Optogenetic β cell interrogation in vivo reveals a functional hierarchy directing the Ca2+ response to glucose supported by vitamin B6

Abstract

Coordination of cellular activity through Ca²⁺ enables β cells to secrete precise quantities of insulin. To explore how the Ca²⁺ response is orchestrated in space and time, we implement optogenetic systems to probe the role of individual β cells in the glucose response. By targeted β cell activation/inactivation in zebrafish, we reveal a hierarchy of cells, each with a different level of influence over islet-wide Ca²⁺ dynamics. First-responder β cells lie at the top of the hierarchy, essential for initiating the first-phase Ca²⁺ response. Silencing first responders impairs the Ca²⁺ response to glucose. Conversely, selective activation of first responders demonstrates their increased capability to raise pan-islet Ca²⁺ levels compared to followers. By photolabeling and transcriptionally profiling β cells that differ in their thresholds to a glucose-stimulated Ca²⁺ response, we highlight vitamin B6 production as a signature pathway of first responders. We further define an evolutionarily conserved requirement for vitamin B6 in enabling the Ca²⁺ response to glucose in mammalian systems.

INTRODUCTION

Pancreatic β cells sustain glucose homeostasis by secreting precise amounts of insulin in proportion to energy-rich nutrients such as sugars. Insulin secretion in response to a stepped increase in glucose involves two phases (1). During the first phase, β cell depolarization triggers a rapid and transient increase in intracellular Ca²⁺ and the release of insulin. During the second phase, insulin is released in a more sustained and pulsatile manner. People with prediabetes or impaired glucose tolerance have a diminished first phase of secretion, while those with established type 2 diabetes (T2D) lose the first phase and show reduced second-phase insulin secretion (2). The pulsatility of secretion is also lost in T2D (3). Hence, it is of critical importance to understand the mechanisms of coordinated insulin secretion to develop effective therapies for diabetes.

Because pancreatic β cells undergo a synchronized influx of Ca²⁺ in response to glucose, they are often considered as a functional syncytium in which all of the cells work as a unit (4, 5). Pioneering work from the Benninger (6) as well as the Stožer and Rupnik (7) groups reported intercellular heterogeneity during the first and second phases and also demonstrated the importance of gap-junction coupling and sugar metabolism for β cell synchronization. These findings were later extended to the human islet by the Rutter group (8). Using isolated mouse islets, Ca²⁺ imaging, and optogenetics, Westacott et al., Johnston et al., and Stožer et al. further identified a subpopulation of highly connected β cells (“hubs”) that serve as possible pacemakers and regulate the Ca²⁺ waves during the oscillatory second phase of insulin release (6, 7, 9). Subsequently, leader cells, which initiate repeated Ca²⁺ waves during the second phase, were also identified and examined in mouse islets engrafted into the anterior chamber (6, 10, 11). In parallel, we have defined a critical subpopulation of β cells, which lead the Ca²⁺ increase during the first-phase response in zebrafish (11). These cells are referred to here as “first responders,” according to the recent terminology proposed by Kravets et al. and Stožer et al. in mouse islets (12, 13). The laser-ablation of first responders in zebrafish can cause a decrease in the ability of the β cell population to mount an effective Ca²⁺ response to glucose (11). Recently, by analogous laser-cell ablation experiments of first responders in cultured mouse islets, Kravets et al. confirmed the importance of first responders for the first-phase response (12) and strengthened the view that the control over islet function is highly conserved across vertebrates. Nevertheless, the existence of controlling β cells and their properties has become the focus of intense investigation using both mouse and human islets (14,15) and has generated significant debates in the field, as reviewed recently (16–19). Therefore, the analysis of β cells in their natural environment using intravital microscopy in zebrafish and mice could help to address the outstanding questions about the functional heterogeneity of β cells and the underlying molecular differences.

While the zebrafish islet differs in its smaller size and relative simplicity of Ca²⁺ oscillations compared to mouse and human islets (11), the in vivo environment provides significant advantages when recording β cell activity where normal blood flow, innervation, and paracrine interactions are all preserved. Upon blood glucose infusion, zebrafish exhibit a very prominent and synchronized initial Ca²⁺ response, yet they show fewer secondary Ca²⁺ oscillations that are otherwise observable in mouse islets cultured under persistently high glucose (11, 20). A recent study, however, has shown that mouse islets imaged intravitally in the anterior chamber of the eye exhibit a Ca²⁺ response to glucose with a prominent first-phase response but lacking marked secondary oscillations, as long as persistent hyperglycemia is avoided, akin to the situation in zebrafish (21).

The invasive approaches for studying β cell connectivity, such as irreversible cell ablation, have limited capability to interrogate the connectivity of multiple cells in the same tissue. To overcome this challenge, here, we deploy optogenetic tools, allowing us to depolarize or hyperpolarize individual cells in vivo in the zebrafish while assessing the impact on the first-phase Ca²⁺ response. In addition, we develop transgenic zebrafish lines enabling us to stably mark β cells that differ in the threshold to glucose-stimulated Ca²⁺ influx and, hence, to probe their gene expression differences. We find that first-responder cells represent a functionally and molecularly distinct group of β cells that control islet activity via the initiation of the first-phase Ca²⁺ response to elevated glucose. Across species, we identify a subpopulation of β cells expressing pyridoxamine 5′-phosphate oxidase (pnpo), a rate-limiting enzyme involved in vitamin B6 production, and show that this subpopulation coincides with the first-responder β cells in vivo. Using pharmacology and genetics, we reveal that pnpo and the vitamin B6 pathway support the coordinated Ca²⁺ response of β cells to glucose in zebrafish and mouse islets.

RESULTS

A proximity relationship between first-responder and follower cells directs the Ca²⁺ response in vivo

To characterize the pattern of glucose-stimulated Ca²⁺ response in vivo, we used double transgenic Tg(ins:GCaMP6s); Tg(ins:cdt1-mCherry) larvae and studied their glucose responsiveness by performing calcium imaging (Fig. 1A). The ins promoter drives specific expression of genetically encoded calcium indicator 6s (GCaMP6s) in the zebrafish β cell. Upon glucose injection into larvae, all cells exhibit an increase in GCAMP6s fluorescence (Fig. 1B). The first-responder cell is the cell with the fastest average Ca²⁺ response from three consecutive stimulations with glucose, whereas the follower cells are the ones with slower responses (see Methods) (Fig. 1, C to E). Upon glucose stimulation, we found that the cumulative time of response for all β cells was 7.62 s after glucose injection (±6.05 s SD, n = 12 islets/larvae) (assessed as the time taken to achieve a GCaMP6 signal >25% above baseline, T₂₅) (Fig. 1D). Plotting the relative speed of response of each cell as a function of its distance to the first-responder cell revealed a positive correlation between the time of response of each cell and its distance to the first-responder β cell coefficient of determination (R²) = 0.57 (± 0.24, n = 12 islets) (Fig. 1, F and G, and movie S1). These findings show that a proximity relationship between first-responder and follower cells underlies the Ca²⁺ response to glucose.

Optogenetic silencing of single β cells can inhibit the islet responses to glucose

To be able to perform reversible and repeated interrogation of multiple β cells and assess the impact on intracellular Ca²⁺, we constructed several transgenic lines with optogenetic actuators with different optical spectra and capabilities (table S1) (22–26). The Halorhodopsin from the archaea Natronomonas (NpHR) is a light-gated chloride pump, which is activated by green light (λ = 560 to 590 nm) and can inhibit cell depolarization robustly (25). In principle, activating NpHR during a glucose stimulus should inhibit the glucose-induced influx of Ca²⁺ as a result of the electrical silencing of β cells. To test whether this approach can be applied in vivo in the fish setting, we generated transgenic larvae expressing eNpHR3.0-mCherry in β cells (26). Notably, we were able to temporally inhibit the glucose-induced influx of Ca²⁺, as reported by a reduction in the GCaMP signal (fig. S1 and movie S2), indicating that NpHR excitation effectively blocks depolarization and Ca²⁺ influx even in the presence of high glucose.

Having demonstrated the feasibility of NpHR for β cell inhibition in vivo, we dissected the role of individual cells in coordinating Ca²⁺ dynamics across optical sections of the islet. We used Tg(ins:GCaMP6s); Tg(ins:eNpHR3.0-mCherry) double transgenic larvae. First, we induced a pulse of glucose to identify the first-responder cell (Fig. 2A). Subsequently, we gave additional pulses of glucose and simultaneously activated NpHR by illumination with the green laser (λ = 561), targeting either the presumptive first responder or a follower cell (Fig. 2, B and C). First-responder inhibition produced the strongest dampening of calcium influx as compared to follower-cell inhibition (Fig. 2, D and E). In the example shown in Fig. 2C, a small group of β cells showed an increase in GCaMP6s fluorescence upon inhibition of the first-responder cell, suggesting that a different cell had initiated the Ca²⁺ response albeit with lower efficiency, leading to the generation of a partial response (Fig. 2C and movies S3 to S5).

An optogenetic system for in vivo activation of β cells identifies cells with disproportionate control over Ca²⁺ dynamics

We next asked whether first-responder activation by light-regulated depolarization would trigger an influx of Ca²⁺ in the follower cells even in the absence of glucose stimulation. To this end, we generated a transgenic line expressing the blue-light–driven cation pump CheRiff under the control of the zebrafish insulin promoter (23). We then combined the CheRiff transgenic line with a red calcium reporter, Tg(ins:K-GECO1). The genetically encoded red calcium indicator K-GECO1 has a similar design to the GCaMP calcium reporter (27), and its use here allowed us to image Ca²⁺ simultaneously while controlling membrane potential via optogenetic regulation of CheRiff–green fluorescent protein (GFP). First, we explored whether CheRiff-based optogenetic activation is sufficient to induce Ca²⁺ influx in the islet. We imaged the islets of Tg(ins:K-GECO1);Tg(ins:CheRiff-GFP) larvae and then activated the CheRiff by illuminating the entire islet with a blue laser (λ = 470 nm) for 30 s while simultaneously recording Ca²⁺ dynamics. We found that it was possible to trigger an influx of calcium upon pan-islet optogenetic activation in vivo (movie S6).

Having established this optogenetic system, we performed single-cell depolarization experiments (Fig. 3, A to E, and movies S7 to S9). We selected the center of the cell using a region of interest (ROI) covering a circle with an approximate diameter of 10 μm, ensuring that a single cell in the plane of illumination is activated. We found that, upon individual activation, ~45% of β cells did not co-activate more than their immediate neighbor β cells. Around 24% of β cells activated four to five additional cells (25 to 75% of the cells). Notably, ~14% of the cells showed high potential for coordination, as they co-activated most of the cells in the imaging plane (>75% of cells; >5 cells) (Fig. 3, F and G) (n = 10 islets). Thus, we find evidence for a functional hierarchy among β cells using two different optogenetic systems.

First-responder cells trigger Ca²⁺ activity in follower cells

Our noninvasive technology enabled us to address whether β cells with the highest capability for co-activating cells in the islet coincide with the population of first-responder β cells (11). To this end, we first gave three pulses of glucose while recording the calcium responses to identify the first-responder cell (Fig. 4, A to C, and movie S10). We then proceeded to interrogate individual β cells via CheRiff optogenetics in the same islet (Fig. 4, D to F, and movies S11 and S12). We found that, in five of the seven larvae, the first-responder cell was that with the highest potential of recruitment of other β cells, as its activation triggered co-activation of most of the cells in the imaging plane (39 cells from seven different animals) (Fig. 4, G and H). In general, the analysis of multiple islets revealed a weak but positive correlation between the temporal order of response and the ability of the cell to co-activate other cells i.e., the cells that tend to respond faster can co-activate more cells (Fig. 4I) (R² = 0.35 and ± SD 0.34). Consistent with our previous optogenetic experiment (Fig. 3), only 11% of β cells (4 of the 39 cells) triggered co-activation of >75% of the cells in the imaging plane (fig. S2) (n = 7 islets). These results demonstrate the stronger capability of first-responder cells compared to follower cells to trigger Ca²⁺ activity in the islet…..