PTH and FGF23 Exert Interdependent Effects on Renal Phosphate Handling: Evidence From Patients With Hypoparathyroidism and Hyperphosphatemic Familial Tumoral Calcinosis Treated With Synthetic Human PTH 1–34

Abstract

Parathyroid hormone (PTH) and fibroblast growth factor 23 (FGF23) both influence blood phosphate levels by regulating urinary phosphate reabsorption. Clinical data suggest that adequate renal phosphate handling requires the presence of both FGF23 and PTH, but robust evidence is lacking. To investigate whether the phosphaturic effects of PTH and FGF23 are interdependent, 11 patients with hypoparathyroidism, which features high blood phosphate in spite of concomitant FGF23 elevation, and 1 patient with hyperphosphatemic familial tumoral calcinosis (HFTC), characterized by deficient intact FGF23 action and resulting hyperphosphatemia, were treated with synthetic human PTH 1–34 (hPTH 1–34). Biochemical parameters, including blood phosphate, calcium, intact FGF23 (iFGF23), nephrogenic cAMP, 1,25(OH)₂ vitamin D (1,25D), and tubular reabsorption of phosphate (TRP), were measured at baseline and after hPTH 1–34 treatment. In patients with hypoparathyroidism, administration of hPTH 1–34 increased nephrogenic cAMP, which resulted in serum phosphate normalization followed by a significant decrease in iFGF23. TRP initially decreased and returned to baseline. In the patient with HFTC, hPTH 1–34 administration also increased nephrogenic cAMP, but this did not produce changes in phosphate or TRP. No changes in calcium were observed in any of the studied patients, although prolonged hPTH 1–34 treatment did induce supraphysiologic 1,25D levels in the patient with HFTC. Our results indicate that PTH and FGF23 effects on phosphate regulation are interdependent and both are required to adequately regulate renal phosphate handling. Published 2021. This article is a U.S. Government work and is in the public domain in the USA.

Introduction

Phosphate homeostasis is regulated by a complex endocrine axis, primarily controlled by fibroblast growth factor 23 (FGF23) and parathyroid hormone (PTH).^(1,2) Both FGF23 and PTH exert their main phosphaturic action by acting on the same cells in the proximal renal tubule, leading to the internalization and deactivation of sodium-phosphate co-transporters (Na-Pi 2a and 2c) and subsequently lowering renal phosphate reabsorption.⁽³⁾ 1,25(OH)₂ vitamin D (1,25D) also has a role in phosphate regulation as it both mediates intestinal absorption of phosphate and calcium and stimulates FGF23 secretion.⁽⁴⁾ Although PTH and FGF23 share an agonistic phosphaturic effect, they have opposite regulatory actions on 1,25D; PTH increases and FGF23 decreases the 1-alpha-hydroxylation of 25-OH-vitamin D needed to produce 1,25D.⁽²⁾ There are also contradictory data regarding the effects of circulating calcium on renal phosphate handling; a recent clinical study suggests that it is inversely associated with TmP/GFR and that this effect is at least partially independent of PTH.⁽⁵⁾ Thus, evaluation of phosphate abnormalities requires a careful interpretation of an intricate interplay of factors.

Clinical observations of patients with rare disorders of PTH or FGF23 deficiency suggest that, in individuals with normal renal function, correction of hyperphosphatemia through phosphaturia may require the presence of both FGF23 and PTH. Patients with hypoparathyroidism typically exhibit increased renal phosphate reabsorption leading to hyperphosphatemia, despite elevated FGF23 levels. This suggests that renal tubule cells may be resistant to the phosphaturic effects of FGF23 in the absence of adequate PTH.⁽⁶⁾ In hyperphosphatemic familial tumoral calcinosis (HFTC), a genetic disorder characterized by a functional deficiency of FGF23, patients also exhibit hyperphosphatemia due to high renal phosphate reabsorption and high 1,25D. In this population, PTH levels are usually low-normal in response to increased intestinal calcium absorption and frank or relative hypercalcemia. It is not known if high PTH levels can induce phosphaturia independently without FGF23 in this population.⁽⁷⁾

To further investigate the interdependent effects of FGF23 and PTH on renal phosphate handling, we studied phosphate homeostasis in a group of patients with hypoparathyroidism and a patient with HFTC, before and after hPTH 1–34 therapy.

Subjects and Methods

Patients

Hypoparathyroidism cohort

Eleven patients with hypoparathyroidism, enrolled in a study investigating the effects of hPTH 1–34 therapy in hypoparathyroidism (NCT00395538), were evaluated at the National Institutes of Health (NIH). The duration of hypoparathyroidism was of at least 1 year. Patients were excluded if they were allergic to tetracyclines, had received prior synthetic human PTH therapy, had significant liver or kidney disease (estimated glomerular filtration rate <25 mL/min/1.73 m²), were affected with a disorder that could potentially alter mineral metabolism other than hypoparathyroidism, were pregnant, or had a history of chronic steroid or bisphosphonate use.

After an optimization period on conventional therapy (calcitriol and calcium supplements), patients were assigned to receive 0.2 mcg/kg/dose of hPTH 1–34 via subcutaneous injections b.i.d. Calcitriol was discontinued on the day before initiating hPTH 1–34.

Baseline testing, immediately before the first hPTH 1–34 injection, included routine blood and urine chemistries, iPTH (electrochemiluminescence immunoassay on Roche Cobas e601 analyzer; NIH, Bethesda, MD, USA), 1,25D (radioimmunoassay, Mayo Medical Laboratories, Rochester, MN, USA), intact FGF23 (iFGF23) (ELISA, Immutopics International, San Clemente, CA, USA), and nephrogenic cAMP (NcAMP = urine cAMP–serum cAMP normalized for creatinine NcAMP) (high-performance liquid chromatography, Mayo Medical Laboratories). Tubular reabsorption of phosphate (TRP), typically 85% to 90% in the setting of normal blood phosphate,⁽⁸⁾ was calculated using the following formula: TRP% = 100 × (1 – [(urine phosphate/urine creatinine) * (serum creatinine/serum phosphate)]). Blood calcium, phosphate, iFGF23, 1,25D, in addition to TRP and NcAMP were also measured 4, 8, 12, and 24 hours after the first injection. A second hPTH 1–34 injection was administered 12 hours after the first dose.

HFTC patient

A patient with HFTC was evaluated to analyze the potential effects of hPTH 1–34 on renal phosphate regulation in the setting of FGF23 deficiency. The patient had an extreme progressing and debilitating phenotype. PTH was administered after all maneuvers to lower blood phosphate, including a low-phosphate diet and the combined administration of sevelamer, acetazolamide, and probenecid, had failed. He was initially treated with a single dose of 40 mcg of hPTH 1–34 and 1 month after with 20 mcg b.i.d. for 6 days. The following parameters were measured at baseline and at 4, 8, 12, and 24 hours after a single subcutaneous hPTH 1–34 injection: blood calcium, phosphate, NcAMP, and TRP.PTH and 1,25D were measured only at baseline and 24 hours after administration. During the second hPTH 1–34 treatment course, blood calcium, phosphate, NcAMP, TRP, PTH, and 1,25D were measured at baseline and every 2 days (fasting morning samples). iFGF23 and C-terminal FGF23 (ELISA, Immutopics International) were measured before initiating hPTH 1–34 treatment.

The studies were approved by the Institutional Review Board of the National Institute of Dental and Craniofacial Research (NIDCR). Written informed consent was obtained from all patients.

Statistics

Age in hypoparathyroid patients is presented as median and range. Normality of sequentially measured variables was assessed through the Shapiro–Wilk test in hypoparathyroid patients. Comparisons of mean or mean rank were performed at each time point compared with baseline through one-way ANOVA or Kruskal–Wallis test, respectively, according to the variables’ distribution. Multiplicity adjusted p values were calculated through Dunnet’s (one-way ANOVA) or Dunn’s (Kruskal–Wallis) multiple comparison tests. Significance was set at a p value <0.05. Statistical analyses were performed, and figures were prepared, with Prism 8 (GraphPad Software, La Jolla, CA, USA).

Results

Baseline characteristics

All 11 patients with hypoparathyroidism were White. Median age was 42 years (range 20 to 59). All were female except one. Etiologies of hypoparathyroidism included postsurgical (n = 9) and idiopathic (n = 2). Mean ± SD PTH values in the cohort were 11.6 ± 5.8 pg/mL (normal range 15 to 65). The patient with HFTC was a 29-year-old Middle Eastern man who had a homozygous variant in GALNT3 (c.1584_1585 A insertion).⁽⁷⁾

Effects of hPTH 1–34 on iFGF23, phosphate, calcium, 1,25D, TRP, and NcAMP in hypoparathyroid patients

Four hours after the first hPTH 1–34 injection, there was a significant but transient increase in NcAMP. This was accompanied by a significant decrease in TRP. The nadir TRP value occurred 4 hours after the PTH injection, the first time point at which it was measured, rose thereafter, remained significantly lower than baseline at the 8-hour time point, and returned to baseline by 24 hours. These changes were associated with a significant decrease in blood phosphate at 4 hours, slight rise at 8 hours with a steady decrease to 24 hours, at which time it was significantly lower compared with baseline. iFGF23 was elevated at baseline and trended down over the 24-hour period until it was significantly lower than baseline and within the normal range. Calcium and 1,25D levels were unchanged throughout (Fig. 1). {FIG1}

Fig. 1.

Sequentially measured phosphate-associated biochemical parameters after hPTH1–34 treatment in hypoparathyroid patients including (A) nephrogenous cAMP (NcAMP), (B) tubular reabsorption of phosphate (TRP), (C) intact FGF23 (iFGF23), (D) blood phosphate, (E) 1,25(OH)₂ vitamin D (1,25D), and (F) blood calcium. Values are expressed as mean ± standard deviation (SD) (B and E) or as median ± interquartile (IQ) range (A, C, D, and F) according to variable distribution. The x axes display hours after the first hPTH 1–34 injection. Normal ranges are displayed in the gray areas between the dashed lines. A second hPTH1–34 dose was administered at 12 hours. Significant differences versus baseline (0 hour): *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Effect of hPTH 1–34 on phosphate, calcium, 1,25D, TRP, and NcAMP in a patient with HFTC

Before initiating hPTH 1–34 treatment, iFGF23 was 3 pg/mL, whereas C-terminal FGF23 was 545 RU/mL, corroborating the patient’s functional FGF23 deficiency due to excessive cleavage. The single 40 mcg hPTH 1–34 injection (Fig. 2A) {FIG2} led to an increase in NcAMP at 4 hours, which decreased thereafter, consistent with intact renal responsiveness to PTH. Despite this, neither phosphate nor TRP changed. PTH was elevated at baseline but normalized at 24 hours, whereas 1,25D did not fluctuate and remained within normality at both time points. During the second treatment course (Fig. 2B), phosphate and TRP exhibited a similar response with persistent hyperphosphatemia and ineffective renal phosphate excretion despite an evident and sustained NcAMP elevation. In contrast to the single dose, daily administration of hPTH 1–34 led to supraphysiological 1,25D levels. On the other hand, PTH values were normal throughout the second treatment course despite elevated 1,25D levels. Calcium remained within normality across all time points during both treatment courses (data not shown).

Fig. 2.

Sequentially measured phosphate-associated biochemical parameters after hPTH 1–34 treatment in a patient with hyperphosphatemic familial tumoral calcinosis. Evaluated parameters included: nephrogenous cAMP (NcAMP), blood phosphate, tubular reabsorption of phosphate (TRP), PTH, and 1,25(OH)₂ vitamin D (1,25D) measured during: (A) Twenty four hours after a single dose of 40 mcg hPTH1–34 and (B) every other day after 20 mcg b.i.d. for 6 days. Calcium remained within the normal range across all time points (data not shown). Boxed areas depict normal ranges. In both instances, PTH administration resulted in a marked increase in NcAMP, consistent with an intact renal response to PTH. Also, in both instances, although there was a suggestion of an effect on TRP, there was no clinically significant decrease in blood phosphate. Note that between the two PTH trials, blood phosphate had increased, in spite of maximal phosphate-lowering therapy. Furthermore, there was a marked increase in 1,25D in response to PTH, a clinically undesirable response further supporting that PTH treatment is not likely an effective treatment for HFTC.

Discussion

In an effort to better understand the interdependent physiology of the phosphaturic hormones PTH and FGF23 in phosphate homeostasis, we studied the phosphate dynamics in response to PTH in a group of patients with hypoparathyroidism and a patient with HFTC. As is typical, patients with hypoparathyroidism had relatively high blood phosphate levels in the setting of elevated iFGF23, consistent with a state of relative FGF23 resistance. The administration of hPTH resulted in a rapid and significant renal response to PTH evidenced by an increase in NcAMP and a marked although transitory decrease in TRP that ultimately resulted in a sustained and significant decrease in blood phosphate (Fig. 1). PTH administration overcame FGF23 resistance, suggesting the full phosphaturic effect of FGF23 is PTH-dependent and demonstrating the interdependence of PTH and FGF23 in phosphate physiology.

Consistent with the above observations and speculations, the patient with HFTC showed no response to PTH while he was in a low iFGF23 state (Fig. 2). The rise in NcAMP confirmed there was renal response to PTH but that PTH/cAMP signaling was insufficient to bring about a phosphaturic effect in the absence of sufficient FGF23.

Previous studies support the hypothesis that PTH is necessary for effective FGF23-mediated phosphaturia. McKenna and colleagues recently examined the contribution of different biochemical factors on renal phosphate handling in a cohort of patients with a variety of phosphate disturbances, including FGF23-mediated hypophosphatemia, chronic kidney disease (CKD), and hypoparathyroidism.⁽⁵⁾ Their study found that FGF23, PTH, and ionized calcium levels were all inversely correlated and exerted independent effects on TmP/GFR. PTH and FGF23 interaction was also assessed across a spectrum of PTH values ranging from deficient (patients with hypoparathyroidism) to elevated (patients with FGF23-mediated hypophosphatemia and CKD). Findings were in line with our results, concluding that FGF23’s phosphaturic effect was more potent in the presence of elevated PTH and was attenuated when PTH was low. Furthermore, studies that have evaluated cinacalcet, a calcium-sensing agonist, in patients with tumor-induced osteomalacia (TIO)⁽⁹⁾ and X-linked hypophosphatemic rickets (XLH)^(10,11) indicate that pharmacologically induced hypoparathyroidism in the setting of FGF23-mediated hypophosphatemia improves and even corrects phosphate tubular reabsorption in spite of concomitant elevated FGF23. Accordingly, iatrogenic hypoparathyroidism after total parathyroidectomy in patients with TIO and XLH produced analogous effects,^(12-14) while hyperphosphaturia returned in a patient with XLH and hypoparathyroidism when treated with synthetic PTH.⁽⁵⁾

Although these studies clearly demonstrate that PTH is required for FGF23-mediated phosphaturia, experimental data examining the role of FGF23, if any, on PTH-mediated renal phosphate action are relatively sparse and limited to animal models. Andrukhova and colleagues developed mice deficient in genes for both Fgf23 and the vitamin D receptor.⁽¹⁵⁾ Despite hypocalcemia with substantial PTH elevation, phosphaturia was absent and the mice exhibited hyperphosphatemia, suggesting that FGF23 is required, in addition to PTH, for effective renal phosphate excretion. Patients with HFTC, who have a functional FGF23 deficiency, are particularly informative in understanding this biochemical relationship. FGF23 deficiency results in hyperphosphatemia and elevated 1,25D with subsequent increased intestinal calcium absorption and low/low-normal PTH levels.^(7,16) By giving hPTH 1–34 to our patient with HFTC, an intervention not previously reported, we showed that raising PTH in the setting of low FGF23 did not increase phosphaturia or lower blood phosphate while it had the effect of stimulating 1,25D production. This finding, though limited to a single patient, strongly supports the premise that FGF23 is also required for PTH-mediated renal phosphate excretion.

The point at which cross-talk between FGF23 and PTH signaling pathways occurs is still unclear, although there is evidence to support it involves their common target: the Na⁺/H⁺ exchange regulatory cofactor 1 (NHERF-1). NHERF-1 is a scaffold protein that promotes NaPi-2 endocytic internalization and degradation when phosphorylated.^(17,18) At the proximal renal tubule cells, FGF23 binds to co-receptor alpha-Klotho and to FGF-tyrosine kinase receptors (mainly FGFR1), with downstream activation of the MAPK/ERK pathway leading to NHERF-1 phosphorylation by the serum/glucocorticoid-regulated kinase-1.⁽¹⁹⁾ PTH binds to the G-coupled protein receptor PTHR1, and through its downstream targets, protein kinases A and C, leads to phosphorylation of NHERF-1.⁽²⁰⁾ The underlying reason for these duplicative yet apparently necessary pathways remains elusive.

The direct role of extracellular calcium on renal phosphate handling is uncertain. Clearly, there is an indirect effect in which PTH suppression due to increased extracellular calcium leads to an increase in TmP/GFR. However, as previously mentioned, McKenna and colleagues found a PTH-independent effect of extracellular calcium on TmP/GFR.⁽⁵⁾ This same phenomenon was also observed in a relevant mouse model.⁽²¹⁾ In our study, neither circulating calcium nor 1,25D experienced significant changes over the first 24 hours after initiating hPTH 1–34. Thus, we can only attribute the changes observed in phosphate and FGF23 to a direct effect of hPTH 1–34 on the kidney.

This study has several limitations. The number of hypoparathyroid patients is small and only one patient with HFTC was evaluated. This small number precluded performing multivariable linear regression, which could better inform the contribution of each evaluated parameter to changes in blood phosphate. In addition, the intervention with hPTH 1–34 and subsequent assessment of phosphate-related biochemical parameters was performed for a short period of time. Therefore, inferences from our results should be taken with caution.

Nevertheless, there are important strengths. Although the number of hypoparathyroid patients was small, it was sufficient in that when the data from these patients were combined with that of the patient with HFTC, we were able to demonstrate, in a mechanistically unified fashion, FGF23/PTH interdependence for full FGF23 action. Importantly, all patients had normal kidney function, thus allowing for adequate interpretation of biochemical parameters without confounding factors due to renal insufficiency. Lastly, although we only evaluated a single patient with HFTC, this novel approach provided orthogonal confirmation of the interdependent relationship between FGF23 and PTH in phosphate homeostasis and that PTH treatment is not a viable option for HFTC treatment.

In conclusion, our study provides direct experimental evidence that FG23 and PTH exert interdependent phosphaturic effects on the proximal renal tubule and deepen our understanding of phosphate mineral homeostasis.

Data Availability Statement

Data available on request due to privacy/ethical restrictions