1,2-dimyristoleoyl-sn-glycero-3-phosphocholine (PC 14:1/14:1) (where 14 and 1 mean the summed number of carbons and double bonds, respectively) purchased from Avanti Polar Lipids (Alabaster, AL). With the exception of ammonia (25%), which was analytical grade, all solvents (e.g., hexane, isopropanol, formic acid, and others) were HPLC grade. All solvents were purchased from Merck (Darmstadt, Germany).

Eleven subjects with T2DM were recruited from Theptarin hospital, Bangkok, Thailand. The inclusion criteria were as follows: 21–60 years of age, male or female, and diagnosed with T2DM. Patients with an acute infection, hepatic/renal disease, or were using insulin or lipid-lowering drugs were excluded. Ten age- and sex-matched control subjects with normal blood glucose and lipid levels were randomly selected from among the researchers’ laboratory staff. The informed consent forms and study protocol were in accordance with the Declaration of Helsinki and were reviewed and approved by the ethical committee on research involving human subjects from Theptarin Hospital, Bangkok, Thailand.

Subjects fasted overnight for 12 h prior to the designated study day. During the day of the study, no drugs or medications were taken until the experiment was completed. After submitting fasting blood samples, the subjects were given a single standard high-fat meal in the form of a milkshake and buttered bread. The meal contained 40 g fat/m² body surface area and had an energy distribution of 72% fat, 24% carbohydrate, and 4% protein. The polyunsaturated fatty acid (PUFA)/saturated fatty acid (SFA) ratio and monounsaturated fatty acid (MUFA)/SFA ratio were 0.08 and 0.46, respectively. The meal was ingested within 20 min and water was allowed ad libitum. Blood samples were collected again 4 h after ingestion of the meal. A 4 h time point was selected based on in vitro and in vivo studies in erythrocytes that showed rapid exchange or incorporation of meal-derived fatty acid into lipid fractions of erythrocytes [12, 16] and because postprandial triacylglycerol is at its peak (i.e., 4–5 h after a lipid-enriched meal administration) [17, 18].

Serum glucose and lipids were measured by standard enzymatic techniques. Low-density lipoprotein (LDL) cholesterol was calculated using the Friedewald equation [19]. Hemoglobin A₁C (HbA₁C) was analysed by an immunoturbidity assay. Serum insulin was determined by using the immunochemiluminometric assay. Insulin resistance was estimated by the homeostasis model assessment insulin resistance (HOMA-IR) index derived from fasting plasma insulin and glucose concentrations (insulin (µU/ml) × glucose (mg/dl)/405) [20].

Venous blood samples were taken from subjects and placed into a tube pre-treated with EDTA. Plasma was immediately separated by centrifugation at 3,000 rpm for 10 min at 4°C. After removing plasma, the erythrocytes were washed with an ice-cold isotonic solution containing 0.15 M NaCl. The buffy coat was removed by aspiration after each wash. The packed erythrocyte sample was divided into small aliquots of 0.5 ml and kept under nitrogen at −80°C.

A Tensor 27 FTIR spectrometer (Bruker Optics GmbH, Ettlingen, Germany), equipped with a high throughput extension (HTS-XT) accessory, was used. Erythrocytes were prepared according to the methods described by Petibois et al. with slight modification [21]. Twenty microlitres of washed, packed erythrocytes were diluted with 180 µl of water and mixed for 60 s. Duplicated dry films were made for each sample by applying 5 µl of the hemolysate to 96 wells in a silicon microplate. The hemolysate films were left to dry at 22°C in desiccators under vacuum to prevent oxidation. After the films were dried, the microplate was mounted in the FTIR-HTS-XT to enable acquisition of infrared spectra. The system was purged with dry nitrogen to reduce water vapour and CO₂. All samples were analysed simultaneously. The infrared spectra were collected from 400 to 4000 cm⁻¹ using a Deuterated triglycine sulphate detector. Each acquisition consisted of 512 interferogram scans with a spectral resolution of 4 cm⁻¹. A Blackman–Harris three-term apodisation and a zero-filling factor of two were applied. Before recording the erythrocyte spectra, the spectrum of a blank well of a silicon plate was acquired and subtracted later from the spectra of the dried erythrocyte films.

The data were analysed using Optics User Software, Version 6.0 (Bruker Optics GmbH, Ettlingen, Germany). Second derivative and vector normalisation were applied to all spectra to resolve and enhance the intensity of the weak bands, especially the olefinic band, and also to reduce the variation in film thickness. The second derivative of the original spectra was used to identify the peak frequencies of characteristic components. Since the band intensity or integrated area derived from the second derivative spectra is directly proportional to the concentration [22], the integrated areas of C-H stretching bands were calculated from the second derivative spectra for preliminary investigation of the changes relative to the fasting state. Assignments for the major bands observed in FTIR spectra of erythrocytes are given in Table 1.

An aliquot of packed erythrocytes was prepared for lipid extraction according to the method of Blight and Dyer [23] with slight modification. PC 14:1/14:1 was added to the samples as an internal standard. After extraction, the organic phase was collected and dried by evaporation under a nitrogen stream. The lipid extracts were stored dry at −20°C until analysis. An HP 1100 series HPLC system (Agilent Technologies, Palo Alto, CA) was used. PC was separated on the diol column Nucleosil 100-OH (Macherey-Nagel, Düren, Germany; 250 mm × 3.0 mm id. × 5-µm particle size). A linear solvent gradient was patterned after the method used by Wang [24] with slight modification. Hexane/isopropanol (4:1, v/v) was used as mobile phase A, and mobile phase B consisted of isopropanol/water/formic acid/ammonia (89.3:10:0.2:0.5, v/v/v/v). Separation was obtained by using a gradient elution starting at 30% B and increasing to 60% over 22 min then maintained for another 2 min. After that, mobile phase B was increased to 80% over 11 min and maintained at that level for an additional 28 min. Finally, solvent B was decreased to 30% in the space of 2 min and the column was re-equilibrated for approximately 5 min prior to the next injection. The flow rate was 0.45 ml/min and the column temperature was 35°C. The injection volume was 5 µl.

The HPLC system was directly coupled to an electrospray ion-trap mass spectrometer or ESI-IT MS (Esquire HCT, Bruker Daltonics GmbH, Bremen, Germany) The ESI capillary potential was set at 4.5 kV. The dry nitrogen gas-flow rate was 8.0 L/min at 300°C. The MS data were collected in the negative ion mode. Mass range scan was set from 500–1,000 m/z at a rate of five spectra for each time point. All PC molecular species were well-detected as the formate adduct, [M + 45]⁻, as previously described [25]. After a correction for overlap of the C¹³ isotope peak, each postprandial PC molecular species was calculated and expressed as a percent distribution relative to its total concentration in the fasting state. The data acquisition and analysis were expedited with DataAnalysis^TM software version 3.2 (Bruker Daltonics GmbH, Bremen, Germany).

Values were expressed as the mean ± SEM. Normality was tested by using the Kolmogorov-smirnov test. Paired and two-sample t tests were used to compare parameters in the fasting state to the 4 h postprandial time point within groups and between the control and T2DM groups, respectively. A value of p<0.05 was considered statistically significant. The statistical analyses were performed using SPSS software version 11.5 (SPSS Inc, Chicago, IL).

No differences in baseline characteristics were observed in the T2DM group compared to the control group (Table 2). Plasma biochemical parameters in the fasting state and 4 h after the meal are summarised in Table 3. As expected, fasting glucose, HbA₁C, and HOMA-IR in T2DM subjects were significantly higher than those in control subjects; however, no differences in lipid parameters were observed between the two groups. Four hours after the meal, both groups showed a significant increase in triacylglycerol concentration compared to that in the fasting state. Postprandial glucose and insulin levels were also significantly greater among diabetic subjects than control subjects. Moreover, the change in insulin concentration relative to the fasting state was significantly higher in the diabetic group compared to that in the control.

The average second derivative spectra of erythrocytes used to identify the C-H stretching region in control and diabetic groups are shown in Fig. 1A and 1B, respectively. The C-H stretching region arising from PLs in erythrocytes was originated mainly from olefinic ν = (CH) (3020–3000 cm⁻¹), ν_as (CH₃) (2990–2945 cm⁻¹), ν_as (CH₂) (2945–2905 cm⁻¹), ν_s (CH₃) (2885–2860 cm¹), and ν_s (CH₂) (2860–2840 cm⁻¹). The integrated areas of observed bands in the control and T2DM groups obtained from the fasting state and 4 h after the meal are shown in Table 4. After the experimental period, all C-H stretching bands in the control group seemed to increase whereas all were found to have decreased or remained unchanged in the T2DM group when compared to the fasting state. In addition, the percent change in olefinic ν = (CH), ν_as (CH₃), ν_as (CH₂) and ν_s (CH₂) bands in the diabetic group were found to be significantly lower than those in the control group. Specifically, a marked difference was observed in the percent change in the integrated area of olefinic ν = (CH) relative to the fasting state compared to what was seen in the control group (−2.4 ± 5.3 vs 74.3 ± 9.0, p<0.001, respectively).

The typical profile of PC molecular species in erythrocytes obtained from LC-ESI-MS is shown in Fig. 2 Changes observed in the control and T2DM groups compared to the fasting state are presented in Fig. 3A and 3B, respectively. The predominant PC molecular species of erythrocytes from both groups were PC 34:2 and PC 34:1, which accounted for approximately 28% and 24% of the total PC molecular species, respectively. In the T2DM group, four PC molecular species were found to significantly change after the high-fat meal. However, the control group showed marked changes of eight major PC molecular species, particularly PC34:1 (16:0/18:1 as major species). As shown in Fig. 4, the total PC molecular species with SFA and MUFA after the meal period in the control group significantly increased when compared to the fasting state, in which there were no significant differences in the T2DM group.