This is an analysis of the INTERFAST-2 study, a single-center, randomized, controlled trial, investigating the effect of intermittent fasting in people with T2DM already injecting insulin (INTERFAST-2). This study was conducted at the University Hospital Graz, Austria, and approved by the ethics committee of the Medical University of Graz, Austria (EK 30–350 ex 17/18). This research adhered to the tenets of the Declaration of Helsinki, and GCP-ICH guidelines, and complied with the protocol and requirements of the relevant regulatory authorities. The study population consisted of individuals with T2DM, aged 18 to 75 years, with glycated hemoglobin A1c ≥ 7.0% (≥ 53 mmol/mol). The primary inclusion criteria were as follows: a total daily insulin dose of ≥ 0.3 units per kilogram of body weight and stable body weight over the previous three months (weight change < ± 3 kg). Participants had to be willing to comply with the study procedures, attend the study site, participate in the required protocols, and adhere to the fasting protocols.

Major exclusion criteria included active known malignancy within the past year (excluding prostate, gastrointestinal, and basal cell carcinoma intraepithelial neoplasia), pregnancy or intent to become pregnant, lactation, and any chronic disease that might interfere with the interpretation of study results. In addition, participants were excluded if they had started a new hormonal supplement or changed their hormonal contraceptive in the previous two months. Participants with type 1 diabetes mellitus or other forms of diabetes mellitus, acute or chronic inflammatory diseases, or who consumed more than 15 standard alcoholic drinks per week were also excluded. In addition, individuals who worked night shifts or used illicit substances were not included.

Four weeks before the start of the dietary intervention, participants were switched to the same basal insulin regimen. A trained physician made dose adjustments for participants in both groups during the intervention period. Further information can be found in the published study protocol [15].

A registered dietitian provided an educational intervention focused on individualized dietary strategies to promote optimal health outcomes. The session emphasized on creating a balanced and varied plate with a rainbow of vegetables. Participants were encouraged to reduce their added sugars and salt intake while adding more whole grains and legumes to their meals. All patients had the same number of interactions with the dietitian during both on-site and telephone visits. Adherence to the diet was continuously monitored voluntarily.

Blood samples were collected from the participants at the outset of the study, which occurred after the insulin switch phase but before the commencement of the intervention. The second blood draw was conducted after 12 weeks of intervention. Blood samples were drawn after a minimum of 8 h of overnight fasting.

ApoB-Depletion of serum

Apolipoprotein B (apoB) was depleted from serum using a polyethylene glycol (PEG) precipitation method. A 20% (w/v) stock solution of PEG (P1458, Sigma-Aldrich, Darmstadt, Germany) was prepared by dissolving it in 200 mmol/L glycine buffer. Forty microliters (µL) of this PEG solution were then added to 100 µL of serum. The mixture was gently mixed and incubated at room temperature for 20 min. Following incubation, the samples were centrifuged at 10,100 × g for 30 min at 4 °C. The resulting HDL-containing apoB-depleted serum was collected and stored at − 70 °C for further analysis.

Lecithin–cholesteryl acyltransferase (LCAT) activity

LCAT activity was assessed using a commercially available kit (MAK107, Merck, Darmstadt, Germany) following the manufacturer’s guidelines. The serum samples were incubated with the LCAT substrate for four hours at 37 °C. The fluorescent substrate emits at 470 nm, and upon LCAT-mediated hydrolysis, a monomer with fluorescence at 390 nm is released. LCAT activity was quantified by monitoring the change in the ratio of emission intensities at λ = 470 nm and λ = 390 nm over time.

Arylesterase activity of Paraoxonase

The Ca2+-dependent arylesterase activity of PON1 was determined using a photometric assay involving phenylacetate substrate, following a previously described protocol [16]. Briefly, 1.5 µL of 1:10 phosphate-buffered saline diluted apoB-depleted serum was added to a 200 µL buffer solution containing 100 mM Tris, 2 mM CaCl2 (pH 8.0), and 1 mM phenylacetate. The hydrolysis of phenylacetate was monitored at a wavelength of 270 nm. The enzymatic activity was determined using the Beer-Lambert law, with a molar extinction coefficient of 1,310 L mol-1 cm-1.

Cholesterol Ester Transfer Protein (CETP) activity

Serum CETP activity was determined using a commercially available kit (MAK106, Merck, Darmstadt, Germany) following the manufacturer’s instructions. The CETP Activity Assay Kit uses a proprietary substrate to measure CETP-mediated neutral lipid transfer. 3 µl of serum samples diluted 1:10 in phosphate-buffered saline are incubated with the donor and acceptor molecules at 37 °C for three hours. The reaction produces a fluorescent signal (λEx = 465 nm/λEm = 535 nm) proportional to CETP activity.

Serum levels of apolipoprotein M

Serum levels of apoM were quantified using a sandwich enzyme-linked immunosorbent assay method described in a prior study [17]. For apoM measurement, capture antibody (clone 1G9) (Abnova, Taipei City, Taiwan) detection antibody (clone EPR2904) (Abcam, Cambridge, UK), and HRP-conjugated anti-rabbit IgG antibody (cat. No. PO448) (DAKO, Glostrup, Denmark) were used. Briefly, a high-binding ELISA plate (Corning, Arizona, US) was coated with a capture antibody overnight and blocked with 2% bovine serum albumin. Serum samples (10 µl) were treated with 1,4-dithiothreitol (Sigma-Aldrich) and iodoacetamide (Sigma-Aldrich) to cleave disulfide bonds in apoM. The 1:50 diluted samples (in tris-buffered saline + 1% bovine serum albumin) were incubated overnight in the ELISA plate. After washing and the addition of detection and secondary antibodies, the absorbance of the colorimetric reaction was measured at 492 nm to determine the apoM concentration.

Cholesterol efflux capacity

The cholesterol efflux capacity of apoB-depleted serum was determined following established protocols [18, 19]. In brief, J774.2 cells (Sigma Aldrich, Darmstadt, Germany) were labeled with 0.5 µCi/mL radiolabeled [3 H]-cholesterol (Hartmann Analytic, Braunschweig, Germany) in DMEM media containing 2% FBS, 1% penicillin/streptomycin, and 8(4-chlorophenylthio)-cyclic adenosine monophosphate (0.3 mM) (Sigma-Aldrich, Darmstadt, Germany) overnight. After two washes, the cells were equilibrated for 2 h in serum-free DMEM supplemented with 2% bovine serum albumin from Sigma-Aldrich (Darmstadt, Germany). The cells were then rinsed and incubated with 2.8% apoB-depleted serum samples for 3 h. Cholesterol efflux capacity was expressed as the ratio of radioactivity in the media to the total radioactivity in the media and lysed cells.

NMR analysis

HDL subclasses and composition were assessed using a Bruker 600 MHz Avance Neo NMR spectrometer (Bruker, Rheinstetten, Germany) according to the Bruker IVDr Lipoprotein Subclass Analysis Protocol. Lipoprotein quantification was performed by analyzing the data using the Bruker IVDr Lipoprotein Subclass Analysis (B.I.LISATM) method as described previously [20]. The Bruker IVDr Lipoprotein Subclass Analysis identifies four HDL subclasses, labeled HDL-1 through HDL-4, based on increasing density and decreasing size. The defined density ranges for these subclasses are HDL-1 (1.063 to 1.100 kg/L), HDL-2 (1.100 to 1.112 kg/L), HDL-3 (1.112 to 1.125 kg/L), and HDL-4 (1.125 to 1.210 kg/L). For simplicity, these subclasses are designated as L-HDL (HDL-1), M-HDL (HDL-2), S-HDL (HDL-3), and XS-HDL (HDL-4).

Statistical analysis

All statistical analyses were performed with SPSS (version 29.0.0.0) (SPSS, Inc., Chicago, IL, USA) and GraphPad Prism 8.0. A p-value of < 0.05 was used to determine statistical significance. Participant characteristics are reported as means ± standard deviation, median with interquartile range or counts, and proportions. Statistical differences between the groups were calculated using paired t-test or Wilcoxon signed rank test, depending on the normality of the data. Fisher’s exact test was used to identify differences in comorbidities and used medication. The Spearman correlation coefficient was used to assess correlations between HDL functions and clinical parameters. The results are presented as a scatterplot. Differences before and after intervention within each group were calculated using the paired t-test or Wilcoxon test, depending on the normality of the data.