Molecular mechanisms, physiological roles, and therapeutic implications of ion fluxes in bone cells: Emphasis on the cation‐Cl− cotransporters

Abstract Bone turnover diseases are exceptionally prevalent in human and come with a high burden on physical health. While these diseases are associated with a variety of risk factors and causes, they are all characterized by common denominators, that is, abnormalities in the function or number of osteoblasts, osteoclasts, and/or osteocytes. As such, much effort has been deployed in the recent years to understand the signaling mechanisms of bone cell proliferation and differentiation with the objectives of exploiting the intermediates involved as therapeutic preys. Ion transport systems at the external and in the intracellular membranes of osteoblasts and osteoclasts also play an important role in bone turnover by coordinating the movement of Ca2+, PO4 2−, and H+ ions in and out of the osseous matrix. Even if they sustain the terminal steps of osteoformation and osteoresorption, they have been the object of very little attention in the last several years. Members of the cation‐Cl− cotransporter (CCC) family are among the systems at work as they are expressed in bone cells, are known to affect the activity of Ca2+‐, PO4 2−‐, and H+‐dependent transport systems and have been linked to bone mass density variation in human. In this review, the roles played by the CCCs in bone remodeling will be discussed in light of recent developments and their potential relevance in the treatment of skeletal disorders.


| INTRODUCTION
Osteoporosis is the most common form of bone turnover disease (BTD).
It affects ∼10% of the population after the age of 50 and increases mortality by ∼10% when it leads to hip fractures (Rosen, 2000). Another form of BTD is known as osteopetrosis but is far less common. Although bones are denser than normal in this disease, they are still prone to fractures (Cleiren et al., 2001;Josephsen et al., 2009;Kornak et al., 2001;Margolis et al., 2008). Regardless of their types or etiologies, BTDs are all characterized by abnormalities in the function and/or number of osteoblasts, osteoclasts, and/or osteocytes (Rosen, 2000). As such, the molecular players of bone cell proliferation and differentiation have been the object of great attention over the last decade.
Among the players of interest are the estrogen receptor complexes (Pickar et al., 2018) and members of the RANKL/NFκB/NFAT and Wnt/ β-catenin/Runx2 signaling pathways (Baron & Gori, 2018;Lacey et al., 2012). A few of them have even become targets for the treatment of BTDs (Lacey et al., 2012). Ion transport systems are also expressed in bone cells to ensure the flux of Ca 2+ , PO 4 2− , and H + ions in and out of the bone matrix and have been linked to many types of bone disorders in animals (Table 1). Somewhat surprisingly, however, they have been the object of very little interest during the last several years.  In this review, the importance and role of ion transport systems in BTDs will be revisited in light of new developments and hypothetical perspectives. The cation-Cl − cotransporter (CCC) family will be paid special consideration given that six of its members have been detected in the skeleton and that one of them has been found to affect bone mineral density in human (Cheng et al., 2018;Nicolet-Barousse et al., 2005;Wasnich et al., 1983). For these reasons, and because their pharmacological inhibition is well-tolerated, a relevant question that needs to be addressed is whether the CCCs are skeletal targets of underestimated clinical potential Garneau et al., 2017;Garneau et al., 2020;Marcoux et al., 2017).
2 | WHAT ARE THE CCCS?
The Na + -dependent CCCs are known to mediate the electroneutral cotranslocation of Cl − and Na + or K + into cells (NKCC1, NKCC2, and NCC), and the Na + -independent CCCs, that of Cl − and K + out of cells (KCCs). In doing so, they affect the activity of coexpressed ion transport systems including Na + -, K + -, and Cl − -dependent Ca 2+ , PO 4 2− , and H + transporters as well as Na + , K + , Cl − , and water channels Garneau et al., 2017;Garneau et al., 2020;Marcoux et al., 2017). As such, they also affect the global ion concentration, membrane potential, and volume of cells. In fact, they exert several of their roles by acting on these coexpressed ion transport systems Garneau et al., 2017;Garneau et al., 2020;Marcoux et al., 2017;Markadieu & Delpire, 2014;Yamamoto et al., 2002). As for CCC8 and CCC9, they have still not been ascribed a precise transport function as of yet (Caron et al., 2000;Daigle et al., 2009).
The ion-transporting CCCs are all inhibited by loop and/or thiazide diuretics but to various degrees. However, these compounds reach much higher concentrations in the lumen of renal tubules than in other compartments or tissues such that they cannot inhibit the extra-renal CCCs potently without inducing a substantial natriuretic response (Ponto & Schoenwald, 1990). CCC-interacting compounds that have limited access to the tubular ultrafiltrate (Ishizaki et al., 2009) could thus reveal beneficial in the treatment of miscellaneous disorders where excessive cation-Cl − cotransport is believed to play an important pathophysiological role (H. Chen et al., 2005;Dzhala et al., 2005;Oppermann et al., 2006;Solymosi et al., 2013;Steffensen et al., 2018;Weidenfeld & Kuebler, 2017).

| Membrane potential and CCC
If the CCCs can alter membrane potential while they are not primarily electrogenic, it is because they harbor transport sites for at least one anion and one cation while the baseline anion-to-cation conductance ratio of cell membranes is rarely equal to 1.0. For instance, this ratio is higher for Cl − than it is for K + in many types of neurons such that K-Cl influx by a CCC exerts a depolarizing effect, and K-Cl efflux, a hyperpolarizing effect (Delpire & Gagnon, 2018). When the anion-tocation conductance ratio is higher for K + , such as in GABA Aexpressing vestibular and inner ear spiral ganglia (Markadieu & Delpire, 2014;Yamamoto et al., 2002), K-Cl influx by a CCC exerts a hyperpolarizing effect, and K-Cl efflux, a depolarizing effect.
There is evidence to suggest that K + channel activity at the surface of bone cells is an important determinant of membrane conductance (Chow et al., 1984;Edelman et al., 1986;Gu et al., 2001;Kelly et al., 1992;Ravesloot et al., 1990;Sims et al., 1991;Wilson et al., 2011). It could thus play the key role of orienting the movement of many ions by the electrogenic transport pathways of both osteoclasts and osteoblasts in response to a change in CCC activity. As will be discussed later, this possibility is supported further by the opposed repercussions of NCC and NKCC inhibition on bone cell signaling and long-term mineral density. It will be used as a working model to propose an integrated portrait of ion fluxes by the skeletal transportome.

| Osteoclasts
Osteoclasts play a key role in the formation of resorptive pits by secreting H + through their ruffled apical border along with Cl − to neutralize the proton charge. As described in Figure 1 (see top half) and accompanying legend, the ion transport systems at play are a vacuolar H + -ATPase (Feng et al., 2009) and a H + /2Cl − exchanger (Kornak et al., 2001). Ion transport systems of importance for the formation of resorptive pits are also present on the serosal side (Francis et al., 2002;Makihira et al., 2011;Wu et al., 2008). They include a Cl − /HCO 3 − exchanger that allows for higher Cl − i -to-Cl − o 1 concentration gradients to drive H + /2Cl − exchange and for higher HCO 3 − efflux to drive the cellular synthesis of H + .
A few years ago, RT-PCR analyses and protein expression studies led Kajiya et al. (2006) to find that KCC1 and KCC2 were both expressed in primary cultures of mouse osteoclasts while KCC3 and KCC4 were both absent. However, there are several transcripts for KCC3 in human bone RNA databanks. As such, this other isoform could be expressed in osteoblasts and/or osteocytes more specifically or more abundantly.
In their study, Kajiya et al. (2006) also observed that pit formation in calcified dentine slices added with mouse osteoclasts in culture was suppressed by KCC1 antisense oligonucleotides and that Cl − i and H + i in these cells were both increased through pharmacological inhibition of the KCCs. As illustrated through Figure 1 (see middle half), it was thus proposed that the role of KCC1 in osteoclasts was to hamper the transfer of H + from pit to cytosol by providing an added extrusion mechanism for Cl − . In this respect, loss-of-function mutations in the H + /2Cl − exchanger CLCN7 (Table 1) have been found to cause osteopetrosis in both mouse models and human (Cleiren et al., 2001;Kornak et al., 2001).
All of the KCCs can translocate NH 4 + through their K + -transport site (Bergeron et al., 2003) and could thus affect the pH of osteoclasts in doing so. NH 4 + -Cl − cotransport by these carriers would be in fact inwardly directed in that the NH 4 + o -to-NH 4 + i gradient is above 2.0 in most cell types (Evans & Turner, 1998).
Interestingly, the KCCs have also been shown to have similar apparent affinities for NH 4 + and K + based on in vitro studies (Bergeron et al., 2003), implying that they could allow for substantial NH 4 + uptake in H + -ATPase expressing cells. If it were localized on the basolateral side, KCC1 would also provide the vacuolar pump with an additional source of substrate on the apical side.
During bone resorption, Ca 2+ and PO 4 2− are released from the matrix and returned to the circulation. As described in Figure 1 (bottom half) and accompanying legend, these ions are transported from the apical to basolateral side of osteoclasts through transcytotic vesicles (along with digested bone matrix) and mitochondria 2 (Kawahara et al., 2009;Zhao, 2012). As shown again in Figure 1 (bottom half), the presence of transport systems for Ca 2+  There is growing evidence to suggest that ion transport systems do more than merely affect the activity of each other, but that they also affect the activity of signaling intermediates. In osteoclasts, such intermediates-those of the RANKL/NFκB/NFAT pathway in particular (Grossinger et al., 2018;Heeschen et al., 2002;Schwab et al., 2012)must in turn act on many of the expressed ion transport systems for functional H +and Cl − -secreting or Ca 2+and PO 4 2− -absorbing cell units to be formed. Evidence in support of this contention is that the H + /2Cl − exchanger CLCN7 is now known a target gene of NFAT along with cathepsin K and TRAP (Park et al., 2017;Sasaki et al., 2009).
The functional relevance of K + -Cl − cotransport in osteoclasts could have been assessed more readily by characterizing the available mouse models or the known human disorders of KCC inactivation or overactivation (Howard et al., 2002;Rust et al., 2007) through relevant phenotyping studies. As it stands, however, there are no findings reported in the literature on this matter. Bone-specific conditional KCC mouse models do not appear to be available either but would allow determining whether K + -Cl − cotransport in osteoclasts affects bone resorption directly or systemically.

| Osteoblasts
An essential step in osteoformation is the skeletal uptake of Ca 2+ and PO 4 2− from the circulation (see Figure 2a and  cotransporters (Lundquist, 2002;Wang et al., 2013) at the basolateral membrane of osteoblasts with the aid of Na + /K + -ATPases to ensure a favorable (inside negative) gradient for the movement of Ca 2+ and PO 4 2− ions and a favorable Na + o -to-Na + i gradient for the movement of PO 4 2− ions (Francis et al., 2002). A large fraction of the absorbed Ca 2+ and PO 4 2− ions is also taken up by intracellular matrix vesicles from the cytosol (see bottom of Figure 2a and legend) through additional Na + -PO 4 2− cotransporters (Nielsen et al., 2001;Suzuki et al., 2006) and Ca 2+ -ATPases in the membrane of these organelles (Balcerzak et al., 2008;Kirsch et al., 1997;Kirsch, 2005;Z. Xiao et al., 2007).
Another essential step in osteoformation is the transfer of Ca 2+ and PO 4 2− ions from osteoblasts to osteoid bone (see Figure 2a  As will be seen below, NCC and NKCC1 are both expressed at the surface of osteoblasts and are thus likely to affect the activity of coexpressed Ca 2+ , PO 4 2− , and H + transport systems. However, they are not predicted to do so analogously based on their effect on K + conductance given that one is K + -dependent while the other K +independent (compared Figure 2a with 2b). The same could be said of the KCCs compared to NKCC1 given that K + movement by these carriers is in the opposite direction (compare Figure 2b with 2c).
Among the various ion-transporting CCCs, NCC is the isoform that has drawn the most interest in the field of BTD. Inter alia, its inactivation in human through long-term administration of thiazides or homozygous loss-of-function mutations has been found to prevent bone mass loss in a number of observational studies. A meta-analysis by Cheng et al. (2018) has recently confirmed that thiazides could be beneficial in the treatment of osteoporosis but also led to the conclusion that higher-quality studies were required to obtain stronger evidence to this effect.
Thiazides have been said to preserve bone mass because of their positive impact on Ca 2+ homeostasis, that is, because NCC inhibition in the renal and intestinal epithelia causes these cell linings to exhibit higher levels apical Ca 2+ conductance and basolateral Na + /Ca 2+ exchange (Alexander & Dimke, 2017;Cheng et al., 2018;Hsu et al., 2015;Nicolet-Barousse et al., 2005). Yet, there is also evidence to suggest that thiazides could preserve bone mass by acting on the skeleton directly (Dvorak et al., 2007;Hsu et al., 2015;Nicolet-Barousse et al., 2005). In particular, a study by Dvorak et al. (2007) has shown that inactivation of Na + -Cl − cotransport in cultured osteoblasts led to increased cell differentiation and nodule formation.
A direct effect of thiazides on osteoblastogenesis would suggest more specifically that it is relayed through the involvement of differentiating factors (such as those of the Wnt/β-catenin/Runx2 pathway for instance) and that the activity, expression or distribution of such factors would thus be sensitive to changes in intracellular ion concentration or cell volume. In this regard, interestingly, human brain vascular smooth myocytes are prevented from proliferating and undergoing Wnt/β-catenin/Runx2 activation in parallel when their H + /2Cl − exchanger CLCN2 is inhibited pharmacologically (Lu et al., 2018).
As mentioned already, thiazides could also play a role in bone mineralization by altering the activity of Ca 2+ -dependent transport systems in osteoblasts given that they are known to do so in renal and intestinal epithelial cells. Based on our working model, and as illustrated through Figures 2a and 3a, they could exert part of this effect by eliciting the following series of events: ↑ K + uptake by NKCC1 → ↑ K + i → ↑ outward K + conductance → ↑ inward negativity → ↑ apical Ca 2+ uptake through conductive Ca 2+ channels and → ↑ basolateral Ca 2+ exit through electrogenic Na + /Ca 2+ exchangers. NKCC1 has also drawn attention in the field of BTD in that its inhibition by loop diuretics has been shown to be a risk factor for bone mass loss (Arampatzis et al., 2013;Bokrantz et al., 2020;Kubota et al., 2006;Lim et al., 2005;Norenberg, 1979;Ooms et al., 1993;Taggart, 1988). Although the mechanisms at cause are undetermined, many have incriminated the inhibitory effect of these drugs on the renal tubular reabsorption of Ca 2+ , Mg 2+ , and PO 4 2− ions (Kubota et al., 2006;Rejnmark et al., 2006). More recent studies have now shown that NKCC1 could affect bone turnover because of its presence in osteoblasts. However, its subcellular localization in bone cells does not appear to have been determined as of now.
The physiological relevance of cation-Cl − cotransport in osteoblasts has not been assessed either through the skeletal characterization of bone-specific conditional CCC mouse models. It would be most convincingly established by examining osteoformation while the activity of either isoform (that of NCC, NKCC1, KCC1, KCC3, or KCC4 in particular) is ablated or overexpressed in bone forming cells and while Ca 2+ , Mg 2+ , and PO 4 2− homeostasis is kept under strict balance.

| Osteocytes
Based on the studies available, the surface of osteocytes appears to harbor a variety of high-conductance K + channels and ion transport systems that are not present in osteoblasts (Gu et al., 2001;Ravesloot et al., 1990). Among the CCC family members, NCC is the only one to have been detected in this cell type but has been the object of no characterizations at this location (Dvorak et al., 2007). Osteocytes reside in a poorly accessible lacunocanalicular network such that their in vivo electrochemical properties are not easily amenable to light.

| Mitochondria of bones cells
Several lines of evidence suggest that the mitochondria of osteoblasts, osteoclasts, and osteocytes play an active role in bone remodeling. In particular, mitochondrial diseases have been found to cause impaired osteogenesis and accelerated age-related bone loss (Dobson et al., 2020). Given that the ion transport systems of mitochondria sustain oxidative phosphorylation, cell death coordination, and other key operations in these organelles, they should thus be seen as additional targets of interest in the treatment of BTD.

| SUMMARY PICTURE BASED ON PERSPECTIVE PRESENTED
Based on the evidence discussed, we propose that NCC inhibition in osteoblasts increases osteoformation by allowing these cells to differentiate and express robust levels of basolateral-to-apical Ca 2+ / PO 4 2− transport activity. In this regard, observational studies have shown that thiazides protect against bone fractures. Whether a decrease in Na + -Cl − cotransport at the surface osteoblasts could affect the function of osteoclasts secondarily has not been determined.
F I G U R E 3 Presumed effect of CCC inhibition on membrane potential in bone cells. The basolateral membranes of osteoclasts and osteoblasts are presumed to be more conductive to cations than anions under normal circumstances due to the presence of several types of K + channels including Kir2.1, Kv1.3, IK Ca , SK Ca , BK Ca, and a number of P2X to name a few (Edelman et al., 1986;Gu et al., 2001;Kelly et al., 1992;Ravesloot et al., 1990;Sims et al., 1991;Wilson et al., 2011).  Even if there is evidence to suggest that K + channel activity in bone cells is an important determinant of membrane conductance and that the movement of Na + , H + , and Cl − is opposed that of K + (Chow et al., 1984;Hirukawa et al., 2008), another limitation of the transport schemes proposed is that the general electrochemical properties of bone cells are poorly defined. In addition, the Na + /K + -ATPase is known to be an important determinant of membrane conductance in many cell types, K + channel activity, to vary during bone turnover and a number of K + channels subtypes, to undergo rectification.

| CONCLUSIONS
The CCCs could very well play crucial roles in bone turnover through their presence in osteoblasts and osteoclasts by coordinating the vectorial movement of H + and Cl − in one direction and that of Ca 2+ and PO 4 2− in the other. The CCCs should thus be seen as targets of interest in the treatment of BTDs, all the more so that they should be amenable eventually to bone-targeted inhibition or perhaps even activation through relevant molecular preys and isoform-specific drugs (Chew et al., 2019;Ishizaki et al., 2009;S. Liu et al., 2019).
During the last years, many fields of research appear to have been driven by the immense enthusiasm that the involvement of

CONFLICT OF INTEREST
The authors declare no conflict of interest.

ORCID
Paul Isenring http://orcid.org/0000-0002-5569-6258 ENDNOTES 1 In the reminder of the text, "i" in subscript will refer to intracellular concentration and "o" in superscript to extracellular concentration. | 4363 5 Surprisingly, proteomic studies have failed to identify PIT1 and PIT2 in matrix vesicles thus far (Balcerzak et al., 2008;Z. Xiao et al., 2007). NPT3/SLC17A2 is another candidate PO 4 2− transport system given that it is present in bone. Several other annexins have also been identified.