Characterization of native FGF23 protein and mutant forms causing autosomal dominant hypophosphatemic rickets and familial tumoral calcinosis

The regulation of phosphate metabolism is a complex process that is
still only partly understood. At the end of the eighties, studies
in a mouse model for hypophosphatemic rickets provided evidence
that phosphate wasting could not be explained by a primary defect
of the kidney but rather by an unknown circulating factor with
phosphaturic properties. X-linked hypophosphatemia (XLH), autosomal
dominant hypophosphatemic rickets (ADHR), and tumor induced
osteomalacia (TIO) are three well defined human disorders of
isolated renal phosphate wasting. XLH and ADHR are mendelian
diseases while TIO is caused by rare, mostly benign tumors. The
opposite phenotype, hyperphosphatemia due to increased renal
phosphate reabsorption is associated to the recessive disorder
familial tumoral calcinosis (FTC). At the beginning of this work
the genes mutated in XLH and ADHR were cloned. One gene codes for
the endopeptidase PHEX, the other for the fibroblast growth factor
FGF23. Both proteins are probably involved in a novel common
pathway of the regulation of phosphate homeostasis. Missense
mutations in FGF23 causing phosphate wasting in patients with ADHR,
overexpression of FGF23 in tumors from patients with TIO, and the
observation that FGF23 plasma levels are elevated in most patients
with XLH provided strong evidence that FGF23 is a hormone with
phosphaturic activity. However, the function of FGF23 in the
regulation of phosphate metabolism is far from understood. The
intention of this study was to investigate the molecular properties
of native FGF23 and its mutant forms. I conducted protein
expression experiments in HEK293 cells which showed that native
FGF23 is a secreted protein partially processed into an N-terminal
fragment and a C-terminal fragment. I provided evidence that this
cleavage occurs during protein secretion and it is performed by
subtilisin like-proprotein convertases (SPCs). In addition, I
determined that native FGF23 undergoes O-linked glycosylation
before secretion by using a deglycosylation assay. Further, RT-PCR
analysis of human tissues showed FGF23 expression in whole fetus,
heart, liver, thyroid/parathyroid, small intestine, testis,
skeletal muscle, differentiated chondrocytes and TIO tumor tissue.
In mouse, FGF23 was expressed in day 17 embryo and spleen. The
FGF23 ADHR mutations replace arginine residues at the SPC cleavage
site (RXXR motif). By expression of the FGF23-R176Q and –R179Q
mutant proteins in HEK293 cells I showed that ADHR mutations
prevent cleavage at the RXXR site and stabilize FGF23. This
alteration in the FGF23 cleavage enhances FGF23 phosphaturic
activity in ADHR. Familial tumoral calcinosis (FTC) with
hyperphosphatemia is a disease considered the mirror image of the
hypophosphatemic condition. It is known that FTC is caused by
mutations in the GALNT3 gene. By performing mutation analysis in
two families with FTC, I could show that FTC can also be caused by
inactivating mutations in the FGF23 gene. To characterize the
FGF23-S71G mutant protein I conducted in vitro expression assays,
immunocytochemistry and ELISA to measure the FGF23 plasma levels in
the patient with FTC. Taken together the results of these
experiments showed that intact FGF23-S71G mutant protein remained
inside the cells and only the C-terminal FGF23 fragment was
secreted. These investigations demonstrate that FGF23 mutations in
ADHR and FTC have opposite effects on phosphate homeostasis. There
is evidence that the endopeptidase PHEX which is mutated in
patients with XLH and FGF23 act in the same pathway. PHEX function
resides upstream of FGF23 and may be involved in the degradation of
FGF23 thereby regulating its phosphaturic activity. I designed an
assay with a recombinant secreted form of PHEX (secPHEX) to prove
whether FGF23 is a substrate of PHEX. Although secPHEX activity
could be demonstrated by degradation of PTHrP107-139, secPHEX
failed to degrade FGF23 in this assay. These results provided
evidence against a direct interaction of PHEX and FGF23.