Endothelium-derived neuregulin protects the heart against ischemic injury

Abstract

Background: Removal of cardiac endothelial cells (EC) has been shown to produce significant detrimental effects on the function of adjacent cardiac myocytes, suggesting that EC play a critical role in autocrine/paracrine regulation of the heart. Despite this important observation, the mediators of the protective function of EC remain obscure. Neuregulin (NRG, a member of the epidermal growth factor family) is produced by EC and cardiac myocytes contain receptors (erbB) for this ligand. We hypothesized that NRG is an essential factor produced by EC, which promotes cardioprotection against ischemic injury.

Methods and results: We demonstrate that human cardiac EC express and release NRG in response to hypoxia-reoxygenation. Under conditions where hypoxia--reoxygenation causes significant cardiac myocyte cell death, NRG can significantly decrease apoptosis of isolated adult ventricular myocytes. Coculturing adult murine myocytes with human umbilical vein, murine lung microvascular, or human coronary artery EC can also protect myocytes against hypoxia--reoxygenation--induced apoptosis. These protective effects are abolished by NRG gene deletion or silencing of NRG expression in EC. Finally, endothelium-selective deletion of NRG in vivo leads to significantly decreased tolerance to ischemic insult, as demonstrated by impaired postischemic contractile recovery in a perfused whole-organ preparation and larger infarct sizes after coronary artery ligation.

Conclusion: Together, these data demonstrate that EC-derived NRG plays an important role in cardiac myocyte protection against ischemic injury in the heart and supports the idea that manipulation of this signaling pathway may be an important clinical target in this setting.

Figure 1

Human cardiac endothelial cells express neuregulin (NRG) and release it in response to hypoxia. (A) Human cardiac autopsy specimens stained by immunofluorescence for lectin (A & D, as a marker of endothelium), NRG (B & E) or both (merged red and green channels, C & F). A representative section is shown with staining in both endocardial (A-C) and vascular (D-F) endothelium (arrowheads). (G) Western blots of cell-associated (“full-length”, top panel) NRG in lysates from isolated human coronary artery endothelial cells or NRG immunoprecipitated from cell culture media (“secreted NRG”, bottom panel) before and after hypoxia/reoxygenation. Representative blots from N=4 experiments are shown.

Figure 2

Effect of neuregulin (NRG) on hypoxia/reoxygenation (H/R)-induced apoptosis in adult murine cardiomyocytes. (A) Representative field of myocytes stained with DAPI or TUNEL after H/R (4h/18h). Bottom panel shows merged image with arrows denoting both DAPI and TUNEL positive nuclei. (B) Representative field of myocytes treated with NRGβ3 and subjected to H/R. (C) Quantification of apoptosis (TUNEL or annexin V staining) and necrosis (propidium iodide, PI) in response to H/R in the absence or presence of NRGβ. (Data are mean ± SD, N=5. *P<0.01 vs. Control (no H/R); †P<0.01 vs. H/R.)

Figure 3

Effect of endothelial cell (EC) co-culture on hypoxia/reoxygenation(H/R)-induced cell death in adult murine cardiomyocytes. (A) Representative field of myocytes after H/R (same field is shown in all 4 panels). (B) Representative field of myocytes, with co-cultured human coronary artery EC (HCAEC) present during H/R (same field is shown in all 4 panels). (C) Quantitation of apoptosis/necrosis in response to H/R in the absence or presence of human umbilical vein EC (HUVEC), murine lung microvascular EC (LMVEC) or HCAEC. (Data are mean ± SD. *P<0.05 vs. Control (no H/R); †P<0.05 vs. H/R for same parameter, N=5/group.)

Figure 4

Deletion of neuregulin (NRG) expression results in loss of cardiac myocyte protection against H/R. (A) Lung microvascular endothelial cells (LMVEC) from NRG homozygously floxed mice treated with adenovirus expressing cre recombinase results in loss of cardiac myocyte protection in co-culture. Adenovirus expressing an irrelevant control protein (DS-Red) had no effect on LMVEC protection of cardiac myocytes. (Inset panel: immunoblot demonstrating dose-dependent reduction in NRG expression (full-length NRG, top) after AdenoCre infection (cre expression, bottom), MOI: multiplicity of infection.) (*P<0.05 vs. TUNEL in normoxic control or H/R treated myocytes co-cultured with AdenoRed treated EC (H/R + EC/AdenoRed); †P<0.001 vs. annexin in normoxic control, LMVEC co-cultured, or AdenoRed treated EC co-cultured myocytes.) (B) Human coronary artery endothelial cells (HCAEC) treated with siRNA directed against NRG (but not scrambled control siRNA) fail to protect cardiac myocytes. (Inset panel: immunoblotting of HCAEC lysates or media (40kD, cleaved product detected by immunoprecipitation) from H/R-exposed HCAEC after siRNA transfection.) (*P<0.01 vs. normoxic control, control siRNA-treated co-cultured or non-treated HCAEC co-cultured myocytes; †P<0.05 vs. control siRNA-treated or non-treated HCAEC co-cultured myocytes; #P<0.001 vs. normoxic control, control siRNA-treated or non-treated HCAEC co-cultured myocytes.)

Figure 5

Endothelial deletion of neuregulin (NRG) impairs post-ischemic recovery in Langendorff perfused hearts. A. Western blot of pooled (3-4 hearts/sample) isolated, purified cardiac endothelial cells from either vehicle or tamoxifen treated mice demonstrating a significant reduction of NRG expression after tamoxifen induction. A reprobe of the same blot using Actin as a loading control is shown in lower panel. B. Left ventricular developed pressures (LVDP) in perfused hearts from vehicle or tamoxifen treated mice. Baseline and ischemic LVDP were not significantly different between groups. Reperfusion (post-ischemic, taken at end of 30 min reperfusion) LVDP was significantly impaired in CFF NRG knockout animals (CFF tamoxifen) (*P<0.05 vs. all other groups, N=5/group). C. Rate-pressure products (RPP) in the same perfused hearts. Reperfusion RPP was also significantly impaired in the NRG knockout animals (*P<0.05 vs. all other groups).

Figure 6

Endothelial deletion of neuregulin (NRG) increases infarct size and blunts erbB4 activation. Mice were pretreated with tamoxifen or vehicle (N=4-5/group) then subjected to left coronary artery ligation. A. Area at risk (AAR, % total ventricle) was not significantly different between the groups; however, the infarct sizes (expressed as % total ventricle or % AAR) were significantly larger in the NRG deleted (CFF tamoxifen) animals compared to tamoxifen treated (C++ tamoxifen) or vehicle treated (CFF vehicle) controls. Administration of exogenous recombinant NRG (rNRG) rescued the defect, leading to smaller infarct sizes in NRG deleted animals (CFF tam + rNRG) (*P<0.05, vs. C++ tamoxifen and CFF tam + rNRG.) B. Immunoprecipitation and Western blotting for phosphorylated erbB4 or total erbB4 in left ventricular lysates from non-infarcted controls or infarcted animals (*P<0.05 vs. CFF tamoxifen infarct)