Double Stranded RNA–Dependent Protein Kinase Deficiency Protects the Heart from Systolic Overload-Induced Congestive Heart Failure

Abstract

Background

Double stranded RNA–dependent protein kinase (PKR) is a eukaryotic initiation factor 2α kinase that inhibits mRNA translation under stress conditions. PKR also mediates inflammatory and apoptotic signaling independent of translational regulation. Congestive heart failure (CHF) is associated with cardiomyocyte hypertrophy, inflammation, and apoptosis, but the role of PKR in left ventricular (LV) hypertrophy and development of CHF has not been examined.

Methods and Results

We observed an increased myocardial PKR expression and translocation of PKR into the nucleus in humans and mice with CHF. To determine the impact of PKR in the development of CHF, PKR knockout and wild-type mice were exposed to pressure overload produced by transverse aortic constriction (TAC). Though heart size increased similarly in wild-type and PKR knockout mice after TAC, PKR knockout mice exhibited very little pulmonary congestion, well preserved LV ejection fraction and contractility, and significantly less myocardial fibrosis as compared to wild-type mice. Bone marrow-derived cells (BMDCs) from wild-type mice did not abolish the cardiac protective effect observed in PKR knockout mice, while BMDCs from PKR knockout mice had no cardiac protective effect in wild-type mice. Mechanistically, PKR knockout attenuated TAC-induced TNF-α expression and leukocyte infiltration, and lowered cardiac expression of pro-apoptotic factors (Bax and Caspase-3) so that PKR knockout hearts were more resistant to TAC-induced cardiomyocyte apoptosis. PKR depletion in isolated cardiomyocytes also conferred protection against TNF-α or LPS-induced apoptosis.

Conclusions

PKR is a maladaptive factor up-regulated in hemodynamic overload that contributes to myocardial inflammation, cardiomyocyte apoptosis and development of CHF.

Introduction

Double stranded RNA dependent protein kinase (PKR) is a ubiquitously expressed stress-induced eIF2α kinase which represses translation initiation under stress conditions by phosphorylating eIF2α at Ser51.¹ Although PKR was initially identified as an anti-viral factor activated by interferons,² it is now evident that PKR is activated or induced by multiple forms of cell stress including oxidative stress,³ metabolic stress,⁴ mechanical stress,⁵ inflammatory signals⁶ and others⁷. In addition to regulating translation initiation, PKR mediates inflammatory signaling through NF-κB activation^{2, 8} and promotes apoptosis through interactions with Fas associated death domain protein (FADD)^{9, 10} and up-regulation of the pro-apoptotic factor Bax ¹¹. As PKR inhibition is recognized as an attractive therapeutic target for diseases such as cancer, inflammation and Alzheimer’s disease, and pharmacologic PKR specific inhibitors are under development, the impact of PKR and the underlying molecular mechanism of PKR on other clinical conditions have been major research topics.

While PKR dependent inflammatory signaling or translation repression in response to viral infection may be beneficial in limiting viral replication and infectivity, activation of PKR in response to sterile forms of cellular stress encountered in the heart with subsequent increases of apoptosis, inflammation, or repression of translation may have the potential to exacerbate pathological conditions such as congestive heart failure (CHF). Notably, many of the factors that contribute to development of CHF, including oxidative stress,¹² Toll receptor activation¹³ and low grade chronic inflammation, are also known to activate PKR. PKR is expressed in the heart, and may play a role in defense against viral myocarditis,¹⁴ but the involvement of PKR in adaptation to hemodynamic overload, a more common cardiovascular stress condition, is unknown.

Here we utilized human CHF patient left ventricular (LV) samples, isolated cardiomyocytes, and PKR knockout mice to investigate the role of PKR in the cardiac adaptation to hemodynamic overload produced by chronic transverse aortic constriction (TAC). Our results identify PKR as a maladaptive factor up-regulated in human and mouse heart failure. We find that PKR contributes significantly to the development of CHF in the setting of hemodynamic overload produced by TAC, ostensibly by exacerbating myocardial inflammation and apoptosis of cardiomyocytes. Together, our findings suggest PKR inhibition may be an attractive new therapeutic target in treating CHF.

Materials and methods

Mice and TAC Procedure

PKR deficient mice and wild-type (WT) controls were obtained from John C. Bell.¹⁵ Mice 8–12 weeks of age were subjected to TAC using a 26G needle to create the aortic constriction as previously described.¹² Experimental studies in mice and human tissues were approved by the Institutional Animal Care and Use Committee and the Institutional Review Board at the University of Minnesota, respectively.

Data Analysis

Whether data were normal distribution were determined using normality test (Shapiro-Wilk) provided by SigmaPlot. If data were normally distributed, the data were presented as mean ± SEM. Student’s t-test was used to test for differences between 2 groups. Two-way ANOVA followed by a Bonferroni correction post-hoc test was used to test for differences among more than 2 groups. If mouse physiological data were not normally distributed or the sample size in one of the experimental groups was less than 10, the data were presented as median (quartiles). Non-parametric test (Mann-Whitney or Kruskal-Wallis) followed by Bonferroni post hoc correction was performed. All pairwise p-values are two-sided. The null hypothesis was rejected at P < 0.05.

Detailed methods are available in the online-only Data Supplement.

Results

Increased Cardiac PKR Expression and Activation in Human and Mouse CHF

While PKR is ubiquitously expressed, its expression level and activity in the setting of CHF are not known. Therefore, LV samples from patients with CHF or normal donor heart samples were examined by western blot for changes in PKR expression. Clinical details on these samples are presented in online-only Data Supplemental Table 1. Interestingly, PKR protein content was elevated 1.7-fold in CHF samples as compared to normal donor samples (p<0.05) (Figure 1A–1D). Expression of the cardiac stress marker atrial natriuretic peptide (ANP) was elevated in CHF samples, while the sarcoplasmic or endoplasmic reticulum calcium 2 ATPase (Serca2) level was reduced, consistent with the CHF phenotype (Figure 1A–1D). To examine whether PKR activity is altered in human CHF, we analyzed PKR phosphorylation status and PKR sub-cellular localization. PKR activation is often evidenced as autophosphorylation at Thr 446 and Thr 451⁷ as well as translocation of PKR from the cytoplasm into the nucleus¹⁶. Though phospho-PKR was not detectable in LV lysates by western blot, immunostaining of LV sections revealed increased nuclear localization of PKR in cardiomyocytes from CHF patients in comparison to normal donors (Figure 1E, 1F and online-only Data Supplement Figure 1A) suggesting increased activity of PKR in CHF. These results suggest both PKR expression and activation are increased in human heart failure.

Figure 1.

PKR expression and activation are increased in human and mouse left ventricle (LV) of congestive heart failure (CHF). A to D, Western blot of ANP, PKR, Serca2 and vinculin (loading control) in the LV from normal donor and CHF patients. E and F, Quantitative data and representative images of PKR immunostaining (red) on the LV from normal donors and CHF patients. G to K, Western blot and quantitative RT-PCR of PKR in whole cell and nuclear extracts from the LV of control mice (Ctr) and CHF mice (TAC). L and M, Quantitative data and representative images of PKR immunostaining (red) on the LV of control mice and CHF mice. Nuclei were stained with DAPI (blue). Myocytes were stained for f-actin with phalloidin (green). n=4 per group for human samples. n=5 per group for mouse samples. *P<0.05 vs donor or control group.

LV pressure overload resulting from hypertension is a common form of pathologic cardiac stress that can lead to the development of CHF. To examine whether PKR is up-regulated in response to hemodynamic overload, WT mice were subjected to LV pressure overload using TAC. Similar to our observations in CHF patient samples, TAC resulted in a 2-fold increase in PKR expression at both protein and mRNA levels in LV total cell lysates (Figure 1G–1I). Western blot analysis of nuclear extracts revealed a ~3-fold increase in nuclear PKR levels in response to TAC (Figure 1J and 1K). Immunohistochemical analysis demonstrated that PKR expression was increased in both cytoplasm and nucleus in cardiomyocytes of the failing hearts (Figure 1L, 1M and online-only Data Supplement Figure 1B). These data indicate that hemodynamic overload causes increased expression and activation of PKR.

PKR Deficiency Attenuated TAC-induced Pulmonary Congestion and Cardiac Dysfunction

To determine whether PKR influences cardiac adaptation to hemodynamic overload, PKR knockout (PKR KO) mice and WT mice were examined under basal conditions and after exposure to chronic pressure overload produced by 9 weeks TAC. Under basal conditions, PKR deficiency had no effect on the ratio of LV, left atrial (LA), lung weight to body weight or myocyte size (Figure 2A–2D and online-only Data Supplement Table 2). In response to TAC, the LV weight to body weight ratio and myocyte size increased similarly in WT and PKR KO hearts (Figure 2A, 2B and online-only Data Supplement Figure 2), suggesting that PKR does not significantly regulate pressure overload induced-LV hypertrophy. However, examination of LA weight to body weight ratio revealed that WT mice developed significantly greater LA hypertrophy than PKR KO mice (Figure 2C). Furthermore, examination of lung weights in WT and PKR KO mice showed that WT mice had developed significantly 68% more pulmonary congestion than PKR KO (Figure 2D). The protective effects of PKR KO against TAC-induced pulmonary congestion and LA hypertrophy were independent of differences in LV weight. Echocardiographic measurements showed that PKR disruption did not affect LV function under basal conditions (Figure 2E–2H). After TAC, however, WT mice exhibited a 34% decline in LV ejection fraction and a 35% increase in LV end-systolic diameter while these changes were significantly less in PKR KO mice (Figure 2E, 2F and online-only Data Supplement Table 3), indicating less reduction of systolic function in PKR KO mice after TAC. In addition, LV end-diastolic diameter was significantly increased (Figure 2G) and LV end-diastolic wall thickness was significantly decreased in WT as compared to PKR KO (Figure 2H) after TAC, reflecting less abnormal diastolic relaxation in PKR KO mice after TAC. Despite the reduced wall stress, PKR KO mice showed the same level of LV hypertrophy as WT mice after TAC, suggesting that overall LV protein synthesis in PKR KO was enhanced under comparable wall stress. This could be the result of an inhibitory function of PKR in mRNA translation under cell stress conditions.

Figure 2.

PKR deficiency attenuates transverse aortic constriction (TAC)-induced pulmonary congestion and cardiac dysfunction. Data were collected from wild-type (WT) and PKR knockout (KO) mice under basal conditions (Ctr) or 9 weeks after TAC. A to D, The ratio of left ventricle (LV), left atria, lung weight to body weight and myocyte size of WT mice and PKR KO mice under basal or TAC conditions. E to H, Echocardiographic measurements of LV ejection fraction, LV end-systolic diameter, LV end-diastolic diameter and LV end-diastolic wall thickness from PKR KO and WT hearts. I to K, Hemodynamics of LV end-diastolic pressure, LV maximum rate of rise of pressure (dP/dt_max) and LV maximum rate of decline of pressure (dP/dt_min) from PKR KO and WT hearts. L to O, Western blot of PKR, ANP, Serca2 and vinculin (loading control) in WT and PKR KO mice under control or TAC conditions. *P<0.05 vs control group; #P<0.05 vs TAC group of WT mice.

Under basal conditions, PKR disruption had no effect on LV end-diastolic pressure (LVEDP), LV maximum rate of rise of pressure (dP/dtmax) or LV maximum rate of decline of pressure (dP/dtmin) in comparison to WT mice (Figure 2I to 2K). In response to TAC, LVEDP increased significantly in WT mice, but this increase was significantly blunted in PKR KO mice (Figure 2I). Similarly, LV dP/dtmax and dP/dtmin were significantly depressed in WT mice in response to TAC, while PKR KO mice exhibited 54% higher LV dP/dtmax and 40% greater dP/dtmin than WT mice (Figure 2J and 2K). Together, these data suggest that PKR activity impairs LV contractility and diastolic function in the setting of pressure overload. The effects of PKR KO on lung weight, LA remodeling, LVEDP and LV dP/dtmin are more impressive than on LV ejection fraction, suggesting that PKR contributes more significantly in TAC-induced LV diastolic dysfunction.

To further examine the impact of PKR in the development of CHF, expression levels of ANP and Serca2 were measured by western blot. Under basal conditions no difference in ANP or Serca2 expression was observed between WT and PKR KO mice. In response to TAC, ANP expression was markedly increased in WT mice, but this TAC-induced ANP expression was significantly attenuated in PKR KO mice. Similarly, expression of Serca2, which is often reduced in CHF, was lower in WT mice exposed to TAC, but preserved in PKR KO mice under the same conditions (Figure 2L–2O). Although PKR is well known as an eIF2α kinase that represses translation initiation by phosphorylating eIF2α^Ser51, PKR KO did not affect phosphorylation of eif2α^Ser51 in the heart (online-only Data Supplement Figure 3). Together, these data indicate that PKR is maladaptive in the setting of systolic overload, so that PKR disruption is protective against TAC-induced LV dysfunction and CHF.

PKR Deficiency Attenuated TAC-induced LV Inflammatory Cytokine Expression and LV Fibrosis

Besides regulating translation, PKR also mediates inflammatory signaling in a variety of pathological conditions.^{4, 17} To determine if PKR plays a role in TAC-induced myocardial inflammation, protein levels of TNF-α were measured in LV lysates from WT and PKR KO mice by western blot and mRNA levels of TNF-α and IL-1β in LV tissues were measured by quantitative RT-PCR. As shown in Figure 3A, TNF-α protein expression as well as TNF-α and IL-1β mRNA levels were significantly increased in the LV of WT mice in response to TAC. These increases were blunted, however, by 51% (TNF-α) and 42% (IL-1β), respectively, in PKR KO hearts (Figure 3A, 3B, 3D and 3E). TGF-β, a pro-fibrogenic cytokine, was also up-regulated in response to TAC in WT mice, but this increase was reduced in PKR KO mice by 39% in protein expression and 24% at mRNA level (Figure 3A, 3C and 3F). Consistent with lower TGF-β expression in PKR KO mice, TAC-induced LV fibrosis as identified by Trichrome staining was also significantly reduced in PKR KO mice as compared to WT mice (Figure 3G and online-only Data Supplement Figure 4). These data suggest that PKR activity promotes myocardial inflammatory cytokine expression and the development of LV fibrosis in response to pressure overload.

Figure 3.

PKR deficiency attenuates transverse aortic constriction (TAC)-induced left ventricular (LV) inflammatory cytokine expression and LV fibrosis. A to C, Western blot of TNF-α and TGF-β in LV lysates of wild-type (WT) and PKR knockout (KO) mice under basal (Ctr) or TAC conditions. D to F, Quantitative RT-PCR results of TNF-α, IL-1β and TGF-β in LV lysates of WT and KO mice under basal or TAC conditions. G, Quantitative data of Trichrome staining for detection of fibrosis in WT and PKR KO mice under basal or TAC conditions. n=5 per group. *P<0.05 vs control group; #P<0.05 vs TAC group of WT mice.

PKR Deficiency Inhibited TAC-induced Cardiomyocyte Pro-apoptotic Protein Expression and Apoptosis

Apoptosis occurs at a low frequency in the heart. However, accumulated loss of cardiomyocytes from a slight but chronic increase in the rate of apoptosis may eventually lead to CHF.^{18, 19} In support of previous studies showing increased cardiomyocyte apoptosis in human heart failure,²⁰ we observed a significantly greater number of TUNEL positive nuclei in LV sections from CHF patients than in those from normal donor hearts (Figure 4A and online-only Data Supplement Figure 5A). PKR expression and activity were also elevated in failing human hearts, but whether PKR promotes cardiomyocyte apoptosis in response to LV stress was not known. Because PKR is elevated in hearts exposed to pressure overload and mediates apoptosis in response to inflammatory signals such as TNF-α^{21, 22} (which we also observed was elevated in response to TAC), we examined whether PKR contributes to cardiomyocyte apoptosis in the setting of hemodynamic overload. Measurement of apoptosis by TUNEL staining showed that TAC significantly increased the percent of TUNEL positive apoptotic nuclei in LV sections of WT mice (0.05% in control versus 0.27% after TAC) (Figure 4B and online-only Data Supplement Figure 5B). This increase was blunted by 50% in PKR KO mice, suggesting that PKR KO mice are more resistant to TAC-induced cardiomyocyte apoptosis.

Figure 4.

PKR deficiency inhibits transverse aortic constriction (TAC)-induced cardiomyocyte apoptosis. A, Quantitative data of TUNEL staining in left ventricular (LV) samples from normal donor and congestive heart failure (CHF) patient. B, Quantitative data of TUNEL staining of LV sections from wild-type (WT) and PKR knockout (KO) mice under basal (Ctr) or TAC conditions. C to E, Western blot of Bax and Caspase-3 in LV lysates from WT and PKR KO hearts under basal or TAC conditions. F and G, Quantitative RT-PCR of Bax and Caspase-3 in LV lysates from WT and PKR KO hearts under basal or TAC conditions. H to J, Western blot of Bax and Caspase-3 in the LV from normal donor and CHF patient LV samples. n=4 to 5 per group. *P<0.05 vs donor or control group; #P<0.05 vs TAC group of WT mice.

PKR mediates apoptosis through multiple mechanisms, including interaction with FADD,^{9, 11} up-regulation of the pro-apoptotic factor Bax, down-regulation of Bcl-2,¹¹ and activation of the caspase-8/caspase-3 pathway.^{9, 10} Interestingly, PKR KO hearts expressed significantly lower levels of pro-apoptotic proteins Bax and Caspase-3 than WT hearts even under unstressed conditions (Figure 4C–4G). Exposure to TAC significantly increased protein expression and mRNA levels of Bax and Caspase-3 in WT hearts, but this increase was significantly blunted in PKR KO hearts (Figure 4C–4G). LV lysates from CHF patients also showed increased expression of Bax and Caspase-3 (Figure 4H–4J), suggesting up-regulation of these proteins is a common response to cardiac stress that increases cardiomyocyte susceptibility to apoptosis. These data demonstrate that PKR disruption limits expression of pro-apoptotic proteins Bax and Caspase-3, and attenuates TAC-induced cardiomyocyte apoptosis.

PKR Deficiency in Bone Marrow-derived Cells (BMDCs) Did Not Attenuate TAC-induced CHF

Because PKR deletion is global, it is possible that loss of PKR expression in bone marrow-derived immune cells attenuates the inflammatory response in pressure overload, and thus contributes to the better adaptation of PKR KO mice to TAC. To determine the influence of BMDC PKR expression in adaptation to pressure overload, we ablated hematopoetic cells in WT mice by lethal irradiation and reconstituted the BMDC population by bone marrow transplantation from either WT or PKR KO donor mice, and then performed TAC 4 weeks later. As demonstrated in online-only Data Supplement Figure 6, PKR expression was readily detectable in white blood cells from mice receiving BMDCs from WT donors, but undetectable in mice that received BMDCs from PKR KO donors. Thirteen weeks after TAC, WT mice containing WT or PKR KO reconstituted BMDCs exhibited similar development of pulmonary congestion and LV dysfunction (Figure 5 and online-only Data Supplement Figure 7), indicating that ablation of PKR activity specifically in BMDCs is not the cause of improved adaptation to hemodynamic overload observed in global PKR KO mice.

Figure 5.

PKR deficiency in bone marrow-derived cells (BMDCs) dose not attenuate transverse aortic constriction (TAC)-induced congestive heart failure (CHF). Samples of blood cells, hearts and lungs, as well as evaluation of echocardiographic measurements were collected under basal conditions (Ctr) or 14 weeks after TAC. A to C, The ratio of left ventricular (LV), LA and lung weight to body weight of wild-type (WT) mice containing WT or PKR knockout (KO) reconstituted BMDCs under basal or TAC conditions. D to F, Echocardiographic measurements of LV ejection fraction, LV end-diastolic diameter, LV end-systolic diameter from WT mice containing WT or PKR KO reconstituted BMDCs under basal or TAC conditions.

WT BMDCs Did Not Exacerbate TAC-induced CHF in Global PKR KO Mice

To confirm that the protective effects of global PKR KO were due to loss of PKR in heart cells, rather than circulating immune cells, we reconstituted irradiated PKR KO mice and WT mice with BMDCs derived from WT mice. As shown in Figure 6A–6J, PKR KO mice reconstituted with WT BMDCs exhibited comparable cardiac phenotypes with WT mice reconstituted with WT BMDCs under basal conditions. In response to TAC, PKR KO mice reconstituted with WT BMDCs developed significantly less reduction of ejection fraction, less LV diastolic dysfunction and less pulmonary congestion than WT mice reconstituted with WT BMDCs (Figure 6A–6F), suggesting that PKR distributed in heart (and possible in lung) contributes to the cardiac protective effect observed in global PKR KO mice.

Figure 6.

Bone marrow-derived cells (BMDCs) from wild-type mice did not abolish the cardiac protective effect observed in PKR knockout (KO) mice. A to C, The ratio of left ventricular (LV), left atrial (LA) and lung weight to body weight of WT mice containing WT or PKR KO reconstituted BMDCs under basal or TAC conditions. D to F, Echocardiographic measurements of LV ejection fraction, LV end-diastolic diameter, LV end-systolic diameter in WT or PKR KO mice reconstituted with BMDCs from wild-type mice under basal or TAC conditions. *P<0.05 vs control group.

PKR was Increased by TNF-α and LPS and PKR Knockdown Inhibited Inflammatory Cytokines, Fibrotic Factors and Apoptosis in Cardiomyocytes

Our in vivo study demonstrated that global PKR KO protected the heart from TAC-induced inflammatory cytokine expression, apoptosis and CHF. PKR is activated by inflammatory cytokines, including TNF-α. The beneficial effect of PKR KO was not due to loss of PKR activity in BMDCs, because TAC-induced CHF was not improved when PKR was disrupted specifically in hematopoetic cells. This suggests PKR in heart cells is contributing to the maladaptive response to pressure overload. To determine whether inflammatory factors regulate PKR specifically in cardiomyocytes, we isolated and exposed cultured neonatal rat cardiomyocytes to TNF-α or LPS. Interestingly, exposure to TNF-α (10 ng/ml) or LPS (100ng/ml) significantly increased cardiomyocyte PKR expression (Figure 7A and 7B). In addition, TNF-α and LPS treatment promoted PKR translocation to the nucleus, suggesting increased activation of PKR (online-only Data Supplement Figure 8). These results indicate that TNF-α and LPS increase cardiomyocyte PKR expression and activation.

Figure 7.

PKR knockdown inhibits TNF-α or LPS-induced inflammation and apoptosis in cardiomyocytes. Neonatal cardiomyocytes were treated with TNF-α (10ng/ml) or LPS (100ng/ml). A and B, Western blot of PKR and vinculin (loading control) in cells treated for 20 hours. C to E, Quantitative RT-PCR of TNF-α, IL-1β and TGF-β in cells treated for 8 hours. F, Quantitative data of apoptosis in cells treated for 20 hours using TUNEL staining. G and H, Apoptosis was measured in cells treated for 4 hours using Annexin V staining followed by flow cytometry. I to L, Western blot of Bax and Caspase-3 in cells after 20 hours of exposure to TNF-α or LPS. M and N, Quantitative data of Bax and Caspase-3 in cells after 8 hours of exposure to TNF-α or LPS. O, Diagram of underlying mechanism. Data are representative of 3 independent experiments. *P<0.05 vs control group; #P<0.05 vs TNF-α or LPS group treated with control siRNA.

Toll-like receptor 4 (TLR4) activation, which increases secretion of TNF-α and IL-1β in the heart is remarkably enhanced in systolic overload-induced CHF.^23,24 PKR deficiency inhibits TLR4-induced TNF-α secretion in mouse embryonic fibroblasts and macrophages.^25–26 Therefore, it is possible that cardiomyocytes secrete inflammatory factors during TAC through PKR-dependent mechanisms. To test this hypothesis, small interfering RNA (siRNA) was used to selectively silence PKR in cultured rat neonatal cardiomyocytes, and TNF-α and IL-1β in cells were measured by quantitative RT-PCR after exposure to LPS for 8 hours. In cells treated with non-targeting control siRNA (siCtr), the mRNA level was increased 9 times for TNF-α and 7 times for IL-1β in response to LPS. PKR knockdown (siPKR) decreased the LPS-induced increase of TNF-α by 50% and decreased the LPS-induced increase of IL-1β by 36% in cells (Figure 7C 7D). Given the fact that TNF-α is an important stimulator for TGF-β secretion ²⁷ we further investigated whether PKR played a role in TNF-α-induced TGF-β secretion in cultured cardiomyocytes. In siCtr-treated cells, the mRNA level of TGF-β was increased 33% in response to TNF-α. PKR knockdown reduced TGF-β in cells by about 50% under basal conditions and after TNF-α treatment (Figure 7E). PKR did not significantly alter eif2α^Ser51 phosphorylation (online-only Data Supplement Figure 9). These results demonstrate that PKR activity promotes inflammatory cytokine secretion and fibrotic factor production in cardiomyocytes.

PKR mediates TNF-α or LPS-induced apoptosis in NIH3T3 cells, MEFs and U937 cells.^{21– 22,28} TNF-α and other inflammatory factors are often increased in CHF²⁹ and TNF-α has been shown to induce apoptosis in cardiomyocytes.^{30, 31} Therefore, it is important to determine if PKR expression increases cardiomyocyte susceptibility to TNF-α-induced apoptosis. To examine the specific role of cardiomyocyte PKR in regulating apoptotic sensitivity to inflammatory factors, apoptosis was measured in cultured rat neonatal cardiomyocytes treated with siCtr or siPKR after exposure to TNF-α or LPS for 20 hours. In cells treated with siCtr, the percent of TUNEL-positive cardiomyocyte nuclei increased from 4.6% under basal conditions to 8.4% in response to TNF-α and to 13.8% in response to LPS. PKR knockdown significantly protected cardiomyocytes from apoptosis, decreasing the percentage of TNF-α or LPS-induced apoptotic cells to 5.5% or 10.3%, respectively (Figure 7F and online-only Data Supplement Figure 10). Further analysis using Annexin V staining by flow cytometry on propidium iodide negative cells (viable cells) confirmed that PKR siRNA treatment significantly decreased the number of cardiomyocytes undergoing apoptosis in response to TNF-α or LPS (Figure 7G, 7H and online-only Data Supplement Figure 11). Consistent with our in vivo data showing that PKR KO mice express less cardiac Bax and Caspase-3, siRNA depletion of PKR specifically in cardiomyocytes also attenuated protein expressions and mRNA levels of Bax and Caspase-3 induced by TNF-α or LPS (Figure 7I–7N). These data indicate that cardiomyocyte PKR mediates increased expression of pro-apoptotic Bax and Caspase-3 expression and promotes cardiomyocyte apoptosis in response to inflammatory factors.

Discussion

Our findings identify PKR as a maladaptive factor up-regulated in hemodynamic overload that promotes the development of CHF. PKR expression and activation were increased both in human and murine failing hearts, while genetic disruption of PKR profoundly attenuated pressure overload-induced CHF. Mechanistically, PKR disruption blunted TAC-induced myocardial inflammatory cytokine expression and protected against cardiomyocyte apoptosis, suggesting that PKR exerts deleterious effects in the overloaded heart by amplifying inflammatory pathways and sensitizing cardiomyocytes to apoptotic stimuli which contribute to the development of CHF.

PKR is expressed ubiquitously in mammals.¹ In addition to its well-known function of mediating the antiviral effects of interferons in leukocytes,² PKR plays an important role in regulating cell inflammation, proliferation and apoptosis in uninfected cell types¹ and may be involved in a variety of cardiovascular diseases associated with chronic inflammation. PKR is activated by multiple forms of cellular stress,¹ including oxidative stress,³² metabolic stress⁴ and inflammation¹⁷. These forms of cellular stress are observed in pressure overload-induced hypertrophy and heart failure, but the involvement of PKR in these conditions has not previously been examined. We observed that both PKR expression and nuclear translocation were increased in mouse and human heart failure samples. PKR auto-phosphorylation⁷ and nuclear translocation are both associated with PKR activation.^{16, 33} While we were unable to detect phosphorylated PKR in LV samples, the increased nuclear translocation of PKR in CHF samples, or in response to TAC and inflammatory stimuli, suggests either cardiomyocyte PKR activation or an alternative kinase independent nuclear function of PKR that is increased under these stress conditions. The specific role of PKR in the nucleus is unclear, but may relate to its nuclear interactions with FADD, which promotes apoptosis in neuronal cells.⁹

Interestingly, despite the role of PKR as an eIF2α kinase,² eIF2α^Ser51 phosphorylation was not reduced in LV lysates of PKR KO mice or in isolated cardiomyocytes depleted of PKR using siRNA. This finding is consistent with previous studies which showed eIF2α^Ser51 phosphorylation was not reduced in mouse embryo fibroblasts derived from PKR KO mice,¹⁵ and suggests that in the heart other eIF2α kinases such as GCN2 or PERK may play a larger role in eIF2α^Ser51 phosphorylation than PKR, or may compensate for loss of PKR. The robust cardioprotective effects of PKR KO, however, in the absence of observable changes in eIF2α^Ser51 phosphorylation, suggest the deleterious effects of PKR in the heart may be due to PKR effects that are independent of eIF2α regulation.

Chronic low grade inflammation is a feature of human CHF that is believed to contribute to worsening heart function. In addition to regulating translation initiation, PKR also mediates inflammatory signaling¹⁷ and activation of cell death pathways^{9, 22}. While PKR expression is induced by inflammatory factors such as TNF-α,³³ PKR activation itself can induce expression of TNF-α and other inflammatory cytokines through inflammasome activity,¹⁷ p38 map kinase,³⁴ and NF-κB activation^35,36. Interestingly, we did not observe an effect of PKR disruption on p38 phosphorylation or NF-κB p65 subunit phosphorylation in cultured cardiomyocytes (online-only Data Supplement Figure 12). Consistent with a role of PKR in inflammation, TAC-induced expression of TNF-α and IL-1β were significantly attenuated in PKR KO hearts. There is experimental evidence that TNF-α^{13, 37} and IL-1β³⁸ contribute to the development of CHF, and that blocking expression or activity of these factors can attenuate the development of CHF in animal models.^13,37,38 Thus, the reduction in TNF-α and IL-1β in PKR KO hearts may partially explain the beneficial effects of PKR disruption in attenuating the development of CHF. Although infiltration of inflammatory cells was very rare in the heart even after TAC, the number of leukocytes was significantly less in PKR KO hearts than that in WT hearts after TAC (online-only Data Supplement Figure 13). Reconstitution of WT mice with PKR KO BMDCs did not influence the deterioration of LV function after TAC, indicating PKR expression in BMDCs does not mediate its maladaptive effects on LV dysfunction. However, reconstitution of PKR KO mice with WT BMDCs could not abrogate the protective effect of global PKR KO in CHF. Decreasing PKR expression in cardiomyocytes by selective gene silencing of PKR attenuated expression of inflammatory cytokines, fibrotic factors and apoptosis in cardiomyocytes, suggesting PKR expressed in cardiomyocytes mediates its maladaptive effects on pressure overload-induced CHF.

PKR has been shown to promote apoptosis through eIF2α phosphorylation³⁹ and through an FADD/Caspase8/Caspase 3 signaling pathway^{9, 10}. PKR can also increase expression of pro-apoptotic factors Bax and p53 and decrease expression of Bcl-2¹¹. We observed that pro-apoptotic factors Bax and Caspase-3 were both increased in LV samples from mouse and human CHF, and were also up-regulated in response to TNF-α in cardiomyocytes. PKR KO mice or PKR depleted cardiomyocytes, however, expressed lower levels of Bax and Caspase-3, suggesting that PKR may sensitize cardiomyocytes to apoptosis by up-regulating expression of these factors. In support of this proposition, PKR KO hearts and PKR depleted cardiomyocytes exhibited increased resistance to TAC or inflammatory factor-induced apoptosis. Our data showed that PKR disruption significantly protected against cardiomyocyte apoptosis both in vivo and in vitro, but this was not related to its effects on eIF2α phosphorylation. The fact that PKR was localized in the nucleus after activation and in line with protein levels, mRNA levels of apoptotic factors were also down-regulated by PKR disruption suggests that PKR are involved in the transcriptional regulation of apoptotic factors. These data suggest that PKR is central in a deleterious cycle of stress-induced inflammation and cardiomyocyte apoptosis, so that PKR KO reduces both chronic inflammation and cardiomyocyte apoptosis. The underlying mechanism is summarized in Figure 7O.

Together our findings indicate PKR is a maladaptive factor in hemodynamic overload that promotes inflammation, cardiomyocyte apoptosis and the development of CHF. Because global PKR KO has no overt deleterious effects under basal conditions, but offers profound protection against the development of LV dysfunction in the setting of systolic overload, pharmacological inhibition of PKR may have therapeutic potential in treating CHF.