The discrepancy in cardiovascular risk between patients taking rosiglitazone or pioglitazone and lack of insight into the underlying causes elicited the present study in mice [16, 17]. Mouse heart tissue was used, since this allowed a comprehensive analysis of similarities and differences between rosiglitazone and pioglitazone. Bioinformatic analysis of the gene expression data identified regulatory pathways and upstream regulators that are specifically influenced by rosiglitazone.
In a high-fat diet-inducible model of combined hyperglycemia and cardiovascular disease we found that rosiglitazone and pioglitazone are effective glucose lowering drugs which markedly improved systemic markers of glucose homeostasis, fasting plasma glucose and insulin, and the urinary excretion of albumin. These findings are in line with observations made in humans [18, 19] and other rodent models [20, 21]. Importantly, only rosiglitazone treatment caused cardiac hypertrophy, defined as a significant increase in heart weight, and tended to increase atherosclerosis in the aortic root area. To our knowledge the disparate capacity of rosiglitazone and pioglitazone to induce cardiac hypertrophy was never studied before in a diet-induced mouse model. In general, two forms of cardiac hypertrophy can be distinguished, physiological and pathological hypertrophy. The latter is a significant independent risk factor for cardiovascular mortality and morbidity (, and references therein). Our study demonstrate that validated markers for pathological cardiac hypertrophy, ANP and BNP , were elevated by rosiglitazone but not by pioglitazone. This demonstrates that only rosiglitazone induced pathological cardiac hypertrophy in these mice within the study period of 7 weeks. Consistent with our observation in mice, rosiglitazone treatment also increased BNP plasma levels in T2DM patients without previous signs of cardiovascular disease , while pioglitazone did not have an effect on BNP levels .
Microarray technology is a powerful technique to analyze the effect of interventions on thousands of genes and across pathways . Transcriptome analysis of the heart tissue revealed that, under conditions at which rosiglitazone and pioglitazone showed comparable hypoglycemic effects, 624 genes were significantly and differently altered by rosiglitazone. Only 31 of these are reportedly controlled by PPARγ. Furthermore, our analysis demonstrates that 354 genes are affected by pioglitazone, only 8 genes of which are target of PPARγ and all are in common with rosiglitazone. This implies, that in case the pathological cardiac hypertrophy is induced through a PPARγ-dependent mechanism, it could be caused by 31 rosiglitazone-modulated genes. However, the role of PPARγ in the development of pathological cardiac hypertrophy and in mediating the effect of rosiglitazone thereupon is not entirely clear yet. Several studies suggest that PPARγ is protective, as mice lacking PPARγ in the heart developed cardiac hypertrophy and dysfunction [26, 27] and treatment with a PPARγ agonist reduced cardiac remodeling and fibrosis in a rat model of hypertension . Yet, cardiac-specific over-expression of PPARγ in mice also resulted in cardiac dysfunction . Another study demonstrated that transgenic mice with enhanced PPARγ activity developed concentric hypertrophy which progressed to dilated cardiomyopathy . On the other hand, a study using cardiomyocyte-specific PPARγ knock-out mice indicated that rosiglitazone can promote the development of myocardial hypertrophy in a PPARγ -independent manner .
The heart relies on a constant high supply of energy, primarily met by the β-oxidation of fatty acids, to maintain the continuous contractile activity . There is increasing evidence that a shift in cardiac energy metabolism towards decreased fatty acid oxidation and increased glucose utilization, can contribute to the development of pathological heart hypertrophy and failure . Transcriptome analysis of the heart revealed that such a switch from fatty acid to glycolytic metabolism may also underlie the rosiglitazone-associated pathological hypertrophic cardiac phenotype in the current study. Rosiglitazone, but not pioglitazone, down regulated PPARα and PGC1α target genes. In PPARα-/- mice, cardiac hypertrophy is induced , and also in mice lacking PGC-1α, cardiac dysfunction becomes evident . Functional enrichment analysis of the genes specifically affected by rosiglitazone indicates that the down regulated genes are related to biological processes ‘Lipid Metabolism’ and ‘Energy Production’ and included CPT1, VLCAD, and LCAD. Inhibition of CPT1 in cardiac tissue has been demonstrated to induce cardiac hypertrophy [33, 34]. Similarly, cardiac hypertrophy was found in mice deficient in VLCAD or LCAD . Further evidence for a putative role of PPARα in cardiac hypertrophy comes from studies on fibroblast growth factor 21 (FGF21), which is expressed in and released by cardiomyocytes through a PPARα - dependent mechanism . Deficiency of FGF21 in the heart was demonstrated to induce of cardiac hypertrophy markers and reduce fatty acid oxidation .