In humans, the pronounced postprandial reduction in bone resorption (decreasing bone resorption markers by around 50%) has been suggested to be caused by gut hormones. Glucose-dependent insulinotropic polypeptide (GIP) is a peptide hormone secreted postprandially from the small intestine. The hormone is known as an incretin hormone, but preclinical studies have suggested that it may also influence bone metabolism, showing both antiresorptive and anabolic effects as reflected by changes in biomechanical measures, microarchitecture, and activity of the bone cells in response to GIP stimulation. Its role in human bone homeostasis, however, is unknown.
To examine the effect of GIP administration on bone resorption in humans.
Plasma samples were obtained from 10 healthy subjects during four conditions: euglycemic (5 mmol/L) and hyperglycemic (12 mmol/L) 90-minute glucose clamps with co-infusion of GIP (4 pmol/kg/min for 15 min, followed by 2 pmol/kg/min for 45 min) or placebo. The samples were analyzed for concentrations of degradation products of C-terminal telopeptide of type I collagen (CTX), a bone resorption marker. Results regarding effects on pancreatic hormone secretion have been published.
During euglycemia, the decremental area under the curve in CTX was significantly (P < .001) higher during GIP infusion (2084 ± 686 % × min) compared to saline infusion (656 ± 295 % × min). During hyperglycemia, GIP infusion significantly (P < .001) augmented the decremental area under the curve to 2785 ± 446 % × minutes, compared to 1308 ± 448 % × minutes during saline infusion, with CTX values corresponding to 49% of basal values.
We conclude that GIP reduces bone resorption in humans, interacting with a possible effect of hyperglycemia.
Glucose-dependent insulinotropic polypeptide (GIP) is a peptide hormone consisting of 42 amino acids. It is secreted postprandially from endocrine K cells predominantly located in the duodenum and proximal jejunum (1–3). After release, it is N-terminally degraded by the ubiquitous enzyme dipeptidyl peptidase 4, resulting in a half-life in plasma of 6–7 minutes (4). Postprandially, the concentration of intact GIP reaches levels of approximately 60–80 pmol/L (4). GIP exerts its functions through activation of a specific G protein-coupled receptor (5) expressed by many different cells, including pancreatic β-cells and adipocytes (6), and apparently also osteoclasts (7) and osteoblasts (8–10).
GIP is best known for its stimulatory effect on the postprandial insulin secretion in normal individuals, but in agreement with the receptor localization, a role in bone remodeling as part of an “entero-osseous axis” has also been suggested (9).
Bone remodeling is a constantly ongoing process that adapts the skeleton to varying mechanical and physical requirements. Bone remodeling also plays a part in calcium and phosphate homeostasis. The bone remodeling process shows a diurnal variation with a pronounced, postprandial reduction in bone resorption (11). The mechanism responsible for this is not known, but an entero-osseous axis has been suggested to exist (12). Thus, Clowes et al (12) showed that the postprandial decrease in bone resorption can be prevented by administration of a somatostatin analog. Because somatostatin inhibits the secretion of gut (and pancreatic) hormones, this supports the notion of a role of enteric hormones in the postprandial fall in bone resorption.
GIP is a candidate for this entero-osseous axis. GIP has been reported to exert an antiapoptotic and stimulating effect on osteoblasts in vitro (8, 13, 14). In animal models, increased numbers of osteoblasts and increased serum concentrations of osteocalcin, a bone formation marker, were observed in response to GIP overexpression (15). Conversely, a decreased concentration of osteocalcin was demonstrated in mice with a deletion of the GIP receptor (16). An antiresorptive effect of GIP was suggested from observations of decreased osteoclast activity in vitro as well as in vivo in GIP receptor knockout (KO) mice. In vitro, there was no direct effect of GIP on osteoclast activity (13), but examined in the presence of PTH or receptor activator of nuclear factor κB ligand, an inhibiting effect was observed (7). In GIP receptor KO mice, increased plasma levels of collagen degradation fragments, such as deoxypyrolidine, were observed, and in histomorphometric analysis, more osteoclasts were observed when compared to wild-type mice (13). In other studies, however, decreased concentrations of collagen degradation fragments were found in the receptor KO mice (17, 18).
In other studies, a less stiff and weaker bone was observed in GIP receptor KO mice in comparison with wild-type mice, and stiffer and stronger bone has been observed in mice with increased circulating concentrations of GIP (15–18). Collectively, these studies quite robustly support a role for GIP in rodent bone metabolism. In contrast, very little is known about GIP effects on bone metabolism in humans. We therefore studied the effect of GIP administration on the bone resorption marker, C-terminal telopeptide of type I collagen (CTX), in healthy volunteers.
Subjects and Methods
We studied the effects of iv infused GIP at fasting and hyperglycemic glucose levels maintained by clamping in healthy volunteers (19).
Subjects
The study population consisted, as described (19), of 10 healthy normal glucose-tolerant male subjects with a median age of 22 (range, 19–30) years and a median body mass index of 22 (range, 20–25) kg/m2. Each subject underwent 6 experimental days within a 2-month period.
Experimental procedures
Results regarding pancreatic endocrine responses were recently published (19). For that part of the study, subjects were studied on 6 different test days. For the present study, we selected plasma samples from 4 of the 6 test days: 2 days with euglycemic clamps (5 mmol/L) and 2 days with hyperglycemic (12 mmol/L) clamps, each with co-infusion of either saline or GIP (4 pmol/kg/min for 15 min followed by 2 pmol/kg/min for 45 min).
Analysis
CTX was measured in plasma samples collected at −10, 0, 30, 60, and 90 minutes using a commercially available sandwich ELISA kit according to the manufacturer's instruction (serum CrossLaps ELISA; Immunodiagnostic Systems Nordic A/S). The ELISA uses highly specific monoclonal antibodies directed against the amino acid sequence EKAHD-β-GGR derived from the C-terminal telopeptide region of collagen 1.
Statistical analysis
Plasma CTX levels are expressed as percentage of fasting level (fasting level = mean of −10 and 0 minutes). The decremental area under the curve (AUC) was calculated using y = 100% as baseline. Differences in CTX levels and differences in decremental AUCs were evaluated by one-way ANOVA for repeated measurements followed by Tukey's post hoc test. P values < .05 were considered significant. Statistical evaluation and graphic presentation were made in GraphPad Prism 5 (GraphPad Software, Inc).
Ethics
The study protocol was approved by the Scientific-Ethical Committee of the Capital Region of Denmark (registration no. H-D-2009-0078), registered with ClinicalTrials.gov (clinical trial no. NCT01048268), and conducted according to the principles of the Helsinki Declaration II. Written informed consent was obtained from all participants before inclusion.
Results
The mean fasting CTX levels were comparable on the 4 days (0.87 ± 0.3, 0.89 ± 0.3, 0.84 ± 0.3, and 0.98 ± 0.5 ng/mL at euglycemia + saline, euglycemia + GIP, hyperglycemia + saline, and hyperglycemia + GIP, respectively).
Plasma CTX levels are shown in Figure 1 (mean ± SD), presented as percentage of the fasting level (baseline). During the euglycemic clamp with saline, there was a gradual decrease in CTX, reaching 86.9 ± 6.8% at 90 minutes (P < .001). With GIP infusion, bone resorption was reduced to 67.3 ± 12.6% of baseline (P < .001). During the hyperglycemic clamp and saline infusion, CTX levels decreased to 74.1 ± 8.6% at 90 minutes (P < .001), whereas GIP infusion resulted in CTX levels of 49.2 ± 8.3% of baseline at 90 minutes (P < .001). The decremental AUCs are shown in Figure 2 (mean ± SD). During euglycemia, the decremental AUC in CTX was significantly (P < .001) higher during co-infusion of GIP (2084 ± 686% × min) compared to saline (656 ± 295% × min). During hyperglycemia, GIP infusion significantly (P < .001) increased the decremental AUC to 2785 ± 446% × minutes compared to 1308 ± 448% × minutes during saline infusion.
Plasma CTX levels (mean ± SD), presented as percentage of the fasting level (baseline) during euglycemia (solid lines) and hyperglycemia (broken lines) with co-infusion of saline (filled symbols) or GIP (open symbols).
Decremental AUCs (mean ± SEM) of CTX calculated using y = 100% as baseline.
*, P < .01; **, P < .001.
Discussion
Preclinical studies have suggested that GIP may influence bone remodeling by inhibiting bone resorption and possibly stimulating bone formation. In this study, we tested the hypothesis that GIP inhibits bone resorption in humans, and our findings support this. By employing glucose clamping, it was possible to study the GIP effects without the influence of the varying glucose concentrations, which would otherwise result from the insulinotropic effects of GIP. The saline-euglycemia clamp experiments showed a slightly, but significantly, decreasing level of CTX, expressed as the percentage of the basal fasting levels (86.9% at 90 min). The mechanism behind this decrease is unclear, but it has previously been observed that even during complete fasting there is a weak diurnal variation in bone resorption as studied by resorption markers, with weak decreases observed during the day time (11). The decreases observed here may reflect this diurnal variation. The infusion of GIP during euglycemia resulted in a significant further reduction in CTX levels, suggesting that GIP inhibits bone resorption. However, this infusion also resulted in a significant but very short-lasting rise in insulin secretion, reaching significance only at time point 5 minutes (19). Therefore, it seems unlikely that this small, short-lasting increase in insulin secretion could be responsible for the subsequent progressive decrease in CTX levels, which were still decreased after the observation period.
In agreement with this, Clowes et al (20) showed that hyperinsulinemia does not affect bone resorption. They subjected healthy young men to euglycemic hyperinsulinemic as well as hypoglycemic hyperinsulinemic clamps. Under euglycemic conditions, hyperinsulinemia did not influence bone resorption, as indicated by a lack of changes in serum CTX levels. However, hyperinsulinemia under hypoglycemic conditions inhibited bone resorption, suggesting that glucose levels rather than insulin levels are important for bone remodeling. The inhibiting effect on bone resorption reported in the study by Clowes et al (20) could also be explained by the acute decrease in PTH due to the hyperinsulinemia or other hormones released in response to hypoglycemia, as suggested by the authors. Taken together, it seems unlikely that the observed changes in CTX levels after GIP infusion in the present study were caused by the small rise in serum insulin.
During the hyperglycemic clamp, there was a larger decrease in CTX compared to euglycemia. The hyperglycemia also resulted in a significant increase in insulin levels, lasting for the duration of the clamp. This suggests that hyperglycemia, or hyperinsulinemia, or both could be responsible. Again, the study by Clowes et al (20) indicates that insulin is unlikely to be responsible, suggesting that plasma glucose levels influence bone resorption. This hypothesis is supported by an in vitro study where the activity of osteoclasts exposed to high glucose concentration was examined, showing that the osteoclasts decrease their activity during hyperglycemia, as reflected by a decrease in a pit formation assay (21).
Addition of GIP to the hyperglycemic clamp caused a further marked lowering of CTX levels, reaching a similar degree of inhibition as observed after intake of large meals, namely to about 50% of basal level (11, 12). GIP, however, also caused a further rise in insulin secretion, but again, on the basis of the studies of Clowes et al (20), it seems unlikely that insulin is responsible. Therefore, our findings strongly support a role for GIP in the postprandial inhibition of bone resorption.
Importantly, the levels of GIP in the present study are physiologically relevant. As already stated, the normal postprandial concentrations of intact GIP are in the range of 60–80 pmol/L (4), whereas the plateau levels obtained here were around 70 pmol/L (19).
During meal ingestion, virtually all of the gut hormone-producing cells are stimulated, and postprandial changes in bone resorption could therefore be influenced by many hormones, including glucagon-like peptide (GLP) 1 and 2. GLP-2 has been demonstrated to significantly lower CTX levels in humans (22), and GLP-1 receptor KO mice have fragile bones (23). These hormones may therefore also play a role in the postprandial inhibition of bone resorption. The present study deals with the isolated effects of iv administered GIP. We have previously shown that iv infusion of GIP in a comparable dose has no effect on plasma levels of GLP-1 and GLP-2 (24). Additional studies looking at GIP effects in models of human bone cells could be beneficial in trying to elucidate the mechanisms underlying the observed effects.
In a previous study from our group, we found no effects of GIP on CTX levels in healthy volunteers (22). However, in that study GIP was administered as a single bolus resulting in a short-lived rise in GIP levels, and CTX levels were followed for only 48 minutes, during which time a CTX decrease of 16% was actually observed. The significance of these results cannot be evaluated because of the lack of a control group in that study. Most likely, however, an effect of GIP was missed in these studies.
Acknowledgments
This work was supported by the Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen.
Author Contributions: A.N., J.J.H., and B.H. planned and designed the study; A.N. and B.H. performed the CTX analyses; M.C., F.K.K., and T.V. designed and performed the clinical studies and provided the plasma samples; A.N. wrote the manuscript; M.C., F.K.K., T.V., J.J.H., and B.H. revised the manuscript. All authors have approved the final version of the manuscript.
Disclosure Summary: The authors declare that they have nothing to disclose associated with this manuscript.
Abbreviations
- AUC
area under the curve
- CTX
C-terminal telopeptide of type I collagen
- GIP
glucose-dependent insulinotropic polypeptide
- GLP
glucagon-like peptide
- KO
knockout.
References
