In this section, we focus on how membrane microdomains participate in cell migration and subsequent fusion at the OC precursor (pOC) stage and how these microdomains mediate the osteolytic and secretory functions of OCs during the mOC stage.

Migration of pOCs: membrane microdomains serve as platforms

During OC culture in vitro, lamellipodia are often observed; they represent the direction of cell extension and are considered the hallmark structure of early OC development.32,33,34 Here, we clarify the basic functions of these membrane microstructures structures and the key proteins that constitute them.

When pOCs migrate, the structure at the leading edge of the cell dynamically extends and retracts.35 OCs without podosomes have spicule-like structures, which are referred to as lamellipodia, as observed after knockout of Cortactin.36 This structure has also been reported in the literature by Akisaka et al.34 The activation and formation of lamellipodia are generally believed to be induced by the Arp2/3 complex consisting of microfilamentous nucleation factors, in which the nucleation-promoting factors Wiskott–Aldrich syndrome protein (WASP) and WASP family verprolin-homologous protein (WAVE) play important roles.34,35 Previous findings suggest that lamellipodia are actin-based structures and that their formation depends on regulatory factors.

Lamellipodia are common in migratory cells such as fibroblasts, and their general characteristics include the broad protrusion of the leading edge and an edge that can roll back from the membrane ruffle.37,38 Whether lamellipodia function similarly in fibroblasts and OCs is unclear. Through the use of transmission electron microscopy (TEM), Domon T et al.33 showed that the morphology of migrating OCs was irregular and flat, and they confirmed that these cells have lamellipodia, indicating that via this migratory structure, OCs can move on dentin.33 Mature osteoclasts (mOCs) also appear to have lamellipodia on their membrane. mOCs cultured in vitro exhibited stretched out lamellipodia that can mechanically decompose substrates and bring the substrates to the surface of the cell body via retraction of the lamellipodia.39 Although lamellipodium-like structures were observed in this study, the specific differences between lamellipodia in the pOC and mOC stages remain unclear. Notably, we initially focus our attention on lamellipodia during the migratory phase.

Lamellipodia determine the direction of cell migration, and lamellipodium stretching requires the actin network; therefore, we need to identify the scaffolding and regulatory proteins that determine the OC membrane structure.40 Structurally, focal adhesion anchors the cell to the matrix and thus provides the mechanical force needed for actin contraction.37 Additionally, cells migrate not only by extending lamellipodia, which are formed by scaffolds via an actin network but also by extending filopodia, which contain only part of the actin bundle and aggregate to form lamellipodia.39,41 Hence, matrix anchoring and actin movement are the two critical components of pseudopod formation and movement.37 We conclude that matrix anchoring is dependent mainly on membrane proteins to form the first structural platform, while the actin that is recruited forms the second platform and mediates contractile motility based on regulatory factors to establish actin flow.37,42,43

In OCs, the initiation site of the lamellipodium membrane microdomain may be composed of integrins or other adhesion receptors (e.g., other receptor tyrosine kinases (RTKs) such as colony-stimulating factor 1 receptor, also known as c-FMS44) that activate downstream regulatory proteins, including GTPases, protein kinases, and phosphatases, to induce the ARP2/3-related actin network and thus form the “second platform”.45,46 Considering the work of Boujemaa–Paterski, Rajaa et al., Geiger, Benjamin et al. and Fukunaga, Tomohiro et al., we propose a model based on integrin adhesion in OCs. (1) Integrins recruit and activate vinculin via talin to form nascent adhesions. (2) Vinculin is recruited and binds highly branched F-actin networks and contracts to establish actin flow, at which point the adhesions mature. (3) Centripetal actin flow at 1–3 μm·min−1 may stimulate the maturation of other nascent adhesions. (4) Eventually, many adhesions accumulate, enhancing mechanical resistance and ultimately leading to expansion of the leading edge of the lamellar pseudopod (Fig. 3).42,43,47 Among the integrins highly expressed in OCs, αvβ348 but not other integrin subunits has been shown to colocalize with vinculin, talin, and arp2/3. RTKs, such as epidermal growth factor receptor (EGFR), can also form the first platform based on its regulation of downstream PI3K, SRC, RAS, and RAC expression, and the modulation of the ARP2/3 complex and WASP affects the formation of the actin network and lamellipodia.32,49,50,51,52 Although evidence to support a role for RTKs in lamellipodium formation in OCs is insufficient, the inhibitory effect of RTK inhibitors on osteoclasts suggests that RTKs may be involved, which warrants further study.53,54 Other adhesion-related proteins, including cadherin, also remain to be investigated as regulators of lamellipodium formation.55,56,57 Notably, many lattice-like protein sheets have been found at the edge of filamentous pseudopodia.58 Although not colocalized with the OC actin network, this protein lattice is tightly bound to the apatite surface and may act as an adhesive rather than an endocytic agent. Therefore, it remains unclear whether this part of the protein lattice can serve as a scaffold for establishing the structural domain of the lamellar pseudopod membrane.

Fig. 3

Lamellipodia and their formation. a Schematic diagram at the macroscopic level: the process of lamellipodium formation. pOCs form filamentous pseudopodia, and their fusion drives lamellipodium formation, which determines the direction of cell migration. b Schematic diagram at the microscopic level: the process of integrin adhesion promoting lamellipodium formation. Longitudinal sections of lamellipodia show that integrins recruit the regulatory proteins talin and vinculin, which regulate actin skeleton remodeling mediated via Arp2/3 to initiate reverse actin flow and mediate pseudopod contraction on the basis of the counteracting force provided by the integrin adhesion bodies. In this process, integrins and regulatory proteins form the scaffolds of the pseudopod membrane microdomains and then integrate actin, leading to the formation of membrane macrostructures

In conclusion, the identification of essential assembly sites in the microstructural domain of the lamellar pseudopod membrane may facilitate the development of locally acting regulators of early OC polarization. Although it remains unclear whether adhesion receptors in addition to integrins are involved in the assembly process, targeted regulation of the “first platform” and “second platform” in lamellipodium formation may facilitate the selective regulation of OC functions.

OC fusion: membrane microdomain interactions with the actin cytoskeleton

The actin cytoskeleton of OCs is a dynamic structure that changes rapidly during cell migration, fusion, and resorption. The membrane microdomains of OCs need to be supported by the actin cytoskeleton, and when cortical actin is reconstructed, the cell membrane structure changes accordingly.

During the fusion phase of the OC life cycle, the actin cytoskeleton promotes the extension of filopodia between cells or actin flow, which results in the formation of the characteristic TNT membrane domain structure (Fig. 4) or ZLS (Fig. 5) to trigger fusion.59,60 In the early stage of fusion, monocytes rely on their TNTs to fuse with another monocyte and thus generate multinucleated cells.60 The later stage is dominated by the formation of ZLSs between multinucleated cells and their fusion partners.61

Fig. 4

Early fusion: OCs fuse through TNTs. a Two mechanisms explain TNT formation: filamentous pseudopods extend between fusion partners or nearby fusion partners that have separated from each other by the action of chemokines, and a TNT is formed at the interconnection of the plasma membrane between fusion partners. b Nuclear translocation is possible when a TNT has (1) a diameter in the range of 5–20 µm and (2) an open interconnection inside the duct. c The processes and mechanisms by which membrane microdomains mediate nucleus transport

Fig. 5

Late fusion: OCs undergo multinucleated cell–multinucleated cell and multinucleated cell–mononuclear cell fusion through ZLS structures. a Fusion partners are closely linked through actin flow, and the actin cytoskeleton forms a ZLS structure at a contact point. b The structure of the ZLS membrane microdomain, including the surface membrane proteins and the internal actin complex

TNT-associated membrane domains, not filopodia, are the keys to early fusion

The conventional view is that at the early stage of OC fusion, pOCs filopodia protrude to initiate fusion with partners. Although the role of filopodia has been demonstrated, many questions, such as how the nucleus is delivered and how filopodia trigger fusion, remain unanswered.59,62 Recently, a TNT, which is a very thin membrane tube, was found at the head of a filopodium where two cells contacted each other.60 Therefore, TNTs, not filopodia, may directly participate in the connection between fusion partners and drive material transport. TNTs are thought to be important in cell communication among bone marrow-derived cells (including macrophages, OCs, and dendritic cells) and in the fusion of macrophages and pOCs. Therefore, to clarify the role of TNT-related membrane domains in OC fusion, we need to define which TNTs can mediate fusion and the key mechanisms through which TNT-related membrane domains are involved in fusion.31,63

How do TNT-associated membrane microdomains form?

First, as described by McCoy–Simandle, Kessler et al., a TNT is identified according to the following three phenotypic criteria: (1) it connects at least two cells, (2) it contains F-actin, and (3) it does not attach to the matrix but extends from filopodia. This definition can be used to distinguish a TNT from any other F-actin-rich structure, and a TNT may be considered a special membrane microdomain.64 A TNT is generated in two situations: when filopodia protrude between fusion partners and when two cells located next to each other are separated under the action of chemokines.65 In these cases, a tube is formed in the plasma membranes where the fused cells are connected and cellular components such as organelles are transported (Fig. 4a).

Which TNTs participate in OC fusion?

TNTs are designated closed or open depending on whether they are connected to the target cell.65 Previous studies have mainly suggested that closed-end TNTs mediate gap junction formation, but after their conversion into open-ended TNTs, TNTs are known to participate in a process similar in virus‒cell membrane fusion or cell‒cell fusion.66,67 In addition, TNTs are classified into two functionally distinct types according to their size: (1) those with a diameter less than 5 µm are thin TNTs and contain only F-actin, and (2) those with a diameter ranging from 5 to 20 µm are thick TNTs and contain both F-actin and microtubules. Previous studies have revealed that large organelles, including lysosomes, mitochondria, and even nuclei, can be transferred only through thick TNTs (Fig. 4b).60,65

Therefore, although TNT-associated membrane microdomains spanning pOCs at the fusion stage have been observed,60 it is thought that TNTs participate in OC fusion only when their diameter is in a specific range (5–20 µm) and when there is intercommunication within the tunneling tube.

How do TNT-associated membrane microdomains participate in OC fusion?

M-Sec is a key factor in the formation of a TNT; its expression is upregulated during osteoclastogenesis, and M-Sec depletion significantly inhibits OC fusion by inhibiting TNT formation.31,68 Nonetheless, the specific mechanism through which TNT-associated membrane microdomains mediate cell fusion remains to be elucidated. pOCs recognize distant fusion partners through long intercellular F-actin structures. When two cells approach each other, thin and short actin protrusions (approximately 10 μm) can be observed on the leading edge of the cells.69 Nuclei have also been observed in these structures,70 suggesting that the nucleus may be transported through tubes formed by the actin cytoskeleton, which may trigger prophase fusion.

Based on these findings, we asked the following question: What is required for TNTs to mediate fusion? (Fig. 4c).

(1)

Membrane proteins, the surface proteins in the membrane domain of TNTs, can recruit the actin cytoskeleton. Two types of membrane proteins involved in TNT function have been found in Ocs: DC-STAMP,71 which shows transport activity, and connexins, including CD3672 and CX-43.73 Although these proteins have been identified, the components of the TNT shell have not been fully characterized, and the exact mechanisms underlying TNT functions remains unclear.

(2)

Actin-related regulatory proteins: Myosin is critical for providing power to F-actin and is often recruited to the membrane domain. Myosin 10 (MyoX) has been identified as a molecular motor that regulates TNT formation. This unconventional myosin is specifically expressed in OCs.74 As shown through in vitro experiments, pOCs remained in a monocyte state after MyoX expression was reduced by shRNA. This result was largely obtained to MyoX binding to microtubules through its MyTH4 tail domain, regulating F-actin cytoskeleton dynamics to promote the formation of an ordered TNT. Moreover, DC-STAMP, a transmembrane protein in the structural domain of the TNT membrane, penetrates other precursor cells by further interacting with the F-actin backbone to achieve migration through TNTs.31

(3)

Interactions between actin and the perinuclear cytoskeleton: The nucleus is sometimes located within the microtubule-actin network, which mediates its transport, and the microtubule–actin filaments usually originate from the perinuclear region, which suggests that the nucleus and F-actin are closely related. Moreover, some regulatory proteins play irreplaceable roles in nucleus-related F-actin dynamics. The actin-binding ARP2/3 complex stabilizes bent and branched actin structures, whereas c-Src and cortactin colocalize with F-actin at the cell periphery, which suggests that the latter may participate in the rearrangement and stabilization of bent and branched F-actin networks.70 In addition, c-Src, cortactin, cofilin, and actin can accumulate around the nucleus, suggesting that their involvement in nuclear movement might partially involve the regulation of nucleus delivery via thick TNTs.70,75,76

Summary

Thus, TNT-associated membrane microdomains facilitate the transport of substances, including nuclei, and this process requires signal recognition mediated by surface molecules, including DC-STAMP, and interactions between F-actin and the perinuclear cytoskeleton. However, only a fraction of the relevant proteins in a TNT have been identified, and the proposed structural domain of the TNT membrane suggests that the scaffolding proteins in this structure not only include marker proteins of intercellular connections but also bind the intracellular actin cytoskeleton to the perinuclear frame. Here, we summarize only some of the components involved in these intercellular linkages, as their specific relation to nuclear transport events via natural scaffolding proteins remains to be discovered.

ZLS-associated membrane microdomains are key for multinucleated cell fusion What are ZLS-associated membrane microdomains?

After single-nucleated precursor cells fuse to form multinucleated cells, they still need to combine with other fusion partners to form multinucleated OCs with more than three nuclei and podosome belts.59,61,77 The experiments conducted by Takito, Jiro et al. revealed that the F-actin cytoskeleton of multinucleated cells agglomerates form a zipper-like F-actin structure when in contact with other multinucleated cells, which has also been shown to be the basic manner through which multinucleated cells fuse.59,78 Therefore, elucidation of the ZLS membrane microdomain is extremely important to clarify the life cycle of OCs.

Formation and function of the ZLS

A ZLS and its associated membrane microdomains have attracted our interest. Membrane proteins in a TNT may recruit actin-related regulatory proteins by downstream signaling to then associate with cortical actin in the perinuclear area. In contrast to TNT-related membrane microdomains, ZLSs appear to mediate the closeness between two precursor OCs through a complex composed of F-actin and regulatory proteins. Force may be critical in directly promoting fusion events. When mononuclear and multinuclear cells collide with actin rings, the cell membranes at the collision site move in response to actin flow, leaving the plasma membranes close together, and F-actin condenses on the plasma membrane to form a cluster of ZLSs (Fig. 5).78 The formation of this structure is the basis for subsequent cell membrane fusion events. Subsequently, when plasma membranes are fused via actin flow, the ZLSs are reconstituted, and cortical actin is cleaved to form a foot vesicle band. Thus, these multinucleated cells fused by ZLSs can give rise to larger OCs (Fig. 5).59,78 Clarifying the mechanism through which the membrane microdomain and F-actin cytoskeleton induce ZLSs will help us elucidate the key mechanisms underlying the later stages of OC fusion.

Motility forces of the actin cytoskeleton and ZLS-associated membrane microdomains

Published studies have not clarified how E-cadherin and integrin β3 on the surface of ZLSs regulate actin flow or stability of the F-actin cytoskeleton. Dufrançais, Ophélie et al. found that these structures are not involved in the early fusion process but may stabilize adhesion points, promote migration, or induce protein hydrolysis in the later fusion phase.60 In contrast, migration and adhesion between multinucleated cells and fusion partners may be facilitated by binding between membrane proteins, which induces downstream signaling and establishes actin flow. Accordingly, we focused on the potential role of the actin cytoskeleton and intracellular motility forces (Fig. 5).

(1)

Actin backbone: Arp2/3 and cortactin are colocalized with actin at the center of a ZLS, and the core framework of a ZLS is based on F-actin and nonmuscle myosin IIA. The periphery of the structure is covered with paxillin and vinculin to regulate its traveling wave motion.78 The cell contact surface is also covered with fusion-related proteins, including zyxin, E-cadherin, CD47-SIRPα, and integrin β3, forming a composite structure consisting of the intracellular cytoskeleton and plasma membrane proteins.79,80

(2)

Actin flow: Actin within a single podosome “foot” undergoes vertical oscillatory motion, which in turn forms a traveling wave. An analysis of the spatiotemporal location of podosomes revealed that the vertical motion is based on two factors, namely, regulatory protein comovement patterns (vinculin and talin show similar vertical oscillations) and actin aggregation and assembly in the podosome core.81 In turn, this traveling wave triggered by the overall vertical oscillation of the actin cytoskeleton moves in such a way that neighboring cells squeeze against each other. Additionally, the distribution of the F-actin bundle at the OC podosome overlaps with that of myoIIA, the activation of which leads to the generation of circumferential forces and helps maintain a balancing effect on actin wave motion.78

Summary

Similar to TNTs, ZLSs contain and regulate actin proteins such as cortactin, paxillin, and vinculin.78 Moreover, the ZLS membrane microdomain similarly lacks a backbone protein that integrates the membrane protein component mediating contact recognition with the actin regulator of traveling wave formation. Importantly, the mechanism through which stomatin is bound by contact partners via exosomes and further mediates contact adhesion suggests that the structural domain microfusion mechanism that we propose may be activated at this stage; this hypothesis was assessed in previous studies, and further investigation of the mechanism underlying ZLS membrane microdomain formation is needed.22

Bone resorption and secretory lysosomes

The classical structural signature of mature OCs is the formation of F-actin-rich adhesion structures on the ventral membrane contacting the bone surface, i.e., the ruffled border.1,58 This membrane is called a ruffled border (RB) due to the large number of folds.1,