NADPH Oxidase-4 Driven Cardiac Macrophage Polarization Protects Against Myocardial Infarction–Induced Remodeling

Visual Abstract


SUMMARY
The reactive oxygen species-generating enzyme NADPH oxidase 4 (Nox4) is up-regulated in the heart after myocardial infarction (MI). Mice with cardiomyocyte-targeted Nox4 overexpression (TG) displayed increased macrophages in the heart at baseline, with skewing toward an M2 phenotype compared with wild-type controls (WT). After MI, TG mice had a higher proportion of M2 macrophages along with higher survival, decreased cardiac remodeling, and better contractile function than wild-type mice. The post-MI increase in cardiac matrix metalloproteinase-2 activity was substantially blunted in TG mice. These results indicate that cardiomyocyte Nox4 modulates macrophage polarization toward an M2 phenotype, resulting in improved post-MI survival and remodeling, likely through the attenuation of cardiac matrix metalloproteinase-2 activity. ("M2") macrophages, respectively (2,5). In the first few days after MI, there is a peak in M1 macrophages, cells that have a high phagocytic activity and produce abundant pro-inflammatory cytokines (2,3). The number of M2 macrophages in the infarct peaks later, and these more heterogeneous cells are believed to tune an anti-inflammatory, reparative response.
Reactive oxygen species (ROS) signaling is involved in the development of many components of the failing heart phenotype, such as cardiomyocyte hypertrophy, cell death, extracellular matrix remodeling, and chamber dilatation (6). NADPH oxidase (Nox) family enzymes are known to be important modulators of redox signaling in heart failure. Unlike ROS sources such as mitochondria or xanthine oxidase, ROS production appears to be the primary function of Nox proteins (7). Of the 5 known Nox family proteins, Nox2 and Nox4 are expressed in the murine heart and have distinct roles in different cardiac pathologies. Nox2 activation contributes to adverse remodeling post-MI, where it promotes cardiomyocyte death, hypertrophy, and extracellular matrix remodeling, as well as being involved in cardiac rupture (6). Nox4 differs from Nox2 in that it is constitutively active, is regulated mainly at the level of abundance, and generates predominantly H 2 O 2 rather than superoxide as its initial product (7,8).
Previous research also showed that Nox4 differs from most other ROS sources by exerting beneficial effects in the heart during chronic pressure overload or starvation, where its expression level increases and enhances specific redox-sensitive signaling (9,10). Nox4 has been linked to anti-inflammatory effects in the vasculature, where it ameliorates inflammation, remodeling, and atherosclerosis (11,12). Myocardial Nox4 levels are increased after acute MI (9,10), but its possible role in modulating inflammation in the heart has not been studied. To quantify initial infarct size after permanent coronary ligation, Evans blue dye (1%) was perfused retrogradely into the aorta 24 h after MI. Hearts were sliced into 5 serial transverse sections and incubated in 1% triphenyltetrazolium chloride to identify infarcted myocardium (13). Infarct area as a percentage of total LV area was quantified by computerized planimetry. Infarct size at 4 weeks' post-MI was determined as the total infarct circumference indexed to total LV circumference (14).     (E) Heart weight/body weight ratio (HW/BW). (F) Cardiomyocyte cross-sectional area quantified in left ventricular sections. n ¼ 5 to 6 per group.
I/R in both groups, but the changes were similar among groups ( Figures 1E and 1F). The total number of leukocytes (CD45 þ ) was significantly and similarly increased after I/R in both TG and WT hearts ( Figure 3A). Although CD11b þ /F4/80 þ macrophages as a percentage of total cells tended to be higher in TG than WT sham (p ¼ 0.057), they increased to a similar level in both groups after I/R ( Figure 3B). The proportion of Ly6C low macrophages relative to total macrophages was similar in the 2 sham groups but was substantially higher in TG than WT after I/R ( Figure 3C). The proportion of MRC1 þ macrophages was significantly higher in TG compared with WT sham, and this were significantly higher in TG myocardium than WT myocardium, both before and after I/R ( Figures 3G and 3H). The proportion of circulating monocytes expressing Ly6C low and MRC1 þ were not significantly different between TG and WT groups either before or after I/R (Supplemental Figure 3).
There was a significant increase in mRNA levels of interleukin (IL)-6 after I/R in WT hearts, which was markedly blunted in TG hearts, whereas levels of IL-10 and tumor necrosis factor-a were not significantly different between groups (Supplemental Figure 4).
We also separately analyzed changes in the infarct region and remote myocardium. Total CD45 þ cells and CD11b þ /F4/80 þ macrophages were significantly higher in the infarct than in the remote region after I/R (Supplemental Figures 5A to 5C Figures 4A and 4C). The proportion of MRC1 þ macrophages also tended to be higher in the remote area of TG compared with WT ( Figures 4B and 4D). The proportion of Ly6C low or MRC1 þ macrophages in the infarct region were similar in WT and TG. The proportion of circulating Ly6C low or MRC1 þ blood monocytes were also similar in WT and TG groups (Supplemental Figure 3).

REMODELING AFTER PERMANENT LAD LIGATION.
TG had a significantly higher survival rate than WT after permanent LAD ligation ( Figure 5A), which was related to a lower rate of cardiac rupture in the first few days after MI. Infarct size quantified 24 h after MI was similar in TG and WT ( Figure 5B). Animals that survived beyond the first week were followed up for 28 days to assess post-MI remodeling. TG mice developed less LV dilatation (lower LV end-diastolic volume) and better preserved ejection fraction than WT after MI (Figures 5C and 5D). Infarct area at  5F). TG had a lower heart/body weight ratio than WT 4 weeks after MI ( Figure 5G) and showed less interstitial fibrosis in the remote myocardium ( Figure 5H, Supplemental Figure 6). Nox2 expression was similar among groups (Supplemental Figure 7).
Taken together, these data indicate that TG are protected against early death and late adverse remodeling after MI. after MI in the WT group, but this increase was substantially attenuated in TG myocardium ( Figures 6A and 6B). MMP-2 mRNA levels were also increased in WT hearts after MI, a response that was blunted in TG ( Figure 6C). MMP-9 was not detected in the cardiac tissue extracts. These results suggest that at least part of the effect of Nox4 on post-MI repair may be linked to changes in MMP activity.

DISCUSSION
In this study, we found that an elevation of cardiomyocyte Nox4 levels alters the baseline "inflam-  In contrast to other Nox proteins such as Nox1 and receptor, CCR2, were also increased in TG myocardium. These proteins are known to be redox sensitive and involved in monocyte recruitment into tissues, and they may therefore be involved in the observed changes in resident macrophage population (19)(20)(21).
Interestingly, previous studies implicated Nox4 in the expression of monocyte chemotactic protein-1 and vascular cell adhesion molecule-1 in endothelial cells (22). Furthermore, recent studies in different mouse models with Nox4 perturbation have suggested a link between Nox4 and inflammatory status (e.g., in atherosclerosis or angiotensin II-induced vascular remodeling) (11,12,23). The present study, however, is the first to examine the details of "inflammatory" changes evoked by Nox4 in the heart and to identify changes in macrophage polarization as a major feature.
Although the changes in monocyte chemo-  (26). NF-kB-dependent changes in polarization toward an M2 phenotype are also described in tumorassociated macrophages (27).
Macrophage polarization is a key aspect of post-MI wound healing and remodeling. M2 macrophages are recruited between days 4 and 7 after MI in the mouse, both in vitro and in atherosclerotic plaques, whereas M2 macrophages produce much lower levels (28). We found that MMP-2 expression and zymographic activity were lower in TG hearts than in WT hearts after ischemia, consistent with the inverted M1/M2 balance found in these animals. It should be noted, however, that other cell types (e.g., cardiomyocytes, fibroblasts, endothelial cells) could also contribute to MMP production. MMP activation is known to be linked to post-MI cardiac rupture and also regulates post-MI remodeling (15,16). Previous studies in mice lacking MMP-2 found that they were protected against post-MI cardiac rupture and adverse remodeling (17).
It is therefore likely that the lower MMP-2 activity in the TG myocardium observed in the present study may have contributed to the higher post-MI survival and better post-MI contractile function and remodeling in these animals.

CONCLUSIONS
In this study, we modeled the effects of a rise in cardiomyocyte Nox4 and its effects on the response to MI. The level of Nox4 overexpression was in a pathophysiologically relevant range (9), and our re-