Obesity and osteoporosis are characterized by high prevalence and mortality rates due to associated complications; therefore, they represent global health problems. Osteoporosis is a systemic, multifactorial disease characterized by the loss of bone mass, impaired metabolism, and disrupted bone tissue microarchitecture, which are accompanied by an increase in skeletal fragility and fracture injuries caused by minimal trauma [1].

The effects of obesity on bone health are very controversial. Increased body weight due to obesity increases the mechanical load on the bones, stimulating bone growth. In subcutaneous and visceral adipose tissue, aromatase synthesizes estrogen, which inhibits osteoclast activity and increases bone mass [2]. Therefore, an increased body mass index (BMI) value in postmenopausal women is associated with estrogen production and increased osteoblast activity, which generally reduces the risk of fractures. However, a critical increase in adipose tissue has been shown to exert a negative effect on bone mass [3]. Thus, osteoporosis is exacerbated by proinflammatory cytokines with proresorptive properties (tumor necrosis factor (TNF)-$\alpha$, interleukin (IL)-15, IL-6, IL-7, IL-8, IL-17, and receptor activator of nuclear factor kappa-B ligand (RANKL)) and some adipokines, the levels of which increase in obesity [4]. The influences of these mediators on signaling pathways (in preosteoblasts, osteoblasts, and stromal cells) lead to a shift in bone remodeling processes toward increased osteoclastogenesis, resulting in bone resorption.

With increasing age, bone marrow mesenchymal stem cells (BMSCs) exhibit a decreased capacity for differentiation into osteoblasts, and adipogenic differentiation increases with an increase in the volume of yellow bone marrow. Furthermore, under visceral obesity conditions, BMSC differentiation into adipocytes is promoted, which has a negative effect on bone strength [5, 6, 7, 8]. Thus, with increasing age against a background of morbid obesity, adipocytes become the predominant stromal cells in bone marrow [9, 10].

The modern view of bone formation and resorption is based on epigenetic regulation mechanisms [11]. The involvement of microRNAs is considered key to the control of BMSC gene expression [12, 13].

MicroRNAs are small noncoding RNAs with lengths of 19–22 nucleotides, and they control biological and pathological processes, including organogenesis, apoptosis, cell proliferation, and differentiation [14]. MicroRNAs bind to specific sequences in the 3’ and 5’ untranslated regions (UTRs) of their target mRNAs to induce translational repression and mRNA deadenylation and decapping [11, 14]. Studies have shown that microRNAs are stable outside the cell and do not degrade when held at room temperature for up to 24 hours or at –80 °C for up to 9 months or after thawing in plasma/serum eight times [15]. Over the past 20 years, the number of publications in the PubMed database (https://pubmed.ncbi.nlm.nih.gov) retrieved with the search keyword “microRNA and cell differentiation” has increased by 780 fold (reaching 1557 articles in 2021). However, the roles of microRNAs in the autocrine and paracrine regulation of BMSC differentiation, including the development of osteoporosis in obesity, have not been adequately studied.

This review is dedicated to analyzing the roles and clinical potential of microRNAs in the epigenetic control of BMSCs and osteoclast differentiation in osteoporosis with obesity.

In the differentiation of mesenchymal stem cells, a cascade of events causes phenotypic and metabolic changes. Decreased expression of stem cell genes and the activation of genes associated with mature phenotypic function are the first steps in the cascade, and they lead to cellular morphological changes. Several studies have shown that factors and signaling pathways that stimulate adipogenesis simultaneously inhibit osteogenesis [16]. The reciprocal balance of the transcriptome between adipogenesis, osteogenesis, and chondrogenesis is essential for skeletal tissue development and the maintenance of homeostasis in a mature organism [17].

In healthy bones, the continuous resorption of old bone mainly by osteoclasts and the formation of new tissue by osteoblasts occur. Osteoclasts originate from the hematopoietic lineage [18] and monocyte/macrophage-derived osteoclast precursors [19]. Osteoblasts originate from BMSCs and can differentiate into adipogenic clones [20]. The microenvironment strictly controls BMSC differentiation, and imbalanced differentiation leads to a higher rate of adipogenic MSC differentiation, which is detrimental to osteogenesis [17, 20, 21].

An increase in adipose tissue mass leads to serious effects on the bone marrow microenvironment due to secreted factors. That is, the secretion of proinflammatory cytokines, adipokines, free fatty acids, exosomes, and microRNAs by adipocytes stimulates osteoclasts and promotes increased bone resorption [22]. In addition, obesity contributes to a disturbance/alteration in adipokine secretion; therefore, the secretion of adiponectin decreases, and the production of leptin and proinflammatory factors increases. Studies of transgenic mice harboring the human adiponectin gene demonstrated the stimulatory effect of adiponectin on bone formation through the activation of the Wingless-type MMTV integration site (WNT)/$\beta$-catenin pathway. Adiponectin action is mediated through receptors (AdipoR1 and AdipoR2) located on BMSCs [23]. In addition, the effects of leptin on osteogenesis are multidirectional. Leptin may directly or indirectly affect bone metabolism through central hypothalamic and peripheral signaling pathways [24]. A study of mice deficient in leptin and its receptors showed a reduced bone formation rate, which increased significantly after the subcutaneous administration of leptin [25]. Leptin receptors were found on red bone marrow MSCs, osteoclasts, and osteoblasts [26]. The administration of leptin by intraventricular infusion resulted in bone loss in both leptin-deficient and wild-type mice [27].

Increased TNF-$\alpha$ production by adipocytes inhibits osteoblastogenesis and is a potent trigger of bone resorption [28]. IL-15 stimulates the secretion of TNF-$\alpha$, which further enhances the osteoclastogenic effect [28]. In preosteoblasts, TNF-$\alpha$ upregulates nuclear factor-$\kappa$B (NF-$\kappa$B) activity by inhibiting the bone morphogenic protein (BMP) 2-induced production of runt-related transcription factor (RUNX) 2.

Simultaneously, TNF-$\alpha$ initiates the expression of RANK [28] and blocks the Wnt pathway by activating the SOST and DKK-1 genes [9]. Razny et al. (2019) [29] showed that patients with obesity and insulin resistance have defects in osteoblastogenesis associated with decreased expression of ß-catenin and microRNAs involved in the regulation of osteoblast differentiation and increased expression of Wnt signaling pathway inhibitor genes (DKK-1, DKK-2, SOST, and SFRP1), leading to a predominance of osteoclastogenesis.

Obesity is associated with increased osteoblast production of RANKL, which regulates osteoclast differentiation, proliferation, survival, and apoptosis. Thus, adipose tissue secretes M-CSF, which induces RANKL-mediated bone resorption by triggering the PI3K/AKT/NFkB signaling cascade, activating NFATC1, or upregulating the expression of the transcription factors HIF-1$\alpha$ and PPARy by activating the PI3K/AKT/mTOR pathway [4, 9]. IL-1, IL-6, IL-11, IL-15, and IL-17 are proresorptive cytokines produced by activated T lymphocytes that stimulate RANKL production [30].

Osteoprotegerin (OPG), which is a member of the TNF receptor superfamily (TNFRSF11B), is secreted by osteoblasts. It is a RANKL decoy receptor that blocks the RANKL/RANK interaction [31, 32]. An increased OPG level with age is considered a homeostatic response to limit bone resorption [33]. It should be considered that high insulin levels in obesity interact with insulin receptors on osteoblasts, blocking the production of OPG and increasing the secretion of osteocalcin [34]. The presence of obesity and type 2 diabetes is associated with the active production of oxidized lipids and carbohydrates in addition to damaged proteins, leading to the formation of enhanced glycation end products (AGEs). AGEs interact with the AGE receptor (RAGE) and trigger a cascade of intracellular responses (NFkB, AP-1, CREB, STAT3, and NFAT), leading to the development of inflammatory responses [35]. The accumulation of AGEs in bone tissue leads to the development of pathological conditions associated with low bone density and osteoporosis [36]. Considering these finding, the triggering of inflammatory responses is a link between metabolic disorders, osteogenesis and bone resorption.

The intracellular mechanisms that determine MSC fate in the context of adipogenic and osteogenic differentiation are under active investigation [17]. The processes of MSC differentiation into adipocytes and osteocytes in vitro can be conditionally classified into two stages: (1) clone predetermination (the commitment stage) from an MSC to a committed progenitor and (2) maturation (the terminal differentiation stage) from the progenitor to a cell with a mature phenotype [16, 21].

Numerous studies have shown that the major signaling pathways involved in regulating MSC lineage commitment are the transforming growth factor-beta (TGF$\beta$)/BMP/specific E3 ubiquitin protein ligase (Smads), WNT, Hedgehog (Hh), Notch, fibroblast growth factor (FGF), Janus kinase/signal transducer and activator of transcription (JAK/STAT), mitogen-activated protein kinase (MAPK), phosphatidylinositol 3-kinase (PI3K/Akt), and extracellular signal-regulated kinase (ERK1/2) pathways [21, 37].

Adipogenic differentiation is associated with increased production of cyclic AMP, leading to the phosphorylation of the transcription factor CCAAT/enhancer-binding protein (C/EBP). Upon activation, C/EBPB binds to regulatory elements and stimulates the transcription of C/EBPA and peroxisome proliferator activated receptor gamma (PPAR$\gamma$). When C/EBPB synthesis is inhibited, the transcription of C/EBPA and PPARG is continuously promoted because C/EBPA binds to regulatory C/EBP elements. PPAR$\gamma$ transcription is also controlled by sterol regulatory element-binding protein (SREBP1) transcription factors and the Kruppel-like factor family [21]. The later stages of adipogenic differentiation are associated with the expression of fatty acid synthase, glycerophosphate dehydrogenase, acetyl-CoA carboxylase, glucose transporter type 4, insulin receptor, and acetyl selective fatty acid binding protein (aP2) [17].

Thus, adipogenesis and osteogenesis are regulated by complex factor cascades, and the role of RNA-dependent epigenetic regulation is extremely important. MicroRNAs form interaction networks that provide additional control over biological processes. In regenerative medicine, MSCs are used in alternative therapies for the treatment of obesity [38, 39] and bone regeneration [17]. Understanding the regulatory pathways involved in cell commitment and investigating the factors that regulate the balance between osteogenic and adipogenic differentiation in MSCs are of great importance for the theoretical understanding of the pathogenesis of osteoporosis in obesity and will enable the identification of molecular targets for the treatment of osteoporosis in the future.

The long-term use of drugs for the treatment of osteoporosis is challenging because of a number of limitations, including severe gastrointestinal side effects, thromboembolism, and increased risk of sarcoma [130]. In this regard, new and safe therapeutic approaches are needed to correct this pathology. MicroRNAs are antagonists of gene expression and can be used to treat osteoporosis and osteoporotic fractures [130]. Notably, existing drugs for the treatment of osteoporosis (e.g., teriparatide and denosumab) may alter the microRNA expression profile. For example, in women with osteoporosis, teriparatide decreased the expression levels of miR-133a and miR-33, and after treatment with denosumab, increases in the expression levels of miR-503 and miR-222 as well as decreases in the activity of the genes related to osteoclastogenesis, namely, RANK and cathepsin K (CTSK), were observed [131].

In the classical approach, microRNAs are used to restore the expression of a missing microRNA. This outcome is often achieved through the use of expression constructs delivered through viral vectors [132].

On the other hand, gene expression can be controlled in other ways. Certain approaches inhibit pathological microRNA expression. The loss of miRNA function can be achieved by using microRNA sponges, microRNA masks, and anti-microRNA oligonucleotides (anti-miRs) [130].

Different types of chemical modifications are used to introduce anti-miR effects [130]. For example, miR-31, miR-106a, and miR-148a inhibit MSC-BM differentiation in osteoblasts by directly targeting RUNX2 [133] (Table 1). After in vitro transfection of anti-microRNA into cells, osteogenic gene expression and ALP activity were increased [133]. Chorionic cells and placental cells were used as alternative sources of MSCs in this study [134]. In addition, inhibition of the expression of specific microRNAs increased the ability of MSCs to undergo osteogenic differentiation culture in vitro, and this strategy can be used to stimulate bone regeneration.

Targeting osteoporosis marker genes for the development of microRNA-based therapies seems to be a promising clinical direction. In addition, curbing excessive osteoclastogenesis can be achieved through the search for molecular switches that direct BMSC differentiation from favoring adipogenesis to favoring osteogenesis [130]. However, no microRNA-based therapy that can be successfully used in the clinic has been found. We analyzed the literature to identify microRNAs with high potential for use as diagnostic markers of osteoporosis, which may also serve as a basis for the development of oligonucleotide therapeutic approaches in the future (Table 4, Ref. [13, 78, 87, 104, 106, 119, 121, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153]).

Clinical trials for the use of microRNAs have been performed to study osteoporosis in women [154], men [155], hyperparathyroid patients with and without osteoporosis, and osteoporosis patients undergoing thyroidectomy [156]. MicroRNAs have been studied as potential biomarkers of coronary artery calcification (CAC), and their relationships with bone mineral density and use as markers of bone metabolism have been investigated [157]. Biomarkers (including microRNAs) in orthopedic and maxillofacial surgery patients have been sought to identify risks associated with bisphosphonate use [158]. A study on the relationships between the gut microbiota and bone microarchitecture has also been conducted [159].

In addition, a study on bone-specific microRNAs in the blood serum of postmenopausal women with osteoporosis against a background of antiresorptive or osteoanabolic therapy has been performed. The results are expected to be reported after the end of the study in 2022 [160].

A number of clinical trials have been performed to test oligonucleotide therapies, including treatment with antisense oligonucleotides [161, 162, 163, 164].

However, despite the many approaches used to stabilize oligonucleotides in cells, oligonucleotide bioavailability is still low, and their efficient delivery to target organs other than the liver remains a challenge [165]. Increasing the concentration of administered oligonucleotides can lead to toxic side effects [166]. Therefore, methods for the delivery of oligonucleotides via exosomes are under development for future application [167].

Osteoporosis remains an incurable and serious multifactorial disease. The relationship between osteoporosis and obesity and the regulation of osteoclast activity, in particular, is detailed in a review by Roy et al. (2016) [9]. Recently discovered miRNAs involved in the epigenetic control of mRNA play important roles in regulating the adipogenic/osteogenic differentiation of BMSCs and osteoclastogenesis [168, 169, 170].

In the present review, we summarize recent findings on the regulatory interactions between mRNAs and miRNAs in the regulation of BMSC differentiation and the activation of osteoclastogenesis in obesity and osteoporosis. We hypothesize that miR-122, miR-21, miR-125, miR-23-3p, miR-140, miR-2861, and miR-133 have high potential for use in the diagnosis of osteoporosis. However, in view of the large body of work devoted to the search for biomarkers of pathologies, this finding remains to be confirmed because the changes in miRNA levels in blood serum/plasma have multidirectional dynamics.

Many studies show the promising possibility of using short-stranded noncoding RNAs for therapeutic purposes [171, 172]. The complex regulation of the expression of genes important for the balance between osteogenesis and osteolysis gives us hope for the development of a safe therapy for osteoporosis. To reduce the severity of osteoporosis, treatment should ideally target the following: the hypoactivation of osteoclasts, the restoration of osteoblast function, the prevention of adipogenic differentiation in BMSCs, and the restoration of bone perfusion. In this regard, to reduce the severity of adipogenesis in BMSCs, we suggest therapy with inhibitors, such as miR-204, miR-211, miR-217, miR-221, miR-222, miR-128, miR-140, miR-16, miR-188, miR-143, miR-145, and miR-532. The positive regulation of osteoblastogenesis in BMSCs can be achieved by increasing the levels of miR-296-3p, miR-2861, miR-3960, miR-335, and miR-29a mimetics in the BMSC microenvironment.

Perhaps, to inhibit osteoclast hyperactivation, therapy with miR-483 inhibitors or miR-338, miR-133a, miR-151, and miR-125 mimetics should be considered. The ambiguous roles of miR-128 and miR-145 in the regulation of osteoclasts require further research.

The relevance of the search for targeted miRNAs stems from the fact that promoting the reparative regeneration of bone tissue (osteosynthesis) under its pathological remodeling conditions in osteoporosis is an unsolved clinical problem. Among the numerous available drugs, BMPs have direct bone tissue regenerative properties [173]. However, when administered systemically, they cause numerous side effects (including bone resorption) in 20–70% of cases, which negates the benefit of their specific osteogenic activity [174]. An obstacle to the widespread use of BMPs is also their high price (due to high costs in the research and development phase), which is disproportionate to the therapeutic efficacy of the drugs.

In this context, the therapeutic use of miRNAs, which modulate the (post)transcriptional activity of genes at the mRNA level, is considered a promising new biotechnological direction in regenerative medicine. However, a number of unresolved issues remain, especially regarding the potential side effects caused by (multiple) crossover effects of microRNAs on intracellular signaling pathways, which may lead to unpredictable responses in cells of nontarget tissues and organs in the long term.

Therefore, the development of local systems for the targeted delivery and release of microRNAs (capsules of different sizes, nanoparticles, scaffolds, etc.) that overcome tissue barriers independently [175] or with the help of carrier cells (e.g., MSCs) [176] is an urgent task to be addressed in the near future.

BMSCs, bone marrow mesenchymal stem cells; MSCs, mesenchymal stem cells; BMI, body mass index; TNF, tumor necrosis factor; IL, interleukin; UTRs, untranslated regions; RANKL, receptor activator of nuclear factor kappa-B ligand; WNT, Wingless-type MMTV integration site; BMP, bone morphogenic protein; RUNX, runt-related transcription factor; NF-kB, nuclear factor-$\kappa$B; TGF, transforming growth factor; Smads, specific E3 ubiquitin protein ligase; JAK/STAT, Janus kinase/signal transducer and activator of transcription; MAPK, mitogen activated protein kinase; ERK, extracellular signal-regulated kinase; C/EBPB, CCAAT-enhancer-binding protein; PPAR$\gamma$, peroxisome proliferator activated receptor gamma; aP2, fatty acid binding protein; SIRT, sirtuin; ALP, alkaline phosphatase; VEGF, vascular endothelial growth factor; TGFBR1, transforming growth factor receptor 1; RICTOR, rapamycin-insensitive companion of mammalian target of rapamycin; mTOR, target of rapamycin; ETS, ETS proto-oncogene; Cbfal, core-binding factor al; HDAC5, histone deacetylase 5; HOXA2, homeobox A2; iPSCs, pluripotent stem cells; I$\kappa$B$\beta$, inhibitor of nuclear factor kappa-B kinase; TNFRSF1B, tumor necrosis factor ligand superfamily member 1B; anti-miR, santi-microRNA oligonucleotides.

MV, NT, KY (Author A, C and D)—original draft preparation; IKh, LL, IK (Author B, F and J)—editing; MV, NT, KY, VM, DS, AK, OKh (Author A, B, C, D, E, G, H)—gathered information and performed material systematization; MV (Author A) visualization. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.

Not applicable.

Not applicable.

This research was funded by the State Assignment Grant number FZWM-2020-0010, the Grant of the President of the Russian Federation number MK-2072.2022.3 and Immanuel Kant Baltic Federal University and Siberian State Medical University Development Program Priority 2030.

The authors declare no conflict of interest.