KCNK1 inhibits osteoclastogenesis by blocking the Ca²⁺ oscillation and JNK–NFATc1 signaling axis

Each member of the two-P K⁺ (K_2P) channels has two P loops and four transmembrane domains, and 15 mammalian genes in the family have been designated as genes encoding the K_2P channels (KCNK genes) (Arrighi et al., 1998). Among them, K_2P1.1 (also called TWIK-1) is encoded by KCNK1. KCNK1 is expressed in virtually all mouse tissues and its functional relevance has been documented in several tissues (Lesage et al., 1996,, 1997; Arrighi et al., 1998; Enyedi and Czirjak, 2010). As a member of the inwardly rectifying K⁺ channel family, KCNK1 drives the membrane potential towards K⁺ balance potential, and it is expressed at significant levels in several tissues including heart, brain, pancreas, lung and placenta (Lesage et al., 1996,, 1997; Arrighi et al., 1998). KCNK1 has been shown to contribute to a large passive K⁺ conductance in rat hippocampal astrocytes, conduct inward leak Na⁺ currents in human cardiac myocytes under pathological hypokalemia, and regulate phosphate and water transport in mouse proximal tubule and medullary collecting duct (Nie et al., 2005; Zhou et al., 2009; Dempster et al., 2013), but its functional involvement in the process of osteoclast differentiation has not been well studied.

Overactivated osteoclasts increase the risk of fractures by weakening bones. This commonly occurs in several osteoclast-related diseases such as osteoporosis, Paget's disease, osteodystrophy, metastatic bone diseases and periodontal diseases (Cummings and Melton, 2002). The differentiation of hematopoietic stem cells into mature multinucleated osteoclast cells (MNCs) is basically regulated by two essential cytokines, macrophage colony-stimulating factor (M-CSF, also known as CSF1) and receptor activator of nuclear factor-κB ligand (RANKL, also known as TNFSF11) (Bucay et al., 1998; Boyce and Xing, 2008). Importantly, RANKL triggers signaling molecules that activate nuclear factor of activated T cells c1 (NFATc1), which subsequently regulates a number of osteoclast-specific genes, including tartrate-resistant acid phosphatase (TRAP, also known as ACP5), osteoclast-associated receptor (OSCAR), the d2 subunit of the V0 v-ATPase (ATP6V0D2) and cathepsin K (Rao et al., 1997; Takayanagi, 2007). RANKL also evokes Ca²⁺ oscillation, which in turn mediates NFATc1 expression and osteoclast differentiation.

In a preliminary study, DNA microarray analysis has revealed that KCNK1 expression was increased by RANKL treatment in osteoclast precursors (data not shown). An increase of KCNK1 expression from neonate to adult animals has also been shown (Arrighi et al., 1998), but its functional involvement in the process of osteoclast differentiation has not been reported to date. Therefore, herein, the relevance of KCNK1 expression to the osteoclastogenesis was investigated through its gain-of-function and loss-of-function studies.

The effect of extracellular K⁺ level on the RANKL-induced osteoclast differentiation was investigated in order to clarify the relevance of K⁺ influx to osteoclast differentiation. Extracellular K⁺ dose dependently attenuated RANKL-induced formation of TRAP⁺ MNCs and TRAP activity (Fig. 1A–C) but not the survival of bone-marrow-derived macrophages (BMMs) (Fig. 1D).

The relevance of K⁺ influx to osteoclast differentiation led us to investigate the genes involved in this process by DNA microarray analysis. When the gene expression levels in BMMs were compared to those in BMMs treated with RANKL for 1 day, a 1.5-fold increase of KCNK1 was found to be induced by RANKL (data not shown). The gradual induction of KCNK1 at its transcript and protein expression levels during RANKL-induced differentiation of BMMs into osteoclasts was further verified by quantitative real-time PCR (qPCR) (Fig. 2A; supplementary material Fig. S1) and western blotting, respectively (Fig. 2B). Osteoclast differentiation was confirmed by evaluating the expression levels of molecules related to osteoclastogenesis including transcription factors, such as c-Fos, NFATc1, and/or its target molecules (TRAP, OSCAR, ATP6V0D2 and cathepsin K) (Jimi et al., 1999; Rao et al., 1997; Takayanagi et al., 2002; Takayanagi, 2007).

The functional relevance of KCNK1 to the osteoclast differentiation was investigated by a retrovirus-based gain-of-function experiment. Retroviral overexpression of KCNK1 in BMMs was confirmed by quantitative reverse transcription PCR (qRT-PCR) (Fig. 3A). Furthermore, a patch-clamp experiment was carried out to evaluate whether overexpressed KCNK1 in BMMs acts functionally. As shown in the left graph of Fig. 3B, inward currents were significantly increased in KCNK1-overexpressing BMMs compared to the controls. An increase in the amplitude of inward K⁺ current following culture time was also observed (the right graph of Fig. 3B); the inward K⁺ current exhibited a mean amplitudes of −274±59, −430±84 and −373±78 pA after a 1-, 2- or 3-day incubation with RANKL, respectively. Patch-clamp recording controls were performed using inward rectifying K⁺ channel (Kir) blocker, Ba²⁺ (1 mM), and different concentrations of extracellular K⁺ (supplementary material Fig. S2).

When KCNK1-overexpressing BMMs were cultured for 4 days with RANKL, the formation of TRAP⁺ MNCs was strongly inhibited (Fig. 3C); the number of TRAP⁺ MNCs and TRAP activity were significantly attenuated by KCNK1 overexpression (Fig. 3D).

To gain insight into the mechanism by which KCNK1 overexpression inhibits osteoclast differentiation, its effects on the RANKL-induced activation of osteoclastogenesis-related early signaling molecules such as MAPKs and Akt were investigated. RANKL induced the phosphorylation of all kinases tested in this study, but KCNK1 overexpression was shown to attenuate the RANKL-induced phosphorylation of JNK (the isoforms p46 and p54 SAPK/JNK) (Fig. 4A).

Given that the RANKL-induced activation of JNK has been shown to subsequently lead to expression of transcription factors (Jimi et al., 1999; Srivastava et al., 1999; Shevde et al., 2000; Takayanagi et al., 2000), we further investigated the effect of KCNK1 overexpression on the expression levels of c-Fos and NFATc1, and their target genes such as TRAP, OSCAR, the v-ATPase d2 subunit and cathepsin K during osteoclast differentiation. RT-PCR revealed that KCNK1 overexpression attenuated the RANKL-induced mRNA expressions of c-Fos and NFATc1, and their target genes (upper images in Fig. 4B). The attenuation of RANKL-induced protein expression of c-Fos and NFATc1 upon KCNK1 overexpression was also confirmed by western blot analysis (bottom images in Fig. 4B).

Intracellular Ca²⁺ oscillation has been reported to be required for the activation of c-Fos and NFATc1 during RANKL-induced osteoclast differentiation (Takayanagi et al., 2002; Sato et al., 2006). Noticeably, an attenuation of RANKL-induced Ca²⁺ oscillation upon KCNK1 overexpression in BMMs was also observed here (Fig. 4C).

Given that KCNK1 mediates K⁺ influx, in the condition of KCNK1 overexpression, higher K⁺ influx might be expected. As with KCNK1 overexpression, extracellular K⁺ also inhibited the RANKL-induced JNK activation (Fig. 4D) and intracellular Ca²⁺ oscillation (Fig. 4E), suggesting that KCNK1-mediated K⁺ influx can inhibit the RANKL-induced osteoclast differentiation by inhibiting JNK activation and intracellular Ca²⁺ oscillation.

Using a retrovirus-based loss-of-function system, we further investigated the effect of KCNK1 knockdown on the RANKL-induced osteoclast differentiation. qRT-PCR revealed the prominent inhibition of KCNK1 mRNA expression in BMMs by its specific retroviral-transduced short hairpin RNA (shRNA) (upper images in Fig. 5A). Importantly, inward K⁺ currents in BMMs were slightly reduced by the downregulation of KCNK1 when compared to those of the control (bottom graph in Fig. 5A).

In contrast to the effect of KCNK1 overexpression in the differentiation of BMMs into osteoclasts, the RANKL-induced formation of TRAP⁺ MNCs was significantly enhanced by KCNK1 knockdown (Fig. 5B,C). Furthermore, KCNK1 knockdown enhanced the RANKL-induced phosphorylation of JNK (p46 and p54 SAPK/JNK), but not p38 MAPKs or ERK1/2 (Fig. 5D), and at the transcript and protein expression level, the induction of NFATc1 was also observed upon KCNK1 knockdown (Fig. 5E).

To investigate whether the enhancing effect of KCNK1 knockdown on the RANKL-induced osteoclast differentiation overcomes the anti-osteoclastogenic action of extracellular K⁺, the effect of K⁺ on osteoclast differentiation was evaluated using KCNK1-knockdown BMMs. Consistent with the results shown in Fig. 1A and Fig. 5B, fewer TRAP⁺ MNCs were formed in the presence of additional 3 mM K⁺ (lower left image in Fig. 6A) and more TRAP⁺ MNCs were formed after KCNK1 knockdown (upper right image in Fig. 6A). Interestingly, there was no significant inhibition of osteoclast differentiation by K⁺ in the condition of KCNK1 knockdown (lower right image in Fig. 6A). The ability of KCNK1 knockdown to reverse K⁺-induced inhibition of osteoclastogenesis was confirmed by counting the number of TRAP⁺ MNCs and measuring TRAP activity (Fig. 6B).

Several K⁺ channels have been reported to be expressed in BMMs, the osteoclast precursors (Vicente et al., 2003), and the inwardly rectifying K⁺ channel has been found in osteoclasts isolated from several species, indicating the presence of functional K⁺ channels in osteoclasts (Sims and Dixon, 1989; Kelly et al., 1992). In addition, osteoclasts with a spread morphology, which represents a motile phase, have been shown to have an inwardly rectifying K⁺ conductance, and rounded osteoclasts, which represent a resorptive phase of osteoclastic activity, have a transient, outwardly rectifying K⁺ conductance (Arkett et al., 1992). Furthermore, the expression of Ba²⁺-sensitive inwardly rectifying K⁺ channels in the early phase of murine osteoclast differentiation has been suggested (Shibata et al., 1996).

Furthermore, the pharmacological possibilities of using K⁺ channel inhibition to the treatment of osteoclast-mediated bone disorders has been recently suggested in several studies. The blockade of K⁺ currents in mature osteoclast by charybdotoxin or apamin (Ca²⁺-activated K⁺ channel blockers) has been shown to decrease the ability of osteoclasts to move and spread on bone substrate as well as to resorb bone (Espinosa et al., 2002), and kaliotoxin (a K⁺ channel Kv1.3 blocker) can act to reduce inflammatory bone resorption in an experimental model of periodontal disease (Valverde et al., 2004). However, additional experiments are still required for validating the clinical application of K⁺ channel blockers to treat osteoclast-mediated bone disorders (Valverde et al., 2005).

Here, we suggest that KCNK1 is a negative regulator of osteoclast differentiation; KCNK1 was gradually induced during osteoclast differentiation, but its overexpression, which induced an inward K⁺ current in BMMs, strongly inhibited the RANKL-induced osteoclast differentiation, JNK activation and expression of c-Fos, NFATc1 and their target genes. In addition, KCNK1 overexpression inhibited the RANKL-induced Ca²⁺ oscillation. Considering that extracellular K⁺ inhibited the osteoclast differentiation accompanied by the attenuation of RANKL-induced Ca²⁺ oscillation and JNK activation, these data suggest that KCNK1 could increase the influx of K⁺ to inhibit osteoclast differentiation by attenuating the RANKL-induced Ca²⁺ oscillation and JNK activation that consequently downregulates NFATc1 expression. This suggestion was also confirmed upon the knockdown of KCNK1, and is supported by the following data.

The RANKL-evoked Ca²⁺ oscillation is well-known to trigger osteoclast differentiation through NFATc1 activation, which induces osteoclast-specific gene expression (Takayanagi et al., 2002; Yang and Li, 2007; Negishi-Koga and Takayanagi, 2009); the transient initial release of Ca²⁺ from intracellular stores and the influx through specialized Ca²⁺ channels controls the dephosphorylation of NFATc1 protein, and leads to its nuclear localization, which is followed by the activation of osteoclast-specific genes. The importance of the RANKL–Ca²⁺-oscillation–NFAT-activation signaling axis during osteoclast differentiation has been also confirmed by a pharmacological inhibition study where it was found that Ca²⁺ chelators inhibit the RANKL-induced osteoclast differentiation through suppressing of NFATc1 nuclear translocation (Negishi-Koga and Takayanagi, 2009). In addition, it has been suggested that intracellular Ca²⁺ levels and the activity of osteoclasts are reduced by the exposure to high K⁺ (Kajiya et al., 2003). These data suggest that functional activation of K⁺ channels might inhibit the RANKL–Ca²⁺-oscillation–NFAT-activation signaling axis during osteoclast differentiation by increasing the influx of K⁺.

As well as the RANKL–Ca²⁺-oscillation–NFAT-activation signaling axis, the RANKL-induced activation of JNK is also essential for osteoclast differentiation (Darnay et al., 1998; Jimi et al., 1999; Kim et al., 1999; Srivastava et al., 1999; Shevde et al., 2000; Takayanagi et al., 2000; Chang et al., 2008). Activated JNK subsequently phosphorylates downstream factors, including c-Fos, which is required for NFATc1 induction. No evidence showing the relationship between exposure of high K⁺ or inwardly rectifying K⁺ channels and JNK activation during osteoclast differentiation had previously been provided, but here we suggest that the anti-osteoclastogenic activity of high K⁺ and KCNK1 overexpression could attenuate RANKL-mediated activation of JNK.

In addition, when KCNK1 was induced by RANKL (which leads to several fold higher levels after 72 h incubation with RANKL), the cultures showed a large number of multinucleated osteoclasts. In contrast to the functional role of KCNK1 as a negative regulator of osteoclastogenesis, KCNK1 might also play a role as a positive regulator to control the motility of mature osteoclasts (Arkett et al., 1992). Moreover, we suggest that the combined actions of ion channels, including KCNK1, with proton pumps could be necessary for the acidification and further bone resorption in mature osteoclasts, although further experiments are needed to substantiate this hypothesis.

In conclusion, we have determined that KCNK1 is a negative regulator of osteoclast differentiation through in vitro gain-of-function and loss-of-function studies; KCNK1 is gradually induced during osteoclast differentiation, but in KCNK1-overexpressing BMMs with the increase of inward K⁺ current, the RANKL-induced osteoclast differentiation was significantly inhibited through the attenuation of the RANKL-induced Ca²⁺ oscillation, JNK activation and NFATc1 expression. Better understanding for the functional roles and mechanisms of K⁺ channels underlying osteoclast-related diseases could be helpful to develop the relevant therapeutic strategy.

This study was carried out in strict accordance with the recommendations in the Standard Protocol for Animal Study of Korea Research Institute of Chemical Technology (KRICT; No. 2012-7D-02-01). The protocol (ID No. 7D-M1) was approved by the Institutional Animal Care and Use Committee of KRICT. All efforts were made to minimize suffering. Five-week-old male ICR mice (Damul Science Co., Deajeon, Korea) were maintained in a room illuminated daily from 07:00 to 19:00 (a 12-h-light–12-h-dark cycle) under controlled temperature (23±1°C) and ventilation (10–12 times per hour). Humidity was maintained at 55±5%, and the mice had free access to a standard animal diet and tap water. To isolate bone-marrow-derived cells (BMCs) from mice, after cervical dislocation, femur and tibia were flushed with α-MEM (Invitrogen Life Technologies, Carlsbad, CA) supplemented with antibiotics (100 units/ml penicillin and 100 µg/ml streptomycin; Invitrogen Life Technologies). BMCs were cultured on a culture dish in α-MEM (including 5.3 mM KCl) supplemented with 10% fetal bovine serum (FBS; Invitrogen Life Technologies) with 10 ng/ml of mouse recombinant M-CSF (R&D Systems, Minneapolis, MN) for 1 day. Then, after non-adherent BMCs were replated on a Petri dish and cultured for 3 days in the presence of M-CSF (30 ng/ml), adherent bone marrow-derived macrophages (BMMs) were used for osteoclast differentiation. For osteoclastogenesis, BMMs (1×10⁴ cells/well in a 96-well plate or 3×10⁵ cells/well in a six-well plate) were seeded in triplicate and cultured in the presence of 10 ng/ml of mouse recombinant RANKL (R&D Systems, Minneapolis, MN) and M-CSF (30 ng/ml) for 4 days to differentiate into mature TRAP-positive MNCs (TRAP⁺ MNCs).

Mature osteoclasts were visualized by staining for TRAP, a biomarker of osteoclast differentiation. Briefly, multinucleated osteoclasts were fixed with 3.7% formalin for 10 min, permeabilized with 0.1% Triton X-100 for 10 min, and stained with TRAP solution (Sigma-Aldrich, St Louis, MO). TRAP⁺ MNCs (≥3 nuclei) were counted. To measure TRAP activity, MNCs were fixed in 3.7% formalin for 5 min, permeabilized with 0.1% Triton X-100 for 10 min, and treated with TRAP buffer (100 mM sodium citrate pH 5.0, 50 mM sodium tartrate) containing 3 mM p-nitrophenyl phosphate (Sigma-Aldrich) at 37°C for 5 min. Reaction mixtures in the wells were transferred to new plates containing an equal volume of 0.1 N NaOH and the absorbance was determined at 405 nm.

BMMs were plated in a 96-well plate at a density of 1×10⁴ cells/well in triplicate. After treatment with M-CSF (30 ng/ml) and various concentrations of K⁺, cells were incubated for 3 days. Then, cell viability was measured using the Cell Counting Kit 8 (CCK-8) according to the manufacturer's protocol.

According to the manufacturer's protocols, total RNA was isolated with TRIzol reagent (Invitrogen Life Technologies), and reverse transcription was performed with 1 µg of RNA using oligo(dT) primers, dNTP, RNase inhibitor and SuperScript II reverse transcriptase (Invitrogen Life Technologies). The cDNA was amplified using an i-Star Taq premix PCR kit (Intron Bio, Seongnam, Korea). The thermal cycling conditions consisted of 30 s denaturation at 94°C, 30 s anneling at 60°C and 30 s extenstion at 72°C. PCR products were electrophoresed on a 1% agarose gel stained with ethidium bromide. Supplemenatry material Table S1 lists primers used in this study. GAPDH and HPRT were used as an internal control. Additionally, SYBR-based real-time quantitative PCR was performed in triplicate to quantify by absolute quantification the exact amount of KCNK1 mRNA by comparison with a DNA standard of KCNK1 using a calibration curve.

Western blotting was carried out as described previously (Choi et al., 2014). Briefly, cells were washed with ice-cold phosphate-buffered saline (PBS) and lysed in lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 1 mM sodium fluoride, 1 mM sodium vanadate, and 1% deoxycholate) supplemented with protease inhibitors. After centrifugation at 15,000 g for 10 min, The protein concentration in the supernatant was determined using the DC protein assay kit (Bio-Rad, CA), Proteins were then boiled in SDS sample buffer for 5 min, subjected to 10% SDS-PAG, and transferred onto a polyvinylidene difluoride membrane (Millipore, Billerica, MA). The membrane was probed with the indicated primary antibody, washed three times for 30 min, incubated with secondary antibody conjugated to horseradish peroxidase (HRP) for 2 h, and washed three times for 30 min. Blots were developed using SuperSignal West Femto Maximum Sensitivity Substrate (Pierce, Rockford, IL), and visualized with a LAS-3000 luminescent image analyzer (Fuji Photo Film Co., Tokyo, Japan). Antibodies against c-Fos, NFATc1 and actin were purchased from Santa Cruz Biotechnology (Dallas, TX). Antibody against KCNK1 was purchased from Abcam. All others antibodies were obtained from Cell Signaling Technology (Beverly, MA). Actin was used as a loading control.

The coding sequence of the KCNK1 gene was prepared by RT-PCR with the following primers: KCNK1 forward primer with BamHI site, 5′-GCTAGGATCCATGCTGCAGTCCCTGGCCGGCA-3′; KCNK1 reverse primer with NotI site, 5′-GCTAGCGGCCGCGTGGTCTGCAGAGCCATCC-3′. KCNK1 was cloned into the retroviral vector pMX-puro (Cell Biolabs, San Diego, CA). The forward shRNA oligonucleotides for sh-KCNK1 was 5′-GATCCCCGCAATTATGGAGTGTCGGTTTCAAGAGAACCGACACTCCATAATTGCTTTTTC-3′, and the reverse was 5′-TCGAGAAAAAGCAATTATGGAGTGTCGGTTCTCTTGAAACCGACACTCCATAATTGCGGG-3′. The oligonucleotides were cloned into the retroviral siRNA vector pSuper-retro-Puro (OligoEngine, Seattle, WA). Retroviral packaging was performed by transfecting the plasmids into Plat-E cells using Lipofectamine 2000 (Invitrogen Life Technologies). Viral supernatant was collected from the culture medium 48 h after transfection. BMMs were incubated with viral supernatant in the presence of polybrene (10 μg/ml). After infection, BMMs were cultured overnight, detached using StemPro^® Accutase^® Cell Dissociation Reagent (Invitrogen Life Technologies, Carlsbad, CA), and further cultured with M-CSF (30 ng/ml) and puromycin (2 μg/ml) for 2 days. Puromycin-resistant BMMs then differentiated into osteoclasts in the presence of M-CSF (30 ng/ml) and RANKL for 4–5 days.

Extracellular fluid containing 117 mM NaCl, 5 mM KCl, 2.5 mM CaCl₂, 1.2 mM MgCl₂, 1.2 mM NaH₂PO₄, 25 mM NaHCO₃, and 11 mM glucose was used for the patch-clamp recording, and continually aerated with 95% O₂ and 5% CO₂ gas to maintain its pH at ∼7.4. The high K⁺ solution was made by equimolar substitution of KCl for NaCl. The pH of pipette (internal) solution containing 150 mM K-Glu, 10 mM HEPES, 5 mM KCl, 0.1 mM EGTA and 2 mM Mg-ATP was adjusted to 7.2 by KOH. Using a gravity-fed perfusion system (BPS-4SG, Ala Scientific Instruments, Farmingdale, NY), complete exchange of solutions occurred within 30 s. To record procedures, BMM cells were transferred into a recording chamber (volume of chamber, 0.5 ml) mounted on an inverted microscope (CK-30, Olympus, Tokyo, Japan). Recording electrodes were prepared from capillary glass tubes (TW150-3, WPI, FL) using a microelectrode pipette puller (PP830, Narishige, Japan), and positioned using a micromanipulator (MPC-200, Sutter Instrument Co., Novato, CA). Patch pipettes filled with the pipette solutions were used at a resistance ranging from 5–8 MΩ. Membrane currents were recorded using Axopatch 200B amplifier (Axon Instruments, Union City, CA) that was connected to a computer using an analog-to-digital converter (Digidata 1322A, Axon Instruments). Currents recording and data analysis (n=10) were performed using pClamp software (Version 9.0, Axon Instruments). Generated currents were filtered with a low-pass eight-pole Bessel filter at 2 kHz. All experiments were performed at room temperature.

BMMs were seeded at a density of 1×10⁵ cells/well in a black clear 96-well plate and cultured in the presence of RANKL (10 ng/ml) and M-CSF (30 ng/ml) for 48 h. For measuring Ca²⁺ oscillations in individual osteoclast precursors, cells were loaded with 5 µM Fura-2-AM (Invitrogen Life Technologies) and 0.05% pluronic F127 (Sigma-Aldrich) for 30 min at room temperature. After washing three times with Dulbecco's modified Eagle's medium including 5.3 mM KCl (Gibco, Grand Island, NY) and changed to Hank's balanced salt solution including 400 mg/ml KCl (Gibco), the fluorescence was recorded at every 2–3 s with 340 and 380 nm excitations and emitted through a 510-nm cut-off filter at 37°C using a BD pathway system (BD Biosciences, San Jose, CA) for 300 s. All results were digitized to the mean of the ratio (340 nm:380 nm).

All quantitative values are presented as mean±s.d. Each experiment was performed three to five times, and results from one representative experiment are shown. Statistical differences were analyzed using a Student's t-test. A value of P<0.05 was considered significant.