8), a membrane protein that has been found in most epithelial cells (37), result in the activation of multiple membrane-anchored ion channels and receptors. Prostasin is made of a precursor form that consists of a 32–amino acid signal peptide and a 311–amino acid proprostasin. Cleavage of proprostasin, between Arg12 and Ile13, generates a 299–amino acid mature prostasin (37, 38). The maturation of prostasin is driven by a protease cascade mediated by matriptase (39). Malfunction of the matriptase-prostasin proteolytic cascade results in compromised epidermal barrier function (40, 41). Here, the ENaC-mediated sodium flux was detected 1 hour after stimulation with high sodium concentration. This proteolytic cascade can be one of the mechanisms used by Nax to activate prostasin within a short period in response to high extracellular sodium. A sustained increase in proinflammatory genes was also found (Fig. 1C) after 16 hours of exposure to high sodium, with increased levels of prostasin mRNA with 4 hours of high-sodium stimulation (Fig. 3E), suggesting that some of the regulation of prostasin is through increased production. The combination of these short- and long-term mechanisms used by Nax provides both rapid and sustained means for the control of prostasin as well as the expression of downstream factors.

Nax likely plays essential roles in epithelial cell differentiation and migration

ENaC is known to control the physiological changes of epithelial cells such as cell migration (42, 43), cell differentiation (44), and cell proliferation (45). As the upstream regulator of ENaC, Nax expression changes may also similarly alter the physiological characteristics of cells. Here, knockdown of Nax from HaCaT cells significantly delayed cell migration compared to wild-type HaCaT (fig. S14, A to C). Knockdown of Nax from keratinocytes also led to a delayed stratification of keratinocytes and less stratified cell layers and the absence of critical keratinocyte differentiation markers (fig. S14, D and E). Combined with the result that showed increased number of layers when the keratinocytes overexpressed Nax (fig. S7), we concluded that Nax is involved in keratinocyte differentiation. The interrupted cell differentiation in Nax knockdown keratinocytes was rescued by periodic treatment with trypsin (10 μg/ml), which activates ENaC activity (fig. S14F). Because trypsin has been widely used as an external ENaC-activating serine protease in many cases (46, 47), it is likely that the application of trypsin recovered the pathway through ENaC in the Nax knockdown cells. This is more indirect evidence supporting the hypothesis that Nax mediates the downstream gene activities by proteolytically activating ENaC.

The wide distribution of Nax makes this channel protein potentially involved in the regulation of many cell activities through mediating sodium homeostasis. Therefore, Nax is a likely candidate to regulate ENaC activity in other organs such as the kidney and sodium flux in the gastrointestinal tract and respiratory epithelium. Given the broad importance of sodium homeostasis in epithelial cellular function, Nax may be a potential therapeutic target for other clinical conditions in addition to hypertrophic scarring and atopic dermatitis.


Study design

The objective of this study is to identify the roles of Nax in barrier dysfunction–related skin diseases. We used both in vitro and in vivo systems to meet this objective. The images for Western blotting and immunofluorescence staining were all processed by ImageJ (48). For Western blot analysis and immunofluorescence, the signal of each band or cells translated into the total intensities. The values of signal intensities were used to show the difference in gene expression. The intensities of the signals were used for Student’s t test analysis. P values less than 0.05 were considered to be significantly different. The number of samples used in the analyses was described in the relevant figure legends. LOF studies of Nax were performed in mouse and rabbit skin. The images of histological slides of sham-DsiRNA– and Nax-DsiRNA–treated samples were analyzed with NIS Elements (Nikon Instruments Inc.) in a blinded manner. The treated and nontreated scars were compared and analyzed with paired t test. Sample numbers are indicated in the figure legends. Normal human skin or hypertrophic scar samples were obtained from the Division of Plastic Surgery at Northwestern Memorial Hospital or the Department of Plastic Surgery of the Union Hospital, Tongji Medical College at Huazhong University of Science and Technology in China. All human samples were obtained under a protocol approved by the Institutional Review Board of Northwestern University or Tongji Medical College, Huazhong University of Science and Technology in China. All the patients involved in this study gave written informed consent before sample collection. Human biopsy samples were taken from patients through scar excision from chest skin. The causes of hypertrophic scar were burn, cut, or surgery. Biopsies were taken from scars that were formed at least 1 year before study enrollment; the patients were of ages 20 to 35 years.

qPCR, Western blot, and immunofluorescence staining

For the HESC samples, epidermal keratinocytes were separated from dermis by incubating the HESC in 0.5 M ammonium thiocyanate at room temperature for 30 min. The separated epidermis layer was homogenized with bead-beating at 5000 rpm for 60 s in TRI Reagent. Total RNA was isolated following a standard protocol, and complementary DNA (cDNA) was synthesized with SuperScript II Reverse Transcriptase (Life Technologies) after removing contaminated DNA with the Turbo DNA-free Kit (Life Technologies). qPCR using SYBR Green I was performed on an ABI 7900HT Fast Real-Time PCR System (Applied Biosystems).

The expression of each gene was normalized to the level of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Gene expression between the groups was calculated using the 2−ΔΔCt method. The primer sets are shown in table S3.

For Western blot analysis, whole cell extracts were prepared using radioimmunoprecipitation assay buffer (50 mM tris, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, and 0.5% sodium deoxycholate, pH 7.5). Electrophoresis, transmembrane blotting, and antibody incubation were performed following standard protocols. The antibodies used in our study included Nax (rabbit x human polyclonal, Abcam, 1:1000 dilution), ENaC-α (rabbit x human polyclonal, Abcam, 1:5000), prostasin (mouse x human monoclonal, R&D, 1:1000), p-WNK1 (rabbit x human polyclonal, Cell Signaling, 1:1000), p-p38 (rabbit x human polyclonal, Cell Signaling, 1:2000), and β-actin (rabbit x human polyclonal, Sigma-Aldrich, 1:5000). Horseradish peroxidase–conjugated anti-rabbit immunoglobulin G (IgG) or anti-mouse IgG (Vector Laboratories) was used as a secondary antibody.

Immunofluorescence staining was performed with slides sectioned from paraffin-embedded tissues. The slides were first deparaffinized and treated with Dako antigen retrieval solution (Dako) at 100°C for 20 min. The slides were then incubated with primary antibodies at 4°C overnight, followed by 1.5-hour incubation with fluorescently labeled secondary antibodies at room temperature. The primary antibodies used for immunofluorescence staining included cytokeratin 5 (rabbit x human polyclonal, Abcam, 1:5000), cytokeratin 10 (mouse x human monoclonal, Dako, 1:5000), α-SMA (rabbit x human polyclonal, Santa Cruz Biotechnology, 1:200), and procollagen I (mouse x human monoclonal, Developmental Studies Hybridoma Bank at University of Iowa, 1:1000).

Keratinocyte cells, HESC, keratinocyte-fibroblast coculture, and mouse samples

Primary keratinocytes were isolated from adult human skin specimens that were obtained from elective abdominoplasties. Epidermis was separated from dermis through treatment with 0.5% Dispase (Life Technologies) in phosphate-buffered saline (PBS) at 4°C overnight. The epidermis layer was digested with 0.5% trypsin (Corning) in PBS at 37°C followed by filtration with a 100-μm cell strainer. The isolated keratinocytes were cultured in Defined Keratinocyte Serum-Free Medium (Life Technologies). Dermal fibroblasts were isolated from human infant foreskin tissues and prepared by the Keratinocyte Core of Northwestern University Skin Disease Research Center (Northwestern University Feinberg School of Medicine). The detailed protocols to make HESC and keratinocyte-fibroblast coculture models were previously described in our earlier studies (2, 18, 49). The mouse tissues for the immunofluorescence staining of Nax and ENaC-α were collected from wild-type C57BL/6 mice.

Gene knockdown in HaCaT cells

The knockdown of target genes was performed by short hairpin RNA (shRNA) through transduction of lentivirus in HaCaT cells. Three shRNA sequences (siNx-1, siNx-2, and siNx-3 in table S1) targeting three different locations in human Nax were selected from the shRNA database maintained by Sigma-Aldrich. The double-stranded shRNA sequences were synthesized from Integrated DNA Technologies (IDT DNA) and cloned in pLKO.1 puro (Addgene). The target sequences are listed in table S2. The procedures for the production of lentivirus and knockdown of selected genes were performed following the protocol from the Web site of Addgene (2). Transduced HaCaT cells were selected by adding puromycin (2 μg/ml) in cell culture media. Knockdown of genes was confirmed by Western blotting. The cells were ready to use after at least three passages of selection. For the knockdown of three target sites, double-stranded shRNA sequences were cloned into pLKO.1-puro, pLKO.1-hydro, and pLKO.1-blast (Addgene). Transduced HaCaT cells were selected in the presence of puromycin (2 μg/ml), hydromycin (300 μg/ml), and blasticidin (10 μg/ml). The efficiency of the Nax knockdown was estimated by the expression levels of downstream genes, IL-1β, IL-8, and COX-2, using qPCR.

Inducible gene expression

HaCaT-TetR cells that stably express the TetR were developed (fig. S3). Briefly, TetR-expressing lentiviruses (pLenti6/TR, Life Technologies) were infected into HaCaT and cultured in the presence of blasticidin (10 μg/ml). Cells were replated at low density, and clones were analyzed for TetR expression (fig. S3D). pLenti CMV/TO-GFP, which contains a CMV/TO promoter regulating GFP, was transfected into a HaCaT-TetR clone that has higher expression of TetR. Induction of GFP by doxycycline was tested in the HaCaT-TetR cell line (fig. S3E). The open reading frame of human Nax (a gift from M. M. Tamkun at the Colorado State University) was cloned into a lentiviral vector (pLenti CMV/TO DEST puro) containing a puromycin resistance gene. Human Nax–expressing lentiviruses were infected into the HaCaT-TetR cell line, and transductants were selected in the presence of puromycin (2 μg/ml). Induction of Nax expression by doxycycline (1 μg/ml) was confirmed by qPCR and Western blot analysis (fig. S3, F and G).

Microarray analyses with HaCaT

Both wild-type and Nax knockdown HaCaT cells were seeded in 12-well plates. Cells were grown to 70 to 80% confluence and then starved with fetal bovine serum (FBS)–free Dulbecco’s modified Eagle’s medium (DMEM) overnight, followed by treatment with control medium (final 150 mM sodium) or medium plus 10% more sodium (final 165 mM sodium). Cells were harvested at 0, 4, and 16 hours after treatment. Total RNA from each sample was isolated using TRI Reagent (Sigma-Aldrich), and DNA was removed with the Turbo DNA-free Kit (Life Technologies). Samples were further cleaned up using an RNeasy Mini Kit (Qiagen). RNA concentration was determined using a NanoDrop spectrophotometer (NanoDrop Technologies Inc.), and sample quality was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies Inc.). cDNA synthesis, dye incorporation, hybridization, and chip scanning were performed at the Interdisciplinary Center for Biotechnology Research Microarray Core, University of Florida. The microarray chips were Affymetrix GeneChip Human Transcriptome Array 2.0 (Affymetrix). The microarray data were processed with Affymetrix Expression Console and Transcriptome Analysis Console (Affymetrix). Gene expression clustering analysis was performed with heatmap.2 in R. Functional clustering analysis was performed with the online version of Data for Annotation, Visualization, and Integrated Discovery (DAVID) (22, 24). The microarray data were deposited in the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) with a serial record GSE65366.

Protein kinase phosphorylation array with HaCaT cells

Keratinocyte protein kinase phosphorylation assay was performed with the Human Phospho-Kinase Antibody Array kit (R&D Systems). Wild-type, Nax knockdown, and ENaC-α knockdown HaCaT cells were cultured in six-well plates to reach 80% confluence. After overnight starvation with FBS-free DMEM, the cells were stimulated with DMEM with 165 mM Na+ for 10, 30, and 60 min. Cells treated with normal DMEM (150 mM Na+) were used as a control. The cells were harvested with the lysis buffer included in the kit to adjust the protein concentration to 1.5 mg/ml. The assay was performed following the user’s manual. The developed film for each array membrane was scanned and analyzed using ImageJ as described earlier for Western blot. For the analysis of protein kinase phosphorylation in Nax-overexpressing HaCaT cells, Nax was conditionally induced by adding doxycycline (1 μg/ml) in HaCaT-TetR-Nax cells cultured under control conditions (150 mM sodium). Cells were harvested at 0 hours (n = 2), 8 hours (n = 3), and 24 hours (n = 3) after induction.

Sodium flux measurements with SIET

The sodium fluxes were measured using sodium-selective electrodes containing a specific ionophore cocktail (Sigma-Aldrich) as previously described (2). The construction of the electrode was described previously (20). Briefly, a 2- to 5-μm-diameter glass tip was made by mechanically pulling a 1.5-mm borosilicate glass capillary tube (World Precision Instruments Inc.) using a Sutter P-97 horizontal puller (Sutter Instrument Co.). Thereafter, the tip was baked at 200°C for 4 to 6 hours and silanized with N,N-dimethyltrimethylsilylamine (Sigma-Aldrich). A 100 mM NaCl solution was backfilled to the tip, and the ionophore was front-filled. An Ag/AgCl 3M KCl dip-type electrode (Microelectrodes Inc.) was used as the reference electrode.

The sodium flux measurement was performed with the self-referencing (SRA) modality (20, 50). The electrode was controlled by a three-dimensional (2D) motion control system (CMC-4, Applicable Electronics Inc.) with a submicron-resolution stepper motor. The position of the electrode was visualized by a video camera attached to a zoom scope with 2D movement. The reading of the signals from the electrode was translated and calculated by the automated scanning electrode technique (ASET) software (Science Wares Inc.).

All the electrodes used in this project were calibrated with a serial dilution of NaCl solutions to confirm nernstian behavior. The sodium flux was measured about 50 μm above the tissue or cells by the electrode and moved 50 μm in the z direction. The difference of the sodium concentrations between the two positions was calculated by the digital differential filtering technique (19). Then, the actual sodium flux was determined using Fick’s first law of diffusion: J = DC)(ΔX−1) (J is the ion flux, D is the ion diffusion coefficient, C is the ion concentration, and X is the electrode excursion distance).

The measurement of ion flux was performed when samples emerged in different types of bathing buffers according to the type of ion that was measured. The HESC or keratinocyte samples were pretreated with normal culture medium (150 mM Na+) or normal medium with 10% additional NaCl (165 mM Na+) for more than 1 hour. Samples were then washed with a sodium-free buffer containing 3.6 mM KCl, 1 mM MgCl2, 0.7 mM CaCl2, 5 mM glucose, 10 mM Hepes, and 312 mM d-sorbitol, and incubated with sodium-free buffer for 30 to 60 min. The sodium fluxes of the samples were measured in measuring buffer with 1 mM NaCl, 3.6 mM KCl, 1 mM MgCl2, 0.7 mM CaCl2, 5 mM glucose, 10 mM Hepes, and 310 mM d-sorbitol.

Sodium imaging with SBFI/AM cell permeant

Cells were seeded on glass coverslips overnight at 37°C with 5% CO2. The buffers used for sodium imaging assay were prepared according to a previously described protocol (51) with minor changes. The coverslips with cells were washed with sodium-free buffer containing 3 mM KCl, 2 mM MgCl2, 5 mM tris, 10 mM glucose, and 313 mM sucrose (pH 7.4, osmolality ∼340 mosmol/liter). Then, the coverslips were incubated with 0.01% Pluronic F-127 and 10 μM SBFI/AM sodium fluorescent dye (sodium-binding benzofuran isophthalate acetoxymethyl ester, Life Technologies) in sodium-free buffer for 1 hour at 37°C for complete dye deesterification. The coverslip was then transferred to an imaging chamber mounted on an inverted epifluorescence microscope (IX71, Olympus) with xenon illumination. The chamber was heated at 34°C. Cell images were captured by a cooled charge-coupled device camera (I-PentaMax, Princeton Instruments) and processed with MetaFluor imaging software (Molecular Devices). The coverslip was then superfused with low-sodium buffer (150 mM NaCl, 3 mM KCl, 2 mM MgCl2, 5 mM tris, 10 mM glucose, and 13 mM sucrose, pH 7.4, osmolality ∼340 mosmol/liter) at a flow rate of 2 to 3 ml/min. The cells were visualized under the microscope with a 40× (numerical aperture 1.35) oil-immersion objective (Olympus). Two excitation filters at wavelengths 340 and 380 nm were used, and the emission was monitored at 500 nm. Ratiometric images (F340/F380) were taken every 30 s.

ALG/CS NP–conjugated DsiRNA preparation

Briefly, DsiRNA was diluted in 18 mM CaCl2 solution. Polyethylenimine (1 mg/ml) was added to the solution with a ratio to DsiRNA of 1:6. The solution was added to 0.075% sodium alginate (pH 4.9) solution dropwise with continuous stirring. Thereafter, 0.05% chitosan in 0.025% acetic acid solution (pH 4.6) was added dropwise to the premixed solution, and the solution was continuously stirred at room temperature for 45 min. The nanoparticle was formed and collected by centrifugation at 15,000g at 4°C for 20 min. The collected nanoparticle with DsiRNA was resuspended with PBS and stocked at 4°C until use.

Functional studies of Nax in rabbit incisional wound repair and hypertrophic scar model

Female New Zealand White rabbits (3 to 4 kg) were used. The rabbit ear incisional wound model was previously reported (18, 27). Briefly, two 2 x 2 cm2 superficial incisional grids were created on the ventral side of each rabbit ear using a scalpel with about 1-mm space between each two incisions. Tegaderm (3M) was applied on top of each ear until it was fully reepithelialized (48 hours after wounding). A total amount of 10 μg of ALG/CS-Nax-DsiRNA or ALG/CS-sham-DsiRNA was mixed with 2% methylcellulose and applied topically on each ear to completely cover all the wounding areas, followed by application of Tegaderm on top of the wound. Samples were harvested 3 days after treatment from four rabbits in each group.

The rabbit ear cutaneous hypertrophic scar model used for the in vivo studies of Nax was described in our earlier studies (5255). Six 7-mm full-thickness punch biopsies were made on the ventral side of each rabbit ear. The ventral side of rabbit ear was covered with Tegaderm to prevent desiccation of the wounds until the wounds were fully reepithelialized (around day 14 after wounding). ALG/CS-Nax-DsiRNA or ALG/CS-sham-DsiRNA (1.5 μg) mixed with 2% methylcellulose was applied topically to each wound. A new Tegaderm was applied on the top of each ear for 4 days after treatment. The rabbit ear scar samples were harvested at day 28 after wounding.

Both incisional and excisional wound samples were fixed in 10% neutralized formalin overnight followed by dehydration and embedding with paraffin. A part of samples from each incisional wound was used for RNA isolation. Histological slides of 5-μm thickness were used for H&E staining and further microscopy analysis. The measurement and evaluation of scar formation are shown in Fig. 5E. The SEI was calculated as described before (52, 54).

In vivo inhibition of Nax in the mouse dermatitis-like model

A mouse dermatitis-like model was created on hairless mice (SKH1-Elite) following a previously described protocol (30) with minor modifications. Briefly, 5% oxazolone in 100% ethanol was applied to the dorsal hindlimbs of a hairless mouse for sensitization. Seven days later, 0.2% oxazolone solution was spread to the dorsum of the mouse every other day for 14 days or until a sign of atopic dermatitis was observed on the back (Fig. 6A). The preparation of Nax-DsiRNA was applied on the atopic dermatitis–like model after signs of atopic dermatitis were visible on the dorsum of the mouse. As described above, the mixture of three Nax-DsiRNA or sham-DsiRNA was conjugated with ALG/CS nanoparticles. Then, the Nax-DsiRNA or sham-DsiRNA was mixed with Safeway Total Moisture Lotion with a ratio of 2-μg DNA in 20-μl lotion. Stimulation with oxazolone was still applied on the back of the mouse 4 hours before the treatment to sustain the inflammation. The lotion containing Nax-DsiRNA or sham-DsiRNA was thereafter gently rubbed symmetrically on the dorsal skin of the mouse on each side of the spine every other day for three times. The skin on the back of the mouse was then harvested for histological and molecular analyses.

Penetration of ALG/CS NP carried DsiRNA (ACN-DsiRNA) into epidermis was estimated by TYE563-labeled DsiRNA control (IDT DNA), which was conjugated with ALG/CS NP (ACN-TYE563-DsiRNA). The skin of the hairless mouse was tape-stripped with packing scotch tape 10 times to remove the stratum corneum to mimic the disturbance of skin barrier function caused by atopic dermatitis. ACN-TYE563-DsiRNA in moisture lotion (2-μg DNA in 20-μl lotion) was applied topically on the skin. Treatment was applied every other day for 6 days, and the skin was harvested at day 6. Skin not treated with ACN-TYE563-DsiRNA was used as a negative control.

Keratinocyte migration assay

Monolayer HaCaT scratch assay was used to study the migration of keratinocyte cell sheet. HaCaT cells were seeded into a 12-well plate until more than 95% confluence. A 500-μm-wide scratch was made with a pipette tip through the middle of each well. At a certain field of each gap, the gap distance was measured at 0, 16, 40, and 64 hours after scratching. At least three observation fields in each gap were monitored in the time course, and the average width of each gap was calculated and used to estimate the cell migration rate.


Fig. S1. Nax immunostaining in keratinocytes.

Fig. S2. Expression of Nax and ENaC-α in mouse tissues (C57BL/6).

Fig. S3. Nax downstream gene expression in human primary keratinocytes, and knockdown and overexpression of Nax in HaCaT cells.

Fig. S4. The skin equivalent models used in the study.

Fig. S5. Immunofluorescence staining of Nax, prostasin, and ENaC-α.

Fig. S6. Change of the intracellular sodium concentrations in HaCaT cells.

Fig. S7. Expression of Nax, prostasin, and ENaC-α in stratified HaCaT cells.

Fig. S8. Cluster analysis of the protein kinase phosphorylation array.

Fig. S9. mRNA expression in WNK1 or p38α knockdown HaCaT cells treated with external serine protease.

Fig. S10. Activation of dermal fibroblasts was modulated by cocultured keratinocytes.

Fig. S11. Expression of prostasin in Nax-DsiRNA– or sham-DsiRNA–treated wounds was detected with its specific antibody and visualized with a fluorescently labeled secondary antibody.

Fig. S12. Inhibition of the Nax pathway by DsiRNA.

Fig. S13. Involvement of PAR-2 in the Nax pathway.

Fig. S14. Knockdown of Nax in HaCaT cells.

Table S1. Sequence information used for RNAi.

Table S2. Differentially expressed genes in wild-type and Nax knockdown keratinocytes under the stimulation with high sodium concentration.

Table S3. Sequence information for qPCR primers.

Source data


  1. Acknowledgments: We would like to acknowledge M. M. Tamkun from the Department of Biochemical Sciences at the Colorado State University, College of Veterinary Medicine, for the full-length human Nax clone. We thank E. Friedrich in the laboratory, A. Mustoe from the University of North Carolina, and K. Stallcup from Northwestern University for the critical reading of the manuscript. Funding: This study was supported by internal funding from the Division of Plastic and Reconstructive Surgery, Northwestern University Feinberg School of Medicine, and in part from the Geneva Foundation a grant to D.J.S. from the JPB Foundation. Author contributions: W.X., S.J.H., D.M.P., D.J.S., K.P.L., R.D.G., and T.A.M. designed the research; W.X., S.J.H., A.Z., S.J., P.X., Z.X., M.Z., J.Z., and S.N.-B. performed the research; D.M.P., D.J.S., and K.P.L. contributed new reagents/analyzed data; and W.X., S.J.H., and T.A.M. wrote the paper. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The microarray data are deposited in the NCBI GEO with a serial record GSE65366.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article