The in vivo effects of the Pro12Ala PPARγ2 polymorphism on adipose tissue NEFA metabolism: the first use of the Oxford Biobank
- First Online:
- 644 Downloads
To investigate the phenotypic effects of common polymorphisms on adipose tissue metabolism and cardiovascular risk factors, we set out to establish a biobank with the unique feature of allowing a prospective recruit-by-genotype approach. The first use of this biobank investigates the effects of the peroxisome proliferator-activated receptor (PPAR) Pro12Ala polymorphism on integrative tissue-specific physiology. We hypothesised that Ala12 allele carriers demonstrate greater adipose tissue metabolic flexibility and insulin sensitivity.
Materials and methods
From a comprehensive population register, subjects were recruited into a biobank, which was genotyped for the Pro12Ala polymorphism. Twelve healthy male Ala12 carriers and 12 matched Pro12 homozygotes underwent detailed physiological phenotyping using stable isotope techniques, and measurements of blood flow and arteriovenous differences in adipose tissue and muscle in response to a mixed meal containing [1,1,1-13C]tripalmitin.
Of 6,148 invited subjects, 1,072 were suitable for inclusion in the biobank. Among Pro12 homozygotes, insulin sensitivity correlated with HDL-cholesterol concentrations, and inversely correlated with blood pressure, apolipoprotein B, triglyceride and total cholesterol concentrations. Ala12 carriers showed no such correlations. In the meal study, Ala12 carriers had lower plasma NEFA concentrations, higher adipose tissue and muscle blood flow, and greater insulin-mediated postprandial hormone-sensitive lipase suppression along with greater insulin sensitivity than Pro12 homozygotes.
This study shows that a recruit-by-genotype approach is feasible and describes the biobank’s first application, providing tissue-specific physiological findings consistent with the epidemiological observation that the PPAR Ala12 allele protects against the development of type 2 diabetes.
KeywordsGenetics Insulin sensitivity Obesity Peroxisome proliferator-activated receptors PPAR 2 Pro12Ala polymorphism Single nucleotide polymorphism Type 2 diabetes
homeostasis model assessment for insulin resistance
peroxisome proliferator-activated receptor
The gene for peroxisome proliferator-activated receptor, PPARG, plays a key role in insulin resistance and type 2 diabetes, as well as adipocyte differentiation and lipid metabolism . The Pro12Ala polymorphism of the PPAR2 isoform has an Ala12 allele frequency of around 0.12 in Caucasians . Epidemiological studies have shown an association of the Pro12 allele with an increased risk of developing type 2 diabetes, with a population-attributable risk of 25% [3, 4]. In vivo in humans, Ala12 allele carriers are more insulin-sensitive than Pro12 homozygotes [5, 6, 7, 8, 9]. Although PPAR activation by thiazolidinediones leads to insulin sensitisation, in vitro studies of the polymorphism have shown that the substitution of Ala for Pro results in reduced transcriptional activity of PPARG [5, 10].
The role of the Pro12Ala polymorphism in the development of obesity is less clear. A meta-analysis suggested that Ala12 carriers were more likely to have a high BMI than Pro12 homozygotes, an effect only identifiable among subjects with a BMI greater than 27 kg/m2 . Thus, Ala12 carriers appear to be heavier than Pro12 homozygotes, yet have a lower risk of developing type 2 diabetes. However, the mechanism underlying this apparent paradox is unclear.
Subjects with type 2 diabetes treated with the PPAR agonist rosiglitazone exhibit a similar unexpected dissociation between weight and insulin sensitivity: increasing weight is associated with an increase in insulin sensitivity. However, some weight gain may be due to fluid retention, as well as increasing fat mass. The exact mechanisms of insulin sensitisation are unclear, although we have previously documented changes in fatty acid metabolism in subjects treated with rosiglitazone .
We hypothesised that Ala12 carriers are more efficient in sequestering fatty acids in adipose tissue, and demonstrate greater metabolic flexibility in the storage and mobilisation of NEFAs in the postprandial period. Such a mechanism would increase insulin sensitivity and decrease the likelihood of developing type 2 diabetes, and would also be consistent with the increased adipose tissue mass observed in Ala12 carriers.
Such a complex genotype–phenotype relationship cannot be identified from stored fasting plasma. A useful approach could be to start by identifying carriers of certain genotypes from large cohort studies (i.e. genotyping many individuals). From such a large database of volunteers, closely matched subjects, differing as far as possible only in their genotype of interest, could be prospectively recruited into a study for detailed phenotyping. This recruit-by-genotype approach would allow the study of the influences of common genetic variants on metabolism. Prospective recruitment allows close baseline matching, along with comprehensive characterisation of the phenotype of interest, and adequately powered studies. Such biobanks have been set up to investigate other complex intermediate phenotypes [13, 14], but never before for insulin resistance and type 2 diabetes.
We therefore set out to establish a biobank, known as the Oxford Biobank (OBB), to allow the prospective investigation of genotype–phenotype relationships in adipose tissue and human metabolism. The OBB included DNA, anthropometric and biochemical data of healthy subjects taken at random from the general population. Recruited subjects would be willing to participate in detailed metabolic phenotyping in order to investigate insulin resistance and type 2 diabetes.
The possible functional impact of the PPAR2 Pro12Ala polymorphism, along with its high prevalence in the Caucasian population, made it a promising candidate for detailed tissue-specific metabolic phenotyping using the recruit-by-genotype approach. As PPARG is highly expressed in adipose tissue  and as subcutaneous abdominal adipose tissue is the major source of circulating NEFA in humans , we chose to phenotype subcutaneous abdominal adipose tissue function in Pro12 homozygotes and Ala12 allele carriers, with particular emphasis on fatty acid and triglyceride metabolism in vivo, in order to test our hypothesis that the Ala12 allele carriers would handle NEFA metabolism more efficiently in adipose tissue and skeletal muscle. These inferences have already been drawn, but have never been tested directly at the level of adipose tissue.
Subjects and methods
Using the comprehensive population register maintained by the UK National Health Service, an age-stratified random sample of 30- to 50-year-old men and women from Oxfordshire (total population 615,200 ) was selected. Exclusion criteria included mental or physical ill health, alcohol- or drug-related problems, and abnormal biochemical data as determined by history, examination, routine blood tests and information obtained from each subject’s primary care physician. Subjects attended a screening visit at the Clinical Research Unit, where blood tests were performed, including plasma glucose, insulin and triglyceride analysis. Basic anthropometric data were recorded, including weight, height and skinfold measurements. DNA was stored from the visit and consent obtained to allow subsequent genotyping for current and future genes of potential importance in adipose tissue metabolism. Samples were genotyped for the Pro12Ala polymorphism by PCR sequence-specific amplification, as previously described , using the allele-specific primers: 5′-TGAAGGAATCGCTTTCTGG-3′ and 5′-TGAAGGAATCGCTTTCTGC-3′ and the consensus primer 5′-GTCACCGGCGAGACAG-3′ to generate an allele-specific product of 357 nucleotides. The products were then visualised on a 1% agarose gel containing ethidium bromide.
Pro12Ala meal studies
From the OBB register, 24 healthy non-diabetic subjects were recruited by genotype. Twelve Ala12 heterozygotes and 12 matched Pro12 homozygotes were selected. No Ala12 homozygotes were used because no suitable subjects were available to participate. As smoking can affect insulin sensitivity, only non-smoking subjects were recruited. The study was approved by the Oxfordshire Research Ethics Committee and all subjects gave their written informed consent.
To assess the interaction of weight and genotype, we deliberately set out to recruit some subjects who were of normal weight and some subjects who were overweight into both genotype groups. Of the 24 subjects, 12 were overweight (six Pro12 homozygotes and six Ala12 carriers) and 12 were of normal weight (six Pro12 homozygotes and six Ala12 carriers). Overweight subjects had a BMI greater than 28.5 kg/m2 and were in the highest quartile of subjects in the OBB by weight. Normal-weight subjects had a BMI less than 25.5 kg/m2 and were in the lower half of subjects in the OBB by weight.
The 24 selected subjects attended the Clinical Research Unit for a full-day metabolic investigation. Vigorous exercise and alcohol were avoided for 24 h and food for 10 h before the study. During the study day, arteriovenous differences in metabolites across adipose tissue and skeletal muscle were measured, along with tissue blood flow. Body fat was measured by bioimpedence using a Bodystat 500 (Bodystat, Douglas, UK).
Metabolic investigation protocol
Recording arteriovenous differences across adipose tissue and skeletal muscle enabled the measurement of differences between the composition of venous blood draining a particular tissue and arterialised blood, thus reflecting that tissue’s metabolic activity. Inclusion of tissue blood flow into the calculations allows quantification of net metabolic activity. On the day of the study, serial blood samples were taken in the fasting state and for 6 h after metabolic provocation with a standardised mixed meal, a technique we have used elsewhere .
Arterialised blood was obtained from a vein draining a heated hand. Venous blood from muscle was taken retrogradely from a vein draining the deep structures of the forearm. Venous blood from adipose tissue was obtained from the superficial epigastric vein, as described previously . This vein drains subcutaneous abdominal adipose tissue with negligible contribution from other tissues . Oxygen saturation and ultrasonography were used to confirm adequate siting of the cannulae. Blood samples were taken simultaneously from all three sites. To prevent contamination of the blood from the forearm vein with blood from the hand, a wrist cuff was inflated to 200 mmHg for 3 min before taking samples. Two sets of baseline blood samples were taken 30 min apart. Patients then consumed a meal of a fat emulsion, skimmed milk and Rice Krispies (Kellogg Company, Manchester, UK) containing 40 g fat and 40 g carbohydrate and 600 mg [1,1,1-13C]tripalmitin. Further blood samples were taken for 6 h after the meal.
Plasma triglyceride, NEFA, glucose and insulin and blood glycerol concentrations were measured as described previously . Adiponectin was measured using ELISA (Linco, St Charles, MO, USA).
The postprandial increase in 13C in the circulation was derived from the [1,1,1-13C]tripalmitin in the meal. NEFAs and triglycerides were extracted from fasting and postprandial arterialised plasma samples for gas chromatography and gas chromatography–combustion–isotope ratio mass spectrometry. Enrichments of [13C]palmitate in plasma NEFA and triglyceride fractions were measured. The results were multiplied by the concentrations of palmitate–NEFA and palmitate-triglyceride, respectively to give the concentrations of the tracer in these fractions, as previously described . The rate of appearance in expired air of 13CO2 from [13C]palmitate was measured using gas chromatography–isotope ratio mass spectrometry, providing an index of whole-body fatty acid oxidation.
Statistical analyses, power and calculations
Adipose tissue blood flow was calculated as described previously . Arteriovenous and venoarterial differences in metabolite concentrations were calculated. Absolute flux was calculated as the product of the arteriovenous or venoarterial difference and adipose tissue blood flow . Lipoprotein lipase rate of action in vivo was calculated from triglyceride removal across each tissue . The rate of action of hormone-sensitive lipase (HSL) in adipose tissue in vivo (i.e. the rate of adipose tissue lipolysis) was calculated from the total adipose tissue glycerol release after subtraction of lipoprotein lipase rate of action .
The study was powered to detect a 12.5% difference in fasting NEFA output from adipose tissue, with a power of 90% at a statistical significance level of 0.05.
Data were analysed using SPSS for Windows v10 (SPSS UK, Chertsey, UK) and statistical significance was set at p<0.05 for all tests. Variables are presented as median (interquartile range) unless otherwise stated. Repeated measures ANOVA was used to identify time effects, drug effects and time b drug interactions. AUCs were compared using Wilcoxon signed rank tests.
Results of Oxford Biobank
Anthropometric, biochemical and haemodynamic data of subjects in the Oxford Biobank
Number of subjects
Frequency of Ala12 allele
Male (349 genotyped)
Female (420 genotyped)
Body mass index (kg/m2)
Waist circumference (cm)
Hip circumference (cm)
Systolic blood pressure (mmHg)
Diastolic blood pressure (mmHg)
The genotype-specific relationship between insulin sensitivity and features of the metabolic syndrome extended beyond biochemical parameters. Systolic and diastolic blood pressures in males correlated with insulin resistance in Pro12 homozygotes (systolic, p=0.001; diastolic, p<0.0001), but not in Ala12 carriers (systolic, p=0.75; diastolic, p=0.38).
Results of Pro12Ala meal studies
Characteristics of subjects in the Pro12Ala meal provocation study
Number of subjects
Body mass index (kg/m2)
Fasting insulin (pmol/l)
Fasting glucose (mmol/l)
Total cholesterol (mmol/l)
Fasting triglyceride (mmol/l)
Plasma apoB (g/l)
Systolic blood pressure (mmHg)
Diastolic blood pressure (mmHg)
Glucose and insulin
Postprandial glucose concentrations were the same in the Ala12 and Pro12 groups (p=0.35). In contrast, insulin concentrations were significantly lower in the postprandial period in Ala12 carriers than in Pro12 homozygotes (p=0.004).
Glucose uptake by forearm muscle and adipose tissue was the same in the Pro12 and Ala12 groups (Pro12 homozygotes, 0.51 μmol/min per 100 g tissue [0.34–0.81] vs Ala12 carriers, 0.67 μmol/min per 100 g tissue [0.45–0.73] [time-averaged AUC], p=0.97).
Hormone-sensitive lipase in adipose tissue
Both fasting and postprandial plasma triglyceride concentrations were similar between the groups (data not shown).
Triglyceride removal in adipose tissue (p=0.86) (data not shown) and forearm muscle (data not shown, p=0.22) did not differ between Pro12 homozygotes and Ala12 carriers.
To quantify the proportion of fatty acids derived from the meal, we measured concentrations of [1-13C]palmitic acid in the triglyceride fraction and NEFA fraction. [13C]Palmitic acid concentrations increased postprandially in the triglyceride fraction (data not shown), and also in the NEFA fraction (data not shown), as would be expected after a meal containing labelled triglycerides. Notably, the rise in [1-13C]palmitic acid concentrations in both the NEFA and the triglyceride fraction were the same in both the Ala12 carriers and the Pro12 homozygotes.
The rate of appearance of exhaled 13CO2 from [13C]palmitate did not differ between the Pro12 homozygotes and Ala12 carriers.
Fasting plasma adiponectin concentrations were not different between Ala12 carriers and Pro12 homozygotes (mean±SEM, 6.5±1.0 vs 6.9±0.9 μg/ml, p=0.70).
Forearm muscle blood flow was also higher in the Ala12 than in the Pro12 group (+22%, p=0.03) (Fig. 6d). This difference in muscle blood flow was only seen in normal-weight Ala12 subjects compared with normal-weight Pro12 subjects (+29%, p=0.04) (Fig. 6e); no difference could be seen between overweight Ala12 and overweight Pro12 subjects (+14%, p=0.30) (Fig. 6f).
This is one of the first studies to use a recruit-by-genotype approach to allow the prospective investigation of genotype–phenotype relationships in adipose tissue metabolism. Here, we show that, for investigation of the impact of a gene that is important in insulin resistance, such an approach was both feasible and successful.
Individuals recruited into the OBB should be considered as a highly motivated subset, although they are similar to those recruited into other large population-based studies (e.g. ). However, because the genotype frequencies of the polymorphism were in Hardy–Weinberg equilibrium, it is likely that a representative sample of the population has been recruited here.
Although there were no overall metabolic differences between Pro12 homozygotes and Ala12 carriers (Table 2), the metabolic response to the presence of insulin resistance did differ, suggesting a gene–environment interaction. Using the OBB register, insulin sensitivity showed the expected correlations with plasma triglyceride, apoB and HDL-cholesterol concentrations among male Pro12 homozygotes, whereas the corresponding correlations were absent among male Ala12 carriers (Fig. 2). Thus, male Ala12 carriers seemed to be protected against the metabolic consequences of insulin resistance. Such findings prompted us to metabolically characterise the Pro12Ala polymorphism in greater detail.
The meal study of the Pro12Ala polymorphism recruited age-, sex- and BMI-matched subjects from the OBB register. Despite this close matching, we observed significant differences in a number of physiological variables. We found that Ala12 carriers had lower NEFA concentrations in the fasting and late postprandial periods than the Pro12 homozygotes. The use of a mixed meal as a physiological stimulus, rather than insulin clamp techniques, allows assessment of normal physiological variations in diurnal NEFA concentrations. Such elevation of plasma NEFA concentrations may play an important role in the pathogenesis of type 2 diabetes [25, 26, 27].
Postprandial excursions of plasma insulin concentrations were lower in Ala12 carriers, despite identical postprandial plasma glucose concentrations. This occurred despite identical fasting concentrations of insulin and glucose. This can only be interpreted as a sign of increased insulin sensitivity in the Ala12 carriers, which is consistent with the findings of other groups [5, 6, 7, 8, 9, 28].
The suppression of HSL activity in adipose tissue is an insulin-sensitive process, and occurs as insulin concentrations rise in the postprandial period. Thus, the measurement of HSL rate of action gives an estimate of the tissue-specific insulin sensitivity. Adipose tissue HSL in Ala12 carriers was suppressed rapidly in the postprandial period, whereas the postprandial suppression HSL of Pro12 homozygotes was not significantly different from baseline, despite higher insulin concentrations than in the Ala12 carriers (Fig. 4). By measuring systemic labelled glycerol concentrations during a clamp, Stumvoll and colleagues found that Ala12 carriers suppressed their systemic concentrations of glycerol at lower insulin concentrations than Pro12 homozygotes . The mechanism of this is likely to be mediated by changes in HSL activity in adipose tissue, as shown by the tissue-specific measurements in our study.
One study found no difference in fasting NEFA concentrations between Pro12 homozygotes and Ala12 carriers . However, by careful phenotyping of matched subjects, we did find subtle differences in NEFA metabolism between Pro12 homozygotes and Ala12 carriers. Fasting and postprandial NEFA concentrations were lower in Ala12 carriers, despite unchanged NEFA release from subcutaneous abdominal adipose tissue. Could a difference in the generation of NEFAs from dietary triglycerides account for the difference in plasma NEFA concentrations observed? By measuring plasma concentrations of stable isotopes derived from the meal containing [1,1,1-13C]tripalmitin, we were able to assess the generation of NEFAs from dietary triglycerides; there was no difference between Ala12 carriers and Pro12 homozygotes. This use of stable isotopes addressed the question of whether there was an increased spillover of fatty acids from the intravascular lipolysis of chylomicron triglycerides. These data did not support that possibility.
Tissue blood flows were also affected by the Pro12Ala polymorphism. Adipose tissue blood flow tended to be higher in Ala12 carriers compared with the Pro12 homozygotes. This difference was due to higher adipose tissue blood flow in the obese Ala12 subgroup. We have previously shown a strong relationship between adipose tissue blood flow and insulin sensitivity , and these findings support the notion that the adipose tissue of Ala12 carriers is more insulin-sensitive than that of Pro12 homozygotes. There was a small but statistically significant increase in forearm muscle blood flow in the Ala12 carriers, but this was accounted for by differences between the lean subgroups. We have previously shown that increasing insulin sensitivity by PPAR activation increases adipose tissue blood flow, but does not affect muscle blood flow . Thus, the explanation for the observed increase in muscle blood flow among lean Ala12 carriers is unclear. PPAR activation is known to modify vascular responsiveness in vivo [31, 32, 33] and in vitro [34, 35, 36], but the effect of PPAR2 Pro12Ala polymorphism has never been assessed previously. Thus, Ala12 carriers appear to have a constitutively higher blood flow compared with Pro12 homozygotes. Such genotype-specific difference in nutritive blood flow is likely to impact on the metabolic function of adipose tissues and muscle [37, 38].
Some authors have suggested that the Pro12Ala polymorphism influences adipose tissue distribution, finding that Pro12 homozygotes have centrally distributed fat, whilst Ala12 carriers have increased peripheral fat [39, 40]. In the OBB database, skinfold and waist:hip measurements were not significantly different between the Pro12 homozygotes and Ala12 allele carriers, irrespective of sex.
Removal of NEFAs from the circulation occurs in muscle and liver. Here, we show that Ala12 carriers modulate their muscle NEFA clearance from fasted to fed and back to the fasted state (Fig. 3c). However, the muscle NEFA uptake of Pro12 homozygotes shows little change at the time of peak postprandial insulinaemia. This flexibility of NEFA metabolism across the forearm muscle suggests a genotype-specific difference in skeletal muscle NEFA handling. At first glance, such an effect is difficult to reconcile, considering that the Pro12Ala polymorphism is specific to PPAR2, whilst it is PPAR1 which is found in skeletal muscle . It is therefore likely that the effect in skeletal muscle is secondary to changes in systemic NEFA concentrations or other metabolic mediators. Such metabolic flexibility reflects the ability of tissues to switch from the use of one fuel to another and has been suggested as a more important marker of insulin sensitivity than absolute concentrations of metabolites [42, 43].
3-Hydroxybutyrate is generated by the liver from fatty acid oxidation and indirectly reflects NEFA handling in the liver; this was lower in Ala12 carriers than in Pro12 homozygotes. Whether this is the primary reason for the difference in plasma NEFA concentrations, or whether the lower NEFA oxidation in the liver merely reflected lower plasma NEFA concentrations is unclear. Interestingly, the measure of whole-body fatty acid oxidation, the rate of appearance in expired air of 13CO2 from [13C]palmitate, did not differ between Ala12 carriers and Pro12 homozygotes. Thus, the lower rate of fatty acid oxidation in the liver of Ala12 carriers may have been compensated for by a lower rate of fatty acid oxidation in the skeletal muscle.
Adiponectin is known to increase fatty acid oxidation and thus might be expected to influence NEFA concentrations. However, reports of the differences in plasma adiponectin concentrations between Ala12 carriers and Pro12 homozygotes are conflicting [44, 45, 46]. The differences are perhaps explained by the race of the subjects studied: among Japanese subjects, Ala12 carriers have lower plasma adiponectin concentrations than Pro12 homozygotes [44, 45], whilst among Caucasian subjects there is no difference between Pro12 homozygotes and Ala12 carriers . Our study of a Caucasian population is consistent with these findings.
The use of a mixed meal as the physiological stimulus to assess the effects of metabolic provocation has benefits compared with non-physiological insulin doses, as occurs when hyperinsulinaemic clamps are used. Many studies look at the association between measures of fasting insulin sensitivity (e.g. HOMA-IR) and the Pro12Ala polymorphism. We can see that in response to the physiological stimulus of a mixed meal, the main effects on insulin sensitisation are observed in the postprandial period. Another advantage of using a physiological stimulus of a meal is that the transition from the fed to the fasted state allows us to explore features of metabolic flexibility, an important characteristic lost in insulin resistance [42, 43]. There were subtle differences in metabolism between healthy Pro12 homozygotes and Ala12 carriers, including differences in plasma NEFA concentrations and tissue blood flow, along with increased insulin sensitivity. These tissue-specific physiological findings are consistent with the epidemiological observation that the PPAR Ala12 allele protects against the development of type 2 diabetes.
G. D. Tan was supported by the Medical Research Council and F. Karpe is a Wellcome Trust Senior Clinical Fellow. This project was supported by a Wellcome Trust Integrative Physiology Award. The Oxford Biobank was established through a research grant from the British Heart Foundation and is now being further developed by funding from EC Framework 6 (MolPAGE consortium LSHG-CT-2004).