α-Klotho as a Regulator of Calcium Homeostasis

See allHide authors and affiliations

Science  15 Jun 2007:
Vol. 316, Issue 5831, pp. 1615-1618
DOI: 10.1126/science.1135901


α-klotho was identified as a gene associated with premature aging–like phenotypes characterized by short lifespan. In mice, we found the molecular association of α-Klotho (α-Kl) and Na+,K+-adenosine triphosphatase (Na+,K+-ATPase) and provide evidence for an increase of abundance of Na+,K+-ATPase at the plasma membrane. Low concentrations of extracellular free calcium ([Ca2+]e) rapidly induce regulated parathyroid hormone (PTH) secretion in an α-Kl- and Na+,K+-ATPase–dependent manner. The increased Na+ gradient created by Na+,K+-ATPase activity might drive the transepithelial transport of Ca2+ in cooperation with ion channels and transporters in the choroid plexus and the kidney. Our findings reveal fundamental roles of α-Kl in the regulation of calcium metabolism.

Klotho mutant mice are short-lived, and the gene responsible for the phenotype is called klotho (1). A homolog has also been identified and named β-klotho (2, 3). To avoid confusion, we refer to the original gene as α-klotho. The α-klotho (α-kl) gene encodes a type I membrane protein with considerable similarity to β-glycosidase that is predominantly expressed in tissues that are involved with calcium homeostasis; that is the parathyroid glands, kidney, and choroid plexus (4, 5) (supporting description 1). Although α-Kl has been predicted to be present on the cell surface, large amounts of α-Kl are detectable in the cytoplasm and the cleaved extracellular domain is secreted into the blood and cerebrospinal fluid (CSF) (6). In this paper, we show a pivotal role of α-Kl in calcium homeostasis (mouse breeding conditions are described in supporting description 2 and fig. S1).

We identified the α1 subunit of Na+,K+-ATPase (α1-Na+,K+-ATPase) as an α-Kl binding protein (7) (supporting description 3) and further confirmed the binding by Western blot analysis and with reciprocal immunoprecipitation (Fig. 1, A and B). α-Kl precipitated by monoclonal antibody to α1-Na+,K+-ATPase contained proteins that migrated as two bands of 120 and 135 kD, indicating that Na+,K+-ATPase binds to α-Kl regardless of its sugar moieties (Fig. 1B) (6). The β subunit of Na+,K+-ATPase, which has a major role in the regulation of membrane recruitment of Na+,K+-ATPase (8), was included in the α-Kl-Na+,K+-ATPase complex (Fig. 1B). These molecules also associated in mouse kidney and in human parathyroid glands (Fig. 1, C and D).

Fig. 1.

Interaction of α-Kl with Na+,K+-ATPase. (A to D) Proteins were immunoprecipitated from lysates of (A) choroid plexi and (C) kidneys from α-kl+/+ and α-kl–/– mice and (D) parathyroid glands from patients with monoclonal antibodies (mAbs) to α-Kl (Mink1, Rink12, and Rink107) and blotted with mAbs to α1-Na+,K+-ATPase (α6F) or α-Kl (KM2076). (B) α-Kl was blotted with KM2076 in reciprocal immunoprecipitation using mAb to α1-Na+,K+-ATPase (C464.6) or polyclonal antibody (pAb) to β1 Na+,K+-ATPase. (E) Proteins from the choroid plexi were surface-biotinylated followed by separation with streptavidin-conjugated resin. Fractions were blotted with one of a series of antibodies (mAbs to calnexin, α-Kl, α1-Na+,K+-ATPase, and β1-Na+,K+-ATPase, and pAb to FGFR2). (NB, nonbinding fraction; B, fraction binding to streptavidin-conjugated resin). (F) Subcellular localization of α-Kl and Na+,K+-ATPase. (Left) Lysates of Hela cells were separated into ER, Golgi-PM, EE, and late endosome (LE) fractions on sucrose gradients, and proteins were blotted either with antibody to α-Kl, EGFP, early endosome antigen 1 (EEA1), Golgi matrix 130 (GM130), or ribophorin1 antibodies. (Right) Analysis of α-Kl coimmunoprecipitated with α1-Na+,K+-ATPase-EGFP.

In these tissues, the majority of α-Kl immunoreactivity was detectable in the cytoplasm, whereas α1-Na+,K+-ATPase immunoreactivity was observed at or near the cell surface and diffusely in the cytoplasm (fig. S2). We verified the above histological observations by the surface biotinylation and subcellular fractionation using the choroid plexus and the HeLa cells introduced with both α-Kl and α1-Na+,K+-ATPase fused with enhanced green fluorescence protein (α1-Na+,K+-ATPase-EGFP), respectively. α-Kl was primarily detected in the nonbiotinylated cytoplasmic fraction and not in the biotinylated cell surface fraction, whereas α1-and β1-Na+,K+-ATPase were readily detectable in both fractions (Fig. 1E). This demonstrated that α-Kl binds to Na+,K+-ATPase in some intracellular organelle and not at the plasma membrane. The α-Kl and α1-Na+,K+-ATPase-EGFP complexes were found in the endoplasmic reticulum (ER), Golgi apparatus-plasma membrane (Golgi-PM), and early endosome (EE) fractions (Fig. 1F) (9). The 120-kD premature form of α-Kl was found solely in the ER fraction, whereas the 135-kD mature α-Kl and α1-Na+,K+-ATPase-EGFP complexes accumulated in the EE and Golgi-PM fractions (6). We conclude that a subset of Na+,K+-ATPase appears to traffic from the ER to the cell surface in conjunction with α-Kl and that α-Kl-Na+,K+-ATPase complexes locate in ER and Golgi apparatus and abundantly accumulate in the EE (including recycling endosome), ready for recruitment to the cell surface.

We hypothesized that α-Kl might directly affect Na+,K+-ATPase activity (supporting description 4 and fig. S3). Because the involvement of α-Kl in calcium homeostasis was strongly suggested, we investigated whether fluctuation of [Ca2+]e results in a change in Na+,K+-ATPase activity in the isolated choroid plexus (10, 11). When tissue was incubated in medium containing low Ca2+, the uptake of 86Rb, representing the activity of Na+,K+-ATPase, increased rapidly (within a few minutes), while decreasing in a high Ca2+ solution (Fig. 2A). Consistent with the 86Rb uptake data, low Ca2+ media significantly increased the surface amount of Na+,K+-ATPase (estimated by tritiated-ouabain labeling) (12, 13), and conversely were decreased in high Ca2+ media (Fig. 2B). This result demonstrates that Na+,K+-ATPase activity is correlated with its amount on the cell surface, both of which respond to [Ca2+]e concentrations, leading us to conclude that the shift of [Ca2+]e triggers a rapid response that accumulates Na+,K+-ATPase at the plasma membrane. Next, we examined the involvement of α-Kl in this system using α-kl–/– mice. The uptake of 86Rb and levels of surface Na+,K+-ATPase were not affected by the change in [Ca2+]e in α-kl–/– mice (Fig. 2, A and B). Notably, the 86Rb uptake in α-kl–/– mice was always comparable to or somewhat lower than that in wild-type (WT) high [Ca2+], the lowest level in WT mice. These indicate that α-Kl is required for the rapid recruitment of Na+,K+-ATPase to the cell surface in response to [Ca2+]e. The amount of the cell surface Na+,K+-ATPase of the extirpated choroid plexus in α-kl–/– mice was slightly but significantly lower than that in WT mice (Fig. 2C), also indicating the involvement of α-Kl in the transport of Na+,K+-ATPase to the cell surface.

Fig. 2.

[Ca2+]e-dependent recruitment of Na+,K+-ATPase. Data are mean +/– standard deviation (SD). Subjects were 8-week-old male mice. (A) Na+,K+-ATPase-dependent 86Rb uptake. Na+,K+-ATPase activity was measured in buffers containing 0.625 mM (L), 1.25 mM (N), or 2.5 mM (H) Ca2+ using bilateral choroid plexi. Uptake for the L and N groups from α-kl+/+ were significantly greater than that for the L, N, and H groups from α-kl–/– animals (P < 0.01, Dunnett's multiple comparison test). Asterisks indicate significant differences. (B) Ratios of tritiated ouabain bound to one lateral choroid plexus after 4 min incubation with artificial CSF (ACSF) containing various concentrations of Ca2+ (L, N, H) to that bound to the contralateral side after incubation with N ACSF. L/N, ratio of binding in L ACSF to that in N ACSF; H/N, ratio of binding in H ACSF to that in N ACSF. N/N (ratio of binding in N ACSF from both sides) was almost equal to 1, confirming the technical reliability. In α-kl+/+,H/N ratio was significantly <1 and L/N ratio was significantly >1 (asterisks, P < 0.01, t-test), whereas both ratios in α-kl–/– were close to 1. (C) Amount of surface α1-Na+,K+-ATPase in the α-kl+/+ and α-kl–/– mouse choroid plexi. Each fragment of the choroid plexus from the lateral ventricles was labeled with tritiated ouabain and counted. Surface α1-Na+,K+-ATPase of α-kl–/– was significantly less than that of α-kl+/+ (asterisk, P < 0.05, Student's t-test). (D) Secretion of α-Kl from isolated choroid plexi of α-kl+/+ incubated with ACSF containing the indicated [Ca2+] for 4 min. at 37°C. Supernatants of plexi from two heads were loaded per lane. (E) Secreted α-Kl during infusion of kidneys. Each band represents secreted α-Kl present in a 30-s fluid fraction. (F) Secretion of α-Kl from the parathyroid gland. Dissected thyroid and parathyroid glands were incubated in low [Ca2+] (0.5 mM Ca2+) and high [Ca2+] (2.0 mM Ca2+) buffers. The supernatants (sup) and precipitates (ppt) were analyzed by Western blotting. (G) Agonist for TRPV4 (4α-PDD) induces the Na+,K+-ATPase recruitment. Shown is ratio of 3H-ouabain-binding of the lateral choroid plexi incubated in ACSF with 4α-PDD (10 μM) for 4 min to that of control plexi incubated without 4α-PDD. The ratio of binding in α-kl+/+ was significantly >1 (asterisk, P <0.01, t-test) but not in α-kl–/–. (H) α-Kl secretion from isolated choroid plexi incubated in the ACSF containing dimethyl sulfoxide (DMSO), 5 μM 4α-PDD, or 5μM 4α-PDD with 2 mM EGTA. Western blots show proteins from α-kl+/+, supernatants of plexi from two heads per lane; n = 4; representative samples are shown. (I) Requirement of α-Kl and TRPV4 for regulated accumulation of Na+,K+-ATPase. The activity and the surface amounts of endogenous Na+,K+-ATPase were estimated using HeLa cells introduced with α-Kl and TRPV4 in various combinations with administration of DMSO or 4α-PDD. 96-well study (n = 12). Two groups transfected with or without α-kl were independently normalized as ratios to the average values in the absence of TRPV4 and 4α-PDD. Multiple comparisons show that the addition of TRPV4 and its agonist significantly induced accumulation of Na+,K+-ATPase only in α-kl-transfected groups (asterisks, P < 0.01, Dunnett's test). The amounts of secreted α-Kl from the cells were analyzed by Western blot. WE, whole extract.

The 130-kD extracellular domain of α-Kl is secreted into serum and cerebrospinal fluid (CSF). Thus, we investigated whether the secretion of α-Kl correlates with surface amounts of Na+,K+-ATPase. Incubation of isolated choroid plexi in buffer containing a low [Ca2+] increased secretion, whereas less secretion was observed in cells exposed to high [Ca2+] (Fig. 2D). In serial infusions with buffers (fig. S4), higher amounts of α-Kl were secreted into the vein of kidney in response to buffer containing a low [Ca2+], and smaller amounts were detected in high Ca2+ buffer (Fig. 2E). Secretion of α-Kl was also induced in response to low [Ca2+]e in the isolated parathyroid glands (Fig. 2F). These results suggest that secretion of α-Kl correlates with Ca2+-dependent accumulation of Na+,K+-ATPase at the cell surface.

Because administration of 4α-phorbol 12,13-didecanoate (4α-PDD) (14), a specific agonist of transient receptor potential vanniloid-4 (TRPV4), induced a significant increase of the cell surface expression of Na+,K+-ATPase and secretion of α-Kl in the choroid plexi derived from WT, but not in plexi from α-kl–/– mice (Fig. 2, G and H, supporting description 5, and fig. S5A), we transfected HeLa cells with plasmids encoding α-Kl and TRPV4 to explore the possible role of α-Kl in recruiting Na+,K+-ATPase (15, 16). We then verified that both α-Kl and activation of TRPV4 were required for the increase in abundance of Na+,K+-ATPase at the plasma membrane and the correlated secretion of α-Kl in HeLa cell experiments (Fig. 2I) (supporting description 5).

To investigate the response of PTH secretion to low [Ca2+]serum (17), we administered a Ca2+ chelating gel (Chelex, Biorad, Hercules, CA) to mice through intraperitoneal injection (fig. S6). The decrease in [Ca2+]serum caused increased secretion of PTH into serum in WT mice (Fig. 3A). The increase in the serum PTH was much smaller in α-kl–/– mice (Fig. 3A), suggesting that α-Kl has a role in promoting regulated secretion of PTH. Secretion of PTH was also increased in cultured thyroid and parathyroid glands exposed to buffer containing a low [Ca2+] (Fig. 3B). In glands from α-kl–/– mice, PTH secretion induced by low Ca2+ was detectable but only about 27% of that observed from WT glands (Fig. 3B). Ouabain, a specific inhibitor of Na+, K+-ATPase, inhibited the secretion of PTH from WT glands (Fig. 3C). There was no significant effect of ouabain on Ca2+-sensitive PTH secretion from α-kl–/– glands (Fig. 3C), which suggests that α-Kl may be required for Na+,K+-ATPase-dependent PTH release. The amount of stored PTH of α-kl–/– mice was ∼53% that of WT mice (fig. S7). However, the amount of responsive PTH secretion during 6 min is validated within 0.33% of stored PTH in α-kl–/– mice, whereas it is within 0.66% in WT mice. These results suggest that there should be sufficient amounts of stored PTH available for secretion even in α-kl–/– mice. Taken together, we suggest that the reduced response of serum PTH in α-kl–/– mice is largely caused by impairment of Na+,K+-ATPase-dependent secretion of PTH (fig. S7) (supporting description 6).

Fig. 3.

Role of α-Kl-Na+,K+-ATPase system in regulated PTH secretion and epithelial Ca2+ transport. Data are mean +/– SD. Dunnett's test was used for multiple comparisons between individual groups. (A) Changes in serum levels of PTH after Chelex injection (1 ml slurry), in α-kl+/+ and α-kl–/–. Chelex injection caused significant elevation of PTH from 2 through 6 min in α-kl+/+, compared with control injection in α-kl+/+ and both Chelex and control injection in α-kl–/– (asterisks, P <0.03). (B) PTH secreted from dissected thyroid and parathyroid glands of α-kl+/+ and α-kl–/– mice incubated in media containing 0.5 mM, 1.25 mM, or 2.0 mM Ca2+. PTH released into 0.5 mM Ca2+ by α-kl+/+ was significantly greater than that in 1.25 mM or 2 mM Ca2+ by α-kl+/+ and in all concentrations in α-kl–/– from 10 through 30 min of incubation, except 0.5 mM of α-kl–/– at 10 min (asterisks; P < 0.01). Alternatively, by multiple comparison within each genotype, the accumulation in 0.5 mM Ca2+ was significantly greater than that in 1.25 and 2.0 mM Ca2+ from 10 through 30 min (P <0.05). (C) Inhibition of PTH secretion by ouabain under low Ca2+ conditions. The accumulation of PTH in 0.5 mM Ca2+ of α-kl+/+ is significantly greater than that in 1.25 mM Ca2+ and 0.5 mM Ca2+ with ouabain of α-kl+/+ and all conditions of α-kl–/– from 10 through 30 min (P < 0.01). In multiple comparisons within α-kl–/–, there was no significant difference between 0.5 mM Ca2+ with and without ouabain from 10 through 30 min (P >0.50). (D) Calcium concentration in CSF or serum of α-kl+/+ and α-kl–/– mice. There was a significant difference between two genotypes in CSF but not in serum (asterisk, P < 0.01, Student's t test).

Since the first report of Brown, it has remained an open question how Na+,K+-ATPase activity responds to low Ca2+ in the parathyroid glands (18). When [Ca2+]e is lowered, Na+,K+-ATPase is quickly recruited to the plasma membrane in correlation with the secretion of α-Kl (Fig. 2F). An electrochemical gradient created by increased Na+,K+-ATPase may cause PTH release. If this pathway is disrupted either by α-kl deficiency or by administration of ouabain, the regulated secretion of PTH is equivalently suppressed (Fig. 3C). In the choroid plexus, the importance of α-Kl in Ca2+ transport was suggested because (i) surface amounts of Na+,K+-ATPase in α-kl–/– mice are lower than that of WT (Fig. 2C), (ii) consistent with this, the concentration of total calcium in CSF ([Ca]CSF) in α-kl–/– mice was 23.6% lower than that of WT mice (Fig. 3D and fig. S8), and (iii) both surface amounts of Na+,K+-ATPase and secretion of α-Kl respond to lowering [Ca2+]e (Fig. 2, B and D, and fig. S9). These results imply that calcium homeostasis is impaired in the CSF of α-kl–/– mice and suggest that α-Kl is involved in regulating the calcium concentration of CSF (supporting description 7) (fig. S10).

The surface expression of Na+,K+-ATPase is usually controlled by the balance of recruitment to the plasma membrane and internalization of Na+,K+-ATPase (the conventional recruitment) (19, 20). We presume that, in α-Kl expressing cells, Na+,K+-ATPase is recruited to the cell surface by a combination of the conventional recruitment and the α-Kl–dependent recruitment. The latter is characterized by Ca2+-dependency and a rapid response (fig. S11). Under normocalcemic conditions, a certain amount of Na+,K+-ATPase may be additionally recruited by the α-Kl–dependent pathway. Therefore, the amount or activity of surface Na+,K+-ATPase in WT mice is significantly higher than that of α-kl–/– mice (Fig. 2, A and C). Low Ca2+ induces an 11.9% increase of the amount of Na+,K+-ATPase at the plasma membrane. In contrast, high Ca2+ led to a 7.8% decrease in the amount of Na+,K+-ATPase at the plasma membrane (fig. S9). Accordingly, [Ca2+]e regulates this additional accumulation of Na+,K+-ATPase at the plasma membrane in correlation with the secretion of α-Kl, leading to PTH secretion and Ca2+ transport across the epithelium (fig. S11) (supporting description 7) (21). By these means, α-Kl participates in the increase of [Ca2+]e in the calcium homeostasis regulatory network. As we reported, α-Kl also invokes a negative feedback system to keep [Ca2+]e within normal range (4). α-Kl, in combination with fibroblast growth factor 23 (Fgf 23), down-regulates the production of 1,25(OH)2D3 by negatively regulating the expression of 1α-hydroxylase encoding rate-limiting enzyme of active vitamin D synthesis (4, 2224). α-Kl may be involved in this negative regulation because Urakawa et al. reported that (i) α-Kl binds to Fgf23 and (ii) α-Kl converts canonical FgfR1(IIIc) to specific receptor for Fgf23, enabling the high-affinity binding of Fgf23 to the cell surface of distal convoluted tubule where α-Kl is expressed (23) (supporting description 8).

α-kl–/– mice show increased production of vitamin D, and altered mineral-ion homeostasis is suggested to be a cause of premature aging–like phenotypes, because the lowering of vitamin D activity, either by dietary restriction (4) or fgf23 ablation (23, 25), rescues the premature aging–like phenotypes and prolongs survival in these mutants (26) (supporting description 8). Thus, our results and others indicate that α-Kl influences calcium homeostasis. This effect can explain phenotypes observed in α-kl–/– mice and the reason that α-Kl is expressed in the tissues related to calcium regulation. The mechanism is different from the previously reported action of α-Kl to interfere with the insulin or insulin-like growth factor (IGF) signal transduction (27). The fundamental role of α-Kl in a multistep calcium homeostasis is further discussed in supporting description 9.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S11


References and Notes

View Abstract

Stay Connected to Science

Editor's Blog

Navigate This Article