Abstract
Purpose
To determine whether dysregulation of circulating concentrations of undercarboxylated osteocalcin (UC-OC) or GLA-carboxylated osteocalcin (GLA-OC) occurs in patients with type 1 diabetes, a condition of insulin deficiency without insulin resistance.
Methods
We measured serum concentrations of UC-OC and GLA-OC in 115 subjects with type 1 diabetes (T1D), ages 14–40 years, and in 55 age-matched healthy control subjects. Relationships between UC-OC and GLA-OC concentrations and patient characteristics (gender, age), indices of glycemic control (HbA1c, fasting plasma glucose, C-peptide concentration, 3-day average glucose measured by a continuous glucose sensor, total daily insulin dose) and circulating indices of skeletal homeostasis [Total calcium, 25-OH vitamin D, parathyroid hormone, IGF-I, type 1 collagen degradation fragments (CTX), adiponectin, leptin] were examined. Between group differences in the concentrations of UC-OC and GLA-OC were the main outcome measures.
Results
Although adiponectin levels were higher in the T1D group, between-group comparisons did not reveal statistically significant differences in concentration of UC-OC, GLA-OC, CTX or leptin between the T1D and control populations. Instead, by multivariate regression modeling, UC-OC was correlated with younger age (p<0.001), higher CTX (p<0.001), lower HbA1c (p=0.013) and higher IGF-I (p=0.086). Moreover, within the T1D subgroup, UC-OC was positively correlated with C-peptide: Glucose ratio (reflecting endogenous insulin secretion); with IGF-I (reflecting intra-portal insulin sufficiency); and with total daily insulin dose.
Conclusions
In T1D, UC-OC appears to correlate positively with markers of insulin exposure, either endogenously produced or exogenously administered.
Keywords: Adiponectin, Leptin, IGF-I, insulin, osteoporosis
INTRODUCTION
Type 1 diabetes (T1D), unlike type 2 diabetes (T2D), is a condition in which insulin sensitivity remains generally intact, but absolute insulin deficiency develops, secondary to progressive, autoimmune-mediated pancreatic β-cell destruction (1). Moreover, a relative IGF-I deficiency co-exists in this disorder (2). Diabetic osteopathy is a significant co-morbidity of T1D (3), characterized by osteoporosis, increased risk for bone fracture, and micro-architectural changes that increase the brittleness of bone (4). Chronic hyperglycemia, hypoinsulinemia, disruption of the growth hormone (GH): IGF-I: IGFBP system and altered vitamin/mineral homeostasis are all thought to contribute to this skeletal pathology (3, 5). However, the specific interplay between osteoblast function and skeletal integrity in T1D is less clear.
Osteocalcin (OC) is a bone-specific protein product of osteoblasts, osteocytes and odontoblasts. Intracellular proosteocalcin undergoes posttranslational modification by the vitamin K- and CO2-dependent synthesis of three gamma-carboxyglutamic (GLA) residues, allowing for α-helical conformation of the protein, a prerequisite for osteocalcin to adsorb to calcium ions on hydroxyapatite crystals in bone (6). Hence, carboxylated osteocalcin (GLA-OC) accumulates in bone (7). In contrast, undercarboxylated OC (UC-OC), with less than three carboxylated Gla residues, does not undergo α-helical conformation and has lesser affinity for Ca2+ and hydroxyapatite. Therefore, while both forms of OC are found in bone, concentrations of total OC in the systemic circulation are relatively smaller, with a greater proportion of serum OC being the undercarboxylated form (8).
Studies which report alterations in serum OC concentrations in various skeletal, as well as non-skeletal diseases are numerous (6). Physiological variations in OC concentrations are also recognized. Specifically, it is known that serum concentrations of OC and UC-OC are higher in children and adolescents (9), and exhibit diurnal variation, being higher in the early morning than in the late afternoon (9–11)
Recently, a direct in vitro stimulatory effect of osteocalcin on insulin expression and secretion by pancreatic islet cells cultured from wild-type mice has been demonstrated (12–13). Consistent with these findings, osteoblast-specific Esp−/− mice, an osteocalcin gain-of-function model, are thought to exhibit increased pancreatic β-cell proliferation, increased insulin secretion and improved insulin sensitivity, leading to fasting hypoglycemia (13). Inversely, transgenic mice over-expressing Esp, or osteocalcin knockout mice (Ocn−/−) display the opposite metabolic phenotype, specifically decreased β-cell proliferation and glucose intolerance (13). These observations have not, as yet, been corroborated in humans by genetically informative clinical conditions. However, cross-sectional studies in humans have demonstrated an association of higher serum UC-OC concentrations with improved glucose tolerance in middle-aged men (14) as well as in patients with T2D (15).
In the present study, we conducted an analysis to answer a related physiological question; specifically, in a condition of absolute insulin deficiency, can dysregulation of circulating concentrations of UC-OC, GLA-OC, or the ratio of UC-OC to GLA-OC be demonstrated. If animal model data is reflective of the human condition, we hypothesized that compensatory upregulation of UC-OC should occur in patients with T1D (i.e., insulin deficiency with normal insulin sensitivity), and specifically in those with suboptimal glycemic control. To examine this question, first-morning, fasting serum concentrations of UC-OC and GLA-OC were compared between T1D subjects and age-matched healthy control subjects.
RESEARCH DESIGN AND METHODS
Study population
Subjects with T1D and age-matched healthy control subjects, ages 14–40 years, were recruited from clinics at the University of Arkansas for Medical Sciences (UAMS), Arkansas Children’s Hospital (ACH) and surrounding communities. Approval was obtained from the Institutional Review Board of UAMS and all subjects provided informed consent (18–40 years) or assent with parental consent (14–17 years). Exclusion criteria included: 1) T2D; 2) history of other chronic systemic inflammatory or autoimmune disease or malignancy; 3) pregnancy; 4) concurrent ketonuria; 5) active infection; or 6) concurrent use of oral glucocorticoid therapy for any reason. Additionally, no subject had evidence of skeletal injury at the time of evaluation.
For all subjects, two first-morning, outpatient evaluations were conducted 3–5 days apart, allowing us to obtain: 1) demographic information and medical history, including insulin daily dose (IDD); 2) two separate fasting venipuncture laboratory measurements of plasma glucose (FPG), HbA1c and C-peptide; and 3) a 3 day interval recording of continuous glucose monitor sensor data (using the Medtronic Minimed® CGMS, MMT-7102, Northridge, CA), indicative of integrated, concurrent glycemic control. Due to specimen volume limitations, serum from visit 1 was used for measurement of GLA-type osteocalcin (GLA-OC), undercarboxylated osteocalcin (UC-OC), adiponectin (Acrp30), leptin, insulin-like growth factor-I (IGF-I) and for quantification of type 1 collagen degradation fragments (CTX); serum from visit 2 was used for measurement of total calcium, 25-hydroxy vitamin D (25OHD) and intact parathyroid hormone (PTH), as previously reported (5). No effort was made to control for or assess dietary calcium or vitamin D intake. However, concurrent medication use, including calcium or vitamin D supplements, was recorded.
Assays
Fasting plasma glucose (FPG) and C-peptide were measured by the UAMS General Clinical Research Center Core Laboratory. HbA1c was measured by LabCorp (Dallas, TX). Serum measurements of total calcium (reference range, 8.4–10.2 mg/dl), intact parathyroid hormone (PTH: reference range, 15–75 pg/ml; electrochemiluminescent immunoassay) and 25-hydroxy vitamin D (25OHD: reference range, optimal level 30–80 ng/ml; insufficiency 20–29 ng/ml; deficiency <20 ng/ml; competitive immunoluminometry) were performed at ARUP Laboratories (Salt Lake City, UT). IGF-I was measured using the Active IGF-I ELISA (#DSL-10-5600, Diagnostic Systems Laboratories, Inc., Webster, TX). Undercarboxylated osteocalcin (UC-OC) and GLA-type osteocalcin (GLA-OC) were measured using Takara Bio Inc., EIA kits (MK118 and MK111, respectively, Madison, WI). As a measure of bone resorption, degradation products of type 1 collagen (CTX) were measured using the serum CrossLAPs ELISA (AC-02F1, Immunodiagnostic Systems, Inc., Scottsdale, AZ). Additionally, concentrations of leptin and adiponectin, as both directly proportional (leptin) and inversely proportional (adiponectin) markers of total body fat mass, were measured using the MILLIPLEX MAP serum adipokine panel B multiplex assay (HADK2-61-B, Millipore Corp. Billerica, MA) and a human total adiponectin ELISA (DRP300, R & D Systems, Inc., Minneapolis, MN), respectively.
Statistical Analyses
Statistical analysis was performed using R software (version 2.12.1; R Development, 2010) and SPSS Statistical software (version 18.0; SPSS Inc., Chicago, ILL). Results for FPG, HbA1c, and C-peptide obtained from the 2 study visits were averaged. Concentrations of UC-OC, GLA-OC, IGF-I, CTX, calcium, PTH, 25OHD, adiponectin and leptin were measured only once per subject. The ratio of UC-OC to GLA-OC (U:G ratio), the ratio of C-peptide to FPG (C-peptide: Glucose ratio) and the ratio of adiponectin to leptin (A:L ratio) were also calculated. Kruskal-Wallis non-parametric tests were used to ascertain differences between multiple groups whereas Mann-Whitney tests were used for specific two-group comparisons. Spearman’s rank correlation coefficients were assessed to estimate correlation between variables of interest. We also used a linear regression model, incorporating the pertinent interaction terms with group (control, T1D) to identify group difference of associations between dependent and independent variables of interest. If necessary, monotonous transformations (e.g., log transformations) of variables were used. Additionally, possible predictor and confounding variables were modeled in a multivariate regression, to identify those factors that best predicted the outcome variable, specifically UC-OC or GLA-OC. Data is presented as median and range for continuous variables. For all analyses, a statistically significant difference was defined by p≤0.05.
RESULTS
The study population included 55 control subjects and 115 subjects with T1D. Clinical characteristics of the study population are shown in Table 1. Aspects of this study population have also been reported elsewhere (5, 16). Control and T1D groups were comparable with respect to gender (56% vs. 53% female, respectively); racial distribution (86% vs. 93% Caucasian, respectively) and baseline BMI (25.2 ± 4.7 vs. 24.9 ± 4.4 kg/m2, respectively). No subjects were receiving oral calcium supplements or vitamin K treatment whereas eight subjects (2 Control, 6 T1D) reported daily intake of a generic multivitamin containing vitamin D. Of the T1D subjects, 99 subjects were categorized as normoalbuminuric at the time of evaluation (24-hour urinary albumin excretion (UAE) ≤30 mg/g Cr), while UAE concentrations were within a microalbuminuria range in 14 subjects (UAE >30–299 mg/g Cr) and within a macroalbuminuria range (UAE ≥300 mg/g Cr) in 2 subjects (5).
Table 1.
Clinical Characteristics of the Study Population
| Control | T1D | All | p-value: Control vs. T1D | |
|---|---|---|---|---|
| n | 55 | 115 | 170 | --- |
| Female: Male | 31:24 | 61:54 | 92:78 | --- |
| Age (years) | 22.3 (14.1, 41.8) | 17.5 (14.1, 41.5) | 18.8 (14.1, 41.8) | <0.001 |
| BMI (kg/m2) | 24.3 (14.9, 38.1) | 23.9 (17.5, 36.7) | 24.1 (14.9, 38.1) | NS |
| Duration of DM (years) | ---- | 8.7 (0.09, 38.3) | ---- | ---- |
| HbA1c (%) | 5.00 (4.40, 5.65) | 7.95 (3.95, 14.80) | 7.22 (3.95, 14.80) | <0.001 |
| CGMS glucose (mg/dL) | 86.8 (64.49, 110.8) | 159.9 (95.81, 335.4) | 131.9 (64.49, 335.4) | <0.001 |
| C-peptide | 0.75 (0.12, 2.82) | 0.09 (0.03, 1.07) | 0.14 (0.03, 2.82) | <0.001 |
| Insulin Daily Dose (u/kg/d) | ---- | 0.81 (0.27, 1.70) | ---- | ---- |
| Total Calcium (mg/dL) | 9.5 (8.4, 10.9) | 9.5 (7.0, 11.0) | 9.5 (7.0, 11.0) | NS |
| 25 OHD (ng/mL) | 34 (12, 75) | 28 (8, 66) | 30 (8, 75) | 0.059 |
| PTH (pg/mL) | 11 (2, 37) | 18 (1, 60) | 17 (1, 60) | <0.001 |
Comparisons are made between control and T1D subjects.
Abbreviations: BMI, body mass index; HbA1c, hemoglobin A1c; CGMS, continuous glucose monitoring system; 25OHD, 25-hydroxy vitamin D; PTH, parathyroid hormone; NS, not significant.
All data are expressed as mean (range). P-values are presented for comparisons between control and T1D subjects.
Serum UC-OC, GLA-OC, CTX, adiponectin and leptin concentrations were first compared between control and T1D subgroups, and further examined by gender, age, and level of glycemic control. No statistically significant differences in concentration of UC-OC, GLA-OC, U:G ratio, CTX or leptin were observed between the T1D and control populations (Table 2). Adiponectin concentrations were, however, significantly higher in the T1D subgroup, and consistent with this, the A:L ratio was also significantly higher in the T1D group [median (range), Control: 1.54 (0.55, 25.79) vs. T1D: 2.08 (1.44, 61.64); p=0.02]. When T1D subjects alone were subdivided into those with HbA1c <7.5% vs. those with HbA1c ≥7.5%, adiponectin levels were also higher in subjects with poorer glycemic control (<7.5%: 10.43 μg/mL (3.91, 25.22) vs. ≥7.5%: 12.93 μg/mL (4.08, 70.86); p=0.002). Again, though, no significant differences were identified based upon differing levels of glycemic control for the other parameters [(<7.5% vs. ≥7.5%), UC-OC: 7.20 (1.51, 56.33) vs. 7.46 (0.35, 59.72), p=0.9; GLA-OC: 7.44 (0.26, 33.62) vs. 5.57 (0.25, 50.19), p=0.5; U:G ratio: 2.05 (0.13, 44.45) vs. 1.14 (0.03, 85.78), p=0.5; CTX: 0.43 (0.17, 1.08) vs. 0.45 (0.10, 2.23), p=0.7; Leptin: 6.08 (0.20, 35.80) vs. 5.32 (0.63, 40.66), p=1.0, respectively).
Table 2.
Bone Markers
| Parameter | Condition | p-value | |
|---|---|---|---|
| Control | T1D | ||
|
|
|||
| UC-OC (ng/mL) | 6.43 (1.56, 45.59) | 7.29 (0.35, 59.72) | NS |
| GLA-OC (ng/mL) | 5.96 (0.25, 21.14) | 6.78 (0.25, 50.19) | NS |
| UC-OC: GLA-OC | 1.38 (0.11, 46.30) | 1.19 (0.03, 85.78) | NS |
| CTX (ng/mL) | 0.47 (0.09, 1.76) | 0.44 (0.10, 2.23) | NS |
| ADIPONECTIN (μg/mL) | 8.23 (1.11, 27.92) | 12.07 (3.91, 70.86) | <0.001 |
| LEPTIN (ng/mL) | 5.47 (0.26, 44.65) | 5.39 (0.20, 40.66) | NS |
|
| |||
| Male | Female | ||
|
|
|||
| UC-OC (ng/mL) | 7.85 (0.64, 59.72) | 5.74 (0.35, 56.27) | <0.007 |
| GLA-OC (ng/mL) | 7.84 (0.25, 50.19) | 5.08 (0.25, 23.99) | <0.002 |
| UC-OC:GLA-OC | 1.38 (0.03, 85.78) | 1.31 (0.10, 46.30) | NS |
| CTX (ng/mL) | 0.61 (0.15, 2.23) | 0.36 (0.09, 1.49) | <0.001 |
| ADIPONECTIN (μg/mL) | 10.27 (2.20, 70.86) | 11.96 (1.11, 37.32) | 0.003 |
| LEPTIN (ng/mL) | 2.18 (0.20, 21.26) | 9.47 (0.28, 44.65) | <0.001 |
|
| |||
| <18 years | ≥18 years | ||
|
|
|||
| UC-OC (ng/mL) | 15.20 (0.64, 59.72) | 4.84 (0.35, 39.37) | <0.001 |
| GLA-OC (ng/mL) | 10.28 (0.25, 50.19) | 4.24 (0.25, 43.10) | =0.001 |
| UC-OC: GLA-OC | 1.97 (0.03, 46.30) | 0.95 (0.10, 85.78) | =0.008 |
| CTX (ng/mL) | 0.74 (0.10, 2.23) | 0.35 (0.09, 1.04) | <0.001 |
| ADIPONECTIN (μg/mL) | 12.00 (4.11, 70.86) | 9.22 (1.11, 37.32) | 0.048 |
| LEPTIN (ng/mL) | 5.12 (0.20, 40.66) | 5.77 (0.26, 44.65) | NS |
Concentrations of UC-OC, GLA-OC, the ratio of UC-OC to GLA-OC, CTX, adiponectin and leptin are compared between: 1) control and T1D subjects; 2) male and female subjects; and 3) subjects < 18 years of age and ≥ 18 years of age.
All data are expressed as median (range).
Mean UC-OC, GLA-OC and CTX concentrations were higher in male subjects, while adiponectin and leptin were higher in females (Table 2). These gender specific differences were maintained within the T1D subgroup for UC-OC (p=0.01), GLA-OC (p=0.002), CTX (p<0.001), adiponectin (p=0.04) and leptin (p<0.0001) but only for CTX (p=0.003), adiponectin (p=0.005) and leptin (p<0.001) within the control group. (Of note, mean UC-OC and GLA-OC concentration were also consistently higher in control male subjects, but the difference was not statistically significant).
Also, UC-OC, GLA-OC, U:G ratio, CTX and adiponectin concentrations were significantly higher in subjects <18 years of age (Table 2). Age differences were maintained within the T1D subgroup for UC-OC (p<0.001), GLA-OC (p=0.001), and CTX (p<0.0001) but not for adiponectin. Age differences were maintained within the control subgroup for UC-OC (p<0.001), U:G ratio (p=0.01) and CTX (p<0.0001), but not for GLA-OC or adiponectin.
Relationships between UC-OC, GLA-OC or CTX (as dependent variable) and: 1) indices of glycemic control (HbA1c, fasting plasma glucose, CGMS glucose, IDD, C-peptide:Glucose); 2) indices of skeletal homeostasis (Total Ca, 25OHD, PTH, IGF-I, CTX); and 3) patient phenotype parameters (age, adiponectin, leptin) were then explored. Correlations between paired variables of interest are shown in Table 3. In both the control and T1D subgroups, UC-OC concentrations were strongly positively correlated with IGF-I and CTX concentrations, yet negatively correlated with age. GLA-OC concentrations were positively correlated with IGF-I and PTH.
Table 3.
Correlation coefficients for Variables of Interest
| Paired Comparison for Spearman Correlation | Control Coefficient (p-value) |
T1D Coefficient (p-value) |
Overall Coefficient (p-value) |
|---|---|---|---|
| UC-OC with | |||
| Age | −0.703 (<0.001) | −0.601 (<0.001) | −0.621 (<0.001) |
| BMI | −0.170 (=0.215) | −0.340 (<0.001) | −0.297 (<0.001) |
| IGF-I | 0.465 (<0.001) | 0.496 (<0.001) | 0.493 (<0.001) |
| C-peptide: Glucose** | −0.101 (=0.465) | 0.251 (=0.007) | 0.047 (=0.541) |
| IDD | N/A | 0.387 (<0.001) | N/A |
| Total Calcium | 0.345 (=0.01) | 0.154 (=0.106) | 0.206 (=0.008) |
| Leptin | −0.089 (=0.522) | −0.254 (=0.006) | −0.208 (=0.007) |
|
| |||
| GLA-OC with | |||
| Age ** | −0.011 (=0.939) | −0.323 (<0.001) | −0.243 (<0.001) |
| IGF-I | 0.351 (=0.019) | 0.467 (<0.001) | 0.424 (<0.001) |
| PTH | 0.369 (=0.006) | 0.329 (<0.001) | 0.364 (<0.001) |
| 25OHD | −0.005 (=0.972) | 0.244 (=0.009) | 0.135 (=0.078) |
|
| |||
| CTX with | |||
| UC-OC | 0.697 (<0.001) | 0.751 (<0.001) | 0.738 (<0.001) |
| GLA-OC** | 0.266 (=0.052) | 0.544 (<0.001) | 0.454 (<0.001) |
| IGF-I | 0.431 (=0.001) | 0.530 (<0.001) | 0.504 (<0.001) |
| PTH | 0.011 (=0.939) | 0.218 (=0.020) | 0.150 (=0.052) |
Spearman rank correlation was assessed as a measure of association between paired variables of interest (delineated in Results).
Data is shown for those associations in which at least one subgroup analysis (control or T1D) was statistically significant, indicated in bold font.
A significant group difference of association (interaction) was confirmed between UC-OC and C-peptide:Glucose (p=0.04), between GLA-OC and CTX (p=0.02), and between GLA-OC and age (p=0.05).
N/A; not applicable.
In the T1D subgroup specifically, UC-OC concentrations were positively correlated with: 1) C-peptide:Glucose ratio (reflective of endogenous insulin secretion); IGF-I (indicative of intra-portal insulin sufficiency; and with 3) total daily dose of insulin (IDD). UC-OC concentrations were also inversely correlated with BMI and leptin, but not correlated with HbA1c, mean fasting glucose, mean CGMS glucose, adiponectin, or the A:L ratio. Hence, in a population of patients without insulin resistance, UC-OC correlated positively with markers of insulin exposure, either endogenously produced or exogenously administered and negatively with higher body fat. GLA-OC concentrations in the T1D subgroup were positively correlated with IGF-I, PTH, 25-OHD, and CTX, while being negatively correlated with age. Finally, a significant group difference of association (interaction) was confirmed between UC-OC and C-peptide:Glucose (p=0.04), between GLA-OC and CTX (p=0.02), and between GLA-OC and age (p=0.05).
To determine those factors most strongly affecting plasma UC-OC or GLA-OC concentrations, we performed multivariate regression modeling, using group (control vs. T1D), gender, age, PTH, HbA1c, CGMS glucose, IGF-I, 25OHD, CTX, leptin, and adiponectin as covariates. After backward elimination from the full model set, significant correlates of UC-OC included CTX (p<0.001, β-coefficient=0.516), subject age (p<0.001, β-coefficient=−0.248) and HbA1c value (p=0.013, β-coefficient=−0.141) with a multiple R2=0.542. IGF-I concentrations were marginally related to UC-OC (p=0.086, β-coefficient=0.109). Specifically, being younger, having higher CTX and lower HbA1c values, and possibly a higher IGF-I concentration all contributed to a higher UC-OC.
Significant correlates of GLA-OC included CTX (p<0.001, β-coefficient=0.412), HbA1c (p=0.009, β-coefficient=0.247), 25OHD (p=0.001, β-coefficient=0.221), PTH (p=0.005, β-coefficient=0.189), IGF-I (p=0.039, β-coefficient=0.147) and gender (p=0.029, β-coefficient=0.147) with a multiple R2 =0.411. Specifically, higher CTX, HbA1c, 25OHD, PTH, and IGF-I levels were all associated with higher GLA-OC concentrations; additionally, male gender contributed to a higher GLA-OC.
CONCLUSIONS
In contrast to our hypothesis, concentrations of UC-OC and GLA-OC were not statistically different between T1D and control populations; and specifically, concentrations of UC-OC were not significantly elevated in the T1D group. In fact, by multivariate analysis, having a higher HbA1c, reflective of poorer glycemic control, was correlated with a lower UC-OC but inversely with a higher GLA-OC. Additionally, UC-OC was inversely correlated with duration of disease (R=−0.299, p=0.001), suggesting that longer duration of T1D related to a lower UC-OC. Consistent with our findings, an inverse association of UC-OC with HbA1c has very recently been demonstrated in a population of over 1500 elderly Japanese men, some with T2D (17), as well as in a population of 180 men, all with T2D (15). Additionally, an inverse correlation of total OC with HbA1c has also been demonstrated in children with T1D (18–19).
Expected gender and age-based differences for UC-OC, GLA-OC and leptin (20–23) were confirmed within our overall study population, as well as within the T1D subgroup. Adiponectin concentrations were also higher in females, both overall and in T1D. Moreover, adiponectin levels were marginally higher in subjects under 18 years of age, though this age difference was not maintained with the T1D group. Increased plasma concentrations of adiponectin have been reported in Japanese adults with T1D (22), while in US pediatric subjects with T1D and in adults with Latent autoimmune diabetes (LADA) (24), adiponectin levels were not different from those in similarly aged healthy control subjects (25). Of note, in our study, the age range (14–40 years) crosses both the pediatric and adult classification, perhaps accounting for our failure to identify differences in adiponectin concentration between a younger and older T1D cohort; none-the-less, overall, adiponectin levels were higher in T1D subjects. To our knowledge, however, gender and age differences for UC-OC and GLA-OC in T1D subjects specifically, have not been demonstrated previously.
Disease-based alterations in either adiponectin or leptin concentrations in T1D, T2D, or as they relate to glucose homeostasis, have been examined extensively by others (24–27), and recreation of these findings was not the intent of our study. Rather, we were interested in examining possible interactions of adiponectin or leptin with UC-OC and GLA-OC concentrations in T1D. Studies in genetically modified mice suggest that UC-OC has a direct affect on adipose tissue to induce adiponectin production, but not leptin (28). Also, in inverse correlation between carboxylated osteocalcin and high molecular weight adiponectin has been reported in healthy children (29). Surprisingly, in the current studies, neither UC-OC nor GLA-OC concentrations were related to adiponectin levels or to the A:L ratio, either by correlation analysis or by multivariate modeling, either overall, or in the T1D subgroup. In contrast, leptin was negatively correlated with UC-OC, but only in T1D subjects. Thus, the interaction of UC-OC with adipokines in humans appears to be more complex than that described in rodents, and these interactions may be different in T1D than in non-diabetic subjects.
In vitro, insulin can up-regulate osteocalcin expression in osteoblasts, through inhibiting the forkhead transcription factor, FoxO1 (30). Furthermore, in genetically modified mice, insulin signaling in osteoblasts appears to enhance osteoclast activity and osteoclasts, in turn, decarboxylate OC (31). Thus, in mice, when insulin signals in osteoblasts, systemic UC-OC production is increased (31). The current studies address this consideration in humans by showing that in type 1 diabetes, UC-OC concentrations were associated with C-peptide:Glucose ratio (i.e., endogenously secreted insulin) and with the total daily dose of insulin (i.e., exogenously administered insulin). Interestingly, the C-peptide:Glucose ratio was not significantly associated with UC-OC in non-diabetic subjects. Nonetheless, the associations in T1D are consistent with the reported effects of insulin on bone-derived UC-OC. Moreover, in both Control and T1D groups, serum IGF-I concentrations were strongly associated with GLA-OC and UC-OC, suggesting that IGF-I, an insulin homolog, also plays a significant role in controlling OC serum concentrations, irrespective of the underlying insulin dynamics.
We also observed a very strong relationship between UC-OC and CTX, both by paired comparisons (Table 3), as well as multivariate regression modeling, supporting a role for bone resorption in the decarboxylation of osteocalcin, as has been suggested by others (32). Alternatively, this association may reflect coupling between bone resorption and bone formation, as has been suggested by studies demonstrating correlation of CTX with total OC in individuals with T1D (33). Consistent with this, we observed a highly significant, though weaker correlation between GLA-OC and CTX in T1D.
A primary limitation of this study is the fact that it was conducted as a retrospective analysis of a preexisting specimen and data set. As such, this analysis represents a cross-sectional, single point-in-time assessment of skeletal and glycemic status. Moreover, we do not have an assessment of bone mineral density available in this dataset. Racial diversity of the study population is also limited, though the racial distribution is comparable across the control and T1D study groups, and not appreciably inconsistent with demographic expectations for type 1 diabetes (34). However, the study group is relatively large, and sufficient in size to confirm many expected physiological determinants of UC-OC and GLA-OC.
In summary, since UC-OC has been shown to increase insulin production and beta cell expansion in mice (31), we anticipated that UC-OC might be elevated in T1D as a compensatory mechanism to improve endogenous insulin production; instead, we found that UC-OC concentrations were similar in control and T1D individuals. However, in T1D, the association of UC-OC serum concentrations with either endogenous production of insulin or exogenous administration of insulin suggests that UC-OC may, at least in part, be regulated through insulin-mediated events. Other parameters also showed strong associations with UC-OC, GLA-OC, or with both, including: age, CTX, HbA1c, IGF-I, PTH and 25OHD, confirming that UC-OC and GLA-OC are regulated by a variety of developmental and metabolic pathways, irrespective of insulin status.
Acknowledgments
Funding Sources: Grants from the National Institutes of Health to KMT (R01-DK62999), to the UAMS General Clinical Research Center (M01 RR14288) and to the Arkansas Children’s Hospital Research Institute (C06RR16517), as well as funds from the Minnie Merrill Sturgis Diabetes Research Fund and the Arkansas Biosciences Institute.
This work was supported by grants from the National Institutes of Health to KMT (R01-DK62999); to the UAMS General Clinical Research Center (M01 RR14288) and to the Arkansas Children’s Hospital Research Institute (C06RR16517). Additional funding was provided by the Minnie Merrill Sturgis Diabetes Research Fund and the Arkansas Biosciences Institute. The authors are also grateful to the study subjects and their families for participation in this research.
Footnotes
CONFLICTS OF INTEREST: The authors have no financial relationships or conflicts of interest to disclose.
References
- 1.Craig ME, Hattersley A, Donaghue K. ISPAD Clinical Practice Consensus Guidelines 2006–2007. Definition, epidemiology and classification. Pediatr Diabetes. 2006;7:343–351. doi: 10.1111/j.1399-5448.2006.00216.x. [DOI] [PubMed] [Google Scholar]
- 2.Thrailkill KM. Insulin-like growth factor-I in diabetes mellitus: its physiology, metabolic effects, and potential clinical utility. Diabetes Technol Ther. 2000;2:69–80. doi: 10.1089/152091599316775. [DOI] [PubMed] [Google Scholar]
- 3.Thrailkill KM, Lumpkin CK, Jr, Bunn RC, Kemp SF, Fowlkes JL. Is insulin an anabolic agent in bone? Dissecting the diabetic bone for clues. Am J Physiol Endocrinol Metab. 2005;289:E735–745. doi: 10.1152/ajpendo.00159.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Nyman JS, Even JL, Jo CH, Herbert EG, Murry MR, Cockrell GE, Wahl EC, Bunn RC, Lumpkin CK, Jr, Fowlkes JL, Thrailkill KM. Increasing duration of type 1 diabetes perturbs the strength-structure relationship and increases brittleness of bone. Bone. 2011;48:733–740. doi: 10.1016/j.bone.2010.12.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Thrailkill KM, Jo CH, Cockrell GE, Moreau CS, Fowlkes JL. Enhanced excretion of vitamin D binding protein in type 1 diabetes: a role in vitamin D deficiency? J Clin Endocrinol Metab. 2011;96:142–149. doi: 10.1210/jc.2010-0980. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Hauschka PV, Lian JB, Cole DE, Gundberg CM. Osteocalcin and matrix Gla protein: vitamin K-dependent proteins in bone. Physiol Rev. 1989;69:990–1047. doi: 10.1152/physrev.1989.69.3.990. [DOI] [PubMed] [Google Scholar]
- 7.Lian JB, Gundberg CM. Osteocalcin. Biochemical considerations and clinical applications. Clin Orthop Relat Res. 1988:267–291. [PubMed] [Google Scholar]
- 8.Motyl KJ, McCabe LR, Schwartz AV. Bone and glucose metabolism: a two-way street. Arch Biochem Biophys. 2010;503:2–10. doi: 10.1016/j.abb.2010.07.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Szulc P, Delmas PD. Biochemical markers of bone turnover: potential use in the investigation and management of postmenopausal osteoporosis. Osteoporos Int. 2008;19:1683–1704. doi: 10.1007/s00198-008-0660-9. [DOI] [PubMed] [Google Scholar]
- 10.Heuck C, Skjaerbaek C, Wolthers OD. Diurnal rhythm of serum osteocalcin in normal children. Acta Paediatr. 1998;87:930–932. doi: 10.1080/080352598750031563. [DOI] [PubMed] [Google Scholar]
- 11.Sokoll LJ, Booth SL, Davidson KW, Dallal GE, Sadowski JA. Diurnal variation in total and undercarboxylated osteocalcin: influence of increased dietary phylloquinone. Calcif Tissue Int. 1998;62:447–452. doi: 10.1007/s002239900458. [DOI] [PubMed] [Google Scholar]
- 12.Ferron M, Hinoi E, Karsenty G, Ducy P. Osteocalcin differentially regulates beta cell and adipocyte gene expression and affects the development of metabolic diseases in wild-type mice. Proc Natl Acad Sci U S A. 2008;105:5266–5270. doi: 10.1073/pnas.0711119105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lee NK, Sowa H, Hinoi E, Ferron M, Ahn JD, Confavreux C, Dacquin R, Mee PJ, McKee MD, Jung DY, Zhang Z, Kim JK, Mauvais-Jarvis F, Ducy P, Karsenty G. Endocrine regulation of energy metabolism by the skeleton. Cell. 2007;130:456–469. doi: 10.1016/j.cell.2007.05.047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hwang YC, Jeong IK, Ahn KJ, Chung HY. The uncarboxylated form of osteocalcin is associated with improved glucose tolerance and enhanced beta-cell function in middle-aged male subjects. Diabetes Metab Res Rev. 2009;25:768–772. doi: 10.1002/dmrr.1045. [DOI] [PubMed] [Google Scholar]
- 15.Kanazawa I, Yamaguchi T, Yamauchi M, Yamamoto M, Kurioka S, Yano S, Sugimoto T. Serum undercarboxylated osteocalcin was inversely associated with plasma glucose level and fat mass in type 2 diabetes mellitus. Osteoporos Int. 2011;22:187–194. doi: 10.1007/s00198-010-1184-7. [DOI] [PubMed] [Google Scholar]
- 16.Thrailkill KM, Bunn RC, Moreau CS, Cockrell GE, Simpson PM, Coleman HN, Frindik JP, Kemp SF, Fowlkes JL. Matrix metalloproteinase-2 dysregulation in type 1 diabetes. Diabetes Care. 2007;30:2321–2326. doi: 10.2337/dc07-0162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Iki M, Tamaki J, Fujita Y, Kouda K, Yura A, Kadowaki E, Sato Y, Moon JS, Tomioka K, Okamoto N, Kurumatani N. Serum undercarboxylated osteocalcin levels are inversely associated with glycemic status and insulin resistance in an elderly Japanese male population: Fujiwara-kyo Osteoporosis Risk in Men (FORMEN) Study. Osteoporos Int. 2011 doi: 10.1007/s00198-011-1600-7. [DOI] [PubMed] [Google Scholar]
- 18.Maggio AB, Ferrari S, Kraenzlin M, Marchand LM, Schwitzgebel V, Beghetti M, Rizzoli R, Farpour-Lambert NJ. Decreased bone turnover in children and adolescents with well controlled type 1 diabetes. J Pediatr Endocrinol Metab. 2010;23:697–707. doi: 10.1515/jpem.2010.23.7.697. [DOI] [PubMed] [Google Scholar]
- 19.Pater A, Sypniewska G, Pilecki O. Biochemical markers of bone cell activity in children with type 1 diabetes mellitus. J Pediatr Endocrinol Metab. 2010;23:81–86. doi: 10.1515/jpem.2010.23.1-2.81. [DOI] [PubMed] [Google Scholar]
- 20.Nimptsch K, Hailer S, Rohrmann S, Gedrich K, Wolfram G, Linseisen J. Determinants and correlates of serum undercarboxylated osteocalcin. Ann Nutr Metab. 2007;51:563–570. doi: 10.1159/000114211. [DOI] [PubMed] [Google Scholar]
- 21.Thrailkill KM, Fowlkes JL, Hyde JF, Litton JC. The effects of co-therapy with recombinant human insulin-like growth factor I and insulin on serum leptin levels in adolescents with type 1 diabetes mellitus. Pediatr Diabetes. 2001;2:25–29. doi: 10.1046/j.1399-543x.2001.00000.x. [DOI] [PubMed] [Google Scholar]
- 22.Imagawa A, Funahashi T, Nakamura T, Moriwaki M, Tanaka S, Nishizawa H, Sayama K, Uno S, Iwahashi H, Yamagata K, Miyagawa J, Matsuzawa Y. Elevated serum concentration of adipose-derived factor, adiponectin, in patients with type 1 diabetes. Diabetes Care. 2002;25:1665–1666. doi: 10.2337/diacare.25.9.1665. [DOI] [PubMed] [Google Scholar]
- 23.Considine RV. Regulation of leptin production. Rev Endocr Metab Disord. 2001;2:357–363. doi: 10.1023/a:1011896331159. [DOI] [PubMed] [Google Scholar]
- 24.Brophy S, Davies H, Stephens JW, Prior SL, Atkinson M, Bain S, Williams R. Adiponectin levels in people with Latent Autoimmune Diabetes-a case control study. BMC Res Notes. 2010;3:317. doi: 10.1186/1756-0500-3-317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Morales A, Wasserfall C, Brusko T, Carter C, Schatz D, Silverstein J, Ellis T, Atkinson M. Adiponectin and leptin concentrations may aid in discriminating disease forms in children and adolescents with type 1 and type 2 diabetes. Diabetes Care. 2004;27:2010–2014. doi: 10.2337/diacare.27.8.2010. [DOI] [PubMed] [Google Scholar]
- 26.Maahs DM, Hamman RF, D’Agostino R, Jr, Dolan LM, Imperatore G, Lawrence JM, Marcovina SM, Mayer-Davis EJ, Pihoker C, Dabelea D. The association between adiponectin/leptin ratio and diabetes type: the SEARCH for Diabetes in Youth Study. J Pediatr. 2009;155:133–135. e131. doi: 10.1016/j.jpeds.2008.12.048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Leth H, Andersen KK, Frystyk J, Tarnow L, Rossing P, Parving HH, Flyvbjerg A. Elevated levels of high-molecular-weight adiponectin in type 1 diabetes. J Clin Endocrinol Metab. 2008;93:3186–3191. doi: 10.1210/jc.2008-0360. [DOI] [PubMed] [Google Scholar]
- 28.Lee NK, Karsenty G. Reciprocal regulation of bone and energy metabolism. Trends Endocrinol Metab. 2008;19:161–166. doi: 10.1016/j.tem.2008.02.006. [DOI] [PubMed] [Google Scholar]
- 29.Prats-Puig A, Mas-Parareda M, Riera-Perez E, Gonzalez-Forcadell D, Mier C, Mallol-Guisset M, Diaz M, Bassols J, de Zegher F, Ibanez L, Lopez-Bermejo A. Carboxylation of osteocalcin affects its association with metabolic parameters in healthy children. Diabetes Care. 2010;33:661–663. doi: 10.2337/dc09-1837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Yang S, Xu H, Yu S, Cao H, Fan J, Ge C, Fransceschi RT, Dong HH, Xiao G. Foxo1 Mediates Insulin-like Growth Factor 1 (IGF1)/Insulin Regulation of Osteocalcin Expression by Antagonizing Runx2 in Osteoblasts. J Biol Chem. 2011;286:19149–19158. doi: 10.1074/jbc.M110.197905. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ferron M, Wei J, Yoshizawa T, Del Fattore A, DePinho RA, Teti A, Ducy P, Karsenty G. Insulin signaling in osteoblasts integrates bone remodeling and energy metabolism. Cell. 2010;142:296–308. doi: 10.1016/j.cell.2010.06.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Clemens TL, Karsenty G. The osteoblast: an insulin target cell controlling glucose homeostasis. J Bone Miner Res. 2011;26:677–680. doi: 10.1002/jbmr.321. [DOI] [PubMed] [Google Scholar]
- 33.Alexopoulou O, Jamart J, Devogelaer JP, Brichard S, de Nayer P, Buysschaert M. Bone density and markers of bone remodeling in type 1 male diabetic patients. Diabetes Metab. 2006;32:453–458. doi: 10.1016/s1262-3636(07)70303-8. [DOI] [PubMed] [Google Scholar]
- 34.Liese AD, D’Agostino RB, Jr, Hamman RF, Kilgo PD, Lawrence JM, Liu LL, Loots B, Linder B, Marcovina S, Rodriguez B, Standiford D, Williams DE. The burden of diabetes mellitus among US youth: prevalence estimates from the SEARCH for Diabetes in Youth Study. Pediatrics. 2006;118:1510–1518. doi: 10.1542/peds.2006-0690. [DOI] [PubMed] [Google Scholar]