Pancreatic samples from seven multiorgan donors without diabetes, six with type 1 diabetes and seven with type 2 diabetes were studied (see Table 1 for details). Approval was obtained from the local ethics committee at the University of Pisa. The type of diabetes was identified based on medical records, anti-GAD autoantibody status and/or the morphological and ultrastructural appearance of pancreatic islets (see electronic supplementary material [ESM] Fig. 1). Pancreatic tissue samples were taken before the islet isolation procedure. Samples were prepared for electron and optical microscopy as previously described [17, 18]. The specimens acquired for electron microscopy were also used to obtain semithin sections (500 nm thickness) for light microscopy analysis.
To study mast cells in semithin sections, staining was performed with a 1:1 mixture of toluidine blue (1% in twice-distilled water) and methylene blue (1% in twice-distilled water) for 15 min at 37°C, with identification of mast cells by light microscopy based on metachromatic staining . For immunohistochemistry studies , specimens were fixed in 10% neutral-buffered formalin (Sigma-Aldrich, St Louis, MO, USA) and embedded in paraffin. Sections (2 μm) were then cut, deparaffinised and rehydrated through passages in xylene (Sigma-Aldrich) and graded ethanol (Sigma-Aldrich) solutions. Afterwards, samples were rinsed in PBS and endogenous peroxidases, and inhibited by treatment for 10 min with 0.5% hydrogen peroxide. The sections were subsequently rinsed in PBS and treated with normal goat serum blocking solution (Vector Laboratories, Burlingame, CA, USA) at 1:50 dilution for 30 min.
To identify mast cells, we used a monoclonal mouse antibody against tryptase (a mast cell marker ) (clone G3; Millipore, Merck, Vimodrone, Italy) at 1:5,000 dilution, with incubation performed overnight at 4°C. After rinsing with PBS, sections were incubated with the biotinylated secondary antibody (Cell Marque, Rocklin, CA, USA) for 15 min, followed by treatment with the horseradish peroxidase streptavidin label (Cell Marque) for an additional 15 min. Sections were then rinsed in PBS and the diaminobenzidine substrate (Vector Laboratories) was applied for 40 s. After rinsing in double-distilled water, sections were counterstained with Mayer’s haematoxylin solution (Sigma-Aldrich), rinsed under running tap water, and then placed in 2% ammonia solution. Sections were finally dehydrated in ethanol, cleared in xylene and covered with mounting media (Sigma-Aldrich). Images were acquired at ×40 magnification using a Leica DM5500 B microscope (Leica, Wetzlar, Germany).
For the electron microscopy studies (sections 60 nm thickness), mast cell identification was based on the particular ultrastructural appearance of these cells, including a monolobed nucleus, surface architecture composed of narrow, elongated folds, the presence of typical cytoplasmic granules and the absence of cytoplasmic glycogen aggregates [21, 22]. Lymphocytes and macrophages were identified by electron microscopy, as previously described [23, 24]. Immune system cell counts per mm2 were performed in semithin sections of pancreatic tissue. The same cells were examined by electron microscopy in consecutive ultrathin section to confirm the identification and assess ultrastructural characteristics (Fig. 1).
Counts of inflammatory cells infiltrating the pancreatic islets and of beta cells per islet (together with the calculation of beta cells with signs of apoptosis) were performed with 58 islets (20 from donors without diabetes, 18 from donors with type 1 diabetes and 20 from donors with type 2 diabetes) and on tissue sections where recognisable islets (independent of the presence of immune system cells) were identified. The number of pancreatic non-consecutive semithin sections (one every ten sections) examined to reach these numbers of islets was 145 in control cases (with 3 ± 1 blocks and 21 ± 8 sections per patient), 568 in samples from donors with type 1 diabetes (10 ± 3 blocks and 95 ± 26 sections per patient) and 240 in samples from donors with type 2 diabetes (5 ± 1 blocks and 34 ± 15 sections per patient). All counts were independently performed by two investigators unaware of sample identity (samples were identified only by randomly assigned numbers).
Quantification of beta cells (including those with apoptotic features) was performed as previously detailed [18, 25–27]. Beta cells were identified based on the presence of typical insulin granules, while apoptotic beta cells were identified based on the appearance of marked chromatin condensation and/or blebs (ESM Fig. 2), as previously reported by us and others [18, 25, 28]. Islets isolated from independent non-diabetic donors were also studied histologically after exposure to histamine (see below). In such cases, in addition to electron microscopy experiments, immunofluorescence analysis was performed to identify insulin-containing cells and TUNEL-positive cells, similar to previously published procedures [17, 29]. To do this, pancreas sections were deparaffinised, rehydrated through passages in xylene and graded ethanol solutions, and rinsed in PBS. Antigen retrieval was then performed using citrate buffer (pH 6) in a microwave at 350 W for 5 min. Insulin immunostaining was performed using a primary, polyclonal guinea pig anti-insulin antibody (Abcam, Cambridge, UK) at 1:100 dilution, applied for 1 h at room temperature, followed by treatment with the secondary antibody (Alexa Fluor 594-conjugated donkey anti-guinea pig IgG; Jackson ImmunoResearch, Pero, Italy) at 1:200 dilution for 1 h at room temperature. The presence of apoptotic cells was confirmed using the In Situ Cell Death Detection Kit (Roche, Mannheim, Germany). Sections were incubated with a reaction solution containing the terminal deoxynucleotidyl transferase enzyme and fluorescein-dUTP labelled nucleotides (TUNEL reaction mixture) (In Situ Cell Death Detection Kit; Roche) for 1 h at 37 C. Analysis was performed at ×400 magnification using a Leica DM5500B microscope. All counts were independently performed by two investigators unaware of sample identity (samples were identified by randomly assigned numbers).
Islet isolation and incubation
Isolated islets were prepared from the pancreases of 18 independent non-diabetic multiorgan donors (age 69 ± 15.5 years; ten male; BMI 26.2 ± 3.1 kg/m2; cause of death: 14 cardiovascular events and four traumas; duration of intensive care unit stay 3 ± 2 days; pancreas cold ischaemia time 18.3 ± 6.2 h) by collagenase digestion followed by density gradient purification, as previously reported [18, 26]. After isolation, the islets were maintained for 2–3 days in M199 medium, containing 5.5 mmol/l glucose, supplemented with 10% serum and antibiotics. Batches of approximately 1,000 islets were then cultured for 72 h in M199 medium supplemented with 10% (vol./vol.) serum and 100 U/ml penicillin, 100 μg/ml streptomycin, 50 μg/ml gentamicin and 750 ng/ml amphotericin B, and either with or without the addition of 100 μmol/l histamine dihydrochloride (all Sigma-Aldrich). In some experiments, the effect of caspase inhibition by 20 μmol/l Z-VAD-FMK (Promega, Madison, WI, USA) was also tested.
Western blot experiments
Isolated human islets were lysed in buffer containing 40 mmol/l Tris, 4% CHAPS, 7 mol/l urea, 2 mol/l thiourea and 1% DTT, and supplemented with a protease inhibitor cocktail (Roche). The lysates were centrifuged to remove cellular debris and the protein amount was evaluated using the Bradford method. A western blot assay of caspase 9 and 3 activation in human islets incubated for 24 h with histamine (see above) was performed as previously described . Briefly, equal amounts of total protein were heated at 100°C for 5 min, resolved by SDS-PAGE and electroblotted onto nitrocellulose membranes. Immunodetection was performed after overnight incubation with antibodies for cleaved caspase 9 and 3 (Cell Signaling, Danvers, MA, USA), and α-tubulin (Cell Signaling) was used as the loading control.
Gene expression studies
Gene expression was determined as previously described [18, 30, 31]. Briefly, total RNA was extracted using the PureLink RNA Mini Kit (Life Technologies, Carlsbad, CA, USA) and quantified by absorbance at A260/A280 nm (ratio >1.9) in a NanoDrop 2000C spectrophotometer (Euroclone Spa, Pero, Italy). RNA integrity was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Cernusco sul Naviglio, Italy). For quantitative PCR experiments, total RNA was reverse-transcripted from 1 μg using a SuperScript VILO cDNA Synthesis Kit (Life Technologies). The oligonucleotides of interest were obtained from assay-on-demand gene expression products (Life Technologies). mRNA levels were quantified and normalised for β-actin using a ViiA 7 analyser (Life Technologies).
The activity of mitochondrial complex I was assayed in control and histamine-treated (72 h) INS-1E cells (kindly provided by C. Wollheim; University of Geneva, Geneva, Switzerland) using a commercially available kit (Mitochondrial Complex I Activity Assay Kit; Millipore, Darmstadt, Germany) and following the manufacturer’s instructions. INS-1E cells were cultured in a humidified atmosphere containing 5% CO2 in complete medium composed of RPMI 1640 supplemented with 10% heat-inactivated FCS, 1 mmol/l sodium pyruvate, 50 μmol/l 2-mercaptoethanol, 2 mmol/l glutamine, 10 mmol/l HEPES, 100 U/ml penicillin and 100 μg/ml streptomycin . The mitochondrial membrane potential in control and histamine-treated INS-1 cells was measured cytofluorimetrically, as previously described , with a FACScan equipped with CellQuest software (BD Biosciences, Franklin Lakes, NJ, USA).
Results are given as means ± SD, and differences between groups were assessed using the two-tailed Student’s t test or ANOVA with Bonferroni correction, as appropriate. Unilinear regression analysis was also performed to evaluate correlations between selected variables. A p value of < 0.05 was considered statistically significant.