Glucokinase activity controls peripherally located subpopulations of β-cells that lead islet Ca2+ oscillations

Abstract

Oscillations in insulin secretion, driven by islet Ca²⁺ waves, are crucial for glycemic control. Prior studies, performed with single-plane imaging, suggest that subpopulations of electrically coupled β-cells have privileged roles in leading and coordinating the propagation of Ca²⁺ waves. Here, we used three-dimensional (3D) light-sheet imaging to analyze the location and Ca²⁺ activity of single β-cells within the entire islet at >2 Hz. In contrast with single-plane studies, 3D network analysis indicates that the most highly synchronized β-cells are located at the islet center, and remain regionally but not cellularly stable between oscillations. This subpopulation, which includes 'hub cells', is insensitive to changes in fuel metabolism induced by glucokinase and pyruvate kinase activation. β-Cells that initiate the Ca²⁺ wave (leaders) are located at the islet periphery, and strikingly, change their identity over time via rotations in the wave axis. Glucokinase activation, which increased oscillation period, reinforced leader cells and stabilized the wave axis. Pyruvate kinase activation, despite increasing oscillation frequency, had no effect on leader cells, indicating the wave origin is patterned by fuel input. These findings emphasize the stochastic nature of the β-cell subpopulations that control Ca²⁺ oscillations and identify a role for glucokinase in spatially patterning 'leader' β-cells.

Keywords: calcium oscillation; cell biology; glucokinase; light-sheet microscope; mouse; pancreatic islets; β-cell.

Figure 1.. Engineering of a light-sheet microscope to image intact islets in 3D.

(A) Schematic of the light-sheet microscope showing the optical configuration. (B) Representative light-sheet (upper panel) and spinning disk confocal images (lower panel) of a mouse pancreatic islet expressing β-cell-specific H2B-mCherry fluorophore at different two-dimensional (2D) focal planes, emphasizing the superior depth penetration of the light-sheet microscope. (C) 3D imaging of β-cells expressing GCaMP6s Ca²⁺ biosensors and nuclei mCherry biosensors. (D) Using Ins1-Cre:ROSA26^{GCaMP6s/H2B-mCherry} islets, the software-identified center of β-cell nuclei (yellow dots) was used to generate GCaMP6s regions of interest (gray spheres). A representative Ca²⁺ time course is displayed in the right panel for an islet stimulated with glucose and amino acids.

Figure 1—figure supplement 1.. Hardware wiring diagram of the light-sheet microscope.

Hardware integration (top panel) for camera-triggered activation of the excitation lasers and piezo z-stage that limits communication to a single instruction from the computer every 3 min. Wiring diagram (bottom panel): two Nikon ‘standard cables’ connect to the NiDAQ card installed in the computer. These two cables link to the laser control box, stage controller, and camera. The images captured are received by the computer through a camera link PCIe card.

Figure 1—figure supplement 2.. NIS-Elements software configuration.

(A) Schematic for comparing continuous acquisition and looped acquisition. The red box indicates the 3-min window where storage speed is higher than imaging speed. (B) Devices linked to NIS-Elements. (C) NIS-Elements JOBS module configured for looped acquisition. (D) Optical configuration for simultaneous GCaMP6s/H2B-mCherry excitation.

Figure 2.. Characterization of single β-cells using three-dimensional (3D) network and phase analysis.

(A) Flow diagram illustrating the calculation of cell degree from pairwise comparisons between single β-cells. (B) An example 3D network for a single β-cell within a representative islet is shown with synchronized cell pairs in blue, cells that have other synchronized pairs in black, and cells that are asynchronous in red. This analysis is repeated for all cells in the islet. (C) Frequency distribution of cell degree for all β-cells analyzed. Top 10% (blue box) and bottom 10% (red box) are high and low degree cells. (D) Representative 3D illustration and Ca²⁺ traces showing the location of high degree cells (blue) and low degree cells (red). (E) Quantification of the normalized distance from the islet center for average degree cells (gray), high degree cells (blue), and low degree cells (red). (F) Flow diagram illustrating the calculation of cell phase, calculated from the correlation coefficient and phase shift. (G) Wave propagation from early phase cells (blue) to late phase cells (red) in 3D space. (H) Frequency distribution of cell phase for all β-cells analyzed. Top 10% (blue box) and bottom 10% (red box) are early and late phase cells. (I) Representative 3D illustration and Ca²⁺ traces showing the location and traces of high phase cells (blue) and low degree cells (red). (J) Quantification of the normalized distance from the islet center for average phase cells (gray), early phase cells (blue), and late phase cells (red). Data represent n = 28,855 cells, 33 islets, 7 mice. Data are displayed as mean ± SEM. ****p < 0.0001 by one-way ANOVA.

Figure 3.. The network of highly synchronized β-cells is consistent between oscillations, while the Ca²⁺ wave axis rotates.

(A) Three-dimensional (3D) representation of the islet showing the location of high degree cells (blue) and low degree cells (red) over three consecutive oscillations (top panel) and their corresponding Ca²⁺ traces (bottom panel). (B) 3D representation of the islet showing the location of early phase cells (blue) and late phase cells (red) over three consecutive oscillations (top panel) and their corresponding Ca²⁺ traces (bottom panel). (C) Quantification of the retention rate of high degree and early phase cells. (D) Relative spatial change in the center of gravity of β-cell network versus the β-cell Ca²⁺ wave. (E) Frequency distribution showing the normalized change in Ca²⁺ wave axis for all islets. Data are displayed as mean ± SEM. ****p < 0.0001 by normality test followed by Paired Student’s t-test or Wilcoxon Signed-Rank Test.

Figure 3—figure supplement 1.. Location of early and late phase cells in an islet with stable wave axis.

Three-dimensional (3D) representation of the islet showing the location of early phase cells (blue) and late phase cells (red) over three consecutive oscillations (top panel) and their corresponding Ca²⁺ traces (bottom panel).

Figure 4.. Cellular and regional consistency of the β-cell network and Ca²⁺ wave quantified by Kullback–Leibler (KL) divergence.

(A) Schematic showing cellular and regional consistency analyses. (B) Schematic depicting the use of KL divergence to determine consistency between consecutive oscillations. Every β-cell in the islet is ranked, with near-zero KL divergence values indicating high consistency between oscillations and near-unity KL divergence indicating randomness. Comparison of cellular versus regional consistency of the network (C) and wave (D) by KL divergence. Data are displayed as mean ± SEM. ****p < 0.0001 by normality test followed by Paired Student’s t-test or Wilcoxon Signed-Rank Test.

Figure 5.. Three-dimensional (3D) analysis is more robust than two-dimensional (2D) analysis.

(A) Example islet showing the locations of the ¼-depth (red) and ½-depth (blue) 2D planes used for analysis. (B) Comparison of wave axis change from 2D and 3D analyses. (C) Comparison of distance from center for average and high degree cells based on either 3D (left panel) or 2D planes (middle and right panels). (D) Comparison of distance from center for average and early phase cells based on either 3D (left panel) or 2D planes (middle and right panels). (E, F) Comparison of cellular and regional consistency of the network (E) and Ca²⁺ wave (F) based on either 3D (left panel) or 2D planes (middle and right panels). Data are displayed as mean ± SEM. *p < 0.05, ****p < 0.0001 by normality test followed by parametric or non-parametric one-way ANOVA (B) or Student’s t-test or Wilcoxon Signed-Rank Test (C–F).

Figure 6.. Effect of glycolytic activators on β-cell oscillations.

(A) Schematic of glycolysis showing the targets of glucokinase activator (GKa) and pyruvate kinase activator (PKa). (B) Illustration indicating the oscillation period, active phase duration, silent phase duration, and duty cycle (active phase/period) calculated at half-maximal Ca²⁺. Sample traces and comparison of period, active phase duration, silent phase duration, and duty cycle before and after vehicle (0.1% DMSO) (n = 11,284 cells, 13 islets, 7 mice) (C), GKa (50 nM RO-28-1675) (n = 6871 cells, 8 islets, 7 mice) (D), and PKa (n = 10,700 cells, 13 islets, 7 mice) (10 µM TEPP-46) (E). Data are displayed as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 normality test followed by Paired Student’s t-test or Wilcoxon Signed-Rank Test.

Figure 7.. Glucokinase activity determines the origin of Ca²⁺ waves in three-dimensional (3D) space.

(A) Illustrations showing the location change of early phase cells (blue) and late phase cells (red) before and after vehicle (left panel), GKa (middle panel), and PKa (right panel). Effect of vehicle, GKa, and PKa on regional consistency of the Ca²⁺ wave (B), wave axis change (C), early phase cell retention (D), late phase cell retention (E), and the time lag between early and late phase cells (F). Data are displayed as mean ± SEM. *p < 0.05 by Student’s t-test.

Figure 7—figure supplement 1.. Effect of glycolytic activators on the β-cell network.

Effect of vehicle, glucokinase activator (GKa), and pyruvate kinase activator (PKa) on regional consistency of the β-cell network (A), high degree cell retention (B), and low degree cells retention (C). Data are displayed as mean ± SEM. *p < 0.05 by Student’s t-test.

Figure 7—figure supplement 2.. Effect of glycolytic activators on the β-cells that depolarize first.

Effect of vehicle, glucokinase activator (GKa), and pyruvate kinase activator (PKa) on the retention (A), regional consistency (B), and wave axis change (C) of the β-cells that depolarize first (Early depolarizer). The retention of β-cells that depolarize first (Early depolarizer) is similar to cells that depolarize and repolarize first (Early phase). Data are displayed as mean ± SEM. *p < 0.05 by Student’s t-test.

Author response image 1.

Author response image 2.