INSULIN RESISTANCE AND PATHOGENESIS OF POSTMENOPAUSAL OSTEOPOROSIS

Abstract

Osteoporosis (OP) is a disease predisposing postmenopausal women to fractures, and often accompanied by insulin resistance (IR) and metabolic syndrome (MetS). Previous studies provided contradictory results concerning prevalence of MetS in postmenopausal OP. To better understand the pathogenesis of IR, we reviewed cellular and molecular aspects and systematically reviewed studies providing homeostasis model assessment (HOMA) index. Bone is an active endocrine organ maintaining its integrity by orchestrated balance between bone formation and resorption. Both osteoblasts and osteoclasts contain receptors for insulin and insulin-like growth factor-1 (IGF-1) operating in skeletal development and in the adult life. Defects in this system generate systemic IR and bone-specific IR, which in turn regulates glucose homeostasis and energy metabolism through osteocalcin. Examination of genetic syndromes of extreme IR revealed intriguing features namely high bone mineral density (BMD) or accelerated growth. Studies of moderate forms of IR in postmenopausal women reveal positive correlations between HOMA index and BMD while correlations with osteocalcin were rather negative. The relation with obesity remains complex involving regulatory factors such as leptin and adiponectin to which the contribution of potential genetic factors and in particular, the correlation with the degree of obesity or body composition should be added.

INTRODUCTION

Osteoporosis (OP) is a multifactorial and debilitating chronic disease characterized by reduced bone strength, predisposing affected individuals to fragility fractures due to diminished bone mass and altered bone architecture. The estimated prevalence of OP stands at 18.3%, signifying that approximately one in three women aged 50 or older will experience osteoporotic fractures during their lifetime (1). An additional serious issue among postmenopausal women, especially with advanced age, is the occurrence of metabolic complications, including obesity, type 2 diabetes mellitus (T2D), and metabolic syndrome (MetS), all hallmarked by insulin resistance (2). This alteration has a complex pathogenesis in which insulin and insulin-like growth factor (IGF)-1, known for maintaining organism fitness, governing growth, glucose homeostasis, and life span, play crucial roles (3).

Despite significant progress, the mechanisms involved in insulin resistance in postmenopausal women and its role in the pathogenesis of OP are not completely understood. When exploring MetS as a clinical manifestation of insulin resistance, the scientific literature presents contradictory results. Some studies suggest a high prevalence of MetS in postmenopausal women with OP, while others report a lower prevalence with a protective effect (4, 5). Several reasons might account for these discrepancies, such as variations in MetS definitions or ethnic differences among populations. However, a central issue appears to be related to specific mechanisms involved in insulin resistance concerning bone homeostasis. In the first part of this article, we will present some clues regarding the cellular and molecular aspects of bone-specific insulin resistance and insights from genetic syndromes of severe insulin resistance. In the second part, we systematically reviewed publications reporting measured insulin resistance in postmenopausal women with OP. In November 2023, we screened Pubmed and Scopus databases using the following string: (HOMA OR HOMA-IR) AND (bone OR TBS OR osteoporosis OR fracture OR osteoblast OR osteoclast OR osteocyte). From 565 articles we selected the studies that included postmenopausal women with measured values of homeostasis model assessment (HOMA-IR) and various bone parameters (such as bone mineral density or bone turnover markers). Studies that included diabetic patients only were excluded, given the fact that diabetes mellitus per se and sometimes its treatment have skeletal consequences beyond those of insulin resistance that may represent confounding factors. We adopted HOMA-IR as a common measure of insulin resistance, considering this index's robustness and reliability as an epidemiological parameter, while acknowledging that euglycemic clamp technique remains the gold standard for studying insulin resistance in vivo. It is crucial to note from the outset that insulin resistance is highly complex. Our research team distinguishes between 'insulin resistance vera', involving defects in insulin signalling at the cellular and molecular level, and other forms of systemic insulin resistance occurring through regulatory mechanisms (e.g. immunological). We will also briefly review aspects of insulin resistance in relation to bone health in rare syndromes characterized by severe insulin resistance due to defects in insulin/IGF-1 receptors. Insulin resistance is defined as a complex condition characterized by reduced tissue response to insulin and underlies various disorders with significant health repercussions, including obesity, T2D, polycystic ovary syndrome (PCOS), and neurodegenerative diseases (6).

As indicated, insulin resistance also constitutes a pivotal component of MetS. This syndrome was successively defined by expert groups such as the World Health Organization (WHO), Adult Treatment Panel (ATP-III) of the National Cholesterol Education Program (NCEP), American Heart Association/National Heart, Lung and Blood Institute (AHA/NHLBI) (NCEP-R), and more recently by the International Diabetes Federation (IDF). The IDF's revised definition considered ethnicity-specific waist circumference (WC) criteria, especially pertinent in Asian populations (7). Variable definitions of MetS have led to challenges in reporting its prevalence. For instance, a 2007 study in France revealed MetS prevalence as 11.7%, 20%, and 26% according to NCEP, NCEP-R, and IDF criteria respectively, underscoring the importance of stringent criteria (8). Despite its variability, the role of MetS in human pathology is well recognized since this syndrome amplifies cardiovascular and metabolic risks in the general population by increasing the likelihood of T2D by fivefold and cardiovascular diseases by twofold (9). Therefore, insulin resistance and MetS have attracted particular attention in postmenopausal women. Furthermore, the relation with bone metabolism and health was the object of several recent reviews in the literature (10-17).

Bone, while mechanically rigid, is metabolically highly active, engaging in numerous biological processes such as mineral storage and endocrine functions (18). The integrity of bone is maintained through a finely controlled balance between bone resorption, carried out by osteoclasts (OC), and bone formation, orchestrated by osteoblasts (OB). Various factors regulate bone modeling during development, with some continuing to operate in adult life, overseeing bone "remodeling" and turnover. Menopause in women is characterized by a significant increase in the rate of bone remodeling, primarily favoring bone resorption (11). This process aligns with estrogen deficiency after menopause, triggering a dramatic escalation (50-100%) in bone turnover, especially in the trabecular bone (such as the lumbar spine and pelvis). Although numerous studies acknowledge the impact of insulin resistance on women as they age, the direct association with OP remains less understood, particularly when examining patients for the presence of MetS. Both endocrine mechanisms (e.g. estrogen deficiency) and bone-specific factors may contribute to the pathogenesis of insulin resistance in postmenopausal women.

Bone metabolism and its role in insulin resistance. Bone is composed of several types of cells with specific functions. Bone formation primarily relies on OB, small cuboidal cells derived from mesenchymal stem cells (MSC), and regulated by the master transcriptional factor runt-related transcription factor 2 (RUNX2), also called core-binding factor subunit alpha-1 (CBF-alpha-1). The effect of transcriptional factor Sp7 (osterix, OSX) potentiates this process. OB undergo proliferation, resulting in mineralization of the osteoid of the bone matrix and eventual apoptosis (19). OB progenitors differentiate under RUNX2 into pro-osteoblasts expressing bone alkaline phosphatase (ALP) and producing alpha-1 type I collagen (COL1A1). Throughout differentiation, various factors, including fibroblast growth factors (FGFs), microRNAs, and connexin 43, play active roles. Besides collagen, OB also secrete osteocalcin (OCN), osteopontin (OPN), and bone sialoprotein (BSP). Experimental knockdown of RUNX2 leads to the absence of OB, directing differentiation towards the adipogenic lineage through the canonical Wnt/β-catenin pathway (20, 21).

OC, large, multinucleated cells, are pivotal in bone resorption. They originate from mononuclear cells under the influence of interleukin-3 (IL-3) and macrophage colony-stimulating factor (M-CSF), which possess a specific receptor (cFMS). The RANKL system and osteoprotegerin (OPG) play essential roles in this process. RANKL activates the expression of specific genes, including tartrate-resistant acid phosphatase (TRAP) and cathepsin K (10, 22). During resorption, cells form a ruffled border that acidifies the milieu necessary for hydroxyapatite dissolution. Cathepsin K and matrix metalloproteinases (MMPs) further contribute to digestion. Additionally, lining cells, flat-shaped cells covering the bone surface, have an uncertain role within bone multicellular units (BMUs). Positioned to prevent close contact between cells, they can secrete osteoprotegerin. Osteocytes, representing 90% of the bone cell population, are long-lived dendritic cells sensitive to mechanical stimuli. Derived from MSCs under podoplanin (E11/gp38, gp36 in humans, Chr 1) influence, they are located in the lacunocanalicular system near blood vessels. Their function involves transforming mechanical stimuli into biochemical signals (known as the "piezoelectric" effect), mediated by Polycystins 1 and 2 and actin-associated proteins such as vinculin, talin, zyxin, and paxillin (10, 23).

All these cellular components can contribute to the pathogenesis of bone-specific insulin resistance or, through inter-organ regulation, impact systemic insulin resistance by modulating energy metabolism. To these, we should add the role of bone marrow that has been demonstrated to play a pivotal role in energy metabolism (24). Indeed, bone marrow adipose tissue (BMAT) is derived from stem cells (different from other adipocytes in the organism) and plays an endocrine role in energy homeosytasis expanding in anorexia nervosa, estrogen deficient states or growth hormone (GH) deficiency. Thus, it was shown that obese postmenopausal women carrying high levels of IGF-1 display lower vertebral marrow fat content, independent of body mass index (BMI) (25).

An important distinction exists between factors regulating bone formation (modelling) during skeletal development and those governing regular functions in the adult life (remodelling). Bone formation during development involves both endochondral processes (seen in long bones, facial bones, vertebrae, and the medial clavicle) and intramembranous processes (observed in the cranium and lateral clavicle). Endochondral ossification begins with the formation of a cartilaginous template followed by mineralization and ossification. In chondrogenesis, within mesenchymal condensation, numerous influential transcription factors and signalling molecules are involved, including the SRY-box transcription factor 9p (Sox9, Chr 17), which plays a crucial role. The proliferation and hypertrophy of chondrocytes are regulated by a balanced interplay among PTH-related peptide (PTHrP), bone morphogenetic protein (BMP), the Wnt system, as well as a series of fibroblast growth factors or FGFs (21, 26, 27). In the ossification center, the osterix lineage in the perichondrium forms the trabecular zone, while collagen lineages of OB contribute to the cortical zone. By contrast, cranial bones involve a direct condensation of mesenchymal cells through OB differentiation (27).

Regarding FGFs, three of them (FGF 1, 2, and 3) operate in adult life during bone remodeling (21, 27). These FGFs bind to specific tyrosine kinase receptors, although FGFR5 lacks enzymatic activity. FGF tyrosine kinase receptors, through the adaptor protein fibroblast growth factor receptor substrate 2 (FRS2α, Chr 12), lead to intracellular activation of STAT 1, 3, and 5. Phosphorylation of FRS2α can activate the MAPK pathway through interaction with GRB2, while the activation of the PI3K-AKT pathway occurs through GAB1. The action of endocrine FGFs necessitates a co-factor for their activity, represented by Klotho proteins. Recently, the role of FGFs in bone formation was reviewed (27). Elevated expression of FGF2 in OB leads to increased trabecular and cortical bone mass, involving the Wnt system and the modulation of FGFR23 expression. Consequently, maintaining a balance between FGF23 and FGF1 plays a crucial role in controlling bone mass and mineralization. Furthermore, FGFs have the ability to regulate OC activity in bone resorption. FGF2 in OC stimulates prostaglandin endoperoxydase synthase-2 (Ptgs2, Chr1), resulting in the production of prostaglandins (Fig. 1).

Figure 1.

Overview of the relationship between systemic insulin resistance and bone-specific insulin resistance.

Insulin/IGF-1 system and bone homeostasis. Bone remodeling in the adult life depends on the balance between the effect of estrogens and growth factors like insulin and IGFs. Insulin plays a pivotal role in glucose metabolism whereas IGFs are secreted both as circulating hormones and in a paracrine manner, participating to mitogenic activity and growth. While IGF-2 is crucial during fetal life, the GH/IGF-1 system is fundamental for achieving normal longitudinal bone mass and the acquisition of peak bone mass during development (28). Action of IGF-1 is exerted directly on bone and involves the essential IGFBP-3 and its acid-labile subunit (ALS). Decrease of IGF-1 levels after menopause might partially explain the reduction of bone mineral density (BMD) (29, 30). Moreover, studies have shown a strong correlation between lower circulating IGF-1 levels and an increased risk of bone fractures in postmenopausal women, independent of BMD (29).

At the molecular level, insulin and IGF-1 exert their effects by binding to specific membrane receptors. IGF2 binds to a specific receptor lacking enzymatic activity, thus its action is explained through binding to IGF-1R (31). The insulin receptor (IR) and IGF-1R are 80% homologous heterotetrameric transmembrane proteins (α2β2) comprising α-subunits binding the ligand and β-subunits containing a tyrosine-specific kinase. Ligand binding triggers a conformational change in the β-subunit. Activation is enhanced through progressive and coordinated autophosphorylation at three Tyr sites, while the phosphorylation of several Ser residues inhibits the Tyr-kinase (31). Activated kinase phosphorylates various substrates (such as insulin receptor substrate or IRS-1, 2 and 3 in humans), leading to the activation of PI3 kinase and PDK1, a Ser/Thr kinase capable of activating protein kinase B (AKT) (32). Downstream, AKT has multiple targets, including FoxO1 and mTOR. FoxO1, expressed in OB, when deleted, induces decreased bone mass. mTOR, also expressed in OB, controls cell proliferation by phosphorylating initiation factor 4E binding protein and p70S6 kinase (33). Glycogen synthase kinase 3 α and β (GSK3), another substrate of AKT, play a crucial role in the regulation of the Wnt signaling pathway. This kinase phosphorylates β-catenin and LRP6, leading to activated β-catenin. Apart from these cellular effects, insulin and IGF-1 stimulate glucose transport by activating GLUT4 (insulin) and GLUT 1 and 3 (IGF-1) (34).

The insulin/IGF-1 system plays a crucial role in chondrocyte differentiation and function, as it includes osteoanabolic hormones, acting primarily via paracrine effects. Global deletion of IGF-1 and IGF1R leads to skeletal defects and a reduction in BMD. Circulating IGF-1, primarily produced by the liver under the influence of GH, is significant in bone development. Liver-specific deletion of IGF-1 results in decreased cortical bone and bone strength, although it does not affect trabecular bone. Insights into the role of insulin/IGF-1 in bone have been gained from genetically modified mouse strains. While their roles in OB and OC are well understood, the impact of IGF-1 on energy metabolism remains more complex (34).

OB contain functional insulin receptors (35). Clinical studies have revealed that Type 1 diabetes (T1D) without insulin is associated with decreased BMD, whereas T2D characterized by insulin resistance and hyperinsulinemia is linked to high BMD. Specific deletion of IR in mice (Ob-dIR) leads to failed OB maturation, reduced expression of Runx2 due to Twist1 downregulation, and decreased osteocalcin (OCN) mRNA. These transgenic mice exhibit reduced trabecular bone later in life and unexpectedly develop adiposity, with a 40% increase in fat mass, alongside systemic insulin resistance and glucose intolerance (36). Loss of IR in OB (Ob-ΔIR) by > 47% affects postnatal trabecular bone along with a 79% decrease in OB numbers. While loss of IGF-1R in OB can be rescued by insulin, OB lacking IR were not rescued by IGF-1. In OC, insulin enhances bone resorption by increasing the expression of cathepsin K, a lysosomal protease, and Tcirg1, a proton pump. Both contribute to the acidification of the lacunar milieu for the decarboxylation of OCN (36). The undercarboxylated OCN is then released into circulation. Murine models have demonstrated that FoxO1 and ATF4 induce Esp and Opg expression (37).

Bone-specific insulin resistance and OCN. Experiments conducted in murine models by the research group led by Karsenty et al. at Columbia University (NY) have revealed the existence of a specific bone insulin resistance, demonstrating its implications in systemic insulin resistance (38). Insulin's action in OB, by inhibiting osteoprotegerin (OPG), contributes to OC differentiation, thereby inducing bone resorption. OC, through the acidification of their internal milieu, facilitate the decarboxylation of OCN, thereby promoting insulin sensitivity in other organs like white adipose tissue (WAT), muscle, and hepatocytes. When transgenic mice were subjected to a high-fat diet (HFD), genetic evidence supported the concept that insulin resistance in bone (OB) can impact whole-body glucose homeostasis. Hence, bone's insulin resistance might reduce OCN secretion, crucial for optimal insulin action in target tissues, although its relationship with concurrent modifications in insulin secretion is not completely understood (39).

Recent experiments involving disruption of the OCN gene in murine models, resulting in altered glucose homeostasis, further emphasize the role of bone in regulating whole-body glucose homeostasis. Furthermore, these studies have highlighted a decrease in insulin receptor activity in OB from mice treated with an HFD. This decrease occurs through ubiquitination mechanism involving the SMURF1 protein, which appears to be specific to these bone cells (39).

It is now widely acknowledged that OCN, a hormone secreted by bones, regulates whole-body glucose homeostasis, influences energy expenditure, male fertility, and even brain development. OCN undergoes post-translational decarboxylation at sites 17, 21, and 24, resulting in the secretion of Glu-OCN, also known as undercarboxylated OCN. Deletion of the OCN gene in mice (Ocn(-/-)) induces hyperglycemia (glucose intolerance), hypoinsulinemia, impaired insulin secretion, and peripheral insulin resistance (40). An unexpected aspect of OCN is its ability to increase visceral fat while reducing serum adiponectin levels. Its role in insulin resistance was further elucidated by the heterozygous deletion of Ocn and AdipoQ, maintaining mice in an insulin-resistant state. Besides its effect on insulin sensitivity, OCN affects glucose metabolism by activating AKT and facilitating GLUT4 translocation to the membrane. OCN also stimulates IL-6 production, which, in turn, increases the expression of RANKL, suppressing osteoprotegerin in OB. Consequently, two significant feed-forward loops were identified involving bone, adipocytes (including adiponectin), and IL-6.

Interesting experiments conducted by Wei et al. 2014 (38) involving the deletion of one allele of InsR (Col1a1-Insr(+/-)) in OB mice exacerbated HFD-induced skeletal insulin resistance. Notably, hyperactivation of mTORC1 leads to the stimulation of S6K1, resulting in serine phosphorylation and subsequent degradation of IRS-1, desensitizing insulin signalling (42). These experiments, among others, indicate that mTORC1 serves as a nutritional integrator of bone development and glucose metabolism (for further details, refer to Ref. 40). Pioneering studies by Ferron et al. 2012 (43) demonstrated that administration of OCN to wild-type mice improved indeed insulin resistance and glucose tolerance. In mice, one regulatory pathway involves Esp gene. In humans, Esp is a pseudogene, and there is no human ortholog. However, it has been postulated that PTP1B might fulfill a similar role by dephosphorylating the IR (44).

At the clinical level, circulating OCN levels are inversely correlated with BMI, fat mass, fasting glucose, HbA1c, and serum adiponectin (40). The association with glucose metabolism appears to be age and sex-dependent. Further studies by the same research group suggested a coordinated regulation between bone mass, whole-body growth, energy metabolism, and reproduction, implicating leptin (45). Leptin, known for its ability to suppress appetite and increase energy expenditure, binds to specific receptors (LEPR) highly expressed in the ventromedial (VMH) neurons and the arcuate nucleus of the hypothalamus (45). Upon activation, LEPR stimulates the JAK2 tyrosine kinase intracellularly. There is now genetic evidence supporting the role of leptin in regulating bone mass via central mechanisms, specifically by enhancing serotonin synthesis, which subsequently acts to regulate bone mass (39).

Estrogen receptors and their role in insulin resistance. Menopause is marked by a sudden decline in estrogen levels, leading to increased osteoclastic activity and bone resorption. The insufficiency of sex steroids, combined with factors like age, contributes to the development of insulin resistance in postmenopausal women. The emergence of insulin resistance later in life promotes the onset of MetS, notably as observed in studies involving "surgical" menopause. Estrogens generally have beneficial effects on insulin resistance and impact various peripheral tissues such as the liver, muscles, cardiac tissues, and endothelium (46). For instance, estrogen treatment in FoxO1 knockout (KO) ovarectomized female mice significantly improves glucose tolerance (47). In adipocytes, estrogen helps prevent the accumulation of visceral abdominal fat. Several other factors also contribute to the development of insulin resistance in postmenopausal women, including significant changes in body composition, age at menarche, and relative hyperandrogenism or follicle-stimulating hormone (FSH) secretion.

At the cellular level, estrogens exert their effects through several mechanisms, including genomic, non-genomic, and mitochondrial pathways (48). Estrogen's impact depends on the availability of receptors on cells and the bioavailability of estrogen not bound to sex hormone-binding globulin (SHBG). Studies have demonstrated, for instance, that a common single nucleotide polymorphism (SNP) (rs6259) slows the plasma clearance of SHBG and is inversely associated with T2D (49). Estrogen's effects are mediated by two types of receptors: ERα and ERβ, and the newly described G protein-coupled E2 receptor 1 (GPER). ERα receptors belong to the nuclear receptor superfamily and are widely expressed in various reproductive and non-reproductive tissues, including muscles, liver, heart, and bone. Conversely, ERβ is predominantly expressed in the ovaries and prostate gland. Recent investigations have indicated that estrogen receptors are expressed in both OB and OC. ERα serves as the principal receptor exerting an osteoprotective effect. Complete deletion of ERα is associated with decreased bone turnover and increased trabecular volume. In these mice, disruption of the feedback loop results in elevated circulating E2 levels, which cannot be restored by estrogen administration (50).

At the cellular level, estrogen receptors act through two mechanisms: a nuclear pathway and a membrane pathway (48). The nuclear effect involves the presence of estrogen responsive elements (ERE) and interactions with AP-1 or SP-1. The membrane pathway, known as membrane-initiated steroid signaling (MISS), also exists. There is also a ligand-independent pathway activated by EGF and IGF-1. Murine models lacking ERα-AF1 cofactors have shown an effect on trabecular bone, while ERα-AF-2 impacts both trabecular and cortical bone. Loss of ERα receptor function due to a point mutation at the palmitoylation site suggests that ERα is effective in OB, explaining the estrogen effect on trabecular bone. The expression of a mutant (E207A/G208A) ERα explains some nonclassical ERα signaling and effects on energy balance. Selective estrogen receptor modulators (SERMs) bind to ER receptors and exhibit estrogen's protective effects. In bone, estrogen modulates the OPG/RANKL system and increases the production of growth factors (IGF-1 and TGF-β) (51).

Lessons from genetic syndrome of insulin resistance. Insight into the relationship between systemic insulin resistance and bone metabolism can be gleaned from the study of various genetic syndromes exhibiting extreme insulin resistance. Classical examples of such syndromes include Type A syndrome, Rabson–Mendenhall syndrome, and leprechaunism, all caused by deleterious mutations in the insulin receptor and typically manifesting Acanthosis Nigricans. Generally, Type A syndrome presents in young girls with accelerated growth, indicating normal bone development and metabolism (52). Although glucose homeostasis is impaired, the diabetes often tends to be mild. Pioneering studies have demonstrated that administering IGF1 at a dosage of 100 microg/kg (twice daily) for six weeks reduces serum insulin levels by 60-80% and improves glucose tolerance (53). The mechanism of action of IGF-1 in these patients is not entirely understood, but the simplest hypothesis suggests a "spillover" mechanism: in the presence of high insulin levels, this hormone would stimulate IGF-1R. For instance, in one patient studied in Montpellier (France), we observed an 80% decrease in insulin binding to circulating erythrocytes (due to a mutation in the α-subunit of IR). Intriguingly, the binding of IGF-1 remained entirely normal (54). Unfortunately, these rare patients were not extensively studied for bone physiology, and data remain limited. A recent study by Kushchayeva et al. (2020) provided valuable insights into 27 patients with severe insulin resistance due to mutations in the insulin receptors (55). The study involved examinations via DXA, kidney ultrasonography, and blood level analysis of PTH, alkaline phosphatase, OCN, and (25-OH) vitamin D. The findings revealed that patients exhibited low BMD and high renal calcium deposition, suggesting that in the absence of insulin signaling, there is an increase in bone resorption. This study supports the model wherein insulin action in bone is essential to prevent bone resorption.

A markedly different perspective on the relationship between bone health and insulin resistance is presented by the study of Congenital Generalized Lipodystrophy (CGL) known as Berardinelli–Seip syndrome. Several forms of CGL exist, primarily caused by four genes: AGPAT2, BSCL2, CAV1, and CAVIN1 (with most cases attributed to the first two genes). Clinical features include the complete absence of subcutaneous fat, Acanthosis Nigricans, muscle hypertrophy, and often hyperandrogenism with PCOS developing after puberty. Research by Lima et al. (56) reported that these patients exhibit notably high BMD despite delayed menarche, low physical activity, and decreased levels of vitamin D. The elevated BMD primarily affects trabecular bone rather than cortical compartments. Most patients demonstrated normal Trabecular Bone Score (TBS), indicating normal bone microarchitecture. Investigations into the bone marrow of these patients revealed a lack of bone marrow adipose tissue (BMAT), resulting in decreased levels of adipokines. Although it remains a topic of debate, it's important to note that CGL2 patients (attributed to mutations in seipin gene) exhibited various bone and dental anomalies, including gingivitis, bone cysts, advanced bone age, diffuse osteosclerosis, elevated serum sclerostin levels, and high bone mineral density in trabecular sites (57).

An additional perspective on the relationship between bone health and insulin resistance can be gained from the study of Alström syndrome, a rare condition characterized by blindness, neuronal deafness, and diabetes mellitus. This syndrome often presents with Acanthosis Nigricans and severe insulin resistance. ALMS1 is the gene responsible for this syndrome (58). Interestingly, despite insulin resistance, these patients tend to exhibit elevated BMD, particularly in the lumbar spine (T scores > 2). Furthermore, among a series of 23 patients, one individual displayed bone marrow obliteration associated with generalized high bone density (59). As for the lipoatrophic diabetes in Alström syndrome, the mechanisms remain unknown. The simplest proposed hypothesis suggests that increased insulin levels could lead to heightened bone mass due to reduced bone turnover, indicating a more substantial reduction in bone resorption compared to bone formation. It's important to note that Alström syndrome, along with Bardet-Biedl syndrome, falls under the category of ciliopathies. ALMS1 protein plays a specific role in intraciliary transport and cell migration, leaving open the possibility that the mutated gene might affect the normal physiology of OC. In conclusion, genetic syndromes characterized by severe insulin resistance offer valuable models to comprehend the pathophysiology of moderate insulin resistance in postmenopausal osteoporosis, as they might provide insights into previously unknown and unexpected mechanisms.

Insulin resistance in OP of postmenopausal woman – relation with measured HOMA index. Studying MetS in postmenopausal women with osteoporosis is of significant importance due to its acknowledged association with a fivefold increase in the risk of Type 2 Diabetes (T2D) and a twofold increase in cardiovascular diseases within the population. Despite these findings, the connection between MetS, particularly insulin resistance, has produced conflicting and unexpected results. Previous investigations exploring the link between MetS and OP in postmenopausal women have generated inconsistent prevalence outcomes. While some studies suggest a higher prevalence of MetS among women with OP, others indicate a lower prevalence, implying a potential protective effect (60). This discrepancy has led to extensive reviews, yet consensus remains elusive regarding the relationship between MetS and OP (4, 5, 60).

Intuitively, MetS might be perceived to worsen conditions in the elderly due to its potential contribution to cardiovascular complications. Therefore, the observed lower or unchanged prevalence of MetS in OP appears somewhat unexpected. There are likely numerous factors accounting for these discrepancies in the literature. An overview of MetS prevalence rates in postmenopausal women with OP indicates significant variations among studies. For instance, in the Rancho Bernardo Study (61), the reported prevalence was 18.2%. In European populations, Muka et al. (2015) reported a prevalence of 45.7% among 1527 women over 55 years in the Rotterdam study (62). In Asian populations, there are also considerable variations. Chen et al. (2018) reported 35.2% prevalence in Eastern China (63). The highest prevalence was observed in Mexicans, at 57.2%, rising to 66% in postmenopausal women (64). While variations in ethnic groups may account for some differences, these previous results suggest potential biases in patient recruitment, among other confounding factors such as differing MetS definitions.

Addressing this issue requires a shift towards examining measured insulin resistance rather than solely focusing on MetS prevalence. Hence, our review emphasizes recent studies that quantified insulin resistance using the HOMA-IR index in Caucasian populations (65-83) and Asian populations (84-97) as indicated in Table 1 and Table 2. Examination of all these studies indicated that HOMA index for insulin resistance is in great majority (80% of studies) associated with higher BMD in non-diabetic patients, but the same correlation is reported by only 28% of studies that also included diabetic patients. The correlation with OCN levels is in general negative in both non-diabetic and diabetic patients (positive correlation was found in only 2 studies that included diabetic patients also). The mean value of BMI index in Asian populations was 24.2 kg/m², value lower than in European populations (mean 28.9 kg/m²). It appears that one major problem in analysis of correlation of systemic insulin resistance with BMD or some other bone turnover markers is the presence of obesity. Indeed, in a great number of studies investigators usually use the BMD adjusted for BMI, under the assumption that there would be a linear correlation between BMI and BMD. This aspects neeeds special attention since the correlation between BMD and obesity and MetS is quite complex (98). In obese women, potential factors such as mechanical loading or hormonal changes (e.g., estrogen production by adipose tissue) might contribute to higher BMD suggesting a protective role. This favorable effect is attributed to the mechanical stress on bones, as increased weight leads to a positive bone balance. While findings regarding this relationship remain somewhat inconsistent, it is widely recognized that adipose tissue and muscle play crucial roles in maintaining the bone mass. Studies over the past decade have emphasized the impact of adipose tissue on bone health, primarily through the endocrine secretion of leptin and adiponectin. Leptin, an adipose-derived hormone, plays a significant role as it can directly enhance the formation of new bone cells (osteoblastogenesis). Additionally, there's a decrease in adiponectin secretion, an increase in pro-inflammatory cytokines like TNF-α and IL-6, as well as altered actions of various factors such as PPAR-gamma (peroxisome proliferator-activated receptor γ), osteopontin and parathyroid hormone (PTH), all of which can impact bone metabolism (99). Of note, several studies indicated that the relation with BMD is accomplished through the proportion of fat mass rather than BMI. Besides mechanical loading, the estrogen levels significantly influence BMD in postmenopausal women. Adipose tissue functions is a substantial source of aromatase, an enzyme crucial for synthesizing estrogens. Postmenopausal women and individuals with obesity often exhibit higher concentrations of estrogens compared to those with a leaner phenotype. Under the name of "obesity paradox," it is highlighted that despite the positive correlation between obesity and BMD, obesity per se doesn't seem to offer protection against fragility fractures in postmenopausal women (100). The relationship between obesity and fragility fractures shows specific effects depending on the anatomical sites. Obese individuals have a reduced risk of hip, vertebral, and wrist fractures, but an increased risk of ankle and humerus fractures. This can be explained by the dynamics of falling, wherein obese individuals are more prone to falling backward, and at the same time might be shielded by the hip fat pad (100). Undoubtedly, the alteration in body composition during menopause significantly contributes to insulin resistance. As women progress through the menopausal phase, there is a noticeable rise in subcutaneous and visceral fat, coupled with a decrease in gluteal fat and lean body mass, a change often strongly correlated with FSH. VAT particularly exhibits a heightened tendency for lipolysis, resulting in an augmented release of pro-inflammatory cytokines, which collectively might contribute to insulin resistance (100, 101). Thus, understanding the pathogenesis of insulin resistance in postmenopausal women implies complex investigations, particularly related to the type of obesity, inflammatory markers, estrogen levels and some other factors. It should be noted that genetic factors may contribute significantly to the intricate relationship between obesity and bone health. For instance, the FTO (fat mass and obesity-associated) gene was demonstrated as being key player in pathogenesis of obesity. Studies involving the deletion of FTO gene in mice have revealed that its absence results in OB death, leading to subsequent bone loss (102). This suggests that individuals harboring mutations in the FTO gene might be predisposed to developing osteoporosis. The relation of FTO gene and OP remains however poorly understood and several mechanisms are possible.

Table 1.

Studies of insulin resistance in postmenopausal women with osteoporosis (n = 20) from Caucasian populations

Study	Patients (Ethnicity)	BMI (kg/m2)	HOMA	Bone parameter	Major findings and specificity
Dennison et al (2004)	444 PMW (UK)	26.8 ±1.2	3.9 ±1.9	LS, FN, TH aBMD, DXA	HOMA positively correlated with LS, FN and TH BMD – NS after adjustment for BMI. Specificity: study also included 465 M; Diabetics included; Not adj. for BMI
Zoico et al (2008)	36 PMW (Italy)	27.68±4.80	2.32±1.28	FN, TH, DXA	Positive associations of insulin, HOMA, DHEAS and BMD (FN, TH, whole body, whole body/height) – NS after adjustment for total body fat mass. Specificity: whole body aBMD and whole body aBMD/height; Non-diabetics; Not adj. for BMI
Jürimäe et al (2008)	88 PMW (Estonia)	27.4 ± 3.6	1.80 ± 1.34	LS, FN aBMD, DXA	HOMA positively associated with whole body bone mineral mass, total body and FN BMD (also with insulin). Specificity: whole-body bone mineral mass, total body; Non-diabetics; Not adj. for BMI
Agbaht et al (2009)	84 PMW (Turkey)	29.4 (25.9-33.8)	1.82 (1.17–2.86)	LS, FN, TH, DXA	HOMA not correlated with BMD. Specificity: forearm aBMD; Non-diabetics; Adj. for BMI
Oros et al (2012)	83 PMW (Romania)	24.10±1.03 31.62±0.87	1.51± 0.54 2.73± 1.07	OCN, CTX, LS, TH aBMD, DXA	In PMW with OP and MetS crosslaps correlated positively with HOMA-IR and OCN with insulinemia. When comparing PMW with OP or MetS, there was higher risk for OP due to higher IR. Diabetics included; Not adj. for BMI
Lu et al (2012)	64 trios (Finland)	28.1 (27.0 –29.2)	2.26 (1.93–2.58)	OCN (cOCN & ucOCN)	OCN not correlated with glucose, insulin and HOMA. Trios included daughter- mother- maternal grandmother; Specificity: Total (tOCN), carboxylated (cOCN), and uncarboxylated (ucOCN) OCN. Non-diabetics; Not adj. for BMI
Srikanthan et al (2014)	254 PMW (USA)	30.0 ± 6.65	1.36 (1.11)	LS, TH, DXA	HOMA associated negatively with composite indices of FN strength, LS and FN BMD. Fasting insulin negatively associated lower bone strength. Specificity: African Americans and European ancestry; study included 717 subjects, of which 51.6% preMW & PMW; Diabetics included; Adj. for BMI
Shanbhogue et al (2016)	146 PMW (USA)	27.5±5.1	2.1±1.5	LS, FN, TH, aBMD, DXA	HOMA associated negatively with bone size and positively with vBMD and bone microarchitecture and TH and LS BMD; NS after adjustement. Specificity: Distal radius & tibia bone geometry, vBMD, microarchitecture; Non-diabetics; Adj. for BMI
Lipovetzki et al (2017)	115 PMW (Israel)	26.4 ± 4.3 26.5 ± 3.9	7.7 ± 10.4 2.3 ± 2.5	OPG (hip fracture)	OPG and HOMA significantly higher in PMW with hip fracture; NS after adjustment. Specificity: 2 groups with and without hip fractures; Diabetics included; Adj. for BMI
Lambrinoudaki et al (2017)	335 PMW (Greece)	26.4 ±3.9	1.47 ± 1.69	Vertebral fracture	Prevalent VFs had lower levels of HOMA compared to women without fractures. Non-diabetics; Not adj. for BMI
Bonneau et al (2017)	129 PMW (Canada)	32.5 ± 4.6	3.8 ± 1.6	uOCN total OCN	Total OCN correlate negatively with fasting glucose, insulin & HOMA. OCN γ-carboxylation on residue Glu17 positively associated with IR and glucose intolerance, and negatively with insulin sensitivity. Non-diabetics; Not adj. for BMI
Mashavi et al. (2017)	114 PMW (Israel)	25.5 ±3.5 27.8± 3.4	1.8 ± 1.6 9.2 ± 12.2	OPG; LS, TH aBMD, DXA	OPG positively associated with HOMA. Diabetics included; Adj. for BMI
Massera et al. (2018)	1,455 PMW (USA)	25.2±4.6 30.0±4.9	2.4±1.12 2.8±2.1	OCN, CTX	HOMA negatively associated with OCN. CTX negatively associated with incident diabetes. Specificity: Caucasian and African-American; Diabetics included; Adj. for BMI
Mesinovic et al. (2019)	43 PMW (Australia)	32.6 (28.4 -39.2)	1.7 (1.0 - 6.0)	Whole body aBMD, DXA, pQCT	HOMA positively correlated with proximal radius periosteal and endosteal circumference (NS after adj) and negatively with proximal radius cortical vBMD; NS after exclusion of DM; Specificity: Radius and tibia trabecular vBMD, radius and tibia cortical vBMD, periosteal/endosteal circumference, stress-strain index (pQCT); Interquartile range; Diabetics included; Not adj. for BMI
Mihai et al. (2019)	61 PMW (Romania)	22.6 (21.2–23.9) 28.8 (26.6–30.1)	3.2 (0.96-5.4) 5.5 (2.0-8.9)	LS, FN aBMD, DXA	HOMA inversely related to LS aBMD. Diabetics included; Adj. for BMI
Campillo-Sánchez et al. (2020)	381 PMW (Spain)	26 ± 4	3.3 ± 4.6	sBMD, TBS, 3D-DXA, QTC	HOMA positively associated with vBMD (cortical and trabecular); NS after adj. TBS correlated negatively with HbA1c, insulin, and HOMA. Specificity: trabecular and cortical vBMD, bone mineral content (BMC), bone mineral volume, cortical thickness (Cth), BMD of cortical surface; Non-diabetics; Adj. for BMI
Shieh et al. (2022)	693 PMW (USA)	NA	2.31	LS, FN aBMD, DXA	When it decreases, IR is associated with BMD preservation; when it increases, IR is associated with BMD loss. Specificity: African American, Chinese, Japanese, Caucasian ancestry; study also included 861 preMW and 571 menopausal transition women; Diabetes status not mentioned; Adj. for BMI
Elmas et al. (2023)	120 PMW (Turkey)	31.7 ± 4.80 33.7 ± 6.4	2.9 (0.4:20.2) 3.7 (0.4:27.7)	LS, FN aBMD, DXA	HOMA not correlated with BMD. Diabetics included; Adj. for BMI not mentioned
Dimitrova et al. (2023)	84 PMW (Bulgaria)	26.5 ± 4.6 27.1 ± 4.0 27.6 ± 3.3	2.33 ± 1.79 1.91 ± 0.92 2.66 ± 1.48	LS; proximal femur aBMD DXA, FRAX	HOMA (Insulin) positively corelated with proximal femur aBMD. No correlations of HOMA (Insulin) with fracture risk in whole population. At HOMA < 2, insulin levels associated with lower 10-year risk of MOF. At HOMA >2, insulin correlated with 10-year risk of MOF. Specificity: 3 groups with OP, osteopenia and normal BMD; Non-diabetics; Adj. for BMI
Sheu et al. (2023)	322 PMW (Australia)	23.0 ± 1.6 30.0 ±4.1	1.1 (0.9–1.5) 3.6 (2.9–4.9)	LS, FN, TH aBMD, DXA	CTX and OCN negatively correlated with insulin. TH BMD positively correlated with insulin and HOMA. Low bone turnover markers were unique to T2D and IR, rather than obesity. Specificity: measured CTX, P1NP, OCN; Diabetics included; Adj. for BMI

PMW, postmenopausal women; preMW, premenopausal women; M, men; OCN, serum osteocalcin; CTX, serum c-terminal telopeptide of Type I collagen; P1NP, serum N-terminal procollagen of Type I collagen; OPG, serum osteoprotegerin; LS, lumbar spine; FN, femoral neck; TH, total hip; aBMD, areal bone mineral density; vBMD, volumetric bone mineral density; TBS, trabecular bone score; DXA, dual X-ray absorptiometry; QCT, quantitative computed tomography; HOMA, Homeostasis Model Assessment; DM, diabetes mellitus; FNW, femoral neck width; Data were expressed as means±SD, ± SEM or median (interquartile range).

Table 2.

Studies of insulin resistance in postmenopausal women with osteoporosis (n = 14) from populations of Asian ancestry

Study	Patients (Ethnicity)	BMI (kg/m2)	HOMA	Bone parameter	Major findings and specificity
Im et al (2008)	339 PMW (Korean)	24.1±3.1 25.8±2.6	1.3±0.7 2.7±1.8	LS aBMD, DXA	OCN negatively correlated with fasting glucose, insulin, HbA1c, and HOMA; OCN significantly lower in T2D than controls. Specificity: measured OCN, CTX; Diabetics included; Not adj. for BMI;
Zhou et al (2009)	180 PMW (Chinese)	25.3±3.4	5.4 (3.7-7.4)	OCN	OCN not correlated with HOMA. In PMW, OCN correlates inversely with age and HbA1c. Specificity: 180 PMW; Diabetics included; Not adj. for BMI;
Lee et al (2012)	214 PMW (Korean)	23.3±2.8	0.84±0.82	OCN	OCN negatively correlated with fasting insulin and HOMA-IR. Non-diabetics; Not adj. for BMI
Kim et al. (2013)	3,279 PMW (Korean)	24.16±0.0	2.65±0.06	LS, FN, TH, DXA	In PMW BMD not correlated with HOMA; in addition, trochanter and intertrochanter aBMD (DXA); Specificity: 3 279 PMW; Diabetics included; Non-adj. for BMI;
Kim et al (2013)	316 PMW (Korean)	24.3± 2.9	NA	OCN	In PMW, OCN negatively correlated with fasting insulin and HOMA- IR; Specificity: there are 300 preMW; Diabetics included; Adj. for BMI
Jung et al (2015)	279 PMW (Korean)	24.4±3.4	3.6±2.6	OCN, CTX	In women ≥50 y, CTX not associated with HbA1c, HOMA and HOMA-β. Specificity: 279 W ≥50 y; Diabetics included; Not adj. for BMI
Zheng et al (2015)	744 PMW (Chinese)	22.8± 3.9 24.2 ±3.8	1.33(1.0-1.8) 1.6 (1.2-1.8)	LS and FN aBMD, DXA	Osteoporosis risk more pronounced in participants with higher levels of HOMA. Specificity: measured OCN, CTX; Non-diabetics; Adj. for BMI
Wang et al (2017)	4171 PMW (Chinese)	24.9±3.4	2.0 (1.4-3.0)	N-term OC, PINP, β-CTX	N-terminal OCN negatively correlated with HOMA and positively correlated with GUTT-ISI and HOMA-β; P1NP positively correlated with HOMA-β, GUTT-ISI, Stumvoll 1st phase and 2nd phase insulin secretion index; β-CTX positively correlated with HOMA-β and GUTT-ISI; Specificity: 919 Males, Diabetics included; Adj. for BMI;
Kalimeri et al (2018)	96 PMW (Chinese Singapore)	22.9 ± 2.6	1.36±0.71	CTX; LS, FN, TH aBMD, DXA, QCT	LS aBMD inversely associated with IR. Composite Strength Indices (CSI) inversely associated with IR; after adjusting for fat mass and age, association remained valid for impact strength index. CSI were lower in participants with high degree of IR; PTH, CTx-1, 25(OH) Vitamin D not associated with HOMA; Investigate also FN width (FNW), hip axis length (HAL); Proximal femur and L3 vBMD (QCT); Non-diabetics; Not adj. for BMI;
Yang et al (2018)	1008 PMW (Korean)	23.6	1.74 (1.3-2.2)	LS, total proximal femur,	HOMA associated negatively with cortical bone volume at proximal femur and bone strength indices at FN and positively with cortical vBMD. Also investigated FN, (inter) trochanter vBMD (QCT); FN axial compression indices; Non-diabetics; Adj. for BMI;
Li et al (2018)	91 PMW (Han Chinese)	25.4±3.7	2.0±0.8	LS, FN, TH aBMD, DXA	HOMA associated positively with bone marrow fat fraction and negatively with LS, FN and TH BMD. FN BMD association lost after adjustment. Additional Marrow fat fraction (MRI); Non-diabetics; Adj for BMI.
Seoung et al (2018)	137 PMW (Korean)	23.8 ± 2.9	0.9 ± 1.1	LS, FN aBMD DXA	HOMA correlated positively with LS BMD. Measured total and undercarboxylated OCN. Non-diabetics; Adj. for BMI
Yang et al (2019)	892 PMW (Chinese)	22.8 (21.4-24.8) 23.6 (21.3-25.8)	1.60 (1.1-2.2) 1.85 (1.3-2.4)	LS, FN aBMD DXA	LS and FN BMD positively associated with HOMA. 1st- degree family history of DM associated with increased BMD, IR, and hyperinsulinemia. Non-diabetics; Not adj. for BMI.
Ye et al (2023)	437 PMW (Chinese)	21.3 (20.0-22.6) 22.8 (21.3-24.6)	0.57 (0.5-0.6) 2.2 (1.8-2.7)	FN aBMD DXA	Greater IR associated with increased BMD; Non-diabetics, Adj. for BMI.

PMW, postmenopausal women; preMW, premenopausal women; M, men; OCN, serum osteocalcin; CTX, serum c-terminal telopeptide of Type I collagen; P1NP, serum N-terminal procollagen of Type I collagen; OPG, serum osteoprotegerin; LS, lumbar spine; FN, femoral neck; TH, total hip; aBMD, areal bone mineral density; vBMD, volumetric bone mineral density; TBS, trabecular bone score; DXA, dual X-ray absorptiometry; QCT, quantitative computed tomography; HOMA, Homeostasis Model Assessment; DM, diabetes mellitus; FNW, femoral neck width; Data were expressed as means±SD, ± SEM or median (interquartile range).

At the clinical level, the analysis of MetS and insulin resistance in OP rises several methodological problems. Usually, HOMA index for insulin resistance or BMD are statistically adjusted for BMI, assuming a linear correlation. However, this statistical strategy might overlook a non-linear correlation or even a decrease in BMD with higher values of BMI values. For instance, a recent study involving healthy Caucasian women (aged 18 to 67 years) observed that BMD was proportionate to the amount of fat mass (up to 33%). Beyond this threshold, the correlation became negative (103). As mentioned earlier, the relationship between insulin resistance, particularly hyperinsulinemia, and bone health is notably intricate. Although in the day-to-day medical practice the most frequent osteoporosis type is primary osteoporosis, one should also consider secondary causes of OP in postmenopausal women (104).

In vitro studies have recognized that insulin exerts an osteoanabolic effect on bone mass. However, conflicting results emerge when assessing insulin resistance through the HOMA index. Some studies indicate a negative association with the HOMA index. The presence of T2D in postmenopausal women holds significant importance because hyperinsulinemia itself may impact bone health before the onset of T2D (105). In the later stages of T2D progression, additional mechanisms come into play, such as the effects of chronic hyperglycemia leading to the accumulation of advanced glycation end products (AGEs) or oxidative stress (72, 105).

In conclusion, the bone is a very active endocrine organ, which may regulate glucose and energy homeostasis at the systemic level. This role of bone tissue has an important impact in understanding the pathogenesis of insulin resistance in postmenopausal women, among other numerous endocrine factors, including degree and type of obesity and body composition. Screening of recent publications concerning the relation between measured insulin resistance and osteoporosis suggest that insulin resistance and metabolic syndrome evaluation should enter in the routine panoply of investigations in postmenopausal women, paying special attention to type 2 diabetes mellitus and its complications.

Conflict of interest

The authors declare that they have no conflict of interest.