Insulin promoter DNA methylation correlates negatively with insulin gene expression and positively with HbA_1c levels in human pancreatic islets

Pancreatic islets from 48 non-diabetic and nine type 2 diabetic deceased donors were obtained from the Human Tissue Laboratory at Lund University Diabetes Centre (Table 1). Islets were prepared by collagenase digestion and density gradient purification. After isolation, islets were cultured free floating in CMRL 1066 culture medium (ICN Biomedicals, Costa Mesa, CA, USA) supplemented with 10 mmol/l HEPES, 2 mmol/l l-glutamine, 50 μg/ml gentamicin, 0.25 μg/ml Fungizone (GIBCO BRL, Gaithersburg, MD, USA), 20 μg/ml ciprofloxacin (Bayer Healthcare, Leverkusen, Germany), and 10 mmol/l nicotinamide at 37°C (5% CO₂) prior to RNA and DNA preparation. The islet purity was similar for type 2 diabetic (68 ± 19%) and non-diabetic (58.7 ± 19%, p = 0.16) donors. Insulin content in homogenised human islets was assessed by ELISA (Mercodia, Uppsala, Sweden) and values were normalised to the total DNA in each sample as determined by a fluorometric assay (Quant-iT Picogreen, Invitrogen Molecular Probes, Stockholm, Sweden). Glucose-stimulated insulin secretion from the human islets was measured in vitro in static incubations as previously described by Rosengren et al. [10]. The donor before death or her/his relatives had, on admission to the intensive care unit, given their consent to donate organs. The local ethics committees approved the protocols.

Beta and alpha cells were purified from pancreatic islets of three human donors (aged 54, 55 and 74 years, with BMI 21.5–23.1 kg/m²), different from the donors described in Table 1, using a method previously described by Parnaud et al. [11]. Dissociation of islet cells was achieved by incubation with constant agitation for 3 min at 37°C in 0.05% (wt/vol.) trypsin-EDTA (Invitrogen) supplemented with 3 mg/ml DNAse I (Roche, Basel, Switzerland) followed by vigorous pipetting. Labelling and FACS sorting of the beta and alpha cell fractions was performed as previously described by Parnaud et al. [11]. Sorted alpha and beta cells were applied to microscope slides and co-immunostained for insulin and glucagon in order to detect the amount of alpha cells in the beta cell fraction, and vice versa. Using this method, a beta cell purity of 89 ± 9% (mean ± SD) was achieved [12].

Clonal rat insulinoma-derived INS 832/13 beta cells were cultured in RPMI medium with 11.1 mmol/l glucose, which is the basal glucose concentration for these cells, supplemented with 10% fetal calf serum, 1 mmol/l sodium pyruvate, 10 mmol/l HEPES, 50 μmol/l β-mercaptoethanol, 100 U/ml penicillin and 100 μg/ml streptomycin at 37°C in a humidified atmosphere containing 95% air and 5% CO₂. The cells were plated out in six well plates and allowed to attach overnight. The medium was then replaced with fresh medium containing either 11.1 or 16.7 mmol/l glucose. The cells were further cultured for 72 h, with the medium exchanged after 48 h. DNA was isolated with the All Prep kit (Qiagen, Hilden, Germany) and kept at −20°C until analysis.

Total RNA was extracted from human islets using All Prep DNA/RNA kit and cDNA was synthesised using QuantiTect Reverse Transcription kit (Qiagen, Hilding, Germany). Insulin mRNA levels were analysed using TaqMan Real-Time PCR with an ABI Prism 7900 HT system and gene-specific probes and primer pairs (Assays-on-demand, Hs02741908_m1; Applied Biosystems, Foster City, CA, USA). The insulin transcript level was normalised to the mRNA level of cyclophilin A (4326316E, Applied Biosystems) and quantified using the ∆∆Ct method.

Sequenom’s MassARRAY EpiTYPER protocol was applied to measure DNA methylation (Sequenom, San Diego, CA, USA). Five different EpiTYPER assays were designed (EpiDesigner), covering 26 CpG sites of the human insulin promoter (successful data were generated for 25 CpG sites, see below). Also, two assays covering ten CpG sites of the rat insulin promoter were designed (successful data were generated for five CpG sites, see below). The primer sequence and the location of the five human assays and the two rat assays are given in Electronic supplementary material (ESM) Table 1. Genomic DNA was first extracted from the human islets and the clonal rat beta cells using the All Prep DNA/RNA kit (Qiagen) as well as from whole blood cells using the Gentra Puregene Blood Kit (Qiagen). Genomic DNA, 500 ng, was bisulfite treated with the EZ DNA Methylation kit (Zymo Research, Orange, CA, USA). Duplicate PCR reactions were performed with bisulfite-specific PCR primers containing a T7-promotor tag and a 10mer tag on the reverse and forward primer, respectively. In vitro transcription and RNase cleavage reaction were conducted using the MassCleave (hMC) kit according to the manufacturer’s recommendations. The cleavage reaction was dispensed onto a 384 element SpectroCHIP bioarray and mass spectra were acquired using a MassARRAY mass spectrometer. The spectra were analysed and the methylation ratios were obtained by the EpiTYPER software v.1.0.1. As a result of the cleavage pattern, four CpG sites were analysed as two separate CpG units (CpG −1,478 together with CpG −1,482 and CpG −1,893 together with CpG −1,897). The results from these CpG units were excluded from further statistical analysis. CpG site −1,565 was excluded from this study because of variation between duplicate samples. CpG site −180 did not generate any methylation data using the EpiTYPER because of low mass of the cleavage product. However, as CpG −180 has previously been suggested to be important for regulation of insulin gene expression via activating transcription factor 2 (ATF2, also known as cAMP responsive element binding protein 2 [CREB2]) binding [9], we used Pyrosequencing (Qiagen) according to manufacturer’s recommendations to analyse DNA methylation at CpG −180 (ESM Table 1). Five CpG sites in the rat insulin promoter assay did not generate any methylation data due to either low or high mass of the cleavage product.

Differences in DNA methylation of the insulin promoter between type 2 diabetes patients and non-diabetic individuals were analysed using multivariate regression analysis including disease status, age, sex and BMI as covariates. Correlations were calculated using the Spearman correlation coefficient. Paired tests were used to analyse differences between beta and alpha cells as well as between clonal rat beta cells cultured in 11.1 or 16.7 mmol/l glucose. Non-parametric two-sample tests were used to analyse differences in insulin mRNA and content as well as glucose-stimulated insulin secretion between individuals with type 2 diabetes and non-diabetic donors. All p values were two-tailed and p values less than 0.05 were considered significant. The presented p values have not been corrected for multiple testing. Statistical calculations were performed using NCSS software (NCSS Statistical Software, Kaysville, UT, USA).