OUP user menu

Multi-parameter in vivo cardiac magnetic resonance imaging demonstrates normal perfusion reserve despite severely attenuated β-adrenergic functional response in neuronal nitric oxide synthase knockout mice

Moriel H. Vandsburger , Brent A. French , Patrick A. Helm , Rene Jack Roy , Christopher M. Kramer , Alistair A. Young , Frederick H. Epstein
DOI: http://dx.doi.org/10.1093/eurheartj/ehm241 2792-2798 First published online: 30 June 2007


Aims The role of neuronal nitric oxide synthase (nNOS) in regulating contractile function remains controversial, and in regulating myocardial perfusion is uninvestigated. We used magnetic resonance imaging (MRI) to phenotype nNOS−/− and wild-type (WT) mice regarding left ventricular (LV) structure, baseline function, β-adrenergic responsiveness, and perfusion reserve.

Methods and results Cine MRI showed higher LV mass to end-diastolic volume ratio (2.3 ± 0.2 mg/µL nNOS−/− vs. 1.7 ± 0.1 mg/µL WT; P=0.032) and LV ejection fraction (64.9 ± 2.1% nNOS−/− vs. 55.8 ± 1.1% WT; P = 0.003) in nNOS−/−. Myocardial tagging demonstrated similar baseline systolic circumferential strain (Ecc) in nNOS−/− and WT. With dobutamine, the normal change in Ecc was nearly absent in nNOS−/− (−0.5 ± 0.3% nNOS−/− vs. −2.2 ± 0.3% WT; P = 0.001), and the systolic strain rate (dEcc/dt) response to dobutamine seen in WT was reduced in nNOS−/− (−29 ± 13%/s nNOS−/− vs. −106±16%/s WT; P = 0.001). Diastolic strain rate increased significantly with dobutamine only in WT. Arterial spin labelling showed that baseline perfusion and perfusion reserve with either dobutamine or an adenosine receptor agonist are normal in nNOS−/−.

Conclusion MRI provides non-invasive in vivo evidence that nNOS does not play a role in basal contractile function or myocardial perfusion, but is required for increasing cardiac inotropy and lusitropy upon β-adrenergic stimulation.

  • Neuronal nitric oxide synthase
  • Cardiac function
  • Heart
  • MRI
  • Adrenergic stimulation
  • Myocardial tagging


Nitric oxide (NO) is generated in normal cardiomyocytes by both neuronal nitric oxide synthase (nNOS) and endothelial NOS (eNOS) and, under different circumstances, can positively or negatively modulate cardiac function.14 However, the specific role of nNOS in basal left ventricular (LV) function and also during β-adrenergic stimulation remains controversial.1 Specifically, one series of studies found that, compared with wild-type (WT) mice, isolated myocytes from nNOS−/− mice have a blunted cell shortening response to β-adrenergic stimulation with isoproterenol,5 and similarly, intact hearts from nNOS−/− mice have a blunted dP/dt response to isoproterenol5 or increased pacing frequency.6 Conversely, other studies found greater shortening under basal conditions7,8 and in response to isoproterenol in the isolated myocytes of nNOS−/− mice.7 Interestingly, in vivo data from this group showed decreased end-systolic elastance (Ees) and preload recruitable stroke work (PRSW) responses to dobutamine in nNOS−/− mice.9 The most recent in vitro data from this series of studies confirmed earlier findings of greater cell shortening at baseline and during β-adrenergic stimulation in nNOS−/− myocytes when compared with myocytes from WT mice, although both WT and nNOS−/− showed similar increases in shortening over baseline.10 Divergent results may be due to methodological differences involving temperature and concentrations of isoproterenol.2,11 Clarification of the role of nNOS is important as recent evidence suggests a significant role for nNOS in heart failure.9,1214

Although the role of nNOS in cardiac function has been examined previously, less is known about the role of nNOS in vascular tone and perfusion. Assessments of the involvement of nNOS in basal vascular tone based upon measurements of systolic blood pressure have shown variability.15 With respect to maintaining vascular homeostasis,16 one study showed preferential augmentation of vascular nNOS expression during cerebral hypoxia.17 In a separate study, hyperoxia caused a temporary depression of regional cerebral blood flow in WT mice which was prolonged in nNOS−/− mice.18 Also, nNOS has been shown to cause vasodilatation of the afferent arteriole in the kidney,15 as well as to contribute to maintenance of renal homeostasis.15 The role of nNOS in myocardial perfusion has not been investigated previously.

Cardiac magnetic resonance (CMR) has become a reference standard modality for non-invasive phenotyping of the mouse heart in vivo.19 In addition to cine MRI20 for quantifying myocardial mass,21 LV volume, wall thickness, and ejection fraction (EF), myocardial tagging20,22 can be used to accurately assess contractile function by measuring myocardial strain and strain rate, and arterial spin labelling (ASL) can uniquely measure myocardial perfusion.23,24 The purpose of the present study was to use these CMR techniques to non-invasively and quantitatively phenotype nNOS−/− mice in terms of cardiac structure, function, responsiveness to β-adrenergic stimulation, and perfusion reserve upon pharmacological vasodilation.



Sixteen WT and 16 nNOS−/− male mice on a 129S4/SvJae * C57BL/6 background (Jackson Laboratory, Bar Harbor, Maine, USA) 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) and were approved by the Animal Care and Use Committee at our institution. The WT mice were F2 hybrids of a B6129SF1/J×B6129SF1/J cross used by Jackson Laboratory to approximate littermate controls. Assessment of basal LV structure and function as well as strain and strain rate was performed at 8 ± 1 weeks of age, 2 days after which systolic blood pressure was measured using a non-invasive tail plethysmography system (Model BP 2000,Visitech Systems, Apex, NC, USA). One week later, ASL was used to assess myocardial perfusion. With peak myocardial circumferential shortening (Ecc) as the primary endpoint, sample sizes were powered to detect a difference in mean ΔEcc with dobutamine of 15% over baseline, with an SD of 7%, power of 0.95 and significance of 0.05.

Cardiac magnetic resonance preparation

Imaging was performed on a 4.7 T MRI system (Varian, Inc., Palo Alto, CA, USA) using a custom-built radiofrequency (RF) coil (Doty Scientific, Columbia, SC, USA). The RF coil, designed for ASL, consisted of an outer 6 cm long cylindrical Litz coil for whole-body RF transmission (used for non-selective inversion in ASL) and an inner 2 cm long cylindrical Litz coil for localized signal reception. Body temperature was maintained at 36.2 ± 0.2°C by circulating thermostated water, and anaesthesia was maintained using 1.25% isoflurane in O2 inhaled through a nose cone. The entire LV function imaging study took 2 h, and the ASL study took 2.5 h. During both experiments the mouse lay prone within the scanner. Heart rate (HR), respiration, and rectal temperature were monitored during imaging using a fibre optic, MR compatible system (Small Animal Imaging Inc., Stony Brook, NY, USA).

Baseline left ventricular structure and function

Baseline LV structure and function were assessed using a black blood cine technique as described previously.20 Six short-axis slices were acquired from base to apex, with slice thickness equal to 1 mm, in-plane spatial resolution of 0.2 × 0.2 mm2, and temporal resolution of 8–12 ms. Baseline EF, end-diastolic volume (EDV), end-systolic volume (ESV), myocardial mass, wall thickness, and wall thickening were measured from the cine images using the ARGUS analysis program (Siemens Medical Solutions, Princeton, NJ, USA). EDV and ESV were then indexed to body mass (EDVI and ESVI, respectively). Mass to volume ratio (MVR) was calculated as the ratio of myocardial mass to EDV. One long-axis slice was also acquired at baseline with identical imaging parameters in seven WT and eight nNOS−/− studies.

Myocardial strain and strain rate

Myocardial tagging was performed on two mid-ventricular short-axis slices as described previously20,22 at baseline and during the constant infusion of dobutamine into the intraperitoneal (IP) cavity at doses of 20 and 40µg/kg×min during the same imaging session. To measure two-dimensional myocardial strain with high spatial resolution, two orthogonal sets of line-tagged images were acquired at each slice location. Strain analysis of the tagged images was performed using the Findtags method25,26 generating circumferential shortening data for three layers of myocardium: subendocardium, mid-wall, and subepicardium. Peak mid-wall circumferential systolic strain (Ecc) and systolic strain rate (dEcc/dt) were used to assess regional contractile function. Diastolic strain rate (dEcc/dt)diastolic was used to assess regional diastolic function. Systolic and diastolic strain rates were calculated as the slopes of the contraction and relaxation portions of the Ecc-time curve, respectively.

Longitudinal shortening

The global baseline shortening in the longitudinal direction was estimated by measuring the length of a line drawn from the center of the mitral valve to the epicardial apex of the LV at end-diastole (ED) and end-systole (ES).

Systolic blood pressure

Systolic blood pressure was measured in WT and nNOS/− mice at baseline (n = 10) and after constant IP infusion of dobutamine at a dose of 20 µg/kg×min (n = 6) using identical anaesthesia and infusion delivery as during the MR procedure. A minimum of five baseline waveforms were obtained for each mouse at both stimulation levels. Systolic blood pressure measurements were obtained from each waveform, and then averaged together for each individual mouse.

Myocardial perfusion

ASL was used to quantify myocardial perfusion at rest and either during IP infusion of dobutamine (20 µg/kg×min) or after bolus IP injection of the highly selective A2A adenosine receptor agonist ATL313 (12.5 µg/kg body weight) (Adenosine Therapeutics, Charlottesville, VA, USA).23,24 The advantage of selective A2A adenosine receptor stimulation is that it achieves maximum vasodilatation without the deleterious side-effects of adenosine.27 Specifically, the ASL method described by Kober et al.23 was used to measure myocardial T1 recovery curves after application of both non-selective and (2 mm thick) slice-selective inversion pulses. A hyperbolic secant RF pulse was used for magnetization inversion. Estimation of T1 following the non-selective (T1ns) and slice-selective (T1sel) inversion pulses was performed using a least-squares fit of the measured data to the theoretical T1 recovery curve. ASL was performed in one midventricular short-axis slice with an imaging slice thickness of 1 mm and in-plane spatial resolution of 0.2 × 0.2 mm2. After estimating T1ns and T1sel, perfusion was calculated using the equation P = (1/T1sel−1/T1ns)×T1ns/T1blood×λ, 24where T1blood was measured as 1.5 s and the partition coefficient (λ) was 0.95.24

Statistical analyses

Statistical analyses of myocardial structure, global function, systolic blood pressure, and perfusion were performed using SigmaStat (Systat Software Inc., Point Richmond, CA, USA). All these comparisons were made using t-tests. Differences in Ecc, dEcc/dt, and (dEcc/dt)diastolic at baseline, 20 µg/kg×min, and 40 µg/kg×min dobutamine, and between WT and nNOS−/− mice were assessed using repeated measures analyses, assuming compound symmetry. Analyses repeated with other assumed covariance structures, including an unstructured covariance matrix, produced nearly identical results. Comparison of dose effects between WT and nNOS−/− mice was performed using F-tests. Comparisons between WT and nNOS−/− mice for the difference between baseline and 20 µg/kg×min dobutamine, or baseline and 40 µg/kg×min dobutamine were examined using contrasts. The analyses were carried out in SAS 9.1.3 PROC MIXED (SAS Inc., Cary, NC, USA). All values in the text, tables, and graphs are presented as mean ± SEM.


Left ventricular structure

Examples of midventricular ED and ES images of WT and nNOS−/− mice are shown in Figure 1. As indicated by lower EDV and ESV, as well as EDVI and ESVI (Table 1), nNOS−/− mice had significantly smaller LV cavity size and lower LV mass when compared with WT mice at 8 weeks. However, MVR was significantly higher (Table 1). No differences were observed in LV wall thickness, body mass, or LV mass indexed to body mass (LVMI) between WT and nNOS−/− mice (Table 1).

Figure 1

Typical black-blood midventricular short-axis end-diastolic and end-systolic images from cine MRI sets of wild-type and nNOS−/− mice. Six contiguous short-axis cine image slices spanning the heart from base to apex are used to assess three-dimensional parameters such as chamber volumes and ejection fraction. Cine MRI detected reduced end-diastolic volume index and end-systolic volume index as well as increased ejection fraction and mass to volume ratio in nNOS−/− mice when compared with WT mice.

View this table:
Table 1

Baseline left ventricular structure for wild-type and nNOS−/− mice

WT (n = 12)nNOS−/− (n = 10)P
End-diastolic volume (µL)56.7 ± 3.235.8 ± 1.80.001
End-systolic volume (µL)25.0 ± 1.512.7 ± 1.20.001
End-diastolic volume index (µL/g)2.1 ± 0.11.5 ± 0.10.001
End-systolic volume index (µL/g)0.9 ± 0.10.5 ± 0.10.001
LV mass (mg)97 ± 580 ± 30.011
MVR (mg/µL)1.7 ± 0.12.3 ± 0.20.032
Wall thickness (mm)1.1 ± 0.11.1 ± 0.10.894
Body mass (g)26.8 ± 1.024.1 ± 0.70.065
LV mass/body mass (mg/g)3.6 ± 0.23.3 ± 0.10.178

Baseline left ventricular function and haemodynamics

Cine CMR revealed that nNOS−/− mice have a significantly lower stroke volume (SV) than WT mice (Table 2). Also, as HR is higher in nNOS−/− mice (Table 2), cardiac output indexed to body mass was similar between WT and nNOS−/− mice (Table 2). In addition, nNOS−/− mice have a higher baseline EF than WT mice (Table 2), which is explained by their higher MVR (Appendix). Indeed, the ratio of EF between WT and nNOS−/− mice (0.86) was essentially equal to the ratio of MVR between WT and nNOS−/− mice (0.87). There were no differences between WT and nNOS−/− mice in wall thickening (Table 2). Baseline systolic blood pressure was lower in nNOS−/− mice (Table 3).

View this table:
Table 2

Baseline left ventricular function for wild-type and nNOS−/−mice

WT (n = 12)nNOS−/− (n = 10)P
Stroke volume (µL)31.7 ± 1.923.0 ± 1.10.001
Heart rate (b.p.m.)419 ± 9490 ± 150.001
Cardiac output index (µL/g×min)492 ± 23470 ± 240.517
Ejection fraction55.8 ± 1.164.9 ± 2.10.003
Wall thickening (%)39.4 ± 3.539.9 ± 3.30.908
View this table:
Table 3

Baseline and dobutamine (20 µg/kg×min) systolic blood pressure for wild-type and nNOS−/−mice

Baseline systolic blood pressure (mmHg) (n = 10)106 ± 494 ± 30.023
Dobutamine systolic blood pressure (mmHg) (n = 6)112 ± 8103 ± 90.501

Baseline strain and strain rate

Examples of ED and ES tagged images of WT mice at baseline and during dobutamine infusion at 20 µg/kg×min are shown in Figure 2. Also, examples of Ecc–time curves are shown for WT and nNOS−/− mice at baseline and with 20 µg/kg×min dobutamine in Figure 3. Baseline Ecc (−13.6 ± 0.4% WT, n = 10 vs. −13.2 ± 0.6% nNOS−/−, n = 10, P = 0.640) was similar for WT and nNOS−/− mice, indicating similar baseline contractile function. Baseline dEcc/dt (−343 ± 15%/s WT, n = 10 vs. −380 ± 13%/s nNOS−/−, n = 10, P = 0.047) was higher in nNOS−/− mice, which, given the baseline Ecc results, reflects the increased heart rate of nNOS−/− mice. Also, both groups had similar baseline diastolic function as assessed by (dEcc/dt)diastolic (205 ± 12%/s WT, n = 10 vs. 238 ± 12%/s nNOS−/−, n = 10, P = 0.111). Baseline longitudinal shortening as measured from the long-axis cine images was also similar between the two groups (10.7 ± 2.8% WT, n = 7, vs. 10.9 ± 2.3% nNOS−/−, n = 8, P = 0.857).

Figure 2

Examples of end-diastolic and end-systolic tagged images from cine data sets of a wild-type mouse acquired at baseline and during dobutamine infusion (20 µg/kg×min). For each slice, two sets of images were acquired with tags applied in orthogonal directions (the second direction is not shown). Orthogonal line-tagged images are combined during image analysis for the computation of two-dimensional midwall circumferential shortening (Ecc).

Figure 3

Ecc vs. time curves showing midwall circumferential shortening (Ecc) at baseline and during dobutamine infusion for wild-type and nNOS−/− mice. Significant increases in Ecc and strain rate (dEcc/dt) are observed in wild-type mice upon dobutamine infusion, however, this increase is essentially absent in nNOS−/− mice.

Dobutamine response

At 20 µg/kg×min dobutamine, there was a four-fold difference in the change in Ecc from baseline between WT and nNOS−/− mice (−2.2 ± 0.3% WT, n = 10 vs. −0.5 ± 0.3% nNOS−/−, n = 10, P = 0.001), which was maintained at 40 µg/kg×min dobutamine (−2.3 ± 0.4% WT, n = 10 vs. −0.4 ± 0.3% nNOS−/−, n = 10, P = 0.001) (Figure 4). In addition, significant increases in dEcc/dt over baseline at 20 (−106 ± 16%/s WT, n = 10 vs. −29 ± 13%/s nNOS−/−, n = 10, P = 0.001) and 40 µg/kg×min dobutamine (−108 ± 18%/s WT, n = 10 vs. −34 ± 18%/s nNOS−/−, n = 10, P = 0.001) seen in WT mice were absent in nNOS−/− mice. Similarly, significant increases in (dEcc/dt)diastolic in WT mice at both 20 (55 ± 9%/s WT, n = 10 vs. 13 ± 7%/s nNOS−/−, n = 10, P = 0.001) and 40 µg/kg×min dobutamine (80 ± 14%/s WT, n = 10 vs. 19 ± 8%/s nNOS−/−, n = 10, P = 0.001) were essentially absent in nNOS−/− mice (Figure 5). While differences in Ecc and dEcc/dt were apparent, WT and nNOS−/− mice displayed similar systolic blood pressure in response to 20 µg/kg×min dobutamine (Table 3). Finally, similar HR responses to dobutamine were observed in WT and nNOS−/− mice at 20 µg/kg×min dobutamine (50 ± 9 b.p.m. WT, n = 10 vs. 36 ± 8 b.p.m. nNOS−/−, n = 10, P = 0.435) and 40 µg/kg×min dobutamine (79 ± 13 b.p.m. WT, n = 10 vs. 72 ± 10 b.p.m. nNOS−/−, n = 10, P = 0.791).

Figure 4

Effect of β-adrenergic stimulation on Ecc. The change in Ecc vs. baseline was greater in wild-type (n = 10) vs. nNOS−/− (n = 10) mice at 20 and 40 µg/kg×min dobutamine.

Figure 5

Effect of dobutamine on (dEcc/dt)diastolic. Wild-type (n = 10) mice show significant increases in (dEcc/dt)diastolic over baseline at 20 and 40 µg/kg × min dobutamine. No significant increase was measured in nNOS−/− mice (n = 10).

Myocardial perfusion

Myocardial perfusion was not significantly different in nNOS−/− mice at baseline (4.3 ± 0.3 mL/g×min WT, n = 16 vs. 5.4 ± 0.6 mL/g×min nNOS−/−, n = 16, P = 0.081), during administration of 20 µg/kg×min dobutamine (6.6 ± 0.8 mL/g×min WT, n = 6 vs. 6.5 ± 0.8 mL/g×min nNOS−/−, n = 6, P = 0.946), or after administration of ATL313 (11.5 ± 1.0 mL/g×min WT, n = 10 vs. 13.2 ± 0.7 mL/g×min nNOS−/−, n = 10, P = 0.187) (Figure 6). However, at baseline there was a trend towards higher perfusion in nNOS−/− mice. Perfusion reserve was similar in the two groups for dobutamine (1.8 ± 0.2 WT, n = 6 vs. 1.8 ± 0.3 nNOS−/−, n = 6, P = 0.976) as well as ATL313 (2.7 ± 0.3 WT, n = 10 vs. 2.2 ± 0.2 nNOS−/−, n = 10, P = 0.151).

Figure 6

Myocardial perfusion as measured by arterial spin labelling is similar in wild-type and nNOS−/− mice at baseline (n = 16), upon infusion of dobutamine (20 µg/kg×min) (n = 6), and upon application of the vasodilator ATL313 (12.5 µg/kg body weight) (n = 10).


The major findings of this study are that: (i) LV cavity size is significantly lower in nNOS−/− when compared with WT mice; (ii) baseline contractile function is similar in WT and nNOS−/− mice despite higher EF in nNOS−/− mice which is attributed to increased MVR (Appendix); (iii) nNOS−/− mice have severely blunted in vivo inotropic and lusitropic responses to dobutamine despite a normal HR response; and (iv) nNOS−/− mice have normal baseline perfusion and perfusion reserve in response to both dobutamine and an adenosine receptor agonist. In contrast to other techniques, such as studying isolated myocytes or using invasive pressure–volume catheters, CMR enables the non-invasive quantitative assessment of in vivo structure, function, and perfusion.

Our finding of decreased LV cavity size in nNOS−/− mice agrees with a trend towards lower ESV and EDV in nNOS−/− mice shown by Khan et al.6 Likewise, our finding of similar LVMI in nNOS−/− and WT mice agrees with previous findings8 including those in mice aged 2 months,5 and 2–3 months,12 but not with increased LVMI found in nNOS−/− mice at 4 months.9 The higher LVMI at older ages reflects progression towards LV hypertrophy in nNOS−/− mice after 8 weeks of age.5,9

Results of previous studies regarding the role of nNOS in basal contractile function have shown considerable variability. Findings in nNOS−/− mice of higher EF,8 ventricular elastance (Ees),9 [(dP/dt)/Pid)5 as well as PRSW in sham-operated nNOS−/− mice9 have previously led to conclusions of elevated basal contractile function.1,3,8 However, in agreement with findings of similar coupling of ventricular to arterial elastance (Ees/Ea)5,6 and whole heart relaxation values,6 our Ecc and (dEcc/dt)diastolic data suggest that basal contractile function and relaxation, respectively, are similar in nNOS−/− and WT mice. Given similar Ecc in WT and nNOS−/− mice, the heightened dEcc/dt in nNOS−/− mice can be explained by increased basal HR. Although our EF results agree with echocardiography,8,9 EF is sensitive to LV geometry.28 Using a cylindrical model of the LV, we demonstrate (Appendix) that the increased MVR of nNOS−/− mice, with maintained circumferential and longitudinal shortening, is sufficient to explain the increased EF in these mice. The divergence of these findings from those previously mentioned may be attributed to the fact that this study was conducted in younger mice (8 weeks). In addition, decreased HR compared with previous studies5,6,8,9,29 or lower baseline systolic blood pressure in nNOS−/− compared with WT mice may have masked subtle differences in baseline contractile function seen in other studies.5,8,9 Also, it is important to note that WT mice used in this study were not littermate controls of the nNOS−/− mice, but separate mice of similar genetic background.

Our finding of a nearly completely absent inotropic response to β-adrenergic stimulation in nNOS−/− mice agrees with the in vivo normalized dP/dt and Ees results of Barouch et al.5 and PRSW and Ees results of Dawson et al.9 Because midwall myofibres are circumferentially oriented, midwall Ecc and dEcc/dt are approximately in vivo measurements of myocyte shortening and shortening velocity, respectively. The increase in midwall Ecc of 17% in WT mice at 20 µg/kg×min dobutamine mirrors the normal dobutamine response in humans.30 In addition, myocardial tagging detected a related change in (dEcc/dt)diastolic with increasing dobutamine dose in WT but not in nNOS−/− mice, mirroring the finding of an attenuated lusitropic response to β-adrenergic stimulation by Dawson et al.9 While the Ecc, dEcc/dt, and (dEcc/dt)diastolic responses were different between WT and nNOS−/− mice, the HR responses of both groups were similar, increasing by approximately 10 and 20% at dobutamine doses of 20 and 40 µg/kg×min, respectively. In addition, the HR increase from 20 to 40 µg/kg×min dobutamine demonstrates that the dose of 20 µg/kg×min is submaximal. Although we cannot exclude the possibility that differences in changes in LV relaxation with dobutamine could result from increased MVR in nNOS−/− mice, the presence of a HR response but absence of Ecc, dEcc/dt, and (dEcc/dt)diastolic responses in nNOS−/− mice supports the hypothesis that NO from nNOS is critical for enhanced contractile function under β-adrenergic stimulation.

The relationship between myocardial perfusion and contractile reserve in nNOS−/− mice has not been investigated previously. Perfusion reserve measurements for WT mice were similar to those reported by other groups.23 Our finding that perfusion reserve is similar in WT and nNOS−/− mice despite an attenuated contractile response to β-adrenergic stimulation in nNOS−/− mice helps to elucidate the role of nNOS in contractile function. Dobutamine serves as a strong β1 and β2 agonist, and a weak α1 agonist.31 The net effect of strong vascular β2 (vasodilatation) and weak vascular α1 (vasoconstriction) stimulation is vasodilatation31 and is manifested in both WT and nNOS−/− mice as increased myocardial perfusion. The impact of a strong cardiac β1 agonist is increased cardiac inotropy, which was seen only in WT mice as increased Ecc and dEcc/dt. Vasodilatation without increased inotropy in nNOS−/− mice indicates that the absence of a contractile reserve is not due to inadequate perfusion reserve. Further, similar increases in myocardial perfusion with ATL313 in WT and nNOS−/− mice show that nNOS−/− mice are not perfusion-limited even under conditions of greater vasodilatation. When examined alongside similar increases in HR and systolic blood pressure in both WT and nNOS−/− mice in response to dobutamine, the attenuated contractile response of nNOS−/− mice indicates a cardiomyocyte-specific role for nNOS with respect to modulation of contractile function.

A limitation of this study is that LV pressure was not measured during imaging, thus any potential differences in LV preload and afterload and consequent impacts on LV contractile function were unaccounted for. A second limitation is that our measurements were made under the influence of isoflurane, which at high concentrations is a vasodilator.23 However, we minimized the dose of isoflurane, and maintained physiological body temperature during MR scanning. Furthermore, both groups of mice were imaged under the same conditions and demonstrated similar perfusion reserve. In addition, the nNOS−/− mice used in this study had systemic, and not cardiac-specific nNOS ablation. Through its participation in autonomic regulation of peripheral targets,15 nNOS may affect in vivo LV function in a manner not seen in in vitro experiments. We were unable to isolate this effect in our experiments, which could be partially responsible for differences seen between in vivo and in vitro results.

In light of the controversy regarding nNOS,13 our findings indicate that systemic nNOS gene deletion does not affect LV systolic function under basal conditions but leads to a severe attenuation of the in vivo contractile response to β-adrenergic stimulation without affecting myocardial perfusion. Our study also indicates that adequate increases in perfusion are available to fuel any increased oxygen demand required to enhance contractile function upon β-adrenergic stimulation. As our study utilized non-invasive myocardial tagging and ASL CMR methods, these data add to our understanding of the in vivo role of nNOS. In agreement with previous in vivo studies our findings indicate that NO from nNOS serves as a critical signalling agent for increasing the inotropic5 and lusitropic9 functional responses to β-adrenergic stimulation. Increased knowledge of the role of nNOS may be clinically important given recent findings of increased nNOS mRNA and protein expression and translocation of nNOS to the sarcolemma in human heart failure.14,32


This work was supported by NIH grant RO1 EB 001763 and by an American Heart Association Established Investigator Award to F.H.E.

Conflict of interest: none declared.


We propose that increased EF found in nNOS−/− mice in this study can be attributed to the increased MVR in this group. It is well known that EF is sensitive to LV geometry, and in particular to the MVR. Following33 we use an incompressible cylindrical model of the LV, and calculate EF as a function of geometry and shortening. Let the inner and outer radii of the cylinder be Ri and Ro, respectively, at ED and ri and ro at ES, respectively. The EF is

Embedded Image

where L is the LV length at ED, l is the length at ES. λL is the extension ratio in the longitudinal direction.

Embedded Image

where %SL is the percent longitudinal shortening and %SCi is the percent circumferential shortening at the inner surface. λCi is the extension ratio in the circumferential direction at the inner surface. Embedded Image

We define a midwall surface which divides the mass of the heart into equal parts, i.e. a radius Rm such that LRo2LRm2 = LRm2LRi2, or Rm2 = (Ro2+Ri2)/2. At ES, myocardium at Rm has moved to rm. Owing to incompressibility, the volume of myocardium between Rm and Ri at ED is preserved at ES.

Embedded Image


Embedded Image

The EF is therefore related to the midwall extension ratio λCm = rm/Rm. Embedded Image Concentric hypertrophy can be quantified by the MVR:

Embedded Image

which can be written as Embedded Image

Substituting this equation into the above relation for EF yields

Embedded Image

The EF can therefore be seen to depend on the longitudinal shortening (λL), the midwall circumferential shortening (λCm) and the degree of concentric hypertrophy (MVR).

In the current study, the contraction in the longitudinal and circumferential directions was found to be similar between WT and nNOS−/− mice. The ratio of the mean WT EF to the mean nNOS−/− EF was 0.86. The ratio of mean (0.5MVR+1) in each group was 0.87. These results agree with the above equation, and support the hypothesis that the increased MVR is sufficient to explain the increased EF.


View Abstract