Insulin storage and glucose homeostasis in mice null for the granule zinc transporter ZnT8 and studies of the type 2 diabetes-associated variants

Abstract

Objective: Zinc ions are essential for the formation of hexameric insulin and hormone crystallization. A nonsynonymous single nucleotide polymorphism rs13266634 in the SLC30A8 gene, encoding the secretory granule zinc transporter ZnT8, is associated with type 2 diabetes. We describe the effects of deleting the ZnT8 gene in mice and explore the action of the at-risk allele.

Research design and methods: Slc30a8 null mice were generated and backcrossed at least twice onto a C57BL/6J background. Glucose and insulin tolerance were measured by intraperitoneal injection or euglycemic clamp, respectively. Insulin secretion, electrophysiology, imaging, and the generation of adenoviruses encoding the low- (W325) or elevated- (R325) risk ZnT8 alleles were undertaken using standard protocols.

Results: ZnT8(-/-) mice displayed age-, sex-, and diet-dependent abnormalities in glucose tolerance, insulin secretion, and body weight. Islets isolated from null mice had reduced granule zinc content and showed age-dependent changes in granule morphology, with markedly fewer dense cores but more rod-like crystals. Glucose-stimulated insulin secretion, granule fusion, and insulin crystal dissolution, assessed by total internal reflection fluorescence microscopy, were unchanged or enhanced in ZnT8(-/-) islets. Insulin processing was normal. Molecular modeling revealed that residue-325 was located at the interface between ZnT8 monomers. Correspondingly, the R325 variant displayed lower apparent Zn(2+) transport activity than W325 ZnT8 by fluorescence-based assay.

Conclusions: ZnT8 is required for normal insulin crystallization and insulin release in vivo but not, remarkably, in vitro. Defects in the former processes in carriers of the R allele may increase type 2 diabetes risks.

FIG. 1.

Expression of ZnT8 in human pancreatic slices (A) and dissociated islet cells (B) and in purified wild-type mouse β- and α-cells (C). A: Pancreatic slices stained for ZnT8; scale bar, 50 μm. Cells in which clear colocalization to non–β-cells was apparent are highlighted with arrows. Essentially identical data were obtained with isolated human islets (not shown). B: Human islets were isolated (26) and dissociated with trypsin to allow the staining of single cells; scale bar, 5 μm. C and D: Mouse pancreatic α- and β-cells collected by flow cytometry from transgenic mice expressing the variant yellow fluorescent protein Venus under the control of the preproglucagon promoter (see supplementary Methods). Three separate preparations of α- and β-cells were analyzed by either microarray analysis (C) or qPCR (D). D: Expression is presented relative to that of β-actin measured in the same sample. Primer and probe sequences are available on request. *P < 0.05 β- versus α-cell normalized mRNA levels. NS, not significant. (A high-quality digital representation of this figure is available in the online issue.)

FIG. 2.

Generation and genotyping of ZnT8^−/− mice. A: See research design and methods for further details. B: Determination of genotype was performed on DNA tail biopsies by PCR amplification of genomic DNA using the following primers: L2: 5′-CTACTTCCATTTGTCACGTCCTGCACG-3′; F4: 5′-TGAAAACGGTGGGAAGCACTTGAGG-3′. The band migrating at ∼4,000 (4,295 bp) corresponds to the endogenous locus (in the +/− case); the band at ∼2,500 (2,593 bp) corresponds to the Cre-mediated excised locus. C: Western immunoblotting and immunocytochemical analysis of mice with the indicated phenotpyes. D: qPCR analysis of ZnT8 mRNA levels in islets from 8- to 12-week-old control and ZnT8^−/+ and ^−/− mice.

FIG. 3.

Pancreatic histology, zinc accumulation, and insulin processing. A: Consecutive 5-μm pancreatic slices from 12-week-old wild-type or ZnT8^−/− mice (London) stained by hematoxylin-eosin (Magnification ×2.5); islets are circled in red. B: Confocal images of pancreatic slices stained for insulin and glucagon (see research design and methods); scale bars, 50 μm. C: Intracellular zinc concentrations were estimated in isolated islets (Toronto) using FluoZin-3 (largely cytosolic; see Fig. 8B) or zinquin (granules; Fig. 8B). Fluorescence intensity per unit area was normalized by subtraction of the background average intensity in an area free of cells on the same coverslip, n = 4–6 islets per condition. D: Isolated islets (London) stained with 0.13 mmol/l dithizone for 5 min before imaging with a 5× objective on a Zeiss Axiovert 40 microscope; scale bar, 50 μm. E: Electron micrographs of isolated islets from ZnT8^−/− animals and wild-type littermate controls at the indicated ages. Toronto (a) London (b) colonies; scale bar, 1 μm. F: Proinsulin conversion to insulin in the β-cells of islets isolated from ZnT8^−/− or littermate control mice (12 weeks, London). (A high-quality digital representation of this figure is available in the online issue.)

FIG. 4.

Blood glucose homeostasis and insulin secretion in ZnT8^−/− mice (Toronto) in (A–D) 6- and (E and F) 12-week-old mice. A: After a 5-h fast, glucose (1.5 g/kg) tolerance was assessed in 6-week-old male littermates. ■, +/+ (n = 12); ○, +/− (n = 8); ▲, −/− (n = 8). **P < 0.05 for the difference between wild-type and ZnT8^−/− mice. B: Insulin levels were measured during IPGTT and the area under the curve (AUC) determined for each genotype. Secretion of insulin (C) and glucagon (D) from freshly isolated islets. Islets were from 3–12 separate mice per genotype. □, 0 mM glucose; ■, 20 mM glucose. *P < 0.05. E and F: Glucose tolerance (1.5 g/kg) assessed in 12-week-old mice as in A. NS, no significant difference in AUC for +/+ or +/− versus −/− mice. ■, +/+ (n = 6); ○, +/− (n = 3); ▲, −/− (n = 4). F: Glucose-stimulated insulin secretion assessed as in A(c). *P < 0.05, **P < 0.01, ***P < 0.001 for the indicated effects.

FIG. 5.

Electrophysiological changes and TIRF imaging of exocytosis in single β-cells. A: Membrane potential, (B) depolarization-induced capacitance changes, and (C) Ca²⁺ currents recorded in pancreatic islet cells: ZnT8^+/+ (▲, n = 17), ZnT8^+/− (○, n = 16), and ZnT8^−/− (■, n = 26) animals (London, 12 weeks). A: Mean values of the whole cell capacitance change versus time are given; data were reduced to show 200 points for clarity. The stimulation protocol is given schematically above the graph. D: TIRF images of β-cells before and after stimulation; scale bar, 5 μm. Release events are indicated by an arrow. Snapshots of a single release event. a–c: Mark the points seen on the graph in E Kinetics of NPY Venus release. Data are normalized to baseline to exclude expression artifacts. Data expressed as the average of 38 versus 43 events (6 vs. 12 separate cells for +/+ vs. −/−; three separate preparations per genotype). F: Proportion of full and incomplete release events.

FIG. 6.

Effect of R325W polymorphism rs13266634 on the predicted molecular structure of ZnT8 and on subcellular localization and stability. A: Western immunoblotting of ZnT8 in mouse islets and clonal MIN6 β-cells. Total lysates were separated by 10% SDS-PAGE and immunoblotted with anti-rat/mouse ZnT8 polyclonal antibodies (Mellitech, France) or anti–β-actin polyclonal antibody (43 kDa; Sigma). B: Dimer formation. Human ZnT8-V5 and ZnT8-EGFP were transfected into HeLa cells. Forty-eight hours posttransfection, cells were lysed and subjected to immunoprecipitation with antibodies against EGFP (upper panel) or V5 (lower panel) tags. After washing the beads, bound proteins were eluted and subjected to immunoblotting, probing with anti-V5 (upper; predicted molecular mass for ZnT8-V5, 46.4 kDa) or anti-EGFP (lower; predicted molecular mass for ZnT8-EGFP, 68.2 kDa) antibodies. “NS” denotes a nonspecific band, likely derived from IgG lost from the sepharose beads. C: Modeling of ZnT8 structure based on YiiP. The human transporter was modeled using the structure of the Escherichia coli transporter YiiP (PDB accession 2QFI) as a template. Left, upper right: Provide views of the W325 variant from the plane of the membrane and the cytoplasmic side of the membrane, respectively; lower right, view of R325 variant from cytoplasmic side of the membrane. The likely locations of zinc ions are shown as red spheres in left, and residues at position 325 are shown in space-filling representation in all panels. D: Western immunoblotting analysis of overexpressed ZnT8 isoforms in MIN6 cells. E: Subcellular distribution of overexpressed ZnT8 isoforms in INS-1(832/13) cells; scale bar, 5 μm. (A high-quality digital representation of this figure is available in the online issue.)

FIG. 7.

Effects of R325W polymorphism of ZnT8 on insulin secretion and processing and on glucose signaling. A: INS-1(832/13) cells were infected with adenovirus encoding either EGFP only or the indicated ZnT8 isoform. Forty-eight hours later, cells were incubated for 30 min at the indicated glucose concentrations (n = 6 separate cultures from two independent experiments). B: Membrane potential and K_ATP channel conductance in MIN6 cells overexpressing the indicated ZnT8 isoform (n = 6 in each case) in pIRES2. C: Typical traces describing changes in intracellular free Ca²⁺ ([Ca²⁺]_i) in individual MIN6 cells overexpressing the indicated ZnT8 isoform from plasmid pIRES2, or empty vector, and stimulated with the indicated concentrations of glucose or KCl (50 mmol/l). D: Combined data from responding cells in C. The graphs describe the increase in cytosolic [Ca²⁺] as AUC after stimulation with 30 mmol/l glucose (upper panel, CTRL n = 13, R325 and W325, n = 8 cells) or 50 mmol/l KCl (lower panel, CTRL n = 46, R325 n = 45, W325 n = 42). *P < 0.05, **P < 0.01 for the indicated effects.

FIG. 8.

Role of ZnT8 in cellular Zn²⁺ transport and effects of the R325W polymorphism. A: ZnT8 expression was reduced by RNAi as assessed by immunocytochemical analysis of cells fixed in 4% (v/v) paraformaldehyde and incubated with rabbit anti-mouse ZnT8 antibody (Mellitech, 1:3,000) and anti-rabbit–conjugated secondary antibody (1:500) plus Hoechst dye (5 μg/ml) before imaging on a Zeiss Axiovert inverted optics microscope. Zn²⁺ uptake into single MIN6 cells was assessed by monitoring changes in FluoZin-3 fluorescence (research design and methods). Under the conditions used in these experiments, the FluoZin-3 fluorescence increase gives an almost linear readout during the perifusion with ZnSO₄. The relative rates of Zn²⁺ accumulation were calculated from the slopes of the corresponding fluorescence changes. *P < 0.05 versus siCon, n = 5 independent experiments, 108–117 cells per condition. Calibrated resting free [Zn²⁺]_i was 600–700 pmol/l (NS, higher vs. lower risk). B: R325 or W325 ZnT8 were overexpressed in MIN6 cells, and transfected cells were identified by fluorescence of cotransfected mCherry-expressing vector before measurements of the initial rate of apparent Zn²⁺ uptake as in A. *P < 0.05 versus control or ZnT8-R, n = 5 independent experiments, 67–115 cells per condition. Zinc accumulation into granules was assessed using zinquin after incubation for 4 h in the presence of 10 μmol/l ZnSO₄. Insets show the localization in single cells of FluoZin-3 and zinquin, as monitored by confocal microscopy. Note the punctuate distribution of zinquin, consistent with accumulation of the dye into secretory granules; scale bar, 5 μm. (C) Localization of overexpressed ZnT8-mCherry chimaeras at the plasma membrane, identified using CD38-EGFP. Note the presence of plasma membrane–associated ZnT8-mCherry fluorescence (red), coincident with CD38-EGFP (green) in the line plots (vertical arrows). (A high-quality digital representation of this figure is available in the online issue.)