Introduction

Type 2 diabetes (T2D) and obesity (OB) are metabolic diseases characterized by systemic chronic inflammation [1]. In particular, serum levels of pro-inflammatory cytokines IL-6 and TNF-α are higher in T2D and OB [2,3]. This chronic inflammatory state contributes to tissue inflammation in insulin target tissues, and it is associated to long-term complications, including bone fragility [4]. Subjects with T2D or OB have an increased risk for fractures despite normal or higher bone mineral density (BMD), suggesting a bone quality impairment. Furthermore, subjects with elevated pro-inflammatory cytokines have been linked to an increased risk of fracture [5–7]. However, data on the direct effect of inflammation on bone fragility in subjects with T2D or OB are lacking. In vitro studies showed that IL-6 is mainly involved in promoting osteoclastogenesis [8]. Moreover, IL-6 contributed to bone loss in vitro, affecting bone marrow mesenchymal stem cells (BMCs) differentiation and cellular senescence in a model of high fat diet-induced obesity [9]. On the other hand, blocking TNF-α in bone-lining forming cells increased cell number and promoted bone formation in a model of T2D [10]. Consistently, Sun M. et al. demonstrated that subjects with T2D and fractures had higher TNF-α levels [3]. Previous data showed that TNF-α promoted bone erosion and inhibited bone formation by increasing the main inhibitors of Wnt canonical pathway [11,12]. Wnt pathway is one of the main regulators of bone metabolism. We previously showed that canonical Wnt pathway is downregulated in T2D, and it is associated with reduced bone formation markers [13] and reduced bone strength [14]. Thus, we hypothesized that increased inflammation may affect bone quality by Wnt canonical pathway in bone of postmenopausal women with T2D or obesity. Results confirmed that subjects with T2D and OB have increased bone inflammation associated with Wnt beta/catenin pathway downregulation and reduced bone strength.

Results

Subject characteristics

Clinical characteristics of study subjects are reported in Table 1. Age and Menopausal age were not different between groups while BMI was higher in T2D and OB compared to controls (T2D: 29.98±5.77 kg/m²; OB: 32.84±2.94 kg/m²; CTRL: 23.11±2.24 kg/m², p<0.0001). As expected, fasting glucose was significantly higher in T2D compared to either OB or controls [T2D: 121.8±20.59 mg/dl; OB: 101.7±10.88 mg/dl; CTRL: 94.98±8.45 mg/dl, p<0.0001). T2D subjects had a mean HbA1c of 6.87±0.79% and a mean disease duration of 13.75±9.54 years. Diabetes medications included monotherapy with metformin (n=15) and combination therapy with metformin plus insulin and glinide (n=3). There were no differences in serum calcium, eGFR (CKD-EPI equation) and serum blood urea nitrogen, as well as in cholesterol values and triglycerides.

Table 1.

Adipose tissue composition and bone mass

All the parameters related to adipose tissue composition and bone mass are reported in Table 2. Trunk total mass was significantly higher in subjects with obesity (p<0.0001) and T2D (p=0.001) compared to controls. Also, trunk fat mass was higher in OB (p=0.003) and T2D (p=0.016) compared to controls. Consistently, percent trunk fat was higher in subjects with obesity (p=0.001) and T2D (p=0.030) as compared to control subjects. Analysis of fat distribution highlighted that OB, but not T2D, had both a higher percentage of android fat (p<0.0001) and of gynoid fat (p=0.022) as compared to controls. Finally, levels of VAT mass and VAT volume were significantly greater in subjects with obesity (p<0.0001, respectively) and T2D (p=0.0002, respectively) compared to controls.

Table 2.

Total body bone mineral density (BMD) and T-score were not different between groups. Lumbar bone mineral density was higher in OB compared to controls (p=0.004) but not in T2D compared to controls. Lumbar T-score was also increased only in OB compared to controls (p=0.003). Femoral BMD was increased not only in OB (p=0.0005) but also in T2D (p= 0.016) compared to controls. In line, femoral T-score was higher in OB (p<0.0001) and T2D (p=0.006) compared to controls.

Bone microarchitecture and strength

Microarchitecture parameters by μCT on trabecular core samples are reported in Table 3. T2D, OB and control subjects had similar trabecular indices, including bone volume density (BV/TV), connectivity density, trabecular thickness, trabecular number, trabecular separation, tissue mineral density and volumetric BMD. Biomechanic analysis (table 3) showed a reduced Young’s modulus in T2D compared to OB (p=0.0395) but not to control subjects, and a trend of reduction of Yield strength in T2D compared to OB (p=0.0552).

Table 3.

Circulating and bone inflammation

Circulating TNF-α levels were significantly higher in T2D (p=0.0004), but they were not higher in OB (p=0.1108) compared to control subjects (Fig. 1A). IL-6 levels were higher in both T2D (p=0.0451) and OB (p=0.0002) compared to controls (Fig. 1B). Adiponectin levels were lower only in T2D compared to OB (p=0.0402), but not in the other groups (Fig. 1C). Of note, circulating TNF-α and IL-6 were strongly positively correlated with VAT mass (r=0.3651, p=0.0310 and r=0.7139, p<0.0001, respectively), as reported in Figure 1-Supplementary figure 1.

Figure 1.

Analysis of gene expression in the femoral head after surgery showed increased TNF-α mRNA levels either in T2D compared to controls (p<0.0001) or in T2D compared to OB (p=0.0161) (Fig. 2A). However, gene expression levels of IL-6 were not different between groups (Fig. 2B), as for IL-8 mRNA levels (Fig. 2C) IL-10 gene expression levels were lower in both T2D (p=0.0285), and OB (p=0.0324) compared to controls (Fig. 2D). Adiponectin (ADIPOQ) mRNA levels were lower in subjects with T2D (p=0.0041) compared to controls (Fig. 2E). On the other hand, SFRP5 mRNA levels were only higher in T2D (p=0.0084) compared to control subjects (Fig. 2F).

Figure 2.

Wnt signaling pathway in bone

The canonical Wnt inhibitor SOST mRNA levels were higher in T2D (p<0.0001) and in OB (p<0.0001) compared to controls (Fig.3A). Interestingly, SOST mRNA levels were positively correlated with fasting blood glucose levels in OB (Figure 3-Supplementary figure 1). In line, gene expression of the canonical ligand WNT10B was lower in T2D (p=0.0071) and in OB (p=0.0196) vs controls (Fig. 3B). Also, gene expression of LEF-1, the Wnt/β-catenin transcription factor, was lower in T2D (p<0.0001) and in OB (p=0.0479) compared to controls (Fig. 3C). Gene expression of non-canonical WNT5A was higher only in T2D (p=0.0025) but it was not different in OB compared to controls (Fig. 3D). Finally, Gene expression of GSK3β was higher either in T2D compared to OB (p=0.0069) or T2D vs controls (p=0.0003) (Fig. 3E).

Figure 3.

Correlation analysis of inflammatory and Wnt target genes

Correlation analysis of inflammatory and Wnt target genes are reported in figure 4 (A-I). TNF-α mRNA levels were positively correlated with SOST (r=0.5121, p=0.0002), WNT5A (r=0.3227, p=0.0396) and GSK3β (r=0.3789, p=0.0146) mRNA levels (Fig. 4A-C). Conversely, TNF-α was negatively correlated with WNT10B (r=-0.3844, p=0.0188) and LEF-1 (r=-0.3310, p=0.0322) mRNA levels (Fig. 4D,E). IL-10 was only negatively correlated with SOST (r=-0.3100, p=0.0457) (Fig. 4F). ADIPOQ mRNA levels were negatively correlated with SOST (r=-0.3864, p=0.0105) and with WNT5A mRNA levels (r=-0.3025, p=0.0515) (Fig. 4G,H). Moreover, SFRP5 mRNA levels negatively correlated with LEF-1 (r=-0.3991, p=0.0131) mRNA levels (Fig. 4I).

Figure 4.

Correlation analysis of inflammatory cytokines and bone strength

Serum levels of TNF-α (r=-0.3473, p=0.0352) and IL-6 (r=-0.3777, p=0.0302) negatively correlated with Young’s Modulus (Fig. 5A-B). There was no difference between circulating inflammatory markers and the other biomechanical parameters.

Figure 5.

Discussion

In this study we showed increased circulating and bone inflammation in postmenopausal women with T2D and obesity. Specifically, we found increased TNF-α mRNA levels in bone of subjects with T2D while the anti-inflammatory cytokine IL-10 was decreased in both subjects with T2D and obesity. Importantly, increased mRNA levels of inflammatory cytokines were associated with downregulation of Wnt pathway related genes. Moreover, serum TNF-α was increased in T2D compared to the other groups and it was negatively associated with bone strength.

The increased TNF-α serum levels in subjects with T2D observed in our study are consistent with the study of Sum et al. [3] that reported increased TNF-α and other chemokines in T2D patients with bone fracture. There are evidence showing that inflammation may affect bone health. In vitro studies showed that TNF-α is involved in the regulation of bone resorption in inflammatory conditions either directly or by increasing RANKL and M-CSF expressions by osteoblasts, stromal cells and even osteocytes [15]. Moreover, blocking TNF-α in bone-lining forming cells increased cell number and enhanced bone formation in a model of T2D [10]. In line, we found that both serum and bone TNF-α were increased in T2D, but not in subjects with obesity. Interestingly, serum TNF-α was negatively correlated with Young’s Modulus, suggesting that increased TNF-α reduces bone strength in subjects with T2D.

Although we confirmed that circulating IL-6 was higher in subject with obesity and it was positively correlated with VAT, we found no difference of IL-6 gene expression levels in bone tissue of either T2D or subjects with obesity. There is only evidence that IL-6 contributed to bone loss in vitro, affecting BMCs differentiation and cellular senescence in a model of high fat diet-induced obesity [9]. However, there is no data supporting the effect of inflammation in bone of subjects with obesity. In line with IL-6, gene expression of pro-inflammatory IL-8 was not altered either in bone of subjects with T2D or obesity. There is no evidence about IL-8 role in bone of either T2D or obesity [8].

On the other hand, we found that the anti-inflammatory IL-10 mRNA levels were decreased in subjects with T2D and obesity. Previous data showed that serum levels of IL-10 were substantially lower in osteoporosis patients [16] and in T2D [17] than in healthy individuals, although to our knowledge, the role of IL-10 on bone of T2D or obesity has not been studied. As expected, both serum and mRNA bone adiponectin levels were lower in T2D [18,19]. In vitro studies indicated that adiponectin promoted osteoblastogenesis, whilst simultaneously inhibiting osteoclastogenesis [20–22] and adiponectin knockout mice have increased bone mass and osteoblast numbers [23].

Moreover, serum adiponectin in postmenopausal women was found inversely associated with bone mass [24]. However, there are no data about adiponectin in bone of subjects with T2D or obesity. We also showed that subjects with T2D had increased SFRP5 bone mRNA levels compared to controls. Data reported that SFRP5, a negative regulator of Wnt canonical signaling pathway, was overexpressed in osteoporosis and was also associated to downregulated bone formation markers [25]. Our data are in line with these findings, supporting that T2D is characterized by low bone formation and downregulated Wnt canonical signaling.

Previous studies showed that inflammation may downregulate Wnt signaling pathway [26], the main regulator of bone metabolism. Consistent with our previous findings [14], we observed that the expression of the Wnt inhibitor SOST was not only higher in bone of T2D but also in subjects with obesity compared to controls. Even though sclerostin is increased during weight loss [27], there are other data showing that serum sclerostin was negatively associated to insulin sensitivity in women with obesity [28]. Conversely, our data showed that SOST mRNA levels were positively correlated with fasting blood glucose in subjects with obesity suggesting that glucose impairment may be a determinant for higher sclerostin levels. Consistent with increased SOST mRNA levels and in line with our previous findings on Wnt signaling regulation in bone of T2D [14], the expression of the Wnt ligand WNT10B and the transcription factor LEF-1 were downregulated in both subjects with T2D and obesity. Also, in vitro studies showed that TNF-α induced the phosphorylation of GSK3β, the deactivated form of GSK3β, which increased nuclear β-catenin and Lef-1 resulting in downregulated Wnt signaling and reduced bone formation [26].Our data confirmed that GSK3βB is increased in subjects with T2D. This is in line with our previous findings in which GSK3β was associated with reduced bone formation and strength in T2D [14].

Furthermore, it is shown that Wnt5a promotes inflammation and insulin resistance in adipose tissue of human and animal models [29,30] and high circulating levels of Wnt5a are detected in subjects with T2D [31] and obesity [32]. Other studies showed that Wnt5a inhibits Wnt3a protein by downregulating beta/catenin-induced reporter gene expression [33]. In line, we observed that gene expression of WNT5A was increased in bone of T2D. This study has some limitations. It is a cross-sectional design with a relatively small number of enrolled subjects.

The main strength of the study is that we investigated for the first time the effect of inflammation on human bone samples measuring several pro and anti-inflammatory genes. Moreover, we found the association between inflammation and bone strength by the regulation of Wnt signaling pathway in subjects with T2D and obesity compared to a control population.

In conclusion, our study showed that there is increased inflammation in bone of subjects with T2D and obesity and it is associated with the downregulation of Wnt signaling and reduced bone strength. These results shed light on the pathophysiology of bone impairment in T2D and obesity and may help understanding the mechanisms of bone fragility in metabolic diseases.

Material and methods

Study subjects

We enrolled for this study 63 postmenopausal women (19 with T2D, 17 with obesity and 27 non-diabetic normal weight controls) undergoing hip replacement surgery for osteoarthritis, consecutively screened to participate in this study between 2018 and 2022. Diabetes status was confirmed by the diabetologist. Participants were diagnosed with type 2 diabetes when they had fasting plasma glucose (FPG) ≥126 mg/dl or 2-h plasma glucose (2-h PG) ≥200 mg/dl during a 75-g oral glucose tolerance test (OGTT); or hemoglobin A1c (HbA1c) ≥6.5% in accordance with the American Diabetes Association diagnostic criteria [34]. Obesity was confirmed based on the BMI ≥30 kg/m², in accordance with the World Health Organization diagnostic criteria [35]. Eligible participants were ≥60 years of age. Exclusion criteria were any diseases affecting bone, any concomitant inflammatory disease or malignancy. Additionally, individuals treated with medications affecting bone metabolism such as estrogen, raloxifene, tamoxifen, bisphosphonates, teriparatide, denosumab, thiazolidinediones, glucocorticoids, anabolic steroids, and phenytoin, and those with hypercalcemia or hypocalcemia, hepatic or renal disorder, hypercortisolism, current alcohol or tobacco use were excluded. The study was approved by the Ethics Committee of Campus Bio-Medico University of Rome (Prot..42/14 PT_ComEt CBM) and all participants provided written informed consent. All procedures were conducted in accordance with the Declaration of Helsinki.

Body Composition/Anthropometric measurements

Body scans were conducted with Lunar ProdigyTM (GE Healthcare©, USA) DXA scanner. Body composition parameters were analyzed using en-CORETM software (version 17, 2016), and VAT was measured with the CoreScanTM (GE Healthcare©, USA). Quality control and calibration of the model were performed each day, following the instruction protocol provided by the manufacturer. For this analysis, the variables of interest derived from the DXA dataset are BMD total body (g/cm²), T score total body, BMD Lumbar L1+L4 (g/cm²), T score femoral, trunk fat mass (gr), trunk total mass (gr), and defined anatomical regions (android, and gynoid % fat) and visceral adipose tissue (VAT mass, g and VAT volume, cm³). The body mass index (BMI) was calculated as weight/(height)² (kg/m²).

Processing of blood Samples

Venous whole blood was collected the day before the surgery in all participants after completing a fasting period of at least 8 hours, using tubes containing clot activator (BD vacutainer®, NJ, USA). Samples were kept at room temperature to allow the blood to clot, and they were centrifuged at 1500 g for 10 minutes for serum isolation. Serum was then transferred to a clean tube and stored at −80°C for further analysis.

Processing of bone samples

Femoral head specimens were obtained during hip replacement surgery. Briefly, trabecular bone specimens were collected fresh and washed multiple times in sterile PBS until the supernatant was clear of blood, as previously described [13]. Trabecular samples were then aliquoted and stored at −80°C for further analysis.

ELISA and ELLA Microfluidic Immunoassay

Adiponectin was quantified in serum using ELISA immunoassay (R&D system, MN, USA) according to the manufacturer instructions. TNF-α and IL-6 were measured in serum by using Simple Plex assays run on the ELLA microfluidic immunoassay system (ProteinSimple, San Jose, CA). 25 μl of serum was diluted at a 1:1 ratio with sample diluent, and 50 μl the solution was added to each sample inlet on the ELLA cartridge, according to manufacturer’s instruction. Sample results were reported using Simple Plex Runner v.3.7.2.0 (ProteinSimple).

RNA extraction and gene expression by RT-PCR

Total RNA from trabecular bone samples was extracted using TRIzol (Invitrogen) following the manufacturer’s instructions. The concentration and purity of the extracted RNA were assessed spectrophotometrically (TECAN, InfiniteM200PRO), and only samples with 260/280 absorbance ratio between 1.8 and 2 were used for reverse transcription using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Carlsbad, CA) according to the manufacturer’s recommendations. Transcription products were amplified using TaqMan real-time PCR (Applied Biosystems, Carlsbad, CA) and a standard protocol (95°C for 10 minutes; 40 cycles of 95°C for 15 seconds and 60°C for 1 minute; followed by 95°C for 15 seconds, 60°C for 15 seconds, and 95°C for 15 seconds). β-Actin expression was used as an internal control (housekeeping gene). Relative expression levels of Tumor necrosis factor alpha (TNF-α), Interleukin 6 (IL-6), Adiponectin (ADIPOQ), Interleukin 8 (IL-8), Interleukin 10 (IL-10), secreted frizzled-related protein 5 (SFRP5), Sclerostin (SOST), Wnt ligands (WNT5A and WNT10B), T-cell factor/ lymphoid enhancer factor 1 (LEF-1), and glycogen synthase kinase 3 beta (GSK3B), were calculated using the 2-ΔCt method.

Micro-computed tomography (μCT) and bone compression tests

We used cylindrical bone specimens of trabecular core (with a diameter of 10 mm and a length of 20 mm) from 19 T2D with obesity, 17 OB non-diabetic and 18 control subjects to measure bone micro-architecture and bone mechanical parameters (Young’s modulus, ultimate strength and yield strength), as previously described [13].

Statistical analysis

Data were analyzed using GraphPad Prism 9.0 (GraphPad Software, San Diego, CA). Patients’ characteristics were described using means and standard deviations (SD) or medians and 25th-75th percentiles, as appropriate, and percentages. Group data are presented in boxplots with median and interquartile range (IQR); whiskers represent maximum and minimum values. We assessed data for normality and Mann-Whitney test was used to compare variables between groups. Data were analyzed using nonparametric Spearman correlation analysis and the correlation coefficients (r) were used to assess the relationship between variables. We used Grubbs’ Test to assess and exclude outliers.

Acknowledgements

This work was supported by an internal Grant of Campus Bio-Medico University of Rome. We wish to thank the structure & strength core at Washington University Musculoskeletal Research Center.

Supplementary figure legends

Figure 1-Supplementary figure 1.

Figure 3-Supplementary figure 1.