Introduction

A number of studies have identified an inflammatory response in the postprandial period.1, 2, 3, 4, 5 Similarly, markers of oxidative stress are elevated following glucose,2, 4, 5 fat4, 6 and mixed-meal challenge.2, 3, 7, 8, 9 Postprandial inflammation and oxidative damage are augmented and prolonged in obese subjects and subjects with type 2 diabetes mellitus (T2DM).3, 8 Postprandial hyperglycaemia and hyperlipidaemia have been identified as independent cardiovascular risk factors,10 and are emerging as possible therapeutic targets.

The concentration of adiponectin, an anti-inflammatory and insulin-sensitizing adipokine, is reduced in obesity and T2DM.11 Preliminary work in our laboratory suggests that adiponectin may be nutritionally regulated,12 supporting the hypothesis that adiponectin may act in an anti-inflammatory manner following food intake. Studies evaluating adiponectin following a mixed-meal have reported variable findings,13, 14, 15, 16, 17, 18, 19 while there are limited data on the postprandial multimeric distribution of adiponectin.20, 21 The high molecular weight (HMW) adiponectin isoform is likely the most metabolically active11 and therefore is of particular interest.

This study evaluates the inflammatory and oxidative stress response following a high-fat mixed-meal in three groups of men: (i) lean (ii) obese non-diabetic and (iii) T2DM, with a focus on postprandial regulation of total and HMW adiponectin.

Materials and methods

Our group has previously published a substudy evaluating hemodynamic responses to a high-fat meal in these same three groups of men.22

Subjects

Caucasian, non-smoking males were recruited by local advertisement and university databases.

Lean subjects (n=10) had a body mass index of 20–25 kg/m2, waist circumference <94 cm and fasting blood glucose <6 mmol/l. Obese subjects (n=10; body mass index ⩾30 kg/m2) completed an oral glucose tolerance test to exclude T2DM. Exclusion criteria in lean and obese groups included any abnormality in the complete blood exam, biochemistry, renal or liver function tests. None of the lean or obese subjects were taking prescribed medications.

Subjects with T2DM were diagnosed according to established criteria23 and there were no body mass index criteria. Subjects were excluded if they had an abnormal resting electrocardiogram or were taking antihypertensive medications. Medications taken by subjects in the T2DM cohort included: sulphonylureas, metformin, statin and proton pump inhibitors. Diabetic therapy was stable for 3 months leading up to the study and statin therapy was withheld for 1 week before the study. Oral hypoglycaemic therapy was not taken before the trials. Subjects with T2DM were excluded if they had known retinopathy, renal dysfunction (serum creatinine concentration above the reference range given by laboratory or microalbuminuria), liver dysfunction (liver enzymes greater than twice the upper limit of normal) or known cardiovascular disease.

All subjects underwent clinical screening and were excluded if any signs or symptoms of cardiovascular disease were identified or if baseline blood pressure was >150/90 mm Hg. Anthropometric data were obtained in the fasting state. Fat mass was assessed by bioimpedance analysis (Bodystat 1500, Isle of Man, British Isles) by one operator.24

All subjects refrained from antioxidants and vitamins during the week before each trial. Subjects were excluded if they had been on topical or systemic steroids within the previous 2 months or had been systemically unwell in the week before the testing.

The study complied with the Declaration of Helsinki. All subjects provided informed, written consent. Ethics approval was obtained through the human research ethics committees at the Princess Alexandra Hospital and the University of Queensland.

Study design

Subjects were asked to avoid alcohol, caffeine and strenuous exercise for 24 h preceding each trial. Subjects presented at 0800 hours following a 12-h fast on two occasions, separated by at least 4 days. Subjects were provided with either water (control trial) or high-fat meal in random order. The meal consisted of a bacon and egg muffin, two hash browns and caramel flavoured milk drink (250 ml of 4% fat milk, four teaspoons skim milk powder and one tablespoon caramel flavour). Energy provided by the high-fat meal was 4136 kJ; the meal consisted of the following proportions of macronutrients: 57.5 g fat and 29 g saturated fat (51.5% total energy); 35 g protein (14.5% total energy) and 83 g carbohydrate (34% total energy). During the control trial, subjects consumed 50 ml of room-temperature water each hour. Subjects were studied for 6 h following both meal and water intervention.

The trials took place in a quiet, temperature-controlled room. Subjects sat upright for meal ingestion but were recumbent during the course of the study.

An indwelling cannula was inserted into the antecubital or brachiocephalic vein and following at least 15 min of rest, baseline blood samples were taken at 0830 hours. Subjects started eating their meal at 0845 hours and finished eating within 15 min. Time zero was taken as 15 min after subjects started eating their meal, or they had consumed the first 50 ml of water (0900 hours). Blood samples were collected every 30 min during the first 3 h of the study and hourly thereafter. Saline was used to keep the line patent after blood sampling and replace the volume of blood taken (<200 ml during the course of each trial).

Biochemical variables

All blood samples were analysed through the Chemical Pathology Department at the Princess Alexandra Hospital, Brisbane, Australia. Plasma glucose and lipid profile (total cholesterol, high-density lipoprotein, and triglycerides) were analysed on Beckman DxC800 general chemistry analysers. The low-density lipoprotein (LDL) was calculated using the Friedewald equation. Serum insulin was analysed on the Beckman DxI800 immunoassay analyser (Beckman Coulter Diagnostics, Sydney, NSW, Australia).

Adiponectin

Serum concentrations of total and HMW adiponectin were measured in duplicate using ELISA (Linco Res Inc., St Charles, MI, USA and Fujirebio Malvern, PA, USA, respectively) according to the manufacturers’ instructions.12, 25 The inter- and intra-assay coefficient of variation (CV) for total adiponectin was 3.4 and 6.6%, respectively. The inter- and intra-assay CV for HMW adiponectin was 9.1 and 7.6%, respectively.

Markers of inflammation—IL6, TNFα, hsCRP

Plasma concentrations of interleukin-6 (IL6) and tumour necrosis factor (TNFα; R&D Systems, Minneapolis, MN, USA) were analysed using Quantikine High-Sensitivity ELISA kits according to the manufacturer’s instructions. IL6 and TNFα were measured in duplicate, and all samples for each participant were measured on the same microplate. High-sensitivity C-reactive protein (hsCRP) was analysed in single using an immunoturbidometric automated technique (Kamiya, Seattle, WA, USA) on an automated analyser (Hitachi 717, Santa Clara, CA, USA). The intra-assay CV for IL6 and TNFα was 5.8 and 8.5%, respectively. The inter- and intra-assay CV for hsCRP was 2.2 and 0.9%, respectively.

NFκB activation in peripheral blood mononuclear cells (PBMCs)

PBMCs were extracted using Ficoll–Paque PLUS gradient (Amersham, Uppsala, Sweden), protein concentration of the extract was obtained using a Bradford-based assay and the binding of nuclear factor (NF)κB p65 was measured on nuclear extracts using the TransAM NFκB p65 Transcription Factor Assay kit (Active Motif, Carlsbad, CA, USA) according to the manufacturer’s instructions. The inter- and intra-assay CV for this assay was <10%.

Markers of oxidative stress—oxLDL, protein carbonyl

Plasma concentrations of oxidised low density lipoprotein (oxLDL) were measured using an ELISA (Mercodia, Uppsala, Sweden) according to the manufacturer’s instructions. Samples were measured in duplicate and all samples for each participant were measured on the same microplate. Plasma concentrations of protein carbonyl were determined using an in-house assay based on the methodology from Cayman Chemical (Ann Arbor, MI, USA). The inter- and intra-assay CV for oxLDL was 4.7 and 6.3%, respectively. The inter- and intra-assay CV for protein carbonyl assay was 8.5 and 4.7%, respectively.

Statistical analysis

Power calculations were undertaken to estimate sample size requirements. These calculations were based on the literature for total adiponectin in healthy subjects,15 using STATA version 10 (StataCorp LP, College Station, TX, USA). The ‘sample size’ command was used, and data for the expected change in adiponectin were imputed. A sample size of 8 subjects was estimated to give 80% power to detect a similar postprandial change in total adiponectin (a 13% postprandial decrease: 7.0±1.8 μg/ml–6.1±1.8 μg/ml), and 10 subjects per cohort were recruited.

Data are given as mean±s.d. (normally distributed) and median, interquartile range (skewed distribution). Pharmacokinetic variables utilized include Cmax (maximum concentration), Tmax (time to Cmax) and the time to return to baseline (defined as±10% of baseline value). The postprandial incremental area under the curve was calculated utilizing the trapezoidal method: area under the curve for water values over 360 min were subtracted from the total postprandial area under the curve. Insulin was only measured at time 0 and 30 min for the water control trial, and remaining values were extrapolated.

Comparison of baseline characteristics was performed using one-way analysis of variance for normally distributed data and Kruskal–Wallis rank test for skewed data. Comparisons within and between groups were done using the paired t-test for normally distributed data and the Wilcoxon signed-rank test for skewed data. In order to evaluate whether differences across groups persisted after correcting for age, analysis of covariance was used with age as a covariate for the relevant parameters. Analyses were performed using STATA version 10 (StataCorp LP) and statistical significance was set at P<0.05; the Holm step-down method was used to correct for multiple comparisons.

The postprandial total adiponectin, HMW adiponectin and markers of inflammation and oxidative stress were also analysed using a random coefficient model in SAS version 9.2 (SAS Institute, Inc., Cary, NC, USA). The response of the model was the analyte of interest (i) following water control, (ii) following a high-fat meal and (iii) the difference between responses to the meal and water control. Independent variables included group, time and the interaction of time and group. Owing to the differences in body composition observed between obese and T2DM subjects, analyses controlling for fat mass were also conducted. For the models that were significant for the meal-water response, age and fat mass were subsequently included in the model as independent variables. Results were unavailable at one time point for IL6 (T2DM subject following water trial) and p65 (lean subject following meal). As baseline values were not consistent between trials in all subjects, change from baseline was used in these analyses. Uncorrected P-values are reported for the random coefficient model.

Results

All 10 subjects from each cohort completed both arms of the study.

Baseline data

The baseline characteristics of the three groups are summarized in Table 1.

Table 1 Baseline subject characteristics
Full size table

Postprandial metabolic response

The postprandial metabolic responses are summarized in Table 2 and Figure 1.

Table 2 Postprandial metabolic responses—glucose insulin and triglycerides in lean (n=10), obese (n=10) and T2DM (n=10)
Full size table
Figure 1
figure 1

Postprandial glucose, insulin and triglycerides following high-fat meal and water control in lean (n=10), obese (n=10) and subjects with T2DM (n=10). Water (··●··), high-fat meal (-▪-); data are represented as mean±s.e.m.

Full size image

Postprandial total and HMW adiponectin

Postprandial total and HMW adiponectin responses are summarized in Figure 2.

Figure 2
figure 2

Postprandial total and HMW adiponectin—change from baseline following high-fat meal and water control in lean (n=10), obese (n=10) and subjects with T2DM (n=10). Water (··●··), high-fat meal (-▪-); data are represented as mean±s.e.m. The total and HMW adiponectin responses changed over time in all groups following both water and meal trials (random coefficient model). The only difference in response between groups was the total adiponectin response following the water trial with the lean group having a significantly different overall response compared with T2DM (P=0.04). When the water data were taken into account there was no overall change following the meal intervention in any group.

Full size image

Total adiponectin responses following water changed over time (P=0.06) and there was a significant difference between the response of lean and T2DM cohorts (P=0.04). Response in the obese group was intermediate but not significantly different compared with either the lean or T2DM cohort. Total adiponectin responses following the high-fat meal also changed over time (P=0.03), however, the response slopes were not significantly different across the three groups. When the difference between water and meal response was evaluated, there were no time or group effects apparent.

The HMW adiponectin responses changed over time following both water and meal (P=0.001 and P=0.05, respectively), however, there was no difference in responses across the three groups. When the difference between water and meal response was evaluated, there were no time or group effects apparent.

Postprandial inflammation—IL6, TNFα, hsCRP and PBMC NFκB activation

Inflammation following water control and high-fat meal are summarized in Figure 3.

Figure 3
figure 3

Postprandial inflammation—mean concentrations following high-fat meal and water control in lean (n=10), obese (n=10) and subjects with T2DM (n=10). Water (··●··), high-fat meal (-▪-); water (white), high-fat meal (black); data are represented as mean±s.e.m. for all parameters except p65, for which a box plot is provided. n=9 for IL6. IL6: when the difference between water and high-fat meal responses were evaluated, there was a significant trend indicating that overall concentrations of IL6 changed over time (random coefficient model) and that there was a difference in response between groups (P=0.098). Overall, concentrations of IL6 increased in lean subjects but decreased in the T2DM cohort (P=0.03); TNFα: there was no change in concentrations of TNFα following water control or high-fat meal (random coefficient model); hsCRP: when the difference between water and high-fat meal responses were evaluated, there was a significant trend indicating that concentrations of hsCRP changed over time (random coefficient model) and that there was a difference in response between groups (P=0.08). The concentrations of hsCRP decreased in lean subjects, but increased in the obese group (P=0.03); p65: there was no change in concentrations of p65 following water control or high-fat meal (random coefficient model).

Full size image

Interleukin-6

The concentrations of IL6 increased following water in all cohorts (P=0.0006). Subjects with T2DM had a greater increase in IL6 following water compared with lean subjects (P=0.04). Concentrations of IL6 following the high-fat meal also increased in all cohorts (P<0.0001), however, there was no difference in response slopes across the three groups. Finally, when the difference between water and high-fat meal responses were evaluated, there was a trend indicating that concentrations of IL6 changed over time and that there was a difference in response between groups (P=0.098). Indeed, concentrations of IL6 increased in lean subjects but decreased in the T2DM cohort (P=0.03). Age and fat mass were not significant in the model evaluating the difference between the water control and high-fat meal response. However, inclusion of fat mass in the model resulted in a loss of the group effect of change in IL6 post water and post meal.

Tumour necrosis factor-α

There was no change in concentrations of TNFα following water control or high-fat meal.

hsCRP

Concentrations of hsCRP following water control changed over time (P=0.02). A decrease in hsCRP concentrations was observed in the obese cohort, which was significantly different compared with both the lean and T2DM groups in whom an increase was seen (P=0.002 and P=0.04, respectively). Concentrations of hsCRP following the high-fat meal increased over time (P=0.0003) in all three groups, however, there was no difference in the response across the cohorts. Finally, when the difference between water and high-fat meal responses were evaluated, there was a trend indicating that concentrations of hsCRP changed over time and that there was a difference in response between groups (P=0.08). The concentrations of hsCRP decreased in lean subjects, but increased in the obese group (P=0.03). Age and fat mass were not significant in the model evaluating the difference between the water control and high-fat meal response. Furthermore, controlling for fat mass did not significantly alter the hsCRP results following water control or high-fat meal in any of the groups.

PBMC NFκB activation (p65)

There was no change in concentrations of p65 following water control or high-fat meal.

Postprandial oxidative stress—oxLDL, protein carbonyl

Markers of oxidative stress following water control and high-fat meal are summarized in Figure 4.

Figure 4
figure 4

Postprandial oxidative stress—mean concentrations of protein carbonyl and oxLDL following a high-fat meal and water control in lean (n=10), obese (n=10) and subjects with T2DM (n=10). Water (··●··), high-fat meal (-▪-); water (white), high-fat meal (black); data are represented as mean±s.e.m. for protein carbonyl while a box plot is provided for oxLDL. oxLDL: there was no change in concentrations of oxLDL following water control or high-fat meal (random coefficient model); protein carbonyl: when the difference between water and meal response was evaluated there were no time or group effects apparent (random coefficient model).

Full size image

oxLDL

There was no change in concentrations of oxLDL following water control or high-fat meal.

Protein carbonyl

Protein carbonyl concentrations increased following both water and high-fat meal in all groups (P=0.001 and P=0.003). However, the responses were not statistically significantly different across the three cohorts and when the difference between water and meal response was evaluated there were no time or group effects apparent.

Discussion

This study evaluated the metabolic, adiponectin, inflammatory and oxidative stress response to a high-fat meal and water control in obesity and T2DM. Overall, there was no significant postprandial change in total or HMW adiponectin. Postprandial IL6 decreased in subjects with T2DM but increased in lean subjects, whereas hsCRP decreased in lean and increased in obese subjects. There was no postprandial change in TNFα, PBMC NFκB activation (p65) or markers of oxidative stress.

This study does not support the induction of significant postprandial inflammation or oxidative stress and regulation of adiponectin was not observed under these conditions.

Postprandial adiponectin

During the water control total adiponectin concentrations changed over time and there was a significant difference between the lean subjects and subjects with T2DM. Although not the primary purpose of this study, the findings from the water control would support the presence of diurnal variation in total adiponectin. Further, this diurnal variation appears to differ in lean compared with T2DM subjects. Published literature reports an early morning nadir and peak in late morning in total adiponectin in lean subjects.26 There are limited data on adiponectin regulation in subjects with metabolic dysfunction, however, a small study identified blunted pulsatility of adiponectin in morbidly obese subjects, which was restored following weight loss surgery.27

There have been inconsistent reports on postprandial concentrations of total adiponectin, however, there is some literature supporting a modest (10–15%) decrease in total adiponectin 4–8 h following a mixed meal.15, 16, 18, 19 There is limited information regarding postprandial HMW adiponectin.20, 21 Only one study that reported postprandial changes in adiponectin concentration of >10% included a control arm (subjects were studied three times following mixed-meals of various macronutrient compositions).15 It is not clear, however, whether the group response following the three meals was in fact statistically significantly different and indeed the results for these postprandial studies may just reflect diurnal variation in adiponectin.

Postprandial inflammation and oxidative stress

Postprandial concentrations of IL6 increased in lean subjects and decreased in subjects with T2DM. Conversely, the obese group showed an overall postprandial increase in hsCRP concentrations, whereas a decrease was observed in lean subjects. There was no discernable change in p65 or TNFα concentrations following water or high-fat meal intervention. Similarly, there was no postprandial change in markers of oxidative damage.

The findings of this study are difficult to reconcile, particularly the divergent IL6 and hsCRP responses seen in the lean cohort. Postprandial IL6 has been evaluated in a number of studies, with mixed results.2, 9, 28, 29, 30 No postprandial increase in IL6 was observed following a fat challenge in two well-designed studies, both of which used meal (but not water) controls.28, 30 Manning et al.28 actually observed a postprandial decrease in inflammatory markers and Dixon et al.31 identified an IL6 response to an indwelling catheter in their meal study (there was minimal postprandial change in samples obtained via venepuncture). Although intervention studies have shown modification of the postprandial inflammatory response following treatment with simvastatin/irbesartan2 and mitiglinide9 in T2DM, these studies lacked a control arm.

Similarly, there have been conflicting reports of postprandial hsCRP1, 32 and TNFα29, 32 concentrations, and there have been fewer studies evaluating postprandial activation of PBMC NFκB.5, 8, 33 A significant 2.5-fold induction of p65 was observed 2 h following a high-fat meal in lean subjects,5 whereas 50 g of high glycaemic-index carbohydrate induced a 1.5-fold induction in p65 activity.33 Patel et al.8 observed minimal postprandial change in PBMC NFκB activation in lean and obese subjects following a high-fat meal and a recent study failed to identify a change in mRNA expression of NFκB in PBMCs following an oral fat load.34

A number of studies have identified elevations in markers of postprandial oxidative stress,2, 4, 6 although not all agree,33, 35 and in some studies, markers of oxidative stress and inflammation change independently,30, 32, 34 which is not easily reconciled. Neri et al.35 report a postprandial increase in IL6, but unchanged TNFα and malonyldialdehyde, while Fernandez-Real et al.36 report increased markers of oxidative stress, decreased antioxidant defence, yet unchanged postprandial hsCRP. Postprandial sampling of interstitial adipose tissue fluid identified a significant postprandial upregulation in secretion of IL6 (5 × fold), but no change in TNFα—serum concentrations of these cytokines were not reported.37 Finally, a recent study evaluated postprandial changes in inflammatory profile expression in PBMCs and identified an upregulation in postprandial IL6 and TNFα, which was not reflected in serum.38

Strengths and limitations

There are few studies that report postprandial metabolism across more than two groups, and a significant strength of this study includes the evaluation of three well-defined patient populations following the same mixed-meal with relatively frequent sampling over a prolonged period of time. The lean and obese subjects had no known medication confounders, whereas the subjects with T2DM were on limited medications and significant co- morbidities were excluded. The use of a water control was critical to the design of the study in light of the potential for catheter-induced inflammation and diurnal regulation of a number of the analytes measured.

This study was limited to males and results cannot be extrapolated to females. Subjects with T2DM were significantly older than lean subjects, and obese subjects had a significantly higher body mass index compared with those with T2DM. Analyses were adjusted for age, however, and fat mass was included as a covariate to further explore potential difference because of differences in bodyweight composition. The group effect of change in IL6 post high-fat meal and water control was lost when fat mass was included as a covariate. However, inclusion of fat mass as an independent variable did not significantly alter the results for the other analytes. Finally, the non-diabetic obese subjects and those with T2DM were selected for their relative ‘health’ and these cohorts therefore may have differed from previously studied subjects with metabolic dysfunction.

It is possible that subtle postprandial alterations were missed because of small sample size. Power calculations undertaken before the commencement of the study were based on published data for adiponectin15 as this was the primary focus of the study. Esposito et al.15 reported a similar inter- and intra-assay CV for both total adiponectin and cytokines. This study largely concentrated on systemic markers of inflammation and oxidative stress, and although cellular response was evaluated through PBMC activation (p65), it is conceivable that postprandial inflammation and oxidative stress status may be tissue specific and best appreciated at a local level.

The markers of inflammation and oxidative stress were chosen following review of the literature. Oxidative stress in particular is difficult to measure, and indirect markers have limitations. It is therefore possible that other markers of oxidative stress and inflammation may have been more useful. In addition, p65 was measured to document NFκB activation in PBMCs, however, there are other members of the NFκB family, which may be differentially regulated and potentially be of interest in the postprandial period. Finally, it is possible that the NFκB data may have been influenced by changes in the ratio of lymphocytes and monocytes in the PBMC pellet, which were not evaluated in this study. However, previous reports suggest39 that there are minimal postprandial changes in lymphocyte and monocyte composition.

Conclusion

This study does not support significant postprandial regulation of total or HMW adiponectin. In addition, there was no consistent significant induction of an inflammatory or oxidative stress response to a high-fat meal in lean subjects, obese subjects or subjects with T2DM. Although not the primary aim of this study, a different profile of basal total adiponectin was observed in subjects with T2DM compared with lean subjects, suggesting that this adipokine might be differentially regulated in the presence of metabolic dysfunction. Our findings highlight the importance of appropriate controls for meal studies and dictate that meal data obtained in the absence of rigorous controls be questioned.