MiR-25 overexpression inhibits titanium particle-induced osteoclast differentiation via down-regulation of mitochondrial calcium uniporter in vitro

Titanium (Ti) implants are commonly used in artificial joint replacements, and the influence of Ti on the periprosthetic environment and the subsequent effects on implant fate have been the subject of much research [1‐3]. Sundfeldt et al. estimated that 10–15% of Ti prosthetics used in artificial joint replacement will loosen and require replacement within 15–20 years after implantation [4]. This is attributed to the inevitable wear of the materials caused by relative motion of the prosthetic and degradation/degeneration of the biomaterials, which will further generate wear debris particles. Particles from all implant types will induce the differentiation of osteoclasts and accelerate an inflammatory response, processes that eventually lead to periprosthetic osteolysis. The final results are joint loosening and prosthesis failures [5, 6].

Signaling mediated via calcium ions (Ca²⁺), an important messenger responsible for cell transduction, is involved in osteoclast differentiation. When osteoclast precursors are continuously stimulated by Ca²⁺ oscillation, the calcium-binding protein calmodulin undergoes a conformational transition, which further induces activation of its downstream effector proteins the Ca²⁺/calmodulin-dependent protein kinases (CaMKs) and calcineurin phosphatase, which is responsible for nuclear translocation of nuclear factor of activated T cells, cytoplasmic 1 (NFATc1), regulation of osteoclast- specific genes, and thus the promotion of osteoclast differentiation [7]. Moreover, Yang et al. demonstrated through down-regulation of regulator of G-protein signaling 10 (RGS10) that Ca²⁺ oscillations influence the process of osteoclast formation in vitro, resulting in increased bone mass, whereas osteoclast formation can be inhibited by removal of Ca²⁺ oscillation [8].

Mitochondria are the most important organelles that regulate Ca²⁺ signals in the cytoplasm [9]. Mitochondria can extract a large amount of Ca²⁺ from a specific subcellular structural domain and regulate Ca²⁺ oscillations, thereby controlling Ca²⁺ signal within cellular solute and affecting cellular functions [10]. Uptake of Ca²⁺ ions into mitochondria is mainly mediated by mitochondrial calcium uniporter (MCU), a calcium channel with high selectivity that excludes other cytosolic cations [11]. This behavior relies on membrane potential, through which Ca²⁺ rapidly moves into mitochondrial matrix along the electrochemical gradient, instead of energy supply from ATP hydrolysis as well as accompanied by co-transformation of other ions and molecules [12].

Mammalian MCU consists of MICU1/MICU2, MCUb, and EMRE (essential MCU regulator) subunits. The selective filter of MCU, which is formed by an α helix that combines Ca²⁺ with acidic amino acid rings, can distinguish Ca²⁺-binding sites, expose the negatively charged surface of the inter-membrane space, reduce the energy barrier for Ca²⁺ flux into the mitochondria, and achieve high selectivity for Ca²⁺ [11, 13, 14]. Through these actions, MCU controls Ca²⁺ flux across the mitochondrial membrane. When the cytosolic calcium concentration increases, MCU can rapidly sense, absorb and transport Ca²⁺ to mitochondrial matrix, directly influencing the frequency and amplitude of calcium signals within cytosol [15]. Therefore, the fate of cells can be mediated by suppressing or enhancing MCU levels to regulate cellular calcium signaling [16].

Because MCU is a ruthenium (Ru)-responsive channel, Ru compounds are commonly investigated as inhibitors of MCU that are able to efficiently enter cells [17]. However, Ru is a retardant for many cations, and thus, it may result in systemic dysfunctions related to these cations [18]. Other forms of MCU regulation have commonly involved posttranscriptional regulations. MicroRNAs (miRNAs) represent an important class of posttranscriptional regulators. Additional research revealed that when cardiac myocytes are exposed to H₂O₂, MCU expression is negatively correlated with miR-25 expression, and miR-25 overexpression can reduce mitochondrial Ca²⁺ overload and inactivate mitochondrial apoptosis, increasing the resistance of cardiac myocytes to oxidative damage [19]. In addition, a bioinformatics prediction algorithm has identified miR-25 as complementary to the 3′ untranslated region (UTR) sequence of MCU, with a 100% match of the informatics nucleotide at the 1075 site. Another experimental study showed that MCU expression is significantly negatively correlated with miR-25 expression, confirming that MCU is a direct target of miR-25 [20]. Thus, we chose miR-25 mimics as a potential regulator of MCU expression in the present study (Table 1).

Accordingly, we hypothesized that osteoclast differentiation could be mediated via regulation of calcium signaling achieved by inhibiting MCU. To test this hypothesis, we established an osteolysis model based on osteoclast differentiation of RAW 264.7 cells induced by Ti particles. miRNA-25 mimics were used to reduce MCU expression and thereby diminish mitochondrial Ca²⁺ oscillations. RAW 264.7 cells were transfected with miRNA-25 mimics, and the effects on MCU expression and Ti particle-induced osteolysis were investigated. The findings of this study are expected to provide a new strategy and target sites for the prevention and treatment of periprosthetic osteolysis.

Artificial joint replacement is an effective treatment strategy for many degenerative joint diseases and arthritis. It is highly successful during the early period post-operation; however, prosthesis loosening during later stages is a major concern. Such loosening is closely related to the increased need for artificial joint replacement. Several risk factors for revision joint surgery have been identified, including advanced age, osteoporosis, high cost, and physicians’ experience [23]. Thus, the prevention and treatment of aseptic loosening of prostheses is of great value for prolonging the service life of joint prostheses and reducing revision arthroplasty. Although the exact mechanism of aseptic loosening of prostheses remains to be determined, it is widely accepted that wear debris particles can induce osteoclast differentiation and osteolysis around a prosthesis [24]. As they are the only cells responsible for bone absorption in vivo, the differentiation and regulation of osteoclasts in the periprosthetic environment is of great significance for treating and preventing PPO.

Bisphosphonates have been used clinically to treat osteolysis induced by osteoporosis and some tumors, as they function to inhibit bone adsorption by osteoclasts. However, these drugs have troublesome side effects, such as gastrointestinal adverse reactions and nephrotoxicity, and accurate data supporting their application to prevent and treat osteolysis are lacking [25, 26]. Two alternatives, osteoprotegerin-Fc (OPG-Fc) and anti-RANKL (receptor activator of NF-κB ligand) therapy represent a major clinical strategy for inhibiting osteoclast differentiation. However, a wide variety of signaling molecules downstream of the RANKL signaling pathway are involved in complicated regulatory pathways with multiple functions [27]. Although clinical trials have not found an increased risk of bone malignancy or deterioration of metastatic bone tumors, anti-RANKL therapy in mice could cause dysfunction and even deficiency of lymphopoiesis, an effect that may be exacerbated by experimental and spontaneous tumor metastasis and more sensitive to chemical carcinogen [28]. Given the limitations of these therapeutic strategies, the Ca²⁺/NFATc1 signaling pathway has become a new target for inhibiting osteoclast differentiation. NFATc1 is known to play a key role in osteoclast differentiation, and with ectopic expression of the NFATc1 gene, bone marrow-derived mesenchymal cells (BMSCs)could differentiate into osteoclasts even without RANKL induction [29]. Moreover, Liu et al. established an in vitro osteolysis model induced by wear debris particles and found that regulation of the Ca²⁺/NFATc1 signaling pathway impacted osteoclast generation; specifically, Ti particle-induced osteoclast differentiation could be suppressed upon inactivation of NFATc1 using 11R-VIVIT peptide [21].

In the process of osteoclast differentiation, when pre-cells are stimulated, calcium channels on plasmatic membrane open, allowing the influx of extracellular calcium into cells, resulting in an increased intercellular calcium concentration. To maintain a calcium equilibrium, adjacent channels or organelles such as the endoplasmic reticulum further release calcium to the cytosol, resulting in nonlinear Ca²⁺ oscillations [30], which directly regulate the process of osteoclast differentiation. MCU plays a great role in this process. Our RT-PCR and western blot analyses confirmed that miR-25 overexpression led to reduced MCU gene and protein expression in RAW264.7 cells, both with and without Ti particle induction of their differentiation into osteoclasts, indicating that miR-25 is a regulator with strong specificity for MCU expression. Meanwhile, with the reduction in MCU levels, Ca²⁺ oscillations decreased and osteoclast formation was reduced, suggesting that down-regulation of MCU led to altered calcium signaling and, consequently, diminished Ti particle-induced osteoclast differentiation.

In the process of Ti particle-induced osteoclast formation in cells overexpressing miR-25, the mitochondrial Ca²⁺ influx was decreased and Ca²⁺ oscillations reduced. Accordingly, activation of NFATc1 was decreased in the absence of continuous stimulation of Ca²⁺ oscillation, and ultimately, osteoclast activity was diminished. In relation to this process, NF-κB and NFATc1 have a similar Rel homology domain (RHD) [31], and Ca²⁺ levels in the cytoplasm are suppressed and cannot be effectively activated to achieve successful nuclear transposition, resulting in reduced mRNA and protein expression of NF-κB p65 and decreased expression of osteoclast gene products. At the same time, due to their structural similarity, activated NF-κB and NFATc1 can directly interact to form a complex, which allows them to play a synergetic role in promoting transcription [32]. In the present study, Ca²⁺ oscillations were inhibited upon down-regulation of MCU, preventing this synergetic effect and finally reducing osteoclast differentiation. Of course, the NF-κB signaling pathway is also an important route for osteoclast differentiation, in which NF-κB is irreplaceable.

In terms of the Ca²⁺/NFATc1 pathway, CaMK family members (especially CaMK II and CaMK IV) are crucial molecules for the transfer of calcium signals that participate in the differentiation and functions of osteoclasts. Ca²⁺ oscillation induces a conformation change in calcium-binding proteins, leading to CaMK activation and regulation of osteoclast differentiation via the CaMK-CREB route. CaMKs can also regulate osteoclast-specific gene expression in cooperation with NFATc1. On the other hand, CaMKs and c-Fos synergistically act on the promoters of NFATc1 genes and trigger a self-amplification mechanism of NFATc1 [33]. In the present study, with the overexpression of miR-25, CaMK II and CaMK IV levels were reduced, resulting in decreased expression of osteoclast-specific genes, minimizing activation of the self-amplification mechanism of NFATc1 gene expression, and ultimately reducing osteoclast activity. In support of these findings, a recent study on neuronal injury caused by acute ischemia/reperfusion found that by blocking the activation of CaMK II/IKK/NF-κB signaling, down-regulation of NF-κB p65 subunits in mice could increase I/R-induced neuronal death, and up-regulation of CaMK II could increase activation of the NF-κB up-regulating pathway [34]. In consideration of the expression of CaMK and NF-κB observed in our study, we speculate that interaction of CaMKs and NF-κB may be another regulatory pathway for osteoclast differentiation.

In this study, transfection with miR-25 mimics decreased MCU expression in RAW 264.7 cells during the process of osteoclast differentiation induced by Ti particles, while also weakening Ca²⁺ oscillation and reducing NFATc1 activation and NF-κB expression. Furthermore, miR-25 expression also decreased CaMK (downstream signaling molecules) levels, inhibited osteoclast activity, and restricted osteoclast formation. These data provide experimental evidence for the theory that osteoclast differentiation can be inhibited through the Ca²⁺/NFATc1 signaling pathway. During this process, overload of intracellular calcium appeared due to the lack of MCU-mediated regulation of intracellular Ca²⁺, resulting in functional disorder and apoptosis of osteoclasts and causing osteoclast absorptive dysfunction. Further studies of the underlying mechanisms of these effects are warranted in support of the development of therapeutic strategies targeting MCU.

In summary, we explored the effect of MCU expression on osteoclast formation. The data obtained in this study confirm that MCU is involved in the process of PPO induction by Ti particles. Transcription mediated by miR-25 can effectively act on MCU, inhibiting osteoclast differentiation possibly via the Ca²⁺/NFATc1 signaling pathway. However, this route for inhibiting osteoclast differentiation might also influence other cellular processes mediated by Ca²⁺ signaling, and thus, further studies are needed to determine the safety of this approach. Our analysis of miR25/MCU/Ca²⁺/NFATc1 interaction and its roles in osteoclast differentiation may provide a foundation for the future development of therapies to prevent and/or treat wear particle-induced PPO.