Measurement of bioactive osteocalcin in humans using a novel immunoassay reveals association with glucose metabolism and β-cell function

Abstract

Osteocalcin (OCN) is a bone-derived hormone involved in the regulation of glucose metabolism. In serum, OCN exists in carboxylated and uncarboxylated forms (ucOCN), and studies in rodents suggest that ucOCN is the bioactive form of this hormone. Whether this is also the case in humans is unclear, because a reliable assay to measure ucOCN is not available. Here, we established and validated a new immunoassay (ELISA) measuring human ucOCN and used it to determine the level of bioactive OCN in two cohorts of overweight or obese subjects, with or without type 2 diabetes (T2D). The ELISA could specifically detect ucOCN concentrations ranging from 0.037 to 1.8 ng/mL. In a first cohort of overweight or obese postmenopausal women without diabetes (n = 132), ucOCN correlated negatively with fasting glucose (r = -0.18, P = 0.042) and insulin resistance assessed by the homeostatic model assessment of insulin resistance (r = -0.18, P = 0.038) and positively with insulin sensitivity assessed by a hyperinsulinemic-euglycemic clamp (r = 0.18, P = 0.043) or insulin sensitivity index derived from an oral glucose tolerance test (r = 0.26, P = 0.003). In a second cohort of subjects with severe obesity (n = 16), ucOCN was found to be lower in subjects with T2D compared with those without T2D (2.76 ± 0.38 versus 4.52 ± 0.06 ng/mL, P = 0.009) and to negatively correlate with fasting glucose (r = -0.50, P = 0.046) and glycated hemoglobin (r = -0.57, P = 0.021). Moreover, the subjects with ucOCN levels below 3 ng/mL had a reduced insulin secretion rate during a hyperglycemic clamp (P = 0.03). In conclusion, ucOCN measured with this novel and specific assay is inversely associated with insulin resistance and β-cell dysfunction in humans.

Keywords: ELISA; carboxylation; diabetes; obesity; osteocalcin; uncarboxylated osteocalcin.

Fig. 1.

Generation of monoclonal antibodies recognizing human uncarboxylated osteocalcin (ucOCN). A: schematic representation of the different forms of OCN used in this study. B: results of the screening strategy used to identify clones specific to the ucOCN epitope or to the OCN C-terminal region. C: recombinant uncarboxylated (ucOCN_1–49) and synthetic partially carboxylated (2XGla OCN_1–49) or fully carboxylated (cOCN_1–49) human osteocalcins were spotted on nitrocellulose membranes and detected by Western blot using anti-ucOCN, anti-C-terminal OCN, and anti-γ-carboxyglutamic acid [Gla (Pan-Gla)] antibodies. Glu, glutamate.

Fig. 2.

Establishment and validation of the human uncarboxylated osteocalcin (ucOCN) ELISA. A: identification of 5 antibody pairs (capture/detection) to detect ucOCN. The results were normalized to the highest signal (100) from the 8H4/4B6 pair. B: schematic representation of the sandwich ELISA described and tested in this study. Human ucOCN is measured using the 8H4 clone as the capture antibody and biotinylated 4B6 as the detection antibody in combination with horseradish peroxidase (HRP)-conjugated avidin. C: measurement of ucOCN, fully carboxylated OCN (cOCN), and partially carboxylated OCN (2XGla OCN) showing that the ucOCN ELISA specifically recognizes the uncarboxylated form of OCN from 0.037 to 1.8 ng/mL. D: ucOCN detection in media from human embryonic kidney 293 cells overexpressing human OCN-V5 cultured with vitamin K₁ (VK₁; 20 μM) or vehicle. E: ucOCN measurements in serum from individuals treated with warfarin (n = 6) or untreated controls (Ctrl; n = 15). **P < 0.01; ***P < 0.001.

Fig. 3.

Association between uncarboxylated osteocalcin (ucOCN) and glucose tolerance in overweight or obese postmenopausal women without diabetes (Cohort A). A: ucOCN concentration within each quartile (Q1–Q4) of the cohort. B and C: glucose excursion during oral glucose tolerance test for Q1 and Q4 (B) and the corresponding area under the curve (AUC; C). *P < 0.05.

Fig. 4.

Association between uncarboxylated osteocalcin (ucOCN) and insulin resistance and diabetes in severely obese subjects with and without type 2 diabetes (T2D; Cohort B). A and B: circulating level of total OCN (tOCN; A) and ucOCN (B) in the nondiabetic (ND) and T2D subpopulations of the cohort. C–E: Pearson’s correlation between ucOCN and glucose (C), glycated hemoglobin (HbA1c; D), and the insulin disposition index (E). **P < 0.01.

Fig. 5.

Association between uncarboxylated osteocalcin (ucOCN) and insulin secretion rate and β-cell function in severely obese subjects with and without type 2 diabetes (Cohort B). A: the population was divided in 2 groups (low and high ucOCN), based on the average circulating ucOCN concentration (3.31 ng/mL). B and C: plasma C-peptide concentration (B) and insulin secretion rate (ISR; C) during stepwise intravenous glucose perfusion in subjects with low and high circulating ucOCN. D: β-cell sensitivity to glucose (based on the relationship between plasma glucose levels and ISR during stepwise intravenous glucose perfusion) in subjects with low and high circulating ucOCN. *P < 0.05.