Impaired renal phosphate reabsorption, as measured by dividing the tubular maximal reabsorption of phosphate by the glomerular filtration rate (TmP/GFR), increases the risks of nephrolithiasis and bone demineralization. Data from animal models suggest that sodium–hydrogen exchanger regulatory factor 1 (NHERF1) controls renal phosphate transport. We sequenced the NHERF1 gene in 158 patients, 94 of whom had either nephrolithiasis or bone demineralization. We identified three distinct mutations in seven patients with a low TmP/GFR value. No patients with normal TmP/GFR values had mutations. The mutants expressed in cultured renal cells increased the generation of cyclic AMP (cAMP) by parathyroid hormone (PTH) and inhibited phosphate transport. These NHERF1 mutations suggest a previously unrecognized cause of renal phosphate loss in humans.
Some reports suggest that NHERF1 can interact with known amino acid sequences at the C-terminal end of NPT2a6,7 and the PTH type 1 receptor (PTH1R).8,9 NHERF1 disruption in mice alters NPT2a targeting,10 and the expression of NHERF1 in cultured cells attenuates PTH-induced cAMP synthesis,8,11 suggesting that a defect in NHERF1 function may increase cAMP generation in response to PTH. Although these results suggest that NHERF1 can modulate renal phosphate transport through various mechanisms, data supporting a role of NHERF1 in phosphate homeostasis in humans have been lacking.
We reasoned that if NHERF1 is a key regulator of renal phosphate reabsorption in humans, mutations in the NHERF1 gene might be detected in patients with unexplained hypophosphatemia and a decreased capacity of the renal tubules to reabsorb phosphate. Here, we describe the sequencing of the coding region of NHERF1 in patients who presented with either renal stones or bone demineralization, together with normal serum PTH concentrations.
Methods
Patients
Ninety-two of 170 patients referred to our clinical investigation department for examination of renal calcium lithiasis or osteopenia between January 2001 and December 2003 had no evidence of any usual causes of these disorders (e.g., hyperparathyroidism, hyperthyroidism, or an increased FGF-23 plasma concentration). These 92 patients (group 1), all of whom were white, provided written informed consent to obtain blood DNA samples for genomic DNA extraction, in accordance with French national guidelines, as did all other participants. The protocol was approved by the institutional review board.
We measured serum phosphate concentrations and calculated the TmP/GFR value12 in group 1. We also calculated the TmP/GFR value in three relatives of a patient in group 1 who had an NHERF1 mutation; these three relatives were not included in group 1, since two had nephrolithiasis and one had no symptoms. In addition, we studied patients who were similar to the patients in group 1 with respect to ethnic background, age range, and GFR but who did not have renal stones or bone demineralization (group 2). The patients in group 2 were referred to our clinical investigation department for renal assessment before undergoing a potentially nephrotoxic treatment for psoriasis.
Sequencing
The sequencing and analysis of both strands of the coding region and the intron–exon junctions of NHERF1 were performed by investigators who were unaware of the TmP/GFR values and the clinical histories of the patients. (The sequences of the intronic primers are provided in the Supplementary Appendix, available with the full text of this article at www.nejm.org.) The two strands of polymerase-chain-reaction products were sequenced with the use of the BigDye Terminator Cycle Sequencing Ready Reaction kit (Perkin Elmer) on an ABI PRISM 3100 sequencer (Applied Biosystems). We analyzed both strands of DNA using the software programs Sequencing Analysis (Applied Biosystems) and Clustal W (Infobiogen). The nucleotide position was numbered according to the starting point of the ATG codon in the complementary DNA (cDNA) (sequence accession number NM_004252 [GenBank] ).
Biochemical Analysis
All biochemical measurements were based on standard clinical laboratory procedures for electrolyte and hormone concentrations in blood and urine. For measurements of the FGF-23 concentration, blood samples were centrifuged within 30 minutes after collection, and plasma samples were stored at –70°C until measurement. FGF-23 concentrations were measured with the use of a commercial enzyme-linked immunosorbent assay for C-terminal FGF-23 (Immutopics).
In Vitro Experiments
We used standard methods for cell-culture conditions; phosphate-uptake experiments; measurements of intracellular calcium concentrations, cAMP, and inositol phosphate concentrations; cell transfection; and plasmid construction (see the Supplementary Appendix).
Statistical Analysis
Results are expressed as means ±SD or medians and ranges. Quantitative variables were compared with the use of an unpaired t-test or Kruskal–Wallis analysis of variance. Post hoc multiple comparisons were performed by means of the Mann–Whitney rank-sum test, when appropriate. The number of comparisons made was adjusted with the Bonferroni correction (
=0.012 when four comparisons were tested). Categorical variables were compared by means of Fisher's exact test. Analyses were performed with the use of JMP software (SAS Institute) and GraphPad software (InStat) for MacIntosh.
Results
Table 1 lists the main relevant characteristics of the 92 patients with calcium-containing renal stones (50 patients), bone demineralization (30), or both (12). We identified three distinct mutations in the NHERF1 gene (Figure 1) in four unrelated patients (Table 1). No mutations in the SLC34A1 gene, which encodes the sodium–phosphate cotransporter NPT2a, were found.
|
|
The four patients with mutations did not have proximal-tubule dysfunction other than the low TmP/GFR value; they did not have glycosuria, and their serum bicarbonate concentrations and blood pH values were normal. These four patients had heterozygous nucleotide substitutions at various exons, resulting in single amino acid changes.
Patients 1 and 2 carried the missense mutation c.328C
G in exon 1, which resulted in the substitution of valine for leucine at position 110 (p.L110V) (Figure 1A). Patient 3 had a c.458GA transition in exon 2, which led to the substitution of glutamine for arginine at position 153 (p.R153Q) (Figure 1B). Three relatives of Patient 3 also agreed to participate in this study. A sister of Patient 3 and the sister's daughter had a history of recurrent renal lithiasis. Both had a low TmP/GFR value (0.50 mmol per liter [1.54 mg per deciliter]) and had the same NHERF1 mutation as Patient 3. Another sister of Patient 3 had no history of renal lithiasis, had a normal TmP/GFR value (0.78 mmol per liter [2.41 mg per deciliter]), and did not carry the mutation. Patient 4 had a mutation in exon 3 (c.673GA), which results in a change from glutamic acid to lysine at amino acid 225 (p.E225K) (Figure 1C). The three mutated amino acids were conserved in various species, suggesting that they play an important role in protein function. We also found a base substitution at codon 167 (CT) that did not change the amino acid. This polymorphism was present in five patients: three with a normal TmP/GFR value and two with a low value.
To determine whether the identified mutations could be detected in a population in which almost all patients have normal TmP/GFR values and normal renal function, we sequenced NHERF1 in group 2, which consisted of 63 patients with no renal lithiasis or bone demineralization. The mean TmP/GFR value in this group was significantly higher than that among patients with renal stones or bone demineralization (Table 1). Only five patients in group 2 had a TmP/GFR value below 0.75 mmol per liter. One patient in group 2 had an NHERF1 mutation (L110V). This patient had a low TmP/GFR value (0.71 mmol per liter [2.19 mg per deciliter]).
Overall, among the 158 patients studied to date (92 patients in group 1, 3 patients who were relatives of a patient in group 1, and 63 patients in group 2), the detected NHERF1 mutations were in 7 of the 46 patients with low TmP/GFR values (<0.75 mmol per liter), and none were in the 112 patients with normal TmP/GFR values (
0.75 mmol per liter; P<0.001 by Fisher's exact test). The mean TmP/GFR value was significantly lower in the 7 patients with mutations than in the other 151 patients (0.60±0.08 vs. 0.83±0.17 mmol per liter [1.85±0.24 vs. 2.56±0.52 mg per deciliter], P<0.001 by an unpaired t-test).
We measured urinary cAMP excretion, which is considered a marker of PTH activity in the kidney, in the four patients in group 1 who had NHERF1 mutations and in eight other patients in group 1 who had a similar GFR and similar serum PTH, serum calcium, and 25-hydroxyvitamin D concentrations — measures known to influence urinary cAMP excretion. Urinary cAMP excretion was significantly higher in the patients with NHERF1 mutations than in the matched patients from group 1 who did not have NHERF1 mutations (Figure 2A).
|
We carried out in vitro experiments to determine the mechanism by which these mutations might decrease renal phosphate transport. Since urinary cAMP excretion was increased in the patients with NHERF1 mutations, we evaluated whether the mutant NHERF1 proteins could alter cAMP synthesis in renal cells under baseline conditions (Figure 2B) or in response to PTH (Figure 2C). We transfected opossum kidney cells, a commonly used model of renal proximal tubular cells that express NPT2a and PTH1R, with plasmid alone or plasmid containing wild-type or mutant NHERF1 cDNA. The baseline accumulation of cAMP was similar in all three types of transfected kidney cells (Figure 2B).
As expected, incubation of cells with PTH stimulated cAMP synthesis in all cell groups. However, cAMP accumulation was significantly higher in the cells transfected with the mutant NHERF1 cDNA than in those transfected with wild-type cDNA (Figure 2C). Transfection efficacy did not account for these differences, since the abundance of wild-type or mutant proteins, as assessed with the use of a Western blot, were similar (see the figure in the Supplementary Appendix). Because PTH1R is coupled to the phospholipase C pathway, we also measured the intracellular calcium concentration and the production of inositol phosphate in response to PTH in opossum kidney cells transfected with wild-type or mutant cDNA. In contrast to bradykinin, which was used as a control, the presence of PTH did not affect the intracellular calcium concentration or inositol phosphate production in these cells.
To evaluate the effect of increased cAMP synthesis on phosphate uptake in vitro, we compared the inhibitory effect of PTH on phosphate uptake in opossum kidney cells transfected with either wild-type cDNA or mutant NHERF1 cDNA. In the absence of PTH, the type of transfected construct did not significantly affect phosphate uptake (Figure 2B). As expected, the addition of PTH induced a significant decrease of phosphate uptake in all cell groups (Figure 2C). However, the effect of PTH on phosphate transport was greater in cells transfected with a mutant NHERF1 than in those transfected with wild-type human NHERF1 cDNA (Figure 2C).
To make sure that these differences were not due to interactions of endogenous NHERF1 with transfected human NHERF1, we performed similar experiments in a subclone (clone 2) of opossum kidney cells in which endogenous NHERF1 expression was almost undetectable on a Western blot (Figure 2D). In this cell line, the results were similar to those described above: PTH decreased phosphate uptake in all groups of cells (Figure 2E). Transfection with the wild-type human NHERF1, as compared with mutant NHERF1, markedly reduced PTH-induced inhibition of phosphate uptake. Phosphate transport in cells transfected with mutant NHERF was not significantly different from that seen in the absence of NHERF1 (Figure 2E).
Discussion
Our results identify NHERF1 mutations as a cause of renal phosphate loss that may increase the risk of renal stone formation or bone demineralization: the mutations were identified only in patients with TmP/GFR values that were below normal and significantly lower than those in patients in whom mutations were not identified. Furthermore, we observed, in one family, that the mutation was associated with a low TmP/GFR value and renal stones. Our in vitro experiments indicate that these mutations had no effect on basal phosphate uptake but, instead, potentiated PTH-induced cAMP generation and, consequently, the inhibition of phosphate transport.
Three lines of evidence support the concept that the NHERF1 mutations reported here affect the phenotype of renal proximal tubular cells through a cAMP-mediated pathway. First, excreted urinary cAMP was increased in patients with mutations; second, the serum calcitriol concentration, which is highly dependent on the PTH–cAMP pathway,13 was higher in patients with NHERF1 mutations than in matched controls; and third, in opossum kidney cells, both PTH-induced cAMP generation and PTH-induced inhibition of phosphate uptake were increased in cells transfected with mutant NHERF1 as compared with those transfected with wild-type NHERF1. The in vivo and in vitro results of our study are in line with those of a study by Mahon et al., who reported that NHERF–PTH1R interaction decreased cAMP synthesis in response to PTH in PS120 cells8 and in opossum kidney cells.11 In contrast, cAMP synthesis and protein kinase C activity in response to a supraphysiological concentration of PTH were similar in cultured NHERF1–/– renal cells and NHERF1+/+ cells14,15 However, when experiments were carried out in NHERF1–/– kidney slices, a decrease in phospholipase C was observed.16 Whether urinary cAMP excretion in vivo is different in NHERF1+/+ and NHERF–/– mice is unknown.
The observation that renal tubular reabsorption of phosphate was impaired in patients with loss-of-function NHERF1 mutations is consistent with the phenotype of mice bearing heterozygous or homozygous disruption of NHERF1. These mice have hypophosphatemia due to decreased renal phosphate transport and nephrolithiasis.10,17 In these mice, the mechanism underlying decreased NPT2a in brush-border membranes is multimechanistic. Decreased phosphate uptake in cultured proximal tubular cells from NHERF1–/– mice14 suggests that NHERF1 is required for proper sorting or activity of phosphate transporters. In contrast, ex vivo experiments conducted in renal brush-border membrane vesicles prepared from NHERF1–/– mice show normal phosphate uptake as compared with vesicles from wild-type animals.16 These latter results might be explained by the interaction between NHERF1 and a signaling pathway triggered by an extrarenal factor such as PTH.
None of the patients we describe had hypercalcemia, most likely because of different patterns of expression of NHERF1 and PTH1R. Indeed, in the kidney, NHERF1 expression is restricted to the proximal tubule,18 whereas PTHR1 is expressed in both the proximal and the distal tubules.19 In bone, NHERF1 and PTH1R are not expressed in the same cell types: NHERF1 is expressed in osteoclasts,20 whereas PTH1R is expressed in osteoblasts.21 As a result, the patients with NHERF1 mutations presented with a selective abnormality of the phosphate balance. This phenotype is strikingly different from both the phenotype encountered in Jansen's metaphyseal chondrodysplasia (characterized by short-limbed dwarfism, bowing of long bones, renal lithiasis, hypercalcemia, mild hypophosphatemia, a low serum PTH concentration, and increased urinary cAMP), which is related to gain-of-function mutations of PTH1R,22 and the phenotype in pseudohypoparathyroidism type 2, in which resistance to PTH results from a post-cAMP abnormality of the PTH signaling pathway.23
In the NHERF1 protein, two structural domains, named PDZ1 and PDZ2, were reported to interact with transmembrane proteins. It is noteworthy that none of the mutations reported here affected the PDZ1 domain in NHERF1; instead, they affected PDZ2 and the region between the two PDZ domains. PDZ1 was reported by Gisler et al. to play a critical role in the interaction between NHERF1 and NPT2a.6 Moreover, a serine residue in PDZ1 was recently shown to play a key role in this interaction.24 In contrast, PTH1R was reported to interact with PDZ1 and PDZ2 with similar affinities,25 although the precise role of these domains in the interaction with PTH1R is not known. Our data suggest an important role of PDZ2 and the inter-region domain in the interaction with PTH1R in humans.
The number of identified genes involved in phosphate homeostasis has increased in recent years. The studies reported here suggest a novel mechanism of renal phosphate loss and hypophosphatemia that may increase the risk of renal stone formation or bone demineralization.
Supported by a grant from INSERM, Université Paris Descartes, Association Laboratoire de Recherches Physiologiques, and l'Agence Nationale pour la Recherche (ANR-07-PHYSIO-017-01).
No potential conflict of interest relevant to this article was reported.
We thank Dr. Paul Landais for valuable help during the preparation of the manuscript, Dr. Françoise Leviel for obtaining cAMP measurements, and Valérie Boitez and Catherine Vernimmen for technical assistance.
Source Information
From INSERM Unité 845, Université Paris Descartes, Faculté de Médecine (Z.K., N.B., R.A., C.L., L.B., G.P., P.U.-T., G.F., D.P.); Hôpital Bichat–Claude Bernard (B. Gérard, C.S., B. Grandchamp), Hôpital Necker–Enfants Malades (P.U.-T., G.F., D.P.), and Hôpital Européen Georges Pompidou (G.F.), Assistance Publique–Hôpitaux de Paris; and INSERM Unité 773, Unité de Formation et de Recherche Médicale (C.S.), and Institut Fédératif de Recherche 02, INSERM (B. Grandchamp), Université Paris Diderot — all in Paris; and Clinique du Landy, Saint Ouen, France (P.U.-T.).
Drs. Friedlander and Prié contributed equally to this article.
Address reprint requests to Dr. Prié at Service d'Explorations Fonctionnelles, Hôpital Necker–Enfants Malades, 149 Rue de Sèvres, Paris 75015, France, or at prie{at}necker.fr.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
NHERF1 Mutations and Responsiveness of Renal Parathyroid Hormone
Bergwitz C., Bastepe M., Rendina D., De Filippo G., Strazzullo P., Prié D., Friedlander G.
Extract |
Full Text |
PDF
N Engl J Med 2008;
359:2615-2617, Dec 11, 2008.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
