Radiochemistry

[11C]FMZ was synthesized according to clinical routine at Uppsala University Hospital, as described previously [36]. Different batches of [11C]FMZ were delivered with molar activity at end of synthesis of approximately 300 MBq/nmol for PET imaging studies and approximately 150 MBq/nmol for in vitro experiments.

Animal preparation and handling

For the PET/CT imaging study, one male pig (weight 24.5 kg) (Yorkshire x Swedish Landrace x Hampshire) was used. The pig was raised at a farm and transported to the PET facility at the day of experiment. Upon arrival it was sedated and prepared for the PET/CT experiment as previously described [37]. Briefly, the pig was sedated by intra-auricular infusion of combined ketamine and fentanyl (free from anesthetic benzodiazepine to prevent any interference with [11C]FMZ and blocking compound). The pig was connected to central venous catheter for the infusion of [11C]FMZ, cold flumazenil and contrast compound. All animal procedures were performed in agreement with the Uppsala University guidelines for animal research (UFV 2007/724) and national legislations. The studies were planned and performed in according to the 3Rs principle and to the ARRIVE guidelines for animal research and were approved by the animal Research Ethical Committee of the Uppsala Region (Ethical approval #5.8.18–15,648/2019).

In vivo biodistribution study with [11C]FMZ in pig model

The imaging procedures were performed using a Discovery MI PET/CT scanner (GE Healthcare), with a 25 cm field of view. A CT acquisition was first performed without contrast as previously described (Cheung et al., 2021) before a bolus intravenous injection of approximately 12 MBq/kg of [11C]FMZ dissolved in 25 mL of PBS (pH 7.4). A dynamic PET imaging of 60 min was subsequently acquired using the following parameters: 30 frames of 12 × 10, 6 × 30, 5 × 120, 5 × 300 and 2 × 600 s, VPFX-S, 3 i/16 s, 256 × 256 × 89 pixels, 3 mm post filter, 500 mm diameter zoom. A whole-body 30 min static PET/CT acquisition was performed after the abdominal scan (CT: 100 kV, 80–400 mA, noise index 10, rotation 0.5′, full spiral, slice thickness 3.75 mm, pitch 0.98:1, recon diameter 50 mm; PET: VPFX-S, 3 i/16 s, 256 × 256 pixels, 3 mm post filter, 500 mm diameter zoom). Three hours after the first injection of [11C]FMZ, cold flumazenil was intravenously infused at a dose corresponding to 0.1 mg/kg of pig body weight. Prior to starting a second PET/CT scan following the procedure previously described, a second dose of approximately 12 MBq/Kg of [11C]FMZ was intravenously injected. At the end of the dynamic PET scan over pancreas, a contrast enhanced CT acquisition was performed following the procedure described previously (Cheung et al., 2021). The PET/CT imaging data for both the baseline scan and with the addition of non-radioactive FMZ were analyzed using the PBAS tool from the PMOD software 4.3 (PMOD Technologies). Manual segmentation of the pancreas and spleen were performed on the co-registered abdominal CT, whereas the brain was delineated on the whole-body CT. The entire organs (pancreas, spleen, brain) were delineated on CT and transferred to the PET images, but care was taken to avoid spill-in of PET signal from surrounding tissues (mainly kidney in this case). In cases where spill-in was suspected, the segmentations were decreased in size to omit confounding signal from other tissues. The same areas of the tissues were segmented on both the baseline and blocking PET/CT scans.

The tissue uptake in kBq/cc was converted to Standardized Uptake Values (SUV) by dividing with the amount of administered radioactivity (in kBq) per body weight (g). SUV thus has the unit g/mL, but is often considered unitless by approximating the density of tissue to 1 g/mL. At the end of the imaging studies the animal was euthanized following intravenous injection of KCl. Organs of interest (pancreas and spleen) were removed, fixed in PFA 4% and embedded in paraffin for further immunostaining experiments.

Human islets and pancreas tissue

Isolated human islets and fractions of exocrine tissue were obtained from the Nordic Network for Clinical Islet Transplantation laboratory in Uppsala, Sweden. Fresh frozen human tissue including spleen and pancreas were provided by Uppsala Biobank. All human sample experiments were approved by the Regional Ethics Board of Uppsala, Sweden (now the Swedish Ethical Review Authority). Characteristics of human tissue and cell fractions obtained from cadaveric donors are listed in Supplementary Table S1. Human islets and fractions of exocrine tissue were hand-picked and collected separately into fresh CMRL1066 medium (# 21530027; Gibco) supplemented with 10% fetal bovine serum (# ref; Gibco), 2 mM L-glutamine (# G7513; Sigma) 100 Units/ml Penicillin, 100 µg/ml Streptomycin (# 1074440 Roche Diagnostics, Mannheim, Germany).

Mouse islets and pancreas tissue

Mice were sacrificed by cervical dislocation. Pancreata were perfused with 1 mg/ml collagenase P (# 11213865001; Sigma Aldrich) extracted and digested in Hank’s balance salt solution (HBSS) (Sigma Aldrich) buffered with 25 mM HEPES (# 15630-056; Gibco) (pH 7.4), supplemented with 0,25% (w/v) bovine serum albumin (BSA) (# A7030; Sigma Aldrich). Islets and exocrine pancreas were isolated and hand-picked under a stereo microscope in cold-HBSS buffered with HEPES (pH 7.4) supplemented with 0.5% BSA (w/v). Mouse pancreatic fractions were then collected separately into fresh RPMI 1640 (#R-0883; Sigma Aldrich) supplemented with 10% fetal bovine serum (# F7524; Sigma Aldrich), 2 mM L-glutamine (# G7513; Sigma Aldrich), 100 Units/ml Penicillin, 100 µg/ml Streptomycin (# 1074440 Roche Diagnostics, Mannheim, Germany).

Human and mouse cell fraction isolation for in vitro homogenates binding assay

Prior to in vitro experimentation, isolated human islets were kept in the same medium to recover overnight in an atmosphere of 95% air and 5% CO2 at 37 °C. The exocrine tissue was processed directly after isolation to prevent any enzymatic degradation of the tissue. Human and mouse endocrine and exocrine pancreatic cell fractions and mouse tissue (cortex and spleen) were freshly collected into ice- cold 0,32 M sucrose + BSA (1 mg per 100 ml 0.32 M sucrose) (# A7030; Sigma Aldrich) and homogenized by hand using a Dounce glass homogenizer (# P7734-1EA; Sigma Aldrich). Sample protein concentration was quantified by Bio-Rad Protein Assay (# 500–0006; Bio-Rad Laboratories) using BSA as a standard. Homogenates from human and mouse samples were snap frozen until used for in vitro homogenate binding assay.

In vitro homogenate binding assay

100 µL of homogenized tissue (endocrine and exocrine pancreatic cell fractions from human and mouse as well as mouse brain cortex and spleen) were incubated for 30 min at RT with 2.5 MBq/mL [11C]FMZ (50 nM) in a final suspension with PBS (pH 7,4) for a total volume of 200µL in a polystyrene round bottom tube. At the end of the incubation time, the samples were filtered via a Brandel harvester for Liquid Scintillation Counter using cold PBS through a Whatman GF/C filter of 1.2 µm particle retention (# 1822021; Cytivia). Filter patches containing cells were dried for 30 min. Radioactivity was measured in a well counter (Uppsala Imanet AB, Uppsala, Sweden) and corrected for radioactive decay. Sample measurements were performed either in duplicate or in triplicate and normalized according to protein concentration.

In vitro autoradiography

Fresh frozen cell fractions (isolated islets, isolated exocrine pancreas) and tissue (pancreas, spleen and cortex) from mouse, pig or human non-diabetic and T2D donors, were embedded into O.C.T (Q Path mounting media, VWR) and sectioned on cryotome (Cryostat NX70, ThermoFisher) in 10 µm or 20 µm thickness. Slices were mounted on SuperFrost Plus glass slides (ThermoFisher) and kept at -80 °C prior to the experiment. Duplicates of tissue and cell fraction sections were immersed in 150 mL of 50 mM cold PBS (Ph7.4) for 10 min. The cell and tissue slides were then incubated for 30 min in cold PBS (pH 7.4) containing low (2,5 nM) and high (5 nM) concentrations of [11C]FMZ corresponding to radioligand activities of 50 MBq and 90 MBq respectively. Three washing steps of 2 min in cold PBS and 1 min in distilled water were performed before drying the cell and tissue slides at 37 °C for 10 min and exposing them for 60 min on a BAS-IP storage phosphor screen (Fujifilm). Reference samples of known radioactivity were included simultaneously with the sections. The resulting digital image readout was obtained using an Amersham Typhoon storage phosphor imager (GE healthcare). Aliquots of each cold PBS (pH 7.4) bath containing low or high concentrations of [11C]FMZ were taken in triplicate to measure [11C]FMZ activity remaining at the end of the experiment and used for data analyses.

Immunohistochemistry

Sections of human, pig, mouse fresh frozen cell fractions (isolated islets and exocrine pancreas) and tissue (pancreas, spleen and cortex) used for Autoradiography assay (with high dose of [11C]FMZ) were taken for a hematoxylin and eosin staining. The cell and tissue sections stored in − 20 °C after ARG assay were placed at room temperature for 15 min, incubated after rehydration in hematoxylin for 5 min, immersed in running water for 10 min, rinsed in distilled water for 10 s, dipped in eosin for 45 s and briefly rinsed in distilled water before dehydration and mounting.

Freshly isolated islets from cadaveric donors (Suppl. Table S1) were harvested, washed in PBS 0.5% BSA, fixed in 4% paraformaldehyde for 1 h at 4 °C, cryoprotected in 15% sucrose solution in PBS for 1 h, then in 30% sucrose solution in PBS overnight. Human islets were then embedded in OCT on dry ice and sectioned in 8 µm. Pig and mouse organs (pancreas or cortex) were harvested, fixed with 10% formalin overnight, washed in water for Three hours, transferred into 70% ethanol solution for 48 h and embedded in paraffin before sectioning in 8 µm. Paraffin-embedded human, pig and mouse tissue sections or OCT-mounted human or mice isolated islets were incubated 30 min or 1 h respectively with blocking buffer (10% fetal bovine serum diluted in Tris–HCl pH7.4, 0.15 M NaCl, 0.1%Triton-X100 [TBST]). For mouse monoclonal primary antibodies, sections were additionally blocked with mouse in mouse reagent (60 min, diluted in PBS) (Vector labs #MKB-2202). Cell fraction or tissue sections were incubated with primary antibodies (Suppl. Table S2) overnight at + 4 °C followed by washes and fluorochrome labelled secondary antibodies for 1 h at room temperature on paraffin-embedded sections or overnight at + 4 °C on OCT-embedded-sections (Suppl. Table S2).

Image analysis

PET imaging data were analyzed with manual segmentation of the brain, pancreas and spleen on sequential transaxial projections using PBAS modeling tool (PMOD technologies LLC, Zurich, Switzerland). Data analyses for PET / CT imaging and quantifications were performed as previously described (Cheung et al., 2021). For Hematoxylin & Eosin staining performed on human, mouse and pig tissue following the autoradiography assay, images were acquired by scanning the complete slide to obtain a total overview of the tissue at × 4 magnification, using a digital pathology slide scanner (PathScan Enabler IV; Meyer Instrument). For higher magnification, H&E staining of the biological tissue was viewed with a Leica microscope LMD6000 (Leica Microsystems) and image acquisition was performed using LMD software.

Statistical analyses

Unless otherwise indicated, statistical analyses were performed and graphics were produced using GraphPad Prism version 9 software (GraphPad Software, Boston, MA, USA). Non-parametric analysis was used to calculate statistical significance, which was defined as p-values < 0.05. All values are presented as mean ± SEM.