Update on the Acute Effects of Glucose, Insulin, and Incretins on Bone Turnover In Vivo

Abstract

Purpose of Review

To provide an update on the acute effects of glucose, insulin, and incretins on markers of bone turnover in those with and without diabetes.

Recent Findings

Bone resorption is suppressed acutely in response to glucose and insulin challenges in both healthy subjects and patients with diabetes. The suppression is stronger with oral glucose compared with intravenous delivery. Stronger responses with oral glucose may be related to incretin effects on insulin secretion or from a direct effect on bone turnover. Glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-2 (GLP-2) infusion acutely suppresses bone resorption without much effect on bone formation. The bone turnover response to a metabolic challenge may be attenuated in type 2 diabetes, but this is an understudied area. A knowledge gap exists regarding bone turnover responses to a metabolic challenge in type 1 diabetes.

Summary

The gut-pancreas-bone link is potentially an endocrine axis. This linkage is disrupted in diabetes, but the mechanism and progression of this disruption are not understood.

Introduction

There has been a resurgence in interest in understanding the role of energy metabolism on bone turnover in the past 15 years, as bone remodeling is an energetically expensive process. Ensuring that osteoblasts, osteocytes, and osteoclasts can siphon the nutrients to generate ATP likely requires that bone communicates its energetic state with other tissues. In line with this notion is an increased appreciation of the existence of a feedback loop between the pancreas and the skeleton, which, in part, includes feed forward communication via insulin and osteocalcin [1]. Glucose is likely, and logically, a mediating factor in this feedback loop. An important characteristic of a feedback loop is the ability to quickly adapt to and regulate changing stimuli. We know that the sensitivity for insulin to stimulate glucose uptake or inhibit lipolysis decreases under conditions of chronic sedentary behavior or overnutrition. This, in turn, suggests the question of whether the sensitivity of the pancreas-bone feedback loop to changes in insulin, osteocalcin, or other potential GI or bone-derived hormones can equally be affected after chronic or multiple acute perturbations. Emerging influencers of the pancreas-bone axis are incretins, gut peptides that are secreted after a meal. We searched the most recent published literature (in the last 5 years) and summarized the findings of these studies to provide a concise review of the effect of glucose, insulin, and incretins (glucagon-like peptide-1 and −2 (GLP-1, GLP-2) and gastric inhibitory polypeptide (GIP)) on bone turnover in people with type 1 and type 2 diabetes in comparison with people without diabetes. We have also provided our opinion on what is unknown in this field that needs to be studied in future to improve our understanding of the gut-bone-pancreas axis. Our search strategy was as follows: “GLP-1, GIP, incretin, GLP-2, GLP” AND “bone turnover,” “type 1 diabetes,” “type 2 diabetes” (separate searches) AND “bone” AND “glucose tolerance” or “insulin.” We also searched “bone” AND “OGTT” (oral glucose tolerance test) or “hyperinsulinemic euglycemic clamp” when not including T1D or T2D. Abstracts of published original articles in the last 5 years were reviewed. Articles that focused on children or adolescents were excluded.

Effects of Insulin and Glucose on Bone

Improved diabetes care and declining mortality from acute complications in both type 1 and type 2 diabetes (T1D and T2D) have resulted in increased longevity in people with T1D and T2D. This increased longevity necessitates a greater understanding of the effect of diabetes on bone. There is mounting evidence that T1D and T2D are associated with an increased risk of fracture, and this has been reviewed elsewhere [2, 3]. In T2D, this is despite having normal or higher than average bone mass. Mechanisms by which diabetes would increase fracture risk are an active area of research. A commonly studied culprit is chronic hyperglycemia. Studies reporting significant associations between chronic hyperglycemia as measured by HbA1c and suppressed resting bone turnover markers (BTMs) are now numerous [1, 4–6]. However, most studies used only one HbA1c measurement, which represents only 3 months of glucose control. Further, the timing of glycemic exposure relative to bone maturity may affect lifelong bone health, at least in T1D [7•]. Cross-sectional comparisons of BTMs have shown that bone turnover is suppressed in both T1D and T2D compared with healthy controls, and this has been previously reviewed [8, 9].

Transitioning from association to causation requires a more hypothesis-driven approach. One such approach is to acutely manipulate blood glucose and/or insulin and test the acute BTM response. Methods of assessing the acute bone turnover response to changes in glucose homeostasis most often include an oral glucose tolerance test (OGTT) because of the relative ease of administering the test, as opposed to a hyperinsulinemic euglycemic clamp (HEC). Presumably, in part due to costs, these studies tend to utilize much smaller sample sizes (e.g., < 20/group), thus requiring a larger effect size to detect a difference. Most recent studies (past 3–5 years) using these approaches to test the BTM response have focused on middle-aged and older adults. For the pancreas-bone axis to become a feedback loop in humans as proposed based on preclinical evidence in mice, there would be a reciprocal regulation of glucose homeostasis by bone. Two studies indicated improved glucose tolerance in women with alendronate, but not denosumab [10, 11]. Although, limited evidence suggests that the incidence of T2D may be reduced with bisphosphonates [12].

The most commonly tested BTMs in the past 5 years include alkaline phosphatase (ALP), total procollagen type 1 N-terminal propeptide (P1NP), osteocalcin, and C-terminal telopeptides of type I collagen (CTX-I). However, these factors have not been proposed to have endocrine functions. Total osteocalcin was classically used as a bone formation marker, until the role of its undercarboxylated form in promoting both insulin secretion and glucose uptake into muscle became more appreciated [13]. Undercarboxylated osteocalcin levels have recently been demonstrated to be lower in those who have metabolic syndrome and later develop T2D [14]. Sclerostin, a well-known inhibitor of bone formation, is becoming more popularly cited, as some preclinical evidence suggests that sclerostin promotes adipogenesis, and knocking out sclerostin improves glucose tolerance and uptake into muscle [15].

Both bone formation (P1NP) and resorption (CTX) markers acutely decreased in response to an OGTT and an intravenous (IV) glucose infusion in healthy middle-aged individuals [16•]. In another study, middle-aged individuals who underwent a 2-h HEC decreased total and undercarboxylated osteocalcin (ucOC) and TRACP5b, but those who underwent 4 h of clamp decreased TRACP5b, and CTX. Increasing the insulin dose for 4 h resulted in decreased ucOC and CTX [17]. When using an intravenous glucose tolerance test (IVGTT), P1NP decreased by ~ 8% within 15 min but returned to baseline by 60 min, whereas CTX linearly declined to ~ 20% decrease [18]. Acute suppression of markers of bone formation and resorption occurs after a meal, oral glucose, and intravenous glucose, indicating that either glucose or endogenous insulin suppresses bone turnover [19]. The magnitude of decrease does appear to depend on the mode of glucose delivery, as OGTT induces a greater suppression of CTX than isoglycemic intravenous glucose infusion. This raises the possibility of a gut hormone contribution to the suppression of bone turnover [20•]. In that study, however, P1NP was not significantly decreased in either condition. In a subsequent study in older men with normal insulin sensitivity and insulin resistance, both P1NP and CTX were suppressed in response to both a meal and an OGTT [21].

There has been a general paucity of studies investigating the pancreas-bone axis in T2D in the past 5 years, and no studies test the bone turnover response to a hyperinsulinemic euglycemic clamp in T1D. In middle-aged women and men with T2D and/or nonalcoholic fatty liver disease (NAFLD), CTX decreased in response to an OGTT and isoglycemic intravenous glucose infusion (IIGI) but to a lesser extent than their healthy counterparts. In healthy individuals, CTX was suppressed to a greater extent with OGTT than with IIGI, and P1NP was suppressed to a similar extent with OGTT and IIGI. In T2D, however, the suppression of CTX was similar between OGTT and IIGI, whereas P1NP was suppressed to a greater extent with IIGI [16•]. This was also observed in postmenopausal women, where decreases in CTX in response to a mixed meal test were lower in women with T2D, whereas OC decreased to a similar extent [22]. Additionally, CTX and OC decreased (7–9%) in response to a HEC in those who were lean or obese insulin-sensitive, but not in those who were insulin-resistant or had T2D. P1NP did not change in any group [23•]. Hyperglycemia is thought to suppress bone turnover in young people with T1D [24, 25]. Further, chronically low levels of insulin (low hepatic insulin concentration compared with peripheral insulin concentration and thus low IGF-1) or frequent hyperglycemia may each slow bone turnover or accrual [26–28]. More work is clearly needed in understanding the pancreas-bone axis in adults with T1D, as it will be important to discover strategies that can be taken to promote healthy remodeling.

In summary, oral and intravenous glucose acutely suppresses bone formation and resorption and suppression is more pronounced with oral glucose in healthy individuals. Acute suppression of bone resorption is observed in people with T2D but to a lesser extent than in people without diabetes, but findings on the effect on bone formation markers are conflicting. In healthy individuals and patients with early T2D, there was no effect of acute insulin on bone turnover during normoglycemic conditions, but small sample size and methodological limitations of these studies make it difficult to interpret these findings. As of this writing, there are no studies testing the bone turnover in response to hyperinsulinemic euglycemic clamp in T1D. The effect of glucose and insulin on bone turnover is illustrated in Fig. 1.

Fig. 1.

Pathways influencing acute bone turnover. Both glucose and insulin may be acutely decreasing bone resorption, whereas effects on bone formation are more varied. Incretins can acutely influence bone turnover, either directly or by increasing insulin production. Intravenous glucose and exogenous insulin have similar effects, but the effects of oral glucose on bone turnover are more robust. The effects of changing glucose or insulin on BTM are attenuated in type 2 diabetes, but whether the effects are changed in type 1 diabetes is less known

Effect of Incretin on Bone Turnover

If glucose regulation is an important underlying goal in the bone-pancreas feedback loop, then discussion of incretins and their role in insulin stimulation and bone turnover becomes critical. As a historical overview, gut extract was demonstrated to reduce urine glucose levels over a century ago, presumably by stimulating insulin secretion [29]. The term “incretin” was coined by La Barre in 1929, when La Barre purified the glucose-lowering fraction from gut and named it “incretin” [30]. The “incretin effect” can be described as a phenomenon in which oral glucose leads to greater secretion of insulin compared with intravenous glucose even when the same plasma glucose levels are achieved. Two gut hormones, glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1), are implicated in the incretin phenomenon. GIP is secreted from intestinal K cells and GLP-1 is cosecreted with GLP-2 from intestinal L cells. Their glucose-lowering effects are extensively studied and GLP-1 is now exploited pharmacologically in diabetes management. GLP-2 is not currently implicated in the incretin phenomenon but does appear to acutely affect bone turnover.

A GIP-induced incretin effect is observed in heathy individuals, but GIP does not reduce glucose in patients with T2D. As a result, GIP treatment is not considered an option for management of T2D [31]. However, GIP does appear to have a positive effect on bone quality in rats and reduces bone resorption in humans [32]. Current evidence in humans suggests that short-term GIP treatment results in suppression of bone resorption in healthy adults, possibly through increased insulin secretion. Nissen et al. [33] investigated the short-term effect of GIP infusion during euglycemic and hyperglycemic conditions on bone turnover markers in 10 healthy volunteers. GIP infusion resulted in a suppression of CTX compared with saline infusion and CTX suppression was augmented during hyperglycemic conditions. The authors did not provide information on changes in bone formation markers with GIP treatment. In another experiment, 12 healthy men underwent an OGTT, isoglycemic intravenous glucose infusion (IIGI), and control conditions after 8 h of fasting on three separate occasions [20•]. OGTT induced 50% suppression of CTX, greater than IIGI and controls. Peak GIP concentration explained 34% of the variability in nadir CTX suggesting that GIP in combination with hyperglycemia may have an acute effect on bone resorption suppression in healthy men. It does appear that GIP has a differential effect on the acute bone turnover response compared with GLP-2. Both appear to substantially suppress CTX (> 50%), although GIP did so more rapidly. However, GIP increased P1NP, whereas GLP-2 transiently decreased PINP (− 12%). Osteocalcin significantly increased ~ 3% with GIP, but then decreased by 6%, whereas GLP-2 only caused a decrease [34•].

To understand the effect of GIP on BTM, GIP was infused in men with and without T1D during insulin-induced hypoglycemia and during hyperglycemia [35••]. GIP progressively suppressed CTX (− 59%) in both conditions and both populations, without inducing an absolute increase in P1NP. The relative (~ 12%) increase in the hypoglycemia condition, however, was significant. The result of this study suggests that GIP may effectively suppress bone resorption in T1D and healthy individuals. Thus far, possible evidence that GIP affects BTMs in response to an OGTT or infused glucose similarly in T2D to healthy controls has been presented in abstract form and mentioned in a recent review, but is not yet published in full form [36].

Though GLP-2 is not currently implicated in the incretin effect, we review it here as it does appear to influence bone resorption. Interestingly, GLP-2 receptor (GLP-2R) has not yet been identified in human bone cells, but has been found in immature human osteoblast cell lines MG-63 and TE-85 [37]. Most studies in patients with short-bowel syndrome demonstrate an increase in bone mineral density with the use of a GLP-2 analog [38–41]. Recent investigation suggested suppression of CTX in postmenopausal women after a single night time dose of GLP-2 [42]. This short study was followed by a 14-day study in which 60 postmenopausal women received either 1.6 or 3.2 mg of parenteral GLP-2 versus a saline control [41]. Compared with saline control, both 1.6 mg and 3.2 mg of GLP-2 resulted in greater suppression of CTX while P1NP remained unaffected.

Many GLP-1 receptor (GLP-1R) analogs are currently approved for management of T2D. There has been increasing use of GLP-1R agonists and dipeptidyl peptidase 4 inhibitors (DPP-4 enzyme cleaves and inactivates GLP-1) in the management of T2D. This has led to increasing interest in understanding the effect of GLP-1R analogs on bone health in hopes of reducing the fracture risk in T2D population. GLP-1 treatment resulted in reduced bone loss in ovariectomized and diabetes mouse models [43, 44]. Studies and meta-analyses indicate no increased risk for fractures with GLP-1R agonist or DPP-4 inhibitors in patients with T2D [45–47]. Recently, injected GLP-1 was shown to acutely decrease CTX in healthy young men and augmented the decrease in CTX in response to and isoglycemic glucose infusion in overweight men, but it did not increase P1NP [48•, 49].

In summary, acute infusion of GIP or GLP-2 suppresses bone resorption (CTX) without much effect on bone formation (P1NP). Long-term effects of GIP and GLP-2R analogs on bone turnover, bone density, and fracture risk are currently unknown. GLP-1R analogs and DPP-4i treatment in patients with T2D did not appear to increase fracture risk but their effect on fracture prevention is unknown. Effects of incretin on bone turnover are illustrated in Fig. 1.

Significant Knowledge Gap About Why Diabetes Impairs the Acute BTM Response to Insulin or Glucose

An important gap in knowledge is in understanding the transition into the attenuated bone turnover response to changes in glucose or insulin in people with diabetes. Much of what we know about the direct effects of high glucose or insulin on signaling cascades in osteoblast, osteoclast, or osteocytes are from in vitro experiments, and these do provide important clues. However, this removes the context of nutrient trafficking, incretin production, and glucose uptake by other tissues. When an OGTT or a HEC is performed, a portion of the glucose will be taken up by the skeleton, but a very large proportion is taken up by skeletal muscle in healthy conditions. What is unclear is whether impaired glucose uptake in skeletal muscle is also indicative of impaired glucose uptake in the skeleton, or if impaired glucose uptake in bone is downstream, timewise, of hyperglycemia induced by impaired glucose uptake by muscle. In either case, incretin responses to a meal are also impaired in T2D. Future work is needed to understand the time course of insulin resistance in muscle versus insulin resistance in bone, and what role incretin responses to food play in insulin resistance in bone.

Significant Knowledge Gap About Sex Differences in BTM Response

It is well-established that there are large sex differences in fracture rates across the lifespan [50, 51]. Boys have higher fracture rates than girls during puberty, which is not entirely explained by differences in lifestyle, but fracture rates are higher in postmenopausal women than in men. Women and men with T2D both have an increased risk of fracture, but men have higher rates of T2D, possibly due to sex differences in the accumulation of visceral fat [2, 52–54]. Sex has been observed as a significant covariate in the association between glycemic control and fracture risk, but we do not yet know if sex is an effect modifier of the T2D-fracture relationship [55–57]. Thus far, T2D status has been observed to affect bone parameters and fracture risk similarly among both sexes [55–57]. T1D prevalence is also higher in men than in women, while T1D may affect BMD disproportionately in women [58, 59]. In one larger study, the effect of sex on the increased fracture risk with T1D was age-dependent, but a meta-analysis indicated that T1D disproportionately increases fracture risk in women [60, 61]. Although most of the recent studies that test the acute bone turnover response to a glucose or insulinemic challenge enroll both women and men, the vast majority do not enroll enough of both sexes to compare the BTM response between sexes. Furthermore, most studies also collapse the results in women and men, preventing attempts at power calculations to determine the sample size needed to detect a sex difference in the BTM response to changes in glucose or insulin. Given that sex appears to modify the T1D-BMD relationship, future studies are needed to test whether sex differences in the BTM response to hyperglycemia or hyperinsulinemia explain sex differences in fracture rates and absolute fracture risk.

Significant Knowledge Gap About Incretin Effects on Bone

The acute effects of incretins on bone may not matter so much if they do not affect long-term bone health. Despite the fact that GLP-1R analogs and DPP-4i showed initial promise in reducing fracture risk, most studies have not confirmed this. No studies with fracture as the primary end point have been done with GLP-1R analogs or dipeptidyl peptidase 4 inhibitors. Therefore, we have been limited to secondary analyses to document fracture outcomes. One such study included a prospective cohort study of 1686 postmenopausal women from Denmark suggesting an association between a functional GIP receptor polymorphism and fracture risk over 10 years [62]. No long-term effect of GIP or GLP-2 has been studied in T1D or T2D. However, meta-analyses thus far have suggested no increased risk for fractures with GLP-R agonist or DPP-4 inhibitors compared with comparators in patients with T2D [45–47].

Conclusion

The gut-pancreas-bone axis appears to be an endocrine axis, as bone turnover markers are acutely influenced by insulin, glucose, and incretins in healthy adults. This axis is disrupted in diabetes, but the mechanism and progression of this disruption are not understood. Evidence that bone regulates glucose homeostasis in humans to close a potential feedback loop exists, but is currently weak. Finally, there needs to be much more attention paid to potential sex differences in the function and dysfunction of this axis.