Heart and Circulatory Physiology

Nanomolar ouabain increases NCX1 expression and enhances Ca2+ signaling in human arterial myocytes: a mechanism that links salt to increased vascular resistance?

Cristina I. Linde, Laura K. Antos, Vera A. Golovina, Mordecai P. Blaustein


The mechanisms by which NaCl raises blood pressure (BP) in hypertension are unresolved, but much evidence indicates that endogenous ouabain is involved. In rodents, arterial smooth muscle cell (ASMC) Na+ pumps with an α2-catalytic subunit (ouabain EC50 ≤1.0 nM) are crucial for some hypertension models, even though ≈80% of ASMC Na+ pumps have an α1-subunit (ouabain EC50 ≈ 5 μM). Human α1-Na+ pumps, however, have high ouabain affinity (EC50 ≈ 10–20 nM). We used immunoblotting, immunocytochemistry, and Ca2+ imaging (fura-2) to examine the expression, distribution, and function of Na+ pump α-subunit isoforms in human arteries and primary cultured human ASMCs (hASMCs). hASMCs express α1- and α2-Na+ pumps. Further, α2-, but not α1-, pumps are confined to plasma membrane microdomains adjacent to sarcoplasmic reticulum (SR), where they colocalize with Na/Ca exchanger-1 (NCX1) and C-type transient receptor potential-6 (receptor-operated channels, ROCs). Prolonged inhibition (72 h) with 100 nM ouabain (blocks nearly all α1- and α2-pumps) was toxic to most cultured hASMCs. Treatment with 10 nM ouabain (72 h), however, increased NCX1 and sarco(endo)plasmic reticulum Ca2+-ATPase expression and augmented ATP (10 μM)-induced SR Ca2+ release in 0 Ca2+, ouabain-free media, and Ca2+ influx after external Ca2+ restoration. The latter was likely mediated primarily by ROCs and store-operated Ca2+ channels. These hASMC protein expression and Ca2+ signaling changes are comparable with previous observations on myocytes isolated from arteries of many rat hypertension models. We conclude that the same structurally and functionally coupled mechanisms (α2-Na+ pumps, NCX1, ROCs, and the SR) regulate Ca2+ homeostasis and signaling in hASMCs and rodent ASMCs. These ouabain/endogenous ouabain-modulated mechanisms underlie the whole body autoregulation associated with increased vascular resistance and elevation of BP in human, salt-sensitive hypertension.

  • human artery
  • α2-Na+ pumps
  • ouabain
  • Na/Ca exchanger-1
  • sarco(endo)plasmic reticulum Ca2+-ATPase
  • receptor-operated channels

the role of salt in the pathogenesis of human essential hypertension is widely accepted, but the specific mechanisms by which salt elevates blood pressure (BP) are still poorly understood (28). Numerous studies link hypertension in rodents to elevated plasma endogenous ouabain (EO), an adrenocortical hormone (6, 24, 30), and to Na+ pumps with high affinity for ouabain (12, 38, 61). Increased expression and function of arterial myocyte Na/Ca exchanger-1 (NCX1) and cation channels with C-type transient receptor potential (TRPC) subunits are also involved (7, 17, 25, 34, 49, 66). We have speculated that similar mechanisms are involved in human hypertension (6, 7). This view is based, in part, upon the evidence that plasma EO levels are elevated in ≈45% of patients with essential hypertension, and in a majority of patients with aldosterone-producing adenomas, i.e., those with a high level of the salt-retaining hormone aldosterone (40, 45, 51).

Na+ pumps are αβ-dimers. The α (catalytic)-subunit contains the ATP hydrolytic activity, the ion transport machinery, and the ouabain binding site (5). Most mammalian cells express two types of Na+ pumps: some with an α1-subunit (which usually predominate) and others with an α2- or α3-subunit. The pumps with an α1-subunit are apparently housekeepers: they maintain the low Na+ and high K+ concentrations in bulk cytosol (20).

Rodent arterial smooth muscle cells (ASMCs) express α1- and α2-, but not α3-, Na+ pumps (39, 53, 62). Rodent α1-Na+ pumps are unusually ouabain resistant (EC50 ∼5 μM), whereas human α1 has much higher ouabain affinity (EC50 <20 nM), reportedly similar to that of all α2- and α3-Na+ pumps (5, 44, 57). Indeed, nanomolar ouabain increases myogenic constriction in isolated rodent small arteries (50, 62, 65). Prolonged (2–4 wk) infusion of low-dose ouabain induces hypertension in rodents (14, 41), and brief infusion elevates BP in humans (42, 52).

Mice with mutant, ouabain-resistant, but otherwise normally functional, α2-Na+ pumps (13) are resistant to ouabain-induced (14) and ACTH-induced forms of hypertension (12, 38). These effects are likely mediated by ASMC α2-Na+ pumps because smooth muscle-specific reduction of α2-Na+ pump expression causes BP elevation (9, 62). In human hypertension, it is not clear whether ASMC α2-Na+ pumps play a similar role because of the high ouabain affinity of human α1 (57).

In rodents, ≈80% of the ASMC Na+ pumps have a α1-subunit; the remainder have a α2-subunit (53, 62). The α2-pumps cluster with NCX1 and receptor- and store-operated cation channels (ROCs and SOCs, respectively) in plasma membrane (PM) microdomains at PM-SR (sarcoplasmic reticulum) junctions, “plasmerosomes” (2, 26, 49, 66). In contrast, α1-Na+ pumps are widely distributed in the PM, but may be excluded from these PM microdomains (26, 27, 39, 54). Thus NCX, ROCs, and SOCs are structurally and functionally coupled to α2-, but not α1-, Na+ pumps in rodent ASMCs (8, 31, 46, 49).

In numerous rodent hypertension models, NCX1 and TRPC6 or TRPC3 proteins (component of some ROCs) in ASMCs are markedly upregulated, and Ca2+ signaling is augmented (7, 17, 35, 49, 55, 66). Indeed, prolonged treatment of primary cultured rat ASMCs, or of normal rats, in vivo, with low-dose ouabain greatly increases arterial myocyte NCX1 and TRPC6 protein expression and Ca2+ signaling mediated by NCX1 and ROCs (49). These changes in the ASMCs are associated with the increased myogenic reactivity of small arteries, as exemplified by arteries isolated from ouabain-hypertensive rats (60) and Milan hypertensive strain rats (34, 66). Moreover, Milan hypertensive rats have elevated plasma EO levels (15, 16). The implication is that, in rodents, the altered protein expression and enhanced Ca2+ signaling in ASMCs, and the elevated BP, are all consequences of the action of EO on α2-Na+ pumps.

Here, we test whether similar mechanisms operate in humans. Specifically, we examine the expression and distribution of the Na+ pump α-subunit isoforms in human ASMCs (hASMCs). We also determine the effects of prolonged nanomolar ouabain treatment on the expression of some Ca2+ transporters, and on Ca2+ signaling in primary cultured hASMCs. The ouabain-induced changes in protein expression and Ca2+ signaling in hASMCs are comparable with results obtained in rodents. This indicates that the rodents are good models for the study of these parameters and that results from the rodent studies are applicable to human arterial smooth muscle. Thus information about the linkage between salt metabolism and enhanced vascular resistance [whole body autoregulation (22)] obtained from rodent hypertension models should be directly relevant to human salt-dependent hypertension.


Human arteries and rodent tissues.

Unused internal thoracic artery segments from coronary bypass patients were used for the experiments in Fig. 1. Un-needed segments of distal superior mesenteric artery, adjacent to the intestinal wall (3rd or 4th order branches), from 11 brain-dead tissue donors were used for all the other studies reported here. The patients were anonymous, but included African Americans, Hispanics, and Caucasians and both men and women, ages 14–68, with a variety of diagnoses. The arteries were obtained with informed consent. All procedures were approved by the University of Maryland School of Medicine (UMSOM) Institutional Review Board.

Fig. 1.

Identification of Na+ pump α-subunit isoforms and Na/Ca exchanger-1 (NCX1) in human arterial smooth muscle cells (hASMCs) and mouse (m) tissues. A: both mouse aorta smooth muscle and primary cultured internal thoracic hASMCs cross-react with anti-HERED (α2) antibodies. hASMCs also cross-react with anti-TSEP (human α1) but not anti-NASE (rodent α1) antibodies; mouse aorta smooth muscle cross-reacts with anti-NASE, but not anti-TSEP. B: mouse brain, but neither hASMCs nor mouse aorta smooth muscle, cross-reacts with anti-TED (α3) antibodies. C: both hASMCs and mouse aorta smooth muscle cross-react with R3F1 (anti-NCX1) antibodies. Anti-β-actin served as a loading control. The expected molecular weights are as follows: α1, 110 kDa; α2, 110 kDa; α3, 110 kDa; NCX1, 116 kDa; and β-actin, 42 kDa. Each panel is representative of replicates from at least 3 different patients or mouse preparations.

In a few experiments, aortae, mesenteric arteries, and whole brains from 25 to 30 g male C57BL/6 mice were harvested. Membrane proteins, prepared from the brains and de-endothelialized arteries, were used for immunoblots. Some unpublished images of primary cultured rat mesenteric artery myocytes, from a study on the effects of ouabain (49), are also shown in results. The procedures were approved by UMSOM Institutional Animal Care and Use Committee.

Dissociation and primary culture of hASMCs.

Human artery segments were placed in ice-cold, sterile HBSS (Gibco, Grand Island, NY), and the fat and connective tissue were dissected away. Each artery segment was then cut lengthwise, and the endothelium was removed by rubbing the lumenal surface with a sterile cotton-tip applicator. Artery segments that were to be used for immunoblotting were then frozen in liquid nitrogen and stored at −80°C until needed (see Immunoblot analysis of membrane proteins).

Primary cultured hASMCs were prepared using a modification of a published method (19). The cleaned, de-endothelialized artery segments were digested for ∼50 min at 37°C (5% CO2-95% air) in Ca2+- and Mg+-free HBSS [HBSS(2); Gibco] containing collagenase type 2 (1 mg/ml). The tissue was washed, the adventitia was stripped off, and segments of muscularis were stored overnight in smooth muscle cell growth medium at 37°C (Cell Applications, San Diego, CA). The next morning, the segments were cut into ∼1- to 2-mm2 pieces and digested for 45–50 min at 37°C (5% CO2-95% air) in HBSS(2) containing collagenase type 2 (1 mg/ml) and elastase type IV (0.5 mg/ml). The digested tissue was washed in 2 ml HBSS(2) containing 4 to 5 drops of culture medium, transferred to 2–5 ml of smooth muscle cell growth medium, and gently triturated with a flame-polished Pasteur pipette. The dissociated cells were centrifuged (1,000 g, 4°C, 5 min), resuspended in growth medium, and plated on either 25-mm coverslips in 6-well plates for immunocytochemistry or Ca2+ imaging, or on 10-cm culture dishes for immunoblotting. The plated cells were maintained in growth medium under a humidified atmosphere of 5% CO2-95% air at 37°C. The medium was changed on days 4 and 7. In cultures in which the effects of ouabain were tested, the standard growth medium was replaced by serum-free growth medium (Cell Applications) 24 h before ouabain was added. Experiments were performed on subconfluent cultures on days 8 and 9 in vitro if not indicated otherwise. Cells were confirmed as hASMCs by immunocytochemistry through cross-reactivity with antibodies directed against smooth muscle α-actin; a cell purity of at least 95–98% was routinely observed.

Immunoblot analysis of membrane proteins.

Cultured hASMCs were harvested in PBS supplemented with protease inhibitor cocktail tablets (Roche Applied Science, Indianapolis, IN). The pellets were resuspended in lysis buffer containing 145 mM NaCl, 10 mM NaH2PO4, 10 mM NaN3, and 1% IGEPAL supplemented with protease inhibitor cocktail tablets. The suspension was centrifuged (5,000 g, 4°C, 30 min). The supernate containing the extracted membrane proteins was mixed with SDS buffer containing 5% 2-mercaptoethanol, and the proteins were separated by SDS-PAGE as described (4).

Some immunoblots were performed on membranes prepared from segments of freshly harvested, cleaned, and frozen human arteries. The frozen artery segments were pulverized with a stainless steel mortar and pestle, and membrane proteins were solubilized and separated by the same methods used for cultured hASMCs.

Mouse arteries and brains were used to compare Na+ pump α-subunit isoform expression in human and mouse tissues (Fig. 1). Details of the methods used for the preparation of membranes from fresh rodent tissues are published (32).

Immunoblot analysis used published methods (20, 32). In brief, the separated proteins on the SDS-PAGE gel were blotted onto a polyvinylidene difluoride membrane, which was then cross-reacted with appropriate primary rabbit polyclonal antibodies (pAbs) and/or mouse monoclonal antibodies (mAbs). After wash, membranes were incubated with secondary anti-rabbit or anti-mouse, horseradish peroxidase-conjugated, IgG (Jackson ImmunoResearch, West Grove, PA) for 1 h at room temperature. The immune complexes on the membranes were detected by enhanced chemiluminescence-plus (Amersham BioSciences, Piscattaway, NJ) and exposure to X-ray film (Eastman Kodak, Rochester, NY). Quantitative analysis of immunoblots was performed with a Kodak DC120 digital camera and one-dimensional Image Analysis Software (Eastman Kodak).

A number of primary antibodies were used for the immunoblot and immunocytochemistry studies. The pAbs directed to the peptide sequences specific to the nonhuman α1 (NASE)-, and all mammalian α3 (TED)-, Na+ pump isoforms were gifts from Dr. T. A. Pressley [Texas Tech University (47); mAb anti-NCX1 antibody (R3F1) was a gift from Dr. K.D. Philipson (University of California, Los Angeles)]. In collaboration with Dr. Pressley, we generated and tested a rabbit polyclonal antiserum directed against the human Na+ pump α1-isoform-specific peptide sequence, TSEP (47). Rabbit pAb directed against the Na+ pump α2-specific epitope (HERED) was purchased from Millipore (Billerica, MA); anti-TRPC6 pAb was obtained from Sigma-Aldrich (PRS3897, St. Louis, MO). The following mAbs were purchased: anti-sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA2; SR Ca2+ pump), anti-PMCA (PM Ca2+ pump), and anti-IP3R1 (inositol triphosphate receptor-1) (all from Affinity BioReagents, ABR, Golden, CO); anti-β-actin (Sigma-Aldrich); and anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Abcam, Cambridge, MA). The anti-GAPDH and anti-β-actin mAbs were used as immunoblot loading controls. Antibody dilutions for immunoblots were anti-NASE and anti-TSEP, 1:2,000; anti-HERED, 1:750; anti-TED, 1:750; anti-TRPC6, 1:500; anti-SERCA2, 1:1,500; anti-PMCA, 1:1,000; anti-IP3R1, 1:750; anti-GAPDH, 1:5,000; and anti-β-actin, 1:10,000.


Primary cultured hASMCs were immunolabeled, as described for rodent cells (2, 66). In a few experiments, immunocytochemistry was performed on freshly dissociated hASMCs, which were prepared using rodent ASMC protocol no. 1 (3), but with enzymatic digestion times of 50–65 min depending upon the size of the human artery segment. The freshly dissociated cells were plated onto coverslips and permitted to settle by incubation for 4 h at 4°C. Both the cultured and the freshly dissociated cells, on coverslips, were fixed in CF fixative: 0.45% (wt/vol) formaldehyde, 75 mM cyclohexylamine, 75 mM NaCl, 10 mM EGTA, 10 mM MgCl2, and 10 mM PIPES. After fixation, the cells were permeabilized in CF fixative containing 0.5% polyoxyethylene 20 cetyl ether (Brij 58) and were then incubated overnight in antibody buffer containing one or more of the following antibodies (see above): anti-TSEP, anti-HERED, anti-NCX1, anti-SERCA2, anti-PMCA, and/or anti-TRPC6. All antibody dilutions for immunocytochemistry were 1:10. Fluorescein isothiocyanate-labeled donkey anti-mouse IgG or Cy3-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch) was used to visualize the primary antibodies. The fluorescence from the secondary antibody in the absence of primary antibody (positive control) did not exceed 2% to 3% of the fluorescence in cells cross-reacted with primary antibodies (see results). To identify SR in arterial myocytes, cells were treated (5 min) with 1 μM ER tracker (Invitrogen Detection Technologies, Eugene, OR). Images were obtained with either a laser scanning confocal microscope [Zeiss 510 Meta with a Zeiss Plan-Apo 63x, 1.4 numerical aperture (NA) oil immersion objective; Carl Zeiss Microscopy, Thornwood, NY] or a widefield microscope (Zeiss Axiovert 100 with a Zeiss Apochromat 63x, 1.2 NA water immersion objective).

Cytosolic Ca2+ signal measurements.

The cytosolic Ca2+ concentration, [Ca2+]CYT was measured with fura-2 by using digital imaging as described (4). Primary cultured hASMCs were loaded with fura-2 by incubation for 40 min in smooth muscle cell growth medium containing 3.3 μM fura-2-AM (20–22°C, 5% CO2-95% air). After dye loading, the coverslips were transferred to a tissue chamber mounted on a microscope stage, where the cells were superfused for 15–20 min (35° to 36°C) with physiological salt solution (PSS) to wash away extracellular dye. The PSS contained (in mM) 140 NaCl, 5.0 KCl, 1.2 NaH2PO4, 5 NaHCO3, 1.4 MgCl2, 1.8 CaCl2, 11.5 glucose, and 10 HEPES (pH 7.4). In Ca2+-free PSS, the 1.8 mM CaCl2 was omitted and 50 μM EGTA was added to chelate residual Ca2+. Cells were studied for 40–50 min during continuous superfusion with PSS (35°C).

The imaging system was based on a Zeiss Axiovert 100 microscope with a Zeiss F-Fluor 40×, 1.3 NA oil immersion objective. The dye-loaded cells were illuminated with a diffraction grating-based system (Polychrome V, TILL Photonics, Munich, Germany). Fluorescent images were recorded with a CoolSnap HQ2 CCD camera (Photometrics, Tucson, AZ). Image acquisition and analysis were performed with a MetaFluor/MetaMorph Imaging System (Molecular Devices, Downingtown, PA). [Ca2+]CYT was calculated by determining the ratio of fura-2 fluorescent emission (510 nm) excited at 380 and 360 nm as described (4, 18, 21). Intracellular fura-2 was calibrated in situ in primary cultured hASMCs as described (4).


Fura 2-AM was obtained from Molecular Probes (Invitrogen Detection Technologies). Collagenase (type 2) was purchased from Worthington Biochemical (Freehold, NJ). Unless otherwise indicated, all other reagents were purchased from Sigma-Aldrich and were analytic grade or the highest purity available.

Statistical analyses.

The numerical data presented in results are means ± SE from n single cells (1 value per cell). Immunoblots were repeated at least three times for each protein. The number of arteries (patients) is presented where appropriate. Data from at least two patients were obtained for all protocols; in almost all cases, arteries from at least three or four patients were subjected to the same protocol. Statistical significance was determined using Student's paired or unpaired t-test, as appropriate. P < 0.05 was considered significant.


Expression of Na+ pump α-subunit isoforms and several Ca2+ transporters in human artery smooth muscle.

A phylogenetic analysis of Na+ pump α-subunit isoforms (47) revealed that residues 489–499 [numbering based on the rat α1-peptide (5)] in the large cytoplasmic loop between transmembrane helices 4 and 5 contain isoform-specific, conserved peptide sequences. For example, the α2-specific sequence, HERED, is conserved in all mammal α2-Na+ pumps, and the α3-specific sequence, TED, is conserved in all vertebrate and some invertebrate α3-pumps. The α1-isoform is more variable: human α1-isoform lacks one α1-specific peptide sequence, NASE, which is conserved in rodents and many other mammals, including dogs, guinea pigs, and opossums. Instead, human α1 and horse α1 contain the peptide sequence, TSEP, which also is not found in α2 or α3 (47).

The immunoblots in Fig. 1A verify that mouse aortae express α1-Na+ pumps with a NASE, but not a TSEP, epitope, as well as α2-Na+ pumps. hASMCs also express α1- and α2-Na+ pumps, but human α1-Na+ pumps cross-react with antibodies raised against the TSEP epitope. Mouse brain (and human neuroblastoma cells, not shown), but neither mouse nor human arteries, expresses α3-Na+ pumps (Fig. 1B). Mouse arteries and hASMCs both express NCX1 (Fig. 1C) and TRPC6 (see next section). Smooth muscle in freshly harvested human mesenteric arteries, too, expresses α1- and α2-Na+ pumps, NCX1, and TRPC6 (Fig. 2), as do mouse mesenteric arteries (62, 63).

Fig. 2.

Expression of α1- and α2-Na+ pumps, NCX1, and C-type transient receptor potential (TRPC6) in smooth muscle from freshly harvested human mesenteric small arteries. Blots were probed with (top to bottom, as indicated) anti-HERED [polyclonal antibodies (pAb)], anti-TSEP (human α1 pAb), and R3F1 [anti-NCX1 monoclonal antibodies (mAb)]; and anti-TRPC6 (pAb). Numbered lanes refer to mesenteric arteries (3rd or 4th order branches) from 4 different patients. Anti-β-actin served as the loading controls for all the blots. These data are representative examples from a total of 10 different patients (7 male, 3 female; 9 Caucasian, 1 Hispanic; age range = 21–63 years; 6 of the 10 patients died from trauma or suicide; no patients were hypertensive).

Two SR Ca2+ transporter proteins examined, SERCA2 and IP3R1, were both expressed in primary cultured hASMCs as well as in freshly harvested arterial smooth muscle (Fig. 3). The expression levels of all the proteins tested, when normalized to GAPDH or ß-actin expression, varied somewhat from patient to patient, in both the fresh tissues and the cultured cells. Expression of NCX1 and SERCA2, relative to β-actin, was greater in actively proliferating cells in media containing FBS than in growth-arrested cultures in which the FBS had been removed (Fig. 3, A and C). An apparent increase in TRPC6 molecular weight was observed in hASMCs incubated in serum-containing growth medium, compared with the TRPC6 from freshly harvested arteries (Fig. 3B).

Fig. 3.

Expression of some Ca2+ transporters in hASMC. The data are representative blots of proteins from freshly harvested human mesenteric artery distal branches (Fresh; left lane) and primary cultured hASMCs (Cult; right lane). A: NCX1. B: TRPC6. C: sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA2). D: inositol triphosphate receptor-1 (IP3R1). Anti-β-actin served as the loading control. Data are representative of duplicate blots of proteins from each of 2 or (NCX1) 3 patients.

It is important that data obtained in primary cultured cells from a single artery under different incubation conditions were reproducible. Thus data from hASMCs from different individuals, when subjected to the same experimental protocols, could be reliably compared. This is exemplified by the effects of prolonged incubation with low-dose ouabain (see Effects of prolonged incubation with nanomolar ouabain on Ca2+ transporter expression and Ca2+ signaling in hASMCs).

Localization of α1- and α2-Na+ pumps in hASMCs, and association of α2-pumps with NCX1 and TRPC6, and with the SR.

In rodents, high ouabain affinity α2-Na+ pumps cluster with NCX1 and ROCs in PM microdomains that overlie closely apposed, junctional SR. In contrast, low ouabain affinity α1-Na+ pumps, although more widely distributed in the PM, may be excluded from the junctional PM microdomains (31). Because human α1-Na+ pumps have high ouabain affinity, it is reasonable to ask whether they have a similar distribution to α2 in hASMCs. Figures 4–6 show that this is not the case. For example, as in rodents, α1-Na+ pumps are widely distributed in the PM; they do not cluster in a reticular pattern, nor do they selectively colocalize with the SR (stained with ER tracker; Fig. 4). PMCA also is diffusely distributed in the PM; it, too, does not distribute in a reticular pattern and does not colocalize with the SR (Fig. 4).

Fig. 4.

Distribution of α1-Na+ pumps, PMCA, and the sarcoplasmic reticulum (SR) in primary cultured mesenteric hASMCs. A: images of a hASMC that was cross-reacted with anti-TSEP human α1 pAb (a) and anti-PMCA mAb (b). The SR was then stained with ER tracker to label the SR (c). d: image (control) of a cell cross-reacted only with anti-rabbit secondary polyclonal antibody labeled with FITC; similar results were obtained with cells labeled with anti-mouse secondary monoclonal antibody labeled with Cy3. Scale bars = 25 μm. N, nucleus. B and C: higher magnification images from the boxed areas in the images in A. B includes the α1 (a; red) and SR (c; green) images and the image overlay (d); b, black-and-white ER tracker (SR) image. C includes the PMCA (a; green) and SR (b; red) images and the image overlay (c). Note that the α1 (TSEP) and PMCA staining patterns both differ markedly from that of ER tracker (an SR stain); this is reflected by the paucity of yellow-orange staining in the overlays.

In contrast, human α2-Na+ pumps, like rodent α2-Na+ pumps, do cluster in reticular patterns and do colocalize with ER tracker-stained SR and with NCX1 (Fig. 5A). Moreover, as illustrated by the confocal images of a freshly dissociated human myocyte (Fig. 6), α2-Na+ pumps colocalize with the adjacent SERCA2 at the cell surface, but only SERCA2 is found in the cell interior. These findings suggest that hASMC α2-Na+ pumps may have a specialized function similar to that in rodents, where these pumps help to regulate Ca2+ signaling by modulating NCX1-mediated Ca2+ transport. TRPC6, too, colocalizes with NCX1 and ER tracker-stained SR in hASMCs (Fig. 5B). This is consistent with a functional linkage of ROCs to NCX1 and α2-Na+ pumps in rodents (34, 46, 49).

Fig. 5.

Distribution of α2-Na+ pumps, NCX1, and SR in primary cultured mesenteric hASMCs. A: hASMCs were cross-reacted with anti-HERED α2 pAb and R3F1 anti-NCX1 mAb; the SR was then stained with ER tracker, as indicated by the labels. Insets: enlargements of the boxed areas. Pseudocolor images of the enlarged α2 (red) and NCX1 (green) regions, and the overlay, are shown at right. B: hASMCs were cross-reacted with R3F1 anti-NCX1 mAb and anti-TRPC6 pAb; the SR was then stained with ER tracker, as indicated. N, nuclei. Insets: enlargements of the boxed areas. Pseudocolor images of the enlarged NCX1 (green) and TRPC6 (red) regions, and the overlay, are shown at right. In A and B, the patterns of staining by both antibodies were very similar to the pattern of ER tracker (i.e., SR) distribution. Scale bars = 30 μm. Note that the α2, NCX1, and TRPC6 staining patterns are all very similar to that of ER tracker. This is reflected by the yellow-orange staining in the A and B overlay panels and indicates that α2-Na+ pumps and NCX1 colocalize and overlie elements of SR.

Fig. 6.

Distribution of α2-Na+ pumps and SERCA2 in a dissociated myocyte from a freshly harvested human internal thoracic artery. Myocytes were cross-reacted with anti-SERCA2 (SERCA2 mAb; A) and with anti-HERED (α2 pAb; B). The confocal images of a freshly dissociated hASMC are shown at left, and enlarged images of the boxed areas are shown at right. C: image overlays at low (left) and high magnification. Scale bars = 5 μm. Note that the α2 and SERCA2 staining patterns are very similar near the surface of the cell where the SR comes close to the plasma membrane (at right of low-power images), whereas only the SERCA2 mAb stains the interior of cell (at left of low-power images). This colocalization is reflected by the yellow-orange staining in the overlay panels; it indicates that the plasma membrane microdomains containing the α2-Na+ pumps overlie junctional elements of SR.

Effects of prolonged incubation with low-dose ouabain on the viability and morphology of hASMCs.

Under control conditions, most primary cultured hASMCs had a fusiform appearance (Fig. 7A), as expected for human cells cultured in serum-free, growth arresting medium (23, 33). The same was true for hASMCs incubated with 10 nM ouabain in serum-free medium for 72 h, although a few cells were rounded up (arrowheads in Fig. 7B). When the ouabain concentration was increased to 100 nM, however, many of the hASMCs died and detached from the coverslips. After 72 h, cell clusters were sparse and contained very few fusiform cells; most of the remaining cells were rounded up, with only long, thin processes extending from the nuclear region (Fig. 7C). This is not surprising because the KD of human α1 Na+ pumps for ouabain is <20 nM (57). A primary function of these very prevalent (housekeeper) pumps is to maintain the normally low-cytosolic Na+ and high-cytosolic K+ concentrations, [Na+]CYT and [K+]CYT, respectively (5, 20). Thus the inhibition by ouabain should cause the cells to gain Na+ and water (by osmosis) and to lose K+ and depolarize, thereby leading to cell death. In contrast, the morphology of primary cultured rat mesenteric artery myocytes is unaffected by a 72-h incubation with 100 nM ouabain (Fig. 7, D and E), presumably because the EC50 of rodent α1-Na+ pumps is >5 μM ouabain (44).

Fig. 7.

Effects of prolonged incubation with 10 or 100 nM ouabain on the morphology of primary cultured human and rat mesenteric small artery myocytes. Morphology of hASMCs were cultured for 10 days in growth medium (A), or in growth medium containing 10 nM (B) or 100 nM (C) ouabain for the final 72 h of culture. In all cultures, the FBS-supplemented medium was replaced by FBS-free medium 24 h before the ouabain was added. After a 72-h incubation without or with ouabain, the cells were loaded with fura-2; the fura-2 images (360 nm excitation) are shown. Arrowheads in B and C point to some of the rounded-up cells; none are seen in A. For comparison, rat mesenteric artery myocytes cultured for 72 h without (D) or with (E) 100 nM ouabain are also shown; for details, see Ref. 49. Scale bars = 50 μm.

Effects of prolonged incubation with nanomolar ouabain on Ca2+ transporter expression and Ca2+ signaling in hASMCs.

Treatment of cultured rat ASMCs with 100 nM ouabain for 48–72 h greatly enhances the expression and function of NCX1 and TRPC6 proteins (49). Immunoblots of hASMCs (Fig. 8) revealed that 72-h treatment with 10 nM ouabain enhanced the expression of NCX1 by 45 ± 10% (n = 5 patients) and SERCA2 by 32 ± 6% (n = 5 patients). TRPC6 was marginally, but not significantly, increased in the hASMCs, and IP3R1 was not affected (Fig. 8).

Fig. 8.

Effects of 72-h incubation with nanomolar ouabain on the expression of NCX1, TRPC6, SERCA2, and IP3R1 in primary cultured hASMCs. Ouabain, 5 nM (A only) or 10 nM (A-D), was added to the culture medium of some cells, and the cells were maintained in culture for 72 h. The data are representative immunoblots (top) and averaged data from 3 to 5 patients (bottom) probed with anti-NCX1 (A), anti-TRPC6 (B), anti-SERCA2 (C), and anti-IP3R1 (D) antibodies. Protein expression was normalized to GAPDH (A, C, D) or β-actin (B) expression (the 2 loading control proteins were used interchangeably because they gave similar results; also see Figs. 1–3). The blots from each patient were run in duplicate or triplicate. NCX1 and SERCA2 were significantly upregulated by 10 nM ouabain; *P < 0.05 vs. control (no ouabain).

These changes in Ca2+ transporter expression were reflected in altered Ca2+ handling by ouabain treatment. The [Ca2+]CYT in the quiescent (resting) cells was not significantly affected by the prior ouabain treatment (Table 1), but agonist-evoked Ca2+ signaling was significantly enhanced even though ouabain was no longer present. Figure 9A illustrates the protocol: it shows the time course of changes in [Ca2+]CYT in two representative hASMCs from a single patient. One hASMC (red line) was from a coverslip treated with 10 nM ouabain for 72 h; the other hASMC (blue line) was from a control coverslip. The fura-2-loaded cells were first exposed to Ca2+-free medium, and 10 μM ATP was then added to activate G protein-coupled purinergic receptors and promote Ca2+ release from the SR (the initial, transient rise in [Ca2+]CYT). Then, after recovery of [Ca2+]CYT to the original basal level, and with ATP still present, extracellular Ca2+ was restored. This induced a secondary rise in [Ca2+]CYT due to Ca2+ entry from the extracellular fluid , presumably mediated, at least in part, by SOCs and ROCs (49). The return to basal [Ca2+]CYT was much slower in this case, but ATP removal before full recovery caused a rapid return to baseline. The summarized data on the cultured hASMCs from three patients are shown in the bar graphs (Fig. 9, B-D, respectively). The 10 nM ouabain treatment increased the SR Ca2+ release (indicated by the initial rise in [Ca2+]CYT in Fig. 9A) by 27 ± 8% and augmented the Ca2+ entry from the extracellular fluid by 96 ± 15%, on average.

View this table:
Table 1.

Resting [Ca2+]CYT in control and ouabain-treated (72 h) primary cultured human mesenteric artery myocytes

Fig. 9.

Effects of 72-h incubation with 10 nM ouabain on Ca2+ signaling in primary cultured hASMCs. A: changes in cytosolic Ca2+ concentration ([Ca2+]CYT) in a representative cell from each of 2 coverslips of hASMCs, 1 treated with 10 nM ouabain (red line) and 1 untreated control (blue line), both from the same patient. As indicated by the bars below the data records, normal physiological salt solution (PSS) was replaced by Ca2+-free PSS 2.5 min before 10 μM ATP was introduced. Then, after an initial, transient rise in [Ca2+]CYT and recovery to the initial basal level, external Ca2+ was restored with ATP still present; this induced a second, more prolonged [Ca2+]CYT transient. Finally, the ATP was removed, usually before the [Ca2+]CYT had returned to the original basal level. None of the solutions used for Ca2+ imaging contained ouabain. B-D: averaged data for hASMCs from 1 mesenteric small artery from each of 3 patients; data in B are from the same patient as the data in A. The bars indicate the mean peak amplitudes ± SE of the first (Ca2+ release) and second (Ca2+ influx) Ca2+ transients, measured from the basal (resting) [Ca2+]CYT (Δ[Ca2+]CYT in nM). Blue bars, control cells; red bars, ouabain-treated cells. *P < 0.05 and **P < 0.01 vs. untreated controls. The basal [Ca2+]CYT levels and numbers of cells (n) in which basal Ca2+ and Ca2+ release were measured are presented in Table 1 (patients 1–3). The expression of α1- and α2-Na+ pumps, NCX1 and TRPC6, in the smooth muscle from the freshly harvested mesenteric arteries of patients 2 and 3 of Fig. 2 correspond to patients 1 and 2, respectively, in this figure and in Table 1.


Effects of prolonged ouabain treatment on arterial myocytes mimic in vivo changes in arteries in many forms of hypertension.

Arterial smooth muscle NCX1 and TRPC6 or TRPC3 expression are greatly increased in many rodent hypertension models (see Table 1 in Ref. 7) including the hypertension induced by prolonged (2 to 3 wk) subcutaneous infusion of ouabain in normal rats (49). Moreover, exposure of primary cultured rodent arterial myocytes to nanomolar ouabain for 48–72 h also increases expression of NCX1 and TRPC6 proteins (49). In this context, it is noteworthy that plasma EO levels are significantly elevated in several rodent hypertension models (15, 16, 29), as well as in 40–50% of humans with essential hypertension and in most patients with aldosterone-linked hypertension (40, 43, 51). We suggested that the increased peripheral vascular resistance in the rodent hypertension models results, in part, from the augmented Ca2+ signaling and vasoconstriction as a consequence of the Ca2+ transporter upregulation in the arterial myocytes (7, 34). In this report, we address the question: Do similar mechanisms function in human arterial myocytes? The affirmative answer appears to be a critical step toward elucidating the cellular and molecular mechanisms that link the tendency to retain salt to the elevated BP in human hypertension.

Human arterial smooth muscle expresses α1- and α2-, but not α3-, Na+ pumps.

The smooth muscle in human arteries (Fig. 1), as in rodent arteries (14, 62), expresses α1- and α2-, but not α3-, Na+ pumps. Both human (Fig. 2) and rodent arteries also express a number of the same Ca2+-transport proteins, including NCX1, TRPC6, SERCA2, and IP3R1. The expression of these Ca2+-transporters, when normalized to GAPDH or β-actin, appeared to vary somewhat from patient to patient. This may have been due to the fact that the recovered arteries were not anatomically identical and the patients′ ages, drug treatments, and diagnoses or causes of death differed markedly. Despite these shortcomings, the immunocytochemical localization of the transporter proteins and the effects of ouabain on the cultured hASMCs were consistent from patient to patient.

The expression of these Na+ and Ca2+ transporter proteins was usually greater in the actively proliferating, primary cultured hASMCs than in the growth-arrested myocytes (Fig. 3A), which have reverted to a more contractile phenotype with fusiform morphology (Fig. 7A) (23, 33). Comparable effects have been reported in rodent arterial myocyte cultures (4).

We observed an apparent increase of ∼10 kDa in the size of TRPC6 (normally a doublet of ∼100 kDa in nonproliferating cells) when hASMCs were cultured in complete growth medium (compare Figs. 2, 3B, and 8B; and see Ref. 4). The explanation is uncertain, but might be due to alternative splicing and/or glycosylation (10, 64).

NCX1 and TRPC6 colocalize with α2-Na+ pumps in PM microdomains at PM-SR junctions.

Immunocytochemical data from both freshly harvested, dissociated cells and primary cultured hASMCs demonstrate that α2-Na+ pumps are confined to PM microdomains adjacent to elements of the SR (Figs. 5 and 6). The SR was identified by staining with anti-SERCA2 pAb and with ER tracker. NCX1 and TRPC6 localize to the same PM microdomains as the α2-Na+ pumps. In contrast, α1-Na+ pumps and PMCA are more diffusely distributed in the hASMC PM, but whether they are excluded from the aforementioned PM microdomains is unclear. These Na+ and Ca2+ transporter protein distribution patterns are very similar to those in rodent arterial myocytes and some other cell types (26, 31, 39, 46, 48, 54). The implication of this structural organization, consisting of units (plasmerosomes) of PM microdomains closely opposed to junctional SR, is that the α2-Na+ pumps, NCX1 and ROCs (of which TRPC6 proteins are components), and the SR are also functionally linked (8). This is discussed in the next section.

The different distribution of the α1- and α2-Na+ pumps in hASMCs implies that they have different functions (20). The more widely distributed α1-pumps, which also have a higher affinity for Na+ (12 vs. 22 mM) (59), likely play the primary housekeeper role in maintaining the low Na+ and high K+ concentrations in bulk cytosol. The Na+ affinity difference is evidence not only for different roles but also for the likelihood that the Na+ concentration is slightly higher in the plasmerosome junctional spaces (between the PM and SR) than in the bulk cytosol. This might also help to explain why NCX1 normally operates in the Ca2+ entry mode much of the time in small artery myocytes under physiological conditions (25, 62, 63).

Prolonged incubation with nanomolar ouabain increases expression of Ca2+ transporters and augments Ca2+ signaling in hASMCs.

Prolonged (72 h) incubation of primary cultured hASMCs with 10 nM ouabain enhanced expression of NCX1 and SERCA2 (Fig. 8). These proteins are also upregulated in rat arterial myocytes treated with 100 nM ouabain for 48–72 h (49). Although TRPC6 expression, too, is enhanced by ouabain in rat myocytes (49), this did not occur in the hASMCs (Fig. 8). Treatment of hASMCs with 100 nM ouabain, however, damaged or killed most of the cells, presumably because all the Na+ pumps, both α1 and α2, were inhibited. If human, like rodent, α2-Na+ pumps have a single ouabain binding site (58) with an EC50 of ∼0.6–1.0 nM (50, 61), 10 nM ouabain should have blocked 90–95% of the α2-pumps. A 72-h exposure to 10 nM ouabain was toxic to few hASMCs, whereas 100 nM ouabain was toxic to most of the cells (see Fig. 7). This suggests that human α1-Na+ pumps have a lower affinity for ouabain (EC50 on the order of 10–20 nM) than do human α2-pumps, albeit not nearly as low as the affinity of rodent α1-pumps for ouabain (EC50 ∼5 μM). With this difference in α1- and α2-Na+ pump ouabain affinity, we might expect circulating levels of EO in humans, in the range of 0.1–0.35 nM (56), to modulate α2-Na+ pump activity and, through NCX1, Ca2+ signaling, with negligible effect on α1.

The augmentation of ATP-evoked SR Ca2+ release in hASMCs cultured with 10 nM ouabain for 72 h (Fig. 9) correlates with the increased NCX1 and SERCA2 expression in these cells (Fig. 8) and with the colocalization of these transporters with α2-Na+ pumps at plasmerosomes (Figs. 5 and 6). The implication is that the prolonged interaction of ouabain with α2-Na+ pumps upregulates NCX1 and SERCA2 expression. Then, because NCX1 operates primarily in the Ca2+ entry mode in arterial myocytes under physiological conditions (25, 62, 63), NCX1-mediated Ca2+ entry is augmented. This elevates local, sub-PM [Ca2+]CYT and increases releasable SR Ca2+ stores. The increased sequestration of Ca2+ in the SR apparently prevents marked elevation of bulk resting [Ca2+]CYT (Table 1), but it enables enhanced Ca2+ release (and, presumably, vasoconstriction) when the myocytes are activated, e.g., by ATP (Fig. 9).

The increased Ca2+ entry, in the ouabain-treated cells, when external Ca2+ is restored in the continued presence of ATP (Fig. 9), is likely mediated by ROCs and SOCs (49). The ROCs should be activated by ATP-induced diacylglycerol synthesis, whereas the SOCs should be opened as a consequence of the SR Ca2+ store depletion. Ca2+ entry through L-type voltage-gated Ca2+ channels is probably not involved under these conditions because the hASMCs should not be depolarized.

It is important to distinguish these slowly developing (long-term) mechanisms by which ouabain augments Ca2+ signaling in arterial myocytes from the short-term mechanisms by which acutely applied nanomolar ouabain elevates the sub-PM Ca2+ concentration and rapidly increases vascular tone (50, 61). Both mechanisms utilize some of the same molecular machinery, including α2-Na+ pumps and NCX1, but the long-term mechanisms discussed in this report involve gene transcription and protein expression and, perhaps, other mechanisms such as protein phosphorylation. The details of how ouabain/EO triggers the process are unknown but might involve ouabain/Na+ pump signaling mediated by a protein kinase [e.g., Src or phosphoinositide-3-kinase (PI3K)] cascade independent of the direct effects of ouabain on cation transport (36, 37). It is interesting that upregulation of these same Na+ and Ca2+ transporter mechanisms has been implicated in many different types of experimental hypertension (1, 7, 17, 34, 55, 66).

The final common path in the pathogenesis of hypertension.

The main conclusion from these studies is that the same ouabain-dependent long-term mechanisms that augment Ca2+ signaling and vascular tone in rodent arterial myocytes (49) are also functional in human arterial myocytes. The implication is that in human hypertension, including a large fraction of patients with essential hypertension and those with hyper-mineralocorticoid syndromes, in whom plasma EO is elevated (40, 43, 51), these Na+ and Ca2+ transport mechanisms promote the increased vascular tone. They also augment responses to (enhanced) sympathetic drive and to humoral vasoconstrictors and, thereby, elevate BP. We suggest that these molecular/cellular mechanisms are responsible for the whole body autoregulation (22) that increases peripheral vascular resistance and enhances vascular contractility, hallmarks of chronic hypertension (11). Indeed, the structural remodeling that is often described in chronic hypertension may be triggered by these functional and molecular changes (7).


This work was supported by National Heart, Lung, and Blood Institute Grants R01-HL-045215 and R01-HL-107555 (to M. P. Blaustein); National Heart, Lung, and Blood Institute Fellowship F32-HL-105052 (to L. K. Antos); National Institutes of Health Training Grants T32-AR07592 and T32-HL07698; and by funds from the UMSOM Dean's Office.


No conflicts of interest, financial or otherwise, are declared by the author(s).


Author contributions: C.I.L., L.K.A., V.A.G., and M.P.B. conception and design of research; C.I.L., L.K.A., and V.A.G. performed experiments; C.I.L., L.K.A., and V.A.G. analyzed data; C.I.L., L.K.A., V.A.G., and M.P.B. interpreted results of experiments; C.I.L., L.K.A., and V.A.G. prepared figures; C.I.L., L.K.A., V.A.G., and M.P.B. edited and revised manuscript; C.I.L., L.K.A., V.A.G., and M.P.B. approved final version of manuscript; M.P.B., C.I.L., L.K.A., and V.A.G. drafted manuscript.


We thank Dr. Thomas A. Pressley (Texas Tech University) for generating and purifying the anti-TSEP antibody and Dr. Hong Song (UMSOM) for validating its selectivity for human α1-Na+ pumps. We also thank Dr. Pressley and Dr. Kenneth D. Philipson (University of California, Los Angeles) for gifts of antibodies. We are very grateful to Drs. Benjamin Philosophe and Bartley P. Griffith (UMSOM), and Linda Romar (UMSOM) and Karen Kennedy (Living Legacy Foundation of Maryland, Baltimore, MD) for the procurement of human artery segments.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
View Abstract