Bone remodeling stages under physiological conditions and glucocorticoid in excess : Focus on cellular and molecular mechanisms

Kyiv City Clinical Emergency Hospital, Bratyslavska st. 3, Kyiv, 02660, Ukraine. Tel.: +38-044-527-69-08. E-mail:bsmp@health.kiev.ua Povoroznyuk, V. V., Dedukh, N. V., Bystrytska, M. A., & Shapovalov, V. S. (2021). Bone remodeling stages under physiological conditions and glucocorticoid in excess: Focus on cellular and molecular mechanisms. Regulatory Mechanisms in Biosystems, 12(2), 212–227. doi:10.15421/022130


Introduction
The skeleton is a unique structure whose function is to maintain skeletal balance and homeostasis. In the formed skeleton, during life, old bone and damaged areas of bone are replaced with newly formed bone tissue. This cycle is called remodeling. Bone remodeling is a complex dynamic process in which local resorption and formation occurring in patterns of basic multicellular units (BMUs) in cortical and cancellous bone are coupled (Siddiqui & Partridge, 2016;Kenkre & Bassett, 2018). The formation of bone structures during remodeling provides a structural adaptation of bone to changes in function. During remodeling, calcium and phosphorus ions are released from bones and play an important metabolic role in mineral homeostasis.
The following cell types are involved in bone remodeling: osteoclasts, osteoblasts, including lining cells, osteomax (bone macrophages), osteocytes, and osteoprogenitor cells (Cho, 2015;Sinder et al., 2015;Novack & Mbalaviele, 2016;Soysa & Alles, 2019). Osteoblasts and osteoclasts, whose activities are considered in strict coupling Kenkre & Bassett, 2018), play a major role in the remodeling. Synchronous cell activity involving blood vessels contributes to the balance of resorption and bone formation in skeletal anatomical regions. However, with age, under conditions of negative factors on the body, with diseases, the remodeling cycle is disturbed; resorption begins to prevail over bone formation and can lead to osteoporosis. Osteoporosis refers to a metabolic disease in which bone mass decreases, bone microarchitecture is disrupted, and the risk of fracture increases. A severe complication in patients who take glucocorticoids is glucocorticoid-induced osteoporosis.
Glucocorticoids (GCs) have significant anti-inflammatory and immunomodulatory properties and are widely used to treat allergic, autoimmune and inflammatory diseases (Hardy et al., 2020). However, while endogenous GCs in physiological concentration support metabolic processes in bone tissue, pharmacological doses lead to loss of bone mass, development of osteoporosis and increased incidence of low-energy fractures early after oral therapy (Cherian et al., 2015). Negative changes appear both in the peripheral and axial skeleton, with a pronounced loss of bone mass in skeletal areas with a high spongy bone content, namely the vertebrae (Chuang et al., 2017). Spongy bone loss of up to 10-20% is reported in the first 6 months after therapy. Changes are also present in the compact substance of the long bones, but the rate of destruction is much lower than in the cancellous bone. In the first year of GC treatment, abnormalities in the compact bone may be only 2-3%, but thereafter, this process progresses (Cherian et al., 2017).
Glucocorticoid-induced osteoporosis (GCO) is quite common due to the large number of patients taking glucocorticoids. An estimated 1% of the US population is treated with glucocorticoids long-term (Buckley et al., 2017). Long-term glucocorticoid use leads to fractures of different localizations. The highest percentage of fractures was recorded in the vertebral bodies and in other parts of the skeleton: proximal femur, diaphyses of long bones, pelvic bones, forearm and ribs. Patients who received GC therapy for 6 months had a one-year incidence of vertebral body fractures of 5.1% and nonvertebral fractures of 2.5%, and after treatment for more than 6 months, the incidence increased by 3.2% and 3.0%, respectively (Amiche et al., 2016). In general, the prevalence of low-energy fractures in patients who received long-term GC treatment is high and ranges from 30% to 50% (Briot & Roux, 2015). In the case of the same bone mineral density values, the risk of fracture in patients receiving GC therapy is significantly higher compared to patients who did not receive therapy, as well as compared to patients with postmenopausal or senile osteoporosis (Saag et al., 2016).
Despite new advances in uncovering the mechanisms of action of GCs on bone, the most frequent and severe complication is GCO (Briot & Roux, 2015;Mazziotti et al., 2016;Cherian et al., 2017;Compston, 2018). The development of GCO is accompanied by impaired bone remodeling through increased resorption and decreased bone formation (Chotiyarnwong & McCloskey, 2020). In these conditions, the number of remodeling units increases in the bones, many BMUs lack bone formation, which leads to bone perforation, disconnection of bone trabeculae, reduction of bone mass and bone strength.
The focus of researchers is to establish the cellular and molecular mechanisms of bone remodeling under physiological conditions and in the development of glucocorticoid-induced osteoporosis (Kameo et al., 2020;Zhang et al., 2020) identifying the group of patients with impaired remodeling and high risk of fracture are key to preventing and developing treatment tactics for this severe complication (Xu et al., 2019).
Multiple signaling pathways that regulate bone remodeling have been identified, but the question of the molecular mechanisms of disruption, requires further study. Multiple molecules and signaling pathways are active at each stage of remodeling, affecting differentiation, activation, and viability of osteoblasts and osteoclasts (Deng et al., 2019;Zhao et al., 2020a, b). Information about the effects of glucocorticoids on bone remodeling stages is fragmentary, further studies on the effects on signaling pathways and cells at different stages of remodeling are needed.
Purpose of the review: to analyze and summarize the information on the cellular and molecular mechanisms acting at different stages of bone remodeling in physiological conditions and after exposure to glucocorticoids in excess on the basis of literature data.

Effect of excess glucocorticoids on bone
The action of GCs on inflammation, organs and tissues of the body, including bone tissue, has genomic and non-genomic effects. The genomic action of GCs is carried out after GC binding in cell cytoplasm to its receptor (GR), with subsequent translocation of the complex into nucleus, action on DNA with regulation of genes. In the mechanism of nongenomic action of GCs, four directions were identified and summarized: membrane-bound GR (mGR) actions, cytosolic GR actions, direct physicochemical interaction of GCs with the cellular membranes, and mitochondrial GR signal transduction (Hartmann et al., 2016;Panettieri et al., 2019). Glucocorticoids are lipophilic molecules that diffuse across the cell membrane into the cytoplasm. In the cytoplasm, GR is bound to immunophilin ligands (Cyp 55, PP5, FKBP51, and FK50652), heat shock proteins (hsp70, hsp90, and hsp23), which function as molecular chaperones that increase the GC ligand affinity and provide GR resistance to ubiquitination and proteosomal degradation (Hartmann et al., 2016). After remodeling of the complex, GR binds to the GC and moves to the nucleus (Fig. 1).
The GR is expressed in osteoblast precursor cells, osteoblasts, osteocytes, and osteoclasts (Bouvard et al., 2009;La Corte et al., 2010). Human (h) GR is the product of a gene located in chromosome 5q31-32 consisting of 10 exons. The hGR has been found to have 3 different functional domains (Oakley & Cidlowski, 2013). The N-terminal immunogenic domain, which contains most of the residues subject to posttranslational modifications, has a transcriptional activation function (AF-1); the central, DNA-binding domain consists of two zinc fingers implicated in DNA binding, translocation into the nucleus, and GC dimerization; the hinge region binds the central DNA-binding domain and the ligand-binding domain; the C-terminal ligand-binding domain (AF-2) interacts with ligand-dependent coregulators (Nicolaides et al., 2010;Oakley & Cidlowski, 2013;Cruz-Topete & Cidlowski, 2015). hsp70, hsp90, and hsp23 -heat shock proteins There are GR protein isoforms (namely, GRα, GRβ, GRγ, GR-A, and GR-P) generated from the same GR gene through alternative splicing and use of alternative translation initiation sites (Smith & Cidlowski, 2010). The most widely expressed isoforms of hGRα (h-human) and hGRβ. As a multiprotein complex, hGRα is present in the cytoplasm of almost all human cells (Kelly et al., 2008). After binding GC to GR, the complex enters the nucleus, where it functions as a ligand-dependent transactivation and transrepression factor directly affecting DNA. hGRβ is mainly located in the nucleus, does not bind glucocorticoids, is transcriptionally inactive, and regulates the expression of numerous of genes independent of GRα activity (Hartmann et al., 2016;Todosenko et al., 2017). GRβ acts as a repressor of cytokine gene transcription through recruitment of histone deacetylase complexes.
To prove the direct effect of glucocorticoids on bone, 50 healthy volunteers of postmenopausal women receiving 5 mg of prednisolone were studied to exclude the action of inflammatory mediators that are expressed in patients with pathology (Ton et al., 2005). The study was carried out at 2, 4, 6 and 8 weeks. There were revealed decreased indices of osteoblast activity in serum: procollagen type I N-terminal propeptide (PINP), procollagen type I carboxy-terminal propeptide (PICP), osteocalcin and bone specific alkaline phosphatase (BSALP) as well as indices of osteoclast activity: serum cross-linked N-telopeptides of type 1 collagen (NTX) and free urinary deoxypyridinoline (DPD). It was found that 5 mg of prednisolone in healthy women suppresses bone formation, which can adversely affect bone strength. This study proved that prednisolone affects bone regardless of inflammation. Inflammatory mediators and other deleterious molecular factors that are expressed in inflammatory diseases, with the summation of the negative effects of GC therapy, lead to poor bone quality (Ton et al., 2005).
Modulation of GC action can be carried out at the tissue level by various mechanisms: variations in the expression and sensitivity of receptors, transmembrane transporters, by enzymatic metabolism. Glucocorticoids modulate the expression of 11β-hydroxysteroid dehydrogenase type1 and type 2 (11β-HSD1, 11β-HSD2) enzymes by osteoblasts, osteoclasts, and osteocytes (Hachemi et al., 2018). These enzymes affect the differentiation and function of bone cells. The most active enzyme is 11β-HSD1, a glucocorticoid action receptor-modulator, one of the functions of which is the conversion of active GCs cortisol and corticosterone into inactive analogues of cortisone and dehydrocorticosterone. It is this enzyme that plays a crucial role in the development of GCO, mediating the harmful effects of GCs on bone cells and bone formation (Fenton et al., 2019). The biosynthesis of this enzyme increases with age and in GCO, which is accompanied by impaired bone quality and an increased risk of fracture. In addition, GCs block the expression of the enzyme 11β-HSD2, which protects osteoblasts and osteocytes from apoptosis. Decreased expression of this enzyme is recorded with age, further enhancing the negative effects of GCs on bone. Modulation of 11β-HSD1 and 11 β-HSD2 systems with age and biosynthesis of proinflammatory cytokines may be key mechanisms of bone response to GCs action and requires further research (Cooper et al., 2003;Hardy et al., 2018).
The active form of vitamin D 1,25(OH)2D3, which is formed in the kidneys, plays a central role in calcium homeostasis, bone metabolism, and bone remodeling. Experimental studies have confirmed the concept that there is a local regulation of vitamin D metabolism in osteoblasts (Siddiqui & Partridge, 2016). Human osteoblasts express vitamin D receptor (VDR) mRNA and they also activate and deactivate enzymes involved in vitamin D conversion (Zaynya et al., 2019). Key enzymes in vitamin D metabolism are members of the cytochrome P450 superfamily (CYP27A1 (as vitamin D 25-hydroxylase), CYP27B1, and CYP24A1). The effects of vitamin D (25(OH)D3 and 1,25(OH)2D3 on CYP27B1 and CYP24A1 enzymes in bone cells have been studied (Wegler et al., 2016). In osteoblast culture, 25(OH)D3 was shown to be converted to the active form 1,25(OH)2D3 under the action of the CYP27B1 enzyme, which stimulates osteoblast proliferation and maturation and increases intercellular mineralization (Pathak et al., 2020). In vitro studies have proven that not only osteoblasts but also osteocytes and osteoclasts have VDR, express CYP27B1, can locally convert vitamin D to the active metabolite 1,25(OH)2D3, and by expressing the enzyme CYP24A1 can locally catabolize vitamin D (Zarei et al., 2016).
Based on their own data, the authors assumed the existence of a previously unknown mechanism of the effect of GCs on human bone metabolism through cytochrome P450 superfamily enzymes in bone cells (Wegler et al., 2016). The obtained data reveal new possibilities of vitamin D influence on metabolic processes in bone cells. It is suggested that one of the mechanisms by which glucocorticoids have a negative effect on bone may be the disruption of vitamin D conversion in osteoblasts (Wegler et al., 2016).
There are negative effects of excess GCs on bones, disruption of GCs function of organs and systems -a combined pathway to the development of osteoporosis (Panettieri et al., 2019). Glucocorticoids negatively affect the expression and metabolism of growth hormone and thyroid hormones, insulin-like growth factor 1 (IGF-1), and insulin-like growth factor-binding protein-1 (IGFBP-1) (Mazziotti et al., 2016). Excess GCs inhibit sex hormone synthesis both indirectly by reducing pituitary hormone levels and adrenal androgen production, and directly by disrupting hormone production by the gonads (Cherian et al., 2017). Under the influence of glucocorticoids, the secretion of luteinizing hormone is reduced. This leads to a decrease in estrogen synthesis by the ovaries, the deficiency of which plays a significant role in the pathogenesis of GIO. The suppression of sex hormone biosynthesis caused by excess glucocorticoids leads to muscle weakness, which is a risk factor for falls and fractures (Briot & Roux, 2015;Cherian et al., 2017).
Under the influence of excess GC in the intestine, absorption of calcium and phosphorus is reduced. This is associated with impaired functionality of the calcitriol receptor (1,25(OH)2D3) in the intestinal mucosa, which leads to decreased biosynthesis of calcium-binding protein (calmodulin). The level of calcium excretion is increased, primarily due to the direct effect of GCs on tubular reabsorption.
Decreased intestinal calcium absorption and increased renal excretion leads to increased serum PTH levels, development of secondary hyperparathyroidism, and increased bone resorption.
In recent years, a new mechanism of osteoporosis development has been described as autophagy (cell recycling process). Autophagy is important for regulating bone remodeling and maintaining bone homeostasis (Shapiro et al., 2014;Shen et al., 2018;Wang et al., 2019a). Regarding the effect of GCs on bone cell autophagy, our knowledge is limited and the available data are contradictory. It has been reported that in glucocorticoidinduced osteoporosis, autophagy inactivation does not affect the number of osteoblasts or the expression of osteoblast-related genes such as osteocalcin, collagen type I alpha 1 (Col1α1) and runt-related transcription factor 2 (RUNX2) (Lin et al., 2016). However, another study has shown that during chronic glucocorticoid therapy, suppression of autophagy in osteoblasts increases their apoptosis (Wang et al., 2019a), promotes progression of GC-induced osteoporosis, and increases the risk of low-energy fractures (Han et al., 2018). When considering autophagy and its role in osteoblast function, at least two factors must be considered: the crosslinking between autophagy and apoptosis, and the GC dose (Wang et al., 2019b). The above-mentioned changes in bone tissue under the influence of GCs in excess, supplemented by the data below in the text, are reflected in bone remodeling.

Physiological bone remodeling cycle
The bone remodeling cycle is usually considered to involve four stages: activation, resorption, reversal and formation (Feng & McDonald, 2011) or additionally two more stages are distinguished, namely the resting stage (resting bone) preceding the activation stage and the terminal stage (Siddiqui & Partridge, 2016) (Fig. 2). According to Parfitt (1980), resorption in the BMU of adult bone takes about 3 weeks and bone formation takes 3-4 months. Overall, about 5-10% of the skeleton is replaced each year, and the entire adult skeleton is renewed within 10 years. The following cell types are involved in bone remodeling: osteoclasts, osteoblasts, bone lining cells, osteomax (bone macrophages), osteocytes and osteoprogenitor cells. Synchronous activity of the cells with the participation of blood vessels contributes to the balance of resorption and bone formation in the anatomical areas of the skeleton.
Various local and systemic factors control the bone remodeling cycle, some of which have been identified. The most important mechanism of remodeling of bone at all stages is the close molecular communication between osteoblasts, osteoclasts and osteocytes, which is provided by many cytokines, hormones and signaling pathways.
Resting stage (resting bone). The trabecular surface is covered bone lining cells, which have a flattened shape and low metabolic rate. Their function is not fully understood (Florencio-Silva et al., 2015). Bone lining cells differ from osteoblasts in phenotype. At this stage, the cells prevent direct interaction between the osteoclast and the matrix.
Activation stage. The activation stage can be considered as a complex of signals affecting the bone and initiating the BMU formation site. A prerequisite for remodeling is the death of osteocytes in the local area, occurring in various ways. This can be mechanical loading associated with deformation, pressure disturbance in the osteocyte lacuno-canalicular system, structural damage to the bone area, bone marrow condition, homeostasis changes, and other factors that activate osteocyte apoptosis.
Resorption stage. Osteoclasts are key players in remodeling bone, skeletal health and disease Novack & Mbalaviele, 2016). Osteoclasts attach to bone at sites of osteocyte death. The resorption is a multistep process consisting of several sub-stages -expression of biologically active molecules by osteoblasts and osteocytes, formation of osteoclast precursor cells, their fusion and osteoclast maturation. The key role is assigned to osteoblasts, but an inverse relationship, the influence of osteoclasts on the maturation and activity of osteoblasts was also revealed (Matsuoka et al., 2014).
An osteoclast is formed from a hematopoietic stem cell. In the early stage of osteoclast differentiation, signaling through macrophage colony stimulating factor (M-CSF) and its macrophage colony stimulating factor receptor (c-FMS) plays a significant role. M-CSF is expressed by mesenchymal stromal cells, cells lining cells, osteoblasts and endothelial cells. Transcription factor PU.1 induces cytokine M-CSF binding to the cellsurface receptor c-FMS, involving micropthalmia-associated transcription factor (MITF) and CCAAT-enhancer binding protein α (C/EBPα) proteins. In committed cell formation, the M-CSF -c-FMS complex activates the cellular oncogene Fos (c-FOS) / nuclear factor kappaB (NF-kB) / nuclear factor of activated T cells (NFATc1) signaling pathway (Huntley et al., 2019).
RANKL are expressed by osteoblasts, bone lining cells, osteoprogenitor cells and osteocytes. There is evidence that compared to osteoblasts, osteocytes express more RANKL and have a greater ability to support osteoclastogenesis than osteoblasts or bone marrow stromal cells (Nakashima et al., 2011). RANKL is also expressed by immunocompetent cells located in the bone marrow, activated T-lymphocytes, B-lymphocytes, monocytes, macrophages, etc. The RANKL protein recognizes and binds to the RANK receptor located on the plasma membrane of osteoclast line cells and promotes their differentiation and activation. The RANKL -RANK connection is considered to be crucial in the activation of osteoclastogenesis.
Blocking of RANKL to RANK is performed by OPG, which is accompanied by a decrease in differentiation and activity of osteoclasts. Osteoprotogerin is expressed by osteoblasts, lining cells and osteocytes, as well as B-lymphocytes located in bone marrow, which account for 64% of total OPG biosynthesis (Li et al., 2007). That is, RANKL / RANK / OPG form an important system that regulates a biological process such as osteoclast differentiation, activation or inhibition. Additionally, the leucinerich repeat-containing G protein-coupled receptor 4 (LGR4) expressed by osteoclasts may be acted on as another decoy receptor for RANKL .
The c-Jun N-terminal kinase (JNK), p38, and ERK signaling pathways that induce osteoclast differentiation are activated in response to signals received by the cell through RANKL binding to the RANK receptor and involvement of the adaptor molecule tumour necrosis factor receptor-associated factor 6 (TRAF6). The JNK signaling pathway mobilizes osteoclast transcription factors associated with c-Fos antigens (Fra-1 and Fra-2) and NFATc1. Other pathways that also loop to NFATc1 are ERK1 / NF-kB / activator protein 1 (АР-1) / MITF and ERK1 / NF-kB / AP-1 (Boyce & Xing, 2008;Soysa & Alles, 2019).
NFAT (NFATc1-4) belongs to the family of transcription calciumdependent factors, which function to coordinate cell growth and differentiation by regulating the production of key molecules of these processes: growth factors, various cellular proteins and cytokines (Mandal et al., 2016). NFAT5, a member of this family, is a calcium-independent factor. In resting cells, NFATs are hyperphosphorylated and in an inactive state. Osteoblast and osteoclasts' differentiation and activity is associated with calcium signaling. An increase in calcium levels in osteoblasts involving bone morphogenetic protein (BMP)-2 activates calcineurin, which dephosphorylates NFATc1, facilitating its import into the nucleus (Negishi-Koga & Takayanagi, 2009). In the nucleus, NFATc1 interacts with active transcription factors such as osterix, AP-1, Pu.1, Fra and, after binding to DNA sites, activates transcription. NFATc1 is important for the regulation of osteoblast differentiation and osteoblast-mediated osteoclast activity (Mandal et al., 2016).
Secreted members of the BMP family, BMP-2 and BMP-7, are involved in comminuted cell fusion and osteoclast formation through Smad1 / 5/9 activation (Omi, 2019). BMP transduce intracellular signaling via the canonical SMAD signaling pathway and non-canonical SMAD signaling pathways such as mitogen-activated protein kinases (MAPK) / phosphoinositide 3-kinase (PI3K) / AKT. Differentiation of macrophages into osteoclasts and activity of mature osteoclasts is enhanced with decreased or loss of Smad1/ 5 expression (Tasca et al., 2018).
Vitamin D metabolites can inhibit the fusion of osteoclast precursors by reducing the expression of NFATc1 and DC-STAMP (Zarei et al., 2016). In addition, smaller osteoclasts are formed with fewer nuclei per osteoclast.
The formation of osteoclasts also involves the Notch signaling pathway, which transmits signals between cells through proteins of the Notch family and involves a large number of different genes, including genes responsible for a variety of cellular functions, including differentiation and proliferation (Zanotti & Canalis, 2016). In osteoclast precursor cells, stimulation of Notch signaling leads to the formation of large osteoclasts with numerous nuclei and increased resorption activity, but the resorptive activity of small osteoclasts is suppressed.
In addition to the previously described pathways, other pathways in the regulation of osteoclastogenesis are being investigated. A new mechanism of osteoclast differentiation has been described and the role of protein kinase D from the kinase family in the regulation of osteoclast activity has been established (Leightner et al., 2020).
PTH has a mediated effect on osteoclast differentiation via the receptor on osteoblasts PTH1R. This link leads to activation of RANKL expression by osteoblasts. In addition, PTH binding to PTH1R on osteo-clasts promotes V-ATPase activity by inducing expression of V-ATPase a3-subunit and d2-subunit, enhances its resorptive properties (Liu et al., 2016). The association of PTH with PTH1R in osteoclasts inhibits the expression of SEMA4, which is accompanied by a decrease in osteoblast activity. Fifteen new genes involved in RANKL-mediated signal transduction for osteoclastogenesis were identified, including Merlot, Tussilagone, and Lrp1 (Yamakawa et al., 2020). The Merlot gene simultaneously inhibits osteoclastogenesis through the GSK3β -NFATc1 axis and induces apoptosis, reducing the lifespan of osteoclasts.
Osteoclast viability is maintained by M-CSF and RANKL (Fig. 5). After M-CSF binds to its c-FMS cognate receptor, two signaling pathways are activated -MARK (Ras / Raf / MEK / ERK) and PI3K / AKT kinase / mammalian target of rapamycin (mTOR) pathways, the latter being also synergistically activated by RANKL -RANK complex (Soysa & Alles, 2019). The M-CSF -c-FMS complex also activates the Ras/Raf/PAK pathway, which leads to increased osteoclast viability, by modulating the expression of Survivin, a protein member of the apoptosis protein inhibitor (IAP) family (Soysa & Alles, 2019). Activated osteoclasts attach to the bone surface via integrins (αVβ3 and others) with the formation of the podosome and subsequent formation of a membrane ruffling (Fig. 6) (Feng & Teitelbaum, 2013;Teitelbaum, 2015). In addition to adhesion, integrin αvβ3 is involved in osteoclast migration (Mellis et al., 2011;Kanakamedala et al., 2019). The formation and function of actin that makes up the osteoclast cytoskeleton is regulated by Rho family proteins (GTPases). Ras homolog family member A (RhoA) proteins are involved in the regulation of actin cytoskeleton of osteoclast: cell division cycle 42 (Cdc42) regulates filopodia formation, Rac protein (subfamily of the Rho family of GTPases), regulates the formation of a membrane ruffling, and RhoA protein increases adhesion and tension of actin fibers (Morel et al., 2018).

Fig. 6.
Schematic representation of active osteoclast: activated osteoclasts attach to the bone surface through integrins, which bind to bone vitronectin, forming an edge seal around the resorption area; through the formed ruffled border of the membrane, hydrogen ions enter the tight seal through a proton pump, chlorine ions through chloride channels, proteases and acid phosphatases are released from lysosomes, which leads to the destruction of the matrix; osteoclasts are characterized by transcytosis; residues of detritus after degradation in lysosomes are removed from the cell and enter the bloodstream; on the opposite surface of the osteoclast there are numerous receptors that receive regulatory signals: FMS, RANK and interleukin-6, TNFR, LGR4, calcitonin receptor, etc.) Resorption is accompanied by the formation of a lacuna approximately 60 μm deep in the bone (Boyce et al., 2018). To resorb the mineral part of bone, osteoclasts produce hydrochloric acid, releasing H + ions through the proton pumps of the cell membrane and Cl-ions through the chloride channels. To destroy the collagen matrix, osteoclasts produce matrix metalloproteinases and cathepsin (Feng & Teitelbaum, 2013).
Reversal stage. The transition to the formation stage occurs through the reversal stage, which includes apoptosis of osteoclasts, preparation of the surface for settlement by osteoblasts, formation of canopies over the resorption area, proliferation and differentiation of osteoblasts. The signaling pathways for switching resorption to the reversal stage are not fully understood.
There are two points of view on the nature of the cells located in the BMU immediately after resorption by osteoclasts, which are involved in preparing the resorbed surface for settlement by osteoblasts. Based on immunohistochemical examination, 97% of the cells were positive for the osteoblast marker RUNX2. They were located adjacent to osteoclasts, which, according to the authors, indicates colonization of the cavity by osteoblasts immediately after resorption (Andersen et al., 2013). According to other authors, the resorption cavity initially contains bone macrophages or osteomax with a characteristic phenotype and the presence of F4/80 marker Cho, 2015;Sinder et al., 2015). The function of cells in the reversal stage is to remove residual mineralized and non-mineralized matrix (detritus) by phagocytosis and prepare the surface of bone for settlement by osteoblasts (Jensen et al., 2011;Sinder, 2015;Kenkre & Bassett, 2018). Bone lining cells produce collagen, which is deposited as a thin layer on the surface of BMU (Matsuo & Irie, 2008).
A feature of this stage is the location above the BMU region of the canopies (Fig. 2), which are a reservoir of osteoprogenitor cells (Kenkre & Bassett, 2018). These cells have a flattened appearance, they are similar to osteoblasts, can proliferate and renew. A significant role in the formation of canopies is played by bone lining cells. Canopies close the bone remodeling BMU compartment, which separates osteoclasts and osteoblasts from bone marrow (Pettita et al., 2008). Activating factors for the formation of osteoprogenitor canopies are: TGF-β, IGF-1, IGF-2, BMPs, PDGF, fibroblast growth factor (FGF), and estrogens (Siddiqui & Partridge, 2016). The canopies of osteoprogenitor cells are permeated with capillaries that provide rapid access to systemic regulatory factors to control osteoclasts and osteoblast precursors that migrate to this area (Andersen, 2009). Osteoprogenitor cells differentiate into osteoblasts and also participate in the settlement of the surface resorbed by osteoclasts.
Formation stage. The main events of this stage are the formation of osteoblasts from mesenchymal stem cells in the resorption cavities, their proliferation, differentiation and mineralization of the osteoid. At this stage joint action (coupling) of osteoclasts and osteoblasts is clearly manifested. In recent years, a number of important regulatory transcription factors contributing to bone formation and mineralization have been identified.
Osteoclasts play an important role in the transition to the reversion and formation stage . One of the ways is bidirectional stimulation of osteoblastogenesis or inhibition of osteoclastogenesis through ephrins (Matsuo & Otaki, 2012;Plotkin et al., 2019). Osteoblasts express ephrin A and B ligands as well as their EphA and EphB receptors, whereas osteoclasts express only ephrin A2, B1, and B2 ligands (Plotkin et al., 2019). Stimulation of osteoblast proliferation through the ephrinB1/ ephrinB2 -EphB4 signaling pathway is carried out through the transcription factors RhoA, RUNX2, PDZ-binding motif (TAZ). Inhibition of osteoclast function is carried out through the c-Fos -NFAT signaling pathway (Fig. 7). These signaling pathways confirm the hypothesis about the coupled mechanism of osteoclast and osteoblast involvement in bone remodeling during the transition to the stage of reversion and formation.
Afamin secreted by osteoclasts plays an important role in osteoblast genesis. At an early stage of differentiation, Afamin, through the Akt signaling pathway, stimulates the migration of preosteoblasts, promotes delivery of Wnt ligands to their receptors on the cell surface and prevents the aggregation of Wnt proteins (Kim et al., 2012). CT-1 attracts mesenchymal stem cells to this area . During bone remodeling, osteoclasts stimulate osteoblast chemotaxis while maintaining spatial segregation.
In bone remodeling at the stage of osteoblastogenesis and formation important signaling pathways are WNT pathways, in which signal transduction by WNT proteins is strictly controlled at several levels. In humans, 19 different Wnt-proteins have been described, which can activate both canonical and non-canonical Wnt-signaling pathways. The canonical Wnt/β-catenin signaling pathway is key to mesenchymal cell differentiation, osteoblast activity, and bone formation through regulation of RUNX2 gene expression (Haxaire et al., 2016). This pathway is related to the interaction of Wnt protein-ligands (Wnt 2, 4, 5, 11, 16) with specific transmembrane protein Frizzled (FZD) and its associated low density lipoprotein 5 and 6 co-receptors (LRP-5/6) (Fig. 8).
The activation of this complex is accompanied by an increase in the function of protein Disheveled (DSH), which inhibits the related proteins: glycogen synthase kinase 3β (GSK3β), the adenomatosis polyposis coli (APC), AXIN, casein kinase 1 (CK1) and decreases GSK3β activity by inhibiting phosphorylation, which is accompanied by β-catenin stabilization and its accumulation in cytoplasm. After translocation into the cell nucleus, β-catenin interacts with the family of transcription factors, RUNX2, and through family of high mobility group (HMG) stimulates the expression of target genes, including Lef1, Tcf7, Nkd2 and Axin2 (Karner & Long, 2017). In addition, in osteoblasts, the Wnt / β-catenin signal transduction pathway controls the expression of OPG, a specific RANKL inhibitor, and thus regulates osteoclastogenesis. Secreted proteins of the Dickkopf family (Dkk) can directly oppose canonical Wnt binding via LRP5 or LRP6 (MacDonald & He, 2012). Proliferation, differentiation of osteoblasts and suppression of apoptosis are promoted by inhibition of Dickkopf protein (DKK4) expression (Hiramitsu et al., 2013). Sclerostin, a Wnt signal transduction antagonist, was found to be highly expressed in osteocytes, after binding to LRP 5/6 competitively affecting Wnt and LRP 5/6 binding. The Wnt signaling pathway also activates signal transduction independent of β-catenin by activating intracellular cascades involving Gproteins Rho and Rac, calcium-calmodulin-dependent kinase 2 (CaMK2), JNK and p38, phospholipase-C, protein kinase C (PKC), protein kinase A (PKA), PI3K / AKT and mTOR (Karner & Long, 2017).
Wnt10b induces osteoblastogenesis by activating transcription factors RUNX2, Dlx5 and osterix (Bennett et al., 2005). One of the mechanisms by which Wnt10b promotes osteoblastogenesis is the suppression of the expression of transcription factor CCAAT-enhancer binding protein (C/EBPα) and PPARγ that promote the adipogenic commitment of mesenchymal stem cells.
BMP-2, transforming growth factor-β (TGF-2) and insulin-like growth factor 1 (IGF-1) are involved in the regulation of Wnt signaling pathway that stimulate osteoblastogenesis (Matsuo & Otaki, 2012;Frenkel et al., 2015;Huntley et al., 2019). IGF-1 affects the stability of β-catenin and stimulates transcriptional activity. TGF-β / BMP signaling plays an important regulatory function in osteoblast differentiation and bone formation. Osteoblastogenesis involves cross-talk between TGF-β / BMP signaling and many other major signaling pathways such as Wnt, MAPK, Smad, Hedgehog, Notch, and FGF (Chen et al., 2012). BMP-mediated signaling results in increased expression of RUNX2 and osterix, mainly through Smad and MAPK signaling pathways. RUNX2 expression is influenced by the Dix5 gene, which is activated by BMP-2. The Dix5 gene protein is considered a key factor in osteoblast maturation (Chen et al., 2012).
RUNX2-dependent transcription is influenced by PTH, BMP, FGF2, mechanical stress transmitted through the matrix (Camilleri & McDonald, 2006). Post-translational changes act through the signaling pathways MAPK, SMAD4, SMAD1 / 5, and protein kinases A and C (PKA, PKC), which activate RUNX2 by phosphorylation and ubiquination (Franceschi & Xiao, 2003). BMP-2, BMP-7, and BMP-9 induce osteoblastogenesis, but BMP-9 has a more pronounced effect (Jann et al., 2020). In addition to the Wnt pathway, the role of Notch signaling pathway was identified. In osteoblastic cells, Notch activation suppresses cell differentiation and causes osteopenia of cancellous bone (Zanotti & Canalis, 2016). Notch can directly interact with RUNX2, by inhibiting Notch function, promotes osteoblast proliferation and differentiation (Infante & Rodríguez, 2018). Stimulation induces juxta and intramembrane cleavage of Notch proteins; the intracellular Notch domain is released and moves towards the nucleus to regulate gene expression. Notch signaling disrupts RUNX2 activity. Control of target gene expression in osteoprogenitor cells and osteoblasts is also carried out by Wnt and Notch influence on signal transduction to TGF-β superfamily (Jann et al., 2020). TGF-β promotes apoptosis of osteoclasts and stimulates chemotaxis of preosteoblasts and osteoblasts.
Osterix/Sp7 involvement is necessary for differentiation of preosteoblasts into osteoblasts. This transcription factor stimulates the expression of type I collagen, osteocalcin, osteopontin, sialoprotein and osteonectin genes in cells.
The role of the transcription factor Forkhead box P1 (FOXP1) in the differentiation of mesenchymal cells in the osteogenic direction was established; FOXP1 inhibits adipocyte differentiation by interacting with the key core-binding factor subunit beta (BFβ).
The attention of researchers is attracted by small non-coding RNA molecules 18-25 nucleotides of microRNA (miRNAs, MiR), which regulate metabolism by modulating the differentiation and activity of osteoblasts and osteoclasts (Xie et al., 2015;Zuo et al., 2015;Sun et al., 2016;Garcia & Delany, 2021). So far, more than 20 species of microRNAs affecting the proliferation and differentiation of osteoblasts and 7 species affecting the differentiation of osteoclasts have been isolated. Data from recent studies regarding miRNA in bone cells are summarized in reviews (Garcia & Delany, 2021). MiR-20a has been shown to upregulate RUNX2, affects TGFβ isoforms, TGFβR1 and TGFβR2 signaling receptors, BMP7, BMP2 and BMPR2 receptor, small mothers against decapentaplegic (SMAD), total SMAD4, and inhibits SMAD7 (Garcia & Delany, 2021). The miR-140 family plays an important role in inhibiting TGFβ signaling by acting on both ligand and receptor. The miR-106 isoforms inhibit osteoblast differentiation via BMP2 and SMAD5. It was found that MiR-29a cluster by regulating Wnt-3a, glycogen synthase kinase 3β and β-catenin signaling enhances osteoblast differentiation through increased expression of RUNX2 osteocalcin, collagen type 1a1, IGF-1, and also stimulates matrix mineralization (Wang et al., 2013;Lian et al., 2019). Thus, miRNAs are considered to be important modulators of bone remodeling.
Other important factors affecting osteogenesis have been identified. These are long non-coding RNA molecules (LncRNAs), more than 200 nucleotides long, which are involved in many biological processes. Most lncRNAs are located in the nucleus, where they typically regulate gene expression by mediating epigenetic changes such as DNA methylation, histone modification, and chromatin remodeling (Zhao et al., 2020b). The cytoplasm contains small amounts of lncRNAs, where they modulate mRNA stability and translation, and interfere with posttranscriptional regulation (Yoon et al., 2014).
LncRNA-DANCR circulating in monocytes is considered as a potential biomarker of postmenopausal osteoporosis (Tong et al., 2015). A recent published review presents the latest data regarding the role of these molecules in osteoporosis (Zhao et al., 2020b). As shown in this review, many lncRNAs (H19, MALAT1, LOC103691336 and ODSM) can regulate osteogenic differentiation and osteoclastogenesis by modulating gene expression at different levels of transcription, and the likes of lnc-ob1, H19, HOTAIR and PGC1β-OT1 can directly or indirectly mediate epigenetic changes. Thus, lncRNA HOTTIP interacts with WDR5, a positive regulator of β-catenin expression, activating Wnt / β-catenin signal transduction pathway, which promotes osteogenic differentiation of mesenchymal stem cells; in particular, lncNKILA affects the RXFP1 / AKT signal pathway, which is a positive regulator of osteogenesis, LncNKILA overexpression inhibits NF-kB expression and increases RUNX2 expression, which promotes osteogenic differentiation of mesenchymal stromal cells (Zhang et al., 2020). Other lncRNAs have also been described that differentially affect osteogenic differentiation (LncRNA TUG1, MALAT1, lncRNA AK077216, etc.).
The family of semaphorins (Sema), membrane and secreted proteins, plays an important role in the regulation of bone formation. In RANKLactivated osteoclasts, the expression of Sema 4D is increased, which, by binding to the Plexin B1 receptor, inhibits osteoblast differentiation, inhibiting the secretion of alkaline phosphatase and osteocalcin, as well as the formation of mineralized nodules in cell culture (Lontos et al., 2018). In addition, Sema 4D, through its Plexin-B1 receptor, inhibits the locomotion of osteoblasts through contact inhibition of locomotion (CIL) (Deb Roy et al., 2017). Blocking Sema 4D has an anabolic effect on bone.
In contrast to the action of Sema 4D, Sema 3A inhibits osteoclastic bone resorption, enhances bone formation, supports osteocyte viability, and mature osteocyte survival; in an estrogen-dependent manner (Hayashi et al., 2019).
On histological examination, different sections of bone trabeculae may show BMUs at different stages of remodeling: osteoblast-filled formation and resorption with the presence of osteoclasts and the formation of canopies (Fig. 9). Osteoclasts in the resorptive cavity die by apoptosis after performing their function. Mechanisms of apoptosis through activation of M-CSF -c-FMS and TNF-α -TNFR signaling pathways have been described. Binding of M-CSF to its receptor c-FMS through the RAS-dependent mechanism and other transcription factors leads to apoptosis of osteoclasts. In the Ras-dependent mechanism, M-CSF activates PAK1, which promotes osteoclast apoptosis by modulating the expression of survivin, a member of the IAP family (Soysa & Alles, 2019).
The binding of TNF-α to the receptor TNFR activates TNF receptorassociated death domain (TRADD) TRADD, which interacts with Fasassociated death domain (FADD) through DD, and subsequently promotes the recruitment of caspase 8, which leads to the formation of a signaling complex that induces osteoclast apoptosis (Peng et al., 2018;Soysa & Alles, 2019) (Fig. 10). The Fas / FasL system is a powerful mechanism of osteoclast apoptosis. Osteoblasts, by expressing RANKL and M-CSF, up-regulate the Fas receptor in osteoclast progenitors via NF-κB and induce Fas-mediated apoptosis (Wu et al., 2005). In mature osteoclasts, other mechanisms of the RANK response function, leading to a reduction in the expression level of Fas and Fas-mediated apoptosis (Wu et al., 2005;Soysa & Alles, 2019). An increase in the Bcl2 / Bax ratio and caspase 3 activity may be a potential role of signaling proapoptotic mechanisms.
Osteocytes are considered as the coordinator of bone remodeling during the transition from resorption to formation stage, which, by producing sclerostin, provide the functioning of two opposite mechanisms involved in the process of adaptive bone remodeling (Sapir-Koren & Livshits, 2014a). Under no load conditions, osteocytes produce high levels of sclerostin, the excess of which disrupts canonical WNT / β-catenin signaling and activates non-canonical WNT signaling pathways aimed at bone resorption. Under mechanical stress, the secretion of sclerostin is reduced, the WNT / β-catenin pathway is activated, which stimulates bone formation. Sclerostin acts as a molecular mechanism, regulating the expression of RANKL and OPG.
Formation stage. At this stage, the metabolism of osteoblasts, which produce collagen type 1, osteocalcin, osteonectin, noncollagen proteins, bone alkaline phosphatase isoenzyme, proteoglycans, and other molecules involved in osteoid formation, becomes more active. In addition, at this stage, the cells secrete vascular endothelial growth factor, which is further involved in the formation of blood vessels in the osteoid (Sims & Martin, 2014).
Each osteoblast synthesizes the matrix around itself, mineralizes it, mounds it, and turns into an osteocyte. With neighbouring cells, osteocytes are connected by canaliculus, forming a unified network.
The newly formed osteoid mineralizes, a multistep process occurring in the BMU. We consider the primary and secondary phases of mineralization, the primary begins 5-10 days after the formation of the osteoid, the secondary occurs after the complete filling of the BMU (compact bone or spongy bone) (Boivin et al., 2009). Osteoblasts secrete membrane-bound matrix vesicles with calcium, phosphate and enzymes to digest mineralization inhibitors (such as pyrophosphate and proteoglycans) (Anderson, 2003). Minerals are deposited between collagen fibers. Osteoblasts are surrounded by the mineralized matrix and turn into osteocytes. During the secondary mineralization stage, the mineral component matures, including an increase in the number and size of crystals. Mineralization ends after 90 days in the cancellous bone and after 130 days in the cortical layer of the compact bone (Fernández-Tresguerres-Hernández-Gil et al., 2006).
A new molecular mechanism of vitamin D influence on mineralization was discovered (Jo et al., 2020). Based on the available data that osteoblasts and osteocytes express DKK (Li et al., 2006), the authors showed in an experiment that under the influence of 1,25(OH)2D3 the levels of DKK1 protein expression increase until day 7 after exposure and decrease on the following terms during differentiation of osteoblasts, as their transformation into osteocytes is associated with mineralization. During mineralization, 1,25(OH)2D3 induces DKK expression through C/EBPβ activation.
Terminal stage. At this stage osteoblasts differentiate into osteocytes. Each osteoblast synthesizes a matrix around itself, mineralizes it, becomes engulfed and transforms into an osteocyte. The BMU is filled with bone tissue. With neighbouring cells located in the mineralized matrix, osteocytes are connected by long processes forming a unified network. Osteocytes express sclerostin, which indicates the completion of the remodeling process. Osteocytes play the leading role at this stage, as they secrete inhibitory factors that slow the rate of bone formation, preventing excessive bone formation. In addition, in osteocytes Notch1 signaling leads to the inhibition of bone resorption, which is considered secondary, as a strong factor in suppressing resorption is osteoprotegerin (Zanotti & Canalis, 2016). In addition, in osteocytes Notch1 can suppress sclerostin biosynthesis with subsequent enhancement of Wnt signaling. A feature of remodeling under normal conditions should be a coupling of bone formation and resorption.

Bone remodeling under the influence of glucocorticoids in excess
Excess GCs disrupt the functioning of many signaling pathways (Fig. 11) that regulate bone cell proliferation and differentiation. While under physiological conditions significant progress has been made in the study of signaling pathways and GC action on the cells involved in remodeling, under conditions of GC excess, this knowledge is fragmentary.
Resting stage. The peculiarity of this stage is the absence of lining cells on the majority of bone trabeculae, which reflects the negative effect of GCs on bone.
Activation stage. Determination of the bone resorption locus occurs at this stage at sites of osteocyte death. Excess GCs induce osteocyte apoptosis by activation of multiple biomarkers and increased expression of proapoptotic genes by increasing the biosynthesis of caspases 3, 7 and 8, which activate the proapoptotic gene Bim and death receptor Fas / FasL (Espina et al., 2008). Osteocyte apoptosis under the influence of excess GCs can be carried out through a receptor-mediated mechanism through rapid activation of pro-apoptotic kinases, which does not require gene transcription. Pyk2 activation leads to reorganization of the cytoskeleton and is accompanied by loss of cell attachment to the extracellular matrix, and JNK activation causes osteocyte death by increasing reactive oxygen species, which eventually leads to apoptosis or anoikis (Plotkin et al., 2007). There is evidence that osteocytes under low doses of GCs undergo autophagy, whereas high doses of GCs lead to osteocyte apoptosis (Yao et al., 2013). Signaling pathway for autophagy may be the activation of the FAS/CD95 via PTH.
Glucocorticoids cause matrix damage by disrupting branched osteocyte canaliculi, promote the release of paracrine factors that increase local angiogenesis, which increases bone resorption (Goldring, 2015).
In addition, during osteocyte apoptosis, RANKL and proinflammatory cytokines (TNF-α and IL-6) mediated by nuclear protein HMGB1 are released under the influence of GCs, and thus, the products of cell degradation also stimulate osteoclast proliferation and differentiation (Hardy et al., 2018;Wang et al., 2019b). As osteocytes die, the number of possible sites of osteoclast attachment to bone increases.
Resorption stage. Glucocorticoids stimulate mononuclear cell fusion and the formation of multinucleated osteoclasts. However, the influence of molecular pathways of GC on this process has not been studied. Glucocorticoids in excess stimulate RANKL expression and decrease OPG expression in osteoblasts and osteocytes, which leads to increased proliferation and differentiation of osteoclasts. The effect of glucocorticoids on RANKL-activated osteoclast progenitor cells leads to increase of their resorptive activity but does not influence the number of formed osteoclasts. Fig. 11. Effects of glucocorticoids in excess on various signaling pathways in bone remodeling: Akt -kinase, BMP -bone morphogenetic protein, GCs -glucocorticoids, GH -growth hormone, IGF-1 -insulin-like growth factor 1, MAPK -mitogen-activated protein kinases, PI3Kphosphoinositide 3-kinase, SMAD -small mothers against decapentaplegic, TGFβ -transforming growth factor beta, Wnt -wingless and Int-1 Other mechanisms of osteoclastogenesis activation by GCs have been described. One mechanism is related to the action of PTH and D3 in combination with GCs on stromal cells, which leads to an increase in RANKL expression, a decrease in OPG and is accompanied by an increase in the number of osteoclasts (Conaway et al., 2019). Another mechanism of GC influence on the proliferation and activation of osteoclasts through the induction of reactive oxygen species (ROS) is presented. The authors consider this method as a potential therapeutic target in diseases associated with bone mass loss (Shi et al., 2015b).
Glucocorticoids affect osteoclastogenesis in different ways, and the mechanisms of influence are being investigated. By definition (Kim et al., 2006(Kim et al., , 2007 a paradoxical situation develops at this stage. On the one hand, under the influence of GCs osteoclastogenesis is reduced, the motor activity of osteoclasts and their ability to attach to the bone surface is impaired. Inhibition of RhoA, Rac and Vav3 proteins by glucocorticoids is accompanied by disruption of osteoclast cytoskeleton formation and its actin ring providing cell adhesion to bone (Kim et al., 2007). Besides, GCs induce cell apoptosis by transactivation of Bim gene, which leads to expression of apoptotic mediators Bax and Bak. Under their influence the mitochondrial membrane potential is disturbed with subsequent release of cytochrome C and Apaf-1 into cytosol, which is accompanied by activation of caspase 9, effector caspases 3, 6, 7 and leads to apoptosis (Fig 12). TCF -T-cell factor; LEF -lymphoid enhancer-binding factor; LRP -low-density-lipoprotein-related protein On the other hand, an increase in the number of osteoclasts and BMU in trabecular and compact bone under the influence of excess GCs is registered. This is associated with an increase in the osteoclast life cycle by inhibition GCs of apoptosis. Excess GCs involving M-CSF -c-FMS signaling pathway through RAS activation leads to the functioning of two independent pathways through Raf and PIK3. In the RANKL -RANK signaling pathway, through cSrc and TRAF6 factors, GCs induce PIK3, AKT and mTOR, which contributes to an increase in the osteoclast life cycle.
Activation of bone resorption by osteoclasts is influenced by increased expression of the cytokine M-CSF, RANKL and IL-6, which stimulate differentiation and proliferation of these cells, against the background of GCs suppression of OPG biosynthesis (Johnson & Kamel, 2007;Kondo et al., 2008). The dimeric GC receptor present in osteoclasts (contains additional β-layer) also stimulates RANKL-mediated bone resorption (Conaway et al., 2016). Under the influence of excess GCs, inhibition of caspase 3, the increased level of which causes apoptosis of osteoclasts, which prolongs their viability, has been recorded (Jia et al., 2006, Johnson & Kamel, 2007Frenkel et al., 2015).
The peculiarity of early phase therapy with high doses of GCs is the increased formation of osteoclasts, while the number of osteoblasts decreases due to the disruption of molecular interaction in the WNT signaling pathway. However, during treatment, bone resorption slows down, resulting in a state of chronic decreased bone metabolism (Johnson & Kamel, 2007).
MiR-17 and miR-20a clusters have an inhibitory effect on osteoclastogenesis by reducing the expression of RANKL in osteoblasts. Decrea-sed expression of these microRNAs under the influence of excess GCs enhances bone resorption (Clayton et al., 2018).
The transition from the resorption stage to the next stage of bone formation depends on the RANKL / OPG ratio, which increases under GCs induction, affecting osteoclast activity and changing the bone resorptionbone formation balance in the direction of increased resorption. Reduction of OPG under the influence of GCs has a complex multilevel mechanism, a link in which is a direct inhibition of GCs expression of OPG gene (Johnson & Kamel, 2007;Kondo et al., 2008).
In addition, two GC receptor gene polymorphisms have been found, whose carriers may also differ in their response to GC (Szappanos et al., 2009). Individuals with the BclI polymorphism (GG genotype) of the GC receptor gene are prone to increased bone resorption under GC therapy, which was found by elevated levels of the serum bone resorption marker β-CrossLaps and the N-terminal collagen telopeptide type I (NTX) in urine.
While under conditions of normal remodeling the resorption phase ends with programmed death of osteoclasts, which is a counteraction to excessive bone mass loss, under conditions of excess of GCs the life cycle of osteoclasts is increased.
Reversal stage is important in the process of bone remodeling, but in both normal bone and GCO conditions, it is insufficiently studied. There is evidence that in the conditions of GCO the disruption of the interaction between resorption and bone formation occurs precisely at this stage (Jensen et al., 2011;Delaisse, 2014). Histomorphometric analysis of iliac crest biopsies from patients who received GCs for a long time showed that over the cavities of osteoclastic resorption in almost the majority of samples no overhangs of osteoprogenitor cells are formed and no bone tissue is formed (Jensen et al., 2015).
The density of osteoprogenitor cells can decrease due to various mechanisms of GC influence on the differentiation of mesenchymal bone marrow progenitor cells into osteoblasts. One way may be GC-influenced differentiation of mesenchymal stromal cells into adipocytes through GC induction of C/EBP family transcription factor expression, which activates PPARγ transcription, leading to a reduction in the number of osteoblasts (Han et al., 2019). These data suggest that the provision of functional osteoprogenitor cells deserves attention when seeking strategies to prevent bone mass loss that occurs during remodeling in pathological conditions, including the effects of GCs (Jensen et al., 2015).
Formation stage. Experimental studies have shown that an excess of GCs, especially over a long period of time, disrupts the molecular mechanisms of the Wnt/β-catenin signaling pathway from binding to receptor molecules on the cell surface to their nuclear translocation (Guañabens et al., 2014;Chen et al., 2016). Excess of GCs increases the expression of sFRP-1, Dkk-1 and sclerostin inhibitors of Wnt signaling pathway, which affects the proliferation and differentiation of osteoblasts (Meszaros & Patocs, 2020), affecting WNT signaling cascade proteins (Han et al., 2019). Wnt signaling through LRP 5/6 is inhibited by Dkk-1 (Meszaros & Patocs, 2020) (Fig. 11). Sclerostin has an inhibitory effect by binding to the extracellular domain of LRP5, and sFRP-1 modulates Wnt signaling by binding to the Wnt ligand and preventing Wnt receptor activation (Meszaros & Patocs, 2020). GC also decreases phosphorylation of Akt and glycogen synthase-kinase 3β (GSK3β), which enhances β-catenin degradation (Komori, 2016).
In addition, GCs have an inhibitory effect on the ligands of WNT7B and WNT10 signaling pathways, which are crucial for the proliferation and differentiation of osteoblasts (Mak et al., 2009). These data show the complexity of negative actions in this pathway caused by GCs: disruption of osteoblast proliferation and differentiation, increased apoptosis of osteoblasts and osteocytes, which negatively affects the structural and metabolic parameters of bone (Rizzoli & Biver, 2015) and leads to osteoporosis (Chen et al., 2014;Briot & Roux, 2015).
Bone remodeling is regulated by Notch1 and Notch2 signaling pathways. GCs can inhibit osteoblast differentiation by increasing the expression of PPARγ and C/EBPβ in two ways, by stimulating the expression of Notch1 and Notch2 mRNA and by inhibiting the expression of Notch target genes such as Hey1, Hey2, and HeyL in osteoblasts (Han et al., 2019;Zanotti et al., 2019). Synergistic action of Notch and GCs leads to the acquisition of adipocyte phenotype by mesenchymal stem cells.
The effect of excess GCs on osteoblasts may be associated with oxidative stress and formation of intracellular ROS, which include free radicals and peroxides (Almeida et al., 2011;Klein, 2015). Oxidative stress leads to impaired osteoblast proliferation, differentiation and apoptosis, which is activated by dexamethasone through the signaling pathway: ROS / PI3K / AKT / GSK3β (Deng et al., 2019). Under oxidative stress, Forkhead Box Protein Os (FOXOs) nuclear transcription factors act as protective molecules by regulating the osteogenic activity of osteoblast precursors by Wnt signaling . FOXOs stimulate osteoblastogenesis through the Wnt/β-catenin signaling pathway.
The role of the transcription hypoxia-inducible factor 1α (HIF-1α) was studied. HIF-1α regulates the expression of the target gene that con-trols cellular metabolism and in bone tissue leads to an increase in osteoblasts, trabecular bone mass and its vascularization (Tseng et al., 2010). Glucocorticoids have been shown to inhibit HIF-1α, resulting in impaired osteoblast metabolism. Amplification of HIF-1α via the phosphoinositidedependent protein kinase 1 (PDK1) / AKT kinase / mTOR signaling pathway acts as an antagonist that may be useful for treating GCO (Xu et al., 2020).
Data from recent studies concerning microRNA in bone cells in GC excess are summarized in a review (Clayton et al., 2015). MicroRNAs can have both positive and negative effects on osteoblastogenesis. In experimental studies on rats it was shown that prolonged GC treatment decreases MiR-29a expression, disrupting Wnt Dkk-1 and Akt (Akt kinase) signaling, and ERK (extracellular-signal-regulated kinase) phosphorylation, which leads to disruption of osteoblast differentiation (Wang et al., 2013). At the same time, there is other evidence that MiR-29a mitigates glucocorticoid induction through its effect on RUNX2 acetylation (Ko et al., 2015). Excess GCs reduce miR-199a-5p, which positively affects WNT, WNT2, and Frizzled4 (FZD4) WNT receptor signaling on the plasma membrane (Shi et al., 2015a). Dexamethasone-inhibited osteogenesis restores miR-216a, promotes osteoblast differentiation, and enhances bone formation by regulating the PI3K / AKT mediated c-Cbl pathway .
Excess GCs modulates miR-34a-5p and miR-199a-5p, which negatively affects the proliferation and differentiation of osteoblastic cells (Komori, 2016). The mechanism of the effect of GCs on microRNA29a was disclosed. Increased doses of glucocorticoids upset the balance between osteogenic and adipogenic differentiation, and microRNA-29a, mitigates GC induction through its effect on RUNX2 acetylation (Ko et al., 2015). In GC-treated mesenchymal stromal cells, miR-204-5p or miR-125a-3p was shown to expand osteogenesis and inhibit adipogenesis by targeting the marker genes PPARγ and C/EBPα in gene expression studies of osterix and osteocalcin .
The role of long non-coding RNAs (lncRNAs) under conditions of GCs in excess has attracted the attention of researchers. Treatment of mesenchymal stromal cells with glucocorticoids promoted upregulation of lncTCONS_00041960, expression of RUNX2, osterix and osteocalcin genes, while expression of adipocyte-specific markers was decreased. In addition, a novel TCONS_00041960-miR-04-5p/miR-125a-3p -RUNX2 / GILZ axis was identified which is involved in the regulation of adipogenic and osteogenic differentiation of GC-treated mesenchymal stromal cells (Zhang et al., 2020). It is suggested that long intergenic nonprotein coding RNA 311 (LINC00311) is thought to induce proliferation and inhibit osteoclast apoptosis through inhibition of delta-like ligand 3 (DLL3) expression in the Notch signaling pathway, ultimately demonstrating that LINC00311 and its target gene DLL3 may serve as independent factors that are involved in osteoporosis development .
The role of long non-coding RNAs (lncRNAs) under conditions of GCs in excess has attracted the attention of researchers. Treatment of mesenchymal stromal cells with GCs, through the regulation of lncTCONS _ 00041960, promoted an increase in the expression of the RUNX2, osterix, and osteocalcin genes, while the expression of specific markers of adipocytes decreased. In addition, a novel lncTCONS_00041960-miR-04-5p / miR-125a-3p / Runx2 / GILZ axis was identified that is involved in the regulation of adipogenic and osteogenic differentiation of GC-treated mesenchymal stromal cells (Zhang et al., 2020).
Studies of the effects of GCs on non-coding small and LncRNA and their role in osteoporosis need further research.
The peculiarity of this stage is the low density of osteoblasts, associated with the induction of apoptosis by excess GCs through genomic, cytoplasmic and caspase cascade signaling (Smith & Cidlowski, 2010;Mollazadeh et al., 2015). Glucocorticoids regulate gene transcription leucine zipper (GILZ), Bim induction via Fox03A/FKHRL1 transcription factor (Smith & Cidlowski, 2010). Cytoplasmic apoptosis signaling is a multicomponent pathway induced by GCs: decreased mRNA levels of antioxidant enzymes, increased ROS, overexpression of GR species specifically targeting mitochondria, etc. Mitochondria also mediate glucocorticoid-induced cytoplasmic effects, including calcium mobilization, release of reactive oxygen species ROS and ceramide, which promote apoptosis.
This transactivation of the Bim gene leads to the expression of apoptotic mediators Bax and Bak. Under their influence the mitochondrial membrane potential is disrupted with the subsequent release of cytochrome C and Apaf-1 into the cytosol, which is accompanied by the activation of caspase 9, effector caspases 3, 6, 7 and leads to apoptosis (Smith & Cidlowski, 2010) (Fig. 13).   Fig. 13. GC-induced mitochondrial Bim -Bax/Bak signaling disrupts mitochondrial function leading to cell apoptosis (adapted from Smith & Cidlowski, 2010) Expression of the key transcription factor of osteoblastogenesis RUNX2 is impaired under the influence of GCs, increases the expression of osterix transcription factor, which inhibits the differentiation and metabolism of osteoblasts and as a consequence decreases the biosynthesis of collagen type 1, proteoglycans, osteocalcin, osteonectin, alkaline phosphatase, β-glycerophosphate and other proteins involved in osteoid formation by osteoblasts. In addition, GCs inhibit the production of prostaglandin E, IGF-1 and TGF-β, which activate osteoblasts, which also leads to a decrease in the expression of bone matrix proteins in osteoblasts, leading to disruption of osteoid formation (Ito et al., 2007).
Excess of GCs leads to changes of gene expression in the Wnt signaling pathway, which influences bone formation, and to changes in genes that regulate functional activity of osteoblasts, osteocytes and mineralization of osteoid. That is, already at the stage of osteoid formation, due to a decrease in the pool of osteoblasts, settlement of resorption cavities by osteoblasts and formation of the organic bone matrix is delayed.
Termination stage. Continuous exposure to excess GCs leads to increased expression of DMP1 and Phex genes, which inhibit mineralization. Glucocorticoids in excess stimulate osteocytic osteolysis, which leads to bone matrix destruction, increases the size of osteocyte lacunae, disrupts mineralization by affecting the key biomarkers (Yao et al., 2013;Matsuo, 2014). Due to the predominance of resorption over bone formation, most resorption cavities remain unfilled with bone tissue.
So, excess GCs lead to a disruption of bone remodeling by negatively affecting the major signaling pathways and active molecules involved in this process.

Conclusion
This review provides a rationale for the cellular and molecular mechanisms of bone remodeling stages under physiological conditions and in glucocorticoid excess. The signaling pathways of cell proliferation, differentiation, viability, and death during remodeling are presented. The main signaling pathways that control bone resorption and formation are RANK / RANKL / OPG, M-CSF -c-FMS, canonical and non-canonical Wnt, Notch, MARK, TGFβ / SMAD ephrinB1/ephrinB2 -EphB4, TNFα -TNFβ, and Bim -Bax/Bak signal pathways. Cytokines, growth factors, prostaglandins, parathyroid hormone, vitamin D, calcitonin, estrogens, non-coding microRNAs and long RNAs also act as regulators of bone remodeling. Excess glucocorticoids negatively affect all stages of bone remodeling, disrupt molecular signaling, induce apoptosis of osteocytes and osteoblasts, increase osteoclast life cycle. Excess glucocorticoids disrupt the reversion stage: the formation of canopies of osteoprogenitor cells over resorption cavities, which is critical for the subsequent stages of remodeling. Under the influence of excess glucocorticoids there is a disturbance of osteoid mineralization. Although many signaling pathways involved in resorption and bone formation have been discovered and described, the temporal and spatial mechanisms of their sequential activation and deactivation in cell proliferation and differentiation require additional research.