Pache.co Information Technologies

One form of cell-cell communication is via gap junctions, the membrane-spanning channels composed of two juxtaposed hemichannels (Goodenough et al., 1996). In addition to direct gap junction intercellular communication (GJIC), halves of gap junctions, hemichannels mediate the communication between bone cells and the extracellular environment (Civitelli, 2008). Connexin-forming hemichannels exhibit relatively low substrate selectivity and permit small molecules (≤1.2 kDa) to pass through (Goodenough et al., 1996; Kumar and Gilula, 1996). Cx43 is the predominant connexin subtype expressed in osteocytes (Yellowley et al., 2000; Cheng et al., 2001a) and osteoblasts (Civitelli et al., 1993). Cx43 hemichannels are highly responsive to mechanical stimulation, and their opening induced by mechanical loading mediates the release of anabolic factors such as PGE₂, adenosine triphosphate (ATP) and nitric oxide (NO) from osteocytes (Jiang and Cherian, 2003; Cherian et al., 2005) and osteoblasts (Thi et al., 2012). Both ATP and NO are related to the production of the bone anabolic agent PGE₂ (Sugiatno et al., 2006; Genetos et al., 2007). In osteocytes, released PGE₂ can act in either a feed-forward or feedback manner in regulating Cx43 expression and function. Mechanical loading-induced PGE₂ release from osteocytic Cx43 hemichannels increases Cx43 expression and Cx43-forming gap junctions (Cheng et al., 2001b). Interestingly, fluid flow-induced accumulation of extracellular PGE₂ leads to the closure of Cx43 hemichannels (Riquelme et al., 2015). The Cx43 hemichannels and extracellular PGE₂ play an inhibitory role in glucocorticoid-induced apoptosis (Kitase et al., 2010). Cx43 hemichannels with PGE₂ release are also essential for normal bone structure (Xu et al., 2015; Zhao et al., 2022a) and the anabolic response of tibias to mechanical loading in vivo (Zhao et al., 2022a; Zhao et al., 2022b). In addition to connexins, pannexins are also capable of forming hemichannels (Plotkin et al., 2017). Pannexin1 is the most widely distributed pannexin in bone cells, and these hemichannels are involved in the release of PGE₂ induced by mechanical stimulation in osteoblasts (Thi et al., 2012). However, the roles of pannexin channels in bone have not been investigated in great detail. In this review, we focus on the function of Cx43 hemichannels in releasing PGE₂, and further PGE₂-regulated skeletal development and cellular signals that drive bone anabolic and bone remodeling responses to mechanical stimulation.

Relationship between Cx43 hemichannels and PGE₂ upon mechanical stimulation in osteocytes

Osteocytes are a rich source of PGE₂ upon mechanical stimulation. Mechanical stress in the form of fluid flow causes a rapid increase in COX-2 expression (Joldersma et al., 2000) and PGE₂ production in osteocytes (Ajubi et al., 1996; Ajubi et al., 1999). Inhibition of COX-2 enzymatic activity with NS-398 inhibitor abolishes the stimulatory effect of fluid flow on PGE₂ secretion from osteocytes (Bakker et al., 2003). During mechanical stimulation, Cx43 hemichannels play a crucial role in the PGE₂ release from osteocytes. Low-density cultures of primary osteocytes and osteocyte-like MLO-Y4 cells with minimal cell-cell contacts, thus void of gap junctions, release more PGE₂ than cells cultured at higher densities (Jiang and Cherian, 2003; Cherian et al., 2005). Inhibition of Cx43 channels by chemical blocker β-glycyrrhetinic acid (Jiang and Cherian, 2003; Cherian et al., 2005) or knocking down Cx43 by siRNA (Genetos et al., 2007) attenuates fluid flow-induced hemichannel activity and PGE₂ production in low-density cultured osteocytes. This experimental evidence suggests that Cx43 hemichannels participate in PGE₂ secretion during mechanical stimulation. To further depict the relationship between Cx43 hemichannels and PGE₂ release, we develop a polyclonal antibody, Cx43 (E2), that targets the second extracellular loop domain of Cx43 and specifically blocks osteocytic Cx43 hemichannels in vitro (Siller-Jackson et al., 2008; Riquelme et al., 2013). Blocking of Cx43 hemichannels by Cx43 (E2) inhibits the opening of hemichannels and the release of PGE₂ induced by flow shear stress in osteocytes (Siller-Jackson et al., 2008). Interestingly, PGE₂ also has a negative feedback regulation on Cx43 hemichannels in response to mechanical stimulation. Extracellular PGE₂ accumulation after the continuous opening of hemichannels by fluid flow acts on EP2/4 receptors to close Cx43 hemichannels (Riquelme et al., 2015). The negative feedback is caused by the PGE₂ activation of p44/42 ERK signaling and direct Cx43 phosphorylation at S279/282 residues thereby leading to the closure of Cx43 hemichannels. (Riquelme et al., 2015). In addition, the released PGE₂ from osteocytes by fluid shear stress promotes Cx43 expression and further increases Cx43 gap junctions (Jiang and Cheng, 2001; Xia et al., 2010). Consistent with the effects of fluid shear stress, direct treatment of the MLO-Y4 cells with PGE₂ similarly increases Cx43 expression and gap junctions. In contrast, inhibition of PGE₂ signaling by indomethacin reduced gap junction formation by fluid shear stress (Cheng et al., 2001b).

Since the PGE₂ secretion by osteocytes depends on the opening of Cx43 hemichannels in osteocytes, it is important to understand how Cx43 hemichannels are regulated by mechanical stimulation (Figure 1). We find that the dendritic processes of osteocytes transmit mechanical signals to the cell body, leading to the opening of Cx43 hemichannels in MLO-Y4 cells and primary osteocytes (Burra et al., 2010). Cx43 is richly present in the cell body, not the dendritic processes of osteocytes. The integrin αVβ3, located at the dendrites of osteocytes, is an important component of the glycocalyx complex that tethers osteocytes to the canalicular wall and amplifies the magnitude of mechanical signals experienced by osteocytes (Wang et al., 2007; Riquelme et al., 2021). Upon fluid flow, the force generated by tethering elements is a magnitude higher than shear stress on the cell surface. Mechanically activated integrin αVβ3 at dendritic processes induces the activation of intracellular PI3K signaling (Batra et al., 2012), which activates the downstream effector AKT (Batra et al., 2014a). Activated AKT directly phosphorylates Cx43 and integrin α5 (Batra et al., 2014a) in the cell body to increase the interaction between these two proteins, opening the Cx43 hemichannels. Upon mechanical loading, activation of α5β1 through its conformational changes opens the Cx43 hemichannels in MLO-Y4 cells. Interestingly, this action is independent of integrin binding to its extracellular substrate, fibronectin (Batra et al., 2012; Batra et al., 2014a). Moreover, the scaffolding molecule 14-3-3θ assists in transporting both Cx43 and integrin a5 from the Golgi apparatus to the plasma membrane to form mechanosensitive Cx43 hemichannels (Batra et al., 2014b). Silencing 14-3-3θ prevents the accumulation of Cx43 on the cell membrane and the opening of hemichannels caused by fluid flow (Batra et al., 2014b). Recently, we find that Piezo1 is co-localized with Cx43 hemichannels on osteocyte cell surface. The activation of the Piezo1 leads to an increase in intracellular Ca²⁺ and the opening of Cx43 hemichannels through PI3K-AKT pathway in osteocytes (Zeng et al., 2022). Interestingly, the release of PGE₂ is upregulated when Piezo1 is activated by either agonist or mechanical stretch (Li et al., 2019; Yang et al., 2022). The PGE₂ release by Cx43 hemichannels is also regulated by extracellular ATP. Mechanical stimulation-induced ATP release through Cx43 hemichannels activates P2X receptors and promotes the Ca²⁺ influx to sustain the activities of Cx43 hemichannels (Zeng et al., 2022). Blockade of P2X7 purinergic receptors prevents PGE₂ release from MLO-Y4 cells. However, the activation of purinergic receptors and the increase in the release of PGE₂ appears to be independent of hemichannel formation (Genetos et al., 2007). Thus, the mechanism for ATP-induced PGE₂ release is yet to be fully understood in osteocytes.

FIGURE 1. A model illustrating the role of PGE₂ released from Cx43 hemichannels under mechanical loading in the regulation of the anabolic response to mechanical stimulation in bone. Upon mechanical loading, osteocytic dendrites sense mechanical stimulation and transduce these signals through integrins αvβ3 to activate intracellular PI3K signaling (Wang et al., 2007; Riquelme et al., 2021). In addition, Ca²⁺ influx through Piezo1 also activates PI3K signaling (Zeng et al., 2022). Activated PI3K activates its downstream effector AKT through protein phosphorylation (Batra et al., 2014a). AKT, in turn, directly phosphorylates both Cx43 and integrin alpha 5 (α5) subunit (Batra et al., 2014a), which is required for the interaction between these two proteins (Batra et al., 2012). Additionally, the scaffolding molecule 14-3-3θ facilitates the delivery of Cx43 and integrin a5 from the Golgi apparatus to the plasma membrane to form mechanosensitive Cx43 hemichannels (Batra et al., 2014b). Upon mechanical stimulation, integrin α5β1 is activated and triggers the opening of hemichannels through the conformational change of the integrin (Batra et al., 2012). The opened Cx43 hemichannels mediate the export of intracellular PGE_2, whose synthesis is greatly increased by mechanical loading (Cherian et al., 2005; Siller-Jackson et al., 2008). Released PGE₂ acts in an autocrine/paracrine manner through EP2/EP4 receptors to activate both cAMP/PKA and PI3k/Akt pathways (Xia et al., 2010). These two pathways prevent osteocyte apoptosis and promote nuclear translocation and accumulation of β-catenin in osteocytes (Kitase et al., 2010), increasing Cx43 expression and gap junction formation (Xia et al., 2010). In addition, increased β-catenin suppresses the sclerostin expression in osteocytes and enhances osteoblast activity and bone formation on endosteal surfaces (Zhao et al., 2022a; Zhao et al., 2022b). Moreover, high extracellular PGE₂ acts on EP2/4 receptors to activate ERK signaling, which directly phosphorylates Cx43 to promote the closure of the Cx43 hemichannels (Riquelme et al., 2015). GJ, gap junction; HC, hemichannel.

PGE₂, Cx43 hemichannels, and mechanical stimulation in other bone cell types

Besides osteocytes, osteoblast is another mechano-responsive bone cell type (Johnson et al., 1996; Romanello et al., 2001; Liu et al., 2022). Earlier studies report that fluid flow stimulates the release of PGE₂ from primary osteoblasts and MC3T3-E1 osteoblastic cells (Duncan and Turner, 1995; Klein-Nulend et al., 1997; Saunders et al., 2001). In contrast to osteocytes, osteoblasts require high magnitudes of shear stress for PGE₂ production (Saunders et al., 2003; Genetos et al., 2005). Lower fluid flow levels induce greater PGE₂ production in MLO-Y4 osteocyte-like cells than in 2T3 osteoblasts (Kamel et al., 2010). Moreover, Genetos and others even show that fluid flow only activates hemichannels in MLO-Y4, leading to the release of PGE₂, but not osteoblastic MC3T3-E1 cells (Genetos et al., 2007). Thus, osteoblasts appear to be less mechanically sensitive to PGE₂ release than osteocytes. Furthermore, increased PGE₂ level (>10 nM) by hypergravity or compressive pressure promotes proliferation but suppresses differentiation of MC3T3-E1 cells (Ozawa et al., 1990; Miwa et al., 1991).

Interestingly, Cx43 hemichannels expressed in osteoblasts (Romanello and D’Andrea, 2001) can be regulated by mechanical stimulation (Romanello et al., 2003). However, PGE₂ release appears not to be driven by Cx43 hemichannels in osteoblasts. Cx43-null calvarial osteoblasts still respond to mechanical stimulation, as evidenced by increased dye uptake and PGE₂ release. In contrast, fluid flow-induced PGE₂ release is abolished in osteoblasts deficient in pannexin1 (Thi et al., 2012). Pannexin1 is a transmembrane channel with a similar topology as connexins that only form hemichannels but not gap junctions (Penuela et al., 2013). These findings suggest that hemichannels formed by pannexin1 and not Cx43 might be responsible for fluid flow-induced PGE₂ release in osteoblasts.

Recent studies indicate that osteoclasts are responsive to mechanical stress. Osteoclasts, after mechanical stretch, polarized to the M2 phenotype associated with YAP activation and nuclear translocation, which facilitates osteogenesis of bone marrow-derived mesenchymal stem cells (BMSCs) (Dong et al., 2021). Jiang and others find that the extracellular PGE₂, acting via EP4 receptors in osteoclasts, activates the Gαs/PI3K/AKT/MAPK signaling pathway and mediates migration and osteoclastogenesis during the progression of osteoarthritis (Jiang et al., 2022). Although it is known that osteoclasts express Cx43 on the plasma membrane and form hemichannels (Vesely et al., 1992; Ilvesaro et al., 2000; Ilvesaro and Tuukkanen, 2003), whether osteoclasts secret PGE₂ through Cx43 hemichannels remains elusive.

The signaling pathways activated by PGE₂ in osteocytes upon mechanical stimulation

PGE₂ can activate four subtypes of G-protein-coupled receptors (GPCRs), named EP1, EP2, EP3, and EP4 (Woodward et al., 2011). EP2 and EP4 are the most extensively studied in bone (Furuyashiki and Narumiya, 2011). EP2 is a mechanosensitive PGE₂ receptor whose expression can be enhanced by fluid flow in osteocytes (Cherian et al., 2003). Inhibition of the EP2 receptor by antagonist AH6809 suppresses the production of PGE₂ and Cx43 expression (Cherian et al., 2003). The fluid flow-induced PGE₂ release from osteocytes exerts autocrine effects on EP2 receptors to activate the cAMP-PKA pathway in osteocytes and increase the Cx43 gap junction formation (Cherian et al., 2003). Further study reveals that PGE₂ released from osteocytes by mechanical stimulation could lead to activation of the PI3K/Akt signaling pathway in addition to the cAMP/PKA pathway. The activation of both PI3K/Akt and cAMP/PKA pathways results in the phosphorylation and inactivation of GSK-3β (Xia et al., 2010), which is responsible for the phosphorylation of β-catenin, resulting in its ubiquitination and degradation by the 26S proteasome complex (Aberle et al., 1997). Consequently, the inactivated GSK-3β causes an increase in nuclear translocation and accumulation of β-catenin in osteocytes (Xia et al., 2010; Lara-Castillo et al., 2015). Increased nuclear β-catenin binds to the promoter region to promote Cx43 expression (Xia et al., 2010). In contrast, inhibition of PGE₂ by a COX-2 inhibitor, Carprofen, blocks the activation of β-catenin nuclear translocation in osteocytes through the PI3K/Akt activation (Lara-Castillo et al., 2015). Canonical Wnt/β-catenin signaling is proven to stimulate anabolic actions in osteocytes (Osório, 2015; Tu et al., 2015). Deletion of β-catenin in osteocytes abolishes the bone anabolic response to mechanical loading (Javaheri et al., 2014; Kang et al., 2016). Thus, osteocytic accumulation of β-catenin may be one mechanism of mechanical-induced bone formation. In addition, the Wnt/β-catenin signaling is a well-known pathway associated with cell apoptosis (Ahmed et al., 1998). Mechanical stimulation-induced PGE₂ in osteocytes blocks glucocorticoid-induced apoptosis through activated β-catenin, a downstream effector of the PI3K/Akt pathway (Kitase et al., 2010). In addition, the cAMP/PKA signaling pathway is involved in PGE₂-mediated osteocyte survival during mechanical stimulation (Kitase et al., 2010). Integrins α5β1, which requires the opening of Cx43 hemichannels in osteocytes, participate in mechanical stimulation-induced osteocyte survival through FAK/Src and the ERK pathway (Plotkin et al., 2005).

The PGE₂ released upon mechanical stimulation exerts a paracrine effect on the EP2/4 receptor to suppress the sclerostin expression through the cAMP/PKA pathway (Galea et al., 2011; Genetos et al., 2011). Sclerostin is an antagonist of canonical Wnt-β-catenin signaling in osteoblasts (Baron and Kneissel, 2013) through binding to the Wnt co-receptor Lrp5/6 (Li et al., 2005) to suppress osteogenesis (Sawakami et al., 2006). Several in vitro studies indicate the roles of osteocyte-derived PGE₂ in promoting osteoblast activity during mechanical stimulation. Increased concentration of PGE₂ released into the conditioned medium by fluid flow-loaded MLO-Y4 osteocytes promotes osteoblast differentiation (Zeng et al., 2019). In a 3D trabecular bone explant co-culture model, dynamic deformational loading can significantly increase the PGE₂ release from osteocytes in their native extracellular matrix environment and promote osteoblastic bone formation (Chan et al., 2009). PGE₂ released from osteocytes regulates osteoblast recruitment and collagen organization in the bone matrix during oscillatory fluid flow (Matsuzaka et al., 2021). Sclerostin not only inhibits bone formation but also promotes osteoclast formation. Osteocytes constitute a significant source of osteoclastogenic cytokine RANKL (Boyle et al., 2003; Xiong et al., 2011), which is also regulated by Wnt/β-catenin signaling (Donald et al., 2005; Nakashima et al., 2011). Sclerostin directly increases the levels of RANKL in osteocytes to regulate osteoclast activity by inhibiting β-catenin in osteoclasts (Wijenayaka et al., 2011). In contrast, sclerostin deficiency is resistant to bone resorption during mechanical unloading (Lin et al., 2009). Together, the PGE₂ secreted from osteocytes upon mechanical stimulation has paracrine effects through EP2 and EP4 receptors to increase β-catenin and suppress sclerostin expression in osteocytes. Increased β-catenin in osteocytes promotes Cx43 expression, gap junction formation, mechanosensitivity, and survival of osteocytes. Suppression of sclerostin secretion promotes osteoblast activity and inhibits osteoblast activity.

Cx43 hemichannels in bone development under physiological level of mechanical stress

Previous studies using Cx43 knockout and transgenic mouse models provide insightful information regarding the importance of Cx43 hemichannels in bone development under physiological mechanical conditions. Although global deleting Cx43 (Cx43^−/−) caused early postnatal death due to an obstruction of the right ventricular outflow tract, they showed retarded intramembranous, endochondral ossification, craniofacial abnormalities, and osteoblast dysfunction (Lecanda et al., 2000; Thi et al., 2010; Chaible et al., 2011; Ishikawa et al., 2016). Moreover, mice with specific deletion of Cx43 by expressing the Cre recombinases in osteoprogenitors (DM1-Cre;Cx43^-/flx) (Watkins et al., 2011), preosteoblasts (Osx1-Cre;Cx43^flx/flx) (Hashida et al., 2014), early osteoblasts/osteocytes (Col1α1-Cre;Cx43^-/flx) (Castro et al., 2003; Chung et al., 2006), mature osteoblasts/osteocytes (OCN-Cre;Cx43^-/flx) (Bivi et al., 2012a), and osteocytes (8-kb DMP1-Cre;Cx43^flx/flx) (Bivi et al., 2012a; Bivi et al., 2012b) all showed thinner cortical thickness, larger marrow area, and total cross-sectional area. The change of cortical bone structure in these cKO mice was due to both increased periosteal osteoblastic bone formation (Watkins et al., 2011; Bivi et al., 2012a; Watkins et al., 2012; Pacheco-Costa et al., 2015) and an even greater unbalanced increase in endosteal osteoclastic bone resorption (Watkins et al., 2011; Bivi et al., 2012a; Watkins et al., 2012; Lloyd et al., 2013). In addition, Cx43 overexpressing in osteocytes (Cx43^OT) preserves osteocyte viability and bone formation to ameliorate age-induced cortical bone loss (Davis et al., 2018). It is worth noting that Cx43 deficiency affects the production and release of PGE₂ in osteoblasts and osteocytes. A lower amount of PGE₂ is found in the primary calvaria cells of Cx43^−/− mice than in wild-type (WT) mice (Grimston et al., 2006). The absence of Cx43 in osteocytes by Cx43 shRNA attenuates PGE₂ synthesis by COX-2 (Bivi et al., 2013). Although Cx43 knockout during bone developmental stages from the early stage (DM1-Cre) to the late stage (DMP1-Cre) of osteoblast differentiation shows the similar bone structure, Cx43 deficiency abolishes both Cx43 gap junctions and hemichannels. Thus, it is impossible to determine whether Cx43 gap junctions or/and hemichannels are responsible for the observed phenotypes. It is worth noting that other cell types may play indirect roles in osteoblast differentiation after knocking out Cx43. For example, Cx43 in osteoblasts/osteocytes indirectly modulates skeletal muscle growth and function (Shen et al., 2015; Li et al., 2021). In turn, skeletal muscle can also influence bone growth by releasing osteogenic myokines (Hamrick et al., 2010). Thus, the roles of Cx43 hemichannel in bone and other tissue crosstalk need to be further studied.

Besides Cx43-deficient mice, several Cx43 gene mutations can lead to a skeletal disease called oculodentodigital dysplasia (ODDD), with phenotypic presentations of syndactyly, craniofacial abnormalities, and long broad bones (Paznekas et al., 2003). To date, four mouse strains (Cx43^I130T/+ (Kalcheva et al., 2007), Cx43^Jrt(G60S)/+ (Flenniken et al., 2005), Cx43^G138R/+ (Dobrowolski et al., 2008) and Cx43^K258Stop/- (Pacheco-Costa et al., 2015; Moorer et al., 2017)) with missense point mutations in one allele of the Cx43 gene are generated to mimic the phenotypes of ODDD. The mutations alter Cx43 protein conformation, thus leading to changed hemichannel activities (Dobrowolski et al., 2007). Increased (Dobrowolski et al., 2008) or decreased (Kalcheva et al., 2007) Cx43 hemichannel functions are found in these Cx43 mutants, indicating that normal Cx43 hemichannel function is crucial in maintaining bone structure.

Our group has generated two transgenic mouse models to dissect the function of Cx43 gap junctions and hemichannels in osteocytes, respectively. A 10 kb-DMP1 promoter drives the two transgenic mouse models R76W and ∆130–136 with the overexpression of dominant negative Cx43 mutants in osteocytes (Xu et al., 2015). The R76W site mutant, of which Cx43 amino acid residue arginine-76 (R) is replaced by tyrosine, inhibits gap junctions. The ∆130–136 mutant with the deletion of seven residues in the cytoplasmic loop of Cx43 protein at amino acids 130–136 inhibits both gap junctions and hemichannels. The bone phenotype in R76W mice is mostly like WT mice, except for increased endosteal osteoclast activity and bone remodeling markers in serum. In contrast, the ∆130–136 mice exhibit increased osteocyte apoptosis, endosteal resorption, and periosteal bone formation, resulting in higher tissue, cortical, and marrow cavity area of femoral midshaft at the femoral mid-diaphysis. The bone phenotypes in ∆130–136 mice are similar, but even more severe, than osteocyte-specific Cx43 cKO mice driven by 8-kb DMP1 promoter (Bivi et al., 2012a), indicating Cx43 deficiency in osteocytes impairs the functions of Cx43 hemichannel in bone development. Compared to WT and R76W, the ∆130–136 mice show lower PGE₂ levels in tibia diaphysis (Zhao et al., 2022a). As discussed above, PGE₂ released by osteocytes mediated by Cx43 hemichannels is proven to maintain osteocyte survival and prevent their apoptosis (Kitase et al., 2010). Consistently, both ∆130–136 and 8-kb DMP1-Cre;Cx43 cKO mice show increased osteocyte apoptosis in cortical bone. Impaired Cx43 hemichannels of osteocytes are also found in osteocyte-specific integrin a5 cKO mice driven by 10-kb DMP1 promoter, which expresses lower serum PGE₂ levels and increased osteocytes apoptosis and cortical thickness in tibias (Zhao et al., 2022b). Thus, under a physiological mechanical environment, Cx43 hemichannels and PGE₂ in osteocytes likely play a predominant role in osteocyte vitality and bone structure.

PGE₂ release through osteocytic Cx43 hemichannels promotes bone anabolism upon mechanical loading

PGE₂ stimulates bone resorption and formation (Jee and Ma, 1997; Blackwell et al., 2010). Continuous PGE₂ treatment decreases cancellous bone mass due to bone resorption exceeding bone formation (Tian et al., 2007). Whereas moderate PGE₂ treatment by intraperitoneal injection (Jee et al., 1985; Ueno et al., 1985) and local metaphyseal injection (Welch et al., 1993; Yang et al., 1993) increases both trabecular and cortical bone mass in growing rats. In addition, PGE₂ prevents bone loss induced by ovariectomy (Mori et al., 1990; Harada et al., 1995), disuse (Jee et al., 1992), and orchidectomy (Li et al., 1995) in rats. Recent studies show that PGE₂ activates EP4 in sensory nerves to promote bone formation and inhibit adipogenesis by inhibiting sympathetic activity through the central nervous system (Chen et al., 2019; Hu et al., 2020). The release of PGE₂, a known direct product of bone mechanical stimulation, has important anabolic effects on the skeleton. In healthy women, a rapid and significant increase of PGE₂ levels in the proximal tibial metaphysis is observed using the microdialysis technique in response to weight-bearing mechanical loading (Thorsen et al., 1996). Inhibition of PGE₂ by a COX-2 inhibitor, NS-398, completely blocks tibial bone formation induced by four-point bending loading in rats (Forwood, 1996; Li et al., 2002). On the contrary, activation of the PGE₂ receptor using ONO-4819 (agonist for prostaglandin E receptor subtype EP4) has an additive effect on bone formation in response to mechanical loading (Hagino et al., 2005). The involvement of Cx43 in the anabolic function of mechanical loading and PGE2 has been reported in earlier studies. Mice with Cx43 deficiency show altered bone anabolic response to mechanical loading. Col1α1-Cre;Cx43 cKO mice show attenuated tibial endosteal bone formation during non-physiological four-point (Grimston et al., 2006) or three-point tibial bending (Grimston et al., 2008). A lower level of PGE₂ is found in the primary calvaria cells from Cx43^−/− mice than in WT mice during mechanical stretching (Grimston et al., 2006). In DM1-Cre;Cx43 cKO mice, axial tibia loading results in a greater decrease of endosteal bone formation compared to WT mice (Grimston et al., 2012). However, Cx43 deficiency has a positive effect on periosteal bone formation. Deletion of Cx43 in osteoblasts/osteocytes (DM1-Cre;Cx43 cKO and OCN-Cre;Cx43 cKO mice) showed an enhanced tibial periosteal response to tibial axial compression (Grimston et al., 2012) or tibial cantilever bending (Zhang et al., 2011). Similarly, deletion of Cx43 in osteocytes (DMP1–Cre; Cx43 cKO mice) showed enhanced periosteal bone formation in response to ulna compression (Bivi et al., 2013). Nevertheless, these cKO mice have both impaired gap junctions and hemichannels, as well as potential channel-independent roles of Cx43. Thus, the specific role of gap junctions and hemichannels formed by Cx43 hemichannels in response to mechanical loading in the bone cannot be resolved with Cx43 deletion models.

Recently, our group find a close relationship between Cx43 hemichannels and PGE₂ release in skeletal response to mechanical loading in vivo. In this study, PGE₂ levels in the diaphysis are significantly increased in WT and gap junction impaired R76W upon tibial cyclic compression loading. Increased PGE₂ level suppresses the sclerostin expression in osteocytes and bone formation on the endosteal surface. However, Δ130-136 mice with impaired gap junctions and hemichannels show unchanged PGE₂ levels and sclerostin expression in osteocytes. As a result, the increased bone mass caused by mechanical loading is not seen in Δ130-136 mice. Attenuated bone formation and increased resorption on the endosteal surface lead to the enlargement of the bone marrow cavity and inhibited bone mass gain (Zhao et al., 2022a). The data points to the role of Cx43 hemichannels in mediating PGE₂ release and the anabolic action of mechanical loading. To further investigate whether the changed mechanical response in Δ130-136 mice is due to Cx43 hemichannels, a specific mouse monoclonal blocking antibody Cx43 (M1) was developed and used to investigate the role of Cx43 hemichannels in vivo. Administration of this particular antibody exhibits similar effects as Δ130-136 mice, including the attenuated PGE₂ level and inhibited anabolic response to mechanical loading on the endosteal surface. PGE₂ administration, however, can rescue the attenuated endosteal osteogenic response to mechanical loading impeded by the Cx43 (M1) antibody. PGE₂ acts in a paracrine manner to suppress sclerostin expression in vitro (Galea et al., 2011; Genetos et al., 2011). These in vivo studies demonstrate that Cx43 hemichannels activated by mechanical stimulation release PGE₂ from osteocytes to suppress sclerostin expression in osteocytes and enhance osteoblast activity and bone formation on endosteal surfaces. We further demonstrate the important role of Cx43 hemichannels and PGE₂ release in bone anabolic response to mechanical loading through the use of osteocyte-specific integrin a5 cKO mice driven by a 10-kb DMP1 promoter (Zhao et al., 2022b). Since the interaction between integrin α5 and Cx43 is essential for the hemichannel opening by mechanical loading (Batra et al., 2012), integrin α5 deficiency impedes load-induced Cx43 hemichannel opening and PGE₂ release (Batra et al., 2012; Zhao et al., 2022b). Integrin α5 cKO mice in osteocytes exhibit attenuated loading effects on catabolic sclerostin reduction and anabolic β-catenin increase, contributing to decreased endosteal osteoblasts and bone formation (Figure 1). Our studies show that the anabolic effect of Cx43 hemichannel-released PGE₂ is on the endosteal surface (Zhao et al., 2022a; Zhao et al., 2022b). Consistently, previous studies also report that the effect of PGE₂ on bone anabolic response to mechanical loading appears to be more on the endosteal surface than the periosteal surface (Forwood, 1996; Li et al., 2002; Hagino et al., 2005).

Interestingly, inhibited Cx43 hemichannels in Δ130-136 mice, Cx43 (M1) group, and integrin α5 cKO mice showed an increase of osteoclast activity on the endosteal surface during mechanical loading. Whether this catabolic response is related to PGE₂ requires further investigation. The mechanosensitivity of bone is gradually diminished during aging (Holguin et al., 2014; Holguin et al., 2016), accompanied by impaired mechanotransduction (Chalil et al., 2015) and decreased Cx43 expression in osteocytes (Kar et al., 2013). We speculate that attenuated Cx43 expression in osteocytes is related to abolished Cx43 hemichannel activity and PGE₂ release in aged bone. Indeed, significantly reduced PGE₂ EP4 receptors in the sensory nerve of aged bone attenuate the sensibility to changes in bone metabolism (Lv et al., 2022).

Cx43 hemichannels play a protective role against bone loss during mechanical unloading

Reduced or no mechanical loading, such as long bed rest (Spector et al., 2009), and astronauts in space missions (Keyak et al., 2009), harms skeletal health, which is associated with an imbalanced bone turnover. Suppressed osteoblastic bone formation and activated osteoclastic bone resorption (Bikle and Halloran, 1999; Kondo et al., 2005) increase bone loss and fracture risk. Previous in vitro studies show that Cx43 hemichannels participate in response to mechanical unloading. Zero-gravity by parabolic flight decreases Cx43 expression in osteocytes (Di et al., 2011). Simulated microgravity by a random position machine (RPM) increases the activity of Cx43 hemichannels and the release of PGE₂ (Xu et al., 2017). Dominant negative integrin β1 mutants driven by an OCN promoter show a lower cancellous bone mass in the distal femoral metaphysis caused by increased bone resorption and decreased bone formation during short-term hindlimb unloading, a model mimicking unloading/disuse (Iwaniec et al., 2005). Given that integrin a5β1 regulates the opening of the Cx43 hemichannels, this study implies the vital role of Cx43 hemichannels in bone response to unloading. High extracellular PGE₂ due to sustained opening of hemichannels during unloading (Xu et al., 2017) resulted in osteoclast resorption and bone loss (Coon et al., 2007; Knippenberg et al., 2007; Tian et al., 2007). Consistently, deletion of Cx43 in osteoblast/osteocyte-specific Cx43 cKO mice driven by the 2.3-kb Col1α1 promoter (Grimston et al., 2011) and OCN promoter (Lloyd et al., 2012; Lloyd et al., 2013) show protection against unloading-induced bone loss. In contrast, our previous study shows that enhanced Cx43 hemichannels in R76W mice protect from osteocyte apoptosis in cortical bone during mechanical unloading (Zhao et al., 2020). PGE₂ is an inhibitor of sclerostin expression, and both sclerostin and PGE₂ inhibitors are known to be associated with cell apoptosis (Ahmed et al., 1998). It is assumed that PGE₂ from Cx43 hemichannels has a protective role against osteocyte apoptosis. Multifaced roles of Cx43 could cause the difference seen between Cx43 deletion and hemichannel-impaired models. Further studies are needed to establish the underlying mechanisms of Cx43 hemichannels and PGE₂ in response to mechanical unloading.

Future perspectives

With research advances and the development of new transgenic mouse models, the inter-relationship between Cx43 hemichannels and extracellular PGE₂ in mediating osteogenic response to mechanical loading is becoming more evident. However, several questions remain, including 1) if extracellular PGE₂ released by hemichannels upon mechanical loading has any direct action on osteoclasts; 2) if the loading-induced bone anabolic response is partly regulated through the influence of osteocyte-released PGE₂ on the nerve system. Early studies reported that temporarily blocking peripheral neurons by bupivacaine reduces bone formation in compressed ulna (Sample et al., 2008). Recently, PGE₂ has been found to act on sensory neurons and affect sympathetic nerve activity, and regulate bone homeostasis (Chen et al., 2019; Lv et al., 2021). It is speculated that the PGE₂ released from Cx43 hemichannels in osteocytes may not only have autocrine and paracrine effects on osteocytes and the other types of bone cells, respectively, but may also systematically influence sensory nerves. The interactions of bone cells and sensory neurons regulated by PGE₂ remain largely unknown. These could all be potential directions for future research. Moreover, further research on unveiling the mechanism of action for hemichannels and PGE₂ may help in the discovery and development of potential therapeutics that aid in treating bone loss, in particular, in the elderly population with lost sensitivity to anabolic responses to mechanical stimulation.

Author contributions

DZ, JW and JJ drafted the manuscript, FA and JJ conceived the idea and revised the manuscript. All authors contributed to manuscript revision read and approved the submitted version.

This work was supported by the National Institutes of Health (NIH) Grants: AR072020 (to JJ) and F32DK134051 (to FA), and Welch Foundation grant: AQ-1507 (to JJ).

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Batra, N., Riquelme, M. A., Burra, S., Kar, R., Gu, S., and Jiang, J. X. (2014b). Direct regulation of osteocytic connexin 43 hemichannels through AKT kinase activated by mechanical stimulation. J. Biol. Chem. 289, 10582–10591. doi:10.1074/jbc.M114.550608

Bivi, N., Nelson, M. T., Faillace, M. E., Li, J., Miller, L. M., and Plotkin, L. I. (2012b). Deletion of Cx43 from osteocytes results in defective bone material properties but does not decrease extrinsic strength in cortical bone. Calcif. Tissue Int. 91, 215–224. doi:10.1007/s00223-012-9628-z

Castro, C. H., Stains, J. P., Sheikh, S., Szejnfeld, V. L., Willecke, K., Theis, M., et al. (2003). Development of mice with osteoblast-specific connexin43 gene deletion. Cell. Commun. Adhes. 10, 445–450. doi:10.1080/cac.10.4-6.445.450

Chaible, L. M., Sanches, D. S., Cogliati, B., Mennecier, G., and Dagli, M. L. (2011). Delayed osteoblastic differentiation and bone development in Cx43 knockout mice. Toxicol. Pathol. 39, 1046–1055. doi:10.1177/0192623311422075

Davis, H. M., Aref, M. W., Aguilar-Perez, A., Pacheco-Costa, R., Allen, K., Valdez, S., et al. (2018). Cx43 overexpression in osteocytes prevents osteocyte apoptosis and preserves cortical bone quality in aging mice. JBMR Plus 2, 206–216. doi:10.1002/jbm4.10035

Dobrowolski, R., Sasse, P., Schrickel, J. W., Watkins, M., Kim, J. S., Rackauskas, M., et al. (2008). The conditional connexin43G138R mouse mutant represents a new model of hereditary oculodentodigital dysplasia in humans. Hum. Mol. Genet. 17, 539–554. doi:10.1093/hmg/ddm329

Grimston, S. K., Watkins, M. P., Brodt, M. D., Silva, M. J., and Civitelli, R. (2012). Enhanced periosteal and endocortical responses to axial tibial compression loading in conditional connexin43 deficient mice. PLoS One 7, e44222. doi:10.1371/journal.pone.0044222

Hagino, H., Kuraoka, M., Kameyama, Y., Okano, T., and Teshima, R. (2005). Effect of a selective agonist for prostaglandin E receptor subtype EP4 (ONO-4819) on the cortical bone response to mechanical loading. Bone 36, 444–453. doi:10.1016/j.bone.2004.12.013

Kalcheva, N., Qu, J., Sandeep, N., Garcia, L., Zhang, J., Wang, Z., et al. (2007). Gap junction remodeling and cardiac arrhythmogenesis in a murine model of oculodentodigital dysplasia. Proc. Natl. Acad. Sci. U. S. A. 104, 20512–20516. doi:10.1073/pnas.0705472105

Kang, K. S., Hong, J. M., and Robling, A. G. (2016). Postnatal β-catenin deletion from Dmp1-expressing osteocytes/osteoblasts reduces structural adaptation to loading, but not periosteal load-induced bone formation. Bone 88, 138–145. doi:10.1016/j.bone.2016.04.028

Lara-Castillo, N., Kim-Weroha, N. A., Kamel, M. A., Javaheri, B., Ellies, D. L., Krumlauf, R. E., et al. (2015). In vivo mechanical loading rapidly activates β-catenin signaling in osteocytes through a prostaglandin mediated mechanism. Bone 76, 58–66. doi:10.1016/j.bone.2015.03.019

Li, J., Burr, D. B., and Turner, C. H. (2002). Suppression of prostaglandin synthesis with NS-398 has different effects on endocortical and periosteal bone formation induced by mechanical loading. Calcif. Tissue Int. 70, 320–329. doi:10.1007/s00223-001-1025-y

Paznekas, W. A., Boyadjiev, S. A., Shapiro, R. E., Daniels, O., Wollnik, B., Keegan, C. E., et al. (2003). Connexin 43 (GJA1) mutations cause the pleiotropic phenotype of oculodentodigital dysplasia. Am. J. Hum. Genet. 72, 408–418. doi:10.1086/346090

Plotkin, L. I., Mathov, I., Aguirre, J. I., Parfitt, A. M., Manolagas, S. C., and Bellido, T. (2005). Mechanical stimulation prevents osteocyte apoptosis: Requirement of integrins, src kinases, and ERKs. Am. J. Physiol. Cell. Physiol. 289, C633–C643. doi:10.1152/ajpcell.00278.2004

Riquelme, M. A., Burra, S., Kar, R., Lampe, P. D., and Jiang, J. X. (2015). Mitogen-activated protein kinase (MAPK) activated by prostaglandin E2 phosphorylates connexin 43 and closes osteocytic hemichannels in response to continuous flow shear stress. J. Biol. Chem. 290, 28321–28328. doi:10.1074/jbc.M115.683417

Sugiatno, E., Samsudin, A. R., Ibrahim, M. F., and Sosroseno, W. (2006). The effect of nitric oxide on the production of prostaglandin E2 by hydroxyapatite-stimulated a human osteoblast (HOS) cell line. Biomed. Pharmacother. 60, 147–151. doi:10.1016/j.biopha.2006.02.008

Thi, M. M., Urban-Maldonado, M., Spray, D. C., and Suadicani, S. O. (2010). Characterization of hTERT-immortalized osteoblast cell lines generated from wild-type and connexin43-null mouse calvaria. Am. J. Physiol. Cell. Physiol. 299, C994–C1006. doi:10.1152/ajpcell.00544.2009

Woodward, D. F., Jones, R. L., and Narumiya, S. (2011). International union of basic and clinical pharmacology. LXXXIII: Classification of prostanoid receptors, updating 15 years of progress. Pharmacol. Rev. 63, 471–538. doi:10.1124/pr.110.003517

Xia, X., Batra, N., Shi, Q., Bonewald, L. F., Sprague, E., and Jiang, J. X. (2010). Prostaglandin promotion of osteocyte gap junction function through transcrip-tional regulation of connexin 43 by glycogen synthase kinase 3/βcatenin signaling. Mol. Cell. Biol. 30, 206–219. doi:10.1128/mcb.01844-08

Zeng, Y., Riquelme, M. A., Hua, R., Zhang, J., Acosta, F. M., Gu, S., et al. (2022). Mechanosensitive piezo1 calcium channel activates connexin 43 hemichannels through PI3K signaling pathway in bone. Cell. Biosci. 12, 191. doi:10.1186/s13578-022-00929-w

Zhang, Y., Paul, E. M., Sathyendra, V., Davison, A., Sharkey, N., Bronson, S., et al. (2011). Enhanced osteoclastic resorption and responsiveness to mechanical load in gap junction deficient bone. PLoS One 6, e23516. doi:10.1371/journal.pone.0023516