https://orcid.org/0000-0002-7121-5899
Abstract
Irisin is a hormonelike molecule that is cleaved and secreted by an unknown protease from fibronectin type III domain-containing protein 5 (FNDC5). It ameliorates bone status and muscle atrophy and influences energy homeostasis. PTH exerts several metabolic effects that may interact with the effects of irisin.
To test the hypothesis that irisin and PTH mutually affect their biological action, we evaluated FNDC5 mRNA and protein expression in myotubes treated with PTH (1–34) and parathyroid hormone receptor (PTH-r) mRNA expression in osteoblasts treated with r-irisin. To confirm the in vivo impact of PTH on irisin, we compared irisin serum concentrations in postmenopausal women with primary hyperparathyroidism (PHPT) and control subjects.
C2C12 myotubes were treated with short-term and continuous 10−10 M teriparatide and MC3T3-E1 osteoblasts with 100 ng/mL r-irisin for 8 hours. In a cross-sectional open-label trial, we enrolled 26 postmenopausal women with PHPT and 31 age-/body mass index (BMI)‒matched control subjects without impairment of calcium/phosphate metabolism.
Teriparatide treatment on myotubes significantly downregulated FNDC5 expression by acting through its own receptor, which in turn activated Erk11/2 phosphorylation. r-Irisin led to a 50% downregulation of PTH-r mRNA expression compared with untreated cells (P < 0.001). Irisin was significantly lower in the PHPT group than in age-/BMI-matched controls (4.5 ± 1.1 vs 12 ± 5.2 µg/mL; P < 0.001). No significant correlation between irisin and bone mineral density or PTH was recorded in the PHPT group.
Preclinical findings suggest the existence of an interplay between PTH and irisin metabolism that seems to be confirmed by the significant reduction of irisin concentration in postmenopausal women with PHPT.
Irisin is a hormonelike molecule that is cleaved and secreted by an unknown protease from fibronectin type III domain-containing protein 5 (FNDC5); it is expressed mainly in the skeletal muscle, heart, adipose tissue, and liver (1). It is produced during physical activity in humans and mice (1). This molecule was initially studied for its ability to lead to trans-differentiation of white into brown adipose tissue (1). In this process, white adipocytes take the features of brown adipocytes, increasing energy expenditure and regulating thermogenesis (1). In 2015, it was discovered that irisin acted not only on adipose tissue but also on bones and muscle (2). Indeed, a dose 35 times lower than that which causes browning acted on bone tissue, increasing bone mineral density (BMD), periosteal circumference, and polar moment of inertia in the long bones of healthy mice (2), pointing out the powerful effect on the cortical bone compartment. It has also been shown that irisin ameliorates bone status and muscle atrophy in hind-limb suspended mice, a mouse model used to study disuse-induced osteoporosis (3). Few cross-sectional studies have investigated the relationship between irisin, bone, and muscle in humans. In particular, it has been documented that irisin serum concentration has a positive correlation with long bone BMD (4, 5) and bone strength (4). These findings suggest that irisin not only has a systemic effect on the bone area most subjected to mechanical stress (5), but it also supposes to mainly act on cortical bone. Clinically, other studies also found a negative correlation between irisin and fragility fractures (6, 7), independent of BMD and body composition (7). Most available evidence also shows that irisin significantly influences glucose and energy homeostasis. Indeed, higher irisin concentrations have been inversely correlated with insulin resistance, obesity, and type 2 diabetes and have been associated with bone health in children with type 1 diabetes (8–10).
PTH exerts several metabolic effects that appear to oppose those of irisin. In particular, subjects with chronic pathologic elevation of PTH experience the catabolic effect of this hormone, resulting in a reduction in BMD, particularly at the cortical site (11).
It has also been suggested that PTH may negatively affect insulin sensitivity and glucose tolerance through reduced expression of insulin receptor substrate 1 and glucose transporter 4 (12). Furthermore, PTH can influence body weight and composition. Indeed, subjects with primary hyperparathyroidism (PHPT) undergo weight gain and body composition changes that have not been clearly elucidated (13, 14). No previous studies have investigated the potential direct relationship between irisin and PTH.
To test the hypothesis that irisin and PTH mutually affect their biological action, we evaluated FNDC5 mRNA and protein expression in myotubes treated with PTH (1–34) and parathyroid hormone receptor (PTH-r) mRNA expression in osteoblasts treated with recombinant irisin. To confirm the in vivo impact of PTH on irisin, we evaluated irisin serum concentrations in postmenopausal women with PHPT compared with concentrations in age-, sex-, and body mass index (BMI)‒matched control subjects with no impairment of calcium/phosphate metabolism.
Material and Methods
Cell cultures
Mouse myoblast C2C12 and preosteoblast MC3T3-E1 cells were used for in vitro experiments in this study. Both cell lines were plated at 10 × 103 cells/cm2 and cultured in Minimum Essential Medium Eagle - Alpha Modification (α-mem) (Gibco; Thermo-Fisher, Waltham, MA) with 10% fetal bovine serum (Gibco; Thermo-Fisher) until they reached confluence in a humidified atmosphere (37°C, 5% CO2) (Hera cell 150; Themo-Fisher). Upon confluence, we induced C2C12 differentiation by replacing the growth medium with the differentiating medium α-mem with 2% fetal bovine serum (Gibco; Thermo-Fisher) to stimulate myotube formation. The cells were grown in differentiating medium for 10 days. To induce differentiation and mineralization of the preosteoblast MC3T3-E1 cells, we cultured them with α-mem medium supplemented with 50 µg/mL ascorbic acid and 10−2 M β-glycerophosphate for 10 days. We changed both growth and differentiating media every 2 days.
Teriparatide treatment
For teriparatide treatment in vitro, we used the Forsteo® Pen (Lilly, Indianapolis, IN) containing 20 μg/80 μL (0.607 × 10−4 M) of PTH (1‒34; corrected for acetate, chloride, and water content), 0.41 mg of glacial acetic acid, 0.1 mg of sodium acetate (anhydrous), 45.4 mg of mannitol, 3 mg of meta-Cresol, and water for injection. In addition, hydrochloric acid solution 10% and/or sodium hydroxide solution 10% may have been added to adjust the product to a pH of 4. We prepared teriparatide working solution at 10−6 M by diluting 1.64 μL of Forsteo Pen solution in 98.36 μL of ddH2O. C2C12 myotubes were then treated with teriparatide working solution 10−6 M to obtain final concentrations of teriparatide in culture ranging from 10−10 M to 10−8 M for 3 and 8 hours. For each time point group, we included the corresponding control treated with vehicle (ddH2O). To evaluate the role of Erk activation in Fndc5 mRNA downregulation, myotubes were pretreated with 50 μM of PD98059 (Sigma Aldrich, St. Louis, MO), a specific Erk1/2 inhibitor, 20 minutes before the addition of teriparatide. At the end of the time point, cells were lysates and were subjected to RNA extraction. To analyze the effects of continuous teriparatide exposure, we treated myotubes with 10−10 M of teriparatide. We refreshed the medium with teriparatide every 48 hours. In respective control cultures, the medium was also changed every 48 hours. We repeated the cycle three times. At the end of three cycles, cells were lysates and were subjected to RNA extraction, as previously reported (15, 16).
For knocking down the PTH-r, C2C12 myoblasts were transfected with PTH-r small interfering RNA (siRNA) or vector-only duplexes (50 nM) for 24 hours using Lipofectamine RNAiMAX Reagent (Invitrogen) and then treated with 10−10 M of teriparatide for 3 hours. PTH-r siRNA (Santa Cruz Biotechnology, Santa Cruz, CA) is a pool of three target-specific 19- to 25-nt siRNAs designed to knock down gene expression of PTH-r. Each vial contains 3.3 nmol of lyophilized siRNA to make a 10-µM solution in RNase-free ddH2O according to manufacturer’s instructions. To monitor PTH-r knockdown, we performed a preliminary time course of transfection for 16, 20, 24, and 28 hours. The complete PTH-r knockdown was obtained at 24 hours after transfection.
Forskolin treatment
C2C12 myotubes were treated with 100 μM of forskolin (Santa Cruz Biotechnology), a potent activator of cAMP for 3 and 8 hours. For each time point group, we included the corresponding control treated with vehicle (dimethyl sulfoxide). At the end of the time point, cells were lysates and were subjected to RNA extraction.
Irisin treatment
MC3T3-E1 cells differentiated for 10 days in differentiating medium were treated with 100 ng/mL of irisin (AdipoGen, Liestal, Switzerland) for 8 hours. Then, irisin-treated osteoblasts and untreated controls were lysates and were subjected to RNA extraction.
Real-time PCR
We extracted RNA from cells with the RNeasy Mini Kit (Qiagen, Hilden, Germany) using spin columns following the manufacturer’s instructions. We did reverse transcription by iScript Reverse Transcription Supermix (Bio-Rad, Hercules, CA). As a thermal cycler, we used MyCycler (Bio-Rad, Hercules, CA)) according to manufacturer’s instructions. Real-time PCR was performed using the SsoFast EvaGreen Supermix (Bio-Rad) on a CFX96 real-time system (Bio-Rad) for 40 cycles (denaturation, 95°C for 5 seconds; annealing/extension, 60°C for 10 seconds) after an initial 30-second step for enzyme activation at 95°C. We used Primer-BLAST to draw primers (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). We chose Gapdh as a housekeeping gene because it was stably expressed in all samples. Primer sequences were as follows: Gapdh (S-acaccagtagactccacgaca, AS-acggcaaattcaacggcacag); Fndc5 (S-ctcgttgtccttgatgata, AS-attgttgtggtcctcttc); and PTH-r (S-agtccatgccagtgtccag; AS-ttccagggattttttgttgc). We calculated relative gene expression using the ΔΔCT method with expression as a fold-change compared with control.
Western blotting
Western blotting was performed on myotube extracts treated with or without 10−10 M teriparatide for 8 and 24 hours to detect FNDC5; 30 μg of cell proteins were subjected to SDS-polyacrylamide gel electrophoresis. Subsequently, proteins were transferred to nitrocellulose membranes (Hybond, Amersham Pharmacia, London, UK). The blots were probed using rabbit polyclonal anti-FNDC5 (Abcam, Cambridge, UK) and IRDye-labeled secondary antibodies (680/800CW) (LI-COR Biosciences, Lincoln, NE). For immunoblotting, the Odyssey IR imaging system was used (LI-COR).
Study population
All subjects were enrolled from the outpatient clinic at the Endocrinology and Diabetes Unit of Campus Bio-Medico University from January to June 2018. We enrolled 26 consecutive postmenopausal women with PHPT and 31 age- and BMI-matched control subjects with adequate calcium, phosphate, magnesium, and PTH levels.
Diagnosis of PHPT was established by the presence of hypercalcemia and concomitant high PTH levels on at least two separate occasions. Control subjects were postmenopausal women participating in a screening visit for osteoporosis who had not been taking calcium or vitamin D supplementations for at least 6 months before enrollment. Women were considered postmenopausal if they had not had a menstrual period for more than 1 year. No subjects regularly participated in physical activity, and study subjects must not have changed their lifestyle (physical activity and diet) for at least 3 months prior to enrollment. Subjects were asked not to do any physical activity the day before the blood sample.
Subjects were excluded if they met any one of the following criteria: male; age younger than 18 years; pregnancy; renal failure with glomerular filtration rate <30 mL/min; any other prior parathyroid disease; any impairment of glucose metabolism such as diabetes or prediabetes; any pharmacological treatment of osteoporosis within the past 2 years; glucocorticoid use within the past 2 years; and a history of Cushing syndrome, uncontrolled thyroid disease, malabsorption syndrome, significant liver disease, multiple endocrine neoplasm, hyperparathyroidism-jaw tumor syndrome, history of skeletal malignancies (primary or metastatic), or prior radiation therapy involving the skeleton.
Blood sampling and bone densitometry evaluation
Fasting blood samples were obtained in the morning from 8:00 to 8:30 am. BMD was measured by bone densitometry at the lumbar spine (L1–L4), total hip, femoral neck, and nondominant forearm (ultradistal radius and one-third radius) using a Discovery A dual-energy X-ray absorptiometer (Hologic Inc; Bedford, MA). Bone mineral density was expressed in grams per square centimeter.
Assays
Irisin serum concentrations were detected using a competitive ELISA kit (AdipoGen, Liestal, Switzerland) with intra-assay coefficient of variation ≤6.9%. This ELISA KIT allows a large range of measurement. The lowest level of irisin that can be detected is 1 ng/mL, and the assay range is 0.001 to 5 µg/mL. The ELISA kit includes a polyclonal antibody recognizing the native irisin and recombinant irisin under competition in irisin-coated plates. We followed the manufacturer’s instructions for all analyses. Colorimetric reaction was measured by using a spectrophotometer (Eon; BioTek, Winooski, VT) at the end of the assay.
Intact PTH (iPTH) was measured by an immunochemiluminometric assay using the automatic analyzer Modular E170 (Roche Diagnostics, Indianapolis, IN). Normal serum iPTH levels ranged between 18 and 65 pg/mL. PTH serum concentrations <10 pg/mL were arbitrarily defined as very low PTH levels. Serum calcium was measured by automated techniques, and normal levels ranged between 8.5 and 10.2 mg/dL. Serum calcium was adjusted for albumin by the following formula: Alb-Ca = (0.8 [4.0 − patient’s albumin] + serum calcium) (17). Serum phosphate, magnesium, and creatinine levels as well as thyroid function were also measured by automated techniques. The estimated glomerular filtration rate was calculated using the CKD-EPI equation. The total procollagen type 1 N-terminal propeptide was measured by ELISA (Biomedica, Vienna, Austria), and a bone resorption marker, the C-terminal telopeptide of type I collagen (CTX), was measured by ELISA (IDS, Boldon, UK). Serum levels of CTX were assayed by the β-CrossLaps (ECLIA; β-CrossLaps/Serum; Roche Diagnostics, Basel, Switzerland), which uses two monoclonal antibodies against β‒cross-linked CTX, according to the manufacturer’s protocol. This assay is a sensitive electro-chemiluminescent detection technology and is formatted for the cobas e601 automated analyzer. Broad analytical measurements ranged from 0.01 to 6 ng/mL.
Ethics
The study was conducted in compliance with the Declaration of Helsinki and the International Conference on Harmonization Principles of Good Clinical Practice. The research protocol was approved by the ethical committee at the University Campus Bio-Medico of Rome, and all participants gave informed consent allowing their anonymized information to be used for data analysis.
Statistical analysis and sample size calculation
The characteristics of the study population were reported using descriptive statistics (mean and SD or median and interquartile range as appropriate for continuous variables; proportions for categorical variables) by diagnosis of hyperparathyroidism. Differences between groups were evaluated using ANOVA or the Wilcoxon test, as appropriate, for continuous variables and the χ2 test for categorical variables. In the two groups, the relationship between irisin and BMD or PTH was evaluated using the Pearson coefficient and linear regression models, both crude and adjusted for age and BMI. In these models, BMD was expressed as milligram per centimeter squared to provide more readable results.
According to our previous findings in postmenopausal women (7), assuming a reduction of irisin of at least 25% in the PHPT group, the estimated sample size to obtain an error 1 rate of 5% with 80% power was 17 participants per group. To account for attrition and possible missing data, we enrolled 25 patients per group.
All analyses were performed using R version 3.5.0 (R Foundation for Statistical Computing, Vienna, Austria).
Results
PTH treatment negatively regulated FNDC5 expression in myotubes
We evaluated the effect of teriparatide on C2C12 myotube cells cultured in differentiating medium for 10 days. We cultured myotubes with teriparatide at concentrations ranging from 10−10 M to 10−8 M for 3 and 8 hours. As shown in Fig. 1(a), Fndc5 mRNA expression was decreased by ∼50% in myotubes treated with all doses of teriparatide for 3 hours with respect to untreated controls. Likewise, FNDC5 protein was rapidly downregulated in myotubes upon treatment with the lowest dose of teriparatide (10−10 M) [Fig. 1(b)]. In view of this rapid action occurring within 3 hours, which was confirmed by five separate experiments [Fig. 1(c)], we tested whether teriparatide acted on the PTH-r by activating the MAPKs Erk1 and Erk2 (pErk1/2) (18), using the specific Erk1/2 phosphorylation inhibitor PD98059. Inhibiting Erk1/2 activation with 50 μM of PD98059 pretreatment prevented PTH-induced downregulation of Fndc5 mRNA [Fig. 1(d)], suggesting that the effect of PTH on irisin/FNDC5 in myotubes occurs through an Erk-dependent pathway.
(a) Quantitative PCR (qPCR) for FNDC5 mRNA of myotubes treated with teriparatide (referred to as PTH), at concentrations ranging from 10−10 M to 10−8 M for 3 and 8 hours, and their corresponding controls treated with vehicle (Untr). (b) Western blot analysis of myotube extracts for expression of FNDC5 after stimulation with 10−10 M PTH for 8 and 24 hours and their corresponding controls treated with vehicle (Untr). (c) qPCR for FNDC5 mRNA of myotubes treated for 3 hours with 10−10 M PTH alone. Data are reported as scatter plots, with the individual data expressed as mean ± SD. (d) qPCR for FNDC5 mRNA of myotubes treated for 3 hours with 10−10 M PTH plus 50 μM of PD98059, a specific Erk1/2 inhibitor, applied 20 minutes before the addition of PTH. (e) qPCR for FNDC5 mRNA expression in myotubes treated with 100 μM forskolin for 3 and 8 hours and their corresponding controls treated with dimethyl sulfoxide (Ctr). (f) qPCR of C2C12 myoblasts expressing either the PTH-r siRNA (siPTH-r) or empty vector (siCTR) showing PTH-induced changes in the expression of FNDC5 mRNAs at 3 hours. (g) qPCR for FNDC5 mRNA of myotubes treated with continuous stimulation of teriparatide (referred to as PTHc). (h) qPCR for PTH-r in MC3T3-E1 osteoblasts differentiated for 10 days and stimulated for 8 hours with 100 ng/mL of irisin. Data are representative of three to five independent experiments, and values are expressed as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001.
The action of PTH through its receptor is mediated mainly by activation of cAMP upon a G protein‒mediated cascade of events (19). Likewise, forskolin is a potent activator of adenylate cyclase leading to intracellular cAMP accumulation. Our results showed that forskolin treatment of myotubes also decreased Fndc5 mRNA expression, although its effect was more persistent than that of teriparatide and lasted until 8 hours [Fig. 1(e)].
To further confirm that teriparatide inhibited Fndc5 expression by acting through the PTH-r, we used siRNA to knock down PTH-r expression in C2C12 myoblasts. Quantitative PCR analysis showed that although teriparatide treatment significantly downregulated Fndc5 mRNA expression in vector-treated cells (siCTR), this effect was completely blunted in PTH-r silenced cells (siPTH-r) [Fig. 1(f)].
Moreover, we sought to determine whether this rapid effect of teriparatide on FNDC5 expression was transient or whether chronic exposure to teriparatide was also effective in reducing FNDC5 expression in myotubes. Therefore, in order to mimic primary hyperparathyroidism, we treated cells with continuous teriparatide (PTHc) treatment, refreshing the medium with teriparatide every 48 hours for 6 days; we found that PTHc was equally effective in decreasing Fndc5 mRNA expression relative to untreated controls, as reported in Fig. 1(g).
Irisin treatment downregulated PTH-r mRNA expression in osteoblasts
We tested the effect of irisin on osteoblasts differentiated for 10 days in vitro in osteogenic medium. As shown in Fig. 1(h), stimulation with irisin (100 ng/mL) for 8 hours led to a 50% downregulation of PTH-r mRNA expression compared with untreated cells.
General characteristics of the study population
The mean age of our population was 65 years (SD, 6.8 years); there were no differences between PHPT and control groups in age, BMI, estimated glomerular filtration rate, CTX, and time from menopause. Compared with the control group, participants with PHPT had higher mean serum concentrations of calcium (10.5 mg/dL, SD 0.2 mg vs 9.6 mg/dL, SD 0.3 mg/dL; P = 0.015); PTH (142.3 pg/mL, SD 41.5 pg/mL vs 51.3 pg/mL, SD 15.1 pg/mL; P < 0.001); 25-hydroxy vitamin D (34.6 ng/mL, SD 5.2 ng/mL vs 28 ng/mL, SD 13.7 ng/mL; P = 0.026). Serum concentration of phosphate was lower in the PHPT group than in the controls (3 mg/dl, SD 0.4 vs. 3.8 mg/dl, SD 0.4, p<0.001), whereas calcium × phosphorous product was higher in the control group (30.1, SD 4.1 vs 36.2, SD 4; P < 0.001) (Table 1). Because 81% (n = 21) of PHPT subjects were taking cholecalciferol supplementation, their 25-hydroxy vitamin D serum concentration was significantly higher than that of controls (34.6 ng/mL, SD 5.2 vs 28 ng/mL, SD 13.7 ng/mL; P = 0.026).
General Characteristics of the Population
| Control (N = 31) | PHPT (N = 26) | Overall (N = 57) | P Value | |
|---|---|---|---|---|
| Age, y | 64.2 (5.8) | 65.8 (7.9) | 65 (6.8) | 0.377 |
| BMI, kg/m2 | 26.4 (5) | 27.2 (4.8) | 26.8 (4.8) | 0.498 |
| Time from menopause, y | 13.2 (6.8) | 15.7 (9.1) | 14.4 (8) | 0.249 |
| eGFR, mL/min/m2 | 87.1 (16) | 82.5 (21.4) | 85 (18.6) | 0.361 |
| Serum calcium concentration, mg/dL; reference range: 8.5–10.2 mg/dL | 9.6 (0.3) | 10.5 (0.2) | 10.1 (0.2) | 0.015 |
| Serum phosphate concentration, mg/dL; reference range: 2.8–4.0 mg/dL | 3.8 (0.4) | 3 (0.4) | 3.4 (0.6) | <0.001 |
| Serum PTH concentration, pg/mL; ref range: 18–65 pg/mL | 51.3 (12.1) | 142.3 (41.5) | 92.8 (54.6) | <0.001 |
| 25-hydroxy vitamin D concentration, ng/mL | 28 (13.7) | 34.6 (5.2) | 31 (11.1) | 0.026 |
| Calcium × phosphorus | 36.2 (4) | 30.1 (4.1) | 33.4 (5) | <0.001 |
| Serum CTX concentration, ng/mL | 0.345 (0.224) | 0.436 (0.168) | 0.385 (0.205) | 0.102 |
| Serum P1NP concentration, μ/L | 51.2 (26.3) | 69.6 (24.5) | 59.2 (26.9) | 0.01 |
| Lumbar BMD, g/m2 | 0.902 (0.144) | 0.844 (0.117) | 0.877 (0.135) | 0.111 |
| Total hip BMD, g/m2 | 0.869 (0.091) | 0.784 (0.094) | 0.832 (0.101) | 0.001 |
| Femoral neck BMD, g/m2 | 0.671 (0.074) | 0.625 (0.07) | 0.651 (0.076) | 0.025 |
| Radial BMD, g/m2 | 0.605 (0.046) | 0.57 (0.067) | 0.59 (0.058) | 0.028 |
| Renal lithiasis, (%) | 0 | 19 | 9 | 0.037 |
| Control (N = 31) | PHPT (N = 26) | Overall (N = 57) | P Value | |
|---|---|---|---|---|
| Age, y | 64.2 (5.8) | 65.8 (7.9) | 65 (6.8) | 0.377 |
| BMI, kg/m2 | 26.4 (5) | 27.2 (4.8) | 26.8 (4.8) | 0.498 |
| Time from menopause, y | 13.2 (6.8) | 15.7 (9.1) | 14.4 (8) | 0.249 |
| eGFR, mL/min/m2 | 87.1 (16) | 82.5 (21.4) | 85 (18.6) | 0.361 |
| Serum calcium concentration, mg/dL; reference range: 8.5–10.2 mg/dL | 9.6 (0.3) | 10.5 (0.2) | 10.1 (0.2) | 0.015 |
| Serum phosphate concentration, mg/dL; reference range: 2.8–4.0 mg/dL | 3.8 (0.4) | 3 (0.4) | 3.4 (0.6) | <0.001 |
| Serum PTH concentration, pg/mL; ref range: 18–65 pg/mL | 51.3 (12.1) | 142.3 (41.5) | 92.8 (54.6) | <0.001 |
| 25-hydroxy vitamin D concentration, ng/mL | 28 (13.7) | 34.6 (5.2) | 31 (11.1) | 0.026 |
| Calcium × phosphorus | 36.2 (4) | 30.1 (4.1) | 33.4 (5) | <0.001 |
| Serum CTX concentration, ng/mL | 0.345 (0.224) | 0.436 (0.168) | 0.385 (0.205) | 0.102 |
| Serum P1NP concentration, μ/L | 51.2 (26.3) | 69.6 (24.5) | 59.2 (26.9) | 0.01 |
| Lumbar BMD, g/m2 | 0.902 (0.144) | 0.844 (0.117) | 0.877 (0.135) | 0.111 |
| Total hip BMD, g/m2 | 0.869 (0.091) | 0.784 (0.094) | 0.832 (0.101) | 0.001 |
| Femoral neck BMD, g/m2 | 0.671 (0.074) | 0.625 (0.07) | 0.651 (0.076) | 0.025 |
| Radial BMD, g/m2 | 0.605 (0.046) | 0.57 (0.067) | 0.59 (0.058) | 0.028 |
| Renal lithiasis, (%) | 0 | 19 | 9 | 0.037 |
Data are presented as mean (SD) unless otherwise noted.
Abbreviations: eGFR, estimated glomerular filtration rate; PiNP, procollagen type 1 N-terminal propeptide.
General Characteristics of the Population
| Control (N = 31) | PHPT (N = 26) | Overall (N = 57) | P Value | |
|---|---|---|---|---|
| Age, y | 64.2 (5.8) | 65.8 (7.9) | 65 (6.8) | 0.377 |
| BMI, kg/m2 | 26.4 (5) | 27.2 (4.8) | 26.8 (4.8) | 0.498 |
| Time from menopause, y | 13.2 (6.8) | 15.7 (9.1) | 14.4 (8) | 0.249 |
| eGFR, mL/min/m2 | 87.1 (16) | 82.5 (21.4) | 85 (18.6) | 0.361 |
| Serum calcium concentration, mg/dL; reference range: 8.5–10.2 mg/dL | 9.6 (0.3) | 10.5 (0.2) | 10.1 (0.2) | 0.015 |
| Serum phosphate concentration, mg/dL; reference range: 2.8–4.0 mg/dL | 3.8 (0.4) | 3 (0.4) | 3.4 (0.6) | <0.001 |
| Serum PTH concentration, pg/mL; ref range: 18–65 pg/mL | 51.3 (12.1) | 142.3 (41.5) | 92.8 (54.6) | <0.001 |
| 25-hydroxy vitamin D concentration, ng/mL | 28 (13.7) | 34.6 (5.2) | 31 (11.1) | 0.026 |
| Calcium × phosphorus | 36.2 (4) | 30.1 (4.1) | 33.4 (5) | <0.001 |
| Serum CTX concentration, ng/mL | 0.345 (0.224) | 0.436 (0.168) | 0.385 (0.205) | 0.102 |
| Serum P1NP concentration, μ/L | 51.2 (26.3) | 69.6 (24.5) | 59.2 (26.9) | 0.01 |
| Lumbar BMD, g/m2 | 0.902 (0.144) | 0.844 (0.117) | 0.877 (0.135) | 0.111 |
| Total hip BMD, g/m2 | 0.869 (0.091) | 0.784 (0.094) | 0.832 (0.101) | 0.001 |
| Femoral neck BMD, g/m2 | 0.671 (0.074) | 0.625 (0.07) | 0.651 (0.076) | 0.025 |
| Radial BMD, g/m2 | 0.605 (0.046) | 0.57 (0.067) | 0.59 (0.058) | 0.028 |
| Renal lithiasis, (%) | 0 | 19 | 9 | 0.037 |
| Control (N = 31) | PHPT (N = 26) | Overall (N = 57) | P Value | |
|---|---|---|---|---|
| Age, y | 64.2 (5.8) | 65.8 (7.9) | 65 (6.8) | 0.377 |
| BMI, kg/m2 | 26.4 (5) | 27.2 (4.8) | 26.8 (4.8) | 0.498 |
| Time from menopause, y | 13.2 (6.8) | 15.7 (9.1) | 14.4 (8) | 0.249 |
| eGFR, mL/min/m2 | 87.1 (16) | 82.5 (21.4) | 85 (18.6) | 0.361 |
| Serum calcium concentration, mg/dL; reference range: 8.5–10.2 mg/dL | 9.6 (0.3) | 10.5 (0.2) | 10.1 (0.2) | 0.015 |
| Serum phosphate concentration, mg/dL; reference range: 2.8–4.0 mg/dL | 3.8 (0.4) | 3 (0.4) | 3.4 (0.6) | <0.001 |
| Serum PTH concentration, pg/mL; ref range: 18–65 pg/mL | 51.3 (12.1) | 142.3 (41.5) | 92.8 (54.6) | <0.001 |
| 25-hydroxy vitamin D concentration, ng/mL | 28 (13.7) | 34.6 (5.2) | 31 (11.1) | 0.026 |
| Calcium × phosphorus | 36.2 (4) | 30.1 (4.1) | 33.4 (5) | <0.001 |
| Serum CTX concentration, ng/mL | 0.345 (0.224) | 0.436 (0.168) | 0.385 (0.205) | 0.102 |
| Serum P1NP concentration, μ/L | 51.2 (26.3) | 69.6 (24.5) | 59.2 (26.9) | 0.01 |
| Lumbar BMD, g/m2 | 0.902 (0.144) | 0.844 (0.117) | 0.877 (0.135) | 0.111 |
| Total hip BMD, g/m2 | 0.869 (0.091) | 0.784 (0.094) | 0.832 (0.101) | 0.001 |
| Femoral neck BMD, g/m2 | 0.671 (0.074) | 0.625 (0.07) | 0.651 (0.076) | 0.025 |
| Radial BMD, g/m2 | 0.605 (0.046) | 0.57 (0.067) | 0.59 (0.058) | 0.028 |
| Renal lithiasis, (%) | 0 | 19 | 9 | 0.037 |
Data are presented as mean (SD) unless otherwise noted.
Abbreviations: eGFR, estimated glomerular filtration rate; PiNP, procollagen type 1 N-terminal propeptide.
There was no difference between the two groups in mean lumbar BMD, whereas the BMD was higher in the control group than in the PHPT group for total hip (0.869 g/cm2, SD 0.091 g/cm2 vs 0.784 g/cm2, SD 0.094 g/cm2; P = 0.001); femoral neck (0.671 g/cm2, SD 0.074 g/cm2 vs 0.625 g/cm2, SD 0.07 g/cm2; P = 0.025); and radium (0.605 g/cm2, SD 0.046 g/cm2 vs 0.57 g/cm2, SD 0.067 g/cm2; P = 0.028). Finally, no cases of renal lithiasis were observed in the control group, whereas it was present in 19% of participants in the PHPT group (P = 0.037).
Irisin in subjects with PHPT
Irisin serum concentration was lower in the PHPT group, with a median (interquartile range) concentration of 11.4 µg/mL (SD 6.7 µg/mL) in the control group and 4.2 µg/mL (SD 1.4 µg/mL) in the PHPT group (P < 0.001) [Fig. 2(a)].
(a) Irisin concentration in patients with primary hyperparathyroidism (PHPT) and controls. The interquartile range (IQR) was calculated by subtracting the 25th % value from the 75th % value. Difference between groups was evaluated using chi-square test. ***P < 0.001. (b) Correlation between PTH and serum irisin concentration was evaluated using linear regression models.
Irisin and BMD
Irisin was not associated with radial BMD in either the control group (β, 2.238; P = 0.168) or the PHPT group (β, 2.382; P = 0.860). These results were confirmed after adjustment for age and BMI (controls: β, 1.772; P = 0.336; PHPT: β, −2.624; P = 0.833) (Table 2). Similarly, irisin did not correlate with total femoral and femoral neck BMD in either the control group (β, 1.389; P = 0.658 and β, 3.114; P = 0.316, respectively) or the PHPT group (β,−10.642; P = 0.523 and β, −11.405; P = 0.410, respectively) (Table 2). No association was found between irisin and lumbar BMD in the PHPT group, whereas a negative correlation was evident in the control group (β,−10.573; P = 0.037) (Table 2); however, after adjustment for potential confounders, no association was documented (controls: β, −4.014; P = 0.431; PHPT: β, −16.907; P = 0.425) (Table 2).
Linear Regression Models of the Association Between Irisin and BMD
| β Crude (P Value) | β Adjusted (P Value)a | |
|---|---|---|
| Control group | ||
| Radial BMD | 2.238 (0.168) | 1.772 (0.336) |
| Total femoral BMD | −4.079 (0.225) | 1.389 (0.658) |
| Femoral neck BMD | 0.527 (0.850) | 3.114 (0.316) |
| Lumbar BMD | −10.573 (0.037) | −4.014 (0.431) |
| β Crude (P Value) | β Adjusted (P Value)a | |
|---|---|---|
| Control group | ||
| Radial BMD | 2.238 (0.168) | 1.772 (0.336) |
| Total femoral BMD | −4.079 (0.225) | 1.389 (0.658) |
| Femoral neck BMD | 0.527 (0.850) | 3.114 (0.316) |
| Lumbar BMD | −10.573 (0.037) | −4.014 (0.431) |
| Hyperparathyroidism group | ||
| Hyperparathyroidism group | ||
| Radial BMD | 2.382 (0.860) | −2.624 (0.833) |
| Total femoral BMD | −21.546 (0.227) | −10.642 (0.523) |
| Femoral neck BMD | −13.690 (0.308) | −11.405 (0.410) |
| Lumbar BMD | −31.054 (0.159) | −16.907 (0.425) |
| Hyperparathyroidism group | ||
| Hyperparathyroidism group | ||
| Radial BMD | 2.382 (0.860) | −2.624 (0.833) |
| Total femoral BMD | −21.546 (0.227) | −10.642 (0.523) |
| Femoral neck BMD | −13.690 (0.308) | −11.405 (0.410) |
| Lumbar BMD | −31.054 (0.159) | −16.907 (0.425) |
Models adjusted for age and BMI.
Linear Regression Models of the Association Between Irisin and BMD
| β Crude (P Value) | β Adjusted (P Value)a | |
|---|---|---|
| Control group | ||
| Radial BMD | 2.238 (0.168) | 1.772 (0.336) |
| Total femoral BMD | −4.079 (0.225) | 1.389 (0.658) |
| Femoral neck BMD | 0.527 (0.850) | 3.114 (0.316) |
| Lumbar BMD | −10.573 (0.037) | −4.014 (0.431) |
| β Crude (P Value) | β Adjusted (P Value)a | |
|---|---|---|
| Control group | ||
| Radial BMD | 2.238 (0.168) | 1.772 (0.336) |
| Total femoral BMD | −4.079 (0.225) | 1.389 (0.658) |
| Femoral neck BMD | 0.527 (0.850) | 3.114 (0.316) |
| Lumbar BMD | −10.573 (0.037) | −4.014 (0.431) |
| Hyperparathyroidism group | ||
| Hyperparathyroidism group | ||
| Radial BMD | 2.382 (0.860) | −2.624 (0.833) |
| Total femoral BMD | −21.546 (0.227) | −10.642 (0.523) |
| Femoral neck BMD | −13.690 (0.308) | −11.405 (0.410) |
| Lumbar BMD | −31.054 (0.159) | −16.907 (0.425) |
| Hyperparathyroidism group | ||
| Hyperparathyroidism group | ||
| Radial BMD | 2.382 (0.860) | −2.624 (0.833) |
| Total femoral BMD | −21.546 (0.227) | −10.642 (0.523) |
| Femoral neck BMD | −13.690 (0.308) | −11.405 (0.410) |
| Lumbar BMD | −31.054 (0.159) | −16.907 (0.425) |
Models adjusted for age and BMI.
Irisin and PTH
There was no association between irisin and PTH in either the control group (β,−0.097; P = 0.863) or the PHPT group (β,−8.448; P = 0.277) [Fig. 2(b);Table 3]. These results were confirmed after adjustment for age and BMI (control group: β, −0.038; P = 0.953; PHPT group: β, 0.056; P = 0.992) (Table 3).
Linear Regression Models of the Association Between Irisin and PTH
| β Crude (P value) | β Adjusted (P value)a | |
|---|---|---|
| Control | −0.097 (0.863) | −0.038 (0.953) |
| Hyperparathyroidism | −8.448 (0.277) | 0.056 (0.992) |
| β Crude (P value) | β Adjusted (P value)a | |
|---|---|---|
| Control | −0.097 (0.863) | −0.038 (0.953) |
| Hyperparathyroidism | −8.448 (0.277) | 0.056 (0.992) |
Models adjusted for age and BMI.
Linear Regression Models of the Association Between Irisin and PTH
| β Crude (P value) | β Adjusted (P value)a | |
|---|---|---|
| Control | −0.097 (0.863) | −0.038 (0.953) |
| Hyperparathyroidism | −8.448 (0.277) | 0.056 (0.992) |
| β Crude (P value) | β Adjusted (P value)a | |
|---|---|---|
| Control | −0.097 (0.863) | −0.038 (0.953) |
| Hyperparathyroidism | −8.448 (0.277) | 0.056 (0.992) |
Models adjusted for age and BMI.
Discussion
We have shown a downregulation of FNDC5 in myotubes under treatment with PTH (1–34) and a reduction in PTH-r mRNA expression in osteoblasts after recombinant irisin exposure. Moreover, the cross-talk between irisin and PTH was sustained by the lower irisin concentration in postmenopausal women with PHPT compared with control subjects.
Over the last 6 years, a few major preclinical and clinical findings have indirectly suggested the existence of a biological interaction between PTH and irisin: Both hormones affect bone, muscle, and adipose tissue, although apparently in opposite ways. In particular, subjects with chronic pathologic elevation in PTH levels, such as those with PHPT, experienced a reduction in BMD, particularly at the cortical site, with partial preservation of the cancellous areas, as shown by the reduction in BMD specifically at the 1/3 distal radius (11); this chronic elevation of PTH levels increases the receptor activator of the nuclear factor κB ligand and reduces osteoprotegerin (OPG), resulting in an increase in the receptor activator of the nuclear factor κB ligand/OPG ratio with the activation of osteoclastogenesis (11). In contrast, irisin can directly promote osteoblast differentiation and proliferation in vitro (20, 21), leading to a major increase in bone mass, primarily at the cortical site, as shown in young male mice treated with recombinant irisin at a low cumulative weekly dose of 100 µg kg−1 (2). Furthermore, in male hind-limb suspended mice, treatment with irisin increased OPG and inhibited sclerostin, thus preventing and restoring cortical bone loss due to immobility (3).
The finding that irisin treatment decreased the expression of PTH-r on osteoblasts suggests that this myokine may exert its anabolic effect on bone, not only by stimulating osteoblast formation and function but also by blunting the catabolic action of PTH on these cells.
Our in vitro results suggest that irisin and PTH can inversely affect themselves. Both short-term (3 hours) and long-term (6 days) treatment with PTH negatively regulated FNDC5 mRNA and protein expression in myotubes. PTH affected FNDC5 expression in myotubes by acting on the PTH-r, which in turn activated Erk1/2 phosphorylation, most likely through an increase in intracellular cAMP.
Our findings may support the fascinating hypothesis that a few PTH-induced metabolic changes may also be mediated by the modulation of irisin. Indeed, on one hand, recent studies have shown that PTH and PTH-related peptide are involved in weight and body composition changes (13) in the “browning of adipose tissue” and cachexia (22, 23); on the other hand, irisin is able to promote fat browning (1), and its higher concentrations are inversely correlated with body weight, body fat mass, insulin resistance, subcutaneous fat area, and type 2 diabetes (8, 24). Therefore, irisin might be one of the keys to interpreting the action of PTH on fat, muscle, and bone (25).
We also found a significant reduction in serum irisin concentration in patients with PHPT, which seems to be supported by other recent clinical studies. In particular, a recent clinical evaluation assessing serum irisin levels in relation to vascular calcification in hemodialysis reported a negative correlation between irisin levels and PTH in patients undergoing hemodialysis treatment (26). In 2014, Anastasilakis et al. (6) aimed to evaluate predictors of circulating irisin levels in postmenopausal women with low bone mass and to assess a potential effect of denosumab or teriparatide treatment for 3 months. They demonstrated that irisin levels were inversely correlated with PTH in postmenopausal women with low bone mass (6). In our in vivo study population, there was a negative trend between irisin and PTH levels that was not statistically significant, probably because of the small number of subjects with PHPT enrolled.
Furthermore, we did not record any significant correlation between irisin and BMD in the study population, as previously shown (6, 7). Because the irisin pathway seems to be implicated in cortical bone health, we expected a significantly positive association between irisin levels and radius BMD. However, our data did not confirm this correlation in the PHPT group. However, it must be noted that irisin concentrations in this group showed very poor variability, and regardless of the actual biologic effect, this is therefore unlikely to explain variations in BMD in this sample.
Our study has several limitations. We did not assess body composition and the rate of physical activity, even though subjects were asked to not perform physical activity the day before blood sampling. Also, our study population included only postmenopausal women. Another limitation is that we treated myotubes with the commercial teriparatide Forsteo Pen, an aqueous solution containing excipients and correctors suitable for injections in human subjects. Therefore, although the solution in the pen was diluted ∼104 to 106 times in ddH2O to treat myotubes, we treated the control wells with ddH2O as vehicle but not with the same diluent that was used for commercial teriparatide. Moreover, in our in vitro study, we sought to determine whether PTH, acting through its receptor, directly downregulates FNDC5 expression by knocking down PTH-r in C2C12 myoblasts. However, further studies performed on myotubes completely lacking PTH-r are needed.
In conclusion, our in vitro findings suggest the existence of interplay between PTH and irisin metabolism, a finding that seems to be confirmed by the significant reduction in irisin concentration in postmenopausal women with PHPT. Additional larger and more robust translational studies are needed to confirm our results and better elucidate the cross-talk between the two hormones.
Abbreviations:
- BMD
bone mineral density
- BMI
body mass index
- CTX
C-terminal telopeptide of type I collagen
- FNDC5
fibronectin type III domain-containing protein 5
- OPG
osteoprotegerin
- PHPT
primary hyperparathyroidism
- PTH-r
parathyroid hormone receptor
- si-RNA
small interfereng RNA
- α-mem
Minimum Essential Medium Eagle – Alpha Modification
Acknowledgments
Author Contributions: A.P. and M.G. contributed to the study design; A.P., L.S., G.C., G.T., A.M.N., R.C., G.B., S.C., S.M., N.N., and M.G. contributed to the conduct of the study; C.P., D.L., G.C., M.G., and A.P. contributed to data analysis. A.P., L.S., G.C., G.T., and M.G. interpreted the data. A.P., L.S., G.T., and G.C. drafted the manuscript, and A.P., L.S., G.C., G.T., A.M.N., R.C., G.B., G.M., S.C., S.M., N.N., and M.G. revised the manuscript content. All authors approved the final version of the manuscript. A.P., L.S., G.C., and M.G. take responsibility for the integrity of the data analysis.
Current Affiliation: L. Sanesi's current affiliation is PhD School in “Tissue and Organ Transplantation and Cellular Therapies,” Department of Emergency and Organ Transplantation, School of Medicine-University of Bari, 70124 Bari, Italy.
Disclosure Summary: The authors have nothing to disclose.
References
Author notes
A.P. and L.S. contributed equally to this study.
