We have created a transgenic model that allows the regulated expression of constitutively active HIF-1α in the adult myocardium. These animals allow us to identify the genes that are regulated by HIF in vivo, as well as the resulting physiologic, metabolic, and anatomic effects of HIF expression.
During HIF-1 expression in adult transgenic mice, ventricular function deteriorated and the heart dilated. Importantly, ventricular function returned to normal 7 days after transgene expression was terminated, indicating that the decrease in ventricular function was reversible and not related to irrevocable myocyte death. A recent study utilizing a cardiac-specific VHL KO −/− mouse also demonstrated a decline in ventricular function, manifesting at three months of age, that was prevented by an additional deletion of HIF-1α 
. Taken together, our findings and these support the direct role of HIF-1α transcriptional activity in ventricular dysfunction, although it is certainly possible that the other study reflects activities of HIF during development or growth, that impair contractility by a distinct mechanism.
Our model allows us to examine the earliest gene regulation induced by HIF in vivo. These early genes are likely to be directly regulated by HIF, and thus provide insight into the pathways that transduce HIF effects. As seen in previous in vitro studies 
, the mRNAs for several glycolytic enzymes, particularly Eno1 and Pfk, were up-regulated. HIF-1α-PPN expression also induced indicators of cellular proliferation, including Mki67 and Mcmd6 mRNAs. These transcripts may be related to the observed angiogenic response, occurring in the endothelial cells and pericytes recruited to new vessels. Several cancer-related transcripts were found to be up-regulated as well, including the mRNA for actin-bundling protein Fscn-1 (reviewed in 
). Bcl-6, a homolog of the up-regulated Bazf gene, has been found to be up-regulated in breast cancer and associated with higher HIF-1α levels 
. Many of the HIF-regulated genes in this study have not been previously recognized to be related to hypoxia, and their putative roles in the hypoxic response will require more evaluation.
The HIF-induced changes in mRNA abundance that we observe are consistent with the anticipated shift from oxidative phosphorylation to glycolytic metabolism. Importantly, we also see a reduction in the abundance of the mRNAs for key components of excitation-contraction coupling. In particular, we have observed reduced abundance of the mRNA for the SERCA2a calcium pump, as well as transcripts for PLB and RYR2. Down-regulation of these three genes would be predicted to produce a reduction in contractility in the heart due to a reduced calcium gradient across the SR membrane, and a reduced release of calcium upon initiation of contraction through the RYR2 channel. We propose that the heart cannot maintain normal contractility because of reduced expression of the SERCA pump, with resulting slowed re-uptake of cytoplasmic Ca++. Our direct measurement of calcium flux supports this mechanism.
In human hearts there is strong evidence for down-regulation of SERCA in ischemia
and heart failure
. There is an ongoing effort to develop methods to treat heart failure by modulating SERCA and other components of Ca regulation
. Our data further emphasizes the importance of reduced SERCA as a primary effector of cardiac contractility, and therefore a worthy target for such strategies. It is interesting in this regard that the gene for SERCA has several potential hypoxia response elements in its promoter and that the response to the single stimulus of HIF induction is rapid. Both observations suggest that HIF-may be directly regulating SERCA, although this promoter was not identified among the ChIP positive fragments.
The paradoxical change in the ATP:ADP ratio suggests that the upregulated glycolytic pathway is able to maintain ATP stores. Reduced calcium cycling from the sarcoplasmic reticulum, which we expect with reduced SERCA, could be responsible for both reducing contractility and maintaining the ATP:ADP ratio by reducing ATP consumption during the cardiac cycle.
In patients with ischemic cardiomyopathy, PET studies have demonstrated that hibernating myocardium, defined as areas of the heart with reduced blood flow but maintained glucose utilization, is less common than “stunned” myocardium with preserved blood flow. Nonetheless, 9% of dysfunctional segments in such patients show the reduced perfusion, (with preserved glucose metabolism), that would be expected to lead to chronic HIF stabilization 
. Parenthetically, it may be that the shift to glycolysis expected with HIF action is playing a role in preserving glucose utilization in these segments of underperfused heart. If HIF is stabilized in these segments it could be responsible for the impaired contractility, as seen in our mouse model 
Our findings suggest that a portion of the reduced contractility in ischemic cardiomyopathy could be directed by HIF. Our model allows us to study HIF in isolation from other transcription factors that may be active during an ischemic insult. The data presented here suggest that HIF, acting alone upon the heart, causes cardiac dysfunction. The mechanism is at least in part mediated by reduction in calcium reuptake into the sarcoplasmic reticulum.
A limitation of our approach is that we have examined the action of HIF in normoxic tissue, whereas the authentic in vivo action of HIF will occur predominately in hypoxic tissue where redox potential, pH, and substrate utilization may modify the transcriptional activity of HIF. However, the in vivo effects that we see are almost certainly more authentic than those obtained in cell culture, and seem likely to accurately recapitulate effects of human cardiac ischemia that occurs in mature adults.
The principal finding of this study, the unexpected, reversible, decrement in ventricular function, must be considered in any effort at HIF-1α-directed therapy. Perhaps of even greater clinical significance, this finding suggests a novel mechanism for ischemic cardiomyopathy. If persistent ischemia produces chronic HIF stabilization in even a portion of the ischemic heart, then the action of this transcription factor, perhaps acting in trans through downstream intra-cardiac signaling pathways, could directly impair ventricular performance. An exciting ramification of this hypothesis is that new therapy directed at HIF action could ameliorate this dysfunction.