Author + information
- Received August 28, 2006
- Revision received June 22, 2007
- Accepted July 31, 2007
- Published online October 30, 2007.
- Wesley D. Gilson, PhD⁎,§,
- Frederick H. Epstein, PhD⁎,†,§,
- Zequan Yang, MD, PhD⁎,§,
- Yaqin Xu, MD, PhD⁎,
- Konkal-Matt R. Prasad, PhD⁎,
- Marie-Claire Toufektsian, PhD⁎,
- Victor E. Laubach, PhD‡,§ and
- Brent A. French, PhD⁎,†,§,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Brent A. French, Department of Biomedical Engineering, University of Virginia, Box 800759, Charlottesville, Virginia 22903.
Objectives We sought to determine the effect of inducible nitric oxide synthase (iNOS) expression on regional contractile function and left ventricular (LV) remodeling after reperfused myocardial infarction (MI).
Background Inducible nitric oxide synthase is known to contribute to global LV dysfunction after a large MI, but the mechanisms underlying this dysfunction remain unclear.
Methods We used immunohistochemistry to investigate the distribution of iNOS expression in wild-type (WT) and iNOS knockout (KO) mice early (day 1) and late (day 28) after reperfused MI. We also used serial cardiac magnetic resonance imaging at baseline and at 1, 7, and 28 days after MI to assess LV volumes, ejection fraction (EF), regional circumferential strain (Ecc), and day 1 infarct size.
Results At baseline, LV volumes and EF were similar between groups. Day 1 infarct size was also similar between groups. Immunohistochemistry revealed that iNOS expression was abundant throughout the heart in WT mice on day 1 after MI, particularly near the infarct borderzone. On day 7 after MI, Eccin KO mice was significantly improved in some borderzone sectors compared with WT. The LV volumes were significantly lower in KO mice at days 7 and 28 compared with WT. The EF on days 7 and 28 was significantly higher in KO mice compared with WT. The circumferential extent of wall thinning was also significantly reduced in KO versus WT mice at days 7 and 28.
Conclusions Expression of iNOS contributes importantly to post-infarction contractile dysfunction and subsequent LV remodeling, suggesting new strategies to combat heart failure resulting from large MI.
Large anterior myocardial infarction (MI) results in acute impairment of left ventricular (LV) function and chronic progression of LV remodeling. Early after MI, elaboration of proinflammatory cytokines and chemokines stimulates the expression of inducible nitric oxide synthase (iNOS), which in turn produces large quantities of nitric oxide. Multiple studies have reported that cardiomyocytes express active iNOS in response to MI; however, previous studies using iNOS knockouts (KO) in mouse models of permanent coronary ligation have reported discordant results regarding the role of iNOS in post-infarction cardiac dysfunction and mortality (1–4). Moreover, no previous study has investigated the role of iNOS in the setting of reperfused MI or the impact of iNOS on regional contractile function or on the time course of post-infarction structural LV remodeling.
Cardiac magnetic resonance imaging (MRI) can serially assess changes in LV size, shape, and function after MI. Cine MRI has previously been used to study LV remodeling (5), cardiac hypertrophy (6), and heart development (7) in mice. Furthermore, contrast-enhanced MRI can accurately and noninvasively determine the size of MI (8), even in mice (9). Finally, myocardial tagging studies have demonstrated the ability to quantify regional intramyocardial contractile function in normal and post-infarction mice (10).
The objective of the present study was to test the hypothesis that iNOS KO mice would exhibit improvements in regional function, reduced LV remodeling, and reduced wall thinning after reperfused MI compared with wild-type (WT) mice.
All animals were studied under protocols approved by the University of Virginia Animal Care and Use Committee. The iNOS KO mice were generated as previously described (11) and were congenic with C57Bl/6. Age- and weight-matched WT C57Bl/6 mice were obtained from Jackson Labs (Bar Harbor, Maine). Myocardial infarction was induced with a 1-h occlusion of the left anterior descending coronary artery followed by reperfusion (5). Four WT and 4 KO male mice were studied by immunohistochemistry at days 1 and 28 after MI. Another 4 WT and 3 KO mice were used in serial studies of plasma nitrite/nitrate (NOx) levels after MI. Twelve WT and 12 KO male mice were studied by cardiac MRI at baseline (Bsl) and at post-MI days 1, 7, and 28.
iNOS expression and activity
Immunostaining for neutrophils, myoglobin, and iNOS was performed on formalin-fixed paraffin-embedded sections using Vecta-Stain Elite ABC (Vector Labs, Burlingame, California). Sections were incubated with rabbit antihuman myoglobin (Dako, Carpinteria, California) or rabbit antimurine NOS2 (sc650, Santa Cruz Biotech, Santa Cruz, California). Myoglobin and iNOS were stained brown with 3,3′-diaminobenzidine (DAB). Sections immunostained for iNOS were counterstained with hematoxylin. Sections first immunostained for myoglobin were incubated with rat monoclonal antimouse neutrophil antibody (Serotec, Raleigh, North Carolina). Neutrophils were stained purple with VIP (Vector Labs).
The metabolic end products of nitric oxide synthase (NOS) activity (NOx) were measured in plasma samples as described previously (12). Briefly, blood was withdrawn from jugular veins. Plasma samples were deproteinized by ultrafiltration (10,000 molecular weight cutoff), and nitrate was reduced to nitrite with nitrate reductase before a spectrophotometric NOx assay using modified Griess reagent (Szechrome NIT, Polyscience, Warrington, Pennsylvania). Plasma NOx concentrations were determined from a calibration curve generated from sodium nitrite standards.
All imaging was performed on a 4.7-T MRI system (Varian, Palo Alto, California). Anesthesia was induced with 2.0% inhaled isoflurane and maintained at 1.0% during imaging. Electrocardiograms (ECGs) and core body temperatures were monitored during imaging using an magnetic resonance (MR)-compatible system for small animals (SA Instruments, Stony Brook, New York). On day 1, an intraperitoneal line was introduced for the injection of gadolinium-diethylenetriamine pentaacetic acid.
For LV volumes, an ECG-triggered 2-dimensional (2D) double inversion recovery black-blood cine sequence (13) was used to obtain 7 to 8 contiguous 1-mm-thick short-axis slices from apex to base. Imaging parameters included: echo time 3.2 ms, repetition time 8 to 10 ms, flip angle 20°, field of view 2.56 × 2.56 cm2, matrix 128 × 128, and 14 to 16 phases per cardiac cycle.
An ECG-triggered 2D double inversion recovery black-blood cine tagging sequence (13) was used to quantify intramyocardial contractile function from 3 1-mm-thick slices centered on the midventricle (referred to as basal, mid, and apical). Imaging parameters included: repetition time 8 to 12 ms, echo time 4.7 ms, flip angle 20°, field of view 2.56 × 2.56 cm2, and matrix 192 × 96. The composite tagging flip angle was 180°, and tag separation was 0.6 mm. For 2D strain analysis, line-tagged images were acquired in orthogonal directions.
To determine infarct size on post-MI day 1, mice were injected intraperitoneally with 0.3 to 0.6 mmol/kg gadolinium-diethylenetriamine pentaacetic acid, and heavily T1-weighted (90° flip angle) gradient echo MR images were acquired 15 min later (9). The effective repetition time was made equal to 2 cardiac cycles, which ranged from 200 to 300 ms (for heart rates from 400 to 600 beats/min).
The LV volume analysis was performed by blinded personnel using Argus (Siemens, Iselin, New Jersey). Endocardial and epicardial borders were planimetered from black-blood cine images. End-diastolic and -systolic volumes, ejection fraction (EF), stroke volume, LV mass, and cardiac output were then computed.
Infarct size was measured from day 1 contrast-enhanced images using software developed in Matlab (Mathworks, Natick, Massachusetts). Infarct regions were defined as areas of hyperenhancement with signal intensities >3 standard deviations above the mean of remote regions and were reported as percentages of total LV mass (9).
Circumferential strain (Ecc) was measured from the basal, mid, and apical tagged images using Findtags (14). Each of these 3 LV slices was divided into 6 equiangular sectors: anterior, lateral, posterior, inferior, septal, and anteroseptal. Mean midwall Eccwas calculated within each sector at each time point for each group and was compared between the 18 different sectors over time as the hearts remodeled. The fraction of each of the above sectors containing contrast enhancement was computed, and sectors with a mean enhancement of >67% were defined as infarcted, those immediately adjacent were designated as borderzone, and those at least 1 sector removed from infarcted sectors were designated as remote.
The circumferential extent of wall thinning for the 3 selected slices was measured from day 7 and 28 end-diastolic images using Findtags. Thinned walls were defined as those with wall thicknesses <500 μm, and the angle encompassing the thinned wall (Fig. 1)was compared between groups.
For volumes, EF, cardiac output, and Ecc, statistical analysis included 2-way analysis of variance to confirm intragroup differences by mouse strain and time point. Temporal and between-group differences were analyzed using pair-wise multiple comparisons subtesting (Tukey test) using SigmaStat 3.0 (Systat Software Inc., San Jose, California). To account for the 58 (of 288) tagged images that could not be analyzed for Eccowing to substandard image quality, a general linear model was used in the analysis of variance. Infarct sizes in WT and KO mice were compared using the unpaired Student ttest. All data are presented as mean ± standard error.
iNOS expression and activity
Double-immunostaining (for neutrophils and myoglobin) (Figs. 2Aand 2C) and serial sections immunostained for iNOS (Figs. 2B and 2D) revealed iNOS expression in cardiomyocytes both inside and outside the infarct zone. Neutrophils were immunostained because they have been implicated in reperfusion injury, and myoglobin was used as a marker of cellular integrity to distinguish between viable and necrotic myocardium (15). At 24 h after reperfusion in mice, neutrophil infiltration into the necrotic myocardium was well underway with few neutrophils located near viable (i.e., myoglobin-positive) cardiomyocytes and abundant neutrophils located in the infarct zone (Figs. 2A [low power] and 2C [high power]). Careful analysis of serial sections immunostained for iNOS (Figs. 2B [low power] and 2D [high power]) revealed that the highest levels of iNOS expression at 24 h after reperfusion were found in nonviable cardiomyocytes located immediately adjacent to viable cardiomyocytes. Figure 2E shows that iNOS immunoreactivity was readily detected in viable myocytes bordering the mature scar tissue as late as 28 days after MI: a time when LV remodeling is essentially complete in mice (5). Figure 2F shows iNOS immunoreactivity in cardiomyocytes bordering the mature scar and in granulation tissue within the scar.
The metabolic end-products of NOS activity (NOx) were monitored in serial plasma samples collected at Bsl and at 1, 7, and 28 days after MI in WT and KO mice. The results (Fig. 2G) indicate that NOx levels in WT mice were significantly elevated at days 1 and 28 after MI, whereas there were no significant changes in NOx plasma levels in iNOS KO mice at any of these time points.
Global LV size, function, and wall thickness
All mice in the MRI study survived surgery and imaging. Infarct size on day 1 after MI was similar in WT and KO mice (42.4 ± 7.0% and 39.7 ± 6.0% of LV mass, respectively; p = NS). The segmental distribution of infarction was also similar between groups (Fig. 3).Infarct sizing was not possible in 1 KO mouse owing to corrupt image datafiles. Mean body weight and heart rates were similar between groups at every time point examined, except body weight was lower in KO mice at day 28 after MI, and heart rate was higher in KO mice at post-MI days 1 and 7 (Table A1 in the Online Appendix). Examples of day 1 midventricular end-systolic cine, contrast-enhanced, and tagged images are shown in Figure 4.
End-diastolic and -systolic volumes, as determined from cine MR images, were similar in WT and KO mice at Bsl and on day 1 (Fig. 5A).Both groups had significant increases in end-diastolic and end-systolic volumes from Bsl to post-MI day 28. However, end-diastolic volume by post-MI day 7 was significantly lower in KO mice (51.8 ± 13.8 μl) compared with WT (65.9 ± 10.7 μl; p = 0.005) and remained lower at day 28 (79.4 ± 14.8 μl in WT vs. 59.8 ± 14.2 μl in KO; p < 0.001) (Fig. 4A). Similarly, end-systolic volume was significantly lower in KO mice at day 7 (41.9 ± 11.0 μl in WT vs. 29.1 ± 11.0 μl in KO; p = 0.005) and day 28 (55.6 ± 19.0 μl in WT vs. 35.1 ± 8.9 μl in KO; p < 0.001) after MI (Fig. 5A). The LV mass was slightly lower in the WT group than in the KO group at Bsl but was equal by day 7 and greater by day 28 (Table A1 in the Online Appendix).
At Bsl, EF was similar between groups (56.8 ± 4.3% in WT and 59.0 ± 8.0% in KO; p = NS). Both groups had a significant decline in EF at day 1 after MI (to 39.1 ± 6.4% in WT and 43.1 ± 7.6% in KO), but no significant difference was found between groups. The EF in WT mice progressively declined on day 7 (36.8 ± 7.6%) and day 28 (31.5 ± 9.6%), whereas KO mice maintained their EF at near day 1 levels (45.1 ± 8.4% on day 7 and 41.1 ± 9.4% on day 28), which led to significant differences between groups at later time points (p = 0.023 on day 7; p = 0.008 on day 28) (Fig. 5B). Both groups exhibited a significant decline in stroke volume on day 1 that partially recovered by day 7. However, stroke volume and cardiac output were similar between groups at all time points (Table A1 in the Online Appendix).
The circumferential extent of wall thinning was significantly reduced in KO mice in the basal, mid, and apical slices of the heart compared with WT mice as early as post-MI day 7 (Fig. 6).In all slices, the circumferential extent of wall thinning progressively increased in the WT mice out to day 28. In contrast, apical and mid slices in the KO mice showed a much slower progression, with the basal (minimally infarcted) slice remaining unchanged from day 7 to 28. Results for 1 of the day 7-KO mice, 2 of the day 28-KO mice, and 1 of the day 28-WT mice were excluded from wall thinning analysis owing to lack of conspicuity.
Regional LV function
At Bsl, Eccin all sectors was similar (Figs. A1 to A3 in the Online Appendix) with mean values between −0.15 and −0.17. At post-MI day 1, Eccsignificantly declined in the basal anterior, lateral, and posterior sectors. Significant declines were also found in all apical and midventricular sectors except the apical septum and midventricular septum. There were no significant differences between groups in any of these sectors at day 1.
In Figure 7,results are presented for those sectors demonstrating a statistically significant difference in Eccbetween WT and KO mice during the 28-day time course. Specifically, the borderzone sectors encompassing the midventricular and apical anterior septum, which were adjacent to nearly completely infarcted anterior sectors, demonstrated a significant decline in function at day 1 in both groups followed by recovery of function at day 7 in KO mice but not in WT mice (p < 0.05). Interestingly, the difference observed at day 7 was transient, because function in these sectors in WT mice recovered to a similar extent as in KO mice by day 28. In the midventricular septum sector, which was remote from the infarct, no significant decline in function was measured at day 1 in either KO or WT mice. However, significantly greater shortening (hypercontractile function) was observed at day 7 in KO but not WT mice (p < 0.05) in this sector, which subsequently returned to normal function by day 28 in both KO and WT mice. In this analysis, a total of 88 basal, 85 midventricular, and 57 apical tagged images were analyzed. Owing to low image quality (attributed to poor cardiac gating, significant wall thinning precluding tag detection, and/or partial volume effect) tagged images from 8 basal slices, 11 midventricular slices, and 39 apical slices could not be analyzed.
The major findings of this study are that, in iNOS KO mice after large reperfused anterior MI: 1) structural LV remodeling is reduced; 2) global LV function is improved; 3) contractile function is transiently improved in some borderzone sectors; and 4) the circumferential extent of wall thinning is reduced compared with WT animals. These differences are evident as early as 7 days after MI. Increases in post-infarction end-diastolic and -systolic volumes were approximately 50% less in KO mice compared with WT mice by day 28.
The cardiac MRI evaluation of iNOS KO mice was stimulated by immunohistochemistry studies that revealed abundant iNOS expression in the WT murine heart, both early (24 h) and late (28 days) after MI (Fig. 2). We have previously shown by Western blot analysis that iNOS protein expression is increased 10-fold in noninfarcted regions and 15-fold in infarcted regions of the mouse heart 24 h after reperfusion (12). At this same time point, immunohistochemistry revealed that neutrophils were largely localized in the necrotic infarct zone (Figs. 2A and 2C). Careful analysis of serial sections immunostained for iNOS revealed that the highest levels of iNOS expression could be found in nonviable cardiomyocytes located immediately adjacent to viable cardiomyocytes (Figs. 2B and 2D). This pattern of expression indicates that iNOS expression is associated with myocellular death and that high-level iNOS expression in nonviable myocytes may serve as a marker of cardiomyocytes that succumb to reperfusion injury (as opposed to those that perished earlier from ischemia).
The early expression of iNOS in the myocardium was complemented by expression detected late (28 days) after MI when LV remodeling is essentially complete (5). Nevertheless, iNOS expression was still readily detectable in cardiomyocytes bordering mature scar tissue. Interestingly, the highest levels of iNOS expression were found in cardiomyocytes surrounded by fibrosis at the very edge of the infarct scar (Fig. 2C). Granulation tissue located within the scar also showed high levels of iNOS expression (Fig. 2D), serving as a convenient internal positive control.
The present study demonstrates the utility of cardiac MRI to quantify initial infarct size and serially assess LV chamber sizes, mass, wall thickness, and wall strain in a mouse model of reperfused MI. Using these methods, infarct size was shown to be no different between iNOS KO and WT mice. The 28-day time course of LV remodeling was then compared between iNOS KO and WT mice. Chamber volumes were smaller and EF was greater in KO mice; however, both WT and KO mice adapted to perturbations in circulatory hemodynamics (16) by increasing heart rate and augmenting shortening in remote myocardium, thereby preserving both stroke volume and cardiac output. The MRI-based measurements over this time course are consistent with the results of Feng et al. (1) who used catheter-based LV pressure measurements to demonstrate improved global LV function at day 5. Although Feng et al. (1) also found decreased mortality at day 30 in iNOS KO mice, the lack of mortality in the present study may be explained by our use of an ischemia/reperfusion model instead of permanent coronary occlusion, because LV remodeling is known to be less severe after reperfused MI (17). Our results diverge from those reported by Sam et al. (2) and Jones et al. (3), who found no functional differences at 1 month after MI. Instead, Sam et al. (2) only detected differences between iNOS KO and WT mice at 4 months after MI. However, the use of permanent coronary ligation in all previous studies precludes direct comparison to the present study, in which reperfused coronary occlusion was used to model the contemporary clinical paradigm in which the vast majority of occluded coronaries are ultimately recanalized.
In recent years, the multifaceted roles of the 3 NOS isoforms in the maintenance of cardiovascular homeostasis have become well appreciated (18). Although most of the cardiovascular actions of nitric oxide are positive, it is important to note that the roles of the NOS isoforms in the settings of acute and chronic MI are exceedingly complex. For example, although the administration of nitrite (19) or augmentation of endogenous eNOS (20) before ischemia is known to reduce MI size, the presence or absence of iNOS had no effect on MI size in the current study (because myocardial iNOS levels are very low in normal mice before ischemia). This finding notwithstanding, iNOS is known to be a critical mediator of the cardioprotection afforded by late pre-conditioning (21). Thus, although the cardioprotective potential of iNOS is well established, iNOS expression must be induced in the heart before ischemia for it to be protective. This property has led some to suggest that chronic iNOS expression in the heart may have clinical potential in protecting patients against MI (22). However, the results of the present study and others (1,2,23) indicate that this strategy might be contraindicated.
Differences in post-MI regional contractile function between WT and KO mice in the present study were localized for the first time to myocardial sectors bordering infarcted sectors (the mid anterior septum and apical anterior septum) using myocardial tagging. This finding suggests that iNOS expression contributes to the post-MI contractile dysfunction that occurs in borderzone myocardium. The circumferential extent of wall thinning was also dramatically reduced in iNOS KO versus WT mice, despite direct proof that infarct sizes were equivalent on day 1 after MI. These results are intriguing in light of studies undertaken in a sheep model of reperfused MI, which found that borderzone myocardium extends during the remodeling process by recruiting previously normal contracting perfused myocardium into an enlarging hypocontractile zone (24,25). The reduced dilation, reduced wall thinning, and transient improvement in borderzone function observed in iNOS KO mice in the present study support the concept of nonischemic infarct extension (24) and indicate that iNOS contributes importantly to LV remodeling by potentiating infarct extension.
Limitations of this study include the lack of serial infarct sizing and the limited temporal resolution of data collection. The current MR technique for infarct imaging in mice lacks the sensitivity necessary to reliably determine infarct size late after MI. However, new contrast-enhanced MRI methods are under development to improve the contrast-to-noise ratio for assessing infarct size in mice (26). Additionally, the present study was prospectively designed to image mice at Bsl and at 1, 7, and 28 days after MI. Although this revealed significant differences between WT and KO mice as early as day 7, evaluating time points between days 1 and 7 after MI may better delineate the early development of post-MI LV dysfunction and remodeling. Furthermore, additional significant differences in regional contractile function between the KO and WT groups might have been evident if image quality in the tagged apical slices was better. Higher resolution methods, such as displacement-encoded imaging (27), may make quantifying contractile function less challenging, particularly in regions with significant wall thinning. One final limitation is that the slice positions were not adjusted during the process of LV remodeling. However, assessment of the change in long-axis dimension suggested a maximum slice shift of only 35% of the slice thickness.
This work demonstrates that iNOS is intimately involved in the LV dysfunction and remodeling that occurs after large MI. Immunohistochemical results and localized modulation of LV dysfunction in borderzone myocardium suggest that high-level iNOS expression in the borderzone may contribute importantly to the LV remodeling process. These results further suggest that selective inhibition of iNOS may be beneficial in retarding the LV remodeling process, thereby protecting patients against subsequent heart failure.
The authors thank R. Jack Roy for assistance in data analysis.
For a table on global parameters and figures showing the circumferential strain for sectors about the apical slice, the mid-ventricular slice, and the basal slice, please see the online version of this article.
Supported by National Institutes of Health grants R01-EB001763 (to Dr. Epstein), R01-HL58582, and R01-HL69494.
- Abbreviations and Acronyms
- circumferential strain
- ejection fraction
- inducible nitric oxide synthase
- left ventricle/ventricular
- myocardial infarction
- magnetic resonance imaging
- Received August 28, 2006.
- Revision received June 22, 2007.
- Accepted July 31, 2007.
- American College of Cardiology Foundation
- Feng Q.,
- Lu X.,
- Jones D.L.,
- Shen J.,
- Arnold J.M.
- Sam F.,
- Sawyer D.B.,
- Xie Z.,
- et al.
- Liu Y.H.,
- Carretero O.A.,
- Cingolani O.H.,
- et al.
- Franco F.,
- Dubois S.K.,
- Peshock R.M.,
- Shohet R.V.
- Wiesmann F.,
- Ruff J.,
- Hiller K.H.,
- et al.
- Kim R.J.,
- Fieno D.S.,
- Parrish T.B.,
- et al.
- Yang Z.,
- Berr S.S.,
- Gilson W.D.,
- Toufektsian M.C.,
- French B.A.
- Laubach V.E.,
- Shesely E.G.,
- Smithies O.,
- Sherman P.A.
- Toufektsian M.C.,
- Yang Z.,
- Prasad K.M.,
- et al.
- Gwechenberger M.,
- Mendoza L.H.,
- Youker K.A.,
- et al.
- Sutton M.G.,
- Sharpe N.
- Boyle M.P.,
- Weisman H.F.
- Schulz R.,
- Rassaf T.,
- Massion P.B.,
- Kelm M.,
- Balligand J.L.
- Guo Y.,
- Jones W.K.,
- Xuan Y.T.,
- et al.
- Li Q.,
- Guo Y.,
- Xuan Y.T.,
- et al.
- Jackson B.M.,
- Gorman J.H.,
- Moainie S.L.,
- et al.
- French B.A.,
- Beyers R.J.,
- Sureau F.C.,
- et al.
- Gilson W.D.,
- Yang Z.,
- French B.A.,
- Epstein F.H.