Epithelial sodium channels in macrophage migration and polarization: role of proinflammatory cytokines TNFα and IFNγ

Abstract

Migration of monocytes-macrophages plays an important role in phagocytosis of pathogens and cellular debris in a variety of pathophysiological conditions. Although epithelial Na⁺ channels (ENaCs) are required for normal migratory responses in other cell types, their role in macrophage migration signaling is unknown. To address this possibility, we determined whether ENaC message is present in several peripheral blood monocyte cell populations and tissue-resident macrophages in healthy humans using the Human Protein Atlas database (www.proteinatlas.org) and the mouse monocyte cell line RAW 264.7 using RT-PCR. We then determined that selective ENaC inhibition with amiloride inhibited chemotactic migration (∼50%), but not phagocytosis, of the mouse monocyte-macrophage cell line RAW 264.7. Furthermore, we generated a cell line stably expressing an NH₂-terminal truncated αENaC to interrupt normal channel trafficking and found it suppressed migration. Prolonged exposure (48 h) of RAW 264.7 cells to proinflammatory cytokines interferon γ (IFNγ) and/or tumor necrosis factor α (TNFα) inhibited RAW 264.7 migration and abolished the amiloride (1 µM)-sensitive component of migration, a finding consistent with ENaC downregulation. To determine if proinflammatory cytokines regulate αENaC protein expression, cells were exposed to proinflammatory cytokines IFNγ (10 ng/mL, last 48 h) and TNFα (10 ng/mL, last 24 h). By Western blot analysis, we found whole cell αENaC protein is reduced ≥50%. Immunofluorescence demonstrated heterogeneous αENaC inhibition. Finally, we found that overnight exposure to amiloride stimulated morphological changes and increased polarization marker expression. Our findings suggest that ENaC may be a critical molecule in macrophage migration and polarization.

INTRODUCTION

Cytokines are integral components of immune responses in health and disease. Proinflammatory cytokines modulate inflammatory cell function to fight infection or scavenge/repair injured tissue and can alter nonimmune cell function including altering ion channel function. Our laboratory is interested in the effects of cytokines on epithelial Na⁺ channel (ENaC) proteins, a family of ion channels with important roles in 1) salt and water transport in the kidney, colon, and lung, 2) cellular migration, and 3) maintenance of vascular tone in the small arteries and arterioles in the brain and kidney (1–10). Multiple studies suggest proinflammatory cytokines, tumor necrosis factor α (TNFα), and interferon γ (IFNγ) inhibit ENaC expression and function in gut epithelial and lung alveolar cells leading to disruption of salt and water reabsorption (11–15). TNFα also suppresses βENaC expression in vascular smooth muscle cells (VSMCs), contributing to cerebrovascular dysfunction, in a rodent model of preeclampsia (16).

A limited number of empirical studies suggest that ENaC subunits or their closely related family members, acid-sensing ion channel (ASIC) subunits, are expressed in peripheral blood mononuclear cells (PBMCs) (17–20). Recently, ENaC subunits were identified in dendritic cells, a population of phagocytic, antigen-presenting immune cells that participate in propagation of an immune response by interacting with and stimulating T-lymphocytes (17, 21). PBMC expression of ENaC subunits is also supported by several large-scale expression studies using flow cytometry and RNA-sequencing showing ENaC message is expressed specifically in human monocytes and macrophages of the peripheral blood and in resident macrophages of the liver, placenta, lung, pancreas, skin, and testes, albeit at a lower level of expression relative to epithelial tissue of the kidney, gut, and lung (www.proteinatlas.org) (22–26). Because ENaC subunit expression in epithelial tissues is regulated by TNFα, a key cytokine involved in activation of macrophages, we questioned whether ENaC subunits might contribute to macrophage function.

Migration of macrophages plays a critical role in phagocytosis of pathogens and cellular debris in a variety of pathophysiological conditions. After an injury, circulating blood monocytes traffic into the injured tissue by chemotactic signals. In the tissue, exposure to other cytokines and chemokines transform monocytes into macrophages that initiate the healing process (phagocytosis followed by extracellular matrix formation). ENaC subunits are required for migration in several cell types including VSMCs, trophoblasts, and keratinocytes; however, their importance in monocyte-macrophage migration has never been addressed (1, 27–30). Based on the impact of proinflammatory cytokines on ENaC expression and the importance of ENaC in migration in nonimmune cells, we hypothesized that proinflammatory cytokines inhibit ENaC expression and possibly inhibit migration in the macrophage system.

To address the importance of proinflammatory cytokines on ENaC expression and their role in macrophage migration, we used the RAW 267.4 cell culture model, Boyden chamber style chemotactic migration, and ENaC-specific inhibitor amiloride (10, 31). Based on this model cell line, we identified a provocative relationship among proinflammatory cytokines, ENaC-dependent migration in macrophages, and their polarization toward proinflammatory phenotype.

METHODS

Macrophage Model

Mouse RAW 264.7 (American Type and Culture Collection TIB-71) cells were used as a macrophage model. Cells were cultured at 37°C, 5% CO₂ in Opti-MEM (Gibco) containing 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin in T75 flasks. Cells were passaged at 1:3 every 3 days. For treatment, cells were plated on T75 flasks and grown to ∼70% confluency.

Tissue Isolation

Lung samples were harvested from 3-mo-old wild-type male C57Bl/6J mice to assess quantitative PCR primer pair-probe specificity. Animals were anesthetized using isoflurane inhalation until respiration ceased. Tissue samples were snap frozen in liquid nitrogen and stored at −80°C until use. The procedure used in this study was reviewed and approved by the Institutional Animal Care and Use Committee of the University of Mississippi Medical Center.

RNA Extraction

Approximately 60 mg of tissue samples were minced on ice and then re-suspended in 500 µL of cold TRIzol. Total RNA was isolated using Quick-RNA Miniprep Kit (Zymo Research). To extract RNA from cultured cells, treated cells from T75 flasks were harvested after a brief rinse with sterile Dulbecco’s physiological buffered saline (DPBS), then gently scraped in 6 mL of ice-cold DPBS, transferred to collection tubes, and centrifuged at 3,000 rpm for 20 min, at 4°C. DPBS was removed and cell pellets were either immediately used for RNA isolation or stored at −80°C. Harvested cells were re-suspended in 300 µL of cold (∼4°C) TRIzol, and total RNA was isolated using Quick-RNA Microprep Kit (Zymo Research). RNA samples were either immediately used for reverse-transcription or stored at −80°C. RNA was reverse transcribed using the iScript Advanced cDNA Synthesis Kit (Bio-Rad, Hercules, CA). Reactions were incubated at 25°C for 5 min, 46°C for 20 min, and 95°C for 1 min.

Quantitative Polymerase Chain Reaction

Quantitative polymerase chain reaction (qPCR) was used to determine mouse SCNN1a, -b, and -g gene expression in RAW cells and assess the impact of ENaC inhibition on macrophage marker expression. Glyceraldehyde 3-phosphate dehydrogenase (Gapdh) gene expression was used as an internal control and to normalize gene expression. TaqMan primer pair-probe [FAM-MGB (SCNN1), FAM-MGB (macrophage markers) and VIC-MGB (Gapdh)] sets were obtained from Applied Biosciences (see Tables 1 and 2). PCR reactions, using Bio-Rad Supermix for Probes (No. dUTP, Cat. No. 1863024), were incubated at 95°C for 10 min, followed by 40–45 cycles of 94°C for 30 s, and 60°C for 1 min, using a ViiA 7 Real-Time PCR System (Applied Biosciences) with MicroAmp Fast Optical 96-Well Reaction Plates [R&D Systems, Inc. (Reference No. 4346906)]. SCNN1 transcript expression in the lung was tested using the reverse transcription reaction equivalent of 10, 100, and 1,000 ng of total RNA. To assess expression of macrophage markers, we used 100 ng of total RNA and TaqMan gene assays listed in Table 2. PCR reactions for SCNN1 and Gapdh with 1,000 ng were separated on 3% agarose gels to ensure that the amplicon of appropriate size was present using GeneRuler Low Range DNA Ladder (Thermo Scientific). All samples were run in triplicate from n = 2 trials.

Table 1.

TaqMan gene expression assay primer-probe pairs for ENaC subunits

Gene	Protein	RefSeq ID	Assay Name	Exon Boundary	Amplicon Length, bp
SCNN1a	αENaC	NM_011324.2	Mm00803386_m1	2–3	69
SCNN1b	βENaC	NM_001272023.1	Mm00441215_m1	4–5	83
SCNN1g	γENaC	NM_11326.2	Mm00441228_m1	6–7	63
Gapdh	GAPDH	NM_008084.2	Mm99999915_g1	3	107

Table 2.

TaqMan gene expression assay primer-probe pairs for proinflammatory macrophage (M1) polarization

Gene	Assay Name
CD68	Mm03047343_m1
CD86	Mm01344638_m1
TNF	Mm00443258_m1
NOS2	Mm00440502_m1
STAT1	Mm01257286_m1

PCR data were analyzed using QuantStudio Real-Time PCR Software v1.6.1. (Applied Biosystems). All amplification curves were examined for logarithmic amplification, baseline stability, noise spikes, and outliers were excluded from analysis on these bases. Cycle threshold (C_T) values were determined by QuantStudio software using the relative analysis setting. A C_T of 40 was considered indicative of no target gene expression. To quantify fold expression, we used the ΔΔC_T method, where fold expression = 2^−ΔΔCT.

Chemotactic Migration

Migratory responses were quantified using a modified Boyden chamber assay. The chamber consists of an upper and a lower chamber separated by a membrane with 8 µm pores. Two hundred fifty thousand cells in 100 µL Opti-MEM media supplemented with 0.4% FBS were loaded into the upper chamber. The lower chamber contains 500 µL of Opti-MEM with 10% FBS as chemoattractant. Cells were prepared by gently scraping from the T75 flask, resuspended in 0.4% FBS media, then counted. To block ENaCs, cells were preincubated with amiloride (0.1–10 µM) or benzamil (0.01–1 µM) for 30 min before and during migration. After the migration period, the upper surface of the migration chamber was gently scraped with a cotton swab to remove unmigrated cells and rinsed with PBS. Inserts were then immersed in 4% paraformaldehyde for 10 min, stained with hematoxylin for 10 min to label nuclei, and examined on an inverted Nikon Eclipse microscope using a ×20 objective. Migration was quantified as the average number of cells from seven fields of view/insert, n = 2–4 inserts per treatment, and normalized to control. Trials were repeated 2–3 times.

Macrophage Activation

Recombinant mouse TNFα and IFNγ (R&D Systems) were reconstituted with 0.1% BSA in sterile DPBS and stored in aliquots at 100 µg/mL that were used for macrophage stimulation/polarization. Macrophages were treated with IFNγ or TNFα at 0, 0.1, 1, and 10 ng/mL for 18 h (IFNγ_18h, TNFα_18h) or IFNγ 10 ng/mL for 48 h ± TNFα 10 ng/mL for the last 24 h (IFNγ_48h/TNFα_24h).

Phagocytosis Assay

To assess the nonopsonized phagocytic ability of RAW cells, we used the Vybrant Phagocytosis Assay (V-6694, Molecular Probes) according to manufacturers’ instructions. Briefly, RAW cells were plated on a 96-well plate and grown to 70% confluency, treated with 1 µM amiloride, 10 ng/mL IFNγ, or TNFα for 18 h. The following day, cells were exposed to the fluorescein isothiocyanate (FITC)-labeled bacterial bioparticles, then incubated for 2 h at 37°C. To control for treatment-induced cell death, live cells were labeled by adding Calcein Red-Orange at 5 µM (Molecular Probes Live/Dead Assay, according to manufacturers’ instructions) and incubated for another hour at 37°C. Nonphagocytosed FITC particles were quenched with trypan blue. Fluorescence signals were collected sequentially using fluorescence microplate reader at excitation/emission of ∼480/520 nm and ∼495/515 nm for phagocytosis and viability assays, respectively. Negative controls, consisting of cells without FITC-bioparticles or methanol mediated cell death, were used for phagocytosis and viability assays. Experimental results were conducted in three independent trials with n = 9 replicates each.

Stable Transfection

To inhibit expression of αENaC in RAW cells, cells were plated on a 6-well plate and transfected at 50% confluency with 4 µg of dominant-negative construct (1:0.7:0.5, DNA, Lipofectamine 3000; Enhancer; DNA) in 300 µL of Opti-MEM. We targeted αENaC because this subunit appeared to be the most abundantly expressed. The construct is formed by the first 112 amino acids of αENaC fused to the COOH-terminus of Enhanced Green Fluorescence Protein (EGFP_αENaC_W112X). Similar NH₂-terminal truncations of ENaC constructs downregulate endogenous subunit expression (1, 32, 33). Despite attempts to optimize transfection, efficiency remained at or below 10%. (The low transfection efficiency rendered siRNA and Crispr approaches undesirable). To enrich the population of cells expressing EGFP_αENaC_W112X, cells were cultured in the presence of Gentacin at 1:100 for 10 passages to generate a stable expressing cell line. EGFP expression was confirmed using fluorescence localization.

Western Blotting and Immunostaining

For Western blot analysis, cells were plated in T75 flasks and treated with IFNγ ± TNFα (10 ng/mL for up to 48 h), where appropriate. Media was removed, cells were rinsed in PBS, then gently scraped, and transferred to 15 mL conical tubes and centrifuged at 1,000 g for 2 min. PBS was removed and cells were lysed into KBO buffer (25 mM Na₃PO₄, 300 mM NaCl, 0.5% Triton X-100, 20 mM octyl glucoside, pH 7.4). Protein concentrations were quantified using Biorad Detergent Compatible Bradford Assay Kit. Protein samples (20 µg) were heated at 100°C for 10 min in 5× Sample Buffer plus DTT (Pierce) then separated on 4%–20% SDS-PAGE and transferred to nitrocellulose. Blots were blocked using Odyssey Blocking Buffer, incubated with mouse anti-β-Actin (loading controls, 1:10,000, Abcam) or mouse anti V5 (1:1,000, Sigma) and rabbit anti-αENaC (1:1,000, StressMarq) overnight at 4°C, rinsed then labeled with donkey anti-rabbit IR680 and donkey anti-mouse IR800 (1:10,000, 1 h). Membrane labeling was visualized using Odyssey Infrared Scanner.

To determine if proteosome and lysosomal degradation contributed to ENaC inhibition with cytokine activation with IFNγ_48h ± TNFα_24h (10 ng/mL each), inhibitors lactacystin (1 µM, Tocris) or MG132 (10 nM, Tocris), or chloroquine (30 µM, Tocris) or Lalistat 1 (100 nM, Tocris) were used. Before being harvested, cell cultures were microscopically examined to ensure viability (membrane blebbing, swelling, or shrinkage) and attachment.

For immunolabeling, cells were plated on 8-well chambered glass slides and treated overnight with TNFα or IFNγ at 10 ng/mL. Cells were rinsed with PBS, fixed in 4% paraformaldehyde for 10 min, and then rinsed in PBS. Cells were blocked in 5% donkey serum for 1 h then incubated with rabbit anti-αENaC (1:100, StressMarq) and human anti-mouse CD86 (1:100, Miltenyi, classically activated macrophage, M1, marker) conjugated to VioBright B515 (excitation/emission 488/525 nm) in 5% donkey serum overnight at 4°C. Samples were rinsed then labeled with donkey anti-rabbit Alexa 546 and anti-mouse Alexa 488 (1:200, 1 h) in 5% donkey serum, rinsed, cover-slipped, and air-dried. Samples were imaged using a Leica TCS SP8 confocal microscope, sequential scanning at 1,024 × 1,024 pixels, and ×63 objective under identical conditions for all samples. Regions of interest were drawn around cell boarders determined by overlaid signal in 25–30 cells from 2 to 3 images/group by a naive, blinded operator. Images were prepared for presentation in Photoshop and represent unmodified original scans. No primary antibody controls had no signal (not shown).

Where appropriate, light microscopy images were collected on a Nikon Eclipse 200 inverted microscope with ×20 objective using a CoolSnap Color camera and captured using Metamorph software.

Statistical Analysis

Data sets with similar standard deviations were analyzed using one-way or two-way ANOVA followed by Dunnett’s or Holm–Sidak post hoc test, adjusted for multiple comparisons, using Prism 9 software, and presented as means ± SE. Immunofluorescence data were analyzed using Kruskal–Wallis Analysis of Variance (following a positive Brown–Forsythe test) followed by a Dunn’s Post Hoc Test in Prism 9 software and presented as median ± quartiles in violin plots. Figure legends identify specific analyses applied. A value of P ≤ 0.05 is statistically significant. Result “trends” are also reported to avoid erroneous true/false conclusions based on bright-line rules. Select P values are provided to demonstrate confidence.

RESULTS

Amplification of SCNN1a, -b, and -g Primer Pair-Probe Sets for qPCR in Mouse Lung Tissue and Macrophages

As a positive control for SCNN1abg expression, a mouse lung homogenate was used (Fig. 1, A and B). Representative agarose gel electrophoresis images show a single amplicon for GAPDH and SCNN1 subunits (Fig. 1A, right) at the expected amplicon length (Table 1). Similar results were also obtained in kidney, liver, and brain homogenates (data not shown). SCNN1a and -b, but not -g, were detected in mouse macrophage RAW 264.7 cell line. Representative agarose gel electrophoresis images show a single amplicon at the expected length (Fig. 1B, right, Table 1).

Figure 1.

Epithelial Na⁺ channel (ENaC) is expressed in and required for migration in unstimulated macrophages. SCNN1 subunit message and GAPDH control expression in the mouse lung homogenates (A) and unstimulated macrophage RAW 264.7 cells (B) using quantitative polymerase chain reaction (qPCR). Left: group data cycle threshold (C_Th) decreases with increase in template content (10, 100, and 1,000 ng). Right: representative images of electrophoretically separated samples with 1,000 ng template. All PCR products are at the predicted amplicon length (Table 1). Similar results obtained in two independent trials. These findings suggest that SCNN1a (αENaC) and SCNN1b (βENaC), but not SCNN1g (γENaC), transcripts are expressed in unstimulated macrophages. All data are means ± SE and represent n = 3 samples. C: chemotactic migration of unstimulated macrophages requires ENaC. Migration of RAW 264.7 cells using a standard Boyden Chamber assay through 8-µm pores is inhibited by the ENaC inhibitors amiloride (0.1, 1.0, and 10 µM) and benzamil (0.01, 0.1, and 1 µM). Cells are treated with inhibitors for 30 min before and during 4-h migration period. The two lowest concentrations of amiloride and benzamil are highly specific for ENaC but not for other channels or transporters. Migration is normalized to control samples with each trial (n = 63 field of view taken from 9 cups/treatment in 3 trials). Data are analyzed using one-way ANOVA (P value provided on graph) followed by Dunnett’s Post Hoc test. *Significantly different from control at P < 0.0001. D: phagocytosis is not altered by ENaC inhibition with 1 µM amiloride (n = 8–27 wells from n = 3 trials). Data were analyzed using a one-way ANOVA (P value provided on graph), followed by a Holm–Sidak post hoc test. P values of specific comparisons are shown.

Role of ENaC in Migration and Phagocytosis of Unstimulated Macrophages

The importance of ENaC channels in chemotactic migration of unstimulated macrophages to a broad-spectrum stimulus of 10% FBS was determined using amiloride and benzamil, shown in Fig. 1C. Migration in RAW 264.7 cells (Fig. 1C) was inhibited by amiloride (0.1–10 µM, left) and benzamil (0.01–1 µM, right). Both amiloride and benzamil are highly selective for ENaC. The lowest concentration of amiloride and benzamil inhibited chemotactic migration to 65.6 ± 5.7% and 54.6 ± 5.8% of control in RAW 264.7 cells, respectively. Higher concentrations of amiloride (10 µM) and benzamil (1 µM) inhibited chemotactic migration by 45.6 ± 4.2% and 55.2 ± 5.0% in RAW 264.7 cells. Our finding that amiloride and benzamil inhibit chemotactic migration at low concentrations indicates ENaC signaling contributes to nearly 50% of migratory responses in unstimulated macrophages.

Phagocytosis is another major function of macrophages. Since cytokines can prune select macrophage cell populations, we normalized phagocytosis to live cell number. We found no effect of amiloride, TNFα, and IFNγ on phagocytosis when normalized to live signal (Fig. 1D). We found that IFNγ increased the phagocytosis of opsonized macrophages, which is consistent with the previous data that IFNγ enhances opsonized macrophage phagocytosis (Fig. 1D) (34). Opsonization is a process where antibody components bind to foreign materials to increase susceptibility to phagocytosis.

EGFP fluorescence, a marker of expression of the truncated αENaC construct (Fig. 2A), is shown in Fig. 2B. EGFP expression was heterogeneous (Fig. 2B). Control macrophages exposed to Gentacin died within 5 days, suggesting that all cells likely express the construct, hence resistance to toxicity, however EGFP levels may be below detection limit. Macrophages expressing EGFP_αENaC_W112X significantly reduced the migratory capacity compared with untransfected control cells. Moreover, amiloride-sensitive migration was abolished in these cells (Fig. 2C).

Figure 2.

Alpha epithelial Na⁺ channel (αENaC) silencing inhibits macrophage migration. A: dominant-negative αENaC construct; EGFP fused to the first 112 amino acids of αENaC. B: EGFP fluorescence in EGFP_αENaC_W112X transfected, but not untransfected, macrophages grown for 10 passages in the presence of Gentacin (1:100). Untransfected macrophages exposed to Gentacin die within 5 days. C: αENaC silencing suppresses migration and abolishes amiloride-sensitive migration of macrophages. Data are means ± SE and represent 14 fields of view taken from 2 cups/treatment in two trials. Data are analyzed by two-way ANOVA, the P values of the factors and their interaction are provided on the graph. Differences among the groups were assessed by the Holm–Sidak post hoc test; P values are shown.

Macrophage Activation Protocol Attenuates Migration

Cells were activated by 48-h exposure to IFNγ ± TNFα during the last 24 h. Expectedly, the combined cytokine exposure was associated with some cell loss and morphological changes, marked by flattening, enlargement, elongation of some cells, and protrusions (Fig. 3A). The 48-h cytokine treatment (IFNγ_48h ± TNFα_24h) inhibited chemotactic migration (Fig. 3B). In contrast, limiting cytokine exposure to the migration period (4 h) stimulated chemotactic migration (last group of Fig. 3B). To determine if ENaC contributed to this stimulated response, we examined the amiloride (1 µM) sensitivity (Fig. 3C) and found the amiloride-sensitive migratory response is similar between control and IFNγ_4h + TNFα_4h cells (Fig. 3D). This latter finding suggests that the enhanced migration response to a brief cytokine exposure does not require ENaC.

Figure 3.

Activation of mouse macrophages into “classical” macrophage-like cells using chronic exposure to interferon γ (IFNγ) and tumor necrosis factor α (TNFα) inhibits migration. Cells are treated with IFNγ (IFNγ_48h, 10 ng/mL) alone for 48 h or with IFNγ_48h plus TNFα (10 ng/mL) during the last 24 h (IFNγ_48h/TNFα_24h), a protocol is used for activation of “classical” macrophages. A: light microscopy images of macrophages following 48 h of IFNγ ± TNFα treatment show morphological changes consistent with “classical” activation, such as increase in size, flattening, and elongation, and formation of projections. B: migration responses presented as % Control to cytokine treatment for 48 h prior to or during (4 h) migration, n = 3–7. The first four columns represent the migration responses to a cytokine exposure for 48 h before migration. The last column represents the effects of exposure to cytokines only during migration. In contrast to the inhibitory effect of a prolonged exposure, limiting cytokine exposure to the migration period stimulates migration. The importance of ENaC channels in the brief response is addressed in (C and D). The amiloride-sensitive component of chemotactic migration in mouse macrophages is identical in unstimulated macrophages during the 4-h migration period. C: chemotactic migration (% Control) ± amiloride (1 µM) and/or IFNγ + TNFα_4h (10 ng/mL) during migration, n = 4–7 migration inserts from 3 trials. Each n represents the mean of 7 fields of view per insert. D: the amiloride-sensitive component of migration in control and IFNγ + TNFα_4h-stimulated cells, n = 4. These data show that the ENaC-dependent component of brief IFNγ + TNFα_4h stimulation is not different from unstimulated control cells. Data in B and C were analyzed using a one-way ANOVA and were followed up with a Holm–Sidak post hoc test. Data in D were analyzed using a two-tail, independent t test. All P values are shown on graph to demonstrate confidence.

Prolonged Cytokine Exposure Inhibits αENaC Protein Expression

To determine if a downregulation of αENaC protein might account for the response, we examined whole cell αENaC expression and found that αENaC (∼85 kDa band) was inhibited by 50%–70% after IFNγ_48h or IFNγ_48hr + TNFα_24hr treatment (Fig. 4, A and B). Similar results were found with βENaC (data not shown). To determine the role of proteasomal or lysosomal degradation to αENaC downregulation, we examined the impact of two proteasome (lactacystin and MG132, Fig. 4, A and B) and lysosome (chorloquine and lalistat 1, Fig. 4, C and D) inhibitors with IFNγ ± TNFα treatment. Proteosome inhibition with lactacystin and MG132 had a marginal effect on preventing αENaC protein degradation, which was only observable in log transformed, absolute αENaC signal (Fig. 4B).

Figure 4.

Proteasome and lysosome mechanisms play a minor role in cytokine-mediated inhibition of alpha epithelial Na⁺ channel (αENaC). A and C: representative Western blots of αENaC (∼85 kDa, top) and β-actin (bottom) in macrophages activated by interferon γ (IFNγ)_48hr ± tumor necrosis factor α (TNFα)_24h (10 ng/mL, each) without and with proteasome (A) and lysosome (C) inhibition. B and D: group data showing mean αENaC signal normalized to β-actin. Neither proteasome (A and B) nor lysosome (C and D) inhibition had a statistically significant impact protecting αENaC protein expression when normalized. Since protein lysate samples were quantitated and equivalently loaded (20 µg/well), we also examined log transformed αENaC expression, proteasome inhibition had a minor protective effect. Sample size is n = 3–6 independent samples from 2 separate trials. Data were analyzed using a one-way ANOVA followed by Holm–Sidak post hoc test. ANOVA and post hoc P values are shown on the graph. MM, molecular mass.

We also examined αENaC protein expression in individual cells using immunolabeling (Fig. 5). As shown in representative images in Fig. 5A and group data violin plots in Fig. 5B, the effects of TNFα_18h and IFNγ_18h on αENaC were heterogeneous. The images show that αENaC expression is reduced in many TNFα_18h- and the majority of IFNγ_18h-treated cells. The violin plots and histograms support this visual representation and show the shift in distribution of cells with low αENaC labeling. In the TNFα_18hr-treated cells, the lower quartile is associated with lower αENaC expression; however, the median value did not reach statistical significance. In the IFNγ_18hr-treated cells, the median was significantly decreased by the downward shifted distribution. IFNγ_18h, but not TNFα_18h, robustly upregulated CD86, a marker of proinflammatory (M1) macrophages (Fig. 5, A and B), giving the cells a large and flat appearance. Although both cytokines contribute to M1 polarization, they do not have the same effectiveness on expression of M1 marker CD86.

Figure 5.

Activation of macrophages with proinflammatory cytokines tumor necrosis factor α (TNFα) and interferon γ (IFNγ) inhibit alpha epithelial Na⁺ channels (αENaC) protein expression. A: representative images of immunolabeling in macrophages stimulated with TNFα or IFNγ (18 h, 10 ng/mL) then labeled with anti-CD86 (green) and anti-αENaC (red) in untreated control (top), TNFα-treated (middle), and IFNγ-treated samples (bottom). In unstimulated control macrophages, CD86 labeling is low and αENaC is modest and homogeneous. After polarization to macrophage-like cells with TNFα, CD86 labeling remains low, however, αENaC labeling is heterogeneous, with a few well expressing cells, but with many low expressing cells. Polarization with IFNγ elicited a robust stimulation of CD86, cell enlargement, and a uniform reduction in αENaC labeling. B: violin plots show that CD86 signal is elevated in IFNγ-treated cells. The histogram plot also shows that the distribution of cell signal intensity shifts to the right. C: violin plots of αENaC show increased variability with TNFα and IFNγ treatment, most notably, more cells with reduced αENaC expression, and a significant drop in median αENaC expression in IFNγ-treated cells. The left-ward shift of the IFNγ histogram curve is consistent with the decreased fluorescence intensity. Violin plots are a method of describing data characteristics without plotting large numbers of individual data points. Data are characterized as minimum, maximum, median, and 25/75% confidence intervals. Data were analyzed using a Kruskal–Wallis analysis of variance for data sets with different standard deviations (P < 0.0001 for both), followed by a Dunn’s post hoc test (P values provided on graph). Data represent median ± quartiles from n = 75 individual cells from three fields of view. RFU, relative fluorescence units.

ENaC-Dependent Migration Is Abolished after IFNγ or TNFα Exposure

To determine if there are remaining functional ENaC channels following proinflammatory cytokine exposure, we examined the effect of IFNγ or TNFα at 10 ng/mL independently after an 18-h exposure on amiloride-sensitive migration responses. As shown in Fig. 6, amiloride inhibited the chemotactic migration of unstimulated cells. In contrast, amiloride had no additional inhibitor effect on migration in IFNγ_18h- or TNFα_18h-polarized cells. These findings suggest that IFNγ/TNFα inhibits migration via suppression of functional ENaC protein expression.

Figure 6.

Epithelial Na⁺ channel (ENaC)-dependent migration is abolished in interferon γ (IFNγ)_18hr- or tumor necrosis factor α (TNFα)_18hr-activated macrophages. Data showing migration responses in unstimulated monocyte-like and IFNγ or TNFα-stimulated cells ± amiloride (1 µM, 18 h), grouped by treatment (unstimulated, stimulated with IFNγ or TNFα at 10 ng/mL). The ENaC-dependent component of migration is abolished following macrophage activation by IFNγ or TNFα. Data represent means ± SE from n = 20–35 fields of view, from 5 to 8 cups/treatment, in three trials. Data were analyzed using a two-way ANOVA (P values for main effects of amiloride, cytokines, and their interaction are provided on the graph), followed by the Holm–Sidak post hoc test. P values of post hoc comparisons within treatment groups are provided on the graph.

Does ENaC-Inhibition Mediate Polarization to Classically Activated Macrophages?

Although proinflammatory cytokines inhibit ENaC protein expression and induce polarization to M1 macrophages, our findings do not prove a causal relationship between ENaC inhibition and polarization. As shown in Fig. 7A, macrophages exhibit subtle cell flattening and more projections after an overnight incubation in amiloride. Using qPCR, we found that ENaC inhibition induced an upregulation of classically activated, proinflammatory, M1 macrophage markers including surface markers (CD86), cytokine (NOS2), and transcription factor (STAT1) (Fig. 7B). CD68, a monocyte-derived cell marker, and TNF message were unchanged. These findings are consistent with a role for ENaC inhibition in macrophage polarization.

Figure 7.

Epithelial Na⁺ channel (ENaC) inhibition potentiates polarization to M1-like macrophages. A: light microscope images of cells after 48 h incubation with amiloride (1 µM). Modest cell flattening and formation of projections are visible after amiloride treatment. B: fold changes in expression of selected macrophage markers. CD68 is a monocyte marker. Of the M1 macrophage markers CD86, NOS2, TNF, and STAT1, message for all but TNF are upregulated in amiloride-treated cells. Data represent n = 9–12 samples, from n = 3 to 4 trials and are presented in violin plots as median and quartiles. Data were analyzed using unpaired t tests. Specific P values are provided on graphs.

DISCUSSION

The monocyte-macrophage system is integral in initiation of an inflammatory response. Circulating monocytes are attracted by chemoattractants [(cytokines, chemokines, and other damage-associated molecular patterns (DAMPs) molecules] released by damaged tissue where they migrate and polarize into macrophages. Proinflammatory cytokines, such as TNFα and IFNγ, polarize macrophages into classically activated M1 type macrophages that phagocytose damaged or necrotic tissues and release cytokines to recruit other inflammatory cells to the site of tissue damage and propagate the immune response. Migration is a very complex response that utilizes many chemoattractants, adhesion, and cytoskeletal molecules (35). ENaC channels contribute to migration in numerous cell types (1, 27–30, 36); however, their role in macrophage migration is unknown. Here, we used an in vitro approach to establish a role for ENaC in macrophage migration using the cell line RAW 264.7 and modified Boyden chamber chemotaxis assay to a broad-spectrum attractant (FBS). Our findings show that ENaC plays a major role in migration of unstimulated macrophages. Polarization with IFNγ and/or TNFα inhibits 1) chemotactic migratory capacity to FBS, 2) αENaC protein expression, and 3) ENaC-dependent chemotactic migration. Unexpectedly, ENaC inhibition increased the expression of select M1 macrophage markers. Thus, our findings suggest that ENaC may play a previously unidentified role in macrophage migration and polarization.

RAW 264.7 Cell Model

We chose an in vitro RAW 264.7 cell model as a starting point to address the role of ENaC channels in migration and regulation of macrophages by proinflammatory cytokines. The cell line is a characterized macrophage model and eliminates complications of activation and isolation of endogenous monocytes and macrophages in a mouse model. We found that RAW 264.7 cells respond to a protocol for macrophage activation (10 ng/mL of IFNγ for 48 h plus 10 ng/mL of TNFα for the last 24 h) and polarization and display characteristic morphological changes (21, 37–40). In the unstimulated state, these cells take on an appearance comparable with monocytes, with a round shape, limited projections, and lack expression of M1 macrophage markers. When the cells are stimulated by proinflammatory cytokines, their appearance changes, sharing macrophage features of an enlarged, flattened cell body with more projections and expression of macrophage markers.

Evidence of the Expression of ENaC and Closely Related Family Member in Inflammatory Cells

Several lines of evidence suggest that ENaC subunits are likely expressed in monocyte-macrophage cells and contribute to their function. First, the Human Protein Atlas database submissions identifies expression of ENaC subunit transcripts, and their closely related family members, Acid Sensing Ion Channel (ASIC) subunits, in select populations of PBMCs and resident tissue macrophages from multiple human healthy populations (22–26). Second, experimental studies support the expression and function of ENaC/ASIC subunits in select inflammatory cell populations. ENaC plays a role in salt sensing and amplifying inflammatory responses in dendritic cells, antigen presenting myeloid cells (17). In addition to the Human Protein Atlas Database, multiple reports indicate that ASIC proteins are expressed in inflammatory cells (18, 41, 42). ASIC subunits are expressed in bone marrow-derived dendritic cells, RAW 264.7 cells, and resident central nervous system macrophages (microglia) (18, 41, 42). At least one study suggests that ASIC1a, the predominant ASIC isoform in microglia, is upregulated by the potent inflammatory modulator lipopolysaccharide (LPS) (41). Moreover, the ASIC1a homomeric channel blocker psalmotoxin inhibits scratch-assay migration responses in unstimulated microglia and RAW 264.7 cells (18, 41, 42). The contribution of ASIC channels to scratch assay migration responses following activation, by LPS or other cytokines, was not examined. Thus, how ASICs contribute to macrophage migration under inflammatory conditions and might inform a role for ENaC in macrophage function is unknown. Third, data presented here suggest that select ENaC transcripts and proteins are expressed in macrophage-like cells, and αENaC and βENaC proteins are downregulated in the polarization process by IFNγ and/or TNFα. Taken together, these findings suggest that ENaC, and their closely related ASIC channels, may have a previously under-appreciated role in immune cell function. A limitation of the current and previous studies is the lack of electrophysiologic evidence of ENaC. Since ENaC, ASIC, and ENaC-ASIC hybrid channels have different gating mechanisms and current characteristics, future studies are necessary to determine the molecular identity of ENaC-ASIC channels in macrophages and whether polarization changes the nature of the channels.

Role of ENaC in Migration of Macrophage-like RAW 264.7 Cells

A primary function of the macrophage system is migration toward signals from tissue injury. Several in vitro assays quantifying cellular migration have been developed including wound-healing/scratch assay (cells migrate into a scratch wound), linear chemotaxis (cells migrate on the horizontal axis toward a concentration gradient), and Boyden Chamber assay (cells migrate on the vertical axis through a porous membrane) (40, 43). Previous studies demonstrate a role for ENaC and ASIC proteins in wound healing and/or chemotactic migration of vascular smooth muscle, astrocyte/glioblastoma, keratinocyte, and trophoblastic cells (1, 7, 27, 28, 30, 36, 44, 45) and a role for ASIC1a homomeric channels in microglia and RAW 264.7 cells (18, 41). However, a role for ENaC channels has not been addressed.

Based on our studies, we make two novel conclusions on the importance of ENaC in macrophage chemotactic migration. First, roughly half of the chemotactic migration response of unstimulated macrophage cells to the broad-spectrum stimulus FBS is ENaC dependent. Second, IFNγ- and TNFα-induced polarization inhibits migration responsiveness to FBS. The remaining migration response is unaffected by amiloride, suggesting that ENaC-dependent migration is abolished following polarization with IFNγ and TNFα. We suspect that inhibition of ENaC, as assessed by αENaC protein expression (Figs. 4 and 5) and amiloride sensitivity (Fig. 3B and Fig. 6), likely accounts for the polarization-induced inhibition of migration. However, we cannot rule out the possibility that reduced macrophage mobility or sensitivity to attractant substance(s) in FBS with polarization, may also contribute. So, what might be the advantage of this relationship between polarization and reduced migratory ability? This might reflect an ideal paradigm: an initial boost in migration may enhance naive macrophage access to the site of injury followed by polarization into a phagocytic macrophage. A less mobile phenotype would be beneficial by limiting phagocytic activity to the injured area.

Amiloride Specificity

To probe for a role of ENaC in macrophage migration, we used low concentrations of amiloride and its analog benzamil to selectively inhibit ENaC channels [IC₅₀ for ENaC is between 0.1–5 and 0.01–5 µM, respectively (10, 31)]. The specificity of amiloride for ENaC at the concentrations used here are essential to our conclusions regarding the role of ENaC in macrophage function. There are three main lines of evidence that amiloride’s effect are primarily signaled through ENaC. The first line of evidence is the specificity of amiloride for ENaC relative to other channels, transporters, and the urokinase-type plasminogen activator (uPA). The 1 µM concentration of amiloride used in these studies is 10-fold greater than the IC₅₀ of amiloride for ENaC and 7- to ∼100-fold less than the IC₅₀ for other channels, transporters (∼100 mM, Na⁺/H⁺ exchanger) or uPA (7–80 µM) (31, 46) Furthermore, studies demonstrating an effect of amiloride on transcriptional splicing events in cancer cells were conducted using 200–500 µM (47–49). The potential for amiloride to promote transcriptional changes based on the use of 200-fold higher concentrations than used here (49). Thus, it is unlikely that off-target effects of amiloride account for the dramatic migration inhibition shown here with 1 µM.

Second, we can rule out significant impacts of amiloride on uPA based on the known importance of uPA in polarization, and phagocytosis. uPA activation induces polarization and phagocytosis (50, 51). If 1 µM amiloride inhibited uPA in our model, we would expect that amiloride should inhibit polarization and phagocytosis. We observed that amiloride facilitated polarization but did not alter phagocytosis (Fig. 7).

Third, we can also safely rule out a significant role of amiloride-mediated uPA inhibition on migration. Our findings in Fig. 1C show that there is no substantial difference in inhibitory effect of 1 µM versus 10 µM amiloride. If amiloride’s inhibition of uPA (IC₅₀ as low as 7 µM) had an impact on migration, the 10 µM amiloride should reveal its impact. However, we observed no difference in migration with 1 µM versus 10 µM. We also used benzamil, a more selective ENaC inhibitor (ENaC IC₅₀ = 10 nM) and found similar migration inhibition at IC₅₀ concentrations. In addition, our findings in Fig. 2 show the importance of αENaC specifically, in migration. Taken together, these findings do not support a role for amiloride inhibition of uPA, other channels/transporters, or transcription events on our results.

Does αβENaC Suppression by IFNγ and TNFα account for the Loss of Macrophage Migration?

To link the IFNγ- and TNFα-induced inhibition of αβENaC expression to the loss of migration, we examined the amiloride sensitivity of the remaining migration responses and found that they were insensitive to amiloride (Fig. 6). This finding suggests that the activated macrophages do not have functional ENaC channels mediating migration. The directional change in ENaC protein expression was consistent with the loss of migration responsiveness. To examine other potential mechanisms for ENaC protein inhibition, we examined proteasomal and lysosomal mechanisms.

Proteasome and lysosome activation are two common mechanisms for silencing protein expression. Proteasomal degradation via ubiquitination, is an established mechanism for ENaC downregulation in epithelial cells. Ubiquitination is a common cytokine signaling mechanism in inflammatory cells (52–54). However, we found proteasomal inhibition with MG132 or lactacystin or lysosomal activity using chloroquine and lalistat 1 and found a similar outcome on αENaC protein; no impact on αENaC normalized to β-actin. Log transformation of absolute αENaC signals had a minor protective impact. We rationalized using the absolute ENaC signal because the blot wells were equally loaded with 20 µg protein and we considered that cytokines might alter expression of many different proteins, including “housekeepers.” In addition, we rationalized transforming the data because the expression values in treated samples are skewed toward much lower values. Analyzed in this manner, proteasomal inhibition with MG132 or lactacystin and lysosomal inhibition with lalistat 1, to a lesser extent, had a minor effect on αENaC levels. The physiological relevance of this difference is not clear. It is possible that the proteasomal and lysosomal inhibition was insufficient to protect against αENaC loss; however, higher concentrations of the agents were cytotoxic in preliminary experiments.

Is ENaC Required for Macrophage Polarization?

In preliminary studies, we observed subtle changes in the cell morphology following amiloride coculture experiments. This led us to question whether ENaC activity might be required for macrophage polarization. Blocking ENaC channels with amiloride overnight elicited an upregulation of 1) phagocytosis, a phenotype enhanced in polarized macrophages, and 2) several markers associated with macrophage activation, suggesting that ENaC inhibition facilitates macrophage polarization.

Perspectives and Significance

In summary, our study demonstrates that ENaC plays a critical role in the polarization and migration properties of macrophages in vitro. Macrophages initiate an immune response leading to a larger, coordinated response and thus may represent an ideal target for immune modulation in chronic diseases. Future in vivo studies are required to determine the role of ENaC in macrophages derived from bone marrow monocytes and resident tissue macrophages and their role in diseases associated with immune dysregulation, such as preeclampsia, obesity, and metabolic disease.

GRANTS

Research reported in this publication is supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health under Award Numbers R01HL136684 (to H. A. Drummond), and P01HL051971 (to H. A. Drummond), and the National Institute of General Medical Sciences of the National Institutes of Health under Award Numbers P20GM104357 (awarded to J. E. Hall, provides support for Drummond, Granger), P20GM121334 (awarded to J. Reckelhoff, provides support for Drummond), and 5U54GM115428 (Granger).

DISCLAIMERS

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

H.A.D. conceived and designed research; Z.N., E.H., N.P., A.B.F., R.W., K.P., and X.B. performed experiments; Z.N., N.P., A.B.F., R.W., X.B., and H.A.D. analyzed data; Z.N. and H.A.D. interpreted results of experiments; H.A.D. prepared figures; Z.N. drafted manuscript; Z.N., J.P.G., M.J.R., and H.A.D. edited and revised manuscript; Z.N. and H.A.D. approved final version of manuscript.