Kidney glycolysis serves as a mammalian phosphate sensor that maintains phosphate homeostasis

Abstract

How phosphate levels are detected in mammals is unknown. The bone-derived hormone fibroblast growth factor 23 (FGF23) lowers blood phosphate by reducing kidney phosphate reabsorption and 1,25(OH)2D production, but phosphate does not directly stimulate bone FGF23 expression. Using PET scanning and LC-MS, we found that phosphate increases kidney-specific glycolysis and synthesis of glycerol-3-phosphate (G-3-P), which then circulates to bone to trigger FGF23 production. Further, we found that glycerol-3-phosphate dehydrogenase 1 (Gpd1), a cytosolic enzyme that synthesizes G-3-P and oxidizes NADH to NAD+, is required for phosphate-stimulated G-3-P and FGF23 production and prevention of hyperphosphatemia. In proximal tubule cells, we found that phosphate availability is substrate-limiting for glycolysis and G-3-P production, and that increased glycolysis and Gpd1 activity are coupled through cytosolic NAD+ recycling. Finally, we show that the type II sodium-dependent phosphate cotransporter Npt2a, which is primarily expressed in the proximal tubule, conferred kidney specificity to phosphate-stimulated G-3-P production. Importantly, exogenous G-3-P stimulated FGF23 production when Npt2a or Gpd1 are absent, confirming that it is the key circulating factor downstream of glycolytic phosphate sensing in the kidney. Together, these findings place glycolysis at the nexus of mineral and energy metabolism and identify a kidney-bone feedback loop that controls phosphate homeostasis.

Renal proximal tubule cells: power and finesse

“…In this setting, recycling of NAD + in the cytosol, by GPD1 and LDH, becomes absolutely essential because oxidative phosphorylation is impaired. In contrast, how might the mechanism of phosphate sensing described by Zhou et al ( 12 ) map to CKD, in which NPT2a and proximal tubular phosphate reabsorption are downregulated and FGF23 levels are constitutively elevated? Perhaps the switch from β-oxidation of fatty acids to glycolysis as the proximal tubule’s primary energy source in CKD ( 22 ) results in sufficient ongoing G-3-P production to continuously stimulate FGF23 production.…”

“…The quest to identify a specific protein akin to the calcium-sensing receptor as the mammalian phosphate sensor has been futile to date. In another paper in the JCI , Zhou et al ( 12 ) discover that proximal tubular glycolysis — a metabolic pathway, not a discrete protein — acts as a phosphate sensor in the kidney.…”

“…Great studies answer critical questions and simultaneously prompt numerous exciting new ones. Zhou et al ( 12 ) solved the previously vexing question of how the bone seems to know to make more FGF23 in response to dietary-phosphate loading without an intermediate change in serum phosphate ( 13 ). The authors also address the previous paradox that FGF23 regulates serum phosphate, but increases in serum phosphate itself do not directly stimulate FGF23 production.…”

“…Indeed, other investigators use calcium, or 1,25-vitamin D, but not phosphate, to stimulate FGF23 expression in vitro ( 14 ). However, the Zhou et al study ( 12 ) did not address whether the proximal tubular G-3-P response was dependent on the phosphate concentration in the tubular lumen or the absolute amount of phosphate entering the proximal tubular cell. The human experiment reported by Zhou et al involved an enormous 1.5-gram oral bolus of phosphate (and the mouse studies used large intravenous doses) that briefly increased the serum phosphate and, thus, the phosphate concentration in the glomerular filtrate ( 12 ).…”

“…Zhou et al ( 12 ) build on a prior discovery by their research team related to increased renal production of G-3-P as the kidney’s distress signal to inform the bone to increase FGF23 levels in ischemic acute kidney injury ( 21 ). In this setting, recycling of NAD + in the cytosol, by GPD1 and LDH, becomes absolutely essential because oxidative phosphorylation is impaired.…”