Progression from prediabetes to type 2 diabetes mellitus induced by overnutrition

Diabetes is associated with β cell failure. But it remains unclear whether the latter results from reduced β cell number or function. FoxO1 integrates β cell proliferation with adaptive β cell function. We interrogated the contribution of these two processes to β cell dysfunction, using mice lacking FoxO1 in β cells. FoxO1 ablation caused hyperglycemia with reduced β cell mass following physiologic stress, such as multiparity and aging. Surprisingly, lineage-tracing experiments demonstrated that loss of β cell mass was due to β cell dedifferentiation, not death. Dedifferentiated β cells reverted to progenitor-like cells expressing Neurogenin3, Oct4, Nanog, and L-Myc. A subset of FoxO1-deficient β cells adopted the α cell fate, resulting in hyperglucagonemia. Strikingly, we identify the same sequence of events as a feature of different models of murine diabetes. We propose that dedifferentiation trumps endocrine cell death in the natural history of β cell failure and suggest that treatment of β cell dysfunction should restore differentiation, rather than promoting β cell replication. Copyright © 2012 Elsevier Inc. All rights reserved.

Insulin resistance is a characteristic feature of type 2 diabetes and plays a major role in the pathogenesis of the disease (1,2). Although β-cell failure is the sine qua non for development of type 2 diabetes, skeletal muscle insulin resistance is considered to be the initiating or primary defect that is evident decades before β-cell failure and overt hyperglycemia develops (3,4). Insulin resistance is defined as a reduced response of target tissues (compared with subjects with normal glucose tolerance [NGT] without a family history of diabetes), such as the skeletal muscle, liver, and adipocytes, to insulin. Because skeletal muscle is the predominant site of insulin-mediated glucose uptake in the postprandial state, here we will focus on recent advances about the time of onset, as well as the mechanism, of the skeletal muscle insulin resistance. RESEARCH DESIGN AND METHODS The euglycemic insulin clamp technique (5) is considered to be the gold standard for measuring insulin action in vivo. With this technique, whole-body insulin action is quantified as the rate of exogenous glucose infusion (plus any residual hepatic glucose production) required to maintain the plasma glucose concentration at euglycemic levels in response to a fixed increment in the plasma insulin concentration. Because 80–90% of the infused glucose is taken up by skeletal muscle under conditions of euglycemic hyperinsulinemia, insulin sensitivity measured with the insulin clamp technique primarily reflects skeletal muscle (6). Another advantage of this technique is that it can be combined with indirect calorimetry to measure different substrate oxidation rates and with muscle biopsy to examine the biochemical/molecular etiology of the insulin resistance. Measurement of insulin sensitivity by the frequently sampled intravenous glucose tolerance test reflects both hepatic and peripheral insulin resistance and correlates well with the insulin clamp technique (7). Because insulin clamp studies are not feasible in large epidemiological studies, other surrogate markers of insulin sensitivity from glucose and insulin values in the fasting state or after an oral glucose tolerance test (OGTT) have been developed (8 –10). The homeostatic model assessment correlates reasonably well with the insulin clamp (10), but it primarily reflects hepatic insulin sensitivity, since the fasting plasma glucose is determined mainly by the rate of hepatic glucose production (HGP) and insulin is the primary regulator of HGP. The correlation between homeostatic model assessment and the insulin clamp also is less robust when analyzed in subgroups of glucose tolerance (11). During an OGTT, significant (∼30–40%) amounts of glucose are taken up by the splanchnic bed, and HGP is less completely suppressed than during the insulin clamp technique (12). As a result, the plasma glucose concentration during OGTT is affected by both hepatic and peripheral (primarily muscle) insulin resistance. Therefore, indexes of insulin resistance from the OGTT, e.g., the Matsuda index, reflect both hepatic and peripheral insulin resistance and correlate well (R value ∼0.70) with insulin sensitivity measured with the euglycemic insulin clamp (9). Normal glucose homeostasis Skeletal muscle is the major site of glucose uptake in the postprandial state in humans. Under euglycemic hyperinsulinemic conditions, ∼80% of glucose uptake occurs in skeletal muscle (13). Studies using the euglycemic hyperinsulinemic clamp and femoral artery/vein catheterization to quantitate glucose uptake have allowed investigators to quantify leg muscle glucose uptake. Because adipose tissue uses 80–90%) of glucose disposal during the euglycemic insulin clamp occurs in muscle, these results demonstrate that a physiologic elevation in the plasma insulin concentration will exacerbate the underlying muscle insulin resistance. Iozzo et al. (53) performed a 240-min euglycemic insulin clamp study with muscle biopsies in healthy volunteers. Subjects then received a low-dose insulin infusion for 72 h (plasma insulin concentration 143 ± 25 pmol/l [21 ± 2 μU/ml]), followed by a repeat insulin clamp with muscle biopsies. After 72 h of sustained physiologic hyperinsulinemia, insulin-stimulated muscle glycogen synthase activity, total body glucose uptake, and nonoxidative glucose disposal (primarily reflects glycogen synthesis in muscle) were significantly reduced. Taken together, these findings indicate that hyperinsulinemia is not only a compensatory response to insulin resistance, but also a self-perpetuating cause of the defect in muscle insulin action. Molecular etiology of the skeletal muscle insulin resistance in genetically predisposed individuals Using the euglycemic insulin clamp with skeletal muscle biopsy, a number of investigators have examined the insulin signal transduction system in human skeletal muscle of type 2 diabetic subjects and consistently demonstrated defects in IRS-1 tyrosine phosphorylation and PI-3 kinase and Akt activation (26,54,55). To examine whether similar defects are present in genetically predisposed individuals, Pratipanawatr et al. (36) examined insulin signaling in NGT subjects with a strong family history of type 2 diabetes and demonstrated that both the basal and insulin-stimulated IRS-1 tyrosine phosphorylation and PI 3-kinase activity associated with IRS-1 were significantly decreased (Fig. A10). Insulin stimulation of PI 3-kinase activity is a requisite for activation of glucose transport and glycogen synthesis. Increased serine phosphorylation of IRS-1 has been shown to impair insulin signaling (tyrosine phosphorylation of both insulin resistance and IRS-1) in type 2 diabetes (56). In lean insulin-resistant NGT offspring of type 2 diabetic parents, increased serine phosphorylation of IRS-1 in skeletal muscle has been documented in association with impaired activation of Akt (57) (Fig. A11). Thus, at the earliest stage in the natural history of type 2 diabetes, i.e., the NGT insulin-resistant offspring of two type 2 diabetic parents, the molecular etiology of the muscle insulin resistance already is well established and is virtually identical to that in their diabetic parents. Relationship between muscle insulin resistance and altered FFA/muscle lipid metabolism Gulli et al. (32) were the first to demonstrate that the NGT offspring of two type 2 diabetic parents demonstrated marked muscle insulin resistance but normal sensitivity to the suppressive effect of insulin on hepatic glucose production. However, a normal basal rate of HGP in the face of fasting hyperinsulinemia could be construed to indicate the presence of hepatic insulin resistance. More impressive was the elevated fasting plasma FFA concentration in the presence of fasting hyperinsulinemia and the impaired suppression of plasma FFA during the euglycemic insulin clamp (Fig. A12). These findings indicate the presence of marked adipocyte resistance to the antilipolytic effect of insulin. Impaired insulin-mediated suppression of whole-body lipid oxidation also was present in the NGT offspring (Fig. A12). Petersen et al. (58) documented an increase in intramyocellular lipid content in the offspring of two type 2 diabetic parents. This observation is of important clinical significance, since diacylglycerol, long-chain fatty acyl CoAs, and ceramides all have been shown to cause serine phosphorylation of insulin resistance and IRS-1 and lead to the development of insulin resistance in skeletal muscle (59,60). Collectively, these results suggest that intramyocellular accumulation of toxic lipid metabolites plays an important role in the pathogenesis of muscle insulin resistance. To further address this question, Kashyap et al. (30) infused a lipid emulsion for 4 days to cause a physiologic elevation in the plasma FFA concentration in NGT insulin-resistant offspring of two type 2 diabetes parents and in NGT insulin-sensitive subjects without any family history of diabetes. Four days of physiological elevation in the plasma FFA concentration in the offspring did not cause any further worsening of insulin-stimulated whole-body glucose disposal, nonoxidative glucose disposal, glucose oxidation, or preexisting defects in insulin-stimulated insulin receptor tyrosine phosphorylation (30). In contrast, in healthy control subjects, chronic lipid infusion was associated with a marked decline in insulin-stimulated glucose uptake and insulin receptor tyrosine phosphorylation (30). When the insulin-resistant offspring were treated with acipimox for 7 days to reduce the plasma FFA concentration and intramyocellular FACoA concentration, a marked improvement in insulin sensitivity was observed (61). These data lend further support to the observation that insulin resistance in skeletal muscle is an early metabolic defect in the pathogenesis of type 2 diabetes and that muscle lipid accumulation plays a central role in the etiology of the muscle insulin resistance. Mitochondria are the main organelles where fatty acids are oxidized and investigators have focused on their structure and function in patients with type 2 diabetes. Studies using the leg balance technique have documented that fat oxidation is reduced in both type 2 diabetic and obese insulin-resistant nondiabetic individuals (62), suggesting that muscle mitochondrial oxidative capacity is impaired. Recently, two groups independently showed that NGT offspring of two type 2 diabetic parents had a reduced expression of key mitochondrial genes involved in the regulation of oxidative metabolism in skeletal muscle (63,64). The most commonly underexpressed functional genes were those coding for energy generation, including multiple glycolytic, tricarboxylic acid cycle, and oxidative phosphorylation genes. Evidence in support of a role for mitochondrial dysfunction as a cause of muscle insulin resistance in the NGT offspring of two type 2 diabetic parents has been provided by Shulman and colleagues. Using 31P-NMR, these investigators demonstrated impaired mitochondrial activity in NGT insulin-resistant offspring of type 2 diabetic parents (57,58,65). Whereas mitochondria from NGT subjects without any family history of diabetes responded to insulin by increasing ATP production by 90%, mitochondria from insulin-resistant offspring increased ATP production by only 5% (Fig. A13). The authors postulated that muscle mitochondrial dysfunction was the primary defect, leading to elevated intramyocellular fatty acid metabolites (as a consequence of reduced fat oxidation) and subsequent insulin resistance (58,66). However, recent studies by Abdul-Ghani et al. (67) have shown that even small increases in palmitoyl carnitine (5–10 μmol/l) can markedly impair ATP synthesis in mitochondria isolated from human muscle. Thus, it is unclear which is the cart and which is the horse: mitochondrial dysfunction leading to increased intramyocellular lipid content and insulin resistance or increased muscle lipid content (i.e., secondary to elevated plasma FFA levels and/or excessive lipid ingestion) leading to mitochondrial dysfunction and insulin resistance. SUMMARY The maintenance of normal glucose homeostasis depends on a finely balanced dynamic interaction between tissue (muscle, liver, and fat) sensitivity to insulin and insulin secretion. Even in the presence of severe insulin resistance, a perfectly normal β-cell is capable of secreting sufficient amounts of insulin to offset the defect in insulin action. Thus, the evolution of type 2 diabetes requires the presence of defects in both insulin secretion and insulin action, and both of these defects can have a genetic as well as an acquired component. When type 2 diabetic patients initially present to the physician, they will have had their diabetes for many years, and defects in insulin action (in muscle, liver, and adipocytes) and insulin secretion will be well established (1,2,39). At this stage, it is not possible to define which defect came first in the natural history of the disease and which tissue is the primary defect responsible for the insulin resistance. Although insulin resistance represents the earliest detectable abnormality in the great majority of type 2 diabetic people, in a minority of individuals (i.e., glucokinase deficiency), it is clear that a β-cell defect initiates the disturbance in glucose homeostasis. Nevertheless, it is now clear that in any given diabetic patient, whatever defect (insulin resistance or impaired insulin secretion) initiates the disturbance in glucose metabolism, it will eventually be followed by the emergence of its counterpart (Fig. A14). Insulin resistance is a nearly universal finding in patients with established type 2 diabetes. In normal-weight and obese individuals with IGT and in type 2 diabetic subjects with mild fasting hyperglycemia (110–140 mg/dl, 6.1–7.8 mmol/l), both the basal and glucose-stimulated plasma insulin levels are increased. Although the first-phase insulin response may be decreased in some, but not all, of these subjects, the first phase consistently is increased in the NGT offspring of two type 2 diabetic parents and the total insulin response is increased in NGT offspring and in IGT subjects. In each of these groups, tissue sensitivity to insulin, measured with the insulin clamp technique, has been shown to be diminished. Prospective studies conclusively have demonstrated that hyperinsulinemia and insulin resistance precede the development of IGT and that IGT represents the forerunner of type 2 diabetes. This scenario has been well documented in Pima Indians, Mexican Americans, and Pacific Islanders. It is noteworthy that all of these populations are characterized by obesity and a younger age at onset of diabetes. Such results provide conclusive evidence that insulin resistance is the inherited defect that initiates the diabetic condition in the majority of type 2 diabetic patients. Studies in NGT first-degree relatives of diabetic individuals and in the offspring of two diabetic parents indicate that the inherited defect in insulin action results from an abnormality in the glycogen synthetic pathway in muscle and more proximal defects in glucose transport/phosphorylation and insulin signal transduction. As the insulin resistance progresses and muscle glucose uptake becomes further impaired, the postprandial rise in plasma glucose concentration becomes excessive, but the increase in basal hyperinsulinemia is sufficient to maintain the fasting plasma glucose concentration and HGP within the normal range. Nonetheless, there is an excessive postprandial rise in plasma glucose concentration, and a longer time is required to restore normoglycemia after each meal. Eventually, however, the insulin resistance becomes so severe that the compensatory hyperinsulinemia is no longer sufficient to maintain the fasting glucose concentration at the basal level. The development of hyperglycemia further stimulates β-cell secretion of insulin, and the resultant hyperinsulinemia causes a downregulation of insulin receptor number and of the intracellular events involved in insulin action, thus exacerbating the insulin resistance. Initially, the hyperglycemia-induced increase in insulin secretion serves a compensatory function to maintain near-NGT. In some individuals, the persistent stimulus to the β-cell to oversecrete insulin leads to a progressive loss of β-cell function. Chronic hyperglycemia (glucose toxicity) and/or disturbances in lipid metabolism (lipotoxicity) may contribute to the defect in insulin secretion. The resultant insulinopenia leads to the emergence/exacerbation of postreceptor defects in insulin action. Many of the intracellular events involved in glucose metabolism depend on the surge of insulin that occurs three to four times per day in response to nutrient ingestion. When the insulin response becomes deficient, the activity of the glucose transport system becomes severely impaired and a number of key intracellular enzymatic steps involved in glucose metabolism become depressed. Additionally, when severe insulinopenia ensues, plasma FFA levels rise, further contributing to the defects in intracellular glucose disposal. There is also compelling evidence that hyperglycemia per se can downregulate the glucose transport system, as well as a number of other intracellular events involved in insulin action (glucose toxicity), and a similar argument can be made concerning the intracellular derangement in lipid metabolism. This pathogenetic sequence can explain all of the clinical and laboratory features observed in type 2 diabetic patients. Insofar as the cellular defect is generalized, both hepatic and peripheral tissues (skeletal muscle and adipocytes), and possibly the β-cells themselves, would manifest insulin resistance, and the numerous metabolic alterations characteristic of the diabetic state could be related to one and the same primary defect. The NGT offspring of two type 2 diabetic parents also manifest marked adipocyte resistance to the suppressive effects of insulin on lipolysis. One could argue, therefore, that the adipocyte represents the primary tissue responsible for the insulin resistance. According to this scenario, the elevated plasma FFA levels produce insulin resistance in muscle and liver and impair β-cell function. Adipocytes in the NGT offspring of two type 2 diabetic parents also secrete excessive amounts of inflammatory and insulin resistance producing adipocytokines that could initiate/exacerbate the insulin resistance in skeletal muscle. As reviewed by Iozzo in this symposium, the adipocyte insulin resistance could be genetic in origin or induced in utero during the third trimester by nutritional deprivation or overfeeding. There is less evidence to support a role for the liver as the organ responsible for the insulin resistance. However, the NGT offspring of two type 2 diabetic parents have a normal rate of HGP in the presence of fasting hyperinsulinemia, suggesting the presence of hepatic resistance to the suppressive effect of insulin on glucose production. Therefore, one could argue that the resultant fasting hyperinsulinemia leads to the development of insulin resistance in skeletal muscle.

In this Perspective, Michael Czech presents evidence for whether hyperinsulinemia occurs before insulin resistance upon overfeeding or high-fat diet feeding, or whether insulin resistance causes hyperinsulinemia, thus attempting to delineate the relationship between hyperinsulinemia, obesity and insulin resistance.