RESEARCH DESIGN AND METHODS

Patients
The study protocol has been described in detail previously (11). Twenty-seven obese patients with insulin-dependent type 2 diabetes mellitus (14 men and 13 women) were included (mean ± SEM: age: 55 ± 2 years, BMI: 37.2 ± 0.9 kg/m2, HbA1c: 7.8 ± 0.2%). At baseline patients used 82 ± 11 units of insulin per day with or without concomitant use of metformin and/or sulfonyl ureum derivates. Exclusion criteria were as follows: smoking, unstable weight during 3 months before inclusion, or any other chronic disease. In an article published previously (11), we described only 12 out of the 27 patients from whom proton magnetic resonance spectroscopy (1H-MRS) scans could be obtained. The local ethics committee approved this protocol. All patients gave written informed consent, and the study was performed in accordance with the Declaration of Helsinki.

Study design
Patients were studied before the start and after completion of the 16-week VLCD. Three weeks before start of the VLCD all oral blood glucose–lowering medication was stopped and insulin therapy was intensified. The day before the start of the VLCD intervention, only short-acting insulin was prescribed. Patients did not use any blood glucose–lowering medication, including insulin, during the 16-week VLCD. The VLCD consisted of three liquid food shakes (Modifast Intensive; provided by Nutrition & Santé, Antwerp, Belgium) containing a total of 450 kcal/day and all essential micro- and macronutrients. Thirteen of the twenty-seven subjects simultaneously followed an exercise program in addition to the VLCD. Because exercise had no effect on outcome parameters (Supplementary Table 1), data of all subjects were pooled for the present analyses.

Hepatic TG content
Hepatic TG content was measured in supine position using 1H-MRS on a 1.5 Tesla whole-body MR scanner (Gyroscan ACS/NT15; Philips, Best, the Netherlands), exactly as described previously (11).

Plasma (apo)lipoprotein and CETP analyses
All plasma samples were obtained after an overnight fast before the start (i.e., after stopping all glucose-lowering medication including insulin) and after the 16-week VLCD protocol, stored in aliquots at −80°C, and analyzed after thawing once in a single laboratory (Leiden, the Netherlands). To ascertain that we could make adequate correlations, all analyses were performed within the same blood samples in the same assay runs. Plasma cholesterol and TG concentrations were determined using enzymatic kits (no. 236691 and 11488872, respectively; Roche Molecular Biochemicals, Indianapolis, IN). Plasma phospholipids were determined using the phospholipids B kit (Wako Chemicals, Neuss, Germany). Plasma CETP concentration was quantified using kit CETP ELISA Daiichi (Daiichi Pure Chemicals, Tokyo, Japan). HDL fractions were obtained after precipitation of apolipoprotein (apo)B-lipoproteins from 50 μL plasma by adding 25 μL 36% polyethylene glycol 6000 (PEG6000, no.81260; Sigma Aldrich) The HDL-cholesterol and phospholipids were determined as described above. Plasma apoAI and apoB100 levels were determined with the Human ApoAI ELISA kit (no. 3710–1H; Mabtech AB, Nacka, Sweden) and Human ApoB ELISA kit (no. 3715–1H; Mabtech AB), respectively.

Cholesterol efflux study
Cholesterol efflux to total plasma and apoB-depleted human plasma were determined using the human monocyte cell line THP-1 as cholesterol donor. THP-1 cells were obtained from European Collection of Cell Cultures and maintained in medium A (RPMI 1640 with 25 mmol/L HEPES buffer, supplemented with 10% fetal bovine serum, 1% l-glutamine, 100 units/mL penicillin, and 100 μg/mL streptomycin) at 37°C in 5% CO2. Before the experiment, THP-1 cells were seeded into 24-well plates at density of 5 × 105 cells per well and differentiated into macrophages with 0.1 μmol/L phorbol 12-myristate-13-acetate (no. P1585; Sigma Aldrich) within 3 days. Macrophages were washed three times with PBS and incubated in medium B (RPMI 1640 with 25 mmol/L HEPES buffer, supplemented with 2% fetal bovine serum, 1% l-glutamine, 100 units/mL penicillin, 100 μg/mL streptomycin, 50 μg/mL acetyl-LDL, and 10 μCi/mL [1α,2α(n)-3H]-cholesterol (no. NET139001MC; Perkin Elmer, the Netherlands) for 1 day at 37°C in 5% CO2. After incubation, cells were washed three times with PBS and the efflux assay was started by adding total human plasma or apoB-depleted human plasma diluted to 1% in medium C (RPMI 1640 with 25 mmol/L HEPES buffer, supplemented with 1% l-glutamine, 100 units/mL penicillin, 100 μg/mL streptomycin, and 0.5 mg/mL BSA). The whole assay was carried out in triplicate. To be able to normalize results between series of experiments and to correct for plate-to-plate variation, efflux to a standard preparation of HDL (50 μg protein/mL) was determined in triplicate. After 4-h incubation, medium was collected and centrifuged. Subsequently, [3H]cholesterol was quantified by liquid scintillation counting. Total cellular 3H-cholesterol was determined after extraction of the cells with 0.1 mol/L NaOH. Cholesterol efflux rate was calculated by dividing the 3H activity in the medium by the sum of the 3H activity in the medium and the cell extract. Background values (the efflux in the absence of plasma) were subtracted.

Statistics
Data are expressed as means ± SEM. Paired t tests were used for the statistical comparisons between measurements at baseline and after 16 weeks of caloric restriction. For correlation analysis, Pearson correlation analysis was used. A P value < 0.05 was considered statistically significant.