The present study identified important roles for PGC-1β in the maintenance of cardiac OXPHOS gene expression and mitochondrial function in non-stressed hearts and cardiac function following pressure overload hypertrophy (POH). PGC-1β plays an important role in maintaining cardiac function in compensated POH, as evidenced by the increased susceptibility of PGC-1β deficient hearts to heart failure. The mechanisms by which PGC-1β contributes to the maintenance of cardiac function in response to pressure overload are complex and cannot be completely accounted for by changes in the expression of nuclear genes that regulate mitochondrial OXPHOS, FAO or glucose metabolism, given the similarity in the transcriptional responses of banded WT and KO mice. Moreover, we identify some circumstances where changes in gene expression do not necessarily correlate with changes at the protein levels as exemplified by hexokinase II in banded PGC-1β KO hearts. Rather, a number of defects became apparent with banding in PGC-1β deficient hearts that include increased oxidative stress that may be precipitated in part by an early loss of anti-oxidant defenses, a precipitous reduction in glucose utilization and reduced cardiac efficiency that precede the subsequent decline in cardiac function. By the time that heart failure ensues, additional defects were noted in β KO hearts such as decreased palmitoyl carnitine ATP generation, increased AMPK phosphorylation and decreased HKII content. It is however challenging to discern if these additional molecular defects are a cause or a consequence of the more severe heart failure in β KO hearts.
In non-stressed hearts, no differences in heart weights were observed between WT and βKO mice and expression of natriuretic peptide precursor type A and B were unchanged. At the ages at which the banded hearts were analyzed, there were no differences in contractile function between sham operated βKO and WT, which is consistent with previous reports 10, 21
. Interestingly however, in young β KO mice a reduction in basal contractility was observed, suggesting that deficiency of PGC-1β led to a constitutive defect in contractile function that did not progress with age. Importantly, inotropic reserve in response to dobutamine was relatively preserved in young β KO mouse hearts, which supports observations by Lai et al. who reported that LV contractile function following exercise was not impaired in β KO mice of similar age 21
. We confirmed a defect in chronotropic reserve in young β KO mice that we previously observed in an older cohort of mice 10
. However, there were no differences in heart rate following banding in β KO mice. Taken together, these data suggest that PGC-1β plays an important role in preserving LV function in response to pressure overload. However, we cannot completely rule out the possibility that this constitutive defect could contribute in part to the more rapid decline in cardiac function when pressure overload was imposed. Cardiac phenotypes have been described for two independently generated lines of PGC-1α KO mice. One model showed normal contractile function under basal conditions 4, 22
and the other exhibited age-dependent contractile dysfunction 23
. Deletion of PGC-1α, results in impaired expression of both FAO and OXPHOS genes in non-stressed hearts 4
. In contrast, we observed that whereas PGC-1β plays an important role in the regulation of OXPHOS gene expression, it does not appear to be necessary for regulating FAO genes, suggesting divergent roles for PGC-1 isoforms in regulating FAO gene expression. Similar changes in mitochondrial gene expression were noted in the hearts of 2 other independently generated PGC-1β deficient mouse models21, 24
. Importantly, repression of OXPHOS and/or FAO gene expression in these models is not sufficient to consistently impair cardiac function in the non-stressed heart.
Deficiency of PGC-1α in the heart reduces maximal palmitoyl-carnitine (PC) supported (VADP
) respiration and increases pyruvate supported VADP
. In the present study, we observed impaired PC-supported VADP
respiration and ATP-synthesis in non-stressed hearts lacking PGC-1β. But in contrast to PGC-1α-KO (α KO) hearts, KO of PGC-1β had no effect on pyruvate-supported mitochondrial function. The basis for the differences in pyruvate-supported respiration in α vs. β KO mice is not fully understood. We speculate that because PGC-1α deficiency significantly represses OXPHOS and FAO gene expression, there is a greater reliance of α KO mitochondria on glucose carbons, which drives adaptive responses that increase pyruvate utilization. In this regard, we observed a reduction in expression levels of the PDHA1 subunit in sham β KO hearts, which could also limit the induction of pyruvate utilization. In contrast, there was no repression of FAO genes in β KO hearts, nor was there any defect in FA utilization in isolated working hearts. PC-supported respiration and ATP synthesis rates were reduced in β KO mitochondria. However, the reductions were not as striking as we previously reported in α KO mitochondria where PC-supported ATP generation was reduced by 65% 4
, which contrasts with the 30% reduction in β KO mitochondrial ATP synthesis, observed in the present study. These findings are consistent with the hypothesis that the greater decline in mitochondrial function with FA substrates in α KO mice reflects the combined effects of reduced OXPHOS and FAO capacity. Electron microscopy revealed normal mitochondrial volume density and number for non-stressed PGC-1β deficient hearts, which is consistent with previous reports that examined younger PGC-1β KO hearts 21
. It is important to note though that mitochondrial volume density declines with age in PGC-1β KO hearts 10
. Thus analyses of non-stressed βKO hearts reveal divergent effects for PGC-1α vs. PGC-1β on the expression of mitochondrial metabolic genes and mitochondrial function. However, both of these models suggest that mitochondrial dysfunction that develops on the basis of loss of either PGC-1α or PGC-1β is not sufficient to impair cardiac function throughout life in the absence of a superimposed stress. They also support the existence of overlapping or redundant roles for both of these transcriptional regulators in the regulation of nuclear-encoded genes that regulate mitochondrial metabolism and suggest that loss of PGC-1α may compensate for the loss of PGC-1β in maintaining contractile function under non-stressed conditions and vice versa. Support for redundant roles for PGC-1α and PGC-1β in maintaining basal mitochondrial function in the heart was also provided by analysis of mice that lack both PGC-1α and PGC-1β, which die of heart failure shortly after birth with bradycardia, small hearts and reduced cardiac output 21
We examined the impact of PGC-1β deletion on pressure overload cardiac hypertrophy (POH). PGC-1β KO mice progressed more rapidly to heart failure than banded WT mice, and this was clearly associated with evidence of energetic stress, namely increased phosphorylation of AMPK. An earlier study demonstrated that mice lacking PGC-1α developed heart failure eight weeks following TAC. This was associated with additional repression of FAO and OXPHOS genes in PGC-1α-KO mice in response to TAC, versus WT-TAC or sham-operated PGC-1α-KO mice 8
. It is important to note that the additional repression of OXPHOS and FAO gene expression in banded PGC-1α-KO mice occurred despite normal expression of PGC-1β. These data suggest that PGC-independent mechanisms must exist that account for reduced expression of mitochondrial target genes in pressure overload hypertrophy. They also suggest that there is a critical threshold of mitochondrial activity that is required to cope with increased workload, which cannot be substituted for or sustained by PGC-1β. The transition to heart failure was not associated with striking changes in mitochondrial morphology in β KO hearts. Although unexpected, it is probable that ultrastructural studies were performed at a time point shortly after the transition to heart failure and that a longer period of observation might be needed before changes in mitochondrial morphology become apparent.
It is also likely that energetic limitations might not be the only mechanism that contributed to heart failure in PGC-1α and PGC-1β KO hearts following TAC. A recent study in banded PGC-1α-KO mice suggested that increased oxidative stress partially drove the LV dysfunction that was observed 9
. In the present study we also obtained evidence of significant oxidative stress that developed in PGC-1β KO hearts and was evident as early as three weeks post-TAC, prior to any differences in LV function. In this regard, it is interesting to note that protein levels of MnSOD were significantly repressed as early as three weeks in banded PGC-1β KO hearts but not in banded WT hearts in which repression of MnSOD was only evident after 8-weeks. It is also noteworthy that preserved MnSOD content, 3-weeks post-TAC correlated with increased PGC-1β expression, despite reduced PGC-1α expression in WT hearts, whereas at 8-weeks, the expression of PGC-1α and β were both reduced in banded WT hearts. Taken together, these data indicate that PGC-1β plays an important role in supporting anti-oxidant mechanisms that may limit oxidative stress in the compensated state of pressure overload cardiac hypertrophy and represents another function that overlaps with that of PGC-1α.
Interesting changes were observed in the expression levels and content of uncoupling proteins, which could also contribute to the cardiac responses observed. UCP3 protein and mRNA were increased in sham β KO hearts, and were repressed by TAC in KO and WT hearts. UCP2 mRNA also exhibited a similar pattern. Uncoupling proteins have been proposed to play a role in reducing mitochondrial membrane potential under conditions where ROS is increased 20
. Thus the increase in UCP2 and UCP3 in the PGC-1β deficient hearts could reflect an adaptation to low levels of oxidative stress that might also be present in non-stressed β KO hearts. The increase in UCPs could also contribute to the increase in MVO2
and reduced cardiac efficiency that characterized these hearts. ROS are potent activators of uncoupling proteins, thus the persistently elevated MVO2
in banded β KO hearts could represent ROS-mediated activation of UCPs in the face of progressive oxidative stress, despite falling levels of UCPs following TAC. Isolated working hearts revealed significant differences in substrate utilization between PGC-1β deficient hearts and WT controls. As reported by others, compensated LVH in wildtype hearts was associated with increased glucose utilization. Unexpectedly, glucose and FA utilization was increased in non-stressed β KO hearts. The molecular mechanisms for these adaptations are not immediately apparent, and will require additional studies in the future. Importantly, these adaptations were not sustained in β KO hearts following TAC. The global reduction in glucose utilization is not accounted for by differences in GLUT4 protein or by increased phosphorylation of PDH. There is the suggestion that hexokinase II (HKII) levels might fall more rapidly in banded β KO hearts and could contribute to decreased glucose utilization, although it is likely that additional mechanisms are also involved. An association between HKII and VDAC in mitochondria plays an important role in cardioprotection 25
. We did not measure mitochondrial localization of HKII in response to banding in β KO hearts. If the decline in HKII in whole heart homogenates parallels a decline in mitochondrial HKII, this could potentially contribute to exacerbated mitochondrial dysfunction.
Our results support a model that PGC-1α and PGC-1β play partially overlapping but somewhat distinct roles in maintaining cardiac mitochondrial energetics in the unstressed heart with both pathways regulating the expression of OXPHOS genes, while PGC-1α predominantly regulates FAO capacity. Both PGC-1 isoforms likely contribute to the maintenance of cardiac function in the context of pressure overload. While it is clear that PGC-1α expression is repressed much earlier in the course of POH than is PGC-1β, it seems that the heart can at least in the short term tolerate a 50% reduction of PGC-1α and complete absence of PGC-1β, with relative preservation of LV function as was observed at 3-weeks. However changes in patterns of substrate utilization and evidence of oxidative stress were already present at this stage, implying that maladaptation as a result of reduced expression of these co-activators likely contributed to the subsequent development of heart failure. The time course for the development of overt heart failure in PGC-1α and PGC-1β KO hearts after TAC are remarkably similar, implying that there is reserve mitochondrial capacity that is capable of maintaining myocardial energetics and contractile function in the face of reduced expression or activity of these transcriptional co-activators. Moreover, our study underscores the potential role of additional mechanisms such as oxidative stress that is exacerbated by deficiency of these transcriptional co-activators that likely contributes to the accelerated transition to decompensated heart failure.
In conclusion, the present study describes the contribution of PGC–1β to mitochondrial function and gene expression in non-stressed hearts and identifies its contribution to maintaining contractile function under pathological increased workload. Thus, modulation of PGC-1 activity may represent a promising target for limiting the transition from pressure overload cardiac hypertrophy to heart failure.