When selectively overexpressed in mouse heart, TNF-α effects the development of a cardiomyopathy that closely mimics that seen in human failing hearts. It has been suggested that two intracellular signaling pathways, the Akt protein kinase and the NF-κB transcription factor, mediated TNF-α signaling. The present experiments assessed the effects of TNF-α overexpression on these two target proteins in vivo. We measured cardiac Akt kinase phosphorylation and NF-κB activity in mice overexpressing TNF-α (TNF1.6). Both basal and insulin-stimulated Akt phosphorylation were reduced by almost 70% by TNF-α overexpression. By contrast, NF-κB was robustly activated. These effects were absent when TNF-α receptor 1 (TNFR1) was selectively ablated. Cardiomyocyte-specific overexpression of the dominant-negative inhibitory κB protein transgene and subsequent inhibition of NF-κB activity attenuated the effects of TNF-α on Akt phosphorylation. NF-κB inhibition also significantly improved fractional shortening and diminished ventricular hypertrophy and survival without affecting infiltrative inflammation or cytokine expression. Thus, while overexpression of TNF-α effected a marked Akt inhibition and NF-κB activation in mouse hearts, inhibition of NF-κB offered salutary benefits mediated at least in part through activation of Akt.
- tumor necrosis factor-α
- insulin signaling
- TNF-α receptor
- protein kinase B
- dominant-negative inhibitory κB protein transgene
tumor necrosis factor-α (TNF-α) is a proinflammatory cytokine that has been implicated in the pathogenesis of heart failure (14). Whereas the normal heart does not express TNF-α, the failing heart produces robust quantities that can be found both in the heart and in the peripheral circulation of patients with heart failure. When selectively overexpressed in the hearts of mice, TNF-α effects the development of a dilated cardiomyopathy that closely mimics the pathophysiology seen in human failing hearts (21, 24). The cardiotoxic effects of TNF-α appear to be mediated through the TNF receptor-1 (TNFR1) as ablation of the TNFR1 receptor attenuated the development of heart failure in TNF-α-over expressing mice. By contrast, ablation of the TNF receptor-2 (TNFR2) actually increased mortality in mice overexpressing TNF-α, suggesting that pathways downstream of TNFR2 might be cardioprotective (18). While the pathophysiological effects of TNF-α overexpression have been well-described, far less is known about the downstream signaling pathways that mediate its biological effects.
One pathway that appears to play a role in TNF-α-mediated signaling is the ubiquitous transcription factor nuclear factor-κB (NF-κB). NF-κB activates the expression of a group of genes, including those that promote immune and inflammatory responses and cell survival, as well as apoptosis, in response to a wide array of extracellular signals (3). That NF-κB plays a role in TNF-α-mediated cardiotoxicity was demonstrated by the finding that NF-κB activity was substantially increased in hearts from TNF-α transgenic mice (25). Furthermore, genomic disruption of the p50 subunit of NF-κB reduced ventricular dilatation and hypertrophy and improved cardiac function in mice overexpressing TNF-α (22). NF-κB is regulated by association with an inhibitory κB protein (IκB) that acts to inhibit NF-κB function (2). Signals including TNF-α inactivate IκB by phosphorylating either serine or tyrosine residues, resulting in IκB degradation or dissociation from NF-κB, respectively. Dissociation of IκB from NF-κB facilitates nuclear localization of NF-κB and its regulation of gene transcription through binding to specific NF-κB sites within the promoter regions of selected genes.
TNF-α has also been shown to stimulate phosphorylation/activation of Akt (also known as protein kinase B), a serine/threonine protein kinase, in neonatal cardiac myocytes (12). Activation of Akt has been associated with increased protein accretion in hypertrophy, and the hypertrophic response of cardiac myocytes to mechanical strain (29). In contrast to the effects of NF-κB, activation of Akt promotes cell survival through inhibition of apoptosis and thus has been proposed as a potential therapeutic target in the treatment of congestive heart failure. However, Akt has also been shown to activate NF-κB through IκB phosphorylation or by more direct effects (17). Thus, while TNF-α-mediated activation of Akt would appear to have adaptive effects on the heart, the subsequent activation of NF-κB by Akt could have deleterious effects. Therefore, in the present experiments, we sought to determine the roles played by both Akt and NF-κB in the development of TNF-α-mediated dilated cardiomyopathy by assessing the effects of TNF-α overexpression on Akt activation in the presence and absence of functional NF-κB.
To disrupt NF-κB activity in previously characterized TNF-α transgenic mice, TNF1.6 (24), we crossed TNF1.6 with IκBαS32A,S36A mice (13). In both lines, overexpression of the transgene (murine TNF-α or the dominant negative form of IκB) are driven by the α-myosin heavy chain (α-MHC) promoter, which restricts expression to cardiac myocytes. TNF1.6 mice were identified by PCR with a sense primer (5′-cca cat tct tca gga ttc tct-3′) specific to the α-MHC promoter exon 2 and an antisense primer (5′-cag cct tgt ccc ttg aag aga-3′) specific to the TNF-α cDNA nucleotides 597 to 599. IκBαS32A,S36A mice were genotyped using an α-MHC-specific primer (5′-aag cct agc cca cac cag aaa tga cag aca-3′) and IκBα-specific primer (5′-agt agc cgc tcc ttc ttc agc ccg tc-3′). All mice were of the same FVB genetic background. TNF1.6 mice with disrupted TNFR1 or TNFR2 genes were generated as mixed genetic background (1:1 ratio of C57BL/6 and FVB) as previously described (18). Akt1 and Akt2 knockout mice in C57B6 background were kind gifts of M. J. Birnbaum (9, 10). Swimming protocols were performed as described previously (32). Briefly, 9-wk-old male mice were subjected to swimming twice daily, 1.5 h in a 30°C water tub. The Institutional Animal Care and Use Committee of Thomas Jefferson University approved the animal use protocol. Freshly isolated mouse heart tissues were either fixed in 4% neutral buffered paraformaldehyde or snap-frozen in liquid nitrogen for later analysis.
Echocardiography was performed to compare cardiac structure and function. M-mode echocardiographic analyses were performed with the use of a 14-MHz transducer (Sequoia C256, Acuson) at baseline and after intraperitoneal injection of the β-adrenergic agonist isoproterenol (30 ng/g body wt) as previously described (21). The percentage of left ventricular fractional shortening (LVFS) was calculated as LVFS (%) = [(LVDD − LVSD)/LVDD] × 100, where LVDD and LVSD indicate left ventricular end-diastolic and end-systolic diameters, respectively.
Enzyme-linked immunosorbent assay.
The protein levels were assessed using kits for mouse TNF-α and IL-1β (Quantikine, R&D Systems) according to manufacturer's instructions. Results were expressed as picograms of target proteins per gram of protein.
Immunofluorescent staining and quantitation of myocardial infiltrates.
After measurement of body and heart weight, hearts were fixed in 2% neutral buffered paraformaldehyde for 90 min at 4°C, and then suspended in 30% sucrose solution overnight at 4°C. Six-micrometer-thick sections were prepared from frozen, fixed hearts and probed with antibodies as previously described (25). Primary antibodies included rat anti-mouse CD45 1:50 (Pharmingen 553076) or rat anti-mouse CD31 [platelet endothelial cell adhesion molecule (PECAM)] 1:50 (Pharmingen 550274). The secondary antibody was goat anti-rat Cy3 (Jackson ImmunoResearch Laboratories). Sections were also labeled with Alexa Fluor 488 phalloidin 1:250 (to identify cells with filamentous actin, predominantly cardiac myocytes) and nuclei labeled with Hoescht 33342 (Sigma). Slides were imaged at ×40 magnification on a confocal microscope (Olympus FluoView 1000).
For hematoxylin-eosin staining and assessment of the degree of cell infiltration, hearts were fixed in 4% neutral buffered paraformaldehyde. To quantify myocardial infiltrates, nuclear density (nuclei/mm2) was determined as described previously (24). In each animal, five independent high-power fields (0.250 × 0.300 mm; 0.075 mm2 area) were analyzed using Image-Pro Plus Software and averaged by investigators blinded to the groups.
NF-κB binding assay.
NF-κB activity in mouse ventricular tissues was performed as previously described (13). Briefly, frozen ventricular tissue was pulverized at liquid nitrogen temperatures and homogenized. The clarified crude nuclear extracts were prepared in buffer with protease and phosphatase inhibitors [20 mM HEPES, pH 7.9; 25% (vol/vol) glycerol; 0.6 mM MgCl2; 0.2 mM EDTA; 0.1 mM PMSF; 0.5 mM DTT; 25 μg/ml leupeptin; 0.2 mM Na orthovanadate]. For electrophoretic mobility shift assay, [32P]dATP-labeled double-stranded 22-bp oligodeoxynucleotide (5′-AGTTGAGGGGACTTTCCCAGGC-3′) containing a consensus NF-κB binding site was incubated with 10 μg of nuclear protein in 10 mM Tris·HCl (pH 7.5), 50 mM NaCl, 1 mM MgCl2, 0.5 mM EDTA, O.5 mM DDT, 4% glycerol (vol/vol), and 1 mg poly(dI-dC). The DNA-protein complexes were separated on 6% nondenaturing polyacrylamide gels in 1× Tris borate-EDTA buffer. Dried gels were exposed to X-ray film at −20°C using intensifying screens. NF-κB activity was determined in at least three mice of each genotype and repeated at least twice.
Kinase assay and immunoblotting.
Akt kinase assay and Akt phosphorylation (Ser473 and Thr308) were measured as described previously (8) with minor modifications for heart tissues. Initial experiments determined that no differences existed between protein extracted from the whole heart (i.e., containing both left and right ventricles) vs. protein extracted from the left ventricular myocardium. Unless stated otherwise, all protein extracts were obtained from the left ventricle. Briefly, frozen tissues were homogenized on ice using a NP40 lysis buffer (25 mM Tris·HCl, pH 7.6, 137 mM NaCl, 10% glycerol, 1% NP40, 10 mM NaF) freshly supplemented with 1 mM sodium pyrophosphate, 5 μg/ml leupeptin, 5 μg/ml aprotinin, 1 mM EDTA, 10 mM PMSF, 1 mM NaVO4, and 1 mM DTT. Clarified cellular lysates were boiled and separated by electrophoresis in a 4–12% SDS-PAGE and transferred onto nitrocellulose membranes. For in vitro kinase assay, cardiac Akt was immunoprecipitated with anti-Akt antibody and protein G beads. Specific Akt kinase activity was determined by measuring in vitro phosphorylation of the recombinant Akt substrate GSK3β. Phosphorylated GSK3β was detected by rabbit polyclonal anti-phospho-GSK3 (ser21/9) antibody according to manufacturer's instructions (Cell Signaling Tech).
For immunoblotting, membranes were blocked for 30 min with Li-Cor blocking buffer and probed with antibodies at 4°C overnight. The blots were subsequently incubated with either IRDye 700 or 800 secondary antibodies conjugated with infrared fluorophores for 60 min. Bands were visualized and directly quantified using the Odyssey Infrared Imaging System (Li-Cor, Lincoln, NE). Direct quantifying labeled secondary antibodies produces a wide linear range. The following antibodies were used at 1:1,000 dilution: anti-phospho-GSK3 (ser21/9) antibody, anti-total Akt rabbit polyclonal, anti-phospho-Akt (Thr308) and anti-phospho-Akt (Ser473) rabbit polyclonal (Cell Signaling Tech), and anti-actin monoclonal (Sigma).
Results are presented as means ± SD. Analysis was performed using SPSS for Windows (version 11.5). Kaplan-Meier survival curves were compared between groups using log-rank tests. Data were analyzed by a one-way ANOVA using Bonferroni's multiple-comparison test. A value of P < 0.05 was considered significant.
To assess the effects of TNF-α overexpression on Akt phosphorylation, we probed heart extracts from 6- and 12-wk-old male TNFl.6 and wild-type mice using antibodies that recognize phosphorylated Akt at Ser473. TNF-α overexpression significantly decreased in basal Akt phosphorylation (Fig. 1A). Indeed, the levels of Akt phosphorylation were 70% less in the TNF1.6 mice than in age- and gender-matched controls (Fig. 1C). Moreover, we measured Akt phosphorylation in newborn heart (Fig. 1, B and C). Our data showed that Akt activity in TNF1.6 mice during the neonatal period is not significantly different from that of wild-type controls. By contrast, Akt activity was decreased significantly at both 6 and 12 wk of age, time points during which the pathophysiological consequences of TNF overexpression begin to be observed.
Not only did TNF-α overexpression markedly diminish basal Akt phosphorylation but there was also a marked inhibition of insulin-mediated Akt phosphorylation as seen in Fig. 2, A and B. We determined the correlation between the Akt phosphorylation sites (Ser473 and Thr308) and cardiac Akt kinase activity (Fig. 2C). To this end, cardiac Akt were immunoprecipitated from insulin-stimulated mouse ventricle extracts. Immunoprecipitated Akt phosphorylated recombinant Akt substrate, GSK3β, in vitro, and the data showed that the Akt phosphorylation and Akt activity are correlated in the heart (Fig. 2C).
The inhibitory effect of TNF-α on Akt phosphorylation appeared to be mediated through TNFR1 as TNF-α overexpression did not lower Akt phosphorylation when TNF-α-overexpressing mice were crossed with mice in which TNFR1 was ablated (Fig. 3A). By contrast, genomic disruption of the TNFR2 receptor had no effect on Akt phosphorylation (Fig. 3A). Unlike TNF-α-induced Akt inhibition, NF-κB activation was markedly enhanced in mice overexpressing TNF-α (Fig. 3B). This activation was mediated exclusively through the TNFR1 receptor as deletion of the TNFR1 receptor abrogates TNF-α-mediated activation of NF-κB, whereas deletion of the TNFR2 receptor has no discernible effect (Fig. 3C).
To assess the role of NF-κB on TNF-α-mediated inhibition of Akt phosphorylation, transgenic mice overexpressing TNF-α were crossed with mice carrying the IκBS32A, S36A transgene (13), a potent dominant-negative against NF-κB activation (Fig. 4A). That NF-κB mediates at least in part the TNF-α-mediated decrease in Akt phosphorylation was demonstrated by the fact that dominant-negative IκB transgene expression in TNF1.6 mice significantly attenuated the effects of TNF-α on both basal (Fig. 4B) and insulin-activated (Fig. 4C) Akt phosphorylation. These effects could not be attributed to the amount of TNF-α-mediated interstitial inflammation as neither cardiac expression of TNF-α or IL-1β nor levels of inflammatory infiltrates were altered by expression of the IκBS32A,S36A transgene (Fig. 5). Furthermore, dominant-negative IκB overexpression in cardiac myocytes had no effect on the level of inflammatory exudates as measured by nuclear density, and this is confirmed by staining the myocardium with the mature nucleated hematopoietic lineage marker CD45 molecule (Fig. 6). As expected, cardiac myocyte expression of IκB also did not affect immunostaining pattern of PECAM-1, an endothelial cell marker (Fig. 6D). The infiltrate population is consistent with that which has been described previously to consist predominantly of CD4-positive lymphocytes and macrophages in the TNF-α transgenic (26).
Because NF-κB is a nuclear transcription factor and overexpression of nuclear-targeted Akt can enhance survival of cardiomyocytes (41), we determined endogenous Akt phosphorylation in both cytoplasmic and nuclear extracts after insulin stimulation (Fig. 7A). As seen in Fig. 7, anti-Akt antibody was not able to detect endogenous nuclear Akt, but a minute amount of phospho-Akt was detected, suggesting that almost all endogenous Akt reside in the cardiac cytoplasm. We further evaluated the role of Akt in the cardiac response to stress using Akt2 knockout mice. Of the three Akt subtypes (Akt1, Akt2, and Akt3), Akt1 and Akt2 are equally expressed in the heart (Fig. 7B) while Akt3 is expressed at only low levels. To address the possibility that Akt could effect the development of hypertrophy and remodeling in murine hearts, we assessed the effects of swimming in mice in which Akt1 or Akt2 has been ablated. As seen in Fig. 7C, mice could not mount a physiological response to the stress of forced swimming as mortality was over 50–80% in mice with ablated Akts. Thus Akt1 and Akt2 appear to play a critical role in the heart's ability to respond to physiological (swimming) stress.
To determine whether IκB dominant-negative expression and the resultant increase in Akt phosphorylation had salutary benefits on the hearts of TNF 1.6 mice, we assessed the effects of overexpression of the IκB dominant-negative protein in TNF l.6 mice. As seen in Fig. 8A, inhibition of NF-κB activity in 13-wk-old mice significantly improved fractional shortening and diminished ventricular hypertrophy as evidenced by a significant decrease in the ventricular weight/body weight ratio (Fig. 8B). In addition, overexpression of the dominant-negative IκB transgene significantly improved survival over a period of 32 wk (Fig. 9).
The potential therapeutic benefits of activation of the phosphatidylinositol (PI)-3 kinase and Akt pathway in the heart have been well documented and include studies of Akt-overexpression transgenics (11, 28, 40), construction of constitutive-active and dominant-negative PI-3 kinase (32, 39), and in vivo cardiac gene transfer studies (15, 30, 41, 43). These studies suggest that Akt activation effects enhanced protein synthesis, myocyte hypertrophy, and cell survival. Indeed, activation of Akt may explain at least in part the cardioprotective and anti-apoptotic effects of insulin. However, there have been relatively few studies of the regulation of Akt and in particular of regulation of Akt in heart failure models (42). In the present study, we demonstrate that the activity of Akt is substantially diminished in failing myocardium from mice overexpressing TNF-α. Because activation of Akt is directly related to cell survival and cell growth, these results suggest that inhibition of Akt phosphorylation might be important in the progressive remodeling and apoptosis that characterizes this model. In addition, the data also support the hypothesis that inhibition of Akt phosphorylation is mediated through activation of NF-κB as expression of an IκB dominant-negative transgene substantially reduced NF-κB activity and was associated with a marked increase in Akt phosphorylation. Both NF-κB activation and Akt inhibition were mediated through the TNFR1 receptor, while the salutary benefits of NF-κB inhibition were independent of changes in either the levels of proinflammatory cytokines or the extent of TNF-α-mediated interstitial inflammation.
The marked phenotypic changes associated with inhibition of NF-κB activity through the cardiac myocyte-selective overexpression of IκB dominant negative are similar to those reported previously when TNF1.6 mice were bred with mice in which the p50 subunit of NF-κB was globally ablated (22). Thus cardiac-selective inhibition of NF-κB is adequate to effect salutary changes in a chronic heart failure model (this study) and ischemia-reperfusion infarct size (7). It should be noted that the development of hypertrophy in the heart is complex in that there are two specific types of cardiac hypertrophy: physiological and pathological (36). In the case of TNF-α overexpression, mice presumably develop pathological hypertrophy. By contrast, Akt activation has been associated with physiological hypertrophy similar to that seen when animals undergo forced exercise. The finding that ventricular weight/body weight ratios decreased in the TNF-α mice after NF-κB ablation despite the finding of an increase in the activity of Akt (although levels only normalized and did not actually “increase”) suggests that NF-κB ablation (and or the normalization of Akt) were able to “normalize” hypertrophy in the hearts of TNF overexpression mice. Furthermore, our results are consistent with the observation that mice overexpressing cardiac TNF-α and elderly heart failure patients developed insulin resistance (38, 45) and with the observation of Ananthakrishnan et al. (1) that Akt levels are diminished in canines with overdrive pacing-induced heart failure and elevated TNF-α (34, 35).
By contrast, our results in the in vivo myocardium differ from earlier studies that assessed the effects of TNF-α on the phosphorylation/activity of Akt in cultured neonatal myocytes in which TNF-α activated Akt (12). This disparity might be attributable to the biological differences between a neonatal myocyte in culture and adult heart muscle in vivo. However, there could be substantial differences between acute and chronic exposure to TNF-α and/or cardiac stress signals. The effects of TNF-α on Akt levels in cardiac myocytes peaked at 15 min and returned to near baseline levels by 60 min (12). Similarly, phosphorylation of Akt increased during acute (10 min) pressure overload in mouse hearts but has not been shown to be altered in chronic stress (5). Furthermore, acute and chronic NF-κB stimulation have variable effects on gene regulation (19, 20, 44). Our results also differ from those seen in noncardiac tissues in which Akt activates NF-κB through sequestration of IκB or nuclear transactivation of NF-κB (16, 27, 37). Thus, taken together, these results suggest that both Akt and NF-κB are regulated in a time- and tissue-selective manner through potentially different regulatory mechanisms.
NF-κB is known to have opposing effects depending on the stimulus. We have shown that blocking NF-κB activation in IκB-dominant-negative (DN) transgenic mice reduces infarct size after ischemia-reperfusion injury, suggesting that NF-κB activation is injurious (7). In a rabbit model of ischemic preconditioning, Zhang et al. (46) showed that NF-κB activation is associated with cardioprotection. Furthermore, Misra et al. (33) published that blocking NF-κB in mice after permanent coronary occlusion increases infarct size (i.e., NF-κB activation is protective). We have confirmed the results of Misra et al. in our model and in a recent review (20) showed these data and discussed the complicated and often antithetical effects of NF-κB activation. Current data suggest that NF-κB acts in concert with many other transcription factors and cofactors (and via cross-talk and feedback regulation as we have shown in the present studies with Akt), and, depending on the stimulus, different combinations of these factors activate distinct sets of NF-κB-dependent genes. The ultimate effect of NF-κB abrogation depends on the specific set of NF-κB-dependent genes activated. Thus we predict, and are currently studying, whether NF-κB regulates a distinct pro-cell death set of genes after ischemia-reperfusion and a distinct protective set after permanent coronary occlusion. The results of NF-κB blockade shown in this study support that NF-κB activates major injurious genes in hypertrophy, an inhibitor of Akt among them.
The mechanism by which NF-κB blockade leads to Akt activation most likely involves transcriptional feedback via NF-κB-dependent gene products. This type of feedback mechanism has been shown to regulate NF-κB itself. The expression of NF-κB inhibitor genes, IκBα, IκBβ, and A20, depends primarily on NF-κB activation (2, 23, 44). Although Akt has been shown to affect NF-κB activation (12, 27, 37), NF-κB-dependent signals that oppositely affect Akt phosphorylation remain to be elucidated.
It should also be noted that NF-κB has direct effects, separable from Akt signaling, on apoptosis (4, 31) as well as effects on genes encoding calcium handling proteins, stress response genes, inflammatory proteins, cell adhesion, as well as the inducible nitric oxide synthase (20). Thus we expect that the salutatory effects of NF-κB inhibition are mediated by both Akt-dependent and -independent pathways.
Akt is a kinase that, in general, is regulated upstream of the NF-κB transcriptional factor (6). One of the novel, perhaps surprising, findings of this study is that excess NF-κB activity in the myocardium is associated with a decrease in Akt phosphorylation. By contrast, when we inhibited NF-κB activity, the level of Akt phosphorylation returned to near normal. Although these results do not prove a cause-effect relationship as the improvement in the degree of hypertrophy and cardiac dilation could have altered Akt phosphorylation by a separate mechanism, new insights regarding the role of Akt in cardiac remodeling strongly suggest a relationship between these two events. For example, in mice with decreased Akt expression, the heart is unable to cope with the need for adaptive hypertrophy when stressed with swimming. Future studies are needed to identify the specific signaling systems responsible for the relationship between NF-κB and Akt.
In summary, the results of the present study suggest that NF-κB and Akt play an important interacting role in the development of TNF-α-induced cardiomyopathy. In view of the pathological relevance of TNF-α in both animal models and human disease, Akt may provide an interesting new target for therapeutic intervention. Furthermore, the salutary benefits of NF-κB inhibition suggest that cardiac-directed inhibition of NF-κB activation might enhance Akt activation while at the same time altering NF-κB-mediated cardiotoxic pathways. However, the complex interactions between Akt and NF-κB signaling in the heart require further evaluation.
This research was supported by National Heart, Lung, and Blood Institute Grant R01-HL-63034 (W. K. Jones), Pennsylvania Research Formulary Fund (A. M. Feldman), and Pennsylvania Research Formulary Fund and American Heart Association Grant SDG-F64701 (T. O. Chan).
We thank B. McGowan for technical advice.
Present address of Y. Higuchi: Medical Institute of Bioregulation, Kyushu University, Fukuoka, Japan.
↵* Y. Higuchi and T. O. Chan contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 by the American Physiological Society