FC 018HIGH DIETARY PHOSPHATE INTAKE AND INTRA-CARDIAC SYNTHESIS OF FIBROBLAST GROWTH FACTOR 23 SYNERGISTICALLY WORSEN CARDIAC FUNCTION

Background and Aims Left ventricular hypertrophy (LVH) is a major complication of CKD and associates with increased levels of the phosphaturic hormone fibroblast growth factor (FGF) 23. FGF23 induces hypertrophic growth of cardiac myocytes in vitro and LVH in rodents, suggesting that FGF23 can directly affect the heart. Besides the bone, cardiac myocytes express FGF23, too, and recent studies demonstrate that its expression is increased in cardiac and kidney injury, suggesting that cardiotoxicity of FGF23 may be at least partly due to the paracrine effects of heart-derived FGF23. However, it is still questioned whether elevated FGF23 per se is able to induce pathologic alterations in the heart or whether additional factors in CKD, such as Klotho deficiency or hyperphosphatemia are required for FGF23 to tackle the heart. By generating a mouse model with cardiac-specific overexpression of FGF23 via myocardial gene transfer using adeno-associated virus (AAV), we elucidated the cardiotoxic properties of elevated FGF23 in (1) unchallenged mice, unbiased of alterations usually associated with CKD, and (2) in the presence of high dietary phosphate intake, mimicking the exposure of enhanced serum phosphate. Method First, an adeno-associated virus that expresses murine Fgf23 (AAV-Fgf23) under the control of the cardiac troponin T promotor was injected subcutaneously into eight-week-old male C57BL/6 wildtype mice. After four months, cardiac function and geometry was assessed by cardiac magnetic resonance imaging (MRI) and echocardiography and heart tissue was analysed by qPCR, immunoblot and histological analyses. The biological activity of AAV-Fgf23-derived cardiac Fgf23 was determined using isolated neonatal rat ventricular myocytes (NRVM) in vitro. Second, AAV-Fgf23 and control mice were fed a 2% high phosphate diet (HPD) or a 0.8% normal phosphate diet (NPD) and cardiac phenotype was investigated after six months. Results AAV-Fgf23 mice showed increased cardiac-specific Fgf23 expression and synthesis of intact Fgf23 (iFgf23) protein in the heart resulting in enhanced circulating iFgf23 compared to control. Serum of AAV-Fgf23 mice stimulated hypertrophic growth of isolated NRVM and induced pro-hypertrophic gene expression in vitro, indicating that cardiac iFgf23 is biologically active. Likewise, AAV-Fgf23 mice revealed an activation of renal FGFR1/Klotho/MAPK signalling and subsequent down-regulation of renal sodium phosphate transporters NaPi-2a and NaPi-2c, causing reduced tubular phosphate reabsorption. Nevertheless, in unchallenged AAV-Fgf23 mice, impaired cardiac function, LVH and LV fibrosis were lacking. In contrast, HPD stimulated the bone expression of Fgf23 in both AAV-Fgf23 and Ctrl mice, while intra-cardiac Fgf23 mRNA levels were only increased in both AAV-Fgf23 groups irrespective of NPD or HPD. However, HPD in AAV-Fgf23 mice promoted O-glycosylation of cardiac iFgf23, suggesting stabilization of biologically active Fgf23 protein. Echocardiography showed impaired cardiac function in AAV-Fgf23 on HPD compared to its NPD group, demonstrated by enhanced end-systolic and end-diastolic volumes, increased systolic and diastolic LV diameters as well as enlarged LV inner diameters, respectively. Pressure-volume analysis using Millar catheter showed higher end-systolic and end-diastolic blood pressure (ESP, EDP) in AAV-Fgf23 mice on HPD compared to NPD. HPD in Ctrl only enhanced EDP, although this did not reach the level of statistical significance. Conclusion Chronic exposure to biologically active cardiac iFgf23 per se does not tackle the heart, while high intra-cardiac Fgf23 synthesis in the presence of high dietary phosphate promotes cardiotoxicity of Fgf23, which could pose a significant health risk to the general population.