This article describes phenotypes observed in a prediabetic population (i.e. a population with increased risk for type 2 diabetes) from data collected at the University hospital of Tübingen. We discuss the impact of genetic variation on insulin secretion, in particular the effect on compensatory hypersecretion, and the incretin-resistant phenotype of carriers of the gene variant TCF7L2 is described. Imaging studies used to characterise subphenotypes of fat distribution, metabolically healthy obesity and metabolically unhealthy obesity are described. Also discussed are ectopic fat stores in liver and pancreas that determine the phenotype of metabolically healthy and unhealthy fatty liver and the recently recognised phenotype of fatty pancreas. The metabolic impact of perivascular adipose tissue and pancreatic fat is discussed. The role of hepatokines, particularly that of fetuin-A, in the crosstalk between these organs is described. Finally, the role of brain insulin resistance in the development of the different prediabetes phenotypes is discussed.
KeywordsBrain insulin resistance Insulin Liver fat Phenotype Prediabetes Review Secretion Sensitivity
High saturated fat
Impaired glucose tolerance
Restricted saturated fat
Monocyte chemoattractant protein-1
Metabolically healthy obesity
Metabolically unhealthy obesity
Non-alcoholic fatty liver disease
Normal glucose tolerance
Tübingen Family Study
Tübingen lifestyle intervention program
The global increase in type 2 diabetes prevalence over recent decades puts a heavy health and socioeconomic burden on society. Lifestyle intervention with increased physical activity and a healthy diet is considered to be generally effective in preventing the development of diabetes [1, 2, 3, 4]. Unfortunately, the prevention studies carried out so far have shown that a substantial number of prediabetic individuals (that is, those with increased risk for type 2 diabetes) seem to be non-responders to lifestyle interventions; the number needed to treat amounted to 7 in the Finnish Diabetes Prevention Study (DPS) and the US Diabetes Prevention Program (DPP) [1, 2]. Based on these studies and on pathophysiology-based studies in our population [5, 6, 7], it appears that even those individuals who are able to reduce their body fat mass adequately show a lack of improvement in hyperglycaemia and insulin resistance. To improve the efficacy of lifestyle intervention programmes, a precise understanding of the pathophysiology and the different phenotypes of the prediabetic population is needed.
Compensatory hypersecretion of insulin: influence of genetic variation
Using our database, we assessed the extent to which genetic variation determines the ability to produce compensatory hypersecretion of insulin. A large number of type 2 diabetes loci are known today [14, 15, 16] and we studied genotype–phenotype associations for the strongest type 2 diabetes genes [17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41].
We indeed found associations between many of the genetic variants and different features of insulin secretion [17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39]. Several of the gene variants showed interaction with other gene variants, and the degree of the effects was dependent on ligands that induce insulin secretion, such as incretins or fatty acids [17, 18, 19, 20, 21, 22, 23, 24, 25]. However, after quantifying the effect of these gene variants and testing for additive effects between variants and genotype–ligand interaction, it appears that the effects on insulin secretion are very small and often non-significant in insulin-sensitive individuals [30, 31]. In insulin-resistant individuals, who have higher insulin levels due to the compensatory effort of the pancreatic beta cell, the reduction in insulin secretion in association with these gene variants is more pronounced and significant [30, 31]. The quantitative effect, however, remains quite small, reaching only 15–20% of the compensatory response observed in individuals with normal glucose tolerance (NGT) vs impaired glucose tolerance (IGT) vs diabetes. This suggests that the type 2 diabetes genes known so far have only minor effects on insulin secretion and that variation in these genes does not contribute much to the large difference in compensatory hypersecretion of insulin that is seen when NGT, IGT and diabetic individuals are compared (Fig. 1).
This phenomenon has also been described by other groups [43, 44, 45]. A recent pharmacogenetic study showed that homozygous T allele carriers of the rs7903146 SNP of TCF7L2 are partially resistant to therapy with dipeptidyl peptidase-4 (DPP-4) inhibitors, which are known to increase GLP-1 availability . Approximately 10% of individuals in our database are homozygous carriers of this T allele, and the gene variant probably contributes to an inability to upregulate insulin secretion. It is important to note that a reduction in glucose levels through lifestyle intervention can reverse the reduced insulin secretion [23, 47]. Therefore, attempts to lower glucose levels both by lifestyle intervention and by pharmacotherapy might be able to slow down the disease progression in this subgroup of prediabetic individuals. Clinical studies to test this hypothesis are on the way. The role of this gene variant in glucose-induced insulin secretion and glucose metabolism has been addressed in many studies [48, 49, 50, 51], some of which suggest that the gene variant affects glucose-induced insulin secretion and the conversion of proinsulin to insulin  as well as affecting glucose metabolism .
Body fat composition
Metabolically healthy and unhealthy obesity
A genetic predisposition seems to contribute to these phenotypes. Genome-wide association studies have identified a number of genetic variants for NAFLD . The I148M variant of PNPLA3, the gene encoding patatin-like phospholipase domain-containing protein 3, is most strongly associated with increased liver fat content (Fig. 4b) . Surprisingly, however, no insulin resistance is observed in individuals possessing this variant (Fig. 4c) . We found that this genotype appears to cause a different fatty-acid composition in the lipid stores of the hepatocytes—decreased levels of stearate and increased levels of polyunsaturated fatty acids (Fig. 4d) [61, 62].
Role of hepatokines
The different fatty-acid pattern might be responsible for an alteration in the interaction between hepatocytes and the cells of the immune system in the liver. In contrast to the metabolically benign fatty liver, the metabolically malignant fatty liver secretes an altered pattern of hepatokines . The hepatokine fetuin-A is associated with insulin resistance, diabetes and cardiovascular outcomes [63, 64, 65, 66, 67]. Saturated fatty acids, such as palmitate, stearate and myristate, have been found to increase hepatic fetuin-A mRNA and protein expression in the human liver cell line HepG2 by increasing NFkB binding to its promoter . Palmitate dose- and time-dependently increases the secretion of fetuin-A from HepG2 cells . Also, high glucose levels dose-dependently increase fetuin-A mRNA and protein expression in HepG2 cells via activation of the extracellular signal-regulated kinase-1/2 (ERK-1–ERK-2) signalling pathway . Finally, preliminary data suggests that exendin-4 may attenuate the expression of fetuin-A in HepG2 cells by improving palmitate-induced endoplasmic reticulum stress through AMP-activated protein kinase . The exact mode of action of fetuin-A is still not fully understood.
George Grunberger’s group was the first to show that fetuin-A inhibits signalling through the insulin receptor ) and we and others have confirmed fetuin-A’s inhibitory effect on the insulin receptor tyrosine kinase [reviewed in 63, 64, 72]. Downstream effects on stress kinases and NFkB were observed .
Another signalling pathway that is modulated by fetuin-A is the fatty-acid signalling pathway, through the Toll-like receptor. Based on mouse data, it has been proposed that fetuin-A acts as an endogenous ligand of Toll-like receptor 4 to promote lipid-induced insulin resistance . This concept seems to be relevant in humans, as we were able to show that, indeed, circulating fetuin-A levels and NEFA interacted to predict insulin resistance in participants of the TÜF study . These data support the concept that organ crosstalk through hepatokines plays a key role in the pathophysiology of prediabetes.
Perivascular adipose tissue
Perivascular fat cells seem to be a particularly important target of hepatokines, which might function as transducers of organ crosstalk [54, 74, 75, 76, 77, 78]. The whole-body MRI data led us to study another interesting fat compartment—perivascular adipose tissue (Fig. 3). Perivascular adipose tissue is a specific fat depot with impact on organ functions [54, 74, 75, 76, 77, 78]. The adipose tissue surrounding arteries seems to have an influence on whole-body insulin sensitivity. This effect is independent of other fat compartments like hepatic fat and visceral fat . Due to the strong effect on whole-body insulin sensitivity, we hypothesised that perivascular adipocytes would have specific characteristics that distinguish them from visceral or subcutaneous fat cells. This is indeed the case, as perivascular fat cells produce and secrete higher quantities of angiogenic factors, cytokines and chemoattractants like monocyte chemoattractant protein-1 (MCP-1) . Furthermore, these fat cells seem to be particularly susceptible to organ crosstalk signals from the fatty liver . The hepatokine fetuin-A stimulates, together with fatty acids, cytokine release and MCP-1 expression in these cells .
Fat cells are also found around the renal artery in the hilus of the kidney. As the amount of kidney fat in the hilus correlates with hypertension-inducible albuminuria in individuals with prediabetes [77, 78], we speculate that this fat compartment might be relevant in the pathogenesis of diabetic kidney disease. We further observed that these renal fat cells are particularly responsive to crosstalk signals from the fatty liver (i. e. fetuin-A). Thus, the fatty kidney might be a subphenotype of prediabetes that defines a higher risk of developing kidney disease in the context of MUHO and NAFLD. This is of course at the moment a pure speculation that has to be tested in prospective studies.
The fatty pancreas: non-alcoholic fatty pancreas disease
Response to lifestyle intervention
All these observations support a key role for NAFLD in the pathogenesis of insulin resistance and the progression from NGT to IGT and finally to type 2 diabetes. Furthermore, our lifestyle intervention study has clearly shown that lifestyle intervention has limited success in improving glycaemia in prediabetic individuals who have NAFLD . Therefore, other approaches to reduce liver fat content are extremely important.
Dietary intervention is a powerful tool with which to reduce liver fat [81, 82, 83, 84, 85, 86]. However, there is evidence that the susceptibility to carbohydrate-dependent induction of liver fat shows a large inter-individual variation , suggesting the existence of diet non-responders. A study in patients with biopsy-proven NAFLD showed that a 2 week administration of a very-low-carbohydrate diet (20 g/day) vs energy restriction (5000–6300 kJ/day) reduced hepatic triacylglycerol levels by a greater amount (−55% vs −28%, respectively), while weight loss was similar in both groups (−4.0 kg vs −4.6 kg) .
In another study carried out over 16 weeks in 52 individuals with obesity, insulin resistance and suspected NAFLD, a normal carbohydrate (60% carbohydrate, 25% fat, 15% protein) or moderately restricted carbohydrate (40% carbohydrate, 45% fat, 15% protein) diet again resulted in a similar decrease in body weight, daily insulin requirement and plasma liver enzymes levels. However, the moderately restricted carbohydrate intervention was associated with a larger decrease in insulin resistance and liver enzymes .
The effect of an isoenergetic diet with restricted fat, restricted saturated fat (LSAT) and restricted glycaemic index (GI) (LSAT: 23% fat [7% saturated fat], GI < 55) on liver fat content was compared with the effect of a high-fat, high-saturated fat (HSAT) and high-GI (HSAT: 43% fat [24% saturated fat] GI > 70) diet in an elderly population. In the LSAT group, but not in the HSAT group, liver fat content decreased significantly .
In most studies intake of n-3 polyunsaturated fatty-acid supplements is associated with a reduction in liver fat content, with the doses ranging from 0.83 to 6 g/day, and duration of therapy ranging from 8 weeks to 18 months [85, 86].
The microbiome is probably important as well, although targeted interventions are not feasible as yet.
The phenotype of non-response to exercise
As a consequence, differences in the adaptation of fuel oxidation are potentially responsible for these phenotypes.
Insulin action in the human brain: the phenotype of brain insulin resistance
For a long time the brain was not considered to be a classical target organ of insulin action. However, as early as 1978, Roth and colleagues showed that mouse and rat brain express high levels of insulin receptors [99, 100]. Later, around 2000, several groups showed that alteration of the insulin signalling chain in the brain by knockdown of the insulin receptor or docking proteins causes brain insulin resistance, leading to a diabetes-like phenotype in mice [101, 102, 103].
Brain insulin sensitivity: cause or consequence of adipose tissue distribution?
Effects of gestational diabetes on fetal brain: does primary brain insulin resistance exist?
Brain insulin resistance can already be found in young obese people [106, 107, 108, 109, 110, 111]. This observation allows one to speculate that brain insulin resistance might indeed precede the development of obesity. To further test this hypothesis we studied fetal brain development in pregnancies of insulin-sensitive mothers, insulin-resistant mothers and mothers with gestational diabetes [118, 119]. Fetal brain functions were tested by fetal MEG (fMEG) and the findings suggested that indeed the metabolic situation of the mothers might influence fetal brain insulin sensitivity. It seems therefore conceivable that brain insulin resistance is induced already in utero. Further studies are required to show whether this leads to altered behaviour, altered eating habits and altered weight gain in the postnatal life of these children.
Brain insulin resistance: starting point for organ crosstalk defining prediabetic phenotypes?
I wish to thank all my collaborators from the University Hospital and the University of Tübingen, Germany for their consistent support over many years. I am particularly grateful to N. Birbaumer, C. D. Claussen, A. Fritsche, B. Gallwitz, F. Gerst, M. Heni, M. Hrabe de Angelis, A. Königsrainer, S. Kullmann, R. Lehmann, J. Machann, F. Machicao, A. Peter, H. Preißl, K. Rittig, N. Stefan, F. Schick, E. Schleicher, D. Siegel-Axel, H. Staiger, S. Ullrich and C. Weigert.
Duality of interest
The author declares that there is no duality of interest associated with this manuscript.
The author was the sole contributor to this paper.
- 16.DIAbetes Genetics Replication and Meta-analysis (DIAGRAM) Consortium, Asian Genetic Epidemiology Network Type 2 Diabetes (AGEN-T2D) Consortium, South Asian Type 2 Diabetes (SAT2D) Consortium et al (2014) Genome-wide trans-ancestry meta-analysis provides insight into the genetic architecture of type 2 diabetes susceptibility. Nat Genet 46:234–244CrossRefGoogle Scholar
- 24.Stefan N, Fritsche A, Häring H, Stumvoll M (2001) Effect of experimental elevation of free fatty acids on insulin secretion and insulin sensitivity in healthy carriers of the Pro12Ala polymorphism of the peroxisome proliferator—activated receptor-gamma2 gene. Diabetes 50:1143–1148PubMedCrossRefGoogle Scholar
- 35.Stancáková A, Pihlajamäki J, Kuusisto J et al (2008) Single-nucleotide polymorphism rs7754840 of CDKAL1 is associated with impaired insulin secretion in nondiabetic offspring of type 2 diabetic subjects and in a large sample of men with normal glucose tolerance. J Clin Endocrinol Metab 93:1924–1930PubMedCrossRefGoogle Scholar
- 43.Pilgaard K, Jensen CB, Schou JH et al (2009) The T allele of rs7903146 TCF7L2 is associated with impaired insulinotropic action of incretin hormones, reduced 24 h profiles of plasma insulin and glucagon, and increased hepatic glucose production in young healthy men. Diabetologia 52:1298–1307PubMedCrossRefGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.