How does pressure overload cause cardiac hypertrophy and dysfunction? High-ouabain affinity cardiac Na⁺ pumps are crucial

Abstract

Left ventricular hypertrophy is frequently observed in hypertensive patients and is believed to be due to the pressure overload and cardiomyocyte stretch. Three recent reports on mice with genetically engineered Na⁺ pumps, however, have demonstrated that cardiac ouabain-sensitive α₂-Na⁺ pumps play a key role in the pathogenesis of transaortic constriction-induced hypertrophy. Hypertrophy was delayed/attenuated in mice with mutant, ouabain-resistant α₂-Na⁺ pumps and in mice with cardiac-selective knockout or transgenic overexpression of α₂-Na⁺ pumps. The latter, seemingly paradoxical, findings can be explained by comparing the numbers of available (ouabain-free) high-affinity (α₂) ouabain-binding sites in wild-type, knockout, and transgenic hearts. Conversely, hypertrophy was accelerated in α₂-ouabain-resistant (R) mice in which the normally ouabain-resistant α₁-Na⁺ pumps were mutated to an ouabain-sensitive (S) form (α₁^S/Sα₂^R/R or “SWAP” vs. wild-type or α₁^R/R α₂^S/S mice). Furthermore, transaortic constriction-induced hypertrophy in SWAP mice was prevented/reversed by immunoneutralizing circulating endogenous ouabain (EO). These findings show that EO and its receptor, ouabain-sensitive α₂, are critical factors in pressure overload-induced cardiac hypertrophy. This complements reports linking elevated plasma EO to hypertension, cardiac hypertrophy, and failure in humans and elucidates the underappreciated role of the EO-Na⁺ pump pathway in cardiovascular disease.

left ventricular (LV) hypertrophy occurs in about 10-15% of the general adult population (87) and in ~40% of hypertensive patients (18) and is a predictor of adverse cardiovascular events, including heart failure and death, in hypertensive patients (1, 49, 65). It is therefore important to understand the mechanism(s) that cause hypertrophy. In poorly controlled hypertension, it is widely accepted that the “pressure overload” and consequent increase in cardiac workload required to expel blood from the heart initiates the almost inevitable cardiac hypertrophy and dysfunction (16, 45, 72, 113, 120). The myocardial stretch is believed to trigger the cardiac changes (35, 40, 110, 122). However, two landmark, but underappreciated, articles published in this journal, and a more recent report in Circulation Research, in which transaortic constriction (TAC), a common model, was used to induce pressure overload, have demonstrated that this view is not entirely correct. Notably, experiments on mice with genetically engineered cardiac α₂-Na⁺ pumps can exhibit substantial TAC-induced elevation of LV systolic pressure (LVSP) but greatly attenuated cardiac hypertrophy and dysfunction (15, 82, 112). Those observations, which are summarized here, provide the basis for a new explanation for how pressure overload triggers the cardiac structural and functional changes that characterize hypertrophy (Figs. 1 and 2).

Fig. 1.

A: control conditions. Key transporters that link Na⁺ transport to Ca²⁺ signaling in cardiomyocytes; for simplicity, plasma membrane (PM) Ca²⁺ channels and Ca²⁺ pumps (PMCA), are omitted. Na⁺/Ca²⁺ exchanger 1 (NCX1), governed primarily by the α₂-Na⁺ pump (NKA)-generated local Na⁺ electrochemical gradient at the PM-sarcoplasmic reticulum (SR) junction (the PLasmERosome; shaded area) helps regulate the local cytosolic Ca²⁺ concentration ([Ca²⁺]_CYT) and SR Ca²⁺ stores ([Ca²⁺]_SR). B: the hypothesized changes in these mechanisms during pressure overload [transaortic constriction (TAC)]-induced cardiac hypertrophy (see text and Ref. 10). Pressure overload (e.g., via TAC) activates the renin-angiotensin-aldosterone system (RAAS). This increases sympathetic nerve activity (SNA), which directly stimulates cardiomyocytes. RAAS activation also leads, via a hypothalamic-adrenocortical neurohumoral pathway, to elevation of plasma endogenous ouabain (EO), which selectively inhibits high-ouabain affinity α₂-Na⁺ pumps, thereby enhancing NCX1-mediated net Ca²⁺ gain (see Fig. 2, left; acute effects of EO). This increases [Ca²⁺]_SR, enhances Ca²⁺ signaling, and augments contraction (i.e., it increases cardiac workload). Sustained high plasma EO binding to (perhaps nontransporting) α₂-Na⁺ pumps also activates protein kinase (PK) cascades that lead to cardiac hypertrophy (see Fig. 2, right; chronic effects of EO). C: as discussed in this review, the dashed-line boxes indicate where the mechanisms in B are influenced by genetic engineering of α₁- and α₂-Na⁺ pumps and by treatment with Digibind. In C, the two red-bordered boxes surrounding α₂-Na⁺ pumps refer to the evidence that TAC-induced cardiac hypertrophy is attenuated or prevented by the following genetic alterations: mutation of α₂ to an ouabain-resistant form (test 1), knockout of α₂ (test 2), or overexpression of α₂-Na⁺ pumps (test 3). The five green-bordered boxes surrounding α₁-Na⁺ pumps refer to the evidence that hypertrophy is unaffected by overexpression of (normally) ouabain-resistant α₁-Na⁺ pumps (test 4) but is augmented by mutation of α₁ to an ouabain-sensitive form in α₂-ouabain-resistant mice (i.e., α₁^S/Sα₂^R/R or “SWAP” mice) (test 5). The single blue-bordered “Digibind” box surrounding EO refers to the evidence that, in SWAP mice, hypertrophy is prevented or reversed by immunoneutralization of ouabain with Digibind (test 6). See text for further details. RyR, ryanodine receptor.

Fig. 2.

Cellular mechanisms that mediate the acute (left) and chronic (right) effects of ouabain/endogenous ouabain (EO) in the heart. Regulation of the heart by sympathetic nerve activity (SNA) is also shown. As described in the text, in rodents, low-dose EO binds exclusively to, and inhibits, α₂-Na⁺ pumps, which are located in cardiomyocyte (CMC) PLasmERosomes (see Fig. 1A). The rise in local intracellular Na⁺ in the PLasmERosomes ([Na⁺]_PL) rapidly induces Na⁺/Ca²⁺ exchanger (NCX)-mediated Ca²⁺ gain and the cardiotonic effect (left). If this is sustained, it leads to cardiac hypertrophy. EO can be immunoneutralized by commercial Fab fragments raised against digoxin (Digibind or DigiFab). Prolonged plasma EO elevation also activates a α₂-Na⁺ pump-associated protein kinase (PK) cascade, e.g., c-Src mediated, that increases cardiac NCX expression (right; green dotted line). This promotes NCX-mediated Ca²⁺ extrusion during diastole. Prolonged α₂-pump inhibition by EO, however, reduces the Na⁺ gradient driving force so that “global” cytosolic Na⁺ concentration ([Na⁺]_CYT) ([Na⁺]_CYT) [as well as [Na⁺]_PL) and diastolic cytosolic Ca²⁺ concentration ([Ca²⁺]_CYT)] are elevated; consequently, cardiac relaxation is slowed and/or incomplete. Also, cardiac SERCA2 expression is usually reduced in heart failure (perhaps due to the high EO; red dotted line), as are sarcoplasmic reticulum (SR) Ca²⁺ stores and Ca²⁺ transients, and systolic function is impaired. The diastolic dysfunction and attenuated cardiac contraction and stroke volume help explain the heart failure. [Modified from Ref. 10.]

Ouabain-Sensitive Na⁺ Pumps in the Heart

Evidence that cardiac α₂-Na⁺ pumps (or Na⁺-K⁺-ATPase), their high-affinity ouabain-binding site, and its ligand endogenous ouabain (EO) are pivotal elements in the pathogenesis of cardiac hypertrophy and failure was recently reviewed (10, 36). The data demonstrate that hypothalamic mechanisms help regulate cardiovascular function not only via central modulation of sympathetic nerve activity but also via a novel neurohumoral pathway mediated by circulating EO and its cardiac and vascular receptors, the high-ouabain affinity Na⁺ pumps (10, 38).

Na⁺ pumps consist of an α-subunit and a β-subunit. The glycosylated β-subunit is required for catalytic activity, but the much larger α-subunit, of which there are four isoforms (α₁−α₄), contains the catalytic machinery and ouabain [cardiotonic steroid (CTS)]-binding site and mediates Na⁺ and K⁺ transport (8, 47, 104, 119). The Na⁺ pump subunit structure is shown in Fig. 3. Ouabain binds with its lactone ring deep in the cation transport pathway and its sugar moiety (rhamnose) in the wide external vestibule exposed to the extracellular fluid (47). Cardiac Na⁺ pumps are regulated by another small polypeptide, phospholemman (PLM), a member of the FXYD family, sometimes called the γ-subunit (7, 13, 55, 81). Cloning of the Na⁺ pump α- and β-subunit isoforms by Shull et al. (93–95) enabled subsequent mutation of the ouabain-binding site in the α₁- and α₂-isoforms and knockout (KO) or overexpression of individual isoforms in specific tissues. This provided an opportunity to examine the Na⁺ pump-endogenous CTS endocrine system and the different functions of these two α-isoforms.

Fig. 3.

Model of a Na⁺ pump depicts the organization of the α- and-β subunits; as illustrated, the small β-subunit is glycosylated (hexoses shown in green). Also shown is the regulatory γ-subunit, phospholemman (PLM). Ouabain (at top) enters the cation transport pathway lactone ring first and binds close to one of the cation-binding sites in the pathway.

Adult rodent and human cardiomyocytes [and arterial myocytes (54)] express Na⁺ pumps with an α₁-isoform (≈80–95% of pumps) and pumps with an α₂-isoform (≈5–20% of pumps) (6, 23, 41, 115), but human hearts also express a very small number of α₃-pumps (63, 91). Notably, α₁ is widely distributed in the plasma membrane (PM) and is the “housekeeper” that maintains the low global cytosolic Na⁺ concentration ([Na⁺]_CYT) (6, 62). The α₁-to-α₂ ratio (≈5–20:1) is similar in rodents and humans, but rodent α₁-Na⁺ pumps are unusually resistant (R) to ouabain (8, 32, 54, 70), i.e., the wild-type (WT) mouse genotype is α₁^R/Rα₂^S/S (where S indicates ouabain sensitive).

In contrast to α₁-pumps, cardiomyocyte α₂-pumps are largely confined to PM microdomains at PM-sarcoplasmic reticulum (SR) junctions, often in t tubules (Fig. 1A); however, some α₁-pumps also are located at PM-SR junctions (6, 66, 90). Because α₂, but not α₁, has a high affinity for ouabain in rodents, it is likely that low EO concentrations preferentially elevate the local Na⁺ concentration at these PM-SR junctions ([Na⁺]_PL) (19, 107). Moreover, type 1 Na⁺/Ca²⁺ exchangers (NCX1) colocalize with α₂-pumps, close to SR Ca²⁺ pumps (SERCA) at cardiac PM-SR junctions (66, 90, 109), in Ca²⁺ signaling units or “PLasmERosomes” (Fig. 1A) (12, 42). This enables “privileged communication” among the α₂-Na⁺ pumps, NCX1 and SERCA. Thus, the elevated [Na⁺]_PL in the tiny, diffusion-retarded volume of cytosol between the PM and “junctional SR” in the PLasmERosome (114)¹ reduces NCX1-mediated Ca²⁺ extrusion and fosters net Ca²⁺ gain by the cardiomyocyte (Figs. 1B and 2). This underlies the well-documented “cardiotonic effect” of ouabain and other CTSs such as digoxin (12, 19).

EO and Other CTSs and Their Physiological Effects

In 1953, Schatzmann (86) reported that a variety of CTSs selectively inhibit the Na⁺ pump, and Szent-Gyorgi (102) postulated that the natural ligand for the Na⁺ pump CTS receptor (i.e., an endogenous CTS) modulates cardiac contraction. Nearly 25 yr later, but before Na⁺ pump α-subunit isoforms were recognized, two groups proposed that an EO-like substance, a Na⁺ pump inhibitor, behaved as both a hypertensinogenic and natriutretic agent (9, 33). That led to the search for a mammalian CTS and to the discovery of EO, an adrenocortical hormone, in human plasma (37). Mammalian EO has been analytically verified in several independent laboratories (37, 43, 44, 61, 76, 89, 103). A recent update and analysis (38) suggested that flawed chromatographic separation of the unusually polar steroid EO in plasma samples may explain why a few investigators failed to find EO in human plasma (50).

A related CTS, marinobufagenin (MBG), a bufadienolide first identified in amphibia (4), has also been linked to hypertension (24, 25). This linkage is based primarily on immunological (versus analytic) identification and on immunoneutralization by Digibind and DigiFab [commercial anti-digoxin Fab fragments with much higher affinity for digoxin and ouabain than for MBG (79, 80)] (24). Also, MBG binds preferentially to α₁-Na⁺ pumps (25, 111). This article focuses on α₂ and EO.

To determine the significance of the high-affinity α₂-ouabain-binding site and its ligand, Lingrel et al. (21) generated mice in which the binding site was mutated to a low-affinity, “ouabain-resistant” form (α₁^R/Rα₂^R/R or “α₂^R/R” mice). This mutation of just two amino acids in the ouabain-binding site does not affect Na⁺-K⁺-ATPase activity (23), pump-mediated cation transport, or baseline cardiac function (21, 112). Nevertheless, pregnant dams from this line apparently have unusually low blood pressure (BP) during the third trimester of pregnancy (73), which may indicate that the ouabain-binding site and its ligand have a physiological role in the cardiovascular system when the body is stressed. Furthermore, hypertension induced by infused adrenocorticotropic hormone or ouabain or dietary or centrally infused NaCl is prevented or greatly attenuated in α₁^R/Rα₂^R/R mice (20, 22, 48, 58, 105). Interestingly, α₁^R/Rα₂^R/R mice are also learning-impaired and exhibit reduced dopamine-mediated locomotion (85), which implies that EO and its α₂-Na⁺ pump receptor play a physiological role in these behaviors. This is likely related to the fact that all astroglia and some neurons also express α₂-Na⁺ pumps (64).

Pressure Overload-Induced Hypertrophy Depends on α₂-Na⁺ Pump Ouabain Sensitivity

In a study of the role of the ouabain-binding site in the heart, Wansapura et al. (112) made the seminal observation that TAC-induced cardiac dysfunction, measured by echocardiography, was attenuated in α₁^R/Rα₂^R/R mice. WT (α₁^R/Rα₂^S/S) mice exhibited evidence of hypertrophy after 4 wk of TAC: LV end-diastolic and end-systolic diameters (LVEDD and LVESD, respectively) were decreased and anterior wall thickness (AWth) and ejection fraction (EF) were increased versus sham-operated mice (Fig. 4).² In striking contrast, there was no alteration in any of these parameters, versus sham-operated mice, in ouabain-resistant α₁^R/Rα₂^R/R mice (Fig. 4 and see Fig. 1C, test 5), even though LVSP was significantly elevated by TAC in α₁^R/Rα₂^R/R mice (Table 1). This leads to the astonishing conclusion that counters current dogma: namely, that much of the cardiac dysfunction is not due directly to the pressure overload and cardiomyocyte stretch, per se, but to the interaction between α₂-ouabain-binding sites and their endogenous ligand.

Fig. 4.

Ouabain resistance attenuates transaortic constriction (TAC)-induced cardiac hypertrophy. The effects of 4 wk of TAC (black bars) versus sham surgery (white bars) was compared in wild type (WT; α₁^R/Rα₂^S/S), ouabain-resistant (α₁^R/Rα₂^R/R), and SWAP (α₁^S/Sα₂^R/R) mice. Left ventricular (LV) structure and function were measured by echocardiography. Results (shown as means ± SE; n values on LVEDD bars apply to all graphs) were analyzed by two-way ANOVA; groups were compared post hoc using Tukey’s test. LVEDD, LV end-diastolic diameter; LVESD, LV end-systolic diameter; AWth, anterior wall thickness; EF, ejection fraction. *P < 0.05 vs. the corresponding sham; †P <0.05 vs. TAC WT and TAC α1^R/Rα2^R/R. [Reproduced from Ref. 112 with permission.]

Table 1.

Left ventricular systolic pressure in 4 wk post-TAC and sham-operated wild-type (α₁^R/Rα₂^S/S), ouabain-resistant α₂-Na⁺ pump (α₁^R/Rα₂^R/R), and SWAP (α₁^S/Sα₂^R/R) mice

	Wild Type	α1R/Rα2R/R	SWAP
Sham operation	96 ± 3 (5)	104 ± 3 (6)	101 ± 3 (5)
TAC	162 ± 9*† (5)	133 ± 7* (8)	122 ± 12 (8)

Data are means ± SE of left ventricular systolic pressure (in mmHg); values in parentheses are numbers of animals/group. TAC, transaortic constriction. Data were analyzed by two-way ANOVA, and groups were compared post hoc using Tukey’s test.

^*P < 0.05 vs. the corresponding sham group;

^†P < 0.05 vs. the TAC value in the other two groups. [Data from Ref. 112 are presented with permission.]

The effects of TAC on α2^R/R mice, in which the α₁-isoform was mutated to make it ouabain sensitive (α₁^S/Sα₂^R/R or “SWAP” mice), were also determined (112). In sham-operated mice, LVSP was low: only 96-104 mmHg in all three strains (Table 1), and there were no differences in baseline cardiac function. After 4 wk of TAC, LVSP was not as high in SWAP mice (122 mmHg) as in the other two strains (Table 1). Nevertheless, TAC increased LVEDD and LVESD much more in SWAP mice than in the other two strains, and EF was significantly reduced in TAC versus sham-operated mice (Fig. 4 and see Fig. 1C, test 6), i.e., SWAP mice with TAC exhibited evidence of heart failure (Fig. 2). Thus, SWAP mice were even more prone to TAC-induced cardiac dysfunction than were WT mice, but why?

The Number of High-Affinity Ouabain-Binding Sites Is Apparently Crucial

There are ~1 million α₁-Na⁺ pumps and ~100,000 α₂-pumps in a neonatal rat (and, we assume, mouse) cardiomyocyte (115). Let us suppose that, during TAC, a circulating endogenous ligand is elevated to a level sufficient to block two-thirds of ouabain-sensitive pumps. In WT (α₁^R/Rα₂^S/S) mice, 67,000 pumps will be inhibited, but the other 1,037,000 pumps (94%) will still function. The 67% reduction in functional α₂-pumps in WT mice should elevate local [Na⁺]_PL but not α₁-controlled global [Na⁺]_CYT (which may, however, partially buffer the rise in [Na⁺]_PL) and thereby promote NCX1-mediated Ca²⁺ retention and enhanced Ca²⁺ signaling.

In contrast, in TAC-SWAP mice, 670,000 (α₁) pumps will be blocked and only 430,000 (α₁ + α₂) pumps (39%) will be functional, and global [Na⁺]_CYT would be expected to rise. Because PLasmERosomes are encircled by this cytosol and are, presumably, not completely impervious to Na⁺, Na⁺ will diffuse in and raise [Na⁺]_PL. In other words, ouabain-induced, NCX1-mediated Ca²⁺ gain and augmented cardiac contraction in SWAP mice should be regulated by ouabain-sensitive α₁-Na⁺ pumps when α₂-pumps are ouabain resistant (19, 111). In addition, of course, no pumps will be inhibited in α₁^R/Rα₂^R/R mice, and there will be negligible “buffering” (see above) in the case of α₁^S/Sα₂-pumps, as in humans. This raises the possibility that susceptibility to cardiac dysfunction during pressure overload is related to the number of cardiac ouabain-sensitive Na⁺ pumps that are inhibited by the endogenous ligand. [As discussed below, however, ouabain-induced activation of protein kinases may also be involved.]

Immunoneutralization of EO Attenuates Hypertrophy

Wansapura et al. (112) also performed another test of the role of the ouabain-binding site and its endogenous ligand: they administered Digibind to immunoneutralize ouabain (79, 80) in TAC and sham-operated SWAP mice 2 wk after the surgery. Two weeks later, Digibind-treated TAC-SWAP mice showed no alterations versus sham surgery (Fig. 4) or TAC-α₁^R/Rα₂^R/R mice in heart weight, LVEDD, LVESD, AWth, or EF (Fig. 5). In other words, the TAC-induced cardiac dysfunction was prevented or reversed by Digibind (see Figs. 1C, test 6, and 2).

Fig. 5.

Digibind, which immunoneutralizes ouabain, attenuates transaortic constriction (TAC)-induced cardiac hypertrophy in SWAP mice. Ouabain-resistant (α₁^R/Rα₂^R/R; n = 6) mice and SWAP (α₁^S/Sα₂^R/R; n = 7) mice were all infused with Digibind (50 μg/day) by minipump via a jugular vein catheter during the final 2 wk of TAC. The echocardiographic data (means ± SE) after 4 wk of TAC are shown. In contrast to results in the absence of Digibind (Fig. 2), there were negligible differences between the two genotypes. HW, heart weight; LVEDD, LV end-diastolic diameter; LVESD, LV end-systolic diameter; AWth, anterior wall thickness; EF, ejection fraction. [Reproduced from Ref. 112 with permission.]

Fig. 6.

A: cardio-specific knockout of α₂-Na⁺ pumps (α₂-KO) attenuates the increase in the heart weight-to-body weight ratio (HW/BW) after 9 wk of transaortic constriction (TAC). Sham-operated mice (0 wk) are indicated by the white bars and TAC mice by the black bars. The data, (reproduced from Ref. 82 with permission) are shown as means ± SE; numbers of mice are in parentheses. Note that the data from TAC weeks 3 and 6, presented in the original report, are omitted because each group contained only 2−3 mice. These data were included in the original statistical analysis by two-way ANOVA; they may have contributed to the failure to find a significant difference between wild-type (WT) and α₂-KO mice after 9 wk of TAC (a type II error). B–D: cardio-specific α₂-KO also delayed the development of TAC-induced cardiac hypertrophy [end-systolic volume (ESV; B); end-diastolic volume (EDV; C); and ejection fraction (EF; D)]. The echocardiographic results in WT and α₂-KO mice after 3, 6, and 9 wk of TAC (means ± SE; n = 9 of each genotype) were analyzed by two-way ANOVA. Post hoc comparisons indicated that ESV was significantly lower (P < 0.05) in α₂-KO than in WT mouse hearts at 3 and 6 wk and in ESV at 3 wk; EF was significantly greater in α₂-KO than WT mice at 3 and 6 wk. [Reproduced from Ref. 82 with permission.]

These data of Wansapura et al. (112) are compelling evidence that the α₂-ouabain-binding site and its endogenous ligand play a crucial role in the mechanism(s) that directly trigger the hypertrophic changes and the cardiac dysfunction during pressure overload. Indeed, shortly after the discovery of EO, elevated plasma levels were reported in patients in congestive heart failure (31). In those patients, the EO level was inversely correlated with cardiac index, which raised the possibility that the functional impairment may be related to EO. Several subsequent cardiac hypertrophy and heart failure studies indicated that the highest EO levels were associated with the most severe cardiac dysfunction and the worst clinical outcomes (76, 96, 100). Furthermore, LV hypertrophy was directly correlated with plasma EO (100).

KO of Cardiac α₂-Na⁺ Pumps Delays TAC-Induced Cardiac Hypertrophy

The role of α₂-Na⁺ pumps in TAC-induced cardiac dysfunction was also assessed in mice with cardiac-selective KO of α₂ (“cardio-α₂-KO”) by Rindler et al. (82). Heart weight (normalized to body weight) and basal cardiac performance were indistinguishable from normal in cardio-α₂-KO mice (Fig. 6). The implication is that, in unstressed cardio-α₂-KO mice, α₁-Na⁺ pump function is sufficient to regulate NCX1-mediated Ca²⁺ transport and [Na⁺]_PL as well as [Na⁺]_CYT because, despite retarded diffusion, PLasmERosomes are likely not completely impervious to Na⁺.

TAC induced the same pressure gradient across the constriction in WT and KO mice. Nine weeks after TAC surgery, however, the heart weight-to-body weight ratio was ≈50% greater in WT-TAC mice than in KO-TAC mice (Fig. 6A). Also, at 3 and 6 wk post-TAC surgery, cardiac function was less impaired in KO mice than in WT mice, as indicated by significantly smaller LV end-systolic and end-diastolic volumes and significantly greater EF (Fig. 6, B–D). Although functional impairment in KO mouse hearts was similar to that in WT hearts by 9 wk postsurgery (Fig. 6, B–D) despite the heart weight/body weight difference, the data demonstrate that KO of cardiac α₂-Na⁺ pumps delays the development of pressure overload-induced dysfunction (see Fig. 1C, test 2). This confirms the key role of cardiac ouabain-sensitive α₂-Na⁺ pumps in the genesis of the structural and functional impairment.

TAC, the Renin-Angiotensin-Aldosterone System, and EO

If the endogenous α₂-ligand EO plays a role in the pathogenesis of TAC-induced cardiac hypertrophy, as the α₁^R/Rα₂^R/R mouse and Digibind treatment data indicate, we need to explore the role of EO more carefully. As noted, human studies have indicated a direct correlation between plasma EO and LV hypertrophy (100). However, does elevated plasma EO actually contribute to cardiac hypertrophy? In addition, if so, how? Also, why should plasma EO be elevated in mice with TAC?

Much evidence indicates that the renin-angiotensin-aldosterone system is activated in suprarenal aortic coarctation, presumably because renal perfusion pressure and blood flow is reduced (28, 67, 74). In addition to the direct effects of ANG II and aldosterone on the heart (3, 14, 88, 92), these agents act on the brain to increase sympathetic drive to the cardiovascular system (116, 121, 123). Indeed, renin-angiotensin-aldosterone system antagonists and sympatholytic agents at least partially reduce BP elevation and cardiac damage in aortic coarctation (51, 69, 83). Recent observations have demonstrated that ANG II infusion, either into the cerebral ventricles at very low concentration (38) or subcutaneously (11), also activates a novel neurohumoral pathway in the hypothalamus that, likely via the adrenal cortex, elevates plasma EO. Specific blockade of this pathway prevents the ANG II-triggered elevation of plasma EO (38). Thus, although not yet determined (a critical shortcoming), we postulate that plasma EO is elevated during TAC. This view is supported by the observation that immunoneutralization of the endogenous ligand EO with Digibind (Fig. 5) attenuates the cardiac pathology (112).

EO Helps Trigger Cardiac Hypertrophy

The next question is: how does EO affect cardiac function? One way is as a classic, selective Na⁺ pump inhibitor: it raises [Na⁺]_PL (10, 12). This increases the force of contraction [the “positive inotropic response” (10)] and thereby increases the cardiac workload (Figs. 1B and 2, left). A caveat is that some α₁-Na⁺ pumps also colocalize with NCX1 (23, 66) and, especially in SWAP mice and α₂-KO mice, can regulate cardiac contraction (23, 82). Nevertheless, under normal circumstances, α₂, because of its localization and ouabain sensitivity, preferentially modulates Ca²⁺ signaling in rodents (19). In humans, the fact that α₁, too, is ouabain sensitive suggests that α₁-pumps might also mediate the effects of EO and modulate Ca²⁺ signaling, as observed in SWAP mice (with ouabain-resistant α₂) (23). Comparison of WT (i.e., α₂-sensitive) and SWAP (i.e., α₁-sensitive) mouse cardiomyocyte responses to low-dose ouabain indicated, however, that α₂-Na⁺ pumps preferentially regulates SR Ca²⁺ release, Ca²⁺ transients, and cardiac inotropy (19). Nevertheless, the accelerated hypertrophy (and failure) in TAC-SWAP mice (112) (Fig. 4) suggests that α₁-pumps likely also contribute significantly to the pathology when plasma EO is elevated in humans.

A second possible mechanism by which EO affects cardiac structure and function relates to the fact that ouabain triggers cardiomyocyte Na⁺ pump-dependent, but cation transport-independent, protein kinase cascade-mediated hypertrophic signaling (Figs. 1B and 2, right) (26, 56, 57, 118). Furthermore, prolonged (48-72 h) exposure of cultured rat cardiomyocytes to 50 nM to 100 μM ouabain increases NCX1 protein expression (11, 68, 106), a cellular modification often associated with the progression to heart failure (71, 84, 101). Nanomolar ouabain exerts similar effects on protein kinase signaling and NCX1 expression in human (53) and rodent (78, 124) vascular myocytes. Some authors have suggested that low-ouabain affinity α₁-Na⁺ pumps in rodents trigger the cardiomyocyte signaling cascade (17, 56, 57, 118). Rodent α₁-pumps, however, do not respond to the low nanomolar ouabain/EO concentrations (21, 98) that are expected in the circulation of animals stimulated with ANG II (38). The implication is that this low concentration ouabain/EO-activated signaling is mediated by cardiac α₂-Na⁺ pumps, although this needs to be tested directly (e.g., in WT vs. α₂^R/R-cardiomyocytes).

It seems likely that both of the EO-dependent, α₂-mediated mechanisms, the cardiotonic effect and the protein kinase cascade activation, are involved in the generation of cardiac hypertrophy. The cardiotonic effect, as well as the stimulated protein kinase-mediated signaling and “reprogramming” of Ca²⁺ transporter expression, should, early on, enhance cardiac work and promote hypertrophy but, if sustained, should progress to heart failure (Fig. 2) (10).

The Paradox of Cardiac α₂ KO Versus α₂ Overexpression

One might anticipate that, if TAC-induced cardiac dysfunction is delayed and/or attenuated in cardio-α₂-KO mice, cardiac-specific overexpression of α₂-Na⁺ pumps in transgenic (TG) mice should have the opposite effect: it should sensitize the heart to pressure overload. This is not the case, however, as TAC-induced cardiac hypertrophy was also attenuated in cardio-TG(α₂) but not cardio-TG(α₁) mice, i.e., the attenuating effect is α₂ specific (15) (see Fig. 1C, tests 3 and 4). As was true of cardio-α₂-KO mice, baseline cardiac function was normal in cardio-TG(α₂) mice. However, the expected, progressive TAC-induced increase in the ventricular weight-to-body weight ratio was attenuated in TG(α₂) but not TG(α₁) mice after 10−12 wk of TAC (Fig. 7, A and B). Several other measurements also indicate that TG(α₂) but not TG(α₁) mouse hearts were at least partially protected from TAC-induced hypertrophy. For example, the expected TAC-induced increases in lung weight (Fig. 7C), a measure of impending heart failure (52), and cardiomyocyte cross-sectional area (Fig. 7, D and E) were markedly diminished in TG(α₂) mice. Also, the increased expression of hypertrophic markers (e.g., atrial natriuretic factor and brain natriuretic peptide) and decreased myocyte fractional shortening (Fig. 7F) were attenuated in TG(α₂) hearts (15).

Fig. 7.

Cardio-specific overexpression of α₂- but not α₁-Na⁺ pumps attenuates transaortic constriction (TAC)-induced cardiac hypertrophy: echocardiographic and histological evidence. Transgenic α₂ [TG(α₂)] mice exhibited less cardiac hypertrophy than wild-type (Wt) mice or transgenic α₁ [TG(α₁)] mice after 10-12 wk (10w or 12w) of TAC. Sham-operated mice served as controls. This was exemplified by the left ventricular weight-to-body weight ratio (VB/BW; A and B), the lung weight-to-body weight ratio (LW/BW; C), cardiomyocyte cross-sectional area (D and E), and fractional cardiomyocyte shortening (FS; F). The bar graphs (reproduced from Ref. 15 with permission) are presented as means ± SE; n values are shown on the bars. Statistical analysis of the data was performed using unpaired two-tailed t-tests with Prism 5 (Graphpad Software). *P <0.05 vs. sham; #P <0.05 vs. WT TAC.

At baseline, the amplitudes of the Ca²⁺ transients and SR Ca²⁺ load were similarly decreased (vs. WT) in TG(α₁) and TG(α₂) mouse hearts (15). NCX1-mediated Ca²⁺ extrusion, however, was significantly more rapid in TG(α₂) hearts than in hearts of WT or TG(α₁) mice (Fig. 8A). This might indicate that NCX1 is functionally more closely coupled to α₂ than α₁ (16). Interestingly, cardiomyocyte [Na⁺]_CYT was ~33% lower in α₁-TG hearts than in WT or TG(α₂) hearts both at rest and during 2-Hz stimulation (Fig. 8, B and C, respectively).³ This is consistent with other evidence showing that the more abundant α₁-Na⁺ pumps are the “housekeepers” that normally maintain the low global [Na⁺]_CYT (30, 62), even in cardio-TG(α₂) mice, which overexpress α₂ by approximately threefold (15), because α₂ is still a minority (perhaps 25%) of total Na⁺ pumps (115). Thus, it is not surprising that overexpression of the low-abundance α₂-Na⁺ pumps had minimal effect on global [Na⁺]_CYT. In astrocytes, too, which have a similar α₁-to-α₂ ratio, reduced α₂-expression had little effect on global [Na⁺]_CYT but a large effect on Ca²⁺ signaling (30). Also, colocalization of cardiac α₂-pumps with NCX1 at PM-SR junctions helps drive NCX1-mediated Ca²⁺ transport (10). Furthermore, phosphorylated and total PLM are markedly downregulated in TG(α₂) but not TG(α₁) mouse hearts both at baseline and after TAC (15). Phosphorylated PLM negatively regulates both α₂ (39, 75) and NCX (34, 108). Therefore, downregulation of PLM expression and phosphorylated PLM should increase the ability of α₂ to extrude Na⁺ at PM-SR junctions and increase the trans-PM Na⁺ electrochemical gradient at the junctions. Because NCX1 is functionally coupled to α₂-Na⁺ pumps (10), the lower local [Na⁺]_CYT and steeper Na⁺ gradient would explain the enhanced efficiency of NCX1-mediated Ca²⁺ extrusion observed in TG(α₂) mouse hearts (15).

Fig. 8.

A: Na⁺/Ca²⁺ exchanger (NCX)-dependent rate of intracellular Ca²⁺ concentration ([Ca²⁺]_i) decline after caffeine-induced unloading of the sarcoplasmic reticulum was faster in transgenic α₂ [TG(α₂)] than TG(α₁) or wild-type (Wt) mouse cardiomyocytes. NCX was blocked by 10 mM NiCl₂ (black bars), which lowered the rate of [Ca²⁺]_i decline to about the same level in all three genotypes. *P < 0.05 vs. control (no nickel); #P < 0.05 vs. Wt mice with the same treatment. B and C: intracellular Na⁺ concentration ([Na⁺]_i) was reduced by about 33% in TG(α₁) but not TG(α₂) versus sham cardiomyocytes, both at rest (B) and during 2-Hz stimulation (C). This reduction of [Na⁺]_i did not reach statistical significance, likely because too few myocytes were studied (see footnote 3 in the text). [Data reproduced from Ref. 15 with permission.]

So what is the explanation for the paradox that both KO and overexpression of cardiac α₂ result in a similar phenotype: delayed or attenuated development of pressure overload-induced cardiac hypertrophy? As summarized above, mutation of the (α₂) high-affinity ouabain-binding site to an ouabain-resistant form or immunoneutralization of its ligand EO ameliorates hypertrophy (112). Therefore, it seems logical that, if hypertrophy depends on the number of ouabain-bound high-ouabain affinity pumps, KO of α₂ (the EO receptors) should also ameliorate the hypertrophy (82) if NCX1-mediated Ca²⁺ transport is then regulated by ouabain-resistant α₁-pumps. The α₁-pumps can apparently prevent [Na⁺]_PL from rising excessively, and they do not respond to elevated plasma EO. Conversely, making α₁ ouabain sensitive (i.e., greatly increasing the number of high-ouabain affinity pumps: SWAP mice) should have an augmenting effect (112). However, to understand why α₂ overexpression should also ameliorate the hypertrophy, consider how EO works: by binding to α₂-Na⁺ pumps, EO blocks those transporters and should elevate [Na⁺]_PL. With ≈100,000 ouabain-sensitive α₂ and 1,000,000 ouabain-resistant α₁ Na⁺ pumps per cardiomyocyte (115), and postulating, as above, that ≈67% of the α₂ pumps are inhibited when plasma EO rises after TAC, ≈33,000 of the α₂-pumps will still be functional. However, with a threefold overexpression of α₂ (15) and thus a total of 300,000 α₂-pumps, even if 67% are inhibited, ≈100,000 ouabain-sensitive α₂-Na⁺ pumps will remain functional. This is the normal complement and is apparently sufficient to sustain ordinary functions including maintaining [Na⁺]_PL at the normal level, thereby minimizing TAC-induced cardiac hypertrophy and remodeling in TG(α₂) mice.

TG(α₁) Mice Support the Concept of a Restricted Cytosolic Compartment at PM-SR Junctions

Much of the preceding discussion examines the evidence that the binding of ouabain and inhibition of cardiac α₂-Na⁺ pumps play a key role in the development of cardiac hypertrophy. The SWAP mouse data suggest that when α₁ is mutated to an ouabain-sensitive form, it, too, can mediate these pathophysiological processes. As noted above, two- to threefold or more (estimated from immunoblots; quantitation was not given) overexpression of ouabain-resistant cardiac α₁ apparently lowers [Na⁺]_CYT more than does α₂ overexpression (Fig. 8, B and C) but does not ameliorate TAC-induced hypertrophy (15). This fits the view (10) that α₂-Na⁺ pumps and NCX1 have “privileged access” to tiny, diffusion-restricted cytosolic compartments at PM-SR junctions, even though there is no direct evidence (but see Ref. 77) that Na⁺ diffusion is restricted. Additionally, possible differences between ouabain-triggered α₁- and α₂-mediated protein kinase cascade signaling may be a factor. If, as inferred above, however, EO is elevated by TAC, the overexpressed ouabain-resistant α₁ should not be responsive to EO, in contrast to the situation when α₁ is mutated to an ouabain-sensitive form (112).

Summary and Conclusions

The three key reports reviewed here (15, 82, 112) provide a total of six different experimental manipulations related to Na⁺ pumps and ouabain binding in cardiomyocytes (see Fig. 1C): 1) mutation of the α₂-Na⁺ pump high-affinity ouabain-binding site to an ouabain-resistant form (Fig. 1C, test 1); 2) KO of cardiac high-ouabain affinity α₂-Na⁺ pumps (Fig. 1C, test 2); 3) immunoneutralization of circulating EO (Fig. 1C, test 6); 4) overexpression of cardiac myocyte high-ouabain affinity α₂-Na⁺ pumps (Fig. 1C, test 3); 5) overexpression of cardiac myocyte low-ouabain affinity α₁-Na⁺ pumps (Fig. 1C, test 4); and 6) mutation of the normally low-affinity α₁-Na⁺ pump ouabain-binding site to a high-affinity site in α₂-ouabain-resistant mice (“ SWAP” mice; Fig. 1C, test 5). Tests 1–3 and 6 attenuate or delay TAC-induced structural and functional changes that constitute cardiac hypertrophy. While overexpression of ouabain-resistant α₁-Na⁺ pumps (test 4) has no effect, simply making α₁-ouabain-sensitive in α₂^R/R mice (test 5) accelerates the hypertrophy. Taken together, these findings provide compelling evidence that EO and cardiac high-ouabain affinity α₂-Na⁺ pumps play a fundamental role in the pathogenesis of pressure overload-induced cardiac hypertrophy in rodents. Nevertheless, the SWAP mouse results raise the possibility that, in humans, ouabain-sensitive α₁ may also be involved.

These data also suggest that blockade of the EO-binding site on α₂-Na⁺ pumps may be a novel strategy to ameliorate the hypertrophy and eventual heart failure. Indeed, PST2238 (rostafuroxin), a digoxigenin derivative that preferentially blocks ouabain binding but not pump-mediated cation transport (27, 97), appears to be an effective antihypertensive agent in rodents (26) and selected humans (46) with high EO and elevated BP. The use of rostafuroxin in humans may be limited (99), however, because of its relatively low affinity for the Na⁺ pump ouabain-binding site (97). Thus, there is need for a more effective EO antagonist that might be useful for attenuating hypertension-induced cardiac hypertrophy. Recent evidence indicates that very low (quarter) doses of multiple antihypertensive agents may have therapeutic benefit with reduced side effects (5). In this context, it may be worth considering the use of very low-dose digoxin, an ouabain antagonist (60, 97), as adjunct therapy in hypertensives with cardiac hypertrophy.

In summary, the three studies that are the focus of this review (15, 82, 112) complement and extend the numerous reports that have linked elevated plasma EO to hypertension as well as cardiac hypertrophy and failure in humans (summarized in Refs. 10 and 36). They elucidate the underappreciated role of the EO-α₂-Na⁺ pump endocrine pathway in cardiovascular physiology and pathophysiology.

GRANTS

This work was supported in part by American Heart Association Grant 15GRNT24940022 and National Heart, Lung, and Blood Institute Grant HL-107555 and funds from the University of Maryland Foundation (Hypertension Center Account).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

M.P.B. conceived and designed the work, prepared and revised the manuscript and figures, and approved the final version of the manuscript.