Acute effects of 17β-estradiol on myocardial pH, Na+, and Ca2+ and ischemia-reperfusion injury

Steven E. Anderson, Dawn M. Kirkland, Andrea Beyschau, Peter M. Cala

Abstract

Evidence suggests that 1) ischemia-reperfusion injury is due largely to cytosolic Ca2+ accumulation resulting from functional coupling of Na+/Ca2+ exchange (NCE) with stimulated Na+/H+ exchange (NHE1) and 2) 17β-estradiol (E2) stimulates release of NO, which inhibits NHE1. Thus we tested the hypothesis that acute E2 limits myocardial Na+ and therefore Ca2+ accumulation, thereby limiting ischemia-reperfusion injury. NMR was used to measure cytosolic pH (pHi), Na+ (NaMath), and calcium concentration ([Ca2+]i) in Krebs-Henseleit (KH)-perfused hearts from ovariectomized rats (OVX). Left ventricular developed pressure (LVDP) and lactate dehydrogenase (LDH) release were also measured. Control ischemia-reperfusion was 20 min of baseline perfusion, 40 min of global ischemia, and 40 min of reperfusion. The E2 protocol was identical, except that 1 nM E2 was included in the perfusate before ischemia and during reperfusion. E2 significantly limited the changes in pHi, NaMath and [Ca2+]i during ischemia (P < 0.05). In control OVX vs. OVX+E2, pHi fell from 6.93 ± 0.03 to 5.98 ± 0.04 vs. 6.96 ± 0.04 to 6.68 ± 0.07; NaMath rose from 25 ± 6 to 109 ± 14 meq/kg dry wt vs. 25 ± 1 to 76 ± 3; [Ca2+]i changed from 365 ± 69 to 1,248 ± 180 nM vs. 293 ± 66 to 202 ± 64 nM. E2 also improved recovery of LVDP and diminished release of LDH during reperfusion. Effects of E2 were diminished by 1 μM Nω-nitro-l-arginine methyl ester. Thus the data are consistent with the hypothesis. However, E2 limitation of increases in [Ca2+]i is greater than can be accounted for by the thermodynamic effect of reduced NaMath accumulation on NCE.

  • myocardial ischemia
  • Na+/H+ exchange
  • Na+/Ca2+ exchange
  • nuclear magnetic resonance
  • ischemic biology
  • ion channels/membrane transport
  • transplantation

it is clear that our understanding of the effects of endogenous and exogenous estrogen on the cardiovascular system, and in particular on its role in modifying susceptibility to ischemic injury, is deficient. It is more important than ever that fundamental research be conducted to understand the basis for the effects of estrogen (50). Because results of various studies are inconsistent and/or difficult to interpret, we have taken a reductionist approach to testing the acute effects of exogenous 17β-estradiol (E2) on ischemic myocardium within the context of the well-accepted paradigm that ischemic injury is largely the result of the following chain of events. Anaerobic metabolism decreases cytosolic pH (pHi), which stimulates pH-regulatory Na+/H+ exchange (NHE1) to increase Na+ uptake and thereby increase intracellular Na+ (NaMath). Increases in Math decrease the driving force for cytosolic Ca2+ efflux via Na+/Ca2+ exchange (NCE) and thereby cause an increase in cytosolic [Ca2+] ([Ca2+]i). Increases in [Ca2+]i cause a cascade of events ending in injury, apoptosis, or necrosis (3, 7, 31, 37, 46, 49). In the heart, there are questions about whether the bulk of Na+ and Ca2+ entry occurs during ischemia or reperfusion, but there is a growing consensus that much of the Ca2+ entry is Na+ dependent (36). Numerous studies have shown that E2 stimulates NOS activity and/or the release of nitric oxide (NO) in the heart (20, 39, 40) and that female hormone-linked changes in Na+ and Ca2+ accumulation and protection of myocardium from ischemia-reperfusion injury are NO dependent (14–16). Although these studies implicate the sarcolemmal NCE in the gender-dependent response, they do not make it clear whether the effect of female hormones is directly on NCE itself or on other transporters that contribute to changes in the Na+ or Ca2+ gradients or cell membrane potential (Em), the thermodynamic determinants of flux via NCE (38, 46). A study of isolated mouse cardiac myocytes overexpressing NCE in which a metabolic inhibition protocol was used to mimic ischemia argued that Math-dependent increases in [Ca2+]i are diminished by female hormone, but the mechanism of Math accumulation has not been determined (48). More recent experiments performed at the same laboratory led the investigators to conclude that the acute cardioprotective effect of E2 during metabolic inhibition may be mediated by an anti-oxidant effect (47). The latter study, however, was conducted with 100 nM E2. This dose is nearly 100 times greater than peak E2 in cycling human females and nearly 500 times greater than that in mice. Thus the mechanism of the E2 effects on Math and Math and protection from ischemic injury for physiological levels of E2 remains unclear.

Given the aforementioned evidence that E2 stimulates release of NO and that NO inhibits NHE1 in cardiac myocytes (22), we tested the hypothesis that acute exposure of isolated rat hearts to E2 limits Na+ uptake during ischemia-reperfusion and thus limit Na+-dependent increases in [Ca2+]i as well as associated ischemia-reperfusion injury. The hypothesis further predicts that this effect of E2 is inhibited by the NO synthesis inhibitor Nω-nitro-l-arginine methyl ester (l-NAME).

In addition, to test more directly the hypothesis that E2 inhibits NHE1, we used the NH4Cl washout protocol (8) to acidify the heart under normoxic, normal perfusion, HEPES-buffered conditions with or without E2. This protocol allowed us to assess the response of NHE1 to intracellular acidification similar in magnitude to that which occurs during ischemia, but with fewer uncontrolled parameters (28).

The results are consistent with the hypothesis that E2 inhibits NHE1 in a NO-dependent manner. That is, in the HEPES-buffered NH4 washout protocol, E2 diminished the rate of proton efflux. In the ischemia protocol, E2 limited cytosolic Math, Math, and Math accumulation during ischemia. Furthermore, these effects of acute E2 improved recovery of function and limited lactate dehydrogenase (LDH) release during reperfusion. However, E2 limitation of increases in [Ca2+]i during ischemia is greater than can be accounted for by the thermodynamic effect of the Na+ gradient on NCE, suggesting that E2 also attenuates NCE activity.

Thus the results of these studies provide new insights into the mechanisms of gender- or female hormone-dependent effects on myocardial ischemia-induced Math accumulation and injury. They go beyond previous studies to provide strong evidence that acute E2 inhibits NHE1 during ischemia and more generally after intracellular acidification. Furthermore, thermodynamic arguments provide strong circumstantial evidence that E2 also inhibits NCE independently of changes in [Na+]i during ischemia. While the evidence supports the conclusion that E2 inhibits NHE1 via NO, how it might inhibit NCE remains to be investigated. These results were previously reported in abstract form (23).

METHODS

General.

The methods used were modified from those previously reported (3, 6, 31). Briefly, 12- to 13-wk-old ovariectomized Sprague-Dawley rats (OVX) were purchased from Charles River 1 wk after ovariectomy. Animals were treated according to the “Guiding Principles in the Care and Use of Animals” and an Institutional Animal Care and Use Committee-approved protocol. One week after delivery, OVX were anesthetized with pentobarbital sodium (65 mg/kg ip) and heparinized (1,000 USP U/kg iv). The hearts were removed, and the aortas were cannulated and perfused at a constant rate (10–12 ml/min). All experiments were conducted at 36 ± 1°C. Control perfusate contained (in mmol/l) 133 NaCl, 4.75 KCl, 1.25 MgCl2, 1.82 CaCl2, 25 NaHCO3, and 11 dextrose, as well as 10 U/l insulin (44), and was equilibrated with 95% O2-5% CO2, which provided a pH of 7.35–7.45. Left ventricular developed pressure (LVDP; end-systolic minus end-diastolic pressure) was measured using a fluid-filled balloon placed in the left ventricle and secured by means of a ligature immediately proximal to the mitral valve. LVDP and perfusion pressure were monitored continuously by strain gauge transducers connected to a MacLab. Left ventricular volume was adjusted as necessary to maintain end-diastolic pressure at 5 ± 5 cmH2O.

The normoxic acidification protocol was conducted using HEPES-buffered perfusate (HR), which was identical to the perfusate described above, except that 20 mM HEPES acid plus 8 mM NaOH were substituted for 25 mM NaHCO3, and the perfusate was equilibrated with 100% O2 and titrated to pH 7.40 ± 0.05. The protocol consisted of 20-min baseline perfusion with HR, 35-min perfusion with HR + 10 mM NH4Cl, followed by 30-min perfusion with HR (NH4Cl washout). For acute E2 and l-NAME experiments, 1 nM E2 and/or 1 μM l-NAME was added to both the NH4Cl solution (5 min before NH4Cl washout) and the NH4Cl washout solution. 31P NMR was used to measure pHi as described below. Proton efflux rates were calculated as the product of the change in pHi during the first 5 min of recovery after acidification multiplied by cardiac myocyte buffer capacity. The buffer capacity (mM/pH) equals −28 (pHiMath) + 222.6 (28), where pHiMath is the pHi at the beginning of the recovery interval.

The control ischemia protocol consisted of 20-min baseline perfusion, followed by 40-min ischemia, followed by 40-min reperfusion. To limit glycolysis (1), 5.5 mM 2-deoxy-d-glucose was substituted for dextrose 10 min before ischemia. The E2 protocol was identical, except that 1 nM E2 was added to the perfusate used during the baseline and reperfusion intervals. Each of the protocols was also conducted with the addition of 1 μM l-NAME to the perfusate. When l-NAME was added, it was added together with E2 or when E2 would have been added. Initiation of 40 min of ischemia was designated t = 0 min. 23Na, 31P, and 19F NMR were used to measure Math, pHi and high-energy phosphates, and [Ca2+]i, respectively. To measure Math, 7.5 mmol/l dysprosium triethylenetetraminehexaacetic acid was substituted for NaCl in the perfusate isosmotically, and Ca2+ was added until the perfusate concentration reached 1.8–2 mmol/l as measured using Ca2+ electrode (31). To measure [Ca2+]i, hearts were perfused for 30 min with perfusate containing the acetoxymethyl ester of 5-fluoro-1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (FBAPTA) at 8 μmol/l (6). FBAPTA was then washed out of the extracellular space with control solution for 15 min before baseline measurement of [Ca2+]i was performed. After perfusions were complete, hearts used to measure 23Na were weighed wet and dried to constant weight (at least 48 h) at 65°C to determine dry weight. Their wet and dry weights were 1.22 ± .04 and 0.19 ± 0.01 g, respectively, and there were no significant differences between treatments.

NMR spectroscopy.

Experiments were conducted using a Bruker AMX400 spectrometer. 23Na, 31P, and 19F spectra were generated from the summed free induction decays of 1,000, 148, and 1,500 excitation pulses (90°, 60°, and 45°) using 2-, 4-, and 2-kilobyte word data files and ±4,000-, ±4,000-, and ±5,000-Hz sweep widths, respectively. Data files were collected over 5-min intervals, but to improve signal-to-noise ratio for 19F measurement of [Ca2+]i, two 5-min files were added together. Because the NMR signal intensity reflected the average for the interval over which data were collected, data in the figures correspond to the midpoint of the appropriate 5- or 10-min acquisition interval. NMR data reflect the signal collected from the entire heart, including both ventricles and both atria.

Math (in meq/kg dry wt) was calculated from the calibrated area of the “unshifted” intracellular peak after subtracting the shifted extracellular peak as previously described (3, 5, 31). This method measures Math as an amount, and to minimize assumptions, we report it as such. [Ca2+]i (in nmol/l cytosolic water) was calculated as the product of the Ca2+-FBAPTA dissociation constant (Kd) and the ratio of the areas of the Ca2+-bound and Ca2+-free peaks in the FBAPTA spectrum (6). The Kd for Ca2+-FBAPTA was 294 nM. pHi was determined from the chemical shift of the inorganic phosphate (Pi) resonance (with reference to control phosphocreatine) calibrated at 37°C (2). High-energy phosphates are reported as percentage of control peak intensity (6).

To assess ischemia-reperfusion injury, total LDH released was measured from timed collections of perfusate leaving the heart for 40 min of reperfusion as previously described for creatine kinase (43). LDH (in IU/g dry wt) was measured spectrophotometrically (Pointe Scientific, Lincoln Park, MI).

Unless otherwise stated, results are reported as means ± SE. For the ischemia experiments, analysis of variance for repeated measures was used to test for differences between treatments. When differences between treatments were found, the Student-Newman-Keuls multiple-comparison test was used to determine which treatments were different and the times at which differences between treatments occurred. Only the latter are indicated in figures. For the NH4Cl washout experiments, the Bonferroni correction was used to compare proton efflux rates. Because overpowered studies are wasteful of resources (17) and/or raise serious ethical issues when animals are used (30), sample sizes were limited to those required to reject the null hypotheses with P < 0.05.

RESULTS

Estrogen limits proton efflux via NHE after normoxic acidification.

The data in Fig. 1 demonstrate that acute E2 (1 nM) added to the perfusate 5 min before NH4Cl washout (8) diminished the proton efflux rate after normoxic acidification in HEPES buffer (P = 0.024) and that the effect of acute E2 was diminished by l-NAME (P = 0.032). In control OVX, the mean proton efflux rate after NH4Cl washout was 1.58 ± 0.23 mM/min compared with 0.58 ± 0.22 mM/min in OVX+E2 and 1.51 ± 0.30 mM/min in OVX+E2+l-NAME. There were no significant differences in pHi between groups after acidification before pH regulation (pHiMath) in OVX (6.86 ± 0.06 mM/min), OVX+E2 (6.85 ± 0.06 mM/min), or OVX+E2+l-NAME (6.83 ± 0.04 mM/min). Under these conditions, proton efflux is commonly accepted to occur almost completely via NHE1 (28). Thus the data support the hypothesis that E2 limits pH-regulatory NHE1 activity. Again, the observation that addition of 1 μM l-NAME significantly limited the E2 effect supports the hypothesis that the acute E2 effect is mediated by NO.

Fig. 1.

In perfused hearts from ovariectomized rats (OVX), acute 17β-estradiol (E2; 1 nM) diminished the proton efflux rate after normoxic acidification and the effect of E2 is diminished by 1 μM Nω-nitro-l-arginine methyl ester (l-NAME). Proton efflux rate after NH4Cl washout (in mM/min) is plotted vs. treatment. *P < 0.05 vs. OVX+E2; number of experiments is shown in parentheses. Note that in contrast to the ischemia protocol (results summarized in all other figures), 2-deoxy-d-glucose was not used to inhibit glycolysis during the normoxic acidification protocol, the results of which are shown.

Estrogen limits intracellular proton accumulation during ischemia.

The data in Fig. 2 are consistent with the hypothesis that acute E2 limits the accumulation of protons in ischemic myocardium (P < 0.0001) and that this response is mediated by NO because it is diminished by 1 μM l-NAME (P < 0.01). There were no significant differences between pHi in any of the groups, except during ischemia. Under control conditions, mean pHi fell from 6.93 ± 0.03 to 5.98 ± 0.04 during 40 min of ischemia, compared with 6.96 ± 0.04 to 6.68 ± 0.07 for OVX+E2, 6.94 ± 0.05 to 6.34 ± 0.13 for OVX+E2+l-NAME, and 6.92 ± 0.04 to 6.21 ± 0.08 for OVX+l-NAME. Experiments were also conducted in which 1 μM d-NAME was substituted for 1 μM l-NAME in animals exposed to E2. The data are also consistent with the interpretation that l-NAME limits the effect of E2 but d-NAME (not shown) does not. That is, in OVX+E2 hearts, pHi was significantly higher during ischemia with d-NAME than with l-NAME treatment (P < 0.01), and there was no significant difference between OVX+E2 hearts with or without d-NAME. The fact that l-NAME alone limited the ischemia-induced fall in pHi suggests that there is a NO-dependent response to ischemia that tends to lower pHi in the absence of E2. We cannot explain this response, but it does not alter the conclusion that the acute effect of E2 on pHi during ischemia is diminished by l-NAME and thus must be NO dependent.

Fig. 2.

In perfused hearts from OVX, acute E2 (1 nM) diminished the fall in intracellular pH (pHi) during ischemia, and this effect is limited by l-NAME (1 μM). pHi is plotted vs. time before, during, and after ischemia with E2 (triangles) or without E2 (circles) and with l-NAME (closed symbols) or without l-NAME (open symbols). *P < 0.05 vs. OVX control; †P < 0.05 vs. OVX+E2; §P < 0.05 vs. OVX+l-NAME; number of experiments is shown in parentheses. l-NAME±E2 was not significantly different.

Estrogen limits Nai+ accumulation during ischemia.

The data in Fig. 3 are consistent with the hypothesis that E2 inhibits Math accumulation during ischemia (P < 0.05) and that this response is mediated by NO because it is diminished by 1 μM l-NAME (P < 0.01). Under control conditions, mean Math rose from 25 ± 6 to 109 ± 14 meq/kg dry wt during 40 min of ischemia, compared with 25 ± 1 to 76 ± 3 meq/kg dry wt for OVX+E2, 34 ± 10 to 136 ± 6 meq/kg dry wt for OVX+E2+l-NAME, and 31 ± 5 to 129 ± 9 meq/kg dry wt for OVX+l-NAME. There were no significant differences between any groups before or after ischemia, except that Math in the OVX+E2+l-NAME group was greater than that in the OVX+E2 group for the first interval of reperfusion (132 ± 18 vs. 52 ± 5 meq/kg dry wt; P < 0.05).

Fig. 3.

In perfused hearts from OVX, acute E2 (1 nM) diminished the increase in intracellular Na+ (Formula) during ischemia, and this effect was limited by l-NAME (1 μM). Formula (in meq/kg dry wt) is plotted vs. time before, during, and after ischemia with E2 (triangles) or without E2 (circles) and with l-NAME (closed symbols) or without l-NAME (open symbols). *P < 0.05 vs. OVX control; †P < 0.05 vs. OVX+E2; §P < 0.05 vs. OVX+l-NAME; number of experiments is shown in parentheses.

Estrogen limits increases in [Ca2+]i during ischemia.

The data in Fig. 4 demonstrate that E2 limited cytosolic Ca2+ accumulation during ischemia (P < 0.01) and that this response was inhibited by 1 μM l-NAME (OVX+E2+l-NAME vs. OVX+E2; P < 0.05). Mean [Ca2+]i rose during 40 min of ischemia from 365 ± 69 to 1,248 ± 180 nM in control OVX compared with 293 ± 66 to 202 ± 64 nM in OVX+E2, 232 ± 54 to 900 ± 154 nM in OVX+E2+l-NAME, and 486 ± 126 to 1,700 ± 291 nM in OVX+l-NAME. This result is consistent with the hypothesis that an NO-dependent effect of E2 inhibits Math accumulation during ischemia.

Fig. 4.

In perfused hearts from OVX, acute E2 (1 nM) diminished the increase in cytosolic Ca2+ concentration ([Ca2+]i) during ischemia, and this effect was limited by l-NAME (1 μM). [Ca2+]i (in nM) is plotted vs. time before, during, and after ischemia with E2 (triangles) or without E2 (circles) and with l-NAME (closed symbols) or without l-NAME (open symbols). *P < 0.05 vs. OVX control; †P < 0.05 vs. OVX+E2; §P < 0.05 vs. OVX+l-NAME; number of experiments is shown in parentheses.

Estrogen limits LDH release and improves recovery of function after ischemia.

The data are also consistent with the hypothesis that E2, by diminishing cytosolic Math accumulation during ischemia, improves recovery of LVDP during reperfusion (P < 0.05; Fig. 5) and limits injury as indicated by the decrease in LDH release during reperfusion (P < 0.05; Fig. 6). Under control conditions, mean LVDP (%baseline) rose to 2 ± 1% after 40 min of reperfusion, compared with 19 ± 6% for OVX+E2, 4 ± 2% for OVX+E2+l-NAME, and 10 ± 5% for OVX+l-NAME. There were no significant differences in LVDP between any groups before or during ischemia. Again, as predicted by the hypotheses, LDH release during reperfusion was diminished by acute E2 (P < 0.05). During reperfusion, LDH release in the control OVX group was 94 ± 43 IU/g dry wt, compared with 18 ± 7, 52 ± 17, and 113 ± 26 IU/g dry wt in the OVX+E2, OVX+E2+l-NAME, and OVX+l-NAME groups, respectively.

Fig. 5.

In perfused hearts from OVX, acute E2 (1 nM) improved recovery of left ventricular developed pressure (LVDP). LVDP (%baseline) is plotted vs. time during reperfusion with E2 (triangles) or without E2 (circles) and with l-NAME (closed symbols) or without l-NAME (open symbols). *P < 0.05 vs. OVX control; number of experiments is shown in parentheses.

Fig. 6.

In perfused hearts from OVX, acute E2 (1 nM) diminished the release of lactate dehydrogenase (LDH) during reperfusion. Total LDH released during reperfusion (in IU/g dry wt) is plotted vs. treatment. *P < 0.05 vs. OVX+E2; number of experiments is shown in parentheses.

DISCUSSION

The general hypothesis that ischemia-reperfusion injury is largely the result of cytosolic Math accumulation mediated by NCE subsequent to pH-regulatory, NHE1-mediated increases in Math is widely accepted (7, 31, 37, 45). It also has been well demonstrated that E2 increases activity of NOS and NO release during ischemia and reperfusion (20, 39, 41). Furthermore, there is good evidence that the NO donor sodium nitroprusside, as well as 8-bromoguanosine 3′,5′-cyclic monophosphate (8-BrcGMP), inhibit pHi recovery in cardiac myocytes from male rats after NH4Cl washout in HEPES medium (22). The latter supports the conclusion that NO (via cGMP) inhibits pH-regulatory NHE in the heart. Given these findings, we tested the logical, but previously unarticulated, hypothesis that acute exposure to E2 before acidification would cause an NO-dependent limitation of NHE1. In the context of the HEPES-buffered NH4Cl protocol, this would cause an l-NAME-inhibitable decrease in proton efflux. In the context of the ischemia protocol, this would cause an l-NAME-inhibitable decrease in cytosolic Math and Math accumulation. The general hypothesis led us to further postulate that E2-dependent limitation of increases in [Ca2+]i would limit ischemia-reperfusion-induced myocardial injury. The data presented here are all consistent with these hypotheses and thus provide new insight into the mechanisms of estrogen's effects on myocardial Math, Math, and injury during ischemia-reperfusion.

Because the balance of proton production and efflux is so difficult to predict and control during ischemia, it is very hard to unequivocally determine the cause-and-effect relationships between changes in pHi and [Na+]i using that model. Therefore, we tested the hypothesis that E2 will inhibit NHE1 under normoxic, normal flow conditions of intracellular acidification using the NH4Cl prepulse technique under HEPES-buffered conditions (8, 22). Under these conditions, a known amount or bolus of protons can be “injected” into the cell, and the rate of proton efflux during pHi recovery can be ascribed to flux via NHE1 (28). The results of these experiments provide strong evidence that E2 inhibits NHE1 (Fig. 1), and because the effect of E2 is diminished by l-NAME, this result is consistent with the postulate that the E2 effect is NO mediated.

Also consistent with these hypotheses are that acute E2 limits Math and Math accumulation during ischemia and that the effect of E2 is diminished by l-NAME, indicating that it is mediated by NO (Figs. 3 and 4).

The observation that E2 limits the fall in pHi during ischemia might initially seem contradictory to our hypothesis that E2, via NO, inhibits NHE1. However, investigators at our laboratory (31) and others (42) have shown that NHE1 inhibitors either have no effect on or limit the fall in pHi during ischemia. Furthermore, inhibiting ischemia-induced Na+ uptake via other pathways also limits ischemia-induced acidification (9, 24, 34). This is because ischemia-induced Math and Math accumulation stimulates ATP hydrolysis (3, 21, 32) and thus proton production (18). In other words, during ischemia, a positive feedback loop is created whereby increased proton production increases Math and Math accumulation, which stimulates more proton production. If this occurs, inhibiting accumulation of any of the ions will have a tendency to diminish the accumulation of the others; in particular, limiting increases in [Na+]i and [Ca2+]i limits proton production and thereby ischemia-induced acidification (31). The data presented in Fig. 2 fit this paradigm in that acute E2 limitation of Math and Math accumulation (see Figs 3 and 4) is associated with a smaller decrease in pHi during ischemia, and, again, this effect of E2 was diminished by l-NAME. We cannot rule out the possibility that E2 inhibits proton production, e.g., by stimulating ATP-dependent K+ channels (29), independently of that due to inhibition of NHE1 (and thus limits stimulation of NHE1), but the data in Fig. 1 provide strong evidence that E2 inhibits the response of NHE1 to acidification under conditions in which the acidification is not E2 dependent.

Finally, the findings that acute E2 limits ischemia-reperfusion-induced decreases in recovery of LVDP (Fig. 5) and increases in LDH release during reperfusion (Fig. 6) are consistent with the hypothesis that E2-sensitive Math and Math accumulation contribute to ischemia-reperfusion injury. Although the effects are statistically significant, in this model LVDP recovery is modest. Thus whether the effects of E2 observed in this protocol would be important in a clinical setting remains to be determined.

Although the results of these studies are consistent with the hypothesis that acute E2 limits Math and thereby Math accumulation, it should also be noted that with respect to the data shown in Figs. 3 and 4, the E2-dependent limitation of cytosolic Ca2+ accumulation is greater than one would expect based on the reduction in Math accumulation alone. That is, investigators at our laboratory (4) and others (38, 46) have shown that NCE remains near equilibrium under a variety of conditions and most notably during ischemia. This behavior is interpreted as support for the notion that increases in [Ca2+]i during ischemia are mediated by NCE. In other words, if NCE is the dominant Ca2+ transport pathway, the driving force for NCE will be at or near zero; incipient forces will dissipate because of NCE flux, and the pathway will be at equilibrium. If that is the case, the reversal potential ENCE will equal Em. Conversely, if ENCE is not equal to Em, NCE is not the dominant Ca2+ transport pathway. If ENCE is calculated at the end of ischemia after acute E2 (OVX+E2 from the data in Figs. 3 and 4), the result is ENCE = −126 mV (see appendix at the end of this article). Again, this means Em would have to be −126 mV for NCE to be at equilibrium. The negative limit for Em under essentially all physiological conditions, including those in which K+ channels are stimulated, is the K+ equilibrium potential, which in this case is near −65 mV (35). [In fact, Em is likely to be nearer to −50 mV (46).] Because ENCE is far more negative than Em (EmENCE ≈ 61–76 mV), one may safely argue that after acute E2 treatment, NCE is no longer the dominant Ca2+ transport protein during ischemia.

Thus Occam's razor leads one to conclude that during ischemia after acute E2, NCE must be inhibited kinetically and/or that some other Ca2+ transport pathway is stimulated so that NCE is no longer dominant. One could argue that there may be errors in the values of [Na+] and [Ca2+] used in our calculations, but it is unlikely that the errors could be large enough to cause this large discrepancy between ENCE and Em. First, the error in our assumptions for extracellular [Na+] and [Ca2+] are likely to be <5%, and the error in our estimate of Em must be <20 mV (Em cannot be more negative than EK). Second, if the discrepancy were due to errors in measurement of Math or Math, to account for the discrepancy would require an overestimate of at least 100% in our measurement of [Na+]i or an underestimate of [Ca2+]i >90%. Again, while the above discussion is based on formal thermodynamic arguments, the same conclusions can be drawn qualitatively from the data in Figs. 3 and 4, which show that during ischemia after acute E2 treatment, nominal increases in [Na+]i are associated with no change or with decreases in [Ca2+]i. This cannot happen if NCE is the dominant Ca2+ transport pathway.

As mentioned above, one could postulate that NCE is not dominant (at equilibrium), because E2 stimulates some other pathway to remove Ca2+ from the cytosol, e.g., Ca2+-ATPase at the sarcolemma or the sarcoplasmic reticulum (13). However, in order for the ischemia-induced changes in Math and Math to be dissociated by E2 as observed in Figs. 3 and 4, a very large increase in efflux via some other pathways would be required to surpass flux via NCE unless the latter was inhibited. Thus, again, the simplest explanation for the data includes E2 inhibition of NCE activity. How this occurs requires further study.

The results of our experiments are consistent with a number of reports suggesting that female hormones play a role in Ca2+-mediated ischemic injury. These include the studies of Cross and colleagues (14, 15), who showed that in mutant mice that overexpressed either β2-adrenergic receptors or NCE, female mice were protected from ischemic injury, in contrast to males and ovariectomized females, and in the β2-adrenergic receptor overexpressors, the female resistance to ischemic injury was reduced by treatment with l-NAME. In contrast, using a simulated ischemia-reperfusion protocol with isolated rat astrocytes, Matsuda et al. (33) found that sodium nitroprusside and 8-bromo-cGMP exacerbated injury and concluded that NO and cGMP stimulated NCE. This is not consistent with our studies or with those conducted by Cross and colleagues. On the other hand, in a metabolic inhibition study using isolated mouse cardiac myocytes overexpressing NCE, Sugishita et al. (48) reported findings consistent with those reported by Cross and colleagues (14, 15) but included measurements of [Ca2+]i and [Na+]i that were consistent with our findings. That is, increases in [Na+]i and [Ca2+]i during metabolic inhibition were greater in male than in female transgenics, and gender differences were eliminated by treating the males with E2. In a more recent report, Sugishita et al. concluded that “the acute cardioprotective effect of estrogen during [metabolic inhibition] may be mediated by an ER-independent anti-oxidant action, which results in improved function of Na+-K+ ATPase” (47, p. 331). However, in the more recent study, the authors used a dose of E2 100 times greater than the dose that we used, which raises the question whether the anti-oxidant effect was physiological.

The fact that the present study was conducted with a 100-fold lower dose of E2 (near peak estrus in the rat) and that the effect is diminished by l-NAME (but not by d-NAME) provides evidence that the response that we observed is not completely due to an anti-oxidant effect of E2. Furthermore, E2-dependent limitation of increases in Math during ischemia seem to be dependent on the presence of estrogen receptor-α. Zhai et al. (51) reported that in perfused hearts of male mice, estrogen receptor-α knockouts increased Math accumulation, decreased nitrite production, and increased multiple indicators of injury during ischemia-reperfusion. Another study by investigators at the same laboratory (52) showed that in female OVX rats, chronic E2 treatment was associated with the opposite response to ischemia-reperfusion: decreased Math accumulation, increased nitrite production, and decreased injury.

Recent studies in a number of organs are consistent with the postulate that NHE1 may be inhibited by signal transduction pathways associated with increases in NO and specifically the cascade of steps through NO, cGMP, cGMP-dependent kinase (cGK) and p38 MAPK (19). These include studies from isolated hepatocytes that suggest a preconditioning effect limiting increases in [Na+]i during hypoxia is mediated by NO through cGK and p38 MAPK (11) and furthermore that atrial natriuretic peptide diminishes hypoxia-induced increases in [Na+]i by a similar pathway that inhibits NHE1 (10). These results are also consistent with studies that suggest activation of p38 MAPK inhibits the response of NHE1 to angiotensin II in vascular smooth muscle cells by phosphorylating the NHE1 at serine/threonine residues located between amino acids 671 and 714 (27).

Finally, it has been suggested that KH or, more broadly, crystalloid perfused hearts are “hyperoxic” and thus more susceptible to oxidative stress because the perfusate is commonly equilibrated with 95% O2. As is the case for respiration, the pertinent factor is not Po2 per se but rather the oxygen concentration in the solution at the site of the reaction. At the level of cardiac myocytes in the KH-perfused heart, there are at least two arguments that suggest the cells are not hyperoxic. First, there is no change in the extravascular O2 transport pathway, except perhaps for interstitial edema, which would lengthen the pathway. Second, although the actual concentration of intracellular O2 is elusive and not agreed upon, NMR measurements of intracellular O2 in perfused and in situ myocardium indicate that O2 delivery to heart cells in KH-perfused hearts is not greater than in situ (26). Furthermore, when the critical level of intracellular O2 is measured (25), KH-perfused hearts are marginally closer than blood-perfused hearts to hypoxia (12). Nevertheless, to the extent that ischemia-reperfusion injury is the result of oxidative stress to the capillary endothelium, the crystalloid perfused heart model used in the present study may be more susceptible.

Conclusions.

The results of the present study are the first strong evidence that acute exposure to 1 nM E2 (20 min or less) inhibits the response of NHE1 to acidification. That is, acute E2 exposure decreases the rate of proton efflux after NH4 washout in HEPES medium. Furthermore, acute E2 exposure diminishes accumulation of cytosolic HMath, Math, and Math in hearts from OVX during ischemia. Limiting the accumulation of these ions was associated with improved recovery of LVDP and decreased release of LDH during reperfusion. In addition, E2-dependent effects on all ions were diminished by addition of 1 μM l-NAME. These results are consistent with the previously untested hypothesis that an NO-dependent effect of E2 inhibits NHE1 to limit Na+ and thereby Ca2+ uptake during ischemia. However, thermodynamic arguments also illustrate that the effect of E2 on [Ca2+]i is greater than can be accounted for by the effect of limiting Math accumulation on sarcolemmal NCE and thus suggest that E2 inhibits NCE kinetically.

Finally, we speculate that regardless of findings related to chronic effects of E2 (e.g., those related to hormone replacement therapy), there may be a role for acute E2 treatment in therapies aimed at limiting myocardial ischemia-induced injury, including those designed for use in organ preservation or bypass surgery.

APPENDIX

Calculation of reversal potential for NCE.

The NCE reversal potential is calculated as ENCE = 3ENa − 2ECa = 3RT/F ln ([Na+]o/[Na+]i) − RT/F ln ([Ca2+]o/[Ca2+]i) where subscripts o and i correspond to extra- and intracellular, respectively. The assumption is made that the intra- and extracellular volumes in the heart are equal and unchanging during the experiment (e.g., during ischemia, K+ loss equals Na+ uptake) and that cell water is 2.5 l/kg dry wt (3). This allows one to calculate [Na+]i in millimoles from the NMR measurement of Na+ in milliequivalents per kilogram dry weight. Math and [Ca2+]i were measured using NMR as described in methods. [Ca2+]o is assumed to be constant at 1.8 mM throughout the experiment. [Na+]o = perfusate [Na+] = 141 mM at the beginning of the experiment, and during ischemia, the change in [Na+]i is equal and opposite to the change in [Na+]o. Using these assumptions for end ischemia in OVX+E2 shown in Figs. 3 and 4, [Na+]i = 30 mM, [Na+]o = 121 mM, [Ca2+]i = 202 nM, and [Ca2+]o = 1.8 mM, such that ENCE = −126 mV.

GRANTS

The research was supported by a grant from Philip Morris USA and National Heart, Lung, and Blood Institute Grant HL-21179. The NMR spectrometer expense was funded in part by National Institutes of Health Grant RR-02511 and National Science Foundation Grant PCM-8417289.

Footnotes

  • 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. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract