• Open Access

Fibroblast Growth Factor 23 (FGF-23) Concentrations in Cats with Early Nonazotemic Chronic Kidney Disease (CKD) and in Healthy Geriatric Cats

Authors


Corresponding author: N.C. Finch, School of Veterinary Sciences, University of Bristol, UK. e-mail: natalie.finch@bristol.ac.uk.

Abstract

Background

Fibroblast growth factor (FGF-23) has an important role in phosphate regulation. Its clinical relevance in cats with CKD has not been explored previously.

Hypothesis/Objectives

The study objectives were (1) to determine whether FGF-23 concentrations are increased in nonazotemic cats, cats which developed azotemia within 12 months of screening compared with cats that remained non-azotemic, and (2) to evaluate the relationships between FGF-23 and PTH and FGF-23 and glomerular filtration rate (GFR).

Animals

Sixty-two healthy client-owned geriatric cats, 14 of which developed azotemia during the 12-month follow-up period.

Methods

Healthy nonazotemic cats were recruited prospectively into the study and followed for 12 months. At the study end-point, cats were categorized into 3 groups according to plasma creatinine concentration. PTH, FGF-23, and additional biochemical variables were evaluated at baseline and after 12 months. GFR was measured by a corrected slope-intercept iohexol clearance method.

Results

FGF-23 concentrations at baseline were found to be significantly increased in cats that developed azotemia (P = .001) compared with cats that did not develop azotemia. A significant positive relationship was identified between FGF-23 and PTH, whereas the relationship between FGF-23 and GFR was negative.

Conclusions and Clinical Importance

FGF-23 concentrations predicted development of azotemia in geriatric cats. Positive relationships between FGF-23 and PTH suggest an association between FGF-23 and renal secondary hyperparathyroidism.

Abbreviations
CKD

chronic kidney disease

FGF-23

fibroblast growth factor 23

GFR

glomerular filtration rate

IRMA

immunoradiometric assay

PTH

parathyroid hormone

Introduction

Fibroblast growth factor 23 (FGF-23) is a phosphatonin that is synthesized and secreted by osteoblasts and osteocytes. It was first identified in human patients with disturbed phosphate homeostasis. FGF-23 has since been shown to have an important physiological role in the regulation of phosphate and vitamin D metabolism. Actions of FGF-23 include inhibition of sodium-dependent phosphate reabsorption in the proximal tubule and suppression of calcitriol (active vitamin D) production by inhibition of 1-α-hydroxylase activity in the kidney.[1, 2] Calcitriol and dietary phosphate loading increase FGF-23 production in healthy human patients.[1, 3]

FGF-23 concentrations have been shown to be increased in several studies of human patients with CKD. This increase may be a compensatory mechanism secondary to disturbed phosphate homeostasis or may be caused by the development of an FGF-23 resistant state.[4-7] Alternatively, FGF-23 may simply be a marker of GFR reflecting decreased renal clearance of the hormone in patients with chronic kidney disease (CKD).[8, 9]

The phosphaturic actions of FGF-23 are similar to those of PTH, but the relationship between the 2 proteins remains unclear. FGF-23 is increased in renal secondary hyperparathyroidism and positive correlations between PTH and FGF-23 have been demonstrated in human patients with CKD.[3, 8, 10] This increase may be related to an FGF-23-mediated reduction in calcitriol, which is an important negative regulator of PTH. On the other hand, FGF-23 decreases PTH secretion by an action mediated by the FGF receptor complex expressed on parathyroid glands.[2, 11, 12] However, in vivo animal models with overexpression of FGF-23 are associated with parathyroid gland hyperplasia and increased PTH concentrations, although these findings may be confounded by calcitriol deficits.[13] In human patients with renal secondary hyperparathyroidism, expression of FGF-23 receptors in the parathyroid gland was decreased.[14] This may explain why increased FGF-23 concentrations do not prevent an increase in PTH concentration. PTH may increase FGF-23 production by stimulation of osteoblasts, but the role of calcitriol as a potential confounder in these studies requires consideration.[15, 16]

In a previous study, PTH was found to be increased in cats in the early nonazotemic stages of CKD compared with cats that remained nonazotemic.[17] This increase preceded development of hyperphosphatemia and hypocalcemia. A previous study identified increased PTH in the later stages of CKD.[18] The point at which calcitriol deficiency plays a role in the development of renal secondary hyperparathyroidism in cats remains unclear. Recognition of factors contributing to increasing plasma PTH concentrations may have clinical importance and consequences in the approach to the management of renal secondary hyperparathyroidism, because when parathyroid gland hyperplasia occurs, the ability of standard treatments to suppress PTH synthesis and secretion is decreased.

C-terminal FGF-23 fragments, which are not biologically active, may accumulate in patients with CKD. These fragments may compete with and inhibit the actions of the full-length hormone. Assays that measure the intact active FGF-23 protein are considered to be more accurate.[9] An enzyme-linked immunosorbent assay (ELISA) previously has been developed and validated for measurement of intact FGF-23 in human serum samples.[19] An assay for FGF-23 has been validated in the cat, and FGF-23 is increased in cats with azotemic CKD.[20]

The objectives of this study were first to examine the hypothesis that FGF-23 is increased in the early nonazotemic stage of feline CKD and second to investigate the relationships between FGF-23 and PTH and GFR.

Methods

Prospective Study

Healthy nonazotemic (plasma creatinine concentrations <2.0 mg/dL) geriatric cats (>9 years) were recruited from 2 London-based first opinion practices into a prospective longitudinal study. Cats were followed over 12 months and at the study end-point (12 months) categorized according to their plasma creatinine concentration: group 1—plasma creatinine concentration ≤1.58 mg/dL; group 2—plasma creatinine concentration >1.58 mg/dL and ≤2.00 mg/dL or plasma creatinine concentration >2.00 mg/dL and a USG > 1.035; group 3—plasma creatinine concentration >2.0 mg/dL in association with decreased urinary concentrating ability (USG < 1.035).

Blood Sampling and Laboratory Evaluation

Owner consent was obtained before blood sampling, and the study was approved by the ethics and welfare committee of the Royal Veterinary College. Cats were excluded from the study if they had any concurrent medical disorders (except hypertension), evidence of renal lymphoma, were fed a protein- and phosphate-restricted diet, or were receiving treatment with drugs known to affect calcium or phosphate homeostasis. Blood samples were collected by jugular venipuncture into heparinized and EDTA tubes. Plasma was separated, harvested, and transferred to storage at −80°C within a few hours of collection. Heparinized plasma samples were analyzed for routine biochemical variables at a commercial laboratory.1 EDTA plasma was assayed at a commercial laboratory1 for intact PTH concentrations by an immunoradiometric assay (IRMA).2 In addition, intact FGF-23 was measured in EDTA plasma samples by an ELISA3 according to the manufacturer's protocol. This is a second-generation 2-site ELISA that recognizes only the biologically active, intact FGF-23 protein validated for use in the cat.[20]

FGF-23 and GFR

Glomerular filtration rate was measured in an additional 19 geriatric cats with variable renal function by a corrected slope-intercept plasma iohexol clearance method as described previosuly.[21] In addition, measurement of routine biochemical variables and FGF-23 was performed in these cats, but PTH was not measured.

Statistical Methods

Statistical analyses were performed by a statistical software package.4 Data were assessed for normality by the Kolmogorov-Smirnov test and by visual inspection of graphical plots. The assumption of Gaussian distribution was not met and therefore nonparametric testing was used. The following plasma variables were compared at baseline and the 12-month time point between groups by the Kruskal-Wallis test: FGF-23, PTH as well as total calcium, phosphate, and creatinine concentrations. In addition, the calcium × phosphate product and age also were compared between groups. Post hoc testing by the Mann-Whitney U-test with a Bonferroni correction was applied to identify where significant differences lay. Correlations between variables were evaluated by Spearman's correlation coefficient. Significance was set at < .05.

Results

Prospective Study

Sixty-two cats were included in the study: 15 cats in group 1, 33 cats in group 2, and 14 cats in group 3. The following variables were found to be significantly different between the groups at baseline: FGF-23 (= .003) and plasma creatinine concentration (< .001). At the 12-month time point, the following variables were found to be significantly different between the groups: FGF-23 (P < .001) and plasma creatinine concentration (P < .001). The FGF-23 concentrations for the groups at baseline and the 12-month time point are presented in Figure 1a,b. The results for the remaining variables are presented in Table 1.

Table 1. Variables involved in calcium homeostasis, plasma creatinine concentrations, and age at baseline and the 12-month time point in normal healthy cats (groups 1 and 2) and cats that developed azotemia over 12 months (group 3). Data are presented as median (range)
 Group 1Group 2Group 3P Value
  1. a

    Significantly different to group 1.

  2. b

    Significantly different to group 2 (P < .017).

N153314 
PTH baseline (pg/mL)50.5 (10.4–124.0)44.2 (9.0–371.7)84.0 (13.9–3269.2).359
PTH 12 months (pg/mL)67.2 (11.0–206.8)47.18 (16.5–715.5)177.2 (12.6–4425.0).164
Total calcium baseline (mg/dL)10.0 (8.9–10.8)10.0 (9.0–11.4)9.6 (9.0–10.4).066
Total calcium 12 months (mg/dL)9.9 (8.4–10.4)9.8 (8.9–11.00)9.8 (8.3–11.0).895
Phosphate baseline (mg/dL)4.0 (2.7–5.9)4.3 (3.0–6.2)4.5 (2.7–6.2).226
Phosphate 12 months (mg/dL)3.9 (2.8–5.)4.1 (2.8–5.4)4.5 (2.7–10.5).735
Ca × phos baseline (mg/dL)40.9 (27.4–61.6)42.8 (27.8–57.9)42.3 (24.4–63.9).578
Ca × phos 12 months (mg/dL)38.4 (26.7–49.3)39.9 (26.8–54.9)45.4 (27.4–87.3).677
Creatinine baseline (mg/dL)1.4 (1.0–1.6)1.6 (1.4–2.0)a1.8 (1.3–2.0)a<.001
Creatinine 12 months (mg/dL)1.4 (1.0–1.5)1.7 (1.6–2.4)a2.25 (2.0–2.8)a,b<.001
Age baseline (years)14.5 (9.5–19.4)13.0 (9.0–18.4)16.0 (11.0–18.1).090
Figure 1.

FGF-23 concentrations at (a) baseline and (b) the 12-month time point. Cats were categorized according to plasma creatinine concentrations over 12 months. Group 1—plasma creatinine concentration ≤1.58 mg/dL, group 2—plasma creatinine concentration >1.58 mg/dL and ≤2.00 mg/dL or a plasma creatinine concentration >2.00 mg/dL and a USG > 1.035, group 3—plasma creatinine concentration >2.0 mg/dL in association with a decreased urinary concentrating ability (USG < 1.035). The box and whiskers represent median and quartiles, respectively. The filled circles (●) and stars (*) represent outliers. The upward arrow represents a single outlier with very high concentration of FGF-23.

Fibroblast growth factor 23 was positively correlated at baseline with the following variables: PTH (r = 0.376, P = .003) and plasma creatinine concentration (r = 0.380, P = .002). At the 12-month time point, FGF-23 was positively correlated with PTH (r = 0.374, P = .003), plasma creatinine concentration (r = 0.558, P < .001), and calcium × phosphate product (r = 0.273, P = .032). The relationship between FGF-23 and PTH at baseline is presented in Figure 2.

Figure 2.

Relationship between FGF-23 and PTH concentrations in 62 nonazotemic cats at baseline. There is a positive linear relationship (r = 0.376, P = .003).

FGF-23 and GFR

The median (range) for GFR measurement was 1.53 (0.38–3.25) mL/min/kg. FGF-23 was negatively correlated with GFR (r = −0.472, P = .041). The relationship between FGF-23 and GFR is presented in Figure 3.

Figure 3.

Relationship between FGF and GFR in 19 cats with variable renal function. GFR was measured by a corrected slope-intercept method using the marker iohexol. A negative exponential relationship was identified (r = −0.472, P = .041).

Discussion

Results of the present prospective study indicate that FGF-23 is increased in geriatric cats in the early nonazotemic stage of CKD compared with cats that remained nonazotemic. The physiological stimulus for FGF-23 secretion may be phosphate, and the phosphaturic effects of FGF-23 suggest it may defend against phosphate overload in the face of decreased GFR.[22] FGF-23 production may increase in response to phosphate retention, which develops when renal function declines in CKD patients. In early CKD, there are sufficient functioning nephrons to respond to increased FGF-23, increase phosphate excretion, and maintain normal plasma phosphate concentration. In the later stages of disease, a decrease in nephron number leads to decreased capacity to excrete phosphate. In rat models of mild progressive CKD, normal phosphate and low calcitriol concentrations and increased fractional excretion of phosphate were demonstrated.[23] After inhibition of FGF-23 activity in these rats, fractional excretion of phosphate was decreased, serum phosphate concentration increased, and normal calcitriol concentration was restored. These findings suggest that phosphate retention is prevented by compensatory FGF-23 actions in mild CKD. In this study, FGF-23 was not found to be significantly correlated with phosphate at either baseline (P = .157) or the 12-month time point (P = .144). This study included a limited number of cats, and larger studies of human patients with CKD have shown significant correlations with phosphate concentration.[24, 25] Furthermore, FGF-23 concentrations are higher in cats matched for plasma creatinine concentration with higher phosphate concentrations.[20] Interestingly, studies of human patients with normal renal function found FGF-23 to be correlated only with urinary excretion and tubular reabsorption of phosphate and not with serum phosphate concentration.[22] This finding may offer further explanation for the lack of any significant correlation between FGF-23 and plasma phosphate concentration in this study given that 77% (48/62) of cats included in the study were considered to have normal renal function. Studies in which cats are fed a standardized diet to control phosphate intake and daily urinary excretion of phosphate is measured would be required to assess the relationship between FGF-23 concentration and renal handling of phosphate at different levels of renal function. Other investigators have suggested that the increase in FGF-23 concentrations in human CKD patients may be stimulated not only by phosphate retention but also by renal injury itself.[26]

A role for decreased calcitriol concentrations in the development of renal secondary hyperparathyroidism in early CKD is becoming increasingly recognized.[27-29] However, this role currently remains unproven in cats. FGF-23 suppresses 1-α-hydroxylase activity leading to decreased production of calcitriol in the kidney.[30] Clinical studies of human patients found that increases in FGF-23 preceded decreases in calcitriol concentrations.[31] After IV injection of FGF-23 into mice, serum phosphate concentration was decreased within 8 h of administration. Calcitriol concentrations, however, were decreased within 2 h, suggesting that the primary role of FGF-23 may be the immediate control of calcitriol with a secondary delayed effect on regulation of phosphate.[30] Furthermore, substantially lower concentrations of FGF-23 are required to decrease calcitriol concentrations as compared with serum phosphate concentration. Decreased calcitriol concentrations traditionally have been attributed to loss of functioning renal mass and decreased 1-α-hydroxylase activity. Concurrent endocrine complications such as decreased production of erythropoietin also may be expected to occur in the early stages of CKD if insufficient functional renal mass was contributing to calcitriol deficiency, although the site of production of erythropoietin (peritubular fibroblasts) differs from that of calcitriol (proximal tubular cells), and other factors such as hypoxia inducible factor (HIF) affect erythropoietin concentrations. Nevertheless, both anemia and decreased calcitriol concentration prevalence have been evaluated in human patients with variable stages of CKD, and the prevalence of calcitriol deficiency was found to be higher than the prevalence of anemia at all stages.[31] This observation suggests that calcitriol deficiency is present when there is sufficient functioning renal mass to maintain erythropoietin production. Furthermore, nephrectomized rat models of renal failure indicate that 1-α-hydroxylase expression is not decreased as compared with control rats.[32] Increased FGF-23 concentrations could contribute to the early development of calcitriol deficiency. This could not be substantiated in this study because calcitriol concentrations were not measured. The stage at which calcitriol deficiency plays a role in the pathophysiology of renal secondary hyperparathyroidism in cats remains unclear and additional studies are warranted.

In human patients with CKD, positive correlations between FGF-23 and PTH concentrations are documented.[3, 8, 10] This study also identified a weak positive correlation between PTH and FGF-23 in cats at both baseline (r = 0.376, P = .003) and the 12-month time point (r = 0.374, P = .003). Increases in FGF-23 concentrations in human patients with early CKD have been shown to precede changes in serum phosphate, calcium, PTH, and calcitriol concentrations.[26, 31, 33, 34] Serum phosphate concentrations remain normal in human patients with CKD as a result of the compensatory increases in FGF-23 and PTH. The hypothesis of the development of renal secondary hyperparathyroidism is that PTH concentrations become increased in response to disturbed calcium and phosphorous homeostasis and calcitriol deficiency. As renal disease progresses, there is loss of the compensatory mechanisms leading to development of hypocalcemia and hyperphosphatemia. A previous study showed that cats in the early nonazotemic stages of CKD develop renal secondary hyperparathyroidism before hypocalcemia and hyperphosphatemia.[17] The finding that FGF-23 is increased in cats in the early nonazotemic stage of CKD suggests an association between FGF-23 and the development of renal secondary hyperparathyroidism, but additional studies are required to confirm this hypothesis. The suppressive action of FGF-23 on calcitriol production and the decline in calcitriol concentrations would lead to loss of negative feedback on the parathyroid glands and increasing PTH production. Experimental animal models with overexpression of FGF-23 have identified parathyroid gland hyperplasia.[13] Indeed, results from human patients suggest that increases in FGF-23 precede increases in PTH concentrations.[26] In addition, low calcitriol concentrations induced by increased FGF-23 could contribute to hypocalcemia further stimulating PTH production.

Klotho is a transmembrane protein that is required for FGF-23 receptor activation and regulates FGF-23 signaling.[5, 35] Klotho is predominantly expressed in the parathyroid glands and kidneys, but Klotho receptor complexes exist in other tissues including choroid plexus, pituitary gland, and sinoatrial node, although little is known about the actions of FGF-23 at these sites.[4, 36] The limited expression of Klotho produces very tissue-specific actions of FGF-23. For example, activation of the FGF-23-Klotho receptor in the renal tubules decreases expression of the sodium-phosphate cotransporters and increases urinary excretion of phosphate.[35] Klotho is produced in the kidney, which leads to decreased concentrations in patients with CKD. In Klotho knockout mice, the effect of FGF-23 to decrease serum phosphate concentration is not seen.[37] Binding of FGF-23 to the FGF-23 receptor-Klotho complex in the parathyroid glands decreases PTH secretion.[11, 12] Increased concentrations of FGF-23 in early stage CKD do not suppress PTH production, suggesting development of an FGF-23 resistant state. In rat models of CKD, FGF-23 receptor and Klotho expression were decreased, which likely contributes to the resistance of the parathyroid gland to FGF-23.[38, 39] FGF-23 receptor and Klotho expression are decreased in parathyroid glands from human patients with renal secondary hyperparathyroidism.[4, 7] This resistance may lead to suppression of the inhibitory effect of FGF-23 on PTH production, thus contributing to the development of renal secondary hyperparathyroidism. This area warrants additional investigation in cats.

It is recognized that circulating FGF-23 concentrations increase as GFR declines,[8, 9, 31, 33] with an increase in FGF-23 being one of the earliest changes during decline of kidney function in human patients.[33] The accumulation of FGF-23 in cats developing azotemia may be the result of decreased renal clearance of the protein. Indeed, GFR may be the most important determinant for FGF-23 concentrations in patients with CKD.[40] Detection of FGF-23 protein in the urine of human patients indicates that renal clearance is the route of excretion of the peptide hormone.[41] In this study, a negative exponential relationship between FGF-23 concentration and GFR was identified in cats. In addition, FGF-23 concentrations were moderately correlated with plasma creatinine concentration at both time points. Data from human patients suggest that increased FGF-23 secretion causes increased FGF-23 concentration rather than decreased FGF-23 clearance. In healthy human patients, several days of dietary phosphate loading increases FGF-23 concentrations.[22, 42] However, acute changes in serum phosphate concentration did not appear to regulate FGF-23 concentrations.[43] The exact mechanism by which extracellular phosphate concentration regulates FGF-23 secretion remains unclear and, until a suitable in vitro model is established, may be difficult to determine. Nevertheless, increases in FGF-23 concentrations in response to phosphate loading suggest that FGF-23 clearance is not the sole cause of increasing concentrations in CKD patients.[22]

In this study, we documented that FGF-23 concentrations are increased in cats that develop azotemia compared with those with stable renal function. The clinical importance of this finding has yet to be ascertained. In human patients with CKD, increased FGF-23 concentrations are associated with mortality independent of any other variable.[44] Furthermore, FGF-23 was shown to be a much stronger predictor than phosphate, which has a well-established association with mortality in CKD patients. Moreover, FGF-23 concentrations also have been shown to predict progression of disease independent of serum phosphate concentration.[24] A multivariable model to predict the development of refractory secondary hyperparathyroidism in human dialysis patients found FGF-23 concentrations to be the strongest independent predictor.[25] The authors concluded that FGF-23 was a more useful screening test to predict progression of secondary hyperparathyroidism than PTH, calcium, or phosphate concentration. Longitudinal studies exploring survival and risk of adverse outcomes in cats in relation to their FGF-23 concentration would be of interest. Serum phosphate concentration has been shown to be predictive of survival in cats with naturally occurring CKD, and therefore clinical interest in FGF-23 seems appropriate.[45] The question of whether FGF-23 should be both a diagnostic and therapeutic target arises. A study of normophosphatemic human patients with CKD treated with the phosphate binder sevelamer hydrochloride experienced significant decreases in PTH and FGF-23 concentrations.[46] Early studies of cats with CKD have shown that dietary suppression of phosphate and PTH can improve survival.[47] The current recommendations by the International Renal Interest Society (IRIS) are to maintain plasma phosphate concentrations <4.5 mg/dL in cats with CKD. Screening of cats for FGF-23 concentrations may help identify those in which early intervention with phosphate restriction may be of benefit.

In conclusion, FGF-23 concentrations were increased in cats before development of azotemia. An increase in FGF-23 may prove to be a useful biomarker of disruption in phosphate homeostasis in cats with CKD, but additional studies are required. Recognition of a positive relationship between FGF-23 and PTH suggests an association between FGF-23 and renal secondary hyperparathyroidism. Additional studies investigating the potential clinical importance of FGF-23 in cats with CKD are warranted.

Acknowledgments

The study was supported by Royal Canin S.A.S, Aimargues, France and The Waltham Centre for Pet Nutrition, Waltham-on-the-Wolds, Melton Mowbray, UK.

Conflict of Interest Declaration: Authors disclose no conflict of interest.

Footnotes

  1. 1

    Idexx Laboratories Inc

  2. 2

    Diagnostic Systems Laboratories Inc

  3. 3

    Kainos Laboratories, Japan

  4. 4

    SPSS version 17.0

  5. [Correction made after online publication January 30, 2013: the Footnotes have been updated]

[Correction made after online publication January 30, 2013: the References have been updated]

Ancillary