Sodium Channel Nav1.3 Is Expressed by Polymorphonuclear Neutrophils during Mouse Heart and Kidney Ischemia InVivo and Regulates Adhesion, Transmigration, and Chemotaxis of Human and Mouse Neutrophils In Vitro

All nine known pore forming α-subunits of voltage-gated sodium channels are regarded to dictate action potential generation in excitable cells.¹  Accordingly, inhibition or modulation of excitability by local anesthetics seems to be primarily due to an inhibition of voltage-gated sodium channels.²  While the expression pattern of most α-subunits suggests that they have rather specialized tissue-specific functions,¹  recent studies show that so-called “neuronal” or “cardiac” α-subunits are expressed in a variety of tissues and that some α-subunits even seem to be expressed in nonexcitable cells including tumor cells, glia, and immune cells.³  While nonexcitable cells obviously do not generate action potentials, voltage-gated sodium channel α-subunits were reported to perform noncanonical roles and thus regulate diverse cellular functions including cell motility, migration, invasiveness, and phagocytosis, to name some.³  In macrophages, the cardiac α-subunit Nav1.5 regulates endosomal acidification, phagocytosis, and degradation of myelin in multiple sclerosis lesions and possibly acts as a pathogen sensor.^4–6  Moreover, Nav1.5 regulates migration and proliferation of astrocytes after mechanical injury through mechanisms involving a [Ca²⁺]_i transient.⁷  A more recent study suggested that inhibition of several voltage-gated sodium channel α-subunits expressed in macrophages and/or monocytes by phenytoin improves cardiac function after cardiac ischemia and reperfusion in adult rats.⁸  The skeletal muscle α-subunit Nav1.4 was reported to play a crucial role for positive selection of cluster of differentiation (CD) 4-positive T cells,⁹  and the sensory neuronal α-subunit Nav1.7 regulates migration and cytokine responses of dendritic cells¹⁰  and migration of endothelial cells.¹¹  These noncanonical roles of voltage-gated sodium channels in different immune cells indicate that voltage-gated sodium channel inhibitors should have antiinflammatory properties. Indeed, lidocaine and other local anesthetics were shown to induce antiinflammatory and antiapoptotic effects in both rodent and human studies.^12–15  However, this interesting property of primarily lidocaine is usually attributed to a modulation of polymorphonuclear neutrophil function including a reduction of activation, priming, and recruitment.^12,14,16  While several molecular mechanisms have been suggested to account for these lidocaine-induced effects, they are commonly stated to be voltage-gated sodium channel–independent. This statement is probably given mainly because effects on polymorphonuclear neutrophils occur at low micromolar (<10 µM) concentrations of lidocaine, which are not sufficient to robustly inhibit sodium currents generated by voltage-gated sodium channels. As inhibition of voltage-gated sodium channels by local anesthetics obeys a strong state-dependency, however, very low concentrations of lidocaine can inhibit sodium currents at depolarized membrane potentials.² 

Alerted to the increasing number of studies demonstrating diverse noncanonical roles of voltage-gated sodium channel α-subunits in immune cells other than polymorphonuclear neutrophils, we hypothesized that voltage-gated sodium channels are relevant for polymorphonuclear neutrophil function and for lidocaine-induced effects on polymorphonuclear neutrophils. Employing a variety of in vitro and in vivo assays on mouse and human polymorphonuclear neutrophils, we found that the neuronal α-subunit Nav1.3 is expressed in a fraction of polymorphonuclear neutrophils both in vitro and in vivo. Furthermore, pharmacologic inhibition of Nav1.3 strongly abbreviated attachment, transmigration, and chemotaxis of polymorphonuclear neutrophils. We therefore suggest that Nav1.3 might be a relevant target for antiinflammatory effects of voltage-gated sodium channel blocking agents like lidocaine.

This study was approved by the Institutional Review Board (Oldenburg, Germany) and performed in accordance with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (Bethesda, Maryland). For myocardial ischemia (MI) with reperfusion injury, the left coronary artery of 8- to 10-week-old C57BL/6 mice was ligated for 30 min followed by 24 h and 72 h of reperfusion. An average infarct size of the area at risk of 47.5% was confirmed by 2,3,5-triphenyl tetrazolium chloride/Coomassie (Sigma, Germany) staining as shown previously.¹⁷  For histologic analysis, hearts were harvested and cryoembedded at 24 h or 72 h after surgery. For kidney ischemia, 12- to 14-week-old male C57BL/6 mice were anesthetized with isoflurane (Baxter, Germany; 3% for induction and 1.5% for maintenance), and butorphanol (Vetoquinol, Germany) 1 mg/kg was given subcutaneous for analgetic treatment. The left renal pedicle was clamped for 45 min to induce unilateral ischemia reperfusion injury. Mice were euthanized 24 h after ischemia reperfusion injury by total body perfusion with ice-cold phosphate-buffered saline (Lonza, Germany) via the left ventricle in deep general anesthesia resulting in circulatory arrest. One part of the tissue was shock frozen for immunohistochemistry. The rest was rinsed in phosphate-buffered saline and used for leukocyte isolation.

Mouse polymorphonuclear neutrophils were freshly isolated from bone marrow using a modified protocol described by Zhang et al.¹⁸  C57BL/6 mice were euthanized by cervical dislocation. Femur and tibia were separated, soft tissues were removed with a scalpel, and bone marrow cells were flushed out with approximately 5 ml of Hank’s balanced salt solution (Lonza) supplemented with 0.1% bovine serum albumin (GE Healthcare, Germany) per bone using a 20-gauge needle. Cell suspension was dispersed with a 18-gauge needle, filtered through cell strainers (40 µm, Becton Dickinson, USA), and centrifuged for 5 min at 2,200 rpm, and erythrocytes were removed by hypotonic lysis. After centrifugation, cells were resuspended in Hank’s balanced salt solution or medium, depending on the type of assay, and cell viability determined by trypan blue exclusion.

Human polymorphonuclear neutrophils were isolated from peripheral blood obtained from healthy volunteers in accordance with the Hannover Medical School Institutional Review Board, Hannover, Germany. Briefly, EDTA-anticoagulated blood was layered on top of a 1.077 g/ml Biocoll (Biochrom, Germany) density gradient, and after centrifugation at 800g for 20 min, the polymorphonuclear neutrophil-containing band was collected. Erythrocytes were removed by lysis with cold double-distilled H₂O. polymorphonuclear neutrophils were resuspended in Hank’s balanced salt solution, cell viability determined by trypan blue exclusion and a purity of greater than 95% verified by May Grünwald/Giemsa (Sigma) staining of cytospins.

Expression of voltage-gated sodium channels was analyzed by quantitative reverse transcription polymerase chain reaction as described previously.¹⁷  Briefly, total RNA from polymorphonuclear neutrophils attached to cell culture plastic for 24 h, 4 h, or 15 min with or without tumor necrosis factor-α (R&D Systems, Germany) stimulation as indicated was extracted with Tritidy reagent (Applichem, Germany), and 500 ng RNA was used for complementary DNA synthesis (Reverse Transcription Core Kit, Eurogentec, Germany). Primers for the different voltage-gated sodium channel isoforms are listed in table S1 (Supplemental Digital Content, https://links.lww.com/ALN/B636). Polymerase chain reaction was performed on a real-time polymerase chain reaction cycler (Rotorgene 3000, Corbett Life Science, Germany). Signals were generated by SYBR green incorporation (SensiFAST SYBR No-ROX, Bioline, Germany) into the amplified DNA (40 cycles), and visualized on a 2% agarose gel using ultraviolet light in a Bio-Vision gel documentation system (Vilbert Lourmat, Germany).

Immunohistochemistry of cryoembedded myocardial and ischemic kidney tissues was performed as described before.¹⁹  Briefly, 5 µm cryo sections of myocardial tissue 24 h and 72 h after MI with reperfusion injury and 24 h after 45 min of kidney clipping were stained with primary antibodies against F4/80 (1:100, Bio-Rad AbD Serotec, Germany) for detection of macrophages or with GR-1 (1:100, Bio-Rad AbD Serotec) or Ly6B (clone 1A8, 1:100, BioLegend GmbH, Germany) for identification of polymorphonuclear neutrophils and with Nav1.3 (1:500, ASC004 and ASC023 Alomone Labs, Israel), all diluted in phosphate-buffered saline supplemented with 0.1% bovine serum albumin and 5% normal donkey serum at 4°C overnight. Secondary antibodies coupled to Cy3 or Cy5 (Jackson ImmunoResearch, United Kingdom) were diluted 1:100 in phosphate-buffered saline and incubated for 2 h at room temperature. Cell nuclei were counterstained with 1 µg/ml 4′,6-Diamidine-2′-phenylindole dihydrochloride (Sigma) and slides were mounted with Fluorescent Mounting Medium (Biozol, Germany). After allowing mounted slides to dry completely, fluorescence was analyzed on an inverted fluorescence microscope (IX81, Olympus, Germany) using Q-Capture Pro7 software (QImaging, Canada).

Specificity of Nav1.3 antibody was assessed by Western blot analysis as described previously.¹⁷  Briefly, protein lysates were prepared in radio immunoprecipitation assay buffer including protease inhibitors (complete mini, no. 11836170001, Roche Diagnostics, Germany) from protein lysates of HEK293 cells transfected with different sodium channels. The protein amount was assessed by the Pierce BCA protein assay kit (Thermo Scientific, Germany). Proteins were separated by electrophoresis, transferred to polyvinylidene difluoride membranes (Sigma), and identified by Western blot using primary antibody Nav1.3 (Alomone Labs) diluted 1:100. Horseradish peroxidase–coupled rabbit IgGs were used as secondary antibodies. Membrane was reprobed with anti–β-actin (no. 4967, Cell Signaling, Germany) to correct for protein loading.

Adhesion assays were performed as described.²⁰  Briefly, immortalized murine endothelioma cells (f.End5) or human umbilical vein endothelial cells (Lonza) were grown to confluency in collagen-coated six-well dishes in Dulbecco’s Modified Eagle’s Medium (Lonza) supplemented with 10% fetal calf serum (f.End5) or endothelial cell growth medium-2 Bullet-kit (Lonza) with 5% fetal calf serum (human umbilical vein endothelial cells), starved for 8 h in Dulbecco’s Modified Eagle’s Medium or endothelial cell growth medium-2 Bullet-kit without fetal calf serum, and stimulated with 10 ng/ml mouse or human tumor necrosis factor-α (R&D Systems) for 12 h at 37°C. Isolated mouse bone marrow polymorphonuclear neutrophils or human whole blood polymorphonuclear neutrophils were labeled with CellTracker Green (Molecular Probes, USA) and 5 × 10⁵ cells in 1 ml Binding Buffer (Hank’s balanced salt solution supplemented with 1% bovine serum albumin, 2 mM Ca²⁺ and 2 mM Mg²⁺) were allowed to adhere to activated endothelial cells in six-well cell culture plates for 1 h on a tilting table in the presence of 100 nM tetrodotoxin (Biotrend, Germany), 1 µM ICA121431, and 0.5 µM Pterinotoxin 2 (Alomone Labs) to block Nav1.3 channel activity as indicated. Afterward, nonadherent cells were removed by extensive washing with Hank’s balanced salt solution. Each condition was performed in triplicate, and eight randomly selected high-power fields per well were documented with ×200 magnification on an inverted fluorescence microscope (IX81, Olympus) using Q-Capture Pro7 software (QImaging). The number of adherent cells was determined using ImageJ version 1.50g (National Institutes of Health, USA).

About 70% of the renal tissue was used for leukocyte isolation as described previously.²¹  Tissues were homogenized with a gentleMACS dissociator (Miltenyi Biotec, USA) according to the manufacturer’s instructions. Digestion with Collagenase II (500 U/ml, Worthington, USA) was performed before and after homogenization for 20 min at 37°C. A 70-μM cell strainer was used to obtain single cell suspensions. Incubation in erythrocyte lysis buffer (420301, BioLegend, USA) for 1 min was done for erythrocyte lysis. For flow cytometry to analyze the proportion of Nav1.3-expressing neutrophils, cells were gated by forward scatter/sideward scatter and by expression of CD45 (clone 30-F11, BioLegend, USA) and fixable viability dye eFluor 506 (65–0866, eBioscience, USA). Ly6G-FITC (clone 1A8, BioLegend, USA) in combination with Nav1.3 (ASC023, Alomone Labs) was applied for detection of Nav1.3-expressing polymorphonuclear neutrophils. FACS CantoII (BD Biosciences, Germany) was used for flow cytometry, and data analysis was done using Kaluza software 1.3 (Beckmann Coulter, Germany).

Migration of polymorphonuclear neutrophils across a dense layer of endothelial cells or the extracellular matrix protein collagen type I was performed in transwell chambers. Transwell filters, 5 µm (Costar, Omnilab, Germany), were coated either with 50 mg/ml rat tail collagen type I (Corning, Omnilab) alone or with collagen type I and f.End5 or human umbilical vein endothelial cells that were grown to confluency in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal calf serum (f.End5) or endothelial cell growth medium-2 Bullet-kit with 5% fetal calf serum (human umbilical vein endothelial cells), starved for 12 h in Dulbecco’s Modified Eagle’s Medium or endothelial cell growth medium-2 Bullet-kit without fetal calf serum, and stimulated with 10 ng/ml mouse or human tumor necrosis factor-α (R&D Systems) for 4 h at 37°C. Isolated mouse bone marrow polymorphonuclear neutrophils or human whole blood polymorphonuclear neutrophils were diluted to 5 × 10⁵ in 100 µl Dulbecco’s Modified Eagle’s Medium and allowed to transmigrate through transwell filters coated with collagen type I, collagen type I and fEnd.5 cells, or collagen type I and human umbilical vein endothelial cells toward 2 mg/ml N-Formyl-Met-Leu-Phe (Sigma) for 2 h at 37°C. Wells were either left untreated for control or treated with 100 nM tetrodotoxin; 1 µM ICA121431; 0.5 µM Pterinotoxin 2; 10 µM, 100 µM, or 1000 µM lidocaine (Sigma); or 100 µM lidocaine together with 100 nM tetrodotoxin. Afterward, cells migrated to the lower chamber were visualized by staining nuclei with 4′,6-Diamidine-2′-phenylindole dihydrochloride and counted using a fluorescence microscope. Each condition was performed in duplicate or triplicate, and 9 or 12 high power fields/well were counted at 100-fold (mouse polymorphonuclear neutrophils) or 40-fold (human polymorphonuclear neutrophils) magnification.

For analyzing Nav1.3-dependent differences in directionality of cell migration and cell velocity, protocols of Polesskaya et al.²²  and Pepperell and Watt²³ were used in a modified manner. Freshly isolated mouse bone marrow neutrophils were diluted to 3 × 10⁶/ml in Dulbecco’s Modified Eagle’s Medium containing 10% fetal calf serum, and 6 µl was seeded to rat tail collagen type I–coated viewing channels of Ibidi µ-Slides (Ibidi, Germany) with or without addition of 100 nM tetrodotoxin, 1 µM ICA121431, and 0.5 µM Pterinotoxin-2, respectively. After allowing cells to adhere for 30 min at 37°C in a humidified chamber followed by removal of nonadherent cells by washing, an N-Formyl-Met-Leu-Phe gradient was established in the chambers by injecting 10 µM N-Formyl-Met-Leu-Phe in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal calf serum to one of the two ports of the chemoattractant chamber. Neutrophils were allowed to migrate toward N-Formyl-Met-Leu-Phe for 45 min at 37°C before video analysis. Observation of chemotaxis was performed recording videos during 15 min, saving a picture every 10 s with an Olympus IX81 microscope, Retiga EXI fluorescence camera (QImaging), and StreamPix Studio 6 software (NorPix Inc., Canada). Cell tracking was performed using the ImageJ plugin “Manual Tracking.” Tracking data of a total of 60 cells (three repeats with 20 cells each) for every experimental condition were analyzed with the ImageJ plugin “Chemotaxis Tool” (Ibidi) to derive trajectory plots and quantification data including accumulated migration distance, Euclidian distance, directionality of cell migration, forward migration indices, cell velocity, and Rayleigh test results assessing distribution of migration data.

Bone marrow polymorphonuclear neutrophils were stained with 5 µM 2′,7′-Dichlorofluorescin diacetate (Invitrogen, Fisher Scientific, Germany), and this dye was allowed to diffuse into cells for 15 min at 37°C in the dark. Afterward, baseline fluorescence was measured with enzyme-linked immunosorbent assay using a Tecan Spectrafluor Plus (Tecan, Germany). Cells were stimulated with 1 µM PMA, and 2.5 × 10⁵ cells/well were incubated in Hank’s balanced salt solution in a 96-well flat-bottom plate (Greiner, Germany) with or without 100 nm tetrodotoxin, 1 µM ICA121431, and 0.5 µM Pterinotoxin-2. Reactive oxygen species-dependent increase in fluorescence was measured after 5, 10, 20, 30, 40, 50, and 60 min. Data are shown for 60 min only.

Whole cell patch clamp experiments were carried out on bone marrow mouse and human peripheral blood polymorphonuclear neutrophils as well as HEK293 cells stably expressing rat Nav1.3 as described previously.²⁴  In addition, Nav1.3-expressing mouse polymorphonuclear neutrophils were isolated from ischemic kidneys 24 h after 45 min of ischemia reperfusion injury. Nav1.3 antibody (2 µg) was coupled to magnetic Protein G beads (100 µl, Milteny Biotech, Germany) overnight at 4°C and used to capture Nav1.3-expressing cells of cell lysates from one ischemic kidney each for 1 h at room temperature. After washing with phosphate-buffered saline containing 0.5% bovine serum albumin and 2 nM EDTA, cells were cultivated in RPMI-1640 (Gibco, Germany) supplemented with 10% heat-inactivated fetal bovine serum (Lonza), and only small, round cells where between two and five Nav1.3 beads were still attached were immediately used for patch clamp recording. HEK293 cells were cultured in Dulbecco’s Modified Eagle’s Medium supplemented with 25 mM HEPES (Lonza), 100 U/ml penicillin/streptomycin (Lonza), 3 mM taurine (Sigma-Aldrich, Germany), and 10% heat-inactivated fetal bovine serum (Lonza). Cells were cultivated in cell culture flasks at 37°C and 5% CO₂. Membrane sodium currents were explored at room temperature using an EPC10 amplifier (HEKA Instruments Inc., USA). Patch pipettes were made from glass capillaries (Science Products, Germany) on a DMZ-Universal Puller (Zeitz, Germany) and heat-polished to give a resistance of 2.0 to 2.5 MΩ when filled with pipette solution. The pipette solution contained (mmol/l) 140 cesium fluoride (Sigma) 10 sodium chloride (Sigma), 1 EGTA, and 10 HEPES (pH 7.4). The external solution contained (mmol/l) 140 sodium chloride, 3 potassium chloride, 1 calcium chloride (Sigma), 1 magnesium chloride (Sigma), 10 HEPES, and 15 glucose (pH 7.4). Voltage errors were minimized by compensation of the series resistance by 60 to 80%, and capacitance artefacts were canceled using the amplifier circuitry. Currents were filtered at 5 kHz. Linear leak subtraction, based on resistance estimates from four hyperpolarizing pulses applied before the test pulse, was used for all voltage-clamp recordings other than use-dependent block. Data were acquired and stored with Patchmaster v20x60 software (HEKA Instruments Inc., USA).

Statistical analyses were conducted with Prism 7 (GraphPad Inc., USA). Data that passed normality testing for distribution of data were tested with one-way ANOVA and corrected with Bonferroni for multiplicity. Data that did not pass normality testing for distribution of data were tested with the Kruskal-Wallis test and were corrected with Dunn for multiplicity. For migration, the Rayleigh test assessing inhomogenous distribution of data was performed with the ImageJ Chemotaxis and Migration Tool 2.0 (Ibidi). A statistical power calculation to determine the number of experiments needed to achieve significant differences was not performed, as all experiments are considered to be pilot experiments and numbers are based on previous experience. Two out of 12 sets of data for mouse adhesion assays had to be excluded since the difference in adherent polymorphonuclear neutrophils between medium control and tumor necrosis factor-α was not at least 1.5-fold, and we had confirmed previously that in this case, endothelial cells had reached biologic senescence, are no longer activated by tumor necrosis factor-α, and could not be used for adhesion assays anymore. For all adhesion, transmigration, and directed migration assays, data were blinded, and experiment and data analysis were conducted by different persons. All values are presented as mean ± SD, and P < 0.05 was considered significant with the exception of data of polymorphonuclear neutrophil transmigration through collagen that did not pass normality testing and are therefore presented as median ± interquartile range.

Over the past few years voltage-gated sodium channels have been shown to be expressed in several nonneuronal cells.³  Based on these reports and the knowledge that unspecific blockers of voltage-gated sodium channels can modify polymorphonuclear neutrophil function,^12,16  we asked if voltage-gated sodium channels are expressed in mouse and human polymorphonuclear neutrophils. In freshly isolated mouse bone marrow polymorphonuclear neutrophils that were attached to cell culture plastic for 24 h, we were able to show messenger RNA (mRNA) expression of Nav1.1 (very weak band), Nav1.3, Nav1.4, Nav1.5, Nav1.6, and Nav1.7 (weak band). In contrast, no expression of Nav1.2, Nav1.8, and Nav1.9 could be detected (fig. 1A). In isolated polymorphonuclear neutrophils from human blood, however, all nine α-subunits but Nav1.1 seem to be expressed (fig. 1B).

After having identified mRNA of multiple voltage-gated sodium channel α-subunits in both mouse and human polymorphonuclear neutrophils, we investigated which of the voltage-gated sodium channels are detected on a protein level on polymorphonuclear neutrophils. Mouse polymorphonuclear neutrophils attached to culture dish plastic for 24 h were stained with antibodies against different voltage-gated sodium channels. We only detected strong staining for Nav1.3 protein in some but not all GR-1–positive polymorphonuclear neutrophils (fig. 1, C–E), whereas antibodies against Nav1.1, Nav1.5, and Nav1.7 showed no staining (not shown). Next we investigated if cytokines like tumor necrosis factor-α, interleukin-6, and interleukin-1β that activate polymorphonuclear neutrophils during inflammation are able to upregulate Nav1.3 expression. Nav1.3 mRNA expression is enhanced 16-fold in polymorphonuclear neutrophils attached to cell culture plastic for 4 h compared to polymorphonuclear neutrophils in suspension or polymorphonuclear neutrophils attached for 15 min only (fig. 1F). Stimulation with tumor necrosis factor-α, interleukin-6 (not shown), or interleukin-1β (not shown) did not increase Nav1.3 expression of attached polymorphonuclear neutrophils after 4 h, suggesting that adhesion and not inflammation is triggering Nav1.3 expression (fig. 1F).

We next aimed to analyze the functional role of Nav1.3 during attachment of polymorphonuclear neutrophils to activated endothelial cells, an early step of neutrophil recruitment. While Nav1.9, Nav1.8, and Nav1.5 are not blocked by the unselective voltage-gated sodium channel blocker tetrodotoxin, all other α-subunits are blocked by low nanomolar concentrations of tetrodotoxin.^25,26  In order to narrow down the effects to Nav1.3 using a pharmacologic approach, we combined the effects of tetrodotoxin with that of ICA121431 (ICA, 1.0 µM), a synthetic blocker of Nav1.1 and Nav1.3²⁷  and Pterinotoxin-2 (0.5 µM), a synthetic analogon of a spider venom blocking Nav1.3 and Nav1.7. Although this combination of blockers cannot be used to specifically examine Nav1.3, Nav1.3 is the only voltage-gated sodium channel α-subunit blocked by all three inhibitors, and thus their effects on polymorphonuclear neutrophils should deliver a good indication whether or not Nav1.3 is functionally expressed. Indeed, tetrodotoxin significantly reduced mouse and human polymorphonuclear neutrophil adhesion rates to 56 ± 9% and 54 ± 11%, respectively, as compared to untreated cells. Furthermore, both 1.0 µM ICA and 0.5 µM Pterinotoxin-2 induced almost identical effects as tetrodotoxin by reducing mouse polymorphonuclear neutrophil adhesion to 53 ± 10% and 55 ± 9% and human polymorphonuclear neutrophil adhesion to 50 ± 16% and 50 ± 14%, respectively (fig. 2, A and B).

Furthermore, the majority of polymorphonuclear neutrophils that were attached to activated endothelial cells were positively stained for Nav1.3 (fig. 2, C–F). The specificity of the antibody used to visualize Nav1.3 protein expression was verified by Western blotting of protein lysates derived from HEK293 cells transfected with complementary DNA for different voltage-gated sodium channels. Only in Nav1.3-transfected HEK293 cells, a corresponding band with a molecular weight of 226 kd could be observed, whereas Nav1.7 transfected and vector-transfected cells were negative (Supplemental Digital Content, fig. S1, https://links.lww.com/ALN/B636).

In a next step, we asked if voltage-gated sodium channels and especially Nav1.3 are functionally expressed on polymorphonuclear neutrophils that are recruited into inflamed tissue in vivo. First we stained mouse myocardial tissue cryosections obtained after myocardial ischemia for 30 min followed by reperfusion for 24 h or 72 h for Nav1.3 and the polymorphonuclear neutrophil marker GR-1 (24 h and low magnification of 72 h are shown in Supplemental Digital Content, fig. S2, https://links.lww.com/ALN/B636). Nav1.3 staining could be detected on individual cells within the infarcted area. High magnifications of 72 h MI with reperfusion injury sections revealed that Nav1.3 expression is colocalized with GR-1 in 25 ± 7% of all GR-1–positive polymorphonuclear neutrophils, and only 3 ± 2% of all cells were Nav1.3-positive and do not show staining for GR-1 (fig. 3, A and B). Notably, immunochemistry with specific antibodies against Nav1.1, Nav1.5, and Nav1.7 in the same tissue sections only gave intense staining in cardiomyocytes outside the infarct area, but not in inflammatory cells (Supplemental Digital Content, fig. S3, https://links.lww.com/ALN/B636). Thus, although we did not extensively rule out a role of other α-subunits in polymorphonuclear neutrophils, it seems that the expression and function of Nav1.3 in polymorphonuclear neutrophils does not apply for multiple α-subunits.

Since it has been reported that expression of voltage-gated sodium channels on monocytes/macrophages modulates MI with reperfusion injury injury in rats,^8,28  we investigated if Nav1.3 staining occurs on macrophages in the infarcted area. However, at neither 24 h nor 72 h after MI with reperfusion injury could a colocalization between the pan-macrophage marker F4/80 and Nav1.3 be observed, indicating that Nav1.3 protein is not expressed on macrophages in our model of MI with reperfusion injury (fig. 3C and Supplemental Digital Content, fig. S2, https://links.lww.com/ALN/B636). Next we asked if Nav1.3-expressing polymorphonuclear neutrophils are specific for MI with reperfusion injury or if Nav1.3 is present on polymorphonuclear neutrophils in other inflammatory or ischemic organs as well. In a model of kidney ischemia 24 h after the left renal pedicle was clamped for 45 min to induce unilateral ischemia reperfusion injury, we could indeed confirm the MI with reperfusion injury results and demonstrated that Nav1.3 is expressed by a proportion of polymorphonuclear neutrophils in ischemic kidney that was identified by GR-1 labeling in stained cryosections (fig. 3D). Flow cytometry analysis of a single kidney in this model revealed that 12.7% and 19.3% of all polymorphonuclear neutrophils are expressing Nav1.3 in Ly6G high and in Ly6G dim polymorphonuclear neutrophil populations, respectively (fig. 3E). Overall flow cytometry analysis of multiple (n = 6) ischemic kidneys confirmed that 19.3 ± 2% of all polymorphonuclear neutrophils are expressing Nav1.3 and Ly6G (fig. 3F). These data demonstrate a specific Nav1.3 staining that is associated with a fraction of polymorphonuclear neutrophils in mouse heart and kidney tissue after ischemia, pointing toward a common and important function of Nav1.3 during neutrophil recruitment.

As we discovered polymorphonuclear neutrophils expressing Nav1.3 and showed that we could reduce polymorphonuclear neutrophil adhesion with Nav1.3-specific blockers, we wanted to explore if Nav1.3 is influencing other polymorphonuclear neutrophil functions during tissue invasion. Therefore, we addressed in the next step, if transmigration of polymorphonuclear neutrophil through an endothelial cell layer and underlying basement membrane is altered by blocking Nav1.3.^29–31  To mimic this in vitro, transmigration of mouse bone marrow polymorphonuclear neutrophils and human peripheral blood polymorphonuclear neutrophils through a confluent layer of f.End5 and human umbilical vein endothelial cells on collagen-coated transwell filters was recorded. Medium supplemented with N-Formyl-Met-Leu-Phe in the lower chamber was set as the reference (100 ± 24%; n = 6). Coincubation with tetrodotoxin (100 nM) significantly reduced rates of transmigrated polymorphonuclear neutrophils to 63 ± 12% (P = 0.0063, n = 6). The addition of ICA and Pterinotoxin-2 reduced transmigration rates to 66 ± 18% and 65 ± 18%, respectively, again suggesting an important role for Nav1.3 in the transmigration process (fig. 4A).

To eliminate possible effects of tetrodotoxin, ICA, and Pterinotoxin-2 by blocking voltage-gated sodium channels expressed by endothelial cells, transmigration assays were repeated without f.End5 cells on transwell filters coated with collagen-I only. Transmigration on collagen-I matrices was reduced to 64 ± 30% with tetrodotoxin (n = 8), to 40 ± 24% with ICA (n = 9), and to 32 ± 15% with Pterinotoxin-2 (n = 9) compared to medium containing N-Formyl-Met-Leu-Phe (control, set to 100 ± 9%; fig. 4B, data are reported as median ± interquartile range as they did not pass normality test). The data suggest that Nav1.3 expressed by polymorphonuclear neutrophils is responsible for the observed alteration of polymorphonuclear neutrophil transmigration. Nav1.3-dependent regulation of polymorphonuclear neutrophil transmigration could be confirmed in the human system as well, using human peripheral blood polymorphonuclear neutrophils and human umbilical vein endothelial cells. However, in comparison with the mouse system, higher tetrodotoxin concentration of 200 nM and a preincubation of tetrodotoxin, ICA, and Pterinotoxin-2 with human polymorphonuclear neutrophils for 15 min before adding polymorphonuclear neutrophils to human umbilical vein endothelial cells was necessary to achieve a reduction of transmigration to 56 ± 19% by tetrodotoxin, to 62 ± 22% by ICA, and to 59 ± 22% by Pterinotoxin-2 (fig. 4C). Taken together, these data strongly suggest a regulating role for Nav1.3 not only for adhesion, but also for transmigration in both human and mouse polymorphonuclear neutrophils.

Next we investigated which aspects of directed migration toward a chemotactic gradient of N-Formyl-Met-Leu-Phe like accumulated distance, velocity, or direction are affected by blocking Nav1.3 channels on mouse polymorphonuclear neutrophils. Trajectory plots of polymorphonuclear neutrophil migration pathways toward N-Formyl-Met-Leu-Phe throughout the recording time of 20 min in Ibidi µ-slides show a directed migration for N-Formyl-Met-Leu-Phe only, whereas blocking Nav1.3 with tetrodotoxin, ICA, and Pterinotoxin-2 resulted in a nondirected migration pattern (fig. 5A).

By measuring the Euclidean distance, we could demonstrate that the shortest distance between the start- and endpoint of migration was significantly reduced for all three Nav1.3 channel blockers. Using N-Formyl-Met-Leu-Phe as a chemoattractant, the Euclidean distance was shortened from 72 ± 39 µm in control solution to 21 ± 29 µm with tetrodotoxin, 18 ± 17 µm with ICA, and 10 ± 9 µm with Pterinotoxin-2 (fig. 5B). The directness of migration toward N-Formyl-Met-Leu-Phe resembling the ratio of Euclidean and accumulated distance was significantly reduced from 0.53 for N-Formyl-Met-Leu-Phe to 0.19 with tetrodotoxin, 0.14 with ICA, and 0.07 with Pterinotoxin-2 (not shown). These data indicate that Nav1.3 is required for directing the migration of polymorphonuclear neutrophils within the N-Formyl-Met-Leu-Phe gradient. Nav1.3-dependent directionality of polymorphonuclear neutrophil movement was further confirmed by the forward migration indices parallel and forward migration indices perpendicular, where forward migration indices parallel was significantly decreased from 0.43 ± 0.27 for N-Formyl-Met-Leu-Phe alone to 0.10 ± 0.2 with tetrodotoxin, 0.04 ± 0.16 with ICA, and 0.01 ± 0.05 with Pterinotoxin-2 (fig. 5C).

As forward migration indices parallel exceeds forward migration indices perpendicular and the Rayleigh test for homogenous distribution revealed a P value < 0.05, only migration of polymorphonuclear neutrophils toward N-Formyl-Met-Leu-Phe fulfill the criteria for chemotactic migration, whereas addition of tetrodotoxin, ICA, or Pterinotoxin-2 blocked chemotactic movement of polymorphonuclear neutrophils toward N-Formyl-Met-Leu-Phe. Surprisingly, the mean accumulated distance polymorphonuclear neutrophils migrated as well as the velocity were significantly reduced only by the combination of N-Formyl-Met-Leu-Phe and tetrodotoxin (accumulated distance: 106 ± 35 µm, velocity: 0.12 ± 0.04) in comparison to N-Formyl-Met-Leu-Phe (accumulated distance: 135 ± 29 µm, velocity: 0.16 ± 0.03). These data indeed suggest that further α-subunits other than Nav1.3 are functionally expressed in polymorphonuclear neutrophils as well. In contrast, the more specific Nav1.3 blockers ICA and Pterinotoxin-2 caused a slight and not significant reduction of the mean accumulated distance (ICA: 128 ± 36 µm and Pterinotoxin-2: 132 ± 35 µm) and of the velocity (ICA: 0.15 ± 0.04 and Pterinotoxin-2: 0.15 ± 0.04; fig. 5, D and E). Taken together, we were able to show that polymorphonuclear neutrophil movement toward a gradient of N-Formyl-Met-Leu-Phe is negatively affected by voltage-gated sodium channel blockade in vitro.

An obvious interpretation of the in vitro adhesion, migration, and chemotaxis data is that Nav1.3 seems to carry a relevant function in polymorphonuclear neutrophils. In order to verify an inhibitory effect of tetrodotoxin, ICA, and Pterinotoxin-2 on Nav1.3, we performed whole cell patch clamp recordings on HEK293 cells stably expressing Nav1.3. As expected, we observed that 100 nM tetrodotoxin, 1.0 µM ICA, and 0.5 µM Pterinotoxin-2 effectively inhibited sodium inward currents in these cells (fig. 6, A–C).

We also asked if voltage-gated sodium currents could be recorded in isolated mouse and human polymorphonuclear neutrophils attached to cell culture plastic or stimulated with inflammatory cytokines. In whole cell voltage clamp recordings of randomly selected cultured polymorphonuclear neutrophils, however, we were not able to detect any voltage-gated sodium currents in polymorphonuclear neutrophils (data not shown, n = 10 for both species). Being able to identify and sort a population of polymorphonuclear neutrophils evidently expressing Nav1.3 after kidney ischemia, e.g., Nav1.3. in Lys6G high cells, we again employed whole cell patch clamp recordings in order to determine whether or not polymorphonuclear neutrophils that stain for Nav1.3 are able to produce voltage-activated sodium currents. However, the voltage-activated sodium currents of 15 examined polymorphonuclear neutrophils that bound Nav1.3 beads were only minimally increased above baseline recordings (fig. 6D). Thus although polymerase chain reaction and pharmacologic experiments suggest that voltage-gated sodium channels are functionally expressed in both mouse and human polymorphonuclear neutrophils, our patch clamp experiments on polymorphonuclear neutrophils from ischemic organs favor more the possibility that expression of voltage-gated sodium channels is not high enough to generate voltage-activated sodium currents from the same quality as recorded on Nav1.3-transfected HEK293 cells (fig. 6E). Alternatively, expression of Nav1.3 in neutrophils might obey a very tight time window when these cells are activated in vivo, so that functional expression is rapidly lost under culture conditions in vitro.

The unselective voltage-gated sodium channel blocker lidocaine has been shown to reduce myocardial infarct size^15,32  and to inhibit several neutrophil functions like transmigration through endothelial cells.¹²  Therefore, we finally explored if lidocaine has an impact on Nav1.3-dependent polymorphonuclear neutrophil adhesion to endothelial cells and to transmigration on collagen-coated transwell filters. Lidocaine (100 µM) reduced adhesion of mouse polymorphonuclear neutrophils to fEND.5 cells to 60 ± 9% compared to medium controls. Blockade of voltage-gated sodium channels with tetrodotoxin (100 nM, 58 ± 8%), ICA (1 µM, 55 ± 7%), or Pterinotoxin-2 (0.5 µM, 59 ± 7%) in combination with lidocaine revealed no further reduction in adhesion (fig. 7A, n = 10). Transmigration rates were not affected by adding 10 µM lidocaine (data not shown), but significantly reduced to 54 ± 8% by 100 µM lidocaine (fig. 7B, n = 9). Coincubation of 100 µM lidocaine with tetrodotoxin (100 nM, 55 ± 12%), ICA (1 µM, 55 ± 11%), or Pterinotoxin-2 (0.5 µM, 59 ± 7%) did not show any additive effects in reducing the number of transmigrated polymorphonuclear neutrophils (fig. 7B), thus suggesting that lidocaine and the voltage-gated sodium channel blockers tetrodotoxin, ICA, and Pterinotoxin-2 might operate via the same mechanisms.

In the current study, we investigated expression and functional roles of voltage-gated sodium channels in polymorphonuclear neutrophils and show that both mouse and human polymorphonuclear neutrophils express mRNA for most known voltage-gated sodium channel α-subunits. However, only Nav1.3 protein is found in mouse polymorphonuclear neutrophils in vivo when residing in the infarcted myocardial area or in injured kidney tissue after ischemia and reperfusion, suggesting a common function of Nav1.3 for recruited polymorphonuclear neutrophils. Accordingly, both unselective voltage-gated sodium channel inhibitors (e.g., tetrodotoxin and lidocaine) and substances preferentially blocking Nav1.3 (e.g., ICA121431 and Pterinotoxin-2) strongly attenuated polymorphonuclear neutrophil adhesion, transmigration, and chemotaxis in vitro. These data suggest that Nav1.3, and possibly also other voltage-gated sodium channel α-subunits, are required for proper function of polymorphonuclear neutrophils and thus of the innate immune system. These novel findings indicate that voltage-gated sodium channels like Nav1.3 in polymorphonuclear neutrophils might be relevant molecular targets for immunomodulation.

While there is no doubt that one of the main functional roles of voltage-gated sodium channels is to generate the upstroke of the action potential in excitable cells, there is also meanwhile little doubt about the notion that voltage-gated sodium channels also perform diverse noncanonical roles in several types of cells in mammals.³  The expression and functional roles of different voltage-gated sodium channel α-subunits have been described in various types of immune cells including monocytes, astrocytes, dendritic cells, and lymphocytes.^3–5,10  Similar to our data on the role of Nav1.3 in polymorphonuclear neutrophils, different voltage-gated sodium channel α-subunits in immune cells were frequently reported to be important for cell motility and migration.³  Nav1.5 seems to be the predominant subunit in macrophages, where it was reported to regulate endosomal acidification, involved in phagocytosis, but also to act as a pathogen sensor.^4–6  Furthermore, Nav1.5 is involved in migration and proliferation in astrocytes.⁷  Nav1.4, on the other hand, seems to regulate CD4-positive lymphocytes,⁹  and Nav1.7 was reported to regulate migration and cytokine responses of dendritic cells.¹⁰  While an early immunohistologic study reported on expression of Nav1.3 in astrocytes, the functional role of Nav1.3 in these cells was not examined.³³  In contrast to many types of immune cells, the expression of voltage-gated sodium channels in granulocytes has not been described in previous reports. We report here that both mouse and human polymorphonuclear neutrophils express multiple voltage-gated sodium channel α-subunits, and that at least Nav1.3 seems to be relevant for proper function of polymorphonuclear neutrophils, at least in vitro.

In a translational perspective, our data correlate very well with the established immunomodulatory property of the unselective sodium channel blocker lidocaine, and also with the general assumption that an inhibition of polymorphonuclear neutrophil function accounts for this property of lidocaine.^12,14,34  However, several previous reports describing that lidocaine and other local anesthetics abbreviate the activity of polymorphonuclear neutrophils have postulated voltage-gated sodium channel–independent mechanisms. To our knowledge, however, none of these studies explored whether or not polymorphonuclear neutrophils express voltage-gated sodium channels. One argument speaking against the involvement of voltage-gated sodium channels for the effects of lidocaine on polymorphonuclear neutrophils was noted in an early report describing that an inhibition of the production of superoxide anions in polymorphonuclear neutrophils by local anesthetics is not mimicked by tetrodotoxin.³⁵  In fact, we could confirm and also extend this notion. We found that tetrodotoxin in the nanomolar range as well as ICA and Pterinotoxin-2 evoke a slight but not significant reduction in the generation of reactive oxygen species in phorbol ester–stimulated polymorphonuclear neutrophils (Supplemental Digital Content, fig. S4, https://links.lww.com/ALN/B636). Therefore, mechanisms other than voltage-gated sodium channels indeed seem to account for the effect of lidocaine on reactive oxygen species production in polymorphonuclear neutrophils. As Nav1.5, Nav1.8, and Nav1.9 are tetrodotoxin-resistant and ICA as well as Pterinotoxin-2 do not block all α-subunits, one cannot completely rule out a role of voltage-gated sodium channels for reactive oxygen species production in polymorphonuclear neutrophils. Nevertheless, our data clearly indicate that tetrodotoxin as well as ICA and Pterinotoxin-2 rather effectively inhibit adhesion, transmigration, and chemotaxis of polymorphonuclear neutrophils in vitro. We also observed that lidocaine mimicked the effects all three blockers, but that the combination of lidocaine with any of the three substances did not induce additive effects. These results indicate that tetrodotoxin, ICA, Pterinotoxin-2, and lidocaine target the same mechanism (e.g., Nav1.3) on polymorphonuclear neutrophils to regulate properties like adhesion, transmigration, and chemotaxis.

The presented results of this report reveal some weakness when it comes to identifying which subcellular functions and signaling pathways in polymorphonuclear neutrophils depend on voltage-gated sodium channels, and this question urges further studies on this topic. Regulation of adhesion, transmigration, and chemotaxis of course include multiple possible roles for voltage-gated sodium channels in polymorphonuclear neutrophils, making it rather challenging to identify specific mechanisms. Accordingly, despite the large number of published reports on noncanonical roles of voltage-gated sodium channels in different kinds of cells, it is still unclear through which mechanisms they regulate different cellular functions.³  While we detected only minimally increased and no prominent sodium currents in polymorphonuclear neutrophils evidently expressing Nav1.3, the majority of cell types in which noncanonical roles of voltage-gated sodium channels were previously described seem to produce sodium currents when explored in patch clamp experiments. When compared to excitable cells, however, sodium currents and thus the expression level of voltage-gated sodium channels in nonexcitable cells seem to be rather small. For example, the human endothelial cell line human umbilical vein endothelial cells expressing Nav1.5 produce only small currents up to 160 to 200 pA, but these were sufficient to potentiate vascular endothelial growth factor-induced ERK1/2 activation through the PKCα-B-RAF signaling axis.¹¹  As a possible mechanism for voltage-gated sodium channel–dependent effects in nonexcitable cells, it was suggested that sodium influx might regulate the membrane potential by regulating the activity of the sodium/potassium-adenosine triphosphatase.³⁶  It has also been shown that sodium influx can couple to a sodium/Ca²⁺ exchanger, thus regulating the intracellular calcium concentration.³⁷  Calcium is required for neutrophil recruitment as it contributes to upregulation of adhesion molecules, cytoskeletal reorganization, and migration.³⁸  Increases in intracellular calcium subsequent to neutrophil activation via several stimulating agonists are reported to be mainly mediated by store operating calcium entry channels with Orai1 as their most common representative.^39,40  In addition to that, sodium channels seem to cause elevations of intracellular calcium.^29–31  Both dependent and independent of calcium, protein kinase C signaling is required for neutrophil functions like adhesion and transmigration.⁴¹  Interestingly, Nav1.3 was found to be modulated by activation of protein kinase C,⁴²  and lidocaine was shown to affect adhesion and transmigration of polymorphonuclear neutrophils via protein kinase C signaling.^12,16,20,43  Thus, Nav1.3 and other voltage-gated sodium channels in polymorphonuclear neutrophils could conceivably act via those signaling pathways. This notion might seem vague as we detected only minimally above baseline increase, but no prominent sodium currents in polymorphonuclear neutrophils. However, it is still possible that small sodium currents less than or at the detection level are generated in polymorphonuclear neutrophils. Alternatively, it is possible that an intracellular localization or an expression of a splice variant of Nav1.3 may explain the lack of prominent sodium currents in polymorphonuclear neutrophils. In fact, an intracellularly localized splice variant of the α-subunit Nav1.6 was reported to regulate cell motility of macrophages and melanoma cells.⁴⁴  Furthermore, the expression of neonatal splice variants of voltage-gated sodium channels was described in many types of nonexcitable cells including astrocytes and macrophages.^44,45  These possibilities of course remain very speculative and need to be explored in further studies.

Neutrophils are regarded to play both positive and deleterious roles after ischemia and tissue injury. Several studies in the past addressed the function of polymorphonuclear neutrophils and tried to identify mechanisms to be influenced for a better outcome after events of ischemia. This includes research on effects of local anesthetics, e.g., lidocaine,³²  depletion of neutrophils,^26,46  or specific antibodies and knockout strategies targeting adhesion molecules.^47–49  Our data now implicate that voltage-gated sodium channels might be suitable targets in order to control or inhibit the activity of polymorphonuclear neutrophils, and we also suggest that Nav1.3 is likely to be a relevant voltage-gated sodium channel α-subunit at this point. While relatively little is known about the role of Nav1.3 in adult humans, preclinical studies on rodents suggest that it is expressed in the developing nervous system and then strongly downregulated in adulthood. There is evidence for an upregulation of Nav1.3 after peripheral nerve injury, leading to the assumption that Nav1.3 contributes to neuropathic pain after nerve injury or diabetic neuropathy.^50–52  Rare channelopathies of Nav1.3 seem to be associated with epilepsy,⁵³  and Nav1.3 was also shown to be important for left ventricular contraction and sinoatrial nodal automaticity in the mouse heart.^54,55  When taking these possible roles of Nav1.3 in adult mammals into account, it is difficult to foresee whether or not potent and specific blockers of Nav1.3 would be effective modulators of the innate immune system without inducing severe cardiac or central nervous system side effects.

In summary, we report that α-subunits of voltage-gated sodium channels are functionally expressed in a fraction of polymorphonuclear neutrophils invading damaged heart and kidney tissue after ischemia and reperfusion in vivo. The α-subunit Nav1.3 seems to critically regulate pivotal neutrophil functions such as migration, adhesion, and transmigration. Therefore, we suggest that Nav1.3 and maybe other voltage-gated sodium channels are attractive targets for selective sodium channel blocking agents applied for immunomodulation.

The authors thank Kerstin Reher, Marina Golombek, and Silke Deus (Department of Anesthesiology and Intensive Care Medicine, Hannover Medical School, Hannover, Germany) and Herle Chlebush (Department of Nephrology, Hannover Medical School, Hannover, Germany) for excellent technical assistance.

The study was supported by funds from the Department of Experimental Anesthesiology, Hannover Medical School, Hannover, Germany. Marit Poffers, cand.med., and Nathalie Bühne, cand.med., were enrolled in the StrucMed-Program at Hannover Biomedical Research School (HBRS), Hannover, Germany, and Nathalie Bühne received stipends from HBRS.

The authors declare no competing interests.