Fig 1.
Confocal microscopic determination of geometric and affinity parameters required for the calculation of ΔψM in rat primary β-cells.
(A) The mitochondria:cell volume fraction (VF) was calculated from confocal micrographs of live rat primary β-cells loaded with MitoTracker Red CMX and Calcein-AM (a-b). β-Cells were identified by insulin immunocytochemistry (ICC; c). MitoTracker images were high pass filtered (d) and both Calcein and MitoTracker images were binarized (e) to determine cellular and mitochondrial cross sections, respectively. Each β-cell was imaged in one of 10 serial z-planes to provide a balanced representation of all planes, indicated here by the three rows of images of separate cells. Cross sections of non-β cells were manually removed as indicated in the midplane images (the boundaries of the β-cell is marked by the dashed line). VF was calculated by the indicated formula after summing the number of all mitochondrial and all cellular (including mitochondrial) pixels in all cells in a recording of 50–100 cells (represented by column e). (B) aR’ was calculated from TMRM fluorescence after depolarization of mitochondria by FCCP (1 μM), oligomycin (1 μg/ml) and antimycin A (1 μM). aR’ is the slope of the linear relationship between nuclear and mitochondrial fluorescence (F; as marked by the dashed line in the raw and gated fluorescence images, respectively) as TMRM fluorescence intensity decays as it leaks out of the cell, and b is background fluorescence. The identity of β-cells was determined by post-hoc immunofluorescence imaging, revisiting stored coordinates (right). Scale bars, 5 μm.
Fig 2.
Measurement of ΔψM in rodent primary β-cells.
(A) Wide-field fluorescence view field of a dispersed rat islet cell culture loaded with PMPI (left; green) and TMRM (right; red). a) 3 mM glucose baseline; b) 16 mM glucose; c) complete depolarization at the end of the experiment, corresponding to the time courses shown in B-G; d) insulin immunofluorescence (red) and Hoechst 33342 nuclear staining (blue) in the same view field. (B-C) Fluorescence intensity time courses of the potentiometric probes. (D-E) Time courses of calibrated potentials corresponding to B-C were calculated using VF = 0.078 and aR’ = 0.296 and biophysical constants of the probes as previously published [21]. Data are mean±SE of n = 34 cells individually calibrated and pooled in five view fields in a single well, and representative of 3 independent experiments. Potentials were calibrated in single or aggregates of insulin-positive cells as indicated by regions of interests in Aa. (F-G) Calibrated potentials in 3 (groups of) cells indicated in Ab by arrows. Error bars indicate the predicted SE of the calibration calculated by propagation of the errors in determinations of calibration parameters. (H) Calculation of parameters of ΔψP calibration using linear regression on normalized fluorescence intensities and extracellular [K+] ([K+]ec) values during “calibration steps” shown in B. (I) Calculation of parameters of ΔψM calibration using linear regression on normalized fluorescence intensities of TMRM decay. H and I correspond to cells indicated in Aa, F and G. Normalization was performed as previously published [21].
Fig 3.
Stability of ΔψM during culturing of human primary β-cells in whole islet and dispersed cell cultures.
(A) ΔψM in 3 mM glucose (rich medium) as a function of days spent in whole-islet suspension culture before dispersion. ΔψM was measured 3–4 days after dispersing islet cells. (B) ΔψM hyperpolarization evoked by 16 mM glucose corresponding to (A). (C) ΔψM in 3 mM glucose (rich medium) as a function of days spent in dispersed islet cell culture after an arbitrary length of whole islet culturing. (D) ΔψM hyperpolarization evoked by 16 mM glucose corresponding to (C).
Fig 4.
Concentration-dependent hyperpolarization of ΔψM and ΔψP by glucose in rat primary β-cells.
(A-B) Absolute magnitudes of ΔψP and ΔψM. For each [glucose], first the baseline (open squares) was recorded in the presence of 3 mM glucose in a rich medium then [glucose] was elevated to the indicated value (closed circles). (C-D) Depolarization of ΔψP and hyperpolarization of ΔψM relative to baseline. The square marks the zero value, while closed circles mark the relative change from the baseline. Data are mean±SE of n = 3 experiments conducted similarly to the one shown in Fig 2. The response to glucose was determined as the mean of potentials measured at 30–60 min after addition of glucose.
Fig 5.
Biphasic glucose-induced ΔψP depolarization is linked to stronger, but monophasic ΔψM hyperpolarization.
(A) ΔψP in glucose-stimulated rat β-cells. β-Cells were categorized into two groups based on ΔψP with monophasic (open symbols) or biphasic (closed symbols) response to 12–16 mM glucose. (B) ΔψM in β-cells categorized based on ΔψP in (A). Data are mean±SE of n = 4 experiments where individual categorized cells within each experiment were expressed as a single mean value. The p-value compares ΔψM (averaged in the interval of 30–60 min after glucose stimulation) between the two groups using paired t-test for experimental replicates.
Fig 6.
Heterogeneous response of individual rat β-cells to glucose.
(A) ΔψP as a function of ΔψM in the presence of 4.5–16 mM glucose (solid isolines and closed squares) or 50 μM glibenclamide (dotted isolines and open circles). Two-dimensional histograms show the distribution of absolute potentials measured 30–60 min after secretagogue addition. Histograms were compiled from 364 and 55 cells for glucose and glibenclamide, respectively, pooled from 3 experiments (spanning the range of 4.5–16 mM [glucose] in each experiment using different wells, corresponding to data in Fig 4). Symbols indicate mean±SE of potentials calculated by binning cells along the ΔψM axis (n = 3 experiments). (B) ΔψP depolarization relative to baseline as a function of ΔψM hyperpolarization relative to baseline evoked by glucose or glibenclamide calculated from data in (A). (C-D) Predicted uncertainty of single-cell potential calibration when the absolute magnitude of stimulation-evoked potentials (C; corresponding to A) or the relative changes to baseline (D; corresponding to B,E,F) are calculated. The distributions marked by isolines visualize stochastic uncertainty originating from detection error assuming constant (top distribution) or cell-to-cell variable VF (bottom distribution). Straight lines indicate the direction and magnitude of the predicted SE of a stimulation-evoked transient change in the efficiency of collecting single-cell fluorescence of the indicated magnitude. Notably the standard error of the stimulation-evoked total fluorescence change measured using the pH-insensitive fluorescence of superecliptic synaptopHluorin corresponds to the top line. (E) Relationship between relative potential changes evoked by low glucose (4.5–6 mM; dotted isolines and open circles; 79 cells) and high glucose stimulation (8–16 mM; solid isolines and closed squares; 285 cells). (F) Relationship between relative potential changes evoked by high glucose (8–16 mM; 285 cells) in the first phase of stimulation (0–5 min; dotted isolines and open circles) and in the plateau phase of the stimulation (30–60 min solid isolines and closed squares). (A-B,E-F) For each bin, values of ΔψP in the two groups were compared by t-test performed on the experimental repeats (*p<0.05; ns, not significant. p is not indicated for bins where data were insufficient).
Fig 7.
Assaying the effect of oligomycin on ΔψM using rhodamine 123 in quench mode in insulinoma cells results in internally inconsistent data.
(A) Whole cell fluorescence, (B) mitochondrial fluorescence obtained by high pass filtering and (C) fluorescence over the nuclei in rhodamine 123-loaded INS-1E cells following the additions of glucose (30 mM), oligomycin (2 μg/ml) and FCCP (1 μM). Data are mean±SE of n = 3 experiments. Note that the oligomycin-evoked drop in quench-mode whole-cell fluorescence (A) originates from the mitochondria (B) and not from the cytosol (approximated by rhodamine 123 fluorescence in the nucleosol) (C). (D) Representative images of INS-1E cells. From top to bottom: raw fluorescence, high pass filtered fluorescence to selectively transmit mitochondrial signal, and the temporal maximum intensity projection of the latter image series indicating that no mitochondria moved into the ROIs (dashed outlines) used to measure nuclear fluorescence (N). (E) Raw fluorescence, corresponding from top to bottom to baseline, glucose and oligomycin treatments. The red color indicates high intensity. (D-E) Scale Bar, 10 μm. (F) Mitochondrial fluorescence obtained by high pass filtering in INS-1 832/13 cells in the same condition as above. Representative trace of three experiments.
Fig 8.
Glucose stimulation and oligomycin inhibition of INS-1 832/13 cells followed using the absolute ΔψP and ΔψM assay.
(A) ΔψP and (B) ΔψM were assayed by measuring whole-well fluorescence of TMRM and PMPI in a microplate reader. Cultures were stimulated with glucose (10 mM) and treated with oligomycin (2 μg/ml) as indicated. Fluorescence traces were calibrated based on the same principles as shown in Fig 2; the time courses are shown truncated before calibration. Data are mean±SE of n = 10 independent experiments.