Blockade of the natriuretic peptide receptor guanylyl cyclase-A inhibits NF-κB activation and alleviates myocardial ischemia/reperfusion injury (original) (raw)
Upregulation of ventricular ANP and BNP concentration and GC-A activity during myocardial ischemia/reperfusion in mice. We measured tissue levels of ANP and BNP in the ventricle after ischemia-reperfusion by RIA. As shown in Table 1, before ischemia/reperfusion the levels of ventricular ANP and BNP were 12.3 ± 2.7 ng/g and 38.9 ± 5.8 ng/g, respectively, which are about 100-fold higher than those of circulating ANP and BNP in mice. In mice treated with 30 minutes of ischemia followed by 6 hours of reperfusion, ventricular ANP and BNP levels were approximately three times higher than those in sham-operated mice. The ventricular BNP level continued to increase until 2 days after reperfusion, while ventricular ANP level tended to decrease slightly. Next, to examine whether ANP and BNP act on in the ventricle, we measured ventricular cGMP concentration during ischemia/reperfusion in mice. In 2 days of reperfusion, ventricular cGMP levels were elevated significantly compared with those in sham-operated mice (13.3 ± 0.9 pmol/g in operated mice and 9.7 ± 0.9 pmol/g in sham-operated mice; P < 0.05), indicating GC-A is activated during ischemia/reperfusion.
Ventricular ANP and BNP concentrations in mice after ischemia/reperfusion
Decreased myocardial infarct size in GC-A–/– mice. The hearts of GC-A+/+ and GC-A–/– mice were subjected to ischemia/reperfusion. Just before the excision, hearts were perfused with Evans blue dye, sectioned, and incubated in 1.5% TTC as described in Methods. Photographs of representative ventricular sections from wild-type and GC-A–/– mice are shown in Figure 1a. Figure 1b illustrates a scheme showing the infarct area, AAR, and nonischemic area to assist in our understanding. The blue-stained tissue was defined as being perfused, the rest of the section was defined as AAR, and regions within the AAR that failed to exhibit brick-red staining were considered to be infarcts. AAR/LV ratios were not significantly different in the two genotypes (GC-A+/+, 64.5 ± 7.2%; GC-A–/–, 63.2 ± 4.2%) (Figure 1c), indicating that there was no genotype-dependent difference in the extent of AAR produced by complete occlusion of the LAD and indicating that ligation site is the same in both groups. On the other hand, infarct/AAR was significantly smaller in GC-A–/– mice after 2 days of reperfusion (GC-A+/+, 58.2 ± 5.2%; GC-A–/–, 44.6 ± 3.7%; P < 0.05) (Figure 1d). Infarct/LV was also significantly smaller in GC-A–/– mice after 2 days of reperfusion (GC-A+/+, 37.5 ± 3.4%; GC-A–/–, 28.2 ± 1.6%; P < 0.05) (Figure 1e), demonstrating that the amount of myocardial tissue salvaged during reperfusion was larger in the absence of endogenous GC-A.
Evaluation of infarct size and AAR in hearts from GC-A+/+ and GC-A–/– mice subjected to 30 minutes of ischemia and 2 days of reperfusion. (a) Representative photographs of midventricular myocardial tissue from GC-A+/+ and GC-A–/– mice. Note that the infarct area (white color) in the GC-A–/– mouse is smaller than that in the GC-A+/+ mouse. (b) Schema of photographs of a. Infarct area is expressed as light blue, AAR is red, and nonischemic area is blue. (c) There was no significant difference in myocardial AAR/LV ratios in GC-A+/+ and GC-A–/– mice. (d) Myocardial infarct/AAR ratios in GC-A–/– mice were significantly smaller than in GC-A+/+ mice. (e) Infarct/LV ratios in GC-A–/– mice were significantly smaller than in GC-A+/+ mice. *P < 0.05 vs. GC-A+/+ mice.
Decreased PMN infiltration number in GC-A–/– mice. One of the hallmarks of reperfusion injury is the accumulation of PMNs within the injured tissue (34). In fact, interventions aimed at inhibiting PMN accumulation have been shown to exert cardioprotective effects in various experimental models. And in the present study, comparison of Figure 2, b and d, with Figure 2, a and c, illustrates that the number of PMNs infiltrating the myocardium was indeed diminished in GC-A–/– mice. Summarizing the counts from at least ten randomly selected microscope fields, we found that significantly fewer PMNs were present in the myocardium of GC-A–/– mice after both 6 hours (GC-A+/+, 16 ± 6 PMNs per high-powered field; GC-A–/–, 8 ± 3 PMNs per high-powered field; P < 0.01) and 2 days (GC-A+/+, 42 ± 3 PMNs per high-powered field; GC-A–/–, 17 ± 4 PMNs per high-powered field; P < 0.01) reperfusion (Figure 2e).
PMN infiltration in GC-A+/+ and GC-A–/– mice subjected to myocardial ischemia/reperfusion. (a–d) H&E staining of myocardial tissue from GC-A+/+ and GC-A–/– mice obtained after 30 minutes of ischemia and 2 days of reperfusion. (a and c) Perinecrotic area of a GC-A+/+ mouse at magnification of ×100 and ×400, respectively. (b and d) Perinecrotic area of a GC-A–/– mouse at magnifications of ×100 and ×400, respectively. Note that significant numbers of PMNs accumulated in perinecrotic areas of GC-A+/+ mice, while a fewer PMNs infiltrated the GC-A–/– heart. (e) Average numbers of infiltrating PMNs per ×400 field in the myocardium subjected to 30 minutes of ischemia and either 6 hours or 2 days of reperfusion. After both 6 hours or 2 days of reperfusion, PMN numbers were significantly lower in GC-A–/– mice (filled bars) than in time-matched GC-A+/+ controls (open bars). *P < 0.01 vs. 6 hours of reperfusion in GC-A+/+ mice. #P < 0.01 vs. 2 days of reperfusion in GC-A+/+ mice. (f) Myocardial MPO activity in infarct and noninfarct cardiac tissue samples obtained from GC-A+/+ (open bars) and GC-A–/– mice (filled bars) after 30 minutes of ischemia and 6 hours or 2 days of reperfusion. MPO activity is expressed as U/100 mg wet tissue weight, which was decreased significantly in the infarct areas of GC-A–/– mice. *P < 0.05 vs. 6 hours of reperfusion in GC-A+/+ mice. #P < 0.01 vs. 2 days of reperfusion in GC-A+/+ mice.
Decreased MPO activity in GC-A–/– mice. To further quantify the extent of PMN infiltration, we measured the activity of cardiac MPO, an enzyme specific to activated neutrophils. This method is commonly used to assess neutrophil accumulation and has been validated previously in cardiac tissue (27). When myocardial MPO activity was measured in infarcted and noninfarcted tissues after 6 hours or 2 days of reperfusion, we found that the activity time dependently increased within the infarcted areas of both GC-A–/– and wild-type mice (GC-A+/+, 0.46 ± 0.08 U/100 mg tissue; GC-A–/–, 0.31 ± 0.06 U/100 mg tissue after 6 hours of reperfusion; GC-A+/+, 1.68 ± 0.03 U/100 mg tissue; GC-A–/–, 1.17 ± 0.03 U/100 mg tissue after 2 days of reperfusion) but that it was significantly lower in the GC-A–deficient mice at either time point (Figure 2f). Within noninfarct areas, MPO activity was very low in both genotypes.
Decreased P-selectin expression in GC-A–/– mice. Endothelial cell surface expression of P-selectin (CD62), an integral membrane protein rapidly translocated to the plasma membrane during exocytosis of the Weibel-Palade bodies, promotes the binding and “rolling” of neutrophils, which precedes their migration into sites of inflammation (35–41). To investigate the mechanism responsible for the reduced PMN infiltration in GC-A–deficient mice, we examined P-selectin expression after ischemia/reperfusion. Immunohistochemical analysis revealed that after 2 days of reperfusion, P-selectin was expressed in coronary microvascular endothelial cells in perinecrotic areas of both wild-type (Figure 3, a, c, and e) and GC-A–/– (Figure 3, b, d, and f) mice, although the latter were labeled to a lesser degree. In contrast, P-selectin expression was barely detectable in sham-operated hearts. Western blot analysis, moreover, confirmed the significant (P < 0.01) declines in P-selectin expression in GC-A–/– mice after either 6 hours of reperfusion (GC-A+/+, 3.11 ± 0.26 arbitrary units; GC-A–/–, 1.83 ± 0.08 arbitrary units; P < 0.01) or 2 days of reperfusion (GC-A+/+, 5.65 ± 0.40 arbitrary units; GC-A–/–, 3.64 ± 0.53 arbitrary units; P < 0.01) (Figures 3, g and h). Thus, the reduced PMN accumulation during ischemia/reperfusion in GC-A–deficient mice is likely due, at least in part, to suppressed expression of P-selectin in coronary endothelial cells.
P-selectin expression in GC-A+/+ and GC-A–/– mice during myocardial ischemia/reperfusion. (a–f) Immunolabeling of P-selectin in hearts from GC-A+/+ (a, c, and e) and GC-A–/– mice (b, d, and f); N, noninfarct tissue; I, infarct tissue. (a and b) Border zones between necrotic and nonischemic areas. (c and d) Perinecrotic areas. (e and f) High-power fields of perinecrotic areas. Levels of endothelial P-selectin expression in the border zones and in the centers of ischemic areas were lower in GC-A–/– hearts. (g) Western blot analysis showing P-selectin expression in LV of GC-A+/+ (lanes 1, 3, and 5) and GC-A–/– (lanes 2, 4, and 6) mice after 30 minutes of ischemia and 6 hours (lanes 3 and 4) or 2 days (lanes 5 and 6) of reperfusion. Lanes 1 and 2 show P-selectin expression without ischemia/reperfusion. (h) Semiquantitative analysis of P-selectin expression. P-selectin expression was significantly lower in GC-A–/– mice (filled bars) than in GC-A+/+ mice (open bars) after 6 hours or 2 days of reperfusion. *P < 0.01 vs. 6 hours of reperfusion in GC-A+/+ mice. #P < 0.01 vs. 2 days of reperfusion in GC-A+/+ mice.
Reduced NF-κB activity in GC-A–/– mice. NF-κB is a transcription factor that, by regulating the gene expression of multiple cytokines, chemokines, cell adhesion molecules, growth factors, and immunoreceptors, plays a critical role in immune and inflammatory reactions (42, 43).
The P-selectin gene, which contains a functional NF-κB–binding site, is required for its induction in endothelial cells (44). We therefore assessed NF-κB activation using EMSA with a NF-κB binding site–specific probe and nuclear proteins isolated from myocardial tissues of GC-A–/– and GC-A+/+ mice. No activation of NF-κB was detected in sham-operated mice of either genotype (Figure 4b, lanes 1 and 2, respectively). In contrast, incubation with nuclear extracts from wild-type mice after 6 hours of reperfusion yielded a clearly detectable band (Figure 4b, lane 3), the identity of which was confirmed by competition and supershift analyses with a specific Ab against the NF-κB p50 subunit (Figure 4a). NF-κB activity was increased after 6 hours of reperfusion in wild-type mice (Figure 4b, lane 3) and maintained until 2 days after reperfusion (Figure 4b, lane 5). In contrast, NF-κB activation was not detectable in GC-A–/– mice after 6 hours of reperfusion (Figure 4b, lane 4) but elevated to the same level as GC-A+/+ mice after 2 days of reperfusion (Figure 4b, lane 6). Densitometric scanning of the shifted bands revealed a 70% decline in NF-κB activity in GC-A–/– after 6 hours of reperfusion (GC-A+/+, 16.52 ± 1.87 arbitrary units; GC-A–/–, 4.86 ± 1.56 arbitrary units) (Figure 4c), which implies that GC-A might play a distinctive role in modulating ischemia/reperfusion injuries through inhibition of NF-κB activation. IκBα was phosphorylated in GC-A+/+ mice, whereas it is not phosphorylated in GC-A–/– mice after 6 hours of reperfusion. Phosphorylation of IκBα in GC-A+/+ mice and GC-A–/– mice after 2 days of reperfusion were upregulated to the same level. However, there was no significant change in IκBα protein level in GC-A–/– mice and GC-A+/+ mice after both 6 hours and 2 days of reperfusion (Figure 4, d–g).
NF-κB binding to DNA during myocardial ischemia/reperfusion in GC-A+/+ and GC-A–/– mice. (a) The identity of the band obtained from a NF-κB gel shift assay was confirmed by competition and supershift analyses using a specific anti-p50 subunit Ab as a probe. An arrow in left side shows shifted band. An arrow with broken line shows supershifted band. (b) EMSA of NF-κB in GC-A+/+ (lanes 1, 3, and 5) and GC-A–/– (lanes 2, 4, and 6) hearts; lanes 1 and 2, sham operated; lanes 3 and 4, 30 minutes of ischemia and 6 hours of reperfusion; lanes 5 and 6, 30 minutes of ischemia and 2 days of reperfusion. (c) Semiquantitative analysis of the binding of activated NF-κB to DNA in GC-A+/+ (open bars) and GC-A–/– mice (filled bars). (d and e) Western blot analysis showing phosphorylated IκBα (d) and total IκBα protein expression (e) in left ventricles of GC-A+/+ (lanes 1, 3, and 5) and GC-A–/– (lanes 2, 4, and 6) mice after 30 minutes of ischemia and 6 hours (lanes 3 and 4) or 2 days (lanes 5 and 6) of reperfusion. Lanes 1 and 2 show phosphorylated IκBα (d) and total IκBα protein expression (e) without ischemia/reperfusion. (f and g) Semiquantitative analysis of phosphorylated IκBα (f) and total IκBα (g) protein expression. Phosphorylated IκBα was significantly lower in GC-A–/– mice (filled bars) than in GC-A+/+ mice (open bars) after 6 hours of reperfusion (f). *P < 0.01 vs. 6 hours of reperfusion in GC-A+/+ mice.
Decreased myocardial infarct size, PMN infiltration, and P-selectin expression in mice treated with HS. To rule out that these effects were restricted within GC-A–/– mice, we next examined the effects of HS. Blockade of GC-A by intravenous injection of HS (3 mg/kg) was initially confirmed by the evoked decline in basal plasma cGMP levels (11.5 ± 1.7 pmol/ml in saline vs. 7.4 ± 0.6 pmol/ml in HS; P < 0.05). Whereas AAR/LV in both groups did not change (saline, 57.9 ± 3.0%; HS, 57.3 ± 7.2%) (Figure 5a) after 2 days of reperfusion, injection of HS 30 minutes before coronary occlusion significantly reduced infarct/AAR ratio (saline, 56.5 ± 2.3%; HS, 38.7 ± 3.3%; P < 0.05) (Figure 5b) and infarct/LV ratio (saline, 32.7 ± 2.1%; HS, 22.2 ± 2.5%; P < 0.05) (Figure 5c), decreased the number of infiltrating PMNs after 2 days of reperfusion (saline, 17 ± 7 PMNs per high-powered field;, HS, 4 ± 1 PMNs per high-powered field; P < 0.01) (Figure 5d), and decreased the MPO activity after 6 hours of reperfusion (saline, 0.42 ± 0.16 U/100 mg tissue; HS, 0.12 ± 0.08 U/100 mg tissue; P < 0.01) (Figure 5e). Western blot analysis revealed that HS reduced the level of cardiac P-selectin expression seen after 6 hours of reperfusion (saline, 39.6 ± 10.4 arbitrary unit; HS, 8.9 ± 2.8 arbitrary unit; P < 0.01) (Figure 5, f and g), and there was a corresponding decline in the immunolabeling of P-selectin in coronary endothelial cells (data not shown). These findings further support the notion that blockade of endogenous natriuretic peptides decreases myocardial PMN infiltration after ischemia/reperfusion by inhibiting expression of P-selectin on endothelial cells.
Myocardial infarct size, PMN infiltration, MPO activity, and P-selectin expression in HS-treated mice after 30 minutes of ischemia and 6 hours of reperfusion. (a) AAR/LV was not significantly different in the two groups. (b) Infarct/AAR was significantly smaller in HS-treated mice (filled bars) than in saline-treated mice (saline; open bars) after 2 days of reperfusion. *P < 0.05 vs. saline group. (c) Infarct/LV was significantly decreased in HS-treated mice compared with that in saline-treated mice. *P < 0.05 vs. saline group. Numbers of infiltrating PMNs per field (d) and MPO activity (e) were both significantly lower in HS-treated mice. **P < 0.01 vs. saline group. (f) Western blot analysis of P-selectin expression following ischemia/reperfusion in mice pretreated with saline (lane 3) or HS (lane 4); lanes 1 and 2, P-selectin in sham-operated mice treated with saline and HS, respectively. (g) Semiquantitative analysis of P-selectin expression in HS-treated mice. Little P-selectin expression was seen in sham-operated mice. P-selectin expression was increased in all mice subjected to ischemia/reperfusion, although significantly (P < 0.01) less so in HS-treated mice (filled bars) than in saline-treated mice (open bars). **P < 0.01 vs. saline group.
Induction of P-selectin expression in HUVECs by ANP. The aforementioned findings led us to hypothesize that ANP directly influences endothelial P-selectin expression during ischemia/reperfusion. To test this hypothesis, we assessed the effect of ANP on P-selectin expression in cultured HUVECs in the presence or absence of H2O2, a toxic by-product of respiration. Although we found only minimal immunolabeling of P-selectin (stained red) under basal conditions (Figure 6a), several cells were labeled in the presence of 100 μM H2O2 (Figure 6b), indicating that exposure to H2O2 induced expression of P-selectin. When incubated with H2O2 plus 10–6 M ANP, more cells were labeled and to a stronger degree than with H2O2 alone (Figure 6c), while exposure to H2O2, ANP, and 100 μg/ml HS resulted in cells being labeled to the same degree seen with H2O2 alone (Figure 6d).
(a–d) Immunolabeling of P-selectin in HUVECs. (a) Unstimulated cells. (b) Cells treated with H2O2 (100 μM). (c) Cells treated with ANP (1 μM) in the presence of H2O2 (100 μM). (d) Cells treated with ANP (1 μM) in the presence of HS (100 μg/ml) and H2O2 (100 μM). (e and f) Cell surface (e) and total (f) expression of P-selectin was increased by exposure to 100 μM H2O2 and was increased further by exposure to H2O2 plus ANP (P < 0.05 vs. H2O2 alone). The effect of ANP was abolished by HS. *P < 0.01 vs. saline-treated group. #P < 0.01 vs. H2O2-treated group. P < 0.01 vs. H2O2- and ANP-treated group.
Using a specific ELISA, we examined translocation of P-selectin to cell surface and total P-selectin expression of HUVECs. Treating cells for 1 hour with ANP alone dose dependently increased cell surface expression of P-selectin (saline, 1 ± 0.13; ANP 10–8 M, 2.34 ± 0.41; ANP 10–6 M, 3.64 ± 0.23 arbitrary units) (data not shown), and 100 μM H2O2 alone for 1 hour increased cell surface P-selectin expression, as well. Cell surface expression of P-selectin in the presence of ANP was further increased by addition of 100 μM H2O2, which was suppressed by pretreatment with 100 μg/ml HS (Figure 6e). Incubating cells with ANP for 6 hours in the presence of 100 μM H2O2 increased total expression of P-selectin, effects of which again were abolished by prior addition of HS (Figure 6f).
Finally, cell viability assays, in which mitochondrial respiration was assessed as a function of the mitochondria-dependent reduction of WST-8 to formazan, revealed that H2O2, ANP, and HS, either alone or in combination, had no effect on endothelial cell viability (data not shown). Thus, ANP appears to directly increase both cell surface and total expression of P-selectin protein in endothelial cells under oxidative stress.
NF-κB activation by ANP and/or H2O2 in HUVECs. We showed that NF-κB was activated in an ischemia/reperfusion model of wild-type mice, while it was not activated in GC-A–/– mice. Next, we examined NF-κB activation in HUVECs, which were treated with ANP in the presence of H2O2. As shown in Figure 7, a and b, 10–6 M ANP alone did not activate NF-κB in HUVECs. Modest activation of NF-κB was observed in H2O2-treated HUVECs in a good agreement with the previous report (45). Surprisingly, ANP greatly augmented H2O2-induced NF-κB activation in HUVECs. These effects were abolished with HS (Figure 7, a and b). These results strongly suggest that, under oxidative condition as in ischemia/reperfusion, ANP could activate NF-κB in vascular endothelial cells. To further examine the mechanism for NF-κB activation, we performed Western blot analysis using specific Ab for phosphorylated IκBα and Ab for total IκB. As shown in Figure 7, c and e, 10–6 M ANP alone did not phosphorylate IκBα, but H2O2 phosphorylated it. When ANP was added in combination with H2O2, IκBα was phosphorylated more compared with that by H2O2 alone. However, incubation of cells with ANP in the presence of H2O2 did not affect degradation of IκBα (Figure 7, d and f).
NF-κB activation, IκBα phosphorylation, and IκBα degradation in ANP- and/or H2O2-treated HUVECs. (a) EMSA of NF-κB in control cells (lane 1), ANP-treated cells (lane 2), H2O2-treated cells (lane 3), ANP with H2O2-treated cells (lane 4), ANP and HS with H2O2-treated cells (lane 5). An arrow in right side indicates shifted band. (b) Semiquantitative analysis of the binding of activated NF-κB to DNA in HUVECs. Level of NF-κB activation was significantly higher in H2O2-treated cells (H)compared with control cells (N) and was further increased in ANP with H2O2-treated cells (H+A) compared with H2O2-treated cells. This effect was abolished by treatment with HS (H+A+HS). *P < 0.01 vs. control cells. #P < 0.01 vs. H2O2-treated cells. Western blot analysis of IκBα phosphorylation (c) and total IκBα protein expression (d) in HUVECs. Lane 1, control cells; lane 2, ANP-treated cells; lane 3, H2O2-treated cells; lane 4, H2O2 with ANP-treated cells; lane 5, H2O2 with ANP- and HS-treated cells. IκBα was not phosphorylated by ANP alone, but by H2O2, and IκBα phosphorylation was further strengthened by ANP with H2O2. This effect was abolished by HS (c). IκBα protein expression did not change among all groups (d). Densitometric scanning of IκBα phosphorylation (e) and IκBα protein expression (f). Phosphorylation of IκBα in H2O2-treated cells is significantly increased compared with control cells. It further increased significantly in H2O2 with ANP-treated cells compared with H2O2-treated cells (e). There is no significant change in IκBα protein expression among all groups (f). *P < 0.05 vs. control cells. #P < 0.01 vs. H2O2-treated cells.
Incubation with ANP did not affect reactive oxygen species production in HUVECs. To investigate if ANP directly produces reactive oxygen species, we examined intracellular peroxides in HUVECs treated or not treated with ANP. The flow cytometric data shows the occurrence of oxidative stress in HUVECs exposed to hydrogen peroxide. In contrast, exposure of cells to 10–6 M ANP did not affect carboxy-H2DCFDA oxidation, indicating that ANP does not directly induce oxidative stress in endothelial cells (data not shown).