Human Saphenous Vein Endothelial Cells Express a Tetrodotoxin-resistant, Voltage-gated Sodium Current*

Whole-cell patch-clamp electrophysiological investigation of endothelial cells cultured from human saphenous vein (HSVECs) has identified a voltage-gated Na+ current with a mean peak magnitude of −595 ± 49 pA (n = 75). This current was inhibited by tetrodotoxin (TTX) in a concentration-dependent manner, with an IC50value of 4.7 μm, suggesting that it was of the TTX-resistant subtype. An antibody directed against the highly conserved intracellular linker region between domains III and IV of known Na+ channel α-subunits was able to retard current inactivation when applied intracellularly. This antibody identified a 245-kDa protein from membrane lysates on Western blotting and positively immunolabeled both cultured HSVECs and intact venous endothelium. HSVECs were also shown by reverse transcription-polymerase chain reaction to contain transcripts of the hH1 sodium channel gene. The expression of Na+ channels by HSVECs was shown using electrophysiology and cell-based enzyme-linked immunosorbent assay to be dependent on the concentration and source of human serum. Together, these results suggest that TTX-resistant Na+ channels of the hH1 isoform are expressed in human saphenous vein endothelium and that the presence of these channels is controlled by a serum factor.

Whole-cell patch-clamp electrophysiological investigation of endothelial cells cultured from human saphenous vein (HSVECs) has identified a voltage-gated Na ؉ current with a mean peak magnitude of ؊595 ؎ 49 pA (n ‫؍‬ 75). This current was inhibited by tetrodotoxin (TTX) in a concentration-dependent manner, with an IC 50 value of 4.7 M, suggesting that it was of the TTXresistant subtype. An antibody directed against the highly conserved intracellular linker region between domains III and IV of known Na ؉ channel ␣-subunits was able to retard current inactivation when applied intracellularly. This antibody identified a 245-kDa protein from membrane lysates on Western blotting and positively immunolabeled both cultured HSVECs and intact venous endothelium. HSVECs were also shown by reverse transcription-polymerase chain reaction to contain transcripts of the hH1 sodium channel gene. The expression of Na ؉ channels by HSVECs was shown using electrophysiology and cell-based enzyme-linked immunosorbent assay to be dependent on the concentration and source of human serum. Together, these results suggest that TTX-resistant Na ؉ channels of the hH1 isoform are expressed in human saphenous vein endothelium and that the presence of these channels is controlled by a serum factor. Vascular endothelial cells form the primary interface between the blood and the underlying tissue. These cells not only provide a barrier of varying permeability between the blood and the smooth muscle of the vessel wall, but are a major contributor to the processes of vascular growth and repair, vascular autoregulation, and control of vascular tone by secretion of both relaxant and contractile factors (1,2). Endothelial cells are known to possess a broad spectrum of ion channels that open in response to a variety of stimuli, including membrane potential, receptor occupation, elevation of [Ca 2ϩ ] i , and mechanical deformation induced by flow (3). Levels of [Ca 2ϩ ] i are an important factor in the control of endothelial cell function (4), and ion channels, with their ability to allow both Ca 2ϩ entry either directly or indirectly, via control of membrane potential, are critical to this process (5).
Definitive data regarding the exact repertoire of ion channels expressed by endothelial cells are still sparse, particularly in venous endothelium. In this study, we report the presence of a voltage-gated Na ϩ current present in human saphenous vein endothelial cells (HSVECs). 1 This type of channel is normally only expressed by classically excitable cells that generate action potentials such as neurons and cardiac and skeletal muscle. Voltage-gated Na ϩ channels are characterized by their kinetics; voltage dependence; and sensitivity to the guanidinium toxin, tetrodotoxin (TTX). TTX-sensitive Na ϩ channels are blocked by nanomolar concentrations of TTX and are found in tissues such as mature skeletal muscle (6). In contrast, TTX-resistant channels have a substantially lower affinity for the toxin, requiring 0.1-10 M for inhibition (7). TTX-resistant channels are found in a wide variety of tissue types, including cardiac cells (8) and denervated or developing skeletal muscle (9) and corneal endothelium (10). A third class of voltage-gated Na ϩ channels, expressed by embryonic cardiac cells (11) and dorsal root ganglion neurons (12), remain unblocked by TTX concentrations in excess of 100 M and are classified as TTXinsensitive. The voltage-gated sodium current we describe here in HSVECs is TTX-resistant and appears to result from expression of the cardiac Na ϩ channel gene (hH1).

EXPERIMENTAL PROCEDURES
Human Saphenous Vein Endothelial Cell Isolation and Culture-HSVECs were obtained by enzymatic release from saphenous vein harvested during high ligation of varicose veins or bypass surgery. After removal of any residual external connective tissue, the vein was carefully opened along its longitudinal axis with a scalpel blade. HSVECs were obtained by placing the vein luminal face down in a shallow Petri dish containing Ca 2ϩ -and Mg 2ϩ -free phosphate-buffered saline (PBS; 150 mM NaCl, 2 mM NaH 2 PO 4 , and 10 mM Na 2 HPO 4 ) and 1 mg/ml collagenase (Type II, Sigma) and incubating at room temperature (20 -22°C) for 30 min. Cells were placed in culture on fibronectin-coated dishes or flasks as appropriate and grown in M199 medium supplemented with heparin, endothelial cell growth supplement, antibiotic solution (200 units/ml penicillin and 200 g/ml streptomycin), and 10% (v/v) heat-inactivated human serum. Serum was obtained either from non-diabetic patients with peripheral arterial disease (Ͼ65 years old) or from healthy donors (Ͻ30 years old). Cultures, characterized by positive immunostaining for von Willebrand factor, were maintained at 37°C in humidified CO 2 in air atmosphere and used in experiments at passages 0 -3.
Electrophysiological Recording-Experiments were performed at room temperature (20 -22°C) using the whole-cell configuration of the patch-clamp technique (13) on subconfluent HSVECs grown in 35-mm diameter Petri dishes. These were placed on the stage of an inverted microscope (Diaphot 200, Nikon, Tokyo, Japan), visualized with phasecontrast optics, and continuously superfused at 2 ml/min with extracellular solution. The standard pipette and extracellular solutions were designed to isolate I Na (pipette: 120 mM CsCl, 10 mM EGTA, 2 mM MgCl 2 , 5 mM NaCl, 5 mM HEPES, 2 mM Na 2 ATP, and 0.5 mM Na 2 GTP; extracellular: 120 mM NaCl, 4 mM KCl, 2 mM MgCl 2 , 2 mM CaCl 2 , 10 mM TEA-Cl, and 10 mM HEPES (pH 7.3) with CsOH), although all current-clamp and some preliminary experiments employed "quasiphysiological" solutions (pipette: 140 mM KCl, 2 mM MgCl 2 , 1 mM CaCl 2 , 0.05 mM  EGTA, 20 mM HEPES, 2 mM Na 2 ATP, and 0.5 mM Na 2 GTP; extracellular: 135 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 2 mM CaCl 2 , 11 mM glucose,  and 10 mM HEPES (pH 7.3) with NaOH). Patch-clamp pipettes were manufactured from borosilicate glass (GC150-15TF, Clark Electromedical, Reading, United Kingdom) using a two-stage puller (PB7, Narashige, Tokyo, Japan) and fire-polished to give final resistances of 1-3 megaohms when filled with pipette solution. Whole-cell membrane currents (voltage-clamp) and potentials (current-clamp) were recorded using an Axopatch 200A patch clamp amplifier (Axon Instruments Inc., Foster City, CA) with analogue cell capacitance (mean cell capacitance ϭ 35.1 Ϯ 1.4 picofarads, n ϭ 131) and series resistance (routinely Ͻ5 megaohms) maximally compensated. Signals were low pass-filtered with an 8-pole Bessel-type filter at either 2 or 5 kHz prior to digitization at 10 kHz by a Digidata 1200 interface (Scientific Solutions, Solon, OH) and storage on a computer hard disk (486DX2, Opus Technology). Analysis was performed using pClamp 6 software (Axon Instruments Inc.), which was also employed to generate voltage step protocols. When recording I Na linear leakage, currents were subtracted using a 4-subpulse (P/4) method (14), and I Na amplitude was measured as the difference between the maximal inward current and the holding current level. Junction potentials were measured as described previously (15), although only determinations of the I Na reversal potential were corrected.
Reverse Transcription-Polymerase Chain Reaction (RT-PCR)-Total cellular RNA was extracted from HSVECs grown to confluence in T75 culture flasks by incubation with 6 ml of RNAzol reagent (Cinna/ Biotecx Labs Inc.) at 4°C for 10 min. Contaminating DNA was removed (Message Clean kit, GenHunter Corp.) prior to reverse transcription (Reverse Transcriptor kit, R&D Systems Ltd., Abingdon, UK), both according to the manufacturers' protocols. Sodium channel cDNA was amplified by PCR using oligonucleotide primers directed against the 3Ј-untranslated region of the human cardiac hH1 channel (16). The primer sequences used were 5Ј-GACCTGTGACCTGGTCTGGT-3Ј and 5Ј-CCATGTCCATGGAAAAATCC-3Ј (Perkin-Elmer, Warrington, UK). HSVEC RNA (1 g) was used for PCR amplification using 1 M primers for 30 cycles (94°C, 1 min; 50°C, 1 min; and 72°C, 1 min) at a MgCl 2 concentration of 1 mM. The resulting PCR fragments were analyzed on a 2.5% (w/v) agarose gel containing ethidium bromide and sequenced using an ABI 373 automated sequencer to confirm fragment identity.
Immunohistochemistry-Immunohistochemical analysis for the presence of voltage-gated Na ϩ channels in HSVECs was performed on intact tissue sections and cultured HSVECs using a polyclonal antibody raised in rabbits against the highly conserved cytosolic linker region between domains III and IV (peptide sequence TEEQKKYYNAM-KKLGSKKP, amino acids 1490 -1508) of known Na ϩ channel ␣-subunits ("anti-Na␣"; TCS Biologicals, Botolph Claydon, UK). Small lengths of human saphenous vein were fixed overnight in Zamboni's solution (consisting of 1.7% (w/v) paraformaldehyde and 15% (v/v) saturated picric acid in PBS) and washed daily in PBS/sucrose solution (PBS supplemented with 440 mM sucrose) for 7 days prior to mounting in OCT compound (Miles Inc.). Veins were cryosectioned using a microtome (Bright Instruments, Huntingdon, UK) to produce 8 -10-m thick transverse sections, which were transferred to 3-aminopropyltriethoxysilane-coated slides prior to overnight incubation at 4°C with primary antibody (anti-Na␣, 7.5 g/ml in PBS buffer supplemented with 1% (w/v) bovine serum albumin and 1% (v/v) human serum). Specific binding was visualized by alkaline phosphatase staining (Vector Red ® , Vector Labs, Peterborough, UK) subsequent to primary antibody detection using a biotinylated secondary antibody (goat anti-rabbit, 1:200 dilution, 1 h) and a tertiary streptavidin-alkaline phosphatase conjugate incubation (1:200 dilution, 1 h; Dako, High Wycombe, UK). The presence of endothelium on intact vein sections was verified by using a mouse monoclonal antibody directed against thrombomodulin. A protocol similar to that described above was employed for cultured HSVECs, which were, however, fixed by incubation in 100% methanol at 4°C for 2 min. Controls for both intact vein sections and HSVECs were prepared by omitting either the primary or secondary antibodies.
Cell-based Enzyme-linked Immunosorbent Assay-HSVECs (10 5 / well) were seeded onto 24-well plates and grown to confluence over 48 h. HSVECs were fixed by incubation in 100% methanol at 4°C for 2 min and washed with Tris-buffered saline supplemented with 0.5% (w/v) bovine serum albumin. Cells were incubated with the anti-Na␣ antibody (2 g/ml) for 40 min at 37°C, washed, and subjected to two further incubations (30 min each at 37°C) with a biotinylated secondary antibody (1:500) and a final streptavidin-horseradish peroxidase conjugate (1:500). Cells were thoroughly washed prior to assessment of anti-Na␣ antibody binding by colorimetric assay using o-phenylenediamine as the substrate, and the optical density was measured at 492 nm.
Data Analysis and Curve Fitting-Data are expressed as mean Ϯ S.E. (n ϭ number of observations). Normalized activation curves for I Na were calculated by dividing conductances (g Na ), derived from peak currents divided by the Na ϩ driving force (V m Ϫ E Na ), by the largest conductances measured. Steady-state inactivation curves (h ϱ ) and activation curves (m ϱ ) were fitted with a Boltzmann function, where V 0.5 is the midpoint and K V is the slope factor: Kinetic analysis of I Na was only performed if the peak current exceeded 500 pA, and time constants for current activation and inactivation were derived by fitting a Hodgkin-Huxley model (17) to the data as described elsewhere (18). The 50% inhibitory concentration (IC 50 ) for TTX was calculated by fitting the concentration inhibition curve to a logistic plot incorporating Hill coefficients (n H ) using MicroCal Origin (MicroCal Inc., Northampton, MA): bound ϭ [drug] n H/[drug] n H ϩ IC 50 . Reversal potentials (E rev ) were obtained by fitting a second-order polynomial to the current-potential (I-V) plots over the appropriate voltage regions (usually ϩ20 to ϩ80 mV). Where appropriate, results were tested for significance using Student's unpaired t test.
Materials-Culture materials were obtained from Gibco Laboratories (Paisley, UK), and the thrombomodulin antibody was a gift from Dr. J. Amiral (Serbio Research, Paris, France). TTX, purchased from Calbiochem (Nottingham, UK), was dissolved in water prior to addition to the appropriate solution. Unless indicated, all other chemicals were obtained from Sigma (Poole, UK).

RESULTS
Potassium Currents-Under whole-cell current clamp using quasiphysiological K ϩ -containing solutions, the resting membrane potential of single HSVECs was found to be Ϫ28 Ϯ 6 mV (n ϭ 24, range of Ϫ3 to Ϫ71 mV). Hyperpolarizing voltageclamp pulses from a holding potential of Ϫ50 mV produced small inward currents in 14 of the 24 cells (58%) investigated (Fig. 1A). These currents showed marked time-dependent inactivation at strongly negative potentials, inwardly rectified, conducting little outward current, and reversed close to the potassium equilibrium potential (Fig. 1B). Currents were fully eliminated by substituting Cs ϩ for K ϩ in the pipette solution and by the addition of 10 mM TEA in the extracellular solution, suggesting that they were carried by potassium ions.
Sodium Currents-Depolarizing voltage steps from a potential of Ϫ120 mV to more positive potentials (Ϫ40 to ϩ60 mV) elicited transient inward currents in HSVECs (Fig. 2A). These currents were present in 10 of the 24 cells (42%) investigated using K ϩ -containing intracellular solutions, but unlike the inward current described above, these currents remained even when intracellular K ϩ was substituted with Cs ϩ . All further electrophysiological experiments performed in this study used the Cs ϩ /TEA-containing solutions to isolate this transient inward current and to avoid contamination from K ϩ currents.
With the use of the Cs ϩ /TEA solutions, the current was found to be present in 75 of the 131 cells (57%) investigated, with a peak inward amplitude that varied between Ϫ70 and Ϫ1700 pA (mean of Ϫ595 Ϯ 49 pA). The current was voltagegated, activating at Ϫ50 mV and reaching a peak near Ϫ10 mV (Fig. 2, A (inset) and B), and had a mean reversal potential of ϩ68 Ϯ 4.2 mV (n ϭ 36), close to the calculated Nernst potential for Na ϩ of ϩ63 mV (E Na ) for these solutions at 22°C. The fast activation and inactivation kinetics of the inward current and its reversal close to E Na suggested that this was likely to be a Na ϩ current. To confirm this, extracellular Na ϩ was replaced with equimolar choline chloride, which totally abolished the inward current (n ϭ 6) (data not shown). These results indicate that the rapidly activating and inactivating inward current in HSVECs is a voltage-gated Na ϩ current, which we have designated as I Na .
Tetrodotoxin Sensitivity of the Sodium Current-To facilitate comparison of I Na in HSVECs with other known Na ϩ currents, we determined the sensitivity of the current to the guanidinium toxin, TTX. TTX has been shown to block a wide range of Na ϩ channels of different origins (7), which are classified as TTX-sensitive, TTX-resistant, or TTX-insensitive based on the  IC 50 value for TTX blockade. In HSVECs, as the concentration of TTX in the extracellular solution was increased, I Na decreased, with 100% blockade occurring at 30 M (Fig. 2). An equal reduction in the amplitude of I Na was observed across the entire voltage range of activation, suggesting that TTX binding was not voltage-dependent (Fig. 2B). A logistic plot fitted to the concentration inhibition curve yielded an IC 50 value for TTX of 4.7 M (Fig. 2C), suggesting that I Na belongs to the TTXresistant classification of Na ϩ channels (19).
Activation and Inactivation Kinetics of the Sodium Current-The steady-state activation and inactivation properties of I Na were assessed by the construction of normalized m ϱ and h ϱ curves. Fits to the normalized activation curves (m ϱ ; Fig. 3), generated by conversion of the I-V relationship to conductances, gave a mean half-activation voltage (V m ) of Ϫ29.2 Ϯ 1.2 mV (n ϭ 17) and an average K value (slope factor) of 7.2 Ϯ 0.3 mV. The steady-state inactivation parameter (h ϱ ) of I Na was measured using a standard two-pulse protocol. Cells, clamped at Ϫ120 mV, were conditioned with a 1-s prepulse to potentials between Ϫ120 and Ϫ20 mV prior to a 30-ms test pulse to Ϫ10 mV, the potential that routinely evoked the maximal current (Fig. 3A). These data show that significant inactivation was observed at Ϫ80 mV and that I Na was inactivated completely at Ϫ40 mV. h ϱ curves were calculated by normalizing the peak current recorded during the test pulse to the maximum current measured on stepping from Ϫ120 mV to the test potentials and were plotted as a function of the prepulse level. The averaged data indicate that I Na was half-inactivated at Ϫ75.4 Ϯ 1.3 mV (V h ; n ϭ 17), with a slope factor of 5.7 Ϯ 0.2 mV (Fig. 3B). There were not any areas of significant overlap between m and h curves. The time dependence of recovery from inactivation also was evaluated using a double-pulse protocol. Cells were stepped from a holding potential of Ϫ120 mV to 0 mV for 20 ms to elicit and inactivate I Na . The cells were then clamped at Ϫ120, Ϫ100, or Ϫ80 mV for a variable duration of between 4 and 64 ms in 4-ms increments, prior to a second test pulse to 0 mV (Fig. 4A). Recovery from inactivation was found to be strongly voltage-dependent, with complete recovery requiring potentials more negative than Ϫ80 mV (Fig. 4B). By fitting a single exponential function to the data, the recovery time constants were calculated to be 3.3 Ϯ 0.3, 9.4 Ϯ 1.2, and 32.2 Ϯ 1.7 ms for cells held at Ϫ120, Ϫ100, and Ϫ80 mV, respectively (n ϭ 4). Time constants for activation ( m ) and inactivation ( h ) were obtained by fitting a Hodgkin-Huxley model (see "Experimental Procedures") to the inward currents. Both m and h were voltage-dependent, becoming more rapid as the test potential became increasingly more depolarized (Fig. 5). This was much FIG. 3. Steady-state activation and inactivation of I Na in HS-VECs. A, I Na currents elicited at a test potential of 0 mV following a 500-ms hyperpolarizing conditioning prepulse to potentials between Ϫ120 and Ϫ20 mV. B, normalized activation (q) and inactivation curves (2) for I Na in HSVECs. Symbols represent the mean fractional current or conductance (calculated as detailed under "Data Analysis and Curve Fitting") at each potential, with the S.E. indicted by the error bars (n ϭ 22).

FIG. 4. Recovery from inactivation of I Na .
A, series of current traces elicited by a double-pulse protocol. The cell was clamped at Ϫ120 mV, and a 20-ms voltage step to 0 mV was used to elicit and inactivate I Na . The cell was then clamped at Ϫ100 mV for 4 -64 ms in 4-ms increments before a second test pulse to 0 mV. B, recovery from inactivation occurs as a function of the holding potential (V hold ). Plots are of the ratio of the amplitude of the second and first current pulses as a function of the interval between the two. The lines represent a single exponential curve fitted to the data. Time constants estimated from these fits were 3.6 ms at Ϫ120 mV, 9.7 ms at Ϫ100 mV, and 34.6 ms at Ϫ80 mV. more marked with h , which was reduced from 3.77 ms at Ϫ30 mV to 0.43 ms at ϩ50 mV. The hyperpolarized half-maximal inactivation potential (V h ) and the inactivation time course are consistent with those reported for the cardiac isoform of Na ϩ channels (7,20).
The addition of an antibody directed against the cytosolic linker region between domains III and IV of known Na ϩ channel ␣-subunits (anti-Na␣) to the pipette solution (10 g/ml) produced substantial slowing of Na ϩ current inactivation. In four cells, h at 0 mV was increased from 0.71 ms to 1.25 ms within 15 min of establishing whole-cell configuration without any effect on peak current amplitude (data not shown).
Sodium Channel hH1 Transcripts-On the basis of electrophysiological data, particularly the current kinetics and sensitivity to TTX, the voltage-gated Na ϩ current in HSVECs appeared to closely resemble the human cardiac sodium channel, hH1. To test this hypothesis, mRNA isolated from HSVECs was reverse-transcribed, and the resulting cDNA was amplified (RT-PCR) with specific primers targeted against the 3Јuntranslated region of the hH1 cDNA (16). The expected product of 180 base pairs was produced only in those samples that had been reverse-transcribed. In the absence of a RT step, no product was present after PCR (Fig. 6). DNA sequencing of the RT-PCR product confirmed that it was identical to the 3Јuntranslated region of the human hH1 sodium channel.
Immunochemical Detection of Sodium Channels in Saphenous Vein Endothelium-Using Western blotting and employing the anti-Na␣ antibody, we were able to detect Na ϩ channel protein in the membrane (but not the cytosolic) fraction of HSVECs with guinea pig cardiac myocytes acting as a positive control (Fig. 7). The antibody routinely recognized a single protein band that had an apparent molecular mass of 242 Ϯ 9 kDa when separated by SDS-polyacrylamide gel electrophoresis (n ϭ 5). This value is close to the molecular mass of the human hH1 ␣-subunit of 230 kDa as calculated from the deduced amino acid sequence.
Subconfluent HSVECs, which were electrophysiologically confirmed to be expressing I Na , also immunostained positively with the anti-Na␣ antibody (Fig. 8). This antibody was used to demonstrate the presence of Na ϩ channels in the endothelium of freshly excised human saphenous vein. In intact saphenous vein endothelium, the immunostaining for Na ϩ channel ␣-subunits was intermittent, with not all endothelial cells being stained (Fig. 9).
Serum Induction of Sodium Channels in HSVECs-To assess the effects of serum on expression of I Na by subconfluent HSVECs, we measured the magnitude and prevalence of I Na in cells that had been incubated for 48 h in growth medium that was serum-free or supplemented with 10% (v/v) human serum either from peripheral arterial disease patients 65 years of age and over ("aged") or from healthy donors under 30 years of age ("young"). Under serum-free conditions and in medium supplemented with aged serum, I Na was of small magnitude and was found in relatively few cells (Table I). However, medium supplemented with young serum was found to increase significantly both the magnitude of I Na and the number of HSVECs in which I Na was observed (Table I). The stimulatory effect of serum upon I Na expression was confirmed by cell-based enzyme-linked immunosorbent assay of confluent HSVECs using the anti-Na␣ antibody. Inclusion of 2 or 10% (v/v) young serum in the incubation medium of HSVECs for 6 h increased the relative concentration of sodium channel protein 2-and 4-fold, respectively. The absorbance increased from 0.10 Ϯ 0.03 (n ϭ 8) in serum-free medium to 0.19 Ϯ 0.03 (n ϭ 7) and 0.36 Ϯ 0.06 (n ϭ 6) for HSVECs cultured in 2 and 10% (v/v) young sera, respectively.

DISCUSSION
In this investigation, we have used a range of techniques to demonstrate the presence of voltage-gated sodium channels in human saphenous vein endothelium. First, immunohistochemistry, using an antibody directed against the conserved cytoplasmic region of the ␣-subunit, showed the presence of sodium channels in both intact saphenous vein endothelium and cultured HSVECs. Second, this same antibody recognized a 245-kDa protein in Western blot analysis of HSVEC membrane lysates. Third, whole-cell patch-clamp electrophysiology of HS- VECs showed the presence of fast inward voltage-gated sodium currents, which were TTX-resistant and showed similar kinetics to the human heart hH1 channel isoform. RT-PCR analysis also showed HSVECs to contain hH1 transcripts. The expression of this sodium channel in HSVECs was dependent on serum characteristics and concentration.
The expression of voltage-gated sodium channels in human saphenous vein endothelium was unexpected since this type of ion channel classically is associated with action potential generation in excitable cells. This is the first report of the presence of sodium channels in the endothelium of intact human vessels. There has been one previous electrophysiological study suggesting that cultured human endothelium from umbilical vein expressed sodium channels, but the subtype of sodium channel was not identified (21). A potential criticism of this latter study was that the expression of sodium channels was an artifact of placing the cells into culture since this phenomenon has been reported for human coronary myocytes (22).
The antibody used for demonstrating the presence of sodium channels in intact endothelium was also used in electrophysiological studies; when the antibody was applied intracellularly to HSVECs, there was substantial slowing of the current inactivation. The current kinetics and the TTX inhibition studies suggested that I Na in HSVECs closely resembles the principal TTX-resistant, voltage-gated sodium channel found in human  heart, hH1 (8,23). Electrophysiological and immunohistochemical analyses showed that the sodium channel was not present in every endothelial cell. However, the prevalence of sodium currents (57%) was similar to that of inwardly rectifying potassium currents (58%), the most widely distributed channel in endothelial cells (3). The data from cell-based enzyme-linked immunosorbent assays and electrophysiology suggest that the prevalence and expression of sodium channels in HSVECs are serum-dependent. Serum harvested from young healthy volunteers increased the magnitude of I Na 3-4-fold compared with serum from aged patients with peripheral arterial disease. The 2-3-fold increase in HSVEC sodium channels, when serum concentration was increased from 2% to 10%, was similar to the previously reported serum stimulation of the sodium channel in rat leiomyosarcoma cells (24).
Endothelial cells have never been reported to produce action potentials and are classed as non-excitable (3). In keeping with this tenet, the magnitude of I Na in HSVECs is small, with a mean peak current of Ϫ595 Ϯ 49 pA, and I Na requires a membrane potential more negative than Ϫ80 mV to remove inactivation completely (Fig. 4B) when the resting membrane potential (E m ) of cultured HSVECs is around Ϫ30 mV. This would imply that I Na normally would be inactivated and dysfunctional. However, in vivo, it is probable that endothelial cells are more hyperpolarized (E m more negative): the E m of endothelial cells on intact saphenous vein is nearer to Ϫ70 mV. 2 Stimuli that are known to hyperpolarize vascular endothelial cells, such as hemodynamic shear stress, could produce potentials that are sufficiently negative to lead to a partial recovery of I Na from inactivation. The inwardly rectifying K ϩ channel is the predominant channel open at rest in endothelial cells and tends to hold E m close to the potassium equilibrium potential (E K ). As these channels do not pass much repolarizing current due to their poor outward rectification characteristics, they would permit relatively small inward currents carried by other ions to depolarize the endothelial cell. Therefore, even a small magnitude I Na may be able to elicit substantial and rapid membrane depolarization. However, as HSVECs lack the outward potassium currents (delayed rectifier currents) necessary to rapidly repolarize the cells, it is unlikely that these cells could elicit repetitive action potentials. Similar findings have been reported in glial cells, which are also considered to be inexcitable, yet express Na ϩ channels (25).
There are at least two possible physiological functions for voltage-gated sodium channels in vascular endothelium, given that they are unlikely to be involved in action potential generation. First, I Na could have a role in the regulation of intracellular calcium levels ([Ca 2ϩ ] i ). This could occur by several mechanisms. An increase in Na ϩ influx would stimulate Na ϩ /Ca 2ϩ exchange and thus raise [Ca 2ϩ ] i (26). It also has been reported that voltage-dependent Na ϩ channel gating is involved in depolarization-induced activation of G-proteins, a process that could lead to Ca 2ϩ mobilization (27). Also, some capillary endothelial cells have been reported to possess a voltage-dependent, BAY K8644-sensitive Ca 2ϩ current (28,29); thus, I Na could provide the depolarizing stimulus leading to opening of these channels. However, these Ca 2ϩ channels have yet to be described in large vessel endothelium. Second, the electrical coupling between vascular endothelial cells, as well as coupling between endothelial cells and smooth muscle cells (30), raises the possibility that an electrical message, such as depolarization, could be conveyed electrotonically by the endothelium. This process also may participate in regulating [Ca 2ϩ ] i , as it has been shown in capillary endothelium that some cells possess a "pacemaker" function and pass an undetermined message via gap junctions to other cells to initiate Ca 2ϩ oscillations (31). This is the first report of the presence of sodium channels of the hH1 isoform in human vascular endothelium. The regulation and distribution of this sodium channel are the focus of current investigations to assess the role of this channel in endothelial homeostasis.