Ouabain and other cardiotonic steroids (CTS) inhibit Na+ pumps and are widely believed to exert their cardiovascular effects by raising the cytosolic Na+ concentration ([Na+]cyt) and Ca2+. This view has not been rigorously reexamined despite evidence that low-dose CTS may act without elevating [Na+]cyt; also, it does not explain the presence of multiple, functionally distinct isoforms of the Na+ pump in many cells. We investigated the effects of Na+ pump inhibition on [Na+]cyt(with Na+ binding benzofuran isophthalate) and Ca2+ transients (with fura 2) in primary cultured arterial myocytes. Low concentrations of ouabain (3–100 nM) or human ouabain-like compound or reduced extracellular K+ augmented hormone-evoked mobilization of stored Ca2+ but did not increase bulk [Na+]cyt. Augmentation depended directly on external Na+, but not external Ca2+, and was inhibited by 10 mM Mg2+ or 10 μM La3+. Evoked Ca2+ transients in pressurized small resistance arteries were also augmented by nanomolar ouabain and inhibited by Mg2+. These results suggest that Na+ enters a tiny cytosolic space between the plasmalemma (PL) and the adjacent sarcoplasmic reticulum (SR) via an Mg2+- and La3+-blockable mechanism that is activated by SR store depletion. The Na+ and Ca2+ concentrations within this space may be controlled by clusters of high ouabain affinity (α3) Na+ pumps and Na/Ca exchangers located in PL microdomains overlying the SR. Inhibition of the α3 pumps by low-dose ouabain should raise the local concentrations of Na+ and Ca2+ and augment hormone-evoked release of Ca2+ from SR stores. Thus the clustering of small numbers of specific PL ion transporters adjacent to the SR can regulate global Ca2+ signaling. This mechanism may affect vascular tone and blood flow and may also influence Ca2+ signaling in many other types of cells.
- sodium pump
- sodium/calcium exchange
- junctional sarcoplasmic reticulum
sodium pumps in the plasmalemma (PL) of animal cells maintain a low cytosolic Na+ concentration ([Na+]cyt) (16).1 These pumps also are receptors for exogenous cardiotonic steroids (CTS) such as ouabain (16, 44) as well as for an endogenous ouabain-like compound (OLC) (19). Four different Na+ pump catalytic (α) subunit isoforms (α1–α4) have been recognized. Each is the product of a different gene and, in the rat, can be distinguished by its affinity for ouabain and for [Na+]cyt (5,46).
Most cells in the rat express the high Na+/low ouabain affinity isoform (α1) as well as one or more of the independently regulated low Na+/high ouabain affinity isoforms (α2, α3, or α4) (5,26, 46). Moreover, the different isoforms are differently distributed in the cells. For example, the α1-isoform is distributed uniformly in the PL of neurons, astrocytes, and mesenteric arterial myocytes (26). In contrast, the α3- and α2-isoforms, which are also present in neurons and arterial myocytes and in astrocytes, respectively, appear to be confined to PL microdomains adjacent to peripheral or “junctional” (j) elements of the sarco/endoplasmic reticulum (S/ER; jS/ER) (26). Despite the differences in their properties and localization, the functional significance of these isoforms has been unclear.
It is generally believed that CTS exert their effects by inhibiting Na+ pumps, elevating bulk [Na+]cyt, and, in turn, raising cytosolic Ca2+ concentration ([Ca2+]cyt) via Na/Ca exchange (3, 6, 28). Most of the entering Ca2+ is rapidly sequestered in the S/ER (6,28), and this permits additional Ca2+ to be mobilized when the cells are activated (50). These steps have all been verified with relatively high ouabain concentrations, usually 1–1,000 μM (6, 31). A longstanding enigma, however, is that nanomolar (therapeutic) concentrations of CTS augment cardiac (31) and arterial (48) contractions, induce hypertension in rats (34), and have other effects, despite little or no change in whole cell (“bulk”) [Na+]cyt (31) or resting [Ca2+]cyt(51). Such findings are incompatible with the generally accepted view of CTS action mentioned above. Furthermore, this view of CTS action does not account for the coexpression of multiple Na+ pump α-subunit isoforms with different kinetic properties including different affinities for ouabain.
Recently, isoforms of the Na+ pump with high affinity for ouabain were shown to cluster with Na/Ca exchangers in PL microdomains of rat arterial smooth muscle cells (27) and some other cell types (21, 27, 30,36). This clustering implies that these Na+pumps have a specific function. We investigated the effects of low-dose (1–100 nM) ouabain and OLC on [Na+]cytand evoked Ca2+ transients in rat arterial myocytes, which coexpress α1 and α3 (26). The results show that selective block of α3-pumps amplifies Ca2+ transients when Ca2+ is released from intracellular stores in cultured cells and in small resistance arteries. In the cultured cells, these effects occurred under conditions in which there was no measurable rise in bulk [Na+]cyt. These results indicate that the α3-isoform of the Na+ pump has a specific role in Ca2+ signaling and contraction in arterial smooth muscle.
MATERIALS AND METHODS
Cells and solutions.
Rat mesenteric arterial myocytes were grown in primary culture (40) for 4–6 days. The cells, on 25-mm coverslips, were loaded with fura 2 or sodium-binding benzofuran isophthalate (SBFI) (for [Ca2+]cyt or [Na+]cyt determination, respectively) by incubation with 3.3 μM fura 2-AM for 30 min or 10 μM SBFI-AM (the respective acetoxymethyl esters; TEFLABS, Austin, TX) for 1 h at 22–25°C.
Coverslips, mounted in a 1-ml chamber on the stage of an inverted fluorescence microscope, were superfused [32°C, 5 ml/min; chamber washout half-time (t 1/2) ≈30 s] with physiological salt solution (PSS). Normal 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 mM HEPES (titrated to pH 7.4 with NaOH). For the first 40 min, cells were superfused with PSS to remove extracellular dye and to permit the intracellular esterases to cleave the fura 2-AM or SBFI-AM. CaCl2 was omitted from Ca-free (“0” Ca) PSS, and 0.2 mM EGTA was added; 0 Ca-5 Na+ PSS contained 5 mM NaCl and 140 mMN-methyl-d-glucamine. In low K+ PSS, Na+ replaced K+. NaH2PO4 and NaHCO3 were both omitted from PSS containing LaCl3. Solution osmolarity was adjusted to 320 mosM with sucrose. The cells were routinely activated with low agonist concentrations (<EC50; seeresults) to avoid saturation of the agonist receptors or the fura 2. Arginine vasopressin (AVP) was from Peninsula Labs (Belmont, CA). All other compounds (from Sigma, St. Louis, MO, or Fisher, Pittsburgh, PA) were of reagent grade or the highest grade available.
Imaging methods for cultured cells.
[Ca2+]cyt and [Na+]cyt were determined with digital ratiometric imaging (9, 17): fura 2 excitation = 360 and 380 nm, emission = 510 nm; SBFI excitation = 340 and 380 nm, emission = 510 nm. A ×40 1.3-NA Nikon UV Fluor oil immersion objective was used for imaging. Fields containing 5–8 cells were imaged, background-subtracted, and transformed to [Ca2+] or [Na+] images (MetaFluor; Universal Imaging, West Chester, PA). Fluorescence ratio data for each cell were obtained from a 5 × 5-pixel (1.5 × 1.5-μm) nonnuclear region (1/cell). To enhance the signal-to-noise ratio, 32 frames were averaged at video frame rate, except when agonists and/or ouabain were added; then only four frames were averaged to improve temporal resolution. Images were acquired at a rate of two per minute, except during agonist application, when the rate was increased to two per second.
Isolated small mesenteric arteries.
Sprague-Dawley male rats, 200–400 g, were anesthetized and decapitated, and small resistance arteries (length = 3–4 mm, maximum passive diameter = 200–220 μm; seeExperimental protocol for blood vessels) were isolated (35). The arteries were dissected at 5–6°C in a “dissection solution” of the following composition (in mM): 145 NaCl, 5 KCl, 2.5 CaCl2, 1 MgSO4, 1 KH2PO4, 2 pyruvate, and 5 glucose, with 1% bovine serum albumin (pH = 7.4, osmolarity = 310–320 mosM). The vessels were denuded of endothelium by perfusion of the lumen (through a small pipette) with tiny air bubbles for 30 s. Removal of the endothelium was verified by the abolition of a 2 μM acetylcholine-evoked Ca2+ transient or vasodilation.
The arteries were placed in a 2-ml experimental chamber on the microscope stage, and the open ends were mounted on glass cannulas. One cannula was sealed off, and the artery was pressurized to 70 mmHg.
Experimental protocol for blood vessels.
Resistance arteries were superfused (2.5 ml/min, 37°C) with modified Krebs solution of the following composition (in mM): 115 NaCl, 4.7 KCl, 1.2 MgSO4, 1.2 KH2PO4, 25 NaHCO3, 10 glucose, and 1.8 HEPES (pH = 7.4, osmolarity = 310–320 mosM, bubbled with 95% air-5% CO2); the solution also contained 5 μM guanethidine to deplete catecholamines from the nerve terminals. Before the experiments were started, the vessels were incubated for 1 h; maximum passive diameter was then determined by a 20-min wash in Ca2+-free Krebs containing 2 mM EGTA. For each experimental protocol, the Krebs solution also contained 50 μM tetrodotoxin during one entire experiment to verify that the responses were unaffected by the presence of this agent.
To measure intracellular Ca2+, the arteries were loaded with 5 μM fura 2-AM for 3 h in dissection solution at 22–25°C in a 95% O2-5% CO2 atmosphere. After dye loading, the vessel was first mounted on one of the cannulas and was internally perfused to wash out residual dye from the lumen. The other end was then cannulated so that the vessel could be pressurized. In the vessel experiments, fura 2 was excited at 380, 360 (the isosbestic point), and 340 nm, and emission was recorded at 510 nm. Because of uncertainty in the dye calibration in these arteries, the data are reported as the background (bkg)-subtracted ratios of fluorescence (F) emitted when the dye was excited at 360 and 380 nm [i.e., F360/F380 = (F360 − Fbkg 360)/(F380 − Fbkg 380), where Fbkg 360 and Fbkg 380 were obtained when the artery was moved out of the field]. The F360 values were used so that we could keep track of bleaching and minimize movement/contraction artifacts in the fluorescent signals. In each experiment, fluorescent ratio data were obtained from each of three different 5 × 50-pixel (1.5 × 15-μM) areas of the vessel wall; these data were then averaged off-line to obtain a single value for each vessel. To improve the signal-to-noise ratio, 32 frames were averaged at video frame rate to obtain the data, except when agonists and/or ouabain were added; in these latter instances, only four frames were averaged to improve temporal resolution. Image ratios were acquired at a rate of one per minute, except during periods of agonist application when the rate was increased to one per second. When the vessels were stimulated with phenylephrine (PE), the “steady-state” F360/F380ratio was taken as the last data point before washout of PE.
Data (fluorescence ratios and calibrated values) were analyzed and plotted with Sigma Plot software (Jandel, San Rafael, CA). The significance of the differences between means (±SE) for the same groups of cultured cells and for small arteries under various conditions was calculated by Student's paired t-test.
Nanomolar ouabain amplifies evoked Ca2+ transients in cultured arterial myocytes.
Figure 1 A illustrates the protocol and a key result. The responses to four consecutive 30-s applications of serotonin (5-HT) in PSS are shown. The first 5-HT application was a control. During the second and third exposures, ouabain (10 and 100 nM, respectively) and 5-HT were applied simultaneously. Ouabain was omitted during the fourth 5-HT application (“recovery”). The 5-HT-evoked Ca2+ transients were negligibly affected by 10 nM ouabain, whereas 100 nM ouabain reversibly augmented transients 1.8-fold (Fig. 1, A and B; Table 1 contains summary data). With a 60-s preincubation to permit more complete solution exchange and equilibration (see materials and methods), 100 nM ouabain-augmented 5-HT evoked Ca2+ transients 2.6-fold (Fig. 1 B); the apparent EC50 was 8 nM, comparable with the IC50 for block of the rat Na+ pump α3 isoform (5).
The graph in Fig. 1 C shows the recovery from the ouabain-induced augmentation of the peak Ca2+ transients; ouabain could be washed out with a t 1/2 of ≈30 s. In these experiments, the cells were incubated with 100 nM ouabain for 30 s; the ouabain was then washed out for various lengths of time (as indicated on the ordinate) before the 5-HT stimulation (Fig.1 C). These results indicated that ouabain dissociates rapidly from its high-affinity binding sites in these myocytes.
Table 1 A shows summary data for the Ca2+transients evoked by 0.1 nM AVP, by 100 nM PE and 5-HT, and by 1 μM cyclopiazonic acid (CPA), an inhibitor of S/ER Ca2+ pumps and the augmentation of these transients by simultaneous application of 100 nM ouabain. In each case, ouabain reversibly augmented the peak transient by ∼1.8- to 2.0-fold; thus the effect did not appear to be agonist specific. Moreover, inhibition of the Na+ pumps with another CTS, 100 nM digoxin (Table 1 A), or by reduction of the extracellular K+ concentration ([K+]O) from 5 to 1 mM (Table1 B) had similar effects.
Figure 2 shows that the OLC isolated from human plasma (19) also augmented the Ca2+transients. Indeed, the augmentation of 5-HT-evoked Ca2+transients by 10–30 nM human OLC and (plant) ouabain (Fig. 2,A–C) was quantitatively indistinguishable (Fig.2 C-a). This implies that ouabain and OLC have similar potencies in these cells.
Amplification by ouabain requires external Na+ but does not depend on a rise in bulk [Na+]cyt.
When external Ca2+ was removed 30 s before 5-HT application, ouabain amplified the peak Ca2+ transients evoked by 5-HT without increasing the duration of the transients (Fig.1 D; Table 1 A). The augmentation of the peak transients was also unaffected by 10 μM nifedipine (Table1 C). This indicates that the amplified response was due to enhanced mobilization of stored Ca2+, and that the simultaneous entry of Ca2+ via L-type Ca2+channels was not required.
No augmentation was observed, however, if external Na+ and Ca2+ were both removed 30 s before 5-HT application (Fig. 1 E). This dependence on external Na+suggested that the amplification of the Ca2+ transients by ouabain depended on Na+ entry and the elevation of [Na+]cyt. Therefore, it was surprising when direct measurements with SBFI revealed that whole cell (bulk) [Na+]cyt did not rise under these conditions or even when 100 nM ouabain was superfused for 10 min (Fig.3 A) or [K+]O was lowered to 1 mM for 10 min (Fig.3 B). As expected, significantly higher doses of ouabain (1–10 μM) did raise [Na+]cyt (Fig.3 A), as did further reduction of [K+]O (not shown; see Ref. 9). Exposure to 100 nM ouabain for 10 min also had no effect on bulk [Ca2+]cyt in unstimulated cells (not shown).
Role of the Na/Ca exchanger in ouabain-augmented responses.
Despite the finding that low-dose ouabain did not elevate whole cell [Na+]cyt, the Na/Ca exchanger apparently contributed to the augmentation by ouabain. Ca2+ efflux via Na/Ca exchange is a steep, sigmoid function of external Na+concentration ([Na+]O) (Hill coefficient ≈3), with half-maximal activation at [Na+]O ≈20–40 mM (7). As the chamber washout t 1/2 was ≈30 s (seematerials and methods), a 30-s washout of Na+(Fig. 1 E) would have lowered [Na+]O only to ≈75 mM, which would still support Ca2+ extrusion via Na/Ca exchange. Yet, under these conditions, the ability of ouabain to augment evoked Ca2+transients was prevented, presumably because of a marked reduction of Na+ entry (which may be a more linear function of [Na+]O in this concentration range; see Ref.22). With more prolonged (2 min) washout of Na+ to lower [Na+]O to ≈10 mM, Ca2+ exit via Na/Ca exchange was inhibited, and the response to 5-HT was amplified in the absence of ouabain (Fig. 1 F). Thus the inhibition of exchanger-mediated Ca2+ extrusion by a large reduction of [Na+]O (Fig. 1 F) or by inhibition of Na+ pump-mediated Na+extrusion, with either nanomolar concentrations of ouabain (Fig. 1,A, B, and D, and Table 1, Aand C) or reduced [K+]O, augmented evoked Ca2+ transients to a similar extent. Collectively, this series of observations demonstrating the dependence of the ouabain augmentation on Na+ is most paradoxical, because in no instance was there any increase in bulk [Na+]cyt (Fig. 3).
La3+ and elevated Mg2+ selectively block ouabain-induced augmentation.
To address the question of whether this amplification depended on Na+ entry or was due to an action of Na+ at the external surface of the cells, we considered possible Na+entry pathways. Most vascular smooth muscle cells, including mesenteric arterial myocytes, do not possess voltage-gated Na+channels (53). Consistent with the electrophysiological data, neither veratridine (an Na+ channel opener; Fig.3 B) nor tetrodotoxin (50 μM, not shown) affected [Na+]cyt. These myocytes do, however, possess store-operated channels (SOC) that are permeable not only to Ca2+ but also to Na+ (2) and are blocked by 10 μM La3+ and 10 mM Mg2+(2, 23, 52). In unstimulated cells, La3+ and Mg2+ lowered resting [Na+]cyt, possibly by blocking Na+ entry through a small number of SOC that may have been open at rest (2). La3+ and Mg2+blocked both the ouabain- and the 1 mM K+-induced augmentation of 5-HT-evoked Ca2+ transients (Fig. 4 and Table 1, C and D), as well as the augmentation induced by human OLC (Fig. 2 C-b). The data suggest that ouabain- and low [K+]O-induced amplification of the Ca2+ transients depended on Na+ entry via an La3+- and Mg2+-inhibitable pathway, possibly mediated by SOC.
Effects of ouabain on PE-evoked Ca2+transients in pressurized small arteries.
The aforementioned experiments were performed on cultured myocytes from large, muscular mesenteric arteries. In small arteries and arterioles, myogenic tone plays a role in the regulation of vessel diameter and blood flow in response to changes in perfusion pressure. To determine the possible influence of ouabain on the function of these vessels, which are responsible for much of the peripheral vascular resistance, we studied evoked Ca2+ transients in pressurized small mesenteric arteries. In control experiments (not shown), a 60-min exposure to nanomolar ouabain had no detectable effect on [Na+]cyt (measured with SBFI) or [Ca2+]cyt in pressurized preparations. Treatment with 1 mM ouabain, however, markedly increased both [Na+]cyt (i.e., the SBFI F340/380ratio) and [Ca2+]cyt (i.e., the fura 2 F360/380 ratio).
To eliminate Ca2+ signals from the endothelium of the intact vessels, the arteries were denuded of their endothelium (seematerials and methods), loaded with fura 2, and pressurized to 70 mmHg (the standard condition). The vessels were superfused with Krebs solution containing 10 μM nifedipine to minimize Ca2+ entry through L-type Ca2+ channels and 5 μM guanethidine to deplete sympathetic nerve terminals of catecholamines. Under these conditions, stimulation of the vessels with 5 μM of the α-adrenergic agonist, PE, for periods of 15–20 min induced Ca2+ transients (increase in the F360/F380 ratio; see materials and methods). These transients were characterized by an early peak with oscillations and a delayed plateau (Fig.5 A).
The Ca2+ transients were highly reproducible; the peaks and plateaus varied by <7 and 10%, respectively, in five successive applications of PE under control conditions (n = 5 vessels). Superfusion of the vessels for 30 min with 1–100 nM ouabain in the Krebs solution had no effect on the baseline (resting) [Ca2+]cyt (i.e., F360/F380 did not change) but increased both the initial peak of the Ca2+ transient and the subsequent plateau evoked by PE (Fig. 5 A). Figure5 B shows the ouabain dose-response curves for both the peak and the plateau (steady state). The “threshold” ouabain concentrations for amplification of both the peak and the plateau were similar (<1 nM). The two curves also had comparable EC50values (4 and 8 nM) that were also similar to the EC50 for ouabain in cultured cells (8 nM; Fig. 1 B).
Figure 6 shows the effects of extracellular Mg2+ concentration ([Mg2+]O) on the PE-evoked Ca2+ transient. When [Mg2+]O was increased from 1.2 mM (in control Krebs solution) to 10 mM, the peaks of the Ca2+transients evoked by 5 μM PE were unaffected (Fig. 6), whereas the plateau responses were reversibly abolished. The initial Ca2+ transient is thought to reflect primarily Ca2+ release from the sarcoplasmic reticulum (SR) stores, whereas the plateau is likely the result of Ca2+ entry through activated SOC that can be blocked by 10 mM Mg2+ (2, 52). Figure7 shows the effects of [Mg2+]O on the ouabain-induced augmentation of the Ca2+ transient evoked by PE in the vessels. Prolonged incubation with 3.3 nM ouabain reversibly augmented the peak Ca2+ transients evoked by PE (Fig. 7 A). When [Mg2+]O was increased to 10 mM in the presence of ouabain, the magnitude of the evoked transients was reduced to normal (i.e., to the “control” level observed in the absence of ouabain), and the plateau disappeared (Fig. 7, Aand B). The ability of elevated [Mg2+]O to block the amplification of the Ca2+ transients by low-dose ouabain in the arteries resembles the effects observed in the cultured myocytes (Fig. 4 A and Table 1 C). Figure8 is a model depicting key transport proteins involved in the augmentation of Ca2+transients by low-dose ouabain and OLC and low [K+]O.
Rationale for a reexamination of the mechanism of action of ouabain.
The idea that the cardiotonic (and other pharmacological) action(s) of low-dose CTS such as ouabain is initiated by Na+ pump inhibition and a global rise in [Na+]cyt has long been accepted dogma (e.g., Ref. 28). The discovery of the Na/Ca exchanger established a critical link between Na+ pump activity and Ca2+ metabolism and provided the basis (3) for this widely held view. Nevertheless, the very low doses (<100 nM) of CTS that induce cardiotonic (11) and vasotonic effects (48) in the rat inhibit only a small fraction of the Na+ pumps of a cell, so that there should be negligible rise in whole cell [Na+]cyt. Indeed, several investigators have pointed out that the cardiotonic effects of CTS can be observed without any detectable rise in [Na+]cyt (9, 31), but their reports have generally been ignored. Our observations are not only consistent with these reports and support the view that the cardiotonic and vasotonic actions of low-dose CTS can occur in the absence of a rise in whole cell [Na+]cytbut now suggest a plausible mechanism for these effects. Moreover, our results indicate specific and distinct functional roles for the two isoforms of the Na+ pump α-subunit (α1 and either α2 or α3) that are coexpressed but differently distributed in the PL of myocytes from arteries and skeletal muscle (26,30) and perhaps cardiac muscle (25). Here, we discuss our findings from this perspective, and we propose a new theory of CTS action that embraces these previously unexplained observations.
Low-dose ouabain augments Ca2+signaling in arterial myocytes with no elevation of whole cell [Na+]cyt.
In pressurized small mesenteric arteries as well as in cultured myocytes, we found that low concentrations of ouabain (1–30 nM) augmented Ca2+ transients. These same low doses of ouabain also augmented vasoconstrictor responses in small arteries (unpublished data). The augmentation of Ca2+ transients by ouabain depended on extracellular Na+. Similar augmentation was also observed with low-dose digoxin, OLC, and reduced [K+]O in the cultured cells. Thus it appears that this effect was the result of Na+ pump inhibition and the reduction of active Na+ extrusion rather than some other effect of the CTS. For example, the effect cannot be attributed to “slip-mode conductance” of Ca2+ through voltage-gated Na+ channels (38) because mesenteric arterial myocytes do not express such channels (53) and because augmentation occurred in the absence of extracellular Ca2+ (Fig. 1 D).
In rat mesenteric artery, the low doses of ouabain, digoxin, and OLC that we used (≤100 nM) should inhibit only α3 Na+ pumps. The α2-isoform is not expressed in the arterial myocytes, and the much more prevalent α1 isoform (26) has an EC50 more than 1,000-fold greater than the ouabain concentrations used here (5).
It seems paradoxical that low ouabain concentrations or reduction of [K+]O to 1 mM augmented Ca2+signaling without elevating whole cell [Na+]cyt (Fig. 3), even though external Na+ was required (Fig. 1 E). Yet, the absence of a detectable rise in whole cell [Na+]cyt is consistent with two related observations. 1) The α3 Na+ pumps apparently constitute only a small fraction (perhaps 10% or less) of the total Na+ pumps in the arterial myocytes (26). 2) In the cultured myocytes, the effects of low-dose ouabain and OLC were manifested within 30–120 s (Fig. 1, A–E, and Fig. 2). The rapidity of the augmentation with 10–100 nM ouabain could not have been due to elevation of whole cell [Na+]cytbecause even 1 μM ouabain caused only a slow rise in bulk [Na+]cyt (<0.3 mM/min; Fig. 3 A).
A critical observation was that low-dose ouabain amplified Ca2+ signals even in the absence of extracellular Ca2+ (Fig. 1 D). This effect might be explained by reduced Ca2+ extrusion (e.g., via Na/Ca exchange). Consistent with this hypothesis, a 2-min incubation in 0 Ca2+-5 mM Na+ medium (sufficient to reduce [Na+]O to <10 mM) augmented evoked Ca2+ transients in the absence of ouabain (Fig.1 F). When taken together, the data fit the idea that the augmentation was the result of reduced Ca2+ efflux via Na/Ca exchange, because of either a low [Na+]O (as in Fig. 1 F) or a high [Na+]cyt in the vicinity of the Na/Ca exchanger brought about by the inhibition of the α3 Na+pumps by the low-dose ouabain (Fig. 1 D and see below).
Na+ pump catalytic subunit (α) isoforms: their functional significance.
Although Na+ pump inhibition may augment Ca2+signaling by inhibiting, indirectly, Ca2+ extrusion via Na/Ca exchangers, it was necessary to find an explanation for the effect in the absence of elevation of whole cell [Na+]cyt. Recent observations of differential Na+ pump isoform distribution and the localization of Na/Ca exchangers provided important clues that may help to resolve this dilemma. In mesenteric arterial myocytes, both the α3 Na+pumps and the Na/Ca exchangers (27) are confined to discrete domains of PL that overlie the jSR (Fig. 8). In contrast, the low ouabain-affinity α1 Na+ pumps and the PL (ATP-driven) Ca2+ pumps are distributed uniformly over the remainder of the cell surface (27). Restricted localization of the α2- or α3-isoforms of the Na+ pump along with the Na/Ca exchanger in PL microdomains adjacent to jS/ER, have also been observed in other cell types (21, 27, 30,36). Furthermore, transgenic studies reveal that, in mouse cardiac muscle, α1 and α2 are coexpressed, but the latter selectively mediates the cardiotonic effect of low-dose ouabain (25). In addition, several groups have reported that the activity of the cardiac Na/Ca exchanger is tightly coupled to the Na+ pump through a small compartment of cytosol (presumably sub-PL) that has restricted diffusional access to the bulk cytosol (14, 33, 45). Indeed, Wendt-Gallitelli et al. (49) directly demonstrated, in cardiac myocytes, the presence of tiny sub-PL regions of cytosol in which the local [Na+]cyt was as high as 40 mM, whereas bulk [Na+]cyt remained low. Also, microheterogeneity of [Ca2+]cyt in sub-PL regions has been inferred from functional studies in arterial smooth muscle (10, 15, 43,47) as well as in cardiac myocytes (29,41).
We propose that these discrete regions of cytosol are located between the jSR and the overlying PL microdomains that, in arterial smooth muscle cells, contain the Na/Ca exchangers and the α3 Na+pumps (Fig. 8). Indeed, functional evidence indicates that some Ca2+ entry and exit in smooth muscle proceeds through a restricted cytoplasmic compartment, the “superficial buffer barrier” (10, 47), and that Ca2+ released from the jSR into this junctional space is preferentially extruded via the Na/Ca exchangers without affecting bulk [Ca2+]cyt (47). These functional units may be analogous to the triads and dyads of skeletal and cardiac muscle (13, 20) and have been called “PLasmERosomes” (7). If the junctional regions are assumed to be disc shaped, with diameters of ≈200 nm and PL-jSR separation of ≈12 nm (42), a single PLasmERosome should have a volume of ≈5 × 10−19 l; it should contain only ≈2,500–3,000 Na+ if the local [Na+] is comparable with that in bulk cytosol (≃9 mM; Fig. 3). With the use of a turnover number of 150 cycles/s for a single Na+ pump (44), which corresponds to 450 Na+/s, and with greatly restricted diffusion of Na+ (49) and Ca2+(29, 41) between the PLasmERosome and bulk cytoplasm, it is apparent that block of a single Na+ pump by low-dose ouabain could cause [Na+]cyt in the PLasmERosome to rise rapidly. This would inhibit Ca2+extrusion by the nearby Na/Ca exchangers and raise local [Ca2+]cyt. Furthermore, a strictly local increase in [Ca2+]cyt may increase global Ca2+ signaling by sensitizing the inositol trisphosphate receptors (36a) and/or raising the jSR Ca2+ content.
Accordingly, we propose that the widely cited sequence of events involved in the mechanism of action of CTS (6,28) is, in principle, correct: namely, Na+pump inhibition → ↑[Na+]cyt → (via Na/Ca exchange)→ ↑[Ca2+]cyt → ↑[Ca2+]SR, where [Ca2+]SR is the SR Ca2+concentration. Our results suggest, however, that these events are mediated preferentially by specific transporters located in the PLasmERosomes (Fig. 8). The increased Ca2+ released at the jSR should then be sufficient to augment whole cell Ca2+ signaling (Fig. 1, A–D, and Figs. 2, 4, and 5).
Mechanism by which Mg2+ and La3+ antagonize the effect of the low-dose ouabain.
The experiments with Mg2+ and La3+ support the aforementioned hypothesis. In mesenteric arterial myocytes, the SOC opened by SR Ca2+ depletion and through which Ca2+ stores are refilled are about one-tenth as permeable to Na+ as to Ca2+ (2). Because extracellular fluid contains ≈100 times more Na+ than Ca2+, however, paradoxically these channels may admit more Na+ than Ca2+ under physiological conditions. Block of SOC by La3+ or Mg2+ should reduce Na+ entry via this route. Neither La3+ nor Mg2+ is thought of as a specific SOC blocker. Nevertheless, the similarity of their effects at the concentrations used and the fact that they did not affect the 5-HT- or PE-evoked Ca2+transient peaks in the absence of ouabain (Table 1,A–D; Ref. 39 and unpublished data) suggest that they may be relatively selective for SOC under the conditions of our experiments. Moreover, the localization of SOC in the PL microdomains at the PLasmERosome (24) might enhance the sensing of SR Ca2+ store depletion. Thus the effects of La3+and elevated Mg2+ (Fig. 2 C-b and Figs. 4 and 7; Table 1, C and D) provide further support for the participation of SOC as an entry pathway for Na+ that is critical to the ability of ouabain to amplify Ca2+transients in the arterial myocytes.
Significantly, comparable effects of low-dose ouabain and of elevated Mg2+ on Ca2+ signaling were observed in small resistance arteries (Figs. 5-7). Here, the PE-evoked Ca2+ transient in the presence of nifedipine consisted of a large rise and oscillation of [Ca2+]cyt (Figs. 5 A,6 A, and 7 A) that may reflect Ca2+entry through receptor-operated channels (ROC) as well as release of Ca2+ from the SR. This early Ca2+ transient was followed by a decline to a new plateau [Ca2+]cyt that was well maintained for at least 20–25 min (Figs. 5 A, 6 A,7 A). The plateau probably reflects Ca2+ entry through SOC, and not ROC (4), because it was abolished by 10 mM Mg2+ (Figs. 6 and 7).
In normal medium containing 1.4 mM Mg2+, low-dose ouabain augmented both the peak and plateau phases of the Ca2+ transient in the isolated arteries (Fig. 5). This fits with the idea that ouabain may amplify cell Ca2+ signaling. Elevation of [Mg2+]O to 10 mM reduced the peak of the Ca2+ transient augmented by ouabain and completely blocked the augmented plateau phase (Fig. 7). These results are consistent with the view that both components of the ouabain-induced augmentation of the transient depended on Na+ entry through the SOC.
Implications of the local modulation of Ca2+ signaling.
The local regulation of [Na+]cyt and [Ca2+]cyt might affect numerous cell functions that are influenced by Ca2+, including many in neurons and in all types of muscle where PL-jS/ER complexes (13, 20, 42) with a similar complement of PL transport proteins are present. Our results in resistance arteries raise the possibility that local regulation of Na+ and Ca2+ may play an important role in the control of vascular tone and blood flow.
Our observations appear to clarify some aspects of the rapid action of therapeutic concentrations of CTS (48) and OLC (Fig. 2) on vascular tissue. The action of OLC itself may provide an additional rationale for the widespread occurrence of discrete receptors for CTS, the affinities of which vary widely (5, 46). Indeed, a mechanism that can influence, selectively, the responsiveness of many types of cells to a variety of neurohumoral stimuli may have very broad relevance. Furthermore, in addition to ouabain, various agents such as dopamine that modulate Na+ pump activity (1, 8) may exert their effects via this mechanism. Finally, elevated plasma levels of OLC are observed frequently in patients with essential hypertension (38) and congestive heart failure (18). Thus agents that affect the action of ouabain on specific Na+ pump isoforms might not only have novel therapeutic consequences (12) but might also provide additional tools with which to explore the effects of local changes in intracellular Na+.
We thank K. Strauss for cell cultures, V. A. Golovina and G.R. Monteith for help with imaging, V. A. Miriel for help with the small artery preparation, M. Juhaszova for help with cytochemistry and imaging, M. D. Stern for helpful discussion, and P. DeWeer, B. K. Krueger, and W. G. Wier for comments on the manuscript. We thank TEFLABS (Austin, TX) for a gift of fluorescent dyes.
This study was supported by National Heart, Lung, and Blood Institute Grant HL-45215; J. M. Hamlyn was an Established Investigator of the American Heart Association.
Present address of A. Arnon: Dept. of Medicine, Cardiovascular Institute, Loyola Univ. Medical Center, Maywood, IL 60153.
Address for reprint requests and other correspondence: M. P. Blaustein, Dept. of Physiology, Univ. of Maryland School of Medicine, 655 W. Baltimore St., Baltimore, MD 21201.
↵1 Cation (M) concentrations in “bulk” cytosol and in the extracellular fluid are denoted by, respectively, [M]cyt and [M]O, where M = Na+, K+, or Ca2+.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society