Insulin Hypersecretion as Promoter of Body Fat Gain and Hyperglycemia Free

Excess adiposity is a major risk factor for prediabetes and for the risk of progressing from prediabetes to type 2 diabetes (1). The mechanisms responsible for the increased risk of hyperglycemia associated with excess adiposity are not entirely clear, but for decades their descriptions have centered on two key factors: 1) reduced insulin sensitivity and 2) reduced β-cell function. Excess body fat mass is strongly associated with resistance to the stimulatory effect of insulin on skeletal muscle glucose uptake and the inhibitory effects of insulin on hepatic glucose production and adipose tissue triglyceride lipolysis, yet many people with obesity are normoglycemic (1). A common conceptual paradigm posits that normoglycemia is maintained if β-cells secrete enough insulin to compensate for the lower insulin sensitivity in liver and muscle (2–4). It has been proposed that increased release of fatty acids from adipose tissue triglyceride lipolysis in people with obesity causes resistance to insulin action and the compensatory increase in insulin secretion (2–6) (Fig. 1). In this conceptual framework, chronic stress on β-cells may eventually cause insufficient insulin secretion caused by β-cell “exhaustion” or even complete β-cell failure and various degrees of hyperglycemia (2–6). An alternate viewpoint posits that insulin hypersecretion (defined as insulin secretion above and beyond what is normal for the insulin sensitivity status) precedes and causes obesity and insulin resistance (5,7–10). In this alternate model, excess insulin directly stimulates lipogenesis and causes insulin resistance primarily via a decrease in insulin receptors but also through excess cellular lipids (5,11–13) (Fig. 1). Although the results from preclinical research provide support for both paradigms, evaluating the sequence of events to determine cause and effect in people is difficult.

In this issue of Diabetes, Tricò et al. (14) performed deep metabolic phenotyping of 100 children and adolescents with obesity (defined as BMI ≥85th percentile of age- and sex-based norms) to evaluate the metabolic features associated with insulin hypersecretion and the metabolic trajectories of hypersecreters compared with normosecreters. This study represents a follow-up to the authors’ earlier study conducted in normoglycemic adults without obesity and in adolescents with obesity, where they first identified and defined primary insulin hypersecretion as insulin secretion in excess of what would be considered normal for a given insulin sensitivity value obtained during a hyperinsulinemic-euglycemic clamp procedure (15). In the current study (14), all participants completed a frequently sampled oral glucose tolerance test (OGTT) to evaluate fasting plasma glucose, lipid, and hormone concentrations, glucose tolerance, whole-body insulin sensitivity (by using the whole-body insulin sensitivity index), and β-cell function (defined as insulin secretion in relation to plasma glucose). The authors define insulin hypersecreters as participants within the upper tertile of the β-cell function-insulin sensitivity relationship. All participants also completed dual-energy X-ray absorptiometry (DEXA) and magnetic resonance imaging (and magnetic resonance spectroscopy) scans to evaluate total body and regional adipose tissue volumes and intrahepatic triglyceride content. Adipocyte size was evaluated in subcutaneous abdominal adipose tissue biopsy samples. A subset of the participants (n = 59) also underwent a single-stage hyperinsulinemic-euglycemic clamp procedure to evaluate whole-body insulin sensitivity. A smaller subset of participants (n = 17) completed a deuterium oxide study to evaluate de novo lipogenesis of adipose tissue triglycerides. Sixty-eight of the 100 participants repeated the frequently sampled OGTT and DEXA scan after about a 2-year period to evaluate changes in body composition and glycemic status.

The main findings from these two studies (14,15) are summarized in Fig. 2. Briefly, basal insulin secretion rate, insulin secretion rate after glucose ingestion, and β-cell function (insulin secretion rate in relationship to plasma glucose) were all higher in hypersecreters than in normosecreters. In addition, hypersecreters had higher rates of de novo lipogenesis, increased adipocyte size, and more intra-abdominal adipose tissue and intrahepatic triglycerides, whereas basal circulating free fatty acid concentration was not different between hypersecreters and normosecreters. Although there was no difference in total body adiposity at baseline, hypersecreters gained relatively more body fat over the 2-year follow-up period. Insulin sensitivity (assessed by using the whole-body insulin sensitivity index and the hyperinsulinemic-euglycemic clamp procedure) also was not different between hypersecreters and normosecreters at baseline or during the follow-up period, but plasma glucose concentration was higher in hypersecreters, and significantly more hypersecreters than normosecreters progressed from normoglycemia to hyperglycemia.

In the study by Tricò et al. (14), insulin hypersecretion was associated with increased triglyceride synthesis in subcutaneous adipose tissue, adipocyte hypertrophy, intra-abdominal adipose tissue mass, intrahepatic triglyceride content, and greater body fat gain during the 2-year observation period. The difference in body fat gain (∼2% of total fat mass), cumulatively over many more years, could be substantial. These findings are entirely consistent with data from studies conducted in animals that found downregulated insulin production can reverse (8) or prevent (11,16) body fat accumulation, including ectopic fat. Insulin is known to be essential for adipose tissue growth and maintenance and is a key driver of lipogenesis in adipocytes and in the liver even when glucose metabolism is resistant to insulin (17–19). Accordingly, the greater fat depot accumulation in insulin hypersecreters is likely directly attributable to the secretion of copious amounts of insulin and may exacerbate insulin resistance mechanisms, since it has been proposed that local modulation of lipid homeostasis can drive selective insulin resistance (17–19). The observations made by Tricò et al. (14) are important, as they represent some of the first direct evidence in people that insulin hypersecretion, on its own—albeit in an obesogenic context—can promote an increase in adipose tissue. This complements results from large cardiovascular outcome trials in people with type 2 diabetes that found insulin therapy causes weight gain (20). Conversely, it is known that mutations in the insulin receptor that are associated with reduced insulin signaling lead to reduced adiposity (21). Together, these data suggest that excessive amounts of circulating insulin can promote excessive weight gain.

The metabolic phenotype of insulin hypersecreters described by Tricò et al. (14,15) provides important new insights into the regulation of plasma glucose concentration and raises some intriguing questions. The first striking observation is the considerable heterogeneity in β-cell function and insulin secretion at any whole-body insulin sensitivity value (assessed by the hyperinsulinemic-euglycemic clamp method or a surrogate index), and the authors noted that insulin hypersecretion can occur independent of alterations in both insulin sensitivity and total body adiposity. A second interesting observation is that insulin hypersecretion relative to insulin sensitivity is associated with higher (not lower) plasma glucose concentration. The phenotype of insulin hypersecreters described by Tricò et al. (14,15) matches well with the phenotype of mice with genetically reduced insulin production, which initially have lower fasting plasma glucose and only become relatively hyperglycemic and insulin resistant with age (11). Some would suggest that these findings go against the “hyperbolic law of glucose tolerance” that has driven diabetes research for decades and claims plasma glucose concentration simply represents the net balance of insulin sensitivity and the amount of circulating insulin (2–6). We believe glucose homeostasis is more complex than is commonly appreciated. Additionally, the confusion might partially arise from assumptions around the methods used to assess insulin sensitivity.

The hyperinsulinemic-euglycemic clamp procedure is considered the gold-standard method for evaluating insulin sensitivity (22), but it involves intravenous glucose and insulin administration and typically measures insulin action on glucose metabolism at a single fixed insulin dose but not during dynamic changes in insulin as they occur during the transition from postabsorptive to postprandial conditions and back. This has important implications for the interpretation of the results and represents a major limitation of the hyperinsulinemic-euglycemic clamp procedure and surrogates thereof, such as the whole-body insulin sensitivity index the authors used. Insulin affects glucose metabolism and other cellular functions by binding to insulin receptors on plasma membranes, which results in insulin receptor autophosphorylation, initiation of a downstream signaling cascade, and receptor internalization (23,24) (Fig. 3A). The cumulative effect of these processes results in a distinct dose-response relationship between insulin and the metabolic readout. Alterations in specific steps along this pathway cause characteristic changes in the dose-response relationship, including a shift of the dose-response curve toward a higher or lower insulin concentration range without a change in the shape of the curve (i.e., a change in insulin sensitivity) when insulin receptor number or insulin-insulin receptor binding dynamics are altered (Fig. 3B) or a change of the shape/slope of the insulin dose-response curve (i.e., a change in insulin responsiveness) because of alterations in intracellular insulin signal transduction (Fig. 3C) (24). Accordingly, it is difficult to appreciate interindividual differences in insulin action on glucose metabolism from a single measurement (Fig. 3D) and to predict differences in insulin action on glucose metabolism during dynamic physiological conditions with a single clamp-derived measure of insulin action.

Even though Tricò et al. (14,15) report no difference in baseline insulin sensitivity of glucose metabolism between insulin hypersecreters and normosecreters, the data they present suggest that insulin action on glucose metabolism is impaired in insulin hypersecreters, because plasma glucose concentration during the OGTT was higher in hypersecreters despite markedly higher plasma insulin concentration. Furthermore, even though insulin secretion during the OGTT was markedly increased, it provided an insufficient amount of insulin to prevent an increase in plasma glucose. During basal conditions, the data are less clear but also point toward insulin resistance, because in one study (15) basal plasma glucose concentration was higher in insulin hypersecreters compared with normosecreters despite markedly higher insulin secretion and systemic plasma insulin concentration. In the other study (14), systemic plasma insulin concentration was not different between hypersecreters and normosecreters, and basal plasma glucose was ∼2% lower in hypersecreters. However, basal insulin secretion and thus basal hepatic insulin exposure was much (∼20%) higher in hypersecreters compared with normosecreters. There was also insulin resistance of fatty acid metabolism in insulin hypersecreters, because the concentration of nonesterified fatty acids in plasma during the hyperinsulinemic-euglycemic clamp procedure with fixed insulin exposure was higher in insulin hypersecreters and basal plasma fatty acid concentration was not different despite higher basal plasma insulin (15).

A likely primary cause of insulin resistance in insulin hypersecreters is insulin itself. It is now established, but still not widely recognized, that hyperinsulinemia can impair insulin action by causing insulin receptor downregulation (through internalization, excess degradation, and reduced synthesis), decreasing insulin-insulin receptor binding dynamics, blunting intracellular signal transduction, or a combination of these mechanisms (13,25–29) (Fig. 4). To some extent, insulin-mediated insulin resistance may be a normal and even necessary negative feedback mechanism to prevent hypoglycemia; chronically it may also present an adaptive defense mechanism against excess insulin-mediated metabolic stress (30). Fatty acids are commonly considered a main culprit in the development of insulin-resistant glucose metabolism (3,6). Although the basal concentration of nonesterified fatty acids in plasma was not different between insulin hypersecreters and normosecreters (14,15), suggesting fatty acid–mediated insulin resistance was not involved, it cannot be ruled out that the intracellular fatty acid load was higher in hypersecreters because of enhanced tissue fatty acid extraction from the circulation (31,32).

The importance of insulin for the regulation of glucose metabolism cannot be understated, but there are additional factors involved that should not be overlooked (Fig. 5). The higher plasma glucose concentration in insulin hypersecreters compared with normosecreters could be because of alterations in insulin-independent glucoregulatory mechanisms, including other hormones (glucagon, amylin, incretins, etc.), plasma glucose itself, the central nervous system, and gastrointestinal glucose dynamics (gastric emptying, intestinal absorption, and splanchnic extraction) (33–38) that could not be overcome with even a large excess of insulin or may even be a manifestation of insulin resistance (e.g., excessive glucagon because of insulin resistance of pancreatic α-cells). Conversely, it is also possible that these control mechanisms prevented a much greater increase in plasma glucose than would have occurred as a result of the insulin resistance.

A major strength of the studies by Tricò et al. (14,15) is the longitudinal study design, which demonstrated that insulin hypersecretion is associated with an increased risk for developing prediabetes. Over the course of 2–3 years, insulin hypersecreters, compared with normosecreters, were approximately twice as likely to progress from normoglycemia to hyperglycemia, defined as either impaired fasting glucose, impaired glucose tolerance, or combined impaired fasting glucose and impaired glucose tolerance (14,15). The transition from normoglycemia to hyperglycemia was driven primarily by a decrease in glucose tolerance (i.e., increased plasma glucose value at 2 h during the OGTT) (14,15) caused by a worsening of insulin action rather than a worsening of β-cell function, defined as insulin secretion in relationship to plasma glucose. This finding concurs with the results from other studies that found isolated impaired glucose tolerance is primarily a result of insulin resistance (39). It is worth noting that insulin action decreased during the observation period in both insulin hypersecreters and normosecreters without a difference between the two groups, but more hypersecreters crossed the threshold of prediabetes because they had higher glucose concentration values at baseline.

The absolute rates of both insulin secretion during overnight fasted conditions and after an oral glucose challenge, as well as β-cell function, defined as insulin secretion rate in relation to plasma glucose, were significantly higher in insulin hypersecreters (14,15). In human subject research, like the studies by Tricò et al. (14,15), we can only speculate about the cellular mechanisms responsible for the high insulin production. Research with preclinical model systems suggest causes of insulin hypersecretion include nutrient (fatty acids and amino acids)-induced β-cell proliferation and β-cell hyperreactivity (40,41). Although the data presented by Tricò et al. (14,15) appear to rule out fatty acid–mediated insulin hypersecretion because basal plasma fatty acid concentration was not different between insulin hypersecreters and normosecreters, increased fatty acid flux, and consequently β-cell fatty acid exposure, cannot be ruled out (31,32,42). Environmental factors and oxidative stress/reactive oxygen species have also been implicated in causing insulin hypersecretion (43). In addition, insulin itself can stimulate β-cell proliferation (44,45). It is also possible that β-cell–specific insulin resistance (i.e., specific loss of insulin receptors) increases insulin secretion (46) or selective insulin resistance associated with impaired phosphatidylinositol 3-kinase signaling shunts insulin signaling toward proliferative signals (47). Adult β-cells that are induced to replicate have a phenotype similar to that of immature neonatal β-cells, which have a lower set point for glucose-stimulated insulin secretion and higher basal insulin secretion rate (48). Furthermore, insulin secretion is controlled by autocrine and paracrine signals as well as neurotransmitters (40,49). Lastly, insulin hypersecreters in the studies by Tricò et al. (14,15) may have had an exaggerated incretin response to glucose ingestion, given that during an OGTT, about 50% of the insulin secretion can be attributed to the release of incretins (50).

The “discovery” of primary insulin hypersecreters and the metabolic phenotype associated with insulin hypersecretion offers novel insights that appear to defy some common knowledge and are thus expected to propel research in the field in exciting new directions. The most obvious questions relate to the mechanisms involved in insulin-mediated adipogenesis and insulin hypersecretion by β-cells during basal, overnight fasted conditions and postprandially.

Funding. The authors received salary support from National Institutes of Health grants R01 HL159461, R01 HL161829, R01 DK131188, R01 DK121560, and R01 DK 115400 (all to B.M.), a grant from the Canadian Institute for Health Research (PJT-168857 to J.D.J.), and grants from the Novo Nordisk Foundation (NNF19OC0057174), the Swedish Research Council (2022-01033 and 2022-01011), the Cancerfonden (20 0840 PjF), and the Diabetesfonden (DIA2022-753 and DIA2021-631) (to G.S. and P.-A.J.) while writing this article.

The funders had no role in the interpretation of the results.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.