Abstract
Background and Aims
Mineral metabolism imbalances and inflammation are related to the development of vascular calcification (VC). Fibroblast growth factor-23 (FGF23) is the main regulator of phosphate homeostasis and various studies have shown the existence of an association between elevated levels of FGF23 and the appearance of cardiovascular disease (CVD). We conducted a case-control study to test the hypothesis that serum and vascular levels of FGF23 are associated with the presence of VC. In addition, we determined the influence of inflammation in these levels.
Method
One hundred and thirty-three patients diagnosed with clinical atherosclerotic disease undergoing elective vascular surgery, and 20 cadaveric organ donors with no medical history of CVD, were included in this study. Serum levels of intact FGF23 (iFGF23), together with tumor necrosis factor-alpha (TNFα), interleukin (IL)10, were determined by ELISA. Vascular fragments of aorta, carotid and femoral arteries were obtained for assaying the gene expression of FGF23, TNF, IL10 and RUNX2 by qPCR. Immunohistochemical procedures were employed to determine vascular protein levels of FGF23, TNFα and IL10. VC was diagnosed by imaging techniques and confirmed by histological procedures including von Kossa staining.
Results
Case group presented a higher prevalence of hypertension, hypercholesterolemia and reduced estimated glomerular filtration rate, with no differences regarding other parameters. Serum iFGF23 and TNFα/IL10 ratio were higher in the case group (P<0.01 and P<0.001, respectively). Vascular expression of FGF23 was detected in 58.6% of CVD patients vs 35% of donors, with mean expression levels significantly higher in the first group (P<0.01). Vascular expression of TNF/IL10 was also increased (P<0.001) in CVD patients. FGF23 immunoreactivity was detected in 84% of CVD patients and only in 35% of controls. Immunoreactivity for FGF23 and TNFα/IL10 ratio were significantly higher in CVD patients (P<0.001 and P<0.0001, respectively).
Stratified analysis according to serum iFGF23 levels showed a higher prevalence of VC in the upper tertiles. Patients with VC presented increased levels of all the FGF23 variables including serum [1.5 (1.2-1.6) vs. 1.4 (0.9-1.5) pg/mL, P<0.01], vascular mRNA [26.1 (14.3-67.4) vs. 18.8 (8.8-312.9) log AU, P<0.01] and vascular immunoreactivity [4.6 (3.8-4.9) vs. 3.7 (3.1-4.1) log µm2, P<0.05]. Moreover, FGF23 immunoreactivity was detected in 92.3% of fragments with VC and only in 53.6% of those without VC. Serum TNFα/IL10 and RUNX2 mRNA levels were also higher in this group (P<0.01 for both).
Correlation analysis showed associations of serum iFGF23 with serum TNFα (r=0.375, P<0.001), neutrophil/lymphocyte (r=0.142, P<0.05), vascular RUNX2 mRNA (r=0.55, P<0.05), and vascular FGF23 immunoreactivity (r=0.281, P<0.05) in the CVD group. Vascular FGF23 expression correlated with RUNX2 mRNA (r=0.315, P<0.05) and FGF23 immunoreactivity (r=0.254, P<0.05). Multiple regression analysis showed that iFGF23 levels were determined by UAE, HDL, FGe, calcium and TNFα levels (adjusted R2= 0.473, P<0.0001) and that vascular FGF23 mRNA expression was determined by TNFα, PCR, glucose and age (adjusted R2= 0.795, P<0.0001). Multivariate logistic regression, with VC as dependent variable, showed that both iFGF23 and vascular mRNA constitute independent risk factors for the existence of VC [OR (95% CI): 1.05 and 1.12, P<0.05 for both].
Conclusion
Patients with atherosclerosis and VC present significantly higher serum concentrations of FGF23, as well as higher immunoreactivity and gene expression levels in the vascular wall compared to patients without VC. Moreover, both serum and vascular mRNA levels of FGF23 are associated with the inflammatory status. Whether the increase in systemic and vascular FGF23 can directly promote or favor the calcifying process in the vascular bed is currently an issue under discussion.