Role of Gastrointestinal Hormones in Postprandial Reduction of Bone Resorption ABSTRACT

Transcription

1 JOURNAL OF BONE AND MINERAL RESEARCH Volume 18, Number 12, American Society for Bone and Mineral Research Role of Gastrointestinal Hormones in Postprandial Reduction of Bone Resorption DENNIS B HENRIKSEN, 1 PETER ALEXANDERSEN, 2 NINA H BJARNASON, 2 TINA VILSBØLL, 3 BOLETTE HARTMANN, 4 EVA EG HENRIKSEN, 1 INGER BYRJALSEN, 1 THURE KRARUP, 3 JENS J HOLST, 4 and CLAUS CHRISTIANSEN 2 ABSTRACT Collagen type I fragments, reflecting bone resorption, and release of gut hormones were investigated after a meal. Investigations led to a dose escalation study with glucagon like peptide-2 (GLP-2) in postmenopausal women. We found a dose-dependent effect of GLP-2 on the reduction of bone resorption. Introduction: The C-terminal telopeptide region of type I collagen as measured in serum (s-ctx) can be used to assess bone resorption. This marker of bone resorption has a significant circadian variation that is influenced by food intake. However, the mediator of this variation has not been identified. Materials and Methods: We studied the release of the gut hormones glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-2 (GLP-2; a representative of the intestinal proglucagon-derived peptides) after ingestion of glucose, fat, protein, and fructose, as well as their effects after parenteral administration in relation to bone turnover processes in healthy volunteers. Furthermore, we studied the effect on bone turnover of a single subcutaneous injection of GLP-2 in four different dosages (100, 200, 400, or 800 g GLP-2) or placebo in 60 postmenopausal women (mean age, 61 5 years). Results: All macronutrients significantly (p 0.05) reduced bone resorption as assessed by s-ctx (39 52% from baseline), and only the glucagon-like peptides were secreted in parallel. Parenteral administration of GIP and GLP-1 did not result in a reduction of the s-ctx level, whereas GLP-2 caused a statistically significant and dose-dependent reduction in the s-ctx level from baseline compared with placebo (p 0.05). Urine DPD/creatinine, a marker of bone resorption, was significantly reduced by 25% from baseline in the 800- g GLP-2 group (p 0.01). An area under the curve (AUC 0 8h ) analysis for s-ctx after GLP-2 injection confirmed the dose-dependent decrease (ANOVA, p 0.05). The s-osteocalcin level was unaffected by the GLP-2 treatment. Conclusion: These studies exclude both GIP and GLP-1 as key mediators for the immediate reduction in bone resorption seen after a meal. The dose-dependent reduction of bone resorption markers found after subcutaneous injection of GLP-2 warrants further investigation into the mechanism and importance of GLP-2 for the bone turnover processes. J Bone Miner Res 2003;18: Key words: bone resorption, bone turnover markers, nutrients, gut hormones, bone formation Drs Henriksen, Henriksen, and Byrjalsen are employees of Nordic Bioscience. Dr Holst has consulted and received royalties from Nordic Bioscience. Dr Christiansen owns stock in Nordic Bioscience. All other authors have no conflict of interest. INTRODUCTION BONE TISSUE IS continuously remodeled throughout life to adapt and repair the damages endured by the bone through repetitive mechanical loading and age-related hormonal changes. To maintain a constant bone mass during adult life, osteoblastic bone formation and osteoclastic resorption are thought to be closely coordinated. (1,2) Furthermore, the skeleton participates in plasma calcium homeostasis and supports hematopoiesis. The mobilization of energy (proteins) and minerals is brought about by stimulation of osteoclastic bone resorption. Conversely, resorption decreases when dietary supplies of nutrients increase. (3) The dynamics of bone metabolism is reflected by the plasma 1 Nordic Bioscience, Herlev, Denmark. 2 Center for Clinical and Basic Research, Ballerup, Denmark. 3 Department of Internal Medicine F, Gentofte Hospital, Hellerup, Denmark. 4 Department of Medical Physiology, University of Copenhagen, The Panum Institute, Copenhagen, Denmark. 2180

2 GASTROINTESTINAL HORMONES AND REDUCTION OF BONE RESORPTION 2181 concentrations of bone matrix proteins. The osteoclastic bone resorption can be assessed by measuring fragments derived from the degradation of the C-terminal telopeptide region of collagen type I (s-ctx) in serum. (4) Bone formation can be assessed in serum by measuring the concentration of the intact bone matrix protein, osteocalcin (sosteocalcin), which reflects osteoblast activity. (5) A significant circadian variation has been found for the s-ctx level. (6) The circadian variation for s-ctx has the same magnitude in men and women and is not affected by menopausal status in women. (7) Moreover, it has been shown that skeletal unloading obtained through a 5-day bed rest or weightlessness during space flight does not alter the circadian variation of bone resorption. (8,9) Several hormones, including parathyroid hormone (PTH) and cortisol, also exhibit circadian rhythms and thus could be candidates for mediating the circadian variation in bone turnover. The circadian variation in serum cortisol could be partly responsible for the s-osteocalcin level because it has been demonstrated that the infusion of cortisol in the morning depresses s-osteocalcin production during the day. (10) A similar effect was not observed for bone resorption markers, and elimination of the morning peak of cortisol with metyrapone had no effect on the circadian variation of bone resorption marker excretion. (11) Abolishing the circadian variation of serum PTH by continuous infusion of calcium had no effect on the circadian variation of bone resorption. (12) Furthermore, neutralizing PTH with an anti-pth antibody did not change the pattern of the circadian rhythms of the resorptive activity of serum in vitro. (13) However, the circadian variation of bone resorption as assessed by s-ctx is significantly decreased in fasting individuals. (14) In contrast to s-ctx, the s-osteocalcin level is seemingly not influenced by ingestion of nutrients. (14) We have previously shown that the nutrient-induced decrease in s-ctx concentrations could be explained only partly by insulin released after oral glucose. (15) To investigate the hypothesis that gastrointestinal hormones could be involved, we measured markers for bone turnover as well as the plasma concentrations of the gastrointestinal hormones, glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-2 (GLP- 2), in response to the ingestion of glucose, fat, protein, or fructose. GLP-2 was chosen as a representative of the proglucagon-derived peptides that include enteroglucagon and glucagon-like peptide-1 (GLP-1) and GLP-2 because of its comparatively slower rate of elimination. These peptide hormones are secreted in response to glucose as well as to a mixed meal in normal subjects, and they are key regulatory hormones, which transmit information generated during ingestion of food. (16 21) They belong to the glucagonrelated peptide superfamily because of their sequence homology with glucagon (22) (Fig. 1). In addition, we investigated the effects on bone turnover of parenteral administration of these hormones. Both GIP and GLP-1 are classified as incretin hormones because of their insulinotropic effect. (23) GIP is a 42 amino acid peptide secreted from the K-cells in the duodenum. (18,20) As already alluded to, GLP-1 and GLP-2 are derived from pro-glucagon and correspond to the proglucagon sequences and , respectively. (24) The gene encoding proglucagon is FIG. 1. Amino acid sequences of some of the members of the superfamily of glucagon-related peptides. (Standard single letter abbreviations according to IUPAC-IUB commission on Biochemical Nomenclature.) These peptide hormones are classified within this family based on their considerable sequence homology and are produced in the gut, pancreas, and the central and peripheral nervous system, and exhibit a wide variety of biological actions. Glucagon is released from the pancreas, and the main function is to maintain the blood glucose levels during fasting. GIP is released from the K-cells found in the proximal gut, and GLP-1 and GLP-2 are secreted from the L-cells of the distal gut. GIP, GLP-1, and GLP-2 are released in response to the ingestion of a meal. Residues identical to glucagon sequence in the same position are shaded. expressed in endocrine L-cells of the gut, in the A-cells of the pancreas, and in neurons of the brain stem. In the L-cells, processing of proglucagon gives rise to GLP-1 and GLP-2. The best documented physiological effect of GLP-2 is the stimulation of intestinal epithelial growth in mice and rats, (16) and both GLP-1 and GLP-2 are probably involved in the ileal brake mechanism. (25,26) It has been shown that GLP-2 induces a pronounced decrease in the apoptosis rate in the cells of the intestinal epithelia. (27) In the central nervous system, GLP-2 is synthesized predominantly in the caudal brainstem and hypothalamus, and intracerebroventricular infusion of GLP-2 in rats diminish food intake. (28) GLP-2 effects outside the gastrointestinal system and the brain have not been studied in detail. Because, in a pilot study, subcutaneous injections of GLP-2 were found to significant lower s-ctx levels (data not shown), the emphasis in the present investigation was on the secretion and effects of GLP-2. MATERIALS AND METHODS All studies were approved by the local ethical committee, and written informed consent was obtained for all participants. The studies were conducted according to the principles of the Helsinki Declaration II. Studies of bone turnover during ingestion of macronutrients and resulting endogenous secretion of GIP, GLP-1, and GLP-2 Study 1A bone turnover during oral intake of glucose, protein, and long-chain triglycerides: Ten healthy individuals (six men and four women) between 30 and 40 years of age ( ) with a body mass index (BMI) of kg/m 2 were included in a randomized, controlled cross over trial involving ingestion of glucose, protein, long-chain triglycerides, and control (fasting). The subjects were fasted from 10:00 p.m. before an experiment, and blood samples were collected between 7:30 and 8:30 a.m. Immediately thereafter, oral glucose, protein, or fat were ingested as specified. Blood samples were collected at 30 minutes and 1, 2, 3, 6, and 9 h after ingestion in tubes that were

3 2182 HENRIKSEN ET AL. immediately cooled on ice and centrifuged within 20 minutes at 4 C, and serum/plasma was stored at 20 C until analysis. In the control experiment, the individuals were fasting throughout the experiment. A washout period of 1 week was instituted between the four experiments. Oral glucose consisted of 75 g glucose dissolved in 300 ml water with juice of half a lemon added. The fat consisted of a 70-ml emulsion of long-chained triglycerides, corresponding to 35 g pure vegetable fat, originating from peanuts (Calogen; SHS International Ltd., Liverpool, UK), and was energy equivalent with oral glucose. The protein consisted of 40 g protein powder, corresponding to 35 g pure protein originating from milk (Casilan; HJ Heinz Co. Ltd., Hayes, UK) and was dissolved in 600 ml water and given orally. This product contained less than 1 g carbohydrate and less than 2 g fat per 100 g dry weight. Collected serum samples were analyzed for s-ctx, s-osteocalcin, GIP, and GLP-2. Selected bone turnover data from this study have previously been reported. (15) Study 1B bone turnover during oral intake of fructose: A subgroup consisting of four individuals of the participants in study 1A (all men with a mean age of years and a BMI of kg/m 2 ) participated in a study of oral fructose. After fasting from 10:00 p.m. before the experiment, blood samples were obtained between 8:00 and 9:00 a.m. Immediately thereafter, fructose was ingested, and blood samples were collected at 1 and 2 h after ingestion. Blood was sampled as in study 1A. The fructose meal consisted of 75 g fructose dissolved in 300 ml water with juice of half a lemon added. Collected serum samples were analyzed for s-ctx, s-osteocalcin, GIP, GLP-1, and GLP-2. Study 2 bone turnover after subcutaneous injection of GLP-1: Seven healthy subjects (five men and two women), years of age (mean, years) with a BMI of kg/m 2, participated. All the participants had normal serum creatinine levels ( 130 M). After an overnight fast from 10:00 p.m. before the experiment, the subjects were studied in the recumbent position, with two cannulas inserted into the cubital veins: one for blood sampling and one for glucose infusion. At time 0, GLP-1 was injected subcutaneously (rather than intravenously to prolong the duration of its effect, which is known to disappear rapidly after intravenous injection) into the periumbilical region (1.5 nmol GLP-1/kg body weight). The injected volume was between 1.0 and 1.5 ml. At 15 minutes, plasma glucose was elevated to 15 mm by an intravenous glucose bolus (50% wt/vol) administered within 1 min (calculated as follows: (15 mmol/liter fasting plasma glucose 35 mg glucose weight in kilograms). Venous blood was drawn 15 and 0 minutes before and 10, 20, 30, 40, 50, 60, 70, 80, and 90 minutes after GLP-1 administration. Blood was collected into tubes containing EDTA (6 mm) plus aprotinin (500 KIU/ml blood; Trasylol; Bayer, Leverkusen, Germany) for peptide hormone and s-ctx analyses. Tubes were chilled immediately on ice and centrifuged at 4 C within 30 minutes. Plasma was stored at 20 C until analysis. Plasma GLP-1 results from this study were reported previously. (29) Study 3 bone turnover after intravenous bolus injections of GIP: Eight healthy subjects (six men and two women) between 51 and 70 years of age (mean, years) with a BMI of kg/m 2 participated. All the participants had normal serum creatinine levels ( 130 M). The protocol was identical to that of the GLP-1 study above, except that 3 minutes after glucose injection, 7.5 nmol of GIP was injected intravenously as a bolus injection for 2 minutes. Venous blood was sampled 15, 10, and 0 minutes before and 5, 6, 7, 9, 11, 13, 18, 23, 28, 33, and 48 minutes after the intravenous bolus of glucose. Blood was collected into tubes containing EDTA (6 mm) plus aprotinin (500 KIU/ml blood; Trasylol; Bayer) for peptide hormone, s-ctx, and s-osteocalcin analyses. Tubes were chilled immediately on ice and centrifuged at 4 C within 30 minutes. Plasma was stored at 20 C until analysis. Plasma GIP results from this study were reported previously. (30) Study 4 bone turnover after subcutaneous bolus injection of GLP-2: We studied healthy postmenopausal women who had passed natural menopause at least 5 years earlier and did not take any medication known to influence calcium metabolism or gastrointestinal function for at least 6 months before inclusion into the study. All had a blood pressure (at rest) below 160/90 mm Hg and a BMI between 19.3 and 29.7 kg/m 2. At a screening visit performed 2 3 weeks before randomization, all participants underwent a medical program including medical history and current medication, a complete physical examination, general laboratory blood sample (hematology and serum chemistry) and urine sample analyses, blood pressure and pulse, oral temperature, and weight and height to ensure that the participants met the inclusion criteria. A total of 60 women were included and randomly assigned to treatment with either placebo (saline) or GLP-2 100, 200, 400, or 800 g. Allocation was in blocks of five. Placebo (0.5 ml) or GLP-2 ( ml) was given as a subcutaneous injection into the periumbilical region. The study code was blinded for the participants as well as for all persons giving the injection, recording adverse events, and performing physical examinations. One participant in the placebo group did not complete the study, and data from this participant were excluded from the data analysis. The data of all other participants were included in the analysis. On the day of treatment, the women were instructed to report in the morning after an overnight fast from 10:00 p.m. the evening before. Initial blood samples were collected at 8:45 a.m. Injection of GLP-2/placebo was given at 9:00 a.m. (time 0), and blood samples were collected after 15, 30, 60, 90, 120, 180, 240, 360, and 480 minutes. Urine samples were collected at baseline (9:00 a.m.) and at 60 and 180 minutes. During this 480-minute sampling and observation period, the women continued to fast but were allowed to drink water. Blood was collected into tubes containing EDTA (6 mm) for GLP-2, s-ctx, and s-osteocalcin analyses. Tubes were chilled immediately on ice and centrifuged at 4 C within 30 minutes. Blood samples (serum and plasma) and urine samples were stored at 20 C until analysis. Peptides Synthetic GLP-1 (7 36) amide was purchased from Peninsula Europe (Merseyside, UK). Synthetic GIP was from PolyPeptide Laboratories (Wolfenbüttel, Germany). The

4 GASTROINTESTINAL HORMONES AND REDUCTION OF BONE RESORPTION 2183 GLP-2 (1 33) was custom synthesized by PolyPeptide Laboratories. Synthetic human GLP-2, provided as the acetate salt, was mixed with 0.9% (W/V) NaCl, and 0.5% (W/V) ammonia was added until the peptide was dissolved (ph 8 9). The solution was neutralized using 0.1 M acetic acid. All peptide solutions were sterile filtered using a m filter (Millex; Millipore, Bedford, MA, USA), saturated with 20% human albumin (Statens Serum Institut, Copenhagen, Denmark), and washed with sterile water before use. The solution was dispensed into capped vials and heat sterilized for 20 minutes at 100 C. The purity of the peptide preparations was higher than 97%, and the capped vials were analyzed for sterility and kept at 20 C until use. The peptide content and peptide structure was confirmed by high-performance liquid chromatography (HPLC), peptide sequence, MALDI-TOF mass spectrometry, and quantitative amino acid analysis. Analytical methods Bone resorption was assessed by a sandwich assay using two monoclonal antibodies specific for a -aspartate form of the epitope EKAHDGGR derived from the C-terminal telopeptide region of type I collagen (s-ctx; Serum Cross- Laps One Step ELISA; Nordic Bioscience). A complex between C-telopeptide fragments of type I collagen, biotinylated antibody, and peroxidase-conjugated antibody is generated, and the complex binds to the streptavidin surface through the biotin. The amount of bound complex is quantified by the use of a chromogenic peroxidase substrate. The intra- and interassay CVs of the assay are 5.1% and 8.2%, respectively. (4) The bone resorption parameter of urinary excretion of deoxypyridinoline (DPD) cross-links was measured by the Metra DPD EIA kit according to the manufacturer s instruction (Quidel Corp., San Diego, CA, USA). The intra- and interassay imprecisions (CVs) were 5% and 10%, respectively. In the same samples, urine creatinine was measured by routine chemistry analysis for the purpose of normalization. Bone formation was assessed from the concentration of s-osteocalcin (Osteocalcin N-MID ELISA assay; Nordic Bioscience). The assay determines the N-terminal midsegment of the osteocalcin molecule. Intraand interassay CVs of the assay are 3.4% and 6.4%, respectively. (5) All plasma/serum samples were extracted in a final concentration of 70% ethanol before GIP, GLP-1, and GLP-2 measurements to remove unspecific cross-reacting substances. Total GIP was measured using the C-terminally directed antiserum R65, which reacts with the intact GIP (1 42) and the N-terminally truncated metabolite, GIP (3 42). (31) The assay has a detection limit of less than 2 pm and an intra-assay variation of approximately 6%. GLP-2 was measured with a N-terminal specific antiserum code no , measuring only GLP-2 with an intact N terminus, as described elsewhere. (32) For standards, we used recombinant human GLP-2, and the tracer was rat GLP-2 with an Asp 33 3 Tyr 33 substitution, 125 I-labeled using the standard stoichiometric chloramine T method as described elsewhere. (33) This assay has a maximum cross-reaction with synthetic human GLP-2 (3 33) of %. The experimental detection limit was 5 pm, and the intra-assay CV was 5% at a concentration of 40 pm. The recovery of synthetic GLP-2 added to plasma before extraction and assay was 68%. GLP-1 in plasma samples was measured using a radioimmunoassay (RIA) specific for the C terminus of the GLP-1 molecule, using standards of synthetic GLP-1 (7 36) amide and antiserum no (34) The assay cross-reacts less than 0.01% with C-terminally truncated fragments and has a detection limit below 1 pm. Intra-assay and interassay CVs were below 6% and 15%, respectively. Statistical analysis and calculations The biochemical markers of bone resorption and formation, that is, s-ctx and s-osteocalcin were expressed in percent of the baseline value of the individual subject, and the effect of the given intervention compared with control was assessed by two-tailed Student s t-test for unpaired data at each time-point. The area under curves (AUC) in studies 1A and 4 were calculated by the trapezoidal method, and ANOVA was carried out using the General Linear Method Procedure for the AUCs having variance homogeneity (s- CTX and s-osteocalcin) or using NPAR1WAY of the parameters of AUC of GIP and GLP-2. In case of statistical significant differences, the two-tailed Student s t-test for unpaired data was used for assessment of the effect of intervention compared with the control. A multivariate test for repeated measures of variance of s-ctx and s-osteocalcin were carried out by means of the General Linear Model to assess the time course dependency, intervention effect, and the interaction between time and intervention effect. The Statistical Analysis System (SAS Institute, Cary, NC, USA) was used for the analyses. The level chosen to indicate statistical significance was RESULTS Bone turnover and secretion of GIP, GLP-1, and GLP- 2 after ingestion of macronutrients Study 1A bone turnover during oral intake of glucose, protein, and long-chain triglycerides: Oral glucose, triglycerides, and protein resulted in a significant reduction in the s-ctx concentration of 52%, 39%, and 52% from baseline compared with 21% for the fasting control group (glucose and protein, p 0.001; fat, p 0.05; Fig. 2A). AUC 0 3h analyses for s-ctx during the first 3 h show that the glucose and protein groups were significantly different from fasting control (p 0.01 and p 0.001, respectively; Fig. 2C). AUC 0 3h analysis for s-ctx after fat ingestion was not statistically significantly different from fasting control (p 0.08; Fig. 2C). Bone formation as measured by s-osteocalcin was unaffected by the oral glucose and fat and was similar to the fasting control group (Fig. 2B). After ingestion of protein, the s-osteocalcin level was reduced, but the AUC 0 3h analysis for s-osteocalcin was not statistically significantly different for the protein group compared with the fasting control group (AUC 0 3h ; p 0.06; Fig. 2C). The AUC 0 3h analysis for s-osteocalcin did not reveal a difference between groups (ANOVA; p 0.11). The peak serum concentration of GLP-2 after ingestion of glucose, fat, and protein were 37, 38, and 32 pm, respectively (Fig. 3A). The

5 2184 HENRIKSEN ET AL. FIG. 2. Effects of oral glucose, protein, and long-chain triglycerides on bone turnover in healthy individuals. The figure shows results of treatment with glucose, protein, long-chain triglycerides, and fasting control expressed as percentage of baseline for (A) s-ctx, and (B) osteocalcin in healthy men and premenopausal women. (C) AUC 0 3h analysis for s-ctx and osteocalcin responses. Data are presented as mean SE. ANOVA (all) and p values for the individual groups compared with fasting controls are shown: *p 0.05, **p 0.01, and ***p FIG. 3. Effects of oral glucose, protein, long-chain triglycerides, and fasting on serum concentrations of GIP and GLP-2 in healthy individuals. Results of treatment with glucose, protein, long-chain triglycerides, and fasting control on (A) serum GLP-2 and (B) serum GIP in healthy men and premenopausal women are shown. (C) AUC 0 3h analysis for GIP and GLP-2 responses. Data are presented as mean SE. ANOVA (all) and p values for the individual groups compared with fasting controls are shown: *p 0.05, **p 0.01, and ***p peak serum concentration of GLP-2 after fat was delayed, but the peak level was similar to treatments with glucose and protein (Fig. 3A). The peak serum concentration of GIP after ingestion of glucose, fat, and protein were 44, 63, and 31 pm, respectively (Fig. 3B). The GIP response after fat ingestion was elongated compared with the responses seen for the glucose and protein groups. An AUC 0 3h analysis for the GLP-2 and GIP secretions showed a statistically significant difference between the glucose, fat, and protein groups and the fasting group (glucose and fat, p 0.05; protein, p 0.001; Fig. 3C). The GIP AUC 0 3h after ingestion of fat was approximately twice of what was found for the glucose and protein groups (Fig. 3C). Study 1B fructose ingestion: Oral fructose resulted in a decrease in s-ctx concentration of approximately 40% at 2 h after initiation, and this was significantly different from the change observed in control group (ANOVA; p 0.05). Osteocalcin showed no difference between groups (p 0.2). The secretion of GLP-2 was similar to what was seen for oral glucose. The peak concentration at 2 h for GLP-2 was 27 pm. The GIP level did not increase after the oral fructose treatment and remained unchanged throughout the study. Studies of administration of GIP, GLP-1, and GLP-2 and resulting effects on bone turnover Study 2 GLP: The s-ctx level was unaffected by subcutaneous injection of GLP-1 and resulted in a nonsignificant decrease of 20% from baseline (mean, minutes). The intravenous glucose bolus given 15 minutes after the GLP-1 injection to raise the plasma glucose to 15 mm did not affect the s-ctx level (Fig. 4A). The s-osteocalcin level was unaffected (Fig. 4B). The plasma level for GLP-1

6 GASTROINTESTINAL HORMONES AND REDUCTION OF BONE RESORPTION 2185 FIG. 4. Effect of subcutaneous injection of GLP-1 on bone turnover in healthy individuals. Results of subcutaneous injection of GLP-1 expressed as percentage of baseline for (A) s-ctx concentration, (B) s-osteocalcin, and (C) plasma GLP-1 level (expressed as actual values) in healthy men and premenopausal women are shown. Data are presented as mean SE. FIG. 5. Effect of intravenous injection of GIP on bone turnover in healthy individuals. Results of intravenous injection of GIP expressed as percentage of baseline for (A) s-ctx concentration, (B) s-osteocalcin, and (C) plasma GIP level (expressed as actual values) in healthy men and premenopausal women are shown. Data are presented as mean SE. reached 650 pm, approximately 10 times the normal postprandial level (Fig. 4C). Study 3 GIP: The s-ctx level was unaffected by the intravenous injection of GIP and resulted in a nonsignificant decrease of 16% from baseline within the 48 minutes of the experiment (Fig. 5A). The s-osteocalcin level was unaffected (Fig. 5B). The plasma level for GIP reached 775 pm, which was more than 10 times the normal postprandial level (Fig. 5C). Study 4 GLP-2: Baseline characteristics for the study are presented in Table 1. The treatment groups were comparable at baseline, except for some variability in the urine calcium/creatinine. Short-term effect of exogenous GLP-2 on bone turnover in fasting postmenopausal women GLP-2 treatment induced a dose-dependent decrease in s-ctx (Fig. 6A). There was a small reduction in the s-ctx

7 2186 HENRIKSEN ET AL. TABLE 1. DEMOGRAPHICS AND BASELINE CHARACTERISTICS OF THE STUDY POPULATION GLP-2 groups Placebo (saline) (n 11) 100 g (n 12) 200 g (n 12) 400 g (n 12) 800 g (n 12) ANOVA Age (years) Years postmenopause (years) Height (cm) Weight (kg) Serum CTX (ng/ml) Serum osteocalcin (ng/ml) Serum calcium (mmol/liter) Serum phosphate (mg/dl) Urine CTX/creatinine ( g/mmol)* Urine calcium/creatinine* Urine phosphate/creatinine* Mean SD. * Geometric mean SD. level of approximately 10% from baseline through the first 3 h in the placebo group. This reduction was probably related to the normal circadian variation in s-ctx for fasting individuals. (14) However, in all the GLP-2 groups except the 100- g group, a statistically significant (p 0.05) reduction in s-ctx compared with placebo was observed at 3 h after injection, indicating that GLP-2 induces an acute suppression of bone resorption. The effect on the bone resorption was transient, and s-ctx was similar to the placebo group in all treatment groups after 6 h. The bone resorption as assessed by the urinary excretion of deoxypyridinoline (DPD) cross-links and creatinine (for normalization) were measured in urine samples obtained at baseline and 1 and 3 h after GLP-2 (800 g) and placebo (saline) injection. There was a reduction of approximately 25% from baseline in the u-dpd/creatinine level for the 800- g GLP-2 group at 1 h, and this level was sustained at the 3-h sample (Fig. 6A, inset). The reduction in u-dpd/creatinine was statistically significant (p 0.01) compared with placebo at 1 h after injection, indicating that GLP-2 also induces an acute suppression of bone resorption when assessed with this bone resorption parameter. The effect on the u-dpd/creatinine was transient, and the difference between the GLP-2 and the placebo group was no longer significant at 3h(p 0.16). Bone formation measured by s-osteocalcin was unaffected in all GLP-2 treatment groups (p 0.09), and all groups were similar to the placebo control group (Fig. 6B). The plasma level of GLP-2 reflected the dose escalation used in the study (Fig. 6C). The 800- g GLP-2 group reached a peak plasma concentration of 2689 pm, and the GLP-2 concentration was below detection limit in the placebo control group (LDL 25 pm GLP-2 at the degree of dilution to allow for measurement at high concentrations). The GLP-2 plasma concentration increased linearly with the dose. An AUC 0 8h analysis for s-ctx during the 8 h after GLP-2 injection showed a dose-dependent decrease in the placebo subtracted AUC 0 8h (ANOVA; p 0.05; Fig. 7). AUC 0 8h for the 800- g GLP-2 group was statistically significantly different from the placebo group (p 0.05). Furthermore, AUC analysis for s-ctx at 0 1, 0 3, 0 6, and 0 8 h after injection of 200, 400, and 800 g GLP-2 showed an added effect that was statistically significant after3h(p 0.01) for the 800- g group; this was sustained at6and8h(p 0.05). DISCUSSION Circadian variations of the markers studied here, s-ctx and s-osteocalcin, reflect the dynamic nature and the ability of particularly bone resorption to respond to food intake. In healthy individuals, bone resorption is reduced to 50% from baseline after ingestion of 75 g of glucose. In present studies, ingestion of different macronutrients resulted in approximately the same reduction in s-ctx concentrations but in no change in s-osteocalcin levels. The gastrointestinal hormones, GLP-2 and GIP, were secreted within a few minutes after nutritional stimulation. Because GLP-1 and GLP-2 are secreted in parallel and increase in a similar way in plasma after a meal ingestion, GLP-1 concentrations can be assumed to have increases similarly. (32) Both glucose and protein stimulated the release of the gut hormones and at the same time significantly reduced the s-ctx concentration. Long-chain fatty acids, on the other hand, did not reduce the s-ctx to the same level as glucose and protein, but induced the greatest hormone response of all treatments. In a previous study, ingestion of fat or protein had little or no effect on the insulin secretion, (15,35) excluding insulin as a key factor in postprandial regulation of bone resorption in this situation. Furthermore, in experiments involving an insulin clamp technique, induction of hyperinsulinemia gave no significant change in bone turnover from baseline in either s-ctx or s-osteocalcin levels when euglycemia was maintained. (35) This suggests that insulin does not play an important role in mediating the acute effect of feeding on bone resorption. Recently, we have shown that the circadian variation of bone resorption is markedly reduced in fasting individuals. (14) It has been speculated that the effect of feeding on bone resorption is caused by the release of a factor that inhibits bone resorption and to a lesser extent bone formation. (36) This is in agreement with the observation that bone formation, as assessed by the s-osteocalcin level, is unaffected by food intake. (14,37) Osteocalcin is rapidly degraded in circulation, and the half-life in serum is

8 GASTROINTESTINAL HORMONES AND REDUCTION OF BONE RESORPTION 2187 FIG. 7. Accumulated effect of subcutaneous injection of GLP-2 on bone resorption in fasting healthy postmenopausal individuals for the first 8 h after injection given at 9:00 a.m. AUC 0 8h analysis for s-ctx responses after subcutaneous injections of GLP-2 (100, 200, 400, or 800 g). Data are presented as placebo subtracted AUC; mean SE. ANOVA (all) and p values for the individual treatment groups compared with saline controls are shown: *p FIG. 6. Effect of subcutaneous injection of GLP-2 on bone resorption, osteocalcin, and GLP-2 levels in fasting healthy postmenopausal individuals for the time period up to 8 h after injection given at 9:00 a.m. Results of subcutaneous injection of GLP-2 in four different dosages (100, ; 200, Œ; 400, ;or800 g, f) or placebo (F; saline) are shown. (A) s-ctx concentrations were expressed as percentage of baseline. Repeated measures ANOVA, p ; p values for the individual treatment groups compared with saline controls were p 0.10 (100 g GLP-2), p (200 g GLP-2), p 0.08 (400 g GLP-2), and p (800 g GLP-2). (Inset) u-dpd concentrations (creatinine corrected) in the first 3 h after injection for the 800- g GLP-2 and placebo groups, p 0.01 compared with placebo at 1 h. Values were expressed as percentage of baseline; mean SE. (B) s-osteocalcin concentrations were expressed as percentage of baseline. Repeated measures ANOVA, p 0.09; p values for the individual treatment groups compared with saline controls were p 0.29 (100 g GLP-2), p 0.44 (200 g GLP-2), p 0.16 (400 g GLP-2), and p 0.48 (800 g GLP-2). (C) Plasma levels of GLP-2 after subcutaneous only 4 minutes. (38) Thus, even in studies of short duration, any change in bone formation would be reflected in the s-osteocalcin level. Whereas GIP and GLP-2 responses to glucose, fat, and protein were qualitatively similar, fructose did not result in GIP secretion in this study. This is in agreement with several previous studies. (39 43) On the other hand, we observed a significant secretion of GLP-2 after oral fructose ingestion, which was associated with a significant reduction in the s-ctx level of 40% from baseline that was significantly different from the reduction of 19% of the placebo group (p 0.04). A similar effect of fructose on GLP-1 secretion in humans has previously been described. (44) Intravenous bolus injection of GIP did not affect the s-ctx concentration, but because the study did not include a placebo control group, the significance of this result is difficult to evaluate. However, this was approximately identical to that observed in the fasting situation. Postprandial release of GIP in healthy individuals will result in a peak plasma concentration in the range of pm (45) compared with 775 pm after the intravenous bolus of 7.5 nmol GIP in the present study. Similarly, the subcutaneous injection of GLP-1 did not significantly affect the s-ctx level, and the reduction in the s-ctx level was comparable with levels seen in healthy fasting individuals. Again, because the study did not include a placebo control group, the significance of this result is unknown. The peak plasma concentration of GLP-1 after injection was 650 pm, which injection of GLP-2 in fasting healthy postmenopausal individuals for the time period up to 8 h after injection given at 9:00 a.m. Results of subcutaneous injections of GLP-2 (100, 200, 400, or 800 g) or placebo (saline) expressed as actual values in healthy fasting postmenopausal women are shown. Data are presented as placebo corrected; mean SE.

9 2188 HENRIKSEN ET AL. was at least 10 times greater than the peak plasma concentration of pm normally observed after a meal. (34) Thus, these results imply that GIP and GLP-1 are of less importance as mediators of the postprandial reduction of bone resorption. In contrast, subcutaneous injection of GLP-2 resulted in a significant reduction of s-ctx compared with the fasting individuals. This was seen in a pilot experiment (data not shown), and the decrease observed for s-ctx was acute and showed a similar response rate that was seen after ingestion of a meal. To further investigate the effect of GLP-2 on bone turnover, we conducted a dose escalation study in healthy fasting postmenopausal women and the effect was assessed by s-ctx and s-osteocalcin. In vivo studies in humans and pigs have further demonstrated that GLP-2 infusion acutely inhibits gastric secretion and motility, and GLP-2 is believed to be implicated in the ileal brake phenomenon. (24,46) The ileal brake mechanism is an endocrine mechanism that is activated by the presence of nutrients in the ileal lumen, which serves to inhibit gastric motility and secretion. It has been speculated that the hormones of the ileal brake may also participate in the regulation of energy intake in humans. (47,48) Thus, GLP-2 has been examined as a potential therapeutic treatment of patients with short-bowel syndrome. (49) Inarecently published study in short-bowel patients, it was found that treatment with GLP-2 for 5 weeks resulted in a moderate increase in bone density of the spine ( %; p 0.05) and the hip ( ; p 0.06). (50) The beneficial effect of GLP-2 on bone mass remains unexplained and was speculated to be related to an increased mineralization of bone matrix resulting from improved intestinal calcium absorption by the investigators. The intestinal calcium absorption increased by 2.7% (p 0.87), which was not explained by bone mineral density (BMD) increases. (50) In this study, we found that the effect of GLP-2 on bone resorption was dose-dependent, and that bone formation, as assessed by osteocalcin, was unaffected by the exogenous GLP-2 treatment. The dose-dependent reduction in s-ctx levels after subcutaneous GLP-2 was obvious both from the time course (Fig. 6A) and the placebo subtracted AUC 0 8h analysis for the different doses of GLP-2 (Fig. 7). The acute reduction was also found for u-dpd/creatinine after injection of GLP-2, and this marker of bone resorption showed changes of approximately the same magnitude and time course as those found for s-ctx (Fig. 6A, inset). The s-osteocalcin level was unaffected by the treatment, and there was no statistical significant difference between the GLP-2 treated and placebo-treated groups for s-osteocalcin. The physiological significance of these findings remains to be clarified. However, the gastrointestinal hormone, GLP-2, could be involved in an entero-osseous axis, which coordinates bone resorption in response to nutrient intake. Recently, the GLP-2 receptor location in the intestine has been pinpointed to the myenteric ganglia, indicating that afferent nerve fibers might play a role in the regulatory pathway of GLP-2 signaling in the gastrointestinal system. (51) The surprising acute effect of GLP-2 on bone resorption could indicate a direct action of the hormone on bone cells (i.e., osteoclasts and osteoblasts) or an induction of local growth or differentiation factors (i.e., cytokines), which in turn, inhibits osteoclast function. Alternatively, activation of afferent nerve fibers through the receptors in the myenteric plexus (51) could provide a neuronal signal to the bone turnover processes and perhaps explain the rapidity of the onset of the effect. Regardless of the mechanism, we found a dose-dependent effect of exogenous GLP-2 on the bone resorption process, which does not affect the immediate bone formation process as assessed by the osteocalcin level. These present findings warrant further investigation as to the mechanism and relevance of GLP-2 in the bone turnover processes. ACKNOWLEDGMENTS We gratefully acknowledge the excellent technical assistance provided by Pia M Hansen and Lone Bagger. Studies 2 and 3 were supported by The Danish Medical Research Council and the Novo Nordisk Foundation. REFERENCES 1. Parfitt AM 1994 Osteonal and hemi-osteonal remodeling: The spatial and temporal framework for signal traffic in adult human bone. J Cell Biochem 55: Manolagas SC 2000 Birth and death of bone cells: Basic regulatory mechanisms and implications for the pathogenesis and treatment of osteoporosis. Endocr Rev 21: Parfitt AM 2002 Targeted and nontargeted bone remodeling: Relationship to basic multicellular unit origination and progression. Bone 30: Rosenquist C, Fledelius C, Christgau S, Pedersen BJ, Bonde M, Qvist P, Christiansen C 1998 Serum CrossLaps One Step ELISA. First application of monoclonal antibodies for measurement in serum of bone-related degradation products from C-terminal telopeptides of type I collagen. Clin Chem 44: Rosenquist C, Qvist P, Bjarnason N, Christiansen C 1995 Measurement of a more stable region of osteocalcin in serum by ELISA with two monoclonal antibodies. Clin Chem 41: Wichers M, Schmidt E, Bidlingmaier F, Klingmuller D 1999 Diurnal rhythm of CrossLaps in human serum. Clin Chem 45: Qvist P, Christgau S, Pedersen BJ, Schlemmer A, Christiansen C 2002 Circadian variation in the serum concentration of C-terminal telopeptide of type I collagen (serum CTx): Effects of gender, age, menopausal status, posture, daylight, serum cortisol, and fasting. Bone 31: Schlemmer A, Hassager C, Pedersen BJ, Christiansen C 1994 Posture, age, menopause, and osteopenia do not influence the circadian variation in the urinary excretion of pyridinium crosslinks. J Bone Miner Res 9: Caillot-Augusseau A, Vico L, Heer M, Voroviev D, Souberbielle JC, Zitterman A, Alexandre C, Lafage-Proust MH 2000 Space flight is associated with rapid decreases of undercarboxylated osteocalcin and increases of markers of bone resorption without changes in their circadian variation: Observations in two cosmonauts. Clin Chem 46: Nielsen HK, Laurberg P, Brixen K, Mosekilde L 1991 Relations between diurnal variations in serum osteocalcin, cortisol, parathyroid hormone, and ionized calcium in normal individuals. Acta Endocrinol 124: Heshmati HM, Riggs BL, Burritt MF, McAlister CA, Wollan PC, Khosla S 1998 Effects of the circadian variation in serum cortisol on markers of bone turnover and calcium homeostasis in normal postmenopausal women. J Clin Endocrinol Metab 83: Ledger GA, Burritt MF, Kao PC, O Fallon WM, Riggs BL, Khosla S 1995 Role of parathyroid hormone in mediating nocturnal and age-related increases in bone resorption. J Clin Endocrinol Metab 80:

10 GASTROINTESTINAL HORMONES AND REDUCTION OF BONE RESORPTION Lakatos P, Blumsohn A, Eastell R, Tarjan G, Shinoda H, Stern PH 1995 Circadian rhythm of in vitro bone-resorbing activity in human serum. J Clin Endocrinol Metab 80: Christgau S, Bitsch-Jensen O, Bjarnason NH, Henriksen EG, Qvist P, Alexandersen P, Henriksen DB 2000 Serum CrossLaps for monitoring the response in individuals undergoing antiresorptive therapy. Bone 26: Bjarnason NH, Henriksen EE, Alexandersen P, Christgau S, Henriksen DB, Christiansen C 2002 Mechanism of circadian variation in bone resorption. Bone 30: Drucker DJ, Erlich P, Asa SL, Brubaker PL 1996 Induction of intestinal epithelial proliferation by glucagon-like peptide 2. Proc Natl Acad Sci USA 93: Elliott RM, Morgan LM, Tredger JA, Deacon S, Wright J, Marks V 1993 Glucagon-like peptide-1 (7 36)amide and glucosedependent insulinotropic polypeptide secretion in response to nutrient ingestion in man: Acute post-prandial and 24-h secretion patterns. J Endocrinol 138: Fehmann HC, Goke R, Goke B 1995 Cell and molecular biology of the incretin hormones glucagon-like peptide-i and glucosedependent insulin releasing polypeptide. Endocr Rev 16: Holst JJ 2000 Gut hormones as pharmaceuticals. From enteroglucagon to GLP-1 and GLP-2. Regul Pept 93: Meier JJ, Nauck MA, Schmidt WE, Gallwitz B 2002 Gastric inhibitory polypeptide: The neglected incretin revisited. Regul Pept 107: Xiao Q, Boushey RP, Drucker DJ, Brubaker PL 1999 Secretion of the intestinotropic hormone glucagon-like peptide 2 is differentially regulated by nutrients in humans. Gastroenterology 117: Kieffer TJ, Habener JF 1999 The glucagon-like peptides. Endocr Rev 20: Holst JJ 1999 Glucagon-like peptide 1 (GLP-1): An intestinal hormone, signalling nutritional abundance, with an unusual therapeutic potential. Trends Endocrinol Metab 10: Holst JJ 1997 Enteroglucagon. Annu Rev Physiol 59: Hartmann B, Thulesen J, Kissow H, Thulesen S, Orskov C, Ropke C, Poulsen SS, Holst JJ 2000 Dipeptidyl peptidase IV inhibition enhances the intestinotrophic effect of glucagon-like peptide-2 in rats and mice. Endocrinology 141: Wojdemann M, Wettergren A, Hartmann B, Hilsted L, Holst JJ 1999 Inhibition of sham feeding-stimulated human gastric acid secretion by glucagon-like peptide-2. J Clin Endocrinol Metab 84: Lovshin J, Drucker DJ 2000 New frontiers in the biology of GLP-2. Regul Pept 90: Tang-Christensen M, Larsen PJ, Thulesen J, Romer J, Vrang N 2000 The proglucagon-derived peptide, glucagon-like peptide-2, is a neurotransmitter involved in the regulation of food intake. Nat Med 6: Vilsboll T, Krarup T, Madsbad S, Holst JJ 2001 No reactive hypoglycaemia in Type 2 diabetic patients after subcutaneous administration of GLP-1 and intravenous glucose. Diabet Med 18: Vilsboll T, Krarup T, Madsbad S, Holst JJ 2002 Defective amplification of the late phase insulin response to glucose by GIP in obese Type II diabetic patients. Diabetologia 45: Krarup T, Holst JJ 1984 The heterogeneity of gastric inhibitory polypeptide in porcine and human gastrointestinal mucosa evaluated with five different antisera. Regul Pept 9: Hartmann B, Johnsen AH, Orskov C, Adelhorst K, Thim L, Holst JJ 2000 Structure, measurement, and secretion of human glucagonlike peptide-2. Peptides 21: Holst JJ, Bersani M 1991 Assays for peptide products of somatostatin gene expression. Methods Neurosci 5: Orskov C, Rabenhoj L, Wettergren A, Kofod H, Holst JJ 1994 Tissue and plasma concentrations of amidated and glycineextended glucagon-like peptide I in humans. Diabetes 43: Clowes JA, Robinson RT, Heller SR, Eastell R, Blumsohn A 2002 Acute changes of bone turnover and PTH induced by insulin and glucose: Euglycemic and hypoglycemic hyperinsulinemic clamp studies. J Clin Endocrinol Metab 87: Clowes JA, Hannon RA, Yap TS, Hoyle NR, Blumsohn A, Eastell R 2002 Effect of feeding on bone turnover markers and its impact on biological variability of measurements. Bone 30: Schlemmer A, Hassager C 1999 Acute fasting diminishes the circadian rhythm of biochemical markers of bone resorption. Eur J Endocrinol 140: Price PA, Williamson MK, Lothringer JW 1981 Origin of the vitamin K-dependent bone protein found in plasma and its clearance by kidney and bone. J Biol Chem 256: Vozzo R, Baker B, Wittert GA, Wishart JM, Morris H, Horowitz M, Chapman I 2002 Glycemic, hormone, and appetite responses to monosaccharide ingestion in patients with type 2 diabetes. Metabolism 51: Ganda OP, Soeldner JS, Gleason RE, Cleator IG, Reynolds C 1979 Metabolic effects of glucose, mannose, galactose, and fructose in man. J Clin Endocrinol Metab 49: Sirinek KR, Levine BA, O Dorisio TM, Cataland S 1983 Gastric inhibitory polypeptide (GIP) release by actively transported, structurally similar carbohydrates. Proc Soc Exp Biol Med 173: Williams RH, May JM, Biesbroeck JB 1981 Determinants of gastric inhibitory polypeptide and insulin secretion. Metabolism 30: Sykes S, Morgan LM, English J, Marks V 1980 Evidence for preferential stimulation of gastric inhibitory polypeptide secretion in the rat by actively transported carbohydrates and their analogues. J Endocrinol 85: Kong MF, Chapman I, Goble E, Wishart J, Wittert G, Morris H, Horowitz M 1999 Effects of oral fructose and glucose on plasma GLP-1 and appetite in normal subjects. Peptides 20: Vilsboll T, Krarup T, Deacon CF, Madsbad S, Holst JJ 2001 Reduced postprandial concentrations of intact biologically active glucagon-like peptide 1 in type 2 diabetic patients. Diabetes 50: Wojdemann M, Wettergren A, Hartmann B, Holst JJ 1998 Glucagon-like peptide-2 inhibits centrally induced antral motility in pigs. Scand J Gastroenterol 33: Drucker DJ 2001 Glucagon-like peptide 2. J Clin Endocrinol Metab 86: Lavin JH, Wittert G, Sun WM, Horowitz M, Morley JE, Read NW 1996 Appetite regulation by carbohydrate: Role of blood glucose and gastrointestinal hormones. Am J Physiol 271:E209 E Jeppesen PB, Hartmann B, Thulesen J, Graff J, Lohmann J, Hansen BS, Tofteng F, Poulsen SS, Madsen JL, Holst JJ, Mortensen PB 2001 Glucagon-like peptide 2 improves nutrient absorption and nutritional status in short-bowel patients with no colon. Gastroenterology 120: Haderslev KV, Jeppesen PB, Hartmann B, Thulesen J, Sorensen HA, Graff J, Hansen BS, Tofteng F, Poulsen SS, Madsen JL, Holst JJ, Staun M, Mortensen PB 2002 Short-term administration of glucagon-like peptide-2. Effects on bone mineral density and markers of bone turnover in short-bowel patients with no colon. Scand J Gastroenterol 37: Bjerknes M, Cheng H 2001 Modulation of specific intestinal epithelial progenitors by enteric neurons. Proc Natl Acad Sci USA 98: Address reprint requests to: Dennis B Henriksen, MSc, PhD Herlev Hovedgade 207 DK-2730 Herlev, Denmark dbh@nordicbioscience.com Received in original form March 13, 2003; in revised form July 9, 2003; accepted August 1, 2003.