Klotho: a novel phosphaturic substance acting as an autocrine enzyme in the renal proximal tubule

Abstract

Klotho has profound effects on phosphate metabolism, but the mechanisms of how Klotho affects phosphate homeostasis is unknown. We detected Klotho in the proximal tubule cell, brush border, and urinary lumen, where phosphate homeostasis resides. Increasing Klotho in the kidney and urine chronically by transgenic overexpression or acutely by intravenous infusion caused hypophosphatemia, phosphaturia from decreased proximal phosphate reabsorption, and decreased activity and protein of the principal renal phosphate transporter NaPi-2a. The phosphaturic effect was present in FGF23-null mice, indicating a direct action distinct from Klotho's known role as a coreceptor for FGF23. Direct inhibition of NaPi-2a by Klotho was confirmed in cultured cells and in cell-free membrane vesicles characterized by acute inhibition of transport activity followed by decreased cell surface protein. Transport inhibition can be mimicked by recombinant beta-glucuronidase and is associated with proteolytic degradation and reduced surface NaPi-2a. The inhibitory effect of Klotho on NaPi-2a was blocked by beta-glucuronidase inhibitor but not by protease inhibitor. Klotho is a novel phosphaturic substance that acts as an enzyme in the proximal tubule urinary lumen by modifying glycans, which cause decreased transporter activity, followed by proteolytic degradation and possibly internalization of NaPi-2a from the apical membrane.

Figure 1.

Klotho protein and mRNA expression in the kidney, proximal tubular fluid, and urine of mice. A) Representative coexpression of Klotho (green) with NaPi-2a (red) in proximal tubule and with NCX1 (red) in distal tubule (examined in a total of 3 mouse kidneys). Kidney sections of mice were stained with rat anti-Klotho antibody (KM2076) and with rabbit anti-NaPi-2a serum. B–E) Immunoelectron microscopy of Klotho (arrows) in the distal tubule (B, C) and proximal tubule (B, D, E). Klotho protein was detected in the basolateral membrane (B) and the apical brush border (E), as well as the cytoplasm (C) of the proximal tubule. In the distal tubule, Klotho protein is highly expressed in cytoplasm (B, C), the basolateral membrane (B), and the tubular lumen (C). Klotho protein is also present in the capillary lumen adjacent to distal tubules (B). Endo, endothelium of capillary; MV, microvilli; TBM, tubular basement membrane. Similar results were obtained from 3 animals (kidneys). F) RT–PCR of Klotho transcript in microdissected glomeruli and proximal and distal tubules. Na/H exchanger 3 (NHE3, 237 bp) and Na/Ca exchanger 1 (NCX1, 553 bp) served as markers for proximal and distal tubules, respectively, and β-actin (453 bp) served as internal control. Similar results were obtained from 3 animals (samples). G) Klotho immunoreactivity in fresh urine of mice. Lanes 1 and 2 were loaded with 20 and 10 fmol of recombinant mouse Klotho (rMkl), respectively; Lanes 3 and 4 were loaded with 40 μl of fresh urine from bladder of Tg-Kl and WT mice, respectively. Membrane was blotted with rat anti-Klotho antibody (KM2076). Similar results were observed in 3 animals of each group. H) Klotho protein in luminal fluid of proximal tubule obtained by in vivo free-flow micropuncture compared to bladder urine of mice. Membranes were dot-blotted with 100 nl of bladder urine from two different WT and Kl^−/− mice and with 50 nl and 82 nl proximal tubule luminal fluid from WT mice, and 60 nl and 48 nl from Kl^−/− mice, and blotted with KM2076 antibody. A pegylated, streptavidin-conjugated Quantum Dot 655-Sav (Quantum Dot Corp. Hayward, CA) was used as a final fluorescent detector. Calibration was performed with rMKl (Supplemental Fig. 1F).

Figure 2.

Chronic effect of Klotho on phosphate homeostasis and NaPi-2a expression in murine kidney. A) Tg-Kl vs. WT mice. Plasma P_i (black solid circle), FE_phos (red solid square) and phosphate transport (J_phos) (blue open circle) in single in vitro microperfused proximal tubules (blue open spot). n = 5; *P < 0.05, **P < 0.01 vs. WT; unpaired Student’s t test. B) Representative immunoblot of NaPi-2a and β-actin in kidney cortical membranes (cortex) and BBM from 4 WT and Tg-Kl mice/group with KM2076 and anti-β-actin. Expected mobilities of full-length NaPi-2a and NaPi-2a N-terminal are indicated.

Figure 3.

Acute effect of Klotho on phosphate homeostasis and NaPi-2a expression in rat kidney. Six normal Sprague-Dawley rats were given 64 pmol of recombinant mouse Klotho (rMKl; solid line) or vehicle (Veh; dashed line). Blood and urine were collected at 0, 1, 2, 4, and 6 h. A) Plasma P_i in Klotho-treated (solid circle solid line) vs. vehicle-treated (open circle dotted line); n = 6. *P < 0.05, **P < 0.01 vs. Klotho at 0 h; ^#P < 0.05, ^##P < 0.01 vs. vehicle at 0 h; ANOVA followed by Student-Newman-Keuls test. B) FE_phos; symbols and statistical analysis same as for A. C) BBMVs were prepared, and Na⁺-dependent phosphate transport was measured; n = 4. *P < 0.05, **P < 0.01; unpaired Student’s t test. D) Representative immunohistochemical stain for NaPi-2a with rabbit anti-Klotho serum from 3 rats. E) Representative immunoblot for NaPi-2a and β-actin of kidney cortex membranes (cortex) and BBMVs from 3 rats/group. F) FGF23-null mice were intraperitoneally injected with 6.0 pmol of rMKl (solid line and symbols) or vehicle (dashed line and open symbols). Blood and urine were collected at 0 and 3 h and measured for plasma P_i (black) and FE_phos (red); n = 4. *P < 0.05 vs. 0 h; unpaired Student’s t test.

Figure 4.

Klotho effect on Na-dependent phosphate transport activity in OK cells and BBMVs and on NaPi-2a antigen in OK cells. Recombinant mouse Klotho was added to a proximal tubule-like cell line (OK cells) and BBMVs, and Na⁺-dependent P_i uptake was measured. A–D) Dose dependence (A) and time dependence (B) in OK cells; dose dependence (C), and time dependence (D) in BBMVs; n = 6. *P < 0.05 vs. baseline; ANOVA with Student-Newman-Keuls test. E, F) OK cells were transfected with NaPi-2a/eGFP, and live cells were monitored with fluorescent microscopy; timed images show effects of Klotho (E) and vehicle (F).

Figure 5.

Examination of Klotho as a glucuronidase-like enzyme. A) Klotho (0.4 nM), β-glucuronidase (β-Glu) (200 μg/ml), and the β-glucuronidase inhibitor DSAL (1 μM) were added directly to BBMVs for 4 h, and Na⁺-dependent P_i uptake was measured. Results were summarized from 6 samples/group and analyzed by ANOVA followed by Student-Newman-Keuls test. B) Representative immunoblot of NaPi-2a and β-actin in BBMV from 3 preparations. BBMVs were treated as in A, but BBMVs were subjected to immunoblotting for NaPi-2a and β-actin. C) Representative immunoblot of NaPi-2a and β-actin in BBMV from 4 preparations. Experiment was similar to A, but protease inhibitors (PIs) were used instead of β-Glu, and 20 μg protein of BBM was subjected to immunoblot for NaPi-2a and β-actin D) Experiment was similar to C, but Na⁺-dependent P_i uptake was measured in BBMVs. Results were summarized from 6 independent samples/group; significant differences were analyzed by ANOVA followed by Student-Newman-Keuls test. ^aP < 0.05; ^bP < 0.01 vs. 1st bar (no Klotho, DSAL, or PIs); ^cP < 0.05; ^dP < 0.01 vs. 2nd bar (Klotho); ^eP < 0.05; ^fP < 0.01 vs. 5th bar (PIs); ^gP < 0.05; ^hP < 0.01 vs. 6th bar (Klotho+PIs).

Figure 6.

Comparison of Klotho to glucuronidase and effect of inhibition of β-glucuronidase or sialidase in OK cells. A) No effect of DANA on Klotho-inhibited Na-dependent phosphate transport in BBMVs. BBMVs were incubated with recombinant Klotho protein (0.4 nM) and the sialidase inhibitor DANA (20 pM) for 4 h. Results were summarized from 6 samples/group and analyzed by ANOVA followed by Student-Newman-Keuls test. B) Klotho was directly added to OK cells. Sodium-dependent P_i uptake was measured in cells incubated with Klotho (0.4 nM), β-glucuronidase (β-Glu) (200 μg/ml), and the β-glucuronidase inhibitor DSAL (1 μM) for 4 h. Results were summarized from 6 independent samples/group. C) Dose response of the sialidase inhibitor DANA on Na⁺-dependent P_i uptake in OK cells incubated with DANA alone or Klotho (0.4 nM) + DANA for 4 h. While β-glucuronidase inhibition blocked the Klotho-induced inhibition, sialidase inhibition did not. Results were summarized from 6 independent samples/group.

Figure 7.

Proposed model of how Klotho regulates NaPi-2a in the apical membrane of the renal proximal tubule. 1) Klotho functions acutely as a direct extracellular enzyme deglycosylating NaPi-2a protein and/or a putative regulatory protein (gray rectangle) to reduce cotransport activity. 2) NaPi-2a protein is more susceptible to resident proteases in BBM and consequently is proteolytically degraded. 3) Internalization of NaPi-2a protein from BBMVs into the intracellular pool.