Displacement-encoded and manganese-enhanced cardiac MRI reveal that nNOS, not eNOS, plays a dominant role in modulating contraction and calcium influx in the mammalian heart

Abstract

Within cardiomyocytes, endothelial nitric oxide synthase (eNOS) and neuronal nitric oxide synthase (nNOS) are thought to modulate L-type calcium channel (LTCC) function and sarcoplasmic reticulum calcium cycling, respectively. However, divergent results from mostly invasive prior studies suggest more complex roles. To elucidate the roles of nNOS and eNOS in vivo, we applied noninvasive cardiac MRI to study wild-type (WT), eNOS^−/−, and nNOS^−/− mice. An in vivo index of LTCC flux (LTCCI) was measured at baseline (Bsl), dobutamine (Dob), and dobutamine + carbacholamine (Dob + CCh) using manganese-enhanced MRI. Displacement-encoded MRI assessed contractile function by measuring circumferential strain (E_cc) and systolic (dE_cc/dt) and diastolic (dE_cc/dt_diastolic) strain rates at Bsl, Dob, and Dob + CCh. Bsl LTCCI was highest in nNOS^−/− mice (P < 0.05 vs. WT and eNOS^−/−) and increased only in WT and eNOS^−/− mice with Dob (P < 0.05 vs. Bsl). LTCCI decreased significantly from Dob levels with Dob + CCh in all mice. Contractile function, as assessed by E_cc, was similar in all mice at Bsl. With Dob, E_cc increased significantly in WT and eNOS^−/− but not nNOS^−/− mice (P < 0.05 vs. WT and eNOS^−/−). With Dob + CCh, E_cc returned to baseline levels in all mice. Systolic blood pressure, measured via tail plethysmography, was highest in eNOS^−/− mice (P < 0.05 vs. WT and nNOS^−/−). Mice deficient in nNOS demonstrate increased Bsl LTCC function and an attenuated contractile reserve to Dob, whereas eNOS^−/− mice demonstrate normal LTCC and contractile function under all conditions. These results suggest that nNOS, not eNOS, plays the dominant role in modulating Ca²⁺ cycling in the heart.

nitric oxide (no) is a key signaling molecule in both the vasculature and the heart (12, 30). Within healthy cardiomyocytes, two distinct isoforms of NO synthase (NOS), endothelial NOS (eNOS) and neuronal NOS (nNOS), are constitutively expressed (2, 4, 13, 27, 36), whereas an inducible NOS (iNOS) isoform is expressed in the heart under stress (16, 58). Both eNOS and nNOS are spatially localized, with eNOS colocalized with caveolae in close proximity to the L-type Ca²⁺ channel (LTCC) (4) and nNOS colocalized with the ryanodine receptor (RyR) (21, 51) at the sarcoplasmic reticulum (SR). NO generated in healthy cardiomyocytes by eNOS and nNOS has been shown to modulate cardiac Ca²⁺ cycling by modulating LTCC and RyR function, thereby modulating contractile function (3, 12, 30, 57). However, the specific role of each NOS isoform in Ca²⁺ cycling and cardiac function in the healthy heart remains unclear (30). In addition to important roles in the healthy heart, changes in cardiac NO signaling have been associated with arrhythmia (8, 18, 50), diastolic dysfunction (38), and left ventricular (LV) remodeling after myocardial infarction (13, 16, 26, 33, 34).

Prior studies, primarily using in vitro or invasive in vivo techniques in mice, have sought to elucidate the roles of eNOS and nNOS in modulating LTCC function and contractile function. Specifically, several studies have suggested that eNOS plays a dominant role in inhibiting LTCC function either at baseline (4), during β-adrenergic stimulation (50), or during muscarinic cholinergic inhibition of β-adrenergic stimulation (20). nNOS has also been suggested to modulate LTCC function (8, 36), as well as SR Ca²⁺ cycling (4, 27) and phospholamban (PLB) mediated modulation of diastolic function (53) under various physiological conditions. Divergent findings among prior studies have been attributed to differences in background strains, age of mice, anesthesia, temperature, and dose of β-adrenergic agonist (3, 9, 30). Although prior in vitro studies of LTCC current (I_Ca,L) in isolated myocytes have used a range of low stimulation frequencies [ranging from 0.2 Hz (50) to 3 Hz (20)], most invasive in vivo studies of contractility have been performed at much higher frequencies [8 Hz (47) to 12 Hz (27)]. No previous study has examined both LTCC function and contractile function in a completely in vivo and minimally invasive manner and at physiological frequencies.

In recent years, cardiac magnetic resonance (CMR) imaging has emerged as an accurate and versatile modality for noninvasively assessing the mouse heart (15, 46). In addition to cine CMR for imaging LV size and function (6, 39), dynamic manganese-enhanced CMR can probe in vivo LTCC function (24), and displacement-encoded (DENSE) CMR can quantify myocardial strain and strain rate (28). We previously used CMR to study nNOS^−/− mice and showed that they have normal baseline contractile function but attenuated contractile reserve in response to medium and high doses of β-adrenergic stimulation (47). We also showed that the attenuated contractile reserve was not due to reduced perfusion reserve. In the present study, we extended these investigations using novel in vivo CMR techniques to elucidate the roles of both nNOS and eNOS in excitation contraction coupling. Specifically, we used dynamic Mn-enhanced CMR and DENSE CMR to noninvasively characterize nNOS^−/− and eNOS^−/− mice for the first time in terms of in vivo LTCC and contractile function, respectively, under varying pharmacological stimulation.

METHODS

Animals.

Ten wild-type (WT), 10 eNOS^−/−, and 10 nNOS^−/− male mice on a C57BL/6 background (Jackson Laboratory, Bar Harbor, ME) were studied under protocols that conformed to the Declaration of Helsinki as well as the Guide for the Care and Use of Laboratory Animals (NIH publication no. 85-23, revised 1996); the protocols were approved by the Animal Care and Use Committee at our institution. CMR assessment of LTCC function and contractile function were performed at 10 ± 3 wk of age.

CMR preparation.

Mn-enhanced CMR was performed on a 4.7 T MRI system (Varian, Palo Alto, CA), and DENSE imaging of LV contractile function was performed on a 7T ClinScan MRI system (Bruker, Ettlingen, Germany). Cylindrical birdcage RF coils were used with both magnetic resonance scanners. Mn-enhanced CMR was performed on the 4.7T system and not the 7T system to maintain continuity with our initial studies. Body temperature was maintained at 36.4 ± 0.3°C by circulating thermostated water, and anesthesia was maintained using 1.25% isoflurane in O₂ inhaled through a nose cone. Each complete Mn-enhanced CMR study took 2 h, and each CMR study of contractile function took 1.5 h. During both experiments, mice lay prone within the scanner. Heart rate (HR), respiration, and core body temperature were monitored during imaging using a fiber optic, MR compatible system (Small Animal Instruments, Stony Brook, NY). For administration of pharmacological agents or infusion of MnCl₂, an indwelling catheter was inserted into the intraperitoneal cavity as part of CMR preparation.

Mn-enhanced cardiac magnetic resonance.

A series of mid-ventricular short-axis T₁ weighted images were acquired before, during, and after continuous intraperitoneal infusion of MnCl₂ to probe manganese (Mn²⁺) influx kinetics. Mn²⁺ ions enter cardiomyocytes through the LTCC in proportion to Ca²⁺ flux (31), remain sequestered within cardiomyocytes for a period on the order of hours, and shorten the T₁ of nearby water in proportion to their concentration (23–25). Representative T₁-weighted images (Fig. 1) acquired before, during, and after infusion of MnCl₂ in a WT mouse at baseline (A), with dobutamine (B), and in response to nifedipine (C), illustrate the effect of Mn²⁺ accumulation on signal intensity under different pharmacological conditions. The accumulation of Mn²⁺, and subsequent changes in myocardial signal intensity, reflects the integrated effects of rapid LTCC currents over the period of MnCl₂ infusion, therefore enabling the noninvasive in vivo assessment of LTCC function with CMR (24, 25, 46). Mn-enhanced CMR was performed using an ECG-gated saturation recovery pulse sequence with constant saturation time (T_sat). Before acquisition of Mn-enhanced data, a proton density weighted image was acquired using a T_sat of 5 s. All subsequent T₁-weighted images, which were acquired with T_sat of 200 ms, had measurement times of 2 to 3 min, depending on HR. Pulse sequence parameters were chosen to achieve a balance between spatial resolution, signal-to-noise ratio, degree of T₁ weighting, and sampling rate for measuring Mn-enhancement kinetics. Specific sequence parameters included a field of view of 25.6 × 25.6 mm², 1 mm slice thickness, 90° flip angle, and four averages. Images were continuously acquired for 15 min before, during, and for 30 min following a 30-min intraperitoneal infusion of 30 mM MnCl₂ (0.42 mg/kg·min), with specific image acquisition times varying slightly based on HR.

Fig. 1.

A–C: midventricular short-axis T₁ weighted images from a wild-type (WT) mouse at baseline (A), with dobutamine (B), and with nifedipine (C). Images were acquired over 90 min, whereas MnCl₂ was infused only from 20 min until 50 min. Representative images in each series were acquired at ∼15, 30, 50, and 80 min. At baseline, accumulation of Mn²⁺ in cardiomyocytes results in increasing myocardial enhancement throughout the period of MnCl₂ infusion. Myocardial enhancement is increased as a result of heightened L-type calcium channel (LTCC) function in response to dobutamine and decreased as a result of partial LTCC inhibition with nifedipine. D: sample data from a WT mouse demonstrate linear increases in signal-to-noise ratio (SNR) during infusion of MnCl₂. The rate of increase in SNR, measured as the slope of a linear fit (dotted black line) over the points corresponding to MnCl₂ infusion (filled gray symbols between the start and end of infusion), provides an index of integrated Mn²⁺ flux through the LTCC [in vivo index of LTCC flux (LTCCI); bold numbers]. When compared with baseline, LTCCI increases in response to β-adrenergic stimulation with dobutamine and decreases in response to nifedipine. AU, arbitrary units.

Image analysis was performed in MATLAB (Mathworks, Natick, MA), where signal-to-noise ratio (SNR) was quantified by dividing mean myocardial signal intensity in each image by the standard deviation of background noise. SNR was then divided by the mean SNR of corresponding proton density weighted images, yielding the normalized SNR. LTCC index (LTCCI) was quantified as the rate of enhancement of the normalized myocardial SNR per 10⁵ heart beats during MnCl₂ infusion, as shown in Fig. 1D.

We first examined changes in LTCCI in response to pharmacological manipulation of LTCC function in WT mice. LTCCI was measured at baseline (Bsl; n = 10), during β-adrenergic stimulation with an intraperitoneal infusion of a low dose of dobutamine (Dob; 5 μg/kg·min; n = 10), during concomitant muscarinic cholinergic inhibition of β-adrenergic stimulation with an intraperitoneal infusion of Dob and carbamylcholine chloride (CCh; 3 mg/kg·min) (Dob + CCh; n = 10), in response to the LTCC inhibitor nifedipine (10 mg/kg; n = 3), and in response to increased frequency of contraction (14) induced by the A_2A adenosine receptor agonist ATL313 (47) (12.5 μg/kg; n = 4). Finally, to verify a sufficient dynamic range of our dynamic Mn-enhanced CMR technique, LTCCI was measured in response to a higher dose of Dob (Dob 20; 20 μg/kg·min; n = 2). Infusion of Dob and Dob + CCh, or bolus injections of nifedipine and ATL313 occurred through a second indwelling intraperitoneal line. Next, LTCCI was measured in eNOS^−/− (n = 9) and nNOS^−/− (n = 9) mice at Bsl, Dob, and Dob + CCh. Because eNOS is theorized to be involved in muscarinic cholinergic inhibition of β-adrenergic stimulation in the heart, the effect of CCh was evaluated only in combination with Dob. In all cases, 1 wk elapsed in between experiments to allow time for complete washout of manganese, since Bsl, Dob, and Dob + CCh studies were performed in the same mice. Although the effects of both nifedipine and ATL313 are observed for the entire course of the experiment following bolus injection, Dob and Dob + CCh were applied using constant-rate infusion to maintain a constant level of stimulation.

Cine-DENSE CMR.

One week after conclusion of measurement of LTCCI, contractile function was assessed in the same mice using cine DENSE (28, 54, 55), a technique that is related to, but improves upon, myocardial tagging (47). Cine DENSE data were acquired for two midventricular short-axis slices at Bsl, Dob, and Dob + CCh during the same imaging session, using an intraperitoneal infusion of pharmacological agents as stated previously. Care was taken to ensure that all data were acquired at the midventricle in each mouse. Specific acquisition parameters included field of view = 25.6 mm, matrix = 128 × 128, slice thickness = 1 mm, repetition time = 6.9 ms, echo time = 1 ms, number of averages = 4, number of spiral interleaves = 27, and displacement encoding frequency = 1.1 cycles/mm. The total scan time for a single two-dimensional slice was 6–8 min.

For data analysis, segmentation of the heart used a semi-automated technique (41), and displacement and strain were calculated using the phase unwrapping and motion tracking algorithms described previously (40). Peak midwall circumferential systolic strain (E_cc) and systolic strain rate (dE_cc/dt) were used to assess contractile function. Diastolic strain rate (dE_cc/dt)_diastolic was used to assess diastolic function, as described previously (47).

Systolic blood pressure.

Systolic blood pressure was measured in nonanesthetized, conscious mice. A minimum of five baseline waveforms were obtained for each mouse. Measurements were obtained from each waveform, and then averaged together for each individual mouse, using a non-invasive tail plethysmography system (Model BP 2000; Visitech Systems, Apex, NC).

Statistical analysis.

All statistical analyses were performed using SigmaStat (Systat Software, Point Richmond, CA). Differences in body weight and hemodynamic measurements were evaluated using one-way ANOVA. Differences in LTCCI, E_cc, dE_cc/dt, and (dE_cc/dt)_diastolic at Bsl, Dob, and Dob + CCh between WT, eNOS^−/−, and nNOS^−/− mice were assessed using one-way repeated-measures ANOVA, assuming compound symmetry. Analysis was performed after confirmation of normal distribution with a statistical power greater than 0.95 unless otherwise noted. Comparisons of LTCCI at Dob20, ATL313, and nifedipine in WT mice were performed using t-test analysis. All values in the text, tables, and graphs are presented as means ± SE.

RESULTS

Body mass and hemodynamics.

Body mass was lower in eNOS^−/− and nNOS^−/− mice (Table 1), systolic blood pressure was significantly higher in eNOS^−/− mice compared with both WT and nNOS^−/− mice, and basal HR (statistical power = 0.16) did not differ between groups (Table 1).

Table 1.

Body mass and hemodynamics in wild-type, eNOS^−/−, and nNOS^−/− mice

	Wild-type	eNOS−/−	nNOS−/−
n	10	9	8
Body mass, g	29.5 ± 0.8	23.9 ± 0.3∗	22.8 ± 0.4∗
Systolic blood pressure, mmHg	110 ± 3	131 ± 2∗	110 ± 2†
Heart rate, beats/min	520 ± 19	481 ± 13	494 ± 11

Values are means ± SE. eNOS and nNOS, endothelial and neuronal nitric oxide synthase, respectively.

^∗P < 0.05 vs. wild-type;

^†P < 0.05 vs. eNOS^−/−.

LTCC function.

We probed LTCCI with Mn-enhanced CMR under a variety of pharmacological conditions in WT mice. Measurements of LTCCI and HR in WT mice are shown in Fig. 2. In response to β-adrenergic stimulation with a low dose of dobutamine (5μg/kg·min), LTCCI increased by 47% over baseline, reflecting a significant increase in LTCC function. In response to Dob + CCh, LTCCI returned to basal levels, reflecting the effect of muscarinic cholinergic inhibition of β-adrenergic stimulation. LTCCI fell by 26% (P < 0.05 vs. Bsl, statistical power = 0.53) compared with baseline in response to increasing HR with ATL313, mirroring the negative I_{Ca,L_density}-frequency relationship seen in mouse cardiomyocytes (1). Finally, in response to partial LTCC inhibition with nifedipine, LTCCI decreased by 42% (P < 0.05 vs. Bsl). HR increased significantly over baseline in response to Dob and ATL313, while it did not change in response to Dob + CCh or nifedipine (Fig. 2B). In response to a higher dose of dobutamine (20μg/kg·min), LTCCI (1.19 ± 0.03 A.U./10⁵ H.B., P < 0.05 vs. Bsl, Dob) and HR (618 ± 58 B.P.M , P < 0.05 vs. Bsl) increased 96% and 27% over baseline, respectively, demonstrating a dynamic range that extends beyond the low-dose dobutamine response measured in WT mice.

Fig. 2.

Measurement of LTCCI (A) and heart rate (HR; B) in WT mice obtained at Bsl, dobutamine (Dob), dobutamine + carbacholamine (Dob + CCh), ATL313, and nifedipine. Dob-induced increases in LTCCI and HR were reversed with Dob + CCh. With adenosine A_2a receptor stimulation using ATL313, LTCCI decreased, whereas HR increased, reflecting a negative LTCCI-frequency response. Nifedipine, an LTCC inhibitor, decreased LTCCI significantly while not having any effect on HR. *P < 0.05 vs. Bsl; †P < 0.05 vs. Dob. BPM, beats/min; HB, heartbeat.

Mn-enhanced CMR experiments were performed in eNOS^−/− and nNOS^−/− mice to elucidate the role of each NOS isoform in modulating LTCC function. The average pre-Mn²⁺myocardial SNR was similar in WT, eNOS^−/−, and nNOS^−/− mice (0.28 ± 0.01 WT, n = 27 vs. 0.28 ± 0.01 eNOS^−/−, n = 24 vs. 0.29 ± 0.01 nNOS^−/−, n = 24, P = NS), as was the SNR of the proton density weighted images (n = 4 per group, 171 ± 3 WT vs. 170 ± 7 eNOS^−/− vs. 166 ± 11 nNOS^−/−, P = NS). At baseline, LTCCI trended higher in eNOS^−/− mice compared with WT mice (P = 0.06) (Fig. 3). However, LTCCI increased significantly over baseline with Dob, and decreased significantly from Dob levels in response to Dob + CCh similarly in eNOS^−/− and WT mice. In contrast, basal LTCCI was significantly higher in nNOS^−/− mice compared with both WT (P < 0.01) and eNOS^−/− (P < 0.05) mice (Fig. 3). LTCCI trended towards a decrease compared with basal levels (P < 0.1 vs. Bsl) in response to Dob and decreased significantly from both Bsl and Dob in response to Dob + CCh in nNOS^−/− mice. Despite differences in the responses to Dob and Dob + CCh between nNOS^−/− and eNOS^−/− mice that reflect differences at baseline, absolute LTCCI values in animals given Dob and Dob + CCh were similar amongst WT, eNOS^−/−, and nNOS^−/− mice.

Fig. 3.

Measurements of LTCCI in endothelial nitric oxide synthase (eNOS)^−/− and neuronal nitric oxide synthase (nNOS)^−/− mice at Bsl, Dob, and Dob + CCh. Basal LTCCI was highest in nNOS^−/− mice compared with WT and eNOS^−/− mice (WT results copied from Fig. 2A for purposes of comparison). In response to Dob and Dob + CCh, LTCCI was similar among all mice. *P < 0.05 vs. baseline; †P < 0.05 vs. Dob.

Contractile function.

Basal contractile function as quantified by cine DENSE MRI was similar in WT, eNOS^−/− and nNOS^−/− mice. Fig. 4 illustrates representative examples of E_cc-time curves from WT, eNOS^−/− and nNOS^−/− mice at Bsl, Dob, and Dob + CCh. At baseline, peak E_cc was similar in WT, eNOS^−/−, and nNOS^−/− mice (Fig. 5A). WT and nNOS^−/− mice demonstrated similar basal dE_cc/dt (−419 ± 16%/s WT, n = 10 vs. −378 ± 13%/s nNOS^−/−, n = 9, P = 0.17, statistical power = 0.79), dE_cc/dt_diastolic (374 ± 33%/s WT, n = 9 vs. 331 ± 14%/s nNOS^−/−, n = 9, P = 0.17, statistical power = 0.76), and HR (540 ± 16 BPM WT, n = 9 vs. 515 ± 11 BPM nNOS^−/−, n = 9, P = 0.14, statistical power = 0.60). In contrast, dE_cc/dt (−368 ± 11%/s, n = 9, P < 0.05 vs. WT), dE_cc/dt_diastolic (296 ± 17%/s, n = 9, P < 0.05 vs. WT), and HR (495 ± 12 BPM, n = 9, P < 0.05 vs. WT) were significantly lower in eNOS^−/− mice at baseline. No differences were seen between eNOS^−/− and nNOS^−/− mice at baseline in terms of dE_cc/dt or dE_cc/dt_diastolic.

Fig. 4.

Sample midwall circumferential strain (E_cc) vs. time curves from WT (top), eNOS^−/− (middle), and nNOS^−/− (bottom) mice at baseline and in response to infusion of Dob and Dob + CCh. In WT and eNOS^−/− mice, peak E_cc increased significantly from basal levels (dotted gray line) in response to Dob and returned to basal levels following administration of Dob + CCh. In nNOS^−/− mice, no change in E_cc was seen in response to either Dob or Dob + CCh.

Fig. 5.

Measurements of contractile function in WT, eNOS^−/−, and nNOS^−/− mice. A: at baseline, peak E_cc (representing maximal contraction) was similar in all mice. WT and eNOS^−/− mice, but not nNOS^−/− mice, demonstrated a strong inotropic response to Dob as measured by E_cc (B) and systolic strain rate (dE_cc/dt; C), which was reversed in response to Dob + CCh. D: only WT mice demonstrated a strong lusitropic response to Dob, which was reduced with Dob + CCh. *P < 0.05 vs. WT.

In response to Dob, E_cc increased significantly in WT and eNOS^−/−, but not nNOS^−/− mice (Fig. 5B). Mirroring E_cc results, the increase in strain rate (dE_cc/dt) with Dob was significantly lower in nNOS^−/− mice compared with WT and eNOS^−/− mice (Fig. 5C). However, both eNOS^−/− and nNOS^−/− mice demonstrated an attenuated lusitropic response to Dob (Fig. 5D). Changes in HR with Dob were similar in WT (10 ± 8 BPM, n = 9) and eNOS^−/− mice (26 ± 6 BPM, n = 9, P = 0.12 vs. WT, statistical power = 0.60), but were greater in eNOS^−/− compared with nNOS^−/− (−4 ± 11 BPM, n = 9, P < 0.01 vs. eNOS^−/−) mice. In response to Dob + CCh, WT, eNOS^−/−, and nNOS^−/− mice all demonstrated no change from baseline in either E_cc or dE_cc/dt (Fig. 5B,C). The change compared with baseline in dE_cc/dt_diastolic decreased significantly from Dob levels in WT mice, did not change in eNOS^−/− mice, and remained lower in nNOS^−/− mice compared with WT mice (Fig. 5D). Also, the change in HR with Dob + CCh was significantly lower in nNOS^−/− mice (−14 ± 7 BPM, n = 9) compared with both WT (8 ± 6 BPM, n = 9, P < 0.05 vs. nNOS^−/−) and eNOS^−/− mice (26 ± 6 BPM, n = 9, P < 0.01 vs. nNOS^−/−), and trended higher in eNOS^−/− compared with WT mice (P = 0.07).

DISCUSSION

Using noninvasive CMR to assess the functional phenotype of eNOS^−/− and nNOS^−/− mice in vivo, the major findings of this study are that: (i) basal LTCCI is highest in nNOS^−/− mice even though basal contractile function is similar to WT mice, (ii) in response to dobutamine, LTCCI decreased from baseline while contractile function did not change in nNOS^−/− mice, (iii) in nNOS^−/− mice, contractile function remained similar to baseline while LTCCI decreased significantly in response to concomitant β-adrenergic and muscarinic cholinergic stimulation, and (iv) both LTCC function and systolic contractile function are similar in eNOS^−/− and WT mice at baseline, in response to β-adrenergic stimulation, and in response to concomitant β-adrenergic and muscarinic cholinergic stimulation. These novel findings represent the first completely in vivo study of modulation of LTCC function and contractile function by eNOS and nNOS in mice, and demonstrate a dominant role for nNOS in modulating LTCC function and contractile function under physiological conditions.

Whereas traditional measurements of I_Ca,L reflect ultra-fast currents across the plasma membrane of an individual cardiomyocyte, LTCCI measurements represent the integrated effect of Mn²⁺ accumulation over time and over all cardiomyocytes. A prior study by Masumiya et al. demonstrated that I_Ca,L kinetics are mirrored by LTCC-Mn²⁺ current (I_Mn) kinetics, although at a lower magnitude (approx. 8%) (31). Thus, LTCCI is an analog of integrated I_Ca,L density per heart beat over the entire heart, and reflects the cumulative effect of ultra-rapid LTCC currents during MnCl₂ infusion. It is important to note that most prior Mn-enhanced CMR techniques used only ECG gating for both triggering of the saturation pulse and image acquisition (24). This resulted in variable saturation times, thus making the degree of Mn-enhancement a function of both Mn²⁺ uptake and HR. In contrast, our modified dynamic Mn-enhanced CMR method adjusted the timing of a non-selective saturation pulse such that a nearly constant saturation time was maintained, thereby making the rate of Mn-enhancement reflect only Mn²⁺ uptake, and thus enabling quantitative examination of LTCC function. Comprehensive examination of our dynamic Mn-enhanced CMR studies in WT mice indicated that LTCCI measurements are sensitive to pharmacological manipulation of LTCC function, and mirror changes in both I_Mn (31), and I_Ca,L in prior patch-clamp experiments performed in isolated myocytes (20). Increases in LTCCI with Dob and decreases with Dob + CCh were similar to changes in I_Ca,L density in isolated myocytes from WT mice in response to β-adrenergic stimulation with isoproterenol (Iso) (20, 45, 50) and Iso+CCh (20, 45). In response to LTCC inhibition with nifedipine, LTCCI decreased similarly to changes in I_Ca,L in prior studies in isolated guinea pig ventricular myoctyesc (37). Our findings agree with prior studies which demonstrated that Mn²⁺ enhancement of myocardium is sensitive to increased LTCC flux following administration of dobutamine (24), and to inhibition with LTCC inhibitors (24). Finally, to probe for an LTCCI-frequency relationship, we used adenosine A_2a receptor stimulation with ATL313 and observed a decrease in LTCCI compared with basal levels. Several studies performed in isolated mouse myocytes have demonstrated that both peak I_Ca,L and integrated I_Ca,L density, a measure of calcium influx, decrease with increasing pacing frequency (1). Unlike β-adrenergic stimulation, which increases HR and LTCC function simultaneously, A_2A receptor stimulation increases HR (14) without increasing LTCC function (22). With ATL313 induced increases in HR, the integrated Mn²⁺ flux over the period of MnCl₂ infusion should increase. However, since LTCCI is an analog of integrated I_Ca,L density per heart beat, and not per unit time, we interpret our LTCCI results with ATL313 as showing a negative in vivo LTCCI-frequency relationship, in agreement with the negative I_Ca,L density-frequency relationship in isolated myocytes (1).

Traditionally, nNOS has been considered to modulate calcium cycling at the SR level (4, 13, 29, 47), and not the LTCC level (4, 57). However, our finding of changes in LTCCI without accompanying changes in contractile function in nNOS^−/− mice, across different pharmacological stimulation protocols, suggests a role for nNOS in modulating calcium cycling at both the LTCC and SR levels. Previous in vitro studies have demonstrated either elevated basal I_Ca,L (8, 36) and myocyte shortening (2, 29), or normal I_Ca,L (4) and myocyte shortening (4, 27) in nNOS^−/− myocytes. Since measurement of circumferential strain and strain rate approximates in vivo measurements of myocyte shortening and shortening velocity, respectively (47), one would anticipate increased baseline circumferential strain to accompany our measurement of elevated basal LTCCI in nNOS^−/− mice. However, given that I_Ca,L accounts for only ∼5% of systolic calcium in murine myocytes (compared to ∼30% in humans (7)), the increase in basal LTCC function in nNOS^−/− mice may be insufficient to elicit significant changes in contractile function given normal basal SR function, as evidenced by normal baseline SR calcium transients in nNOS^−/− mice amongst prior studies (4, 27). In fact, our finding of normal basal contractile function in nNOS^−/− mice agrees with most previous in vivo studies (4, 8, 27, 33, 36), including both a prior study by Sears et al. which found increased I_Ca,L and myocyte shortening but normal basal in vivo contractility in nNOS^−/− mice (36), as well as our previous myocardial tagging study (47). In response to Dob, nNOS^−/− mice demonstrated a slightly negative LTCCI response (Fig. 3) and no E_cc, dE_cc/dt, or dE_cc/dt_diastolic response (Fig. 5). As HR increased on average 9%, and LTCCI decreased on average 10% with Dob in nNOS^−/− mice, this likely reflects a negative LTCCI-frequency response, similar to that witnessed in WT mice with ATL313, and further suggests that basal LTCC function is already maximal. While these results alone may suggest that the primary role of nNOS is to modulate only LTCC function under basal conditions, our Dob + CCh results suggest a more complex role for nNOS. In WT mice, LTCCI, E_cc, dE_cc/dt, and dE_cc/dt_diastolic returned to baseline levels in response to Dob + CCh. In contrast, nNOS^−/− mice demonstrated a significant reduction in LTCCI without any change in E_cc, dE_cc/dt, or dE_cc/dt_diastolic. These results suggest that while basal LTCC function was maximal, contractile function remained at baseline levels even with Dob and Dob + CCh in nNOS^−/− mice. A number of previous studies have demonstrated reduced SR calcium transients and diastolic SR calcium reuptake in nNOS^−/− mice in response to β-adrenergic stimulation (4, 49), further suggesting that compromised SR function in nNOS^−/− mice may contribute to altered ventricular function observed in the current study. Recent studies have suggested that nitrosylation decreases LTCC function (8, 43) while increasing RyR function (42, 52). Similarly, NO can modulate Ca²⁺ cycling through the NO/cGMP signaling cascade (35, 42, 48, 49, 56). Although nNOS is spatially co-localized with the RyR it is possible that, given the small dimensions of the cardiac dyad, NO from nNOS can modulate the function of both the LTCC and the RyR. With this in mind, our findings of increased basal LTCCI and attenuated contractile reserve in nNOS^−/− mice may suggest a prominent role for nNOS in modulating contractile function separately from LTCC function, as has been suggested by others (3).

Based on its physical proximity to the LTCC, eNOS has been suggested to modulate LTCC function under a variety of physiological conditions (4, 20, 50). Although basal LTCCI was slightly higher in eNOS^−/− compared with WT mice in our study, LTCCI was lower compared with nNOS^−/− mice. This suggests that the elevation in basal LTCCI in eNOS^−/− compared with WT mice may result from hypertension in eNOS^−/− mice (4, 50). Ideally, the impact of hypertension on LTCC function in eNOS^−/− mice could be normalized through pharmacological intervention. However, most methods for reducing blood pressure would themselves alter either LTCC function or the β-adrenergic responsiveness in vivo, thus obfuscating the role of eNOS in such studies. Our finding of similar LTCCI in WT and eNOS^−/− mice with Dob agrees with most in vitro studies showing similar I_Ca,L density in response to β-adrenergic stimulation with Iso (5, 17, 20, 45). Only one prior study has demonstrated greater increases in I_Ca,L density in eNOS^−/− mice in response to Iso at a stimulation frequency of 0.5Hz (50). It is possible that, given the negative I_Ca,L-frequency relationship (1), Wang et al. (50) detected frequency-dependent differences in I_Ca,L which are not apparent at much higher in vivo frequencies. Finally, in agreement with most (5, 17, 45), but not all (20) prior I_Ca,L studies, we detected similar LTCCI in WT and eNOS^−/− mice in response to Dob + CCh. Previously, only Han et al. (20) have detected increased I_Ca,L in eNOS^−/− mice in response to Iso+CCh, possibly from using higher doses of Iso and CCh than prior studies. However, the concentration of CCh in our study (15μM) was greater than that used by Han et al. (10μM) (20). Furthermore, in WT mice, we observed a ratio of LTCCI with Dob and Dob + CCh similar to that of I_Ca,L in response to Iso and Iso+CCh in their study (20). These findings suggest that, in vivo, eNOS does not play a significant role in modulating LTCC function.

A few previous studies have suggested that eNOS may also play a role in modulating contractility (4) and β-adrenergic responsiveness (4, 50). Our finding of similar basal E_cc in WT and eNOS^−/− mice agrees with the majority of in vivo findings of similar (dP/dt)/P_id (4, 19, 27) and in vitro findings of similar myocyte shortening (4, 17, 20, 27, 29, 45, 50) in WT and eNOS^−/− mice. In agreement with the (dP/dt)/P_id findings of Barouch et al. in response to low and medium doses of Iso (4), as well as most in vitro findings of similar myocyte shortening in WT and eNOS^−/− mice (4, 20, 29, 45), we observed a normal inotropic contractile response to Dob in eNOS^−/− mice. Two studies have demonstrated enhanced β-adrenergic dP/dt_max in eNOS^−/− mice with Iso (10, 19). However, dP/dt_max is known to be sensitive to afterload (44), suggesting that the innate hypertension in eNOS^−/− mice may have influenced prior results (10, 19). It is also possible that no differences were detected between eNOS^−/− and WT mice as a result of the relatively low dose of Dob used. However, Martin et al. demonstrated similar sarcomere shortening in eNOS^−/− and control mice in response to several doses of Iso (29). Interestingly, we observed an attenuated lusitropic response to Dob in eNOS^−/− mice, which given the recently discovered role of eNOS in preventing early after-depolarizations during diastole (50), may result from the absence of eNOS mediated regulation of Ca²⁺ handling during diastole. In addition, we also observed greater HR responses in eNOS^−/− mice in response to Dob and Dob + CCh, further suggesting that eNOS may be involved in regulating the chronotropic response in the heart. Finally, our finding of normal contractile function in eNOS^−/− mice in response to Dob + CCh agrees with previous in vitro findings (17, 29, 45), but not with findings of elevated whole heart contractility in response to Iso+CCh by Champion et al. (10). Whereas Champion et al. infused CCh (30 mg/kg·min) intravenously (10), we found that IP infusion of CCh at the same rate was immediately lethal, suggesting that the dose of CCh may have affected our findings. However, the E_cc response to Dob + CCh in WT mice was similar in magnitude to that of Champion et al. (10). Together, these results suggest that eNOS does not play a significant role in modulating contractile function in the mouse heart, but does play a role in diastolic function.

One limitation of our study was that differences in LTCC and NCX expression between WT and nNOS^−/− mice were not investigated, which could affect the rate of Mn-enhancement. However, several prior studies have demonstrated similar expression of both LTCC (18, 36) and NCX (18, 36, 49) proteins in WT and nNOS^−/− mice. Also, relative abundances of eNOS and nNOS in knockout mice were not measured, suggesting that potential genetic compensation of the remaining NOS isoform could obfuscate the role attributed to a given NOS based on results in knockout mice. However, prior studies using similar mice have not found compensatory over-expression or translocation of the available NOS under similar experimental conditions (2, 4, 11, 13, 33, 45). In addition, mice used in this study were global knockouts, and experiments were not repeated using selective nNOS inhibitors (such as L-VNIO) or donors in WT mice. An additional limitation was that LTCCI measurements were not validated using conventional patch clamp methods. This suggests that it is possible that increases in LTCCI in nNOS^−/− mice with Dob were not detected because of a possible maximum sensitivity in our Mn-enhanced technique. However, LTCCI in nNOS^−/− mice at both Bsl and Dob was lower than the maximum LTCCI measured in WT mice at high-dose Dob (20μg/kg·min), suggesting that sensitivity was not a limitation. Also, although measurements of LTCCI and contractile function were made after allowing for IP administration of Dob and Dob + CCh to reach a steady state, prolonged administration of Dob and Dob + CCh could affect abdominal organs, thus producing secondary effects on the heart that are not observed in experiments on isolated myocytes. Additionally, while in vitro experiments typically use a variety of doses of nifedipine to determine a dose-response curve, preliminary dosing experiments with nifedipine proved fatal at doses higher than 10mg/kg. At the chosen dose of nifedipine, we believe that we achieved LTCC inhibition on the order of 40%, in agreement with prior ex vivo data (37). Also, nifedipine is known to be more active on arterial and arteriolar Ca²⁺ channels than on cardiac LTCCs, raising the possibility that the significant decrease in LTCCI in response to nifedipine may have also reflected a decrease in Mn-enhancement of the cardiac vasculature. However, the water within the wall of the cardiac vasculature represents a very small portion of the water, and therefore MR signal, within each voxel compared with that within cardiomyocytes. Thus, our measurements of LTCCI with nifedipine predominantly reflect the impact of nifedipine on cardiomyocytes. Finally, differences in body weight between WT and knockout mice raise the possibility that differences in heart structure and conduction system may have existed, implying that the regions sampled in CMR studies may have been different between different strains. However, care was taken to acquire data on in midventricular slices in all experiments, thereby ameliorating this concern.

Conclusions

Our results demonstrate that nNOS plays a dominant role in modulating both LTCC and contractile function in vivo. Further, our results suggest that nNOS separately modulates LTCC function at baseline, and contractile function during β-adrenergic stimulation. While other groups have demonstrated that eNOS plays a role in preventing ventricular arrhythmia (50), and in modulating the cardiomyocyte response to stretch (32), our results suggest that, in vivo, the role of eNOS in modulating calcium cycling is less dominant than that of nNOS in the normal heart.

GRANTS

This work was supported by American Heart Association predoctoral grant AHA0815242E (to M. H. Vandsburger), and American Heart Association Established Investigator Award 0540060N and National Institutes of Health R01EB001763 (to F. H. Epstein).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

Author contributions: M.H.V., B.A.F., X.Z., and F.H.E. conception and design of research; M.H.V. performed experiments; M.H.V. analyzed data; M.H.V., B.A.F., C.M.K., and F.H.E. interpreted results of experiments; M.H.V. prepared figures; M.H.V. drafted manuscript; M.H.V., B.A.F., C.M.K., X.Z., and F.H.E. edited and revised manuscript; M.H.V., B.A.F., C.M.K., X.Z., and F.H.E. approved final version of manuscript.