Etiology of Insulin Resistance in Youth with Type 2 Diabetes

Abstract

Type 2 diabetes (T2DM), historically an adult disease, is now increasingly prevalent in obese youth. Poor diet and increased sedentary behavior contribute to the increasing rates of obesity in youth, yet not all obese children develop T2DM. In general, T2DM is characterized by both insulin resistance (IR) and pancreatic beta-cell insufficiency. In children, IR is related to elevated body mass index (BMI) and pubertal hormones, along with abnormal fat partitioning, elevated free fatty acids, inflammation, and/or mitochondrial dysfunction. Hyperglycemia and T2DM develop when the pancreas cannot match the increased insulin demands resulting from IR. Unique to youth, IR varies with stage of pubertal development, and some children may have resolution of hyperglycemia post-puberty once the IR of puberty resolves. Further understanding of IR, the progression to T2DM in youth, and later outcomes as adults will help direct future therapies and interventions for youth at risk.

Introduction

Type 2 diabetes mellitus (T2DM) was once thought to be a disease exclusive to adults. However, as rates of obesity in children have risen, the prevalence of pediatric T2DM has also increased [1–3]. T2DM is caused by the combination of insulin resistance (IR), defined as dysfunctional cellular response to insulin combined with pancreatic beta-cell insufficiency. IR, thought to be the initial defect, also underlies the metabolic syndrome. The metabolic syndrome is a cluster of obesity-associated conditions, including dyslipidemia, fatty liver disease, and hypertension, that puts youth at risk for cardiovascular disease [4, 5]. This is especially concerning, as recent reports have indicated that average life expectancy drops by approximately 15 years when adolescents develop T2DM, and chronic complications may occur by the age of 40 [6•]. Multiple factors play a role in the progression from IR to T2DM in youth. It is well known that insulin sensitivity decreases in puberty due to increases in growth hormone, testosterone, and estrogen secretion, and increases fat mass. Our modern, sedentary, high-calorie lifestyle has contributed to the increase in obesity rates overall, but the youth who develop IR or T2DM seem to store this excess lipid in different locations than obese youth who remain insulin sensitive [1]. It may also be that this alternate lipid storage is affected by fetal exposure to maternal obesity or diabetes, and early neonatal feeding patterns [7, 8••]. This review will focus on the unique aspects of IR in progression of T2DM in youth and the lifestyle and molecular mechanisms that contribute to IR in this population.

Assessment of Insulin Resistance in Children

The gold standard to measure insulin sensitivity is the hyperinsulinemic euglycemic clamp, as exogenous insulin administration bypasses reliance on pancreatic insulin release [4, 9, 10]. However, accurate assessment of IR in children is difficult as there are limited normative data and the dynamic relationship between glucose homeostasis and pancreatic beta-cell function is influenced by factors such as exercise, menstrual cycle, diet, and puberty [11, 12]. Using hyperinsulinemic euglycemic clamps, we and others have demonstrated IR in obese, non-diabetic adolescents, and markedly decreased insulin sensitivity in obese, T2DM adolescents [9, 13]. A simpler yet still valid measurement of insulin sensitivity and secretion in individuals with intact beta cells is the frequently sampled intravenous glucose tolerance test (IVGTT) [4]. Hyperinsulinemic clamps and the IVGTT are both labor intensive, somewhat invasive, and costly modalities that are not practical for use in large scale studies for screening. Therefore, surrogate markers are often used. The OGTT-derived whole body insulin sensitivity index and the ratio of glucose and insulin ratios under the curve correlate well with clamp-derived IR in youth, but less so with fasting indices [10, 14]. However, once insulin secretion is decreased, OGTT and IVGTT assessments become problematic, and require further calculations. The product of beta-cell function and insulin sensitivity is the disposition index (DI), which quantifies insulin secretion relative to insulin sensitivity [15]. DI has been shown to predict development of T2DM in adult and adolescent longitudinal studies, with T2DM occurring after beta-cell failure [16••, 17]. A comparison of OGTT-derived DI (oDI) and clamp-derived DI (cDI) in adolescents demonstrated that oDI correlated well with cDI especially among subjects with IGT [16••]. Thus, oDI may be used as an estimate for IR when clamps are not available or feasible, and there is a suspicion of insufficient insulin secretion.

Fasting measures of IR are problematic. Fasting insulin tests must be truly fasting and due to assay variability can only be compared with results run in the same laboratory, making the creation of universal norms difficult. The homeostasis model assessment of IR (HOMAIR) is a IR surrogate widely used for population-based studies that uses fasting insulin and glucose but is only marginally better or even inferior to fasting insulin alone in children [4, 18]. In addition, due to the variability of IR through puberty, a normative range for children is not clear, though a recent study of 6100 children in Mexico reported age-based pediatric HOMA-IR norms as they change throughout childhood, similar to that used for BMI and blood pressure [19•]. Finally, as is the case for the OGTT and IVGTT, measures based on fasting insulin become invalid once insulin secretion is defective, as is the case in IGT, IFG, and T2DM [4, 9]. Alternatively, the fasting triglyceride:HDL ratio has been correlated with insulin suppression-based IR assessment in adults, but has yet to be validated in children [20]. In addition, using waist circumference, HbA1c and triglyceride data, we recently developed an estimate of clamp-based IR in youth for application to large epidemiologic studies, but further research as to its predictive value for long term outcomes is still required [21].

Mechanisms of Insulin Resistance in Children with T2DM

As in adults, IR is thought to be related to the interaction between abnormal fat and glucose metabolism. A Japanese study of children with T2DM showed that OGTT-derived IR correlated with the degree of overall obesity [22]. Abnormal fat partitioning, with increased visceral, hepatic, and intramyocellular lipids and elevated BMI are related to IR in adults [1, 2, 23]. Similarly this relationship between IR and increased visceral, hepatic and intramyocellular lipids are seen in youth with prediabetes and T2DM [13, 24–26]. In obese youth with normal glucose tolerance, increased muscle, hepatic, and visceral fat are also related to the degree of IR [27]. Youth with T2DM can have up to 3 times the amount of liver fat compared with BMI-matched nondiabetic controls [1]. Similarly, youth with non-alcoholic fatty liver disease (NAFLD) have hepatic and peripheral IR as demonstrated by hyperinsulinemic euglycemic clamps [28]. In obese youth with NAFLD, the severity of liver disease is also related to the degree of IR [29]. It is not clear if hepatic steatosis precedes IR, but is known that NAFLD is a strong risk factor for IGT, IFG, and T2DM [1, 13, 25]. However, it is important to remember that not all obese youth go on to develop T2DM, and this data suggests that early storage of fat in the liver, viscera and muscle increases risk of development of T2DM. In addition, genetic factors likely influence the location of storage of excess lipid.

Increased ectopic fat storage may also be related to increased levels of plasma free fatty acids (FFA) which can also directly decrease insulin sensitivity. Excessive processing of nutrients, including FFA by mitochondria may result in an increase of reactive oxygen species (ROS) due to uncoupling of oxidative phosphorylation which in turn leads to altered mitochondrial function and to production of more ROS [30]. These ROS can then lead to endoplasmic reticulum dysfunction and thus defective insulin secretion in pancreatic beta cells [31]. ROS, and/or lipid byproducts may also impair insulin signaling cascades. Obese children are reported to have increased FFA concentrations, related to their degree of IR [32]. Muscle mitochondrial dysfunction in individuals with T2DM has been shown to be associated with muscle IR [33]. A study in obese children found that obesity, per se, was not related to mitochondrial function, but that those with obesity and IR had prolonged ATP synthesis following exercise, a marker of mitochondrial dysfunction [34]. Muscle and hepatic IR is also mediated by peroxisome proliferator-activated receptor (PPAR) receptors, as evidenced by the use of PPAR-y agonists for the treatment of T2DM. In some individuals with T2DM, there may be genetic alterations in PPAR-y expression, mitochondrial function, and/or inflammation related to IR [35], arguing that obesity-induced FFA elevations in genetically predisposed individuals are a primary cause of pediatric T2DM.

Part of the variability in insulin sensitivity in obese adolescents is likely related to hormonal changes in puberty. Growth hormone, estrogen, and androgens can all affect insulin sensitivity [36–38]. Studies have monitored the changes in insulin sensitivity in healthy children throughout the progression of puberty. These studies show that IR starts prior to puberty, even before rises in pituitary gonadotropin secretion, and is partly explained by the accumulation of fat and rising insulin-like growth factor 1 (IGF-1) [12]. IR peaks when children are in mid-puberty, equivalent to Tanner stage 3–4, and decreases at the conclusion of puberty [39], at least in normal-weight youth. These unique, fluctuating influences on IR, which can vary by time of day and time during menstrual cycle in girls, maybe a large reason that the diagnosis of IR, IGT, and IFG can be so challenging to reproduce in adolescents.

The Role of Lifestyle Factors on Development of T2DM

The recent increase in obesity rates has become a costly public health problem. The link between obesity and T2DM has been well reported through many epidemiologic studies [40]. Most studies have focused on the degree of obesity, using BMI as a marker for risk to development of diabetes [41]. The duration of obesity may be equally important as a risk factor for T2DM. Independent of absolute BMI, longer obesity duration is associated with T2DM [42, 43]. The tripling of obesity rates in children over the last 30–40 years implies that the length of time that people will spend obese has also increased. Studies show that time spent at a BMI greater than the 85^th percentile is associated with an increased risk of diabetes [44]. Pubertal IR requires additional insulin secretion to maintain normoglycemia, and superimposed obesity may create too large of a demand on the pancreas in susceptible individuals. Furthermore, obesity, once established, is difficult to reverse. Therefore, being obese at puberty may be especially detrimental. Thus, early interventions, prior to puberty, may be most helpful in preventing the development of T2DM and its complications [45]. Efforts to quell obesity in the very young show promise, as youth at risk for behavior problems who received family intervention at age 4 had a lower BMI and improved health behaviors as they reached adolescence [46].

A sedentary lifestyle is also critical to development of T2DM in youth. There has been a dramatic recent decrease in overall physical activity level, especially in girls, during the transition from childhood to adolescence [47], the time at which overweight youth are at great risk for developing T2DM. Of note, T2DM youth are even more obese and sedentary than T2DM adults [48, 49]. A sedentary lifestyle and low cardiorespiratory fitness are associated with IR, IFG, and T2DM [13, 50]. This predisposing factor is especially important to understand, as once T2DM develops, diabetes itself appears to impair cardiorespiratory fitness, which may further reduce physical activity level [13]. Participation in physical activity can prevent or delay the onset of T2DM and improve blood glucose control in adults [51]. Additionally, in children with a family history of obesity, physical activity is associated with increased insulin sensitivity, independent of level of fitness, or sedentary behavior [52•]. However, in the recently published TODAY study of T2DM youth, an intensive lifestyle intervention including physical activity was unable to decrease treatment failure when added to metformin, arguing that exercise and dietary changes are difficult to make in youth with established T2DM [53]. Therefore, it appears especially important for youth to maintain a physical activity regimen early in life.

The rise in obesity, metabolic syndrome, and diabetes has increased in conjunction with the rise in daily caloric intake [54]. In particular, a western diet high in total fat, saturated fatty acids, and refined grains has been associated with increased risk for metabolic syndrome [55]. In contrast, risk for metabolic syndrome is decreased with diets high in fruits, vegetables and whole grains [56, 57]. In addition, Mediterranean diets high in monounsaturated and polyunsaturated fats and fiber reduced the risk for metabolic syndrome in a controlled clinical trial [58]. The Dietary Intervention Study in Children (DISC) evaluated the effects of a low fat, high fiber diet during adolescence and showed benefit in both glycemic control and blood pressure [59]. A ketogenic very low calorie diet sustained for 60 days can also improve BMI and blood glucose control in children with T2DM [60].

In addition to the effect of childhood diet and activity, the development of IR and obesity may also be influenced by the in utero environment, and by feeding patterns and weight gain in the first year of life. It is known that both low and high birth weight predict metabolic abnormalities later in life [61]. In children exposed to gestational diabetes mellitus (GDM) in utero, the BMI velocity from ages 6–12 is increased [8••]. The risk of metabolic syndrome by age 11 years in children born large for gestational age to GDM mothers was 3.6 times greater than those born average for gestational age to GDM mothers, and maternal obesity, regardless of birth size, increased the risk twofold [7]. Increased maternal BMI prior to conception was also associated with increased neonatal hepatic fat content [62]. At 1 year of age, infants whose mothers had GDM had higher plasma triglycerides, and glucose area under the curve during an OGTT compared with infants from mothers without GDM [63]. IR in these infants was associated with increased weight gain in the first year of life. Furthermore, children who were breastfed at least 6 months in infancy were more likely to be in to be lean and have lower percentages of visceral and subcutaneous adipose tissue [64]. While the mechanisms of these changes require further study, it is likely that changes we are currently seeing in pediatrics may be “second generation” effects of the obesity epidemic which started 30–40 years ago.

Unique Aspects of Progression from Insulin Resistance to T2DM in Children

The progression from IR to IGT/IFG to T2DM is regulated by the relationship between IR and insulin secretion [65]. Hyperglycemia develops once the beta cell insulin secretion is inadequate to match the level of IR [1, 66, 67]. In children, the rate of conversion from IGT/IFG to T2DM appears faster compared with adults and can occur over the span of 12–21 months [1, 68, 69••]. Adults who are IR lose an average of 7 % of their beta cells per year [66]. Case reports have detailed the progression of a child to development of T2DM with a rate of beta-cell deterioration at about 15 % per year [70]. Secretion of insulin is dependent on beta-cell mass and secretory capacity, that are both governed by genetic and environmental factors, including increased FFA levels [1, 66, 71]. A recent study of 700 children found that higher fasted plasma concentrations of FFA were related to decreased first phase insulin secretion following a glucose load [72]. Once IR-induced hyperglycemia develops, inflammation caused by hyperglycemia may promote further apoptosis of beta cells [73]. In longitudinal pediatric population studies, IR worsens already damaged beta cells [29, 74], and cross sectional studies show that children who progress to T2DM have lower insulin levels [67]. In obese youth a >30 % decline in cDI is already apparent at the upper end of normal glucose tolerance (NGT; 120–140 mg/dl) [75]. Furthermore, with a 1 hour OGTT glucose >155 mg/dl, obese youth have a lower DI even if they are NGT at 2 hours [76]. Although overall insulin secretory capacity is greater in youth, and first phase insulin secretion is higher in T2DM youth compared with T2DM adults, there is a step-wise decline in first and second-phase insulin secretion from NGT to IGT to T2DM in obese adolescents [68, 77•].

In contrast to adults with IR, not every obese adolescent with hyperglycemia goes on to develop T2DM. In a cohort of 75 NGT obese adolescents, increased baseline HbA1c and weight gain predicted progression to IGT [77•]. Among 218 obese youth, HbA1c at baseline and 2 hour OGTT level at baseline predicted an increase in 2 hour OGTT. HbA1c >5.7 % predicted later IGT or T2DM [78]. In a study of 117 obese youth, at baseline 84 had NGT, and 33 had IGT. 30.3 % of those with IGT returned to NGT, and only 8 % progressed to T2DM. Severe obesity, IGT, and African American background were the best predictors of developing T2DM [79]. Another study prospectively followed 79 obese white children and adolescents with IGT. 32 % of children after 1 year continued to be IGT, but 66 % converted to NGT. Predictive factors for normalization of IGT were lower weight, HbA1c, 2 hour OGTT glucose at baseline, weight loss, and entering late puberty during follow-up [69••]. In the 3- to 5-year follow-up, 16 % were still IGT, 75 % converted to NGT, and 2 % developed T2DM. Higher 2 hour OGTT level at baseline and weight gain were predictors for remaining at IGT [80]. Some youth, especially Caucasian, or in later stages of puberty, may not progress to overt T2DM.

We propose that there are 4 overall phenotypes of children, as shown in Fig. 1a–d. In lean children (Fig. 1a), insulin sensitivity decreases with puberty, and is matched by adequate up-regulation of pancreatic insulin secretion. In obese children (Fig. 1b) with no family history of T2DM and minimal abnormalities in fat partitioning, there is decreased insulin sensitivity, and again adequate pancreatic up-regulation in insulin secretion through puberty. In those obese children with some alteration in fat partitioning (Fig. 1c), insulin sensitivity also decreases through puberty, and at the nadir of insulin sensitivity, insulin secretion is inadequate, leading to IGT or IFG. However as they complete puberty, insulin sensitivity recovers and their hyperglycemia resolves as their pancreas can again meet the insulin demands. It may be that the transient IGT seen in these teens is similar to that of GDM, implying that these teens will be at increased risk of diabetes later in life, compared with obese teens with no IGT. Finally, in those children who develop T2DM (Fig. 1d), decreased insulin sensitivity and the stress on the pancreas leads to a rapid decline in beta-cell function, so that even as their insulin sensitivity increases after puberty, permanent pancreatic insufficiency continues, and they remain with hyperglycemia into adulthood.

Fig. 1.

Patterns of insulin resistance and pancreatic beta-cell function throughout puberty. Proposed patterns of insulin resistance and beta-cell function are shown for different populations of children. ••••• Denotes beta-cell function, and – Insulin resistance in all graphs. a In lean children, there is a modest increase in insulin resistance as they progress through puberty, which is matched by pancreatic insulin secretions. b The same is seen in obese children with NGT, although they are more IR, with matched increases in insulin release. c In obese children who become NGT or IGT, as they progress through puberty, they return to an NGT state, but maybe at risk for redevelopment of IGT/FGT when they face another physiologic stressor. d In obese children who progress to developing diabetes, they develop IGT/FGT first, and then have continued insulin deficiency, and cannot compensate for their degree of IR, even as they progress to adulthood

Conclusions

In conclusion, pediatric obesity, IR, and T2DM are steadily increasing in prevalence worldwide. Lifestyle factors, including decreased physical activity and diets high in simple carbohydrates and fat, play a major role in the development of IR in obese youth. In youth with a combination of genetic predisposition, in utero exposure to maternal obesity or diabetes, or breastfeeding less than 6 months, this obesity may be associated with abnormal lipid partitioning and potentially beta-cell failure. Insulin sensitivity decreases as children go through puberty, but typically improves once puberty is completed. Not all children with IR go on to develop T2DM, as some return to euglycemic levels. Identifying those youth who will progress to T2DM will be important for targeting therapies and interventions.