A high-throughput assay for the measurement of Ca2+ oscillations and insulin release from uniformly sized mouse β-cell (MIN6) spheroids

Stijn Robben; Patricia Davidson; Rita S Rodrigues Ribeiro; Thomas Voets

doi:10.7554/eLife.110428

. 2026 Oct 5;15:RP110428. doi: 10.7554/eLife.110428

A high-throughput assay for the measurement of Ca²⁺ oscillations and insulin release from uniformly sized mouse β-cell (MIN6) spheroids

Stijn Robben ^1,^2,³, Patricia Davidson ³, Rita S Rodrigues Ribeiro ³, Thomas Voets ^1,^2,^✉

Editors: Leon D Islas⁴, Merritt Maduke⁵

PMCID: PMC13638198 PMID: 42832264

Abstract

Diabetes mellitus is a rapidly growing global health challenge, necessitating the development of more effective anti-diabetic therapies, including drugs that improve insulin release from pancreatic β-cells. Traditional high-throughput screening methods typically rely on 2D β-cell cultures, but such cultures do not mimic the 3D organization and cell-to-cell communication of β-cells in pancreatic islets of Langerhans. Existing 3D β-cell culture models are hindered by high costs, technical complexity, and limited compatibility with high-throughput screening platforms. In this work, we developed an approach for generating 19 homogeneously shaped pancreatic β-cell spheroids in each well of a 96-well plate, using micropatterned polyethylene glycol (PEG)-based hydrogels and murine mouse insulinoma 6 (MIN6) insulinoma cells. The uniform shape and positioning of the individual spheroids enabled the simultaneous, real-time imaging of Ca²⁺ signals in up to 1824 independent spheroids in response to glucose and various test compounds. Using this approach, we show that increasing glucose causes concentration-dependent Ca²⁺ oscillations in individual spheroids, that these Ca²⁺ oscillations are sensitive to modulators of ATP-sensitive K⁺ channels, and that the frequency of Ca²⁺ oscillations correlates with insulin secretion. Finally, we demonstrate that the neurosteroid pregnenolone sulfate, an agonist of the cation channel TRPM3, increases the frequency of glucose-induced Ca²⁺ oscillations and enhances insulin release; the TRPM3 antagonist isosakuranetin inhibited these responses. In conclusion, we established a cost-effective and scalable 3D β-cell platform for high-throughput screening of insulin release-modifying compounds, with potential applications in drug development and personalized medicine for the management of diabetes mellitus.

Research organism: Mouse

Introduction

As one of the fastest growing diseases, diabetes mellitus is estimated to affect 783.2 million adults (20–79 years) worldwide by 2045 (Cho et al., 2018). Characterized by insulin resistance and/or abnormal insulin secretion, this disease leads to increased mortality, blindness, kidney failure, and an overall decreased quality of life (Cole and Florez, 2020). In 2019, it was estimated that 4.2 million adults aged 20–79 years died from diabetes-related complications, with almost half of these deaths (1.9 million) estimated to have occurred in adults below the age of 60 (Saeedi et al., 2020). These premature deaths combined with disability due to diabetes and absenteeism from work contribute to an astronomical economic impact, estimated to a total of 1.31 trillion USD in 2019 alone (Bommer et al., 2017). Hence, the development of effective anti-diabetic drugs, including improved insulin secretagogues, has been a priority for the pharmaceutical sector.

In vitro cell-based high-throughput screening methods often form a starting point for the development of new drugs (American Diabetes Association, 2010; DeFronzo and Tripathy, 2009). For the discovery of insulin secretagogues, currently used screening methods include cytosolic calcium measurements and insulin secretion assays, primarily performed on pancreatic β-cells cultured in a 2D monolayer (Vhora et al., 2020). However, 2D cell models can be suboptimal predictors of medical outcomes in clinical trials due to the lack of physiological or morphological relevance to in vivo conditions (Rupnik, 2009; Wikstrom et al., 2012; Sun et al., 2022). This is particularly relevant for pancreatic β-cells, which are clustered in vivo in the islets of Langerhans. In these highly metabolically active spherical clusters of endocrine cells, dispersed throughout the exocrine tissue of the pancreas, β-cells exhibit coordinated activity in response to glucose, which leads to islet calcium waves and pulsatile insulin release. To date, only a few studies have explored the application of 3D β-cell culture for developing high-throughput-compatible screening methods (Amin et al., 2016; Lee et al., 2018). The main challenges remain the high costs, technical complexity, time consumption, and issues with reproducibility due to variability and lack of standardization (Mohandas et al., 2023). Furthermore, the limited compatibility of these models with high-throughput read-out methods poses a significant barrier to their integration into industrial applications, complicating the development of a standardized and reproducible assay that meets all necessary requirements.

In this work, we introduce a novel approach which allows the precisely patterned creation of uniform 3D pancreatic β-cell spheroids within a 96-well plate, utilizing micropatterned polyethylene glycol (PEG)-based hydrogels. Due to the unique design and properties of the plates, we manage to achieve simultaneous real-time monitoring of insulin release-associated Ca²⁺ fluorescence signals at the individual spheroid level, yielding a total of up to 1824 unique measurements per plate (19 spheroids per well), and directly correlate these signals with insulin release. This method solely relies on conventional cell culture techniques, with results obtainable within 24 hr after cell seeding. As a proof of principle, we illustrate the concentration-dependent effects of modulators of ATP-sensitive K⁺ (K-ATP) and TRPM3 channels on glucose-induced Ca²⁺ oscillations and insulin release.

Results

Hydrogel design

To facilitate spheroid formation while ensuring compatibility with high-throughput calcium screening, 316L stainless-steel molds were designed to produce gels conforming to the dimensions of the well of a 96-well plate and contain 19 identical microwell structures into which the cells will sediment after seeding (Figure 1A–C). The molds were used to stamp multiple gels in a single polymerization step, resulting in a PEG-based hydrogel-covered glass-bottom 96-well plate (Figure 1D). The uniformity in the shape and dimensions of the fabricated microstructured hydrogels was assessed by quantifying the contact point diameters. These contact points (Figure 1E, top) correspond to regions where the molds make direct contact with the glass substrate during polymerization, inhibiting hydrogel formation in those areas. The mean contact point diameter was determined to be 183 µm±23 µm (mean ± SD), with no statistically significant variation observed between samples (p>0.05) (Figure 1E, bottom).

Figure 1. — (A) Schematic top-view representation of the hydrogel, illustrating a honeycomb arrangement of 19 microwells per gel. (B) Schematic cross-sectional depiction of the microwells (not to scale). (C) Computer-aided design (CAD) rendering of the hydrogel positioned within the well of a glass-bottom 96-well plate. (D) Entire glass-bottom 96-well plate with microstructured polyethylene glycol (PEG)-based hydrogel in each well. (E) Microwell with mouse embryonic fibroblasts attached to the contact point, which is a result of the molds being in contact with the glass bottom during polymerization (top). Mean contact point diameter for two independently moulded samples. N=16 wells (304 contact points) per sample. Statistical difference was analyzed using one-way ANOVA and Tukey’s post hoc test, where ns = not significant.

Figure 1—source data 1. Contact point diameters per sample.

elife-110428-fig1-data1.xlsx^{(21.3KB, xlsx)}

Protocol

After determining that the microwells of our hydrogels within the 96-well plate are of consistent shape, we focused on developing a protocol to assess intracellular calcium signaling in the 3D pancreatic β-cell spheroids. Figure 2A illustrates the established timeline, including the tissue formation, loading of the tissues with a Ca²⁺-sensitive dye, and data acquisition. Within 16 hr post-seeding of the mouse insulinoma 6 (MIN6) β-cells onto the hydrogels, 3D tissues had formed within the microwells (Figure 2B), and changes in the Ca²⁺ influx-associated fluorescent signal of individual spheroids could be measured in real time using a fluorescence plate imager (Figure 2C). The contact points exposing the glass substrate facilitate tissue adhesion, preventing spheroid loss during washing steps and liquid deposition or aspiration by the µCELL FDSS. Following all experimental manipulations, 75.0% of the wells retained all 19 spheroids, with no wells losing more than three spheroids despite the extensive washing steps.

Figure 2. — (A) Schematic representation of protocol timeline. 200 μL of the mouse insulinoma 6 (MIN6) pancreatic β-cell suspension is added to each well of the 96-well plate at hour 0. Over the next 16 hr, the cells undergo sedimentation into the microwells and spontaneously self-assemble into spheroids. Two hours prior to data acquisition, 10 μM of Cal-520 is introduced for 1 hr, followed by washing and incubation in a low-glucose KREBS solution for 30 min (hours 17–18). At hour 19, the plates are transferred to the μCELL FDSS for fluorescence-based data acquisition. (B) Representative images of MIN6 β-cells following sedimentation within a microwell, captured 10 min post-seeding (top) and 16 hr post-seeding (bottom). (C) Representative fluorescence image of a section of a 96-well plate containing the Cal-520 AM-loaded MIN6 β-cell spheroids acquired using the µCELL FDSS fluorescence plate imager, distinctly showing individual spheroids arranged in a honeycomb pattern within the microwell array. (D) Histogram showing the number of spheroids still present per well at the time of the control stimulus. N=48 wells on two independent occasions.

Figure 2—source data 1. Number of spheroids still present per well at the time of the control stimulus.

elife-110428-fig2-data1.xlsx^{(15.7KB, xlsx)}

Increased levels of extracellular glucose induce Ca²⁺-oscillatory behavior

Next, we evaluated the spheroid Ca²⁺ signal response to increases in extracellular glucose concentration (Figure 3). At the basal glucose concentration (3 mM), the signal remained stable, free of any oscillatory behavior. However, periodic oscillations emerged at extracellular glucose concentrations exceeding 5 mM, with increased frequency at higher glucose concentrations (Figure 3A and B), whereas no oscillations were observed in 2D cultures of the same cells (Figure 3—figure supplement 1). To quantify the glucose-induced Ca²⁺ responses, we determined two main oscillatory parameters: the frequency and the area under the peaks (AuP), as illustrated in Figure 3A. Both the frequency and AuP increased in a concentration-dependent manner with higher glucose, with concentrations for half-maximal activation (EC₅₀ values) of 7.4±0.4 mM and 7.9±0.4 mM for frequency and AuP, respectively. Figure 3C displays the fluorescence traces of all individual spheroids within a single well, prior to baseline subtraction, and the average of those signals. Figure 3D presents a comparison of the average Ca²⁺ fluorescence signal recorded from a well containing 3D-cultured MIN6 β-cell spheroids and a well containing MIN6 cells cultured under conventional 2D conditions.

Figure 3—figure supplement 1. — (A) Representative Ca²⁺ response of a single 3D mouse insulinoma 6 (MIN6) β-cell spheroid to an increase in extracellular glucose concentration (20 mM), highlighting the visual representations of the peaks and the area under the peaks. The spheroids were initially exposed to basal extracellular glucose levels (3 mM) for 10 min, followed by an increase in extracellular glucose (20 mM) for 30 min. The protocol ended with a control stimulus of 50 mM KCl. (B) Mean frequency and the normalized area under the peaks of the fluorescent signal during the 30 min of exposure to increased concentrations of extracellular glucose. Peak threshold was set to three times the signal’s standard deviation during the initial 10 min of exposure to basal levels of extracellular glucose. Error bars represent the standard error of the mean (SEM). N>50 spheroids per concentration. Signal values for each spheroid were normalized to the amplitude of the control stimulus. (C) Overview of Ca²⁺ responses of each individual MIN6 β-cell spheroid within a single well to an increase in extracellular glucose concentration (20 mM). To preserve the absolute response dynamics and inter-spheroid variability, no baseline correction was applied to the displayed traces. Gray lines represent individual spheroids, while the black line denotes the mean response across all spheroids. (D) Comparison of the mean Ca²⁺ fluorescence response of all individual MIN6 β-cell spheroids within a single-well (solid line) and 2D-cultured MIN6 cells (dashed line) following stimulation with 20 mM extracellular glucose.

Figure 3—source data 1. Source data for Figure 3A, B, C and D.

elife-110428-fig3-data1.xlsx^{(682.8KB, xlsx)}

K_ATP-agonist diazoxide inhibits Ca²⁺-oscillatory response to extracellular glucose

It is well established that glucose-induced insulin release in vivo depends on the inhibition of K-ATP channels, which close when cytosolic ATP levels rise in response to glucose in flux and metabolism. To assess the dependence of glucose-induced calcium oscillations in the MIN6 β-cell spheroids on K-ATP channel closure, we exposed spheroids to different concentrations of diazoxide, a K-ATP channel opener, prior to stimulation with 10 mM glucose. Diazoxide caused a concentration-dependent inhibition of the glucose-induced oscillations (Figure 4A). Both the frequency and AuP decreased with increasing diazoxide, with concentrations for half-maximal inhibition (IC₅₀ values) of 2.9±0.1 μM and 1.9±1.71 μM, respectively.

Figure 4. — (A) Characteristic Ca²⁺ responses to 10 mM extracellular glucose concentrations of individual mouse insulinoma 6 (MIN6) β-cell spheroids with and without 30 min of prior exposure to 100 μM diazoxide. The spheroids were initially exposed to basal extracellular glucose levels for 5 min after which the spheroids were exposed to varying concentrations of diazoxide for 30 min, followed by a 30 min period of extracellular glucose concentrations of 10 mM until a control stimulus of 50 mM KCl. Each signal was normalized over the amplitude of the control stimulus. (B) Average of the frequency and the normalized area under the peaks of the fluorescent signal during the 30 min of exposure to increased concentrations of extracellular glucose. Peak threshold was set to three times the signal’s standard deviation during the initial 10 min of exposure to basal levels of extracellular glucose. Error bars represent the standard error of the mean (SEM). N>40 spheroids per concentration.

Figure 4—source data 1. Source data for Figure 4 A & B.

elife-110428-fig4-data1.xlsx^{(139.5KB, xlsx)}

K_ATP-antagonists induce Ca²⁺-oscillatory behavior

K_ATP-channel antagonists, such as the sulfonylureas and meglitinides, are widely used to increase insulin release in patients with type 2 diabetes. To test the efficacy of K_ATP-channel antagonists in our MIN6 β-cell spheroid assay, we exposed spheroids in 3 mM glucose to increasing levels of glimepiride for 30 min, and then increased extracellular glucose to 10 mM. As illustrated in Figure 5A, at a concentration of 100 nM, the potent sulfonylurea antagonist glimepiride induced robust calcium oscillations. Glimepiride increased both the frequency and AuP in a concentration-dependent manner, with EC₅₀ values of 3.3±0.6 nM and 3.4±1.5 nM, respectively. Similar results, albeit with a different concentration dependence, were obtained with the less potent, first-generation sulfonylurea antagonist tolbutamide (EC₅₀ values of 6.4±6.1 μM for the frequency and 4.3±7.7 μM for the AuP) and with the meglitinide antagonist nateglinide (EC₅₀ values of 44.5±1.23 nM for the frequency and 67.4±0.623 nM for the AuP). Taken together, these results illustrate the ability of the MIN6 β-cell spheroid assay to quantitatively measure the effects of K_ATP modulators on intracellular Ca²⁺ dynamics.

Figure 5. — (A) Characteristic Ca²⁺ responses of individual mouse insulinoma 6 (MIN6) β-cell spheroids to 10 mM of extracellular glucose with and without prior exposure to 100 nM glimepiride. The spheroids were initially exposed to basal extracellular glucose levels for 5 min after which the spheroids were exposed to varying levels of glimepiride for 30 min, followed by a 30 min period of glimepiride+extracellular glucose concentrations of 10 mM until a control stimulus of 50 mM KCl. Each signal was normalized over the amplitude of the control stimulus. (B) Average of the frequency and the normalized area under the peaks of the fluorescent signal during the 30 min of exposure to various concentrations of glimepiride, nateglinide, and tolbutamide. Peak threshold was set to three times the signal’s standard deviation during the initial 10 min of exposure to basal levels of extracellular glucose. Error bars represent the standard error of the mean (SEM). N>50 spheroids per concentration.

Figure 5—source data 1. Source data for Figure 5 A & B.

elife-110428-fig5-data1.xlsx^{(181.7KB, xlsx)}

K_ATP-channel modulators impact insulin secretion by MIN6 spheroids

To determine whether Ca²⁺ oscillation in the MIN6 β-cell spheroids translates into insulin secretion, we compared calcium signals and insulin release in three conditions: 8 mM extracellular glucose in the presence of either vehicle, 100 nM glimepiride, or 100 µM diazoxide. After measuring calcium signals during 60 min (Figure 6A and B), supernatant was aspirated from the wells, and the insulin content of aspirated samples was quantified using ELISA (Figure 6C). The results indicate that treatment with 8 mM glucose caused calcium oscillations and increased insulin secretion. Inclusion of glimepiride (100 nM) not only increased the frequency of oscillations, but also robustly enhanced insulin release. Oppositely, diazoxide (100 μM) abolished calcium oscillations and reduced the insulin in the supernatant to the level of non-stimulated spheroids. These data demonstrate that the frequency of evoked calcium oscillations in MIN6 β-cell spheroids positively correlates with insulin release.

Figure 6. — (A) Characteristic Ca²⁺ responses of individual MIN6 β-cell spheroids. The spheroids were initially exposed to basal extracellular glucose levels for 10 min after which the spheroids were exposed to either 8 mM of extracellular glucose or 8 mM of extracellular glucose supplemented with 10 nM glimepiride or 10 µM diazoxide, followed by a control stimulus of 50 mM KCl. Each signal is normalized over the maximum value of the control stimulus. (B) Average of the frequency and the normalized area under the peaks of the fluorescent signal during the 60 min of exposure to the various modulators. Peak threshold was set to three times the signal’s standard deviation during the initial 10 min of exposure to basal levels of extracellular glucose. Error bars represent the standard error of the mean (SEM). N>50 spheroids per concentration. (C) Corresponding insulin concentrations present in the supernatant of the wells at the 1 hr timepoint quantified using ELISA. Data is presented as means ± SEM (n=4 wells per condition). Statistical difference was analyzed using one-way ANOVA and Tukey’s post hoc test, where ns = not significant, **p<0.01, ****p<0.0001.

Figure 6—source data 1. Source data for Figure 4 A, B & C.

elife-110428-fig6-data1.xlsx^{(200.3KB, xlsx)}

PS induces Ca²⁺-oscillatory behavior and increases insulin secretion

To evaluate whether other signaling pathways can influence Ca²⁺ signals in MIN6 β-cell spheroids, we tested the effect of pregnenolone sulfate (PS), an endogenous neurosteroid and agonist of the cation channel TRPM3. At a concentration of 50 μM, PS strongly potentiated glucose-induced calcium oscillatory behavior (Figure 7A and B), and this was associated with enhanced insulin release (Figure 7C). To evaluate whether the PS-induced increase in calcium oscillations and insulin release is dependent on TRPM3, we performed additional experiments in which PS was applied in the absence or presence of the potent TRPM3 antagonist isosakuranetin (ISN; 10 μM). These experiments revealed that ISN partly reversed effects of PS on calcium oscillations and insulin release, indicating that the positive effects of PS on insulin release are at least partly mediated by TRPM3. To assess the robustness and assay quality of the calcium imaging measurements in MIN6 β-cell spheroids, the Z-factor was calculated and determined to be 0.6.

Figure 7. — (A) Characteristic Ca²⁺ responses of individual mouse insulinoma 6 (MIN6) β-cell spheroids exposed to an extracellular glucose concentration of 8 mM, with and without 50 μM PS. Spheroids were initially exposed to a baseline extracellular glucose concentration of 3 mM for 10 min, followed by a 1 hr incubation with varying glucose concentrations, with or without 50 μM PS. After 60 min, 15 μL of supernatant was collected for insulin level analysis, concluding with a control stimulus of 50 mM KCl. (B) Quantitative analysis of Ca²⁺ oscillation parameters. The frequency and normalized area under the peaks were analyzed for spheroids exposed to glucose concentrations of 3 mM, 8 mM, and 20 mM, with or without 50 μM PS. Peak threshold was set at three times the signal’s standard deviation during the initial 10 min of exposure to basal levels of extracellular glucose. (C) Corresponding insulin concentrations present in the supernatant of the wells at the 1 hr timepoint quantified using ELISA. Data is presented as means ± SEM (n=4 wells per condition). (D) Characteristic smoothed Ca²⁺ responses of individual MIN6 β-cell spheroids exposed to an extracellular glucose concentration of 8 mM+50 μM PS, with and without 10 μM isosakuranetin (ISN). Spheroids were initially exposed to a baseline extracellular glucose concentration of 3 mM for 5 min, followed by a 1 hr incubation with 8 mM of extracellular glucose+50 μM PS with and without 10 μM ISN. After 60 min, 15 μL of supernatant was collected for insulin level analysis, concluding with a control stimulus of 50 mM KCl. (E) Quantitative analysis of Ca²⁺ oscillation parameters. The frequency and normalized area under the peaks were analyzed for spheroids exposed to a glucose concentration of 8 mM with or without 50 μM PS, supplemented with or without 10 μM ISN. Peak threshold was set at five times the signal’s standard deviation during the initial 10 min of exposure to basal levels of extracellular glucose. (F) Corresponding insulin concentrations present in the supernatant of the wells at the 1 hr timepoint quantified using ELISA. Data is presented as means ± SEM (n=4 wells per condition). Statistical difference was analyzed using one-way ANOVA (E and F) or two-way ANOVA (B and C) and Tukey’s post hoc test, where ns = not significant, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 7—source data 1. Source data for Figure 7 A, B, C, D, E & F.

elife-110428-fig7-data1.xlsx^{(320.3KB, xlsx)}

Discussion

In this study, we describe a platform to create homogeneously sized β-cell spheroids in a 96-well plate, and illustrate how these can be used to screen for small-molecule modulators of glucose-induced Ca²⁺ oscillations and insulin release. In particular, within 24 hr after cell seeding, 1824 spheroids (19 per well) uniformly sized spheroids are formed, and their Ca²⁺ signals can be measured simultaneously using commercial fluorescent plate imaging systems. We found that β-cells within each spheroid exhibit synchronized Ca²⁺ kinetics, resulting in spheroid calcium oscillations, whereas the 19 spheroids within the same well show similar but temporally independent Ca²⁺ signals. Analysis of insulin levels in individual wells demonstrated the correlation between oscillatory frequency and insulin secretion.

Pancreatic β-cells in 3D cell culture have been used in earlier studies, with numerous reports highlighting the advantages of spheroid-based approaches to emulate in vivo pancreatic islets (Lee et al., 2018; Diane et al., 2023; Essaouiba et al., 2021). Several methods have been described for the generation of these spheroids, including the use of the hanging-drop method, microwell systems, scaffold-based culture, and ultra-low adherence plates (Białkowska et al., 2020). From these, microwell systems have emerged as the most cost-effective and scalable option for producing uniform and functional β-cell spheroids (Wassmer et al., 2020). One earlier study reported the use of the microwell method combined with microfluidics to form β-cell spheroids, and applied Ca²⁺ imaging to a single spheroid, observing synchronized Ca²⁺ oscillations throughout the spheroid (Lee et al., 2018). However, large-scale, parallel Ca²⁺ imaging of 3D β-cell tissues, with comparable throughput as in 2D screening methods, has not been achieved. We present a method that addresses this gap by providing a platform enabling high-throughput screening performance, as reflected by a Z′-factor of 0.6, indicating excellent assay quality for drug screening, together with detailed analysis of Ca²⁺ influx oscillations in pancreatic β-cells cultured within a highly reproducible 3D environment. This approach enables screening for insulin secretagogues in a 3D environment, enhancing the physiological relevance and increasing the predictive power of drug discovery efforts.

The β-cell spheroids develop by self-organization of MIN6 cells guided by the specific design of our micropatterned hydrogel, which we here implemented within the individual wells of a 96-well plate. The molding technique with which the gels are made within each well exposes the glass through the contact point located at the bottom of each microwell, thus facilitating adhesion of the cells to the glass surface, thereby immobilizing the spheroids during manipulations of the liquid inside the well. The coordinates of the spheroids relative to each other are identical in every well, which allows standardization and automation of the analysis obtained from the fluorescent plate imager. Due to the glass bottom and the transparency of the hydrogel, the plates are also amenable to conventional microscopy. The current design of the mold contains 19 microwells per well and therefore provides multiple replicates per condition, thereby reducing the variability and increasing the statistical power of the approach. The global oscillatory signal implies that the cells within a single spheroid respond to stimuli in a coordinated and synchronized manner (Nielsen et al., 2012), a phenomenon also observed in islets of Langerhans in vivo (Gilon et al., 2002). Importantly, such oscillatory activity was not observed in parallel experiments using MIN6 2D monolayers (Figure 3, Figure 3—figure supplement 1), highlighting the importance of using 3D spheroids.

The results that we obtained using MIN6 β-cell spheroids align both qualitatively and quantitatively very well with published data obtained using isolated islets of Langerhans. For instance, we obtained EC₅₀ values of 7.4±0.4 mM and 7.9±0.4 mM for glucose-induced calcium oscillations, consistent with previous research in which electrophysiological experiments performed on mouse islets of Langerhans provided an EC₅₀ value of 8.7 mM for glucose-induced membrane potential oscillations (Antunes et al., 2000), and EC₅₀ values between 8 and 10 mM for calcium responses in human islets. Likewise, all evaluated pharmacological compounds targeting K_ATP channels modulated Ca²⁺ oscillations with concentration dependence profiles that match their known affinity to the target, with EC₅₀ values in the nanomolar range for glimepiride and nateglinide, and low micromolar IC₅₀/EC₅₀ values for diazoxide and tolbutamide (Shigeto et al., 2006; Broichhagen et al., 2014; Shigeto et al., 2007). Finally, it is well established that insulin secretion from islets is driven by oscillations in β-cell cytoplasmic Ca²⁺ (Islam, 2010), and we indeed found that the level of secreted insulin from MIN6 β-cell spheroids correlated well with the frequency of calcium oscillations.

Finally, we used the MIN6 β-cell spheroid assay to evaluate the effect of the neurosteroid PS on glucose-induced Ca²⁺ signals and insulin release. Earlier work had shown that PS evokes a rise in intracellular Ca²⁺ and enhances glucose-induced insulin release from mouse pancreatic islets and in INS-1 human insulinoma cells, and this effect was largely attributed to PS-induced activation of TRPM3 (Becker et al., 2020; Wagner et al., 2008). We found that PS robustly increases the frequency of glucose-induced Ca²⁺ oscillations in MIN6 β-cell spheroids, associated with a significant potentiation of insulin release. The effects of PS on Ca²⁺ signals and insulin release were only partly abrogated in the presence of the TRPM3 antagonist ISN, suggesting that TRPM3 may not be the only molecular target for PS in these spheroids. In these experiments, PS and ISN were applied simultaneously, meaning that a kinetic offset between TRPM3 activation by PS and the establishment of effective channel blockade by ISN may have contributed to the incomplete inhibition of Ca²⁺ oscillations and insulin release observed in the presence of the antagonist.

While the present study focused on six pharmacological compounds spanning two mechanistically distinct ion channel targets, namely K-ATP channel modulators and TRPM3 modulators, the platform is designed to accommodate a substantially broader compound panel. Extending the assay to include modulators of additional ion channel classes relevant to β-cell physiology, such as voltage-gated Ca²⁺ channel blockers like verapamil or nimodipine, represents a natural next step in demonstrating the full versatility of the platform for high-throughput drug discovery applications. Furthermore, validation of the platform using alternative insulinoma cell lines, such as INS-1 cells, would provide additional evidence of its generalizability across different β-cell models and further strengthen its utility as a broadly applicable screening tool.

In conclusion, we presented a straightforward method for the formation and functional characterization of up to 1824 β-cell spheroids on a single 96-well plate, and illustrate how the method can be used as a platform to identify small-molecules that modulate insulin secretion. The platform is easily amenable for further modifications, for instance, by co-culturing with other cell types found in the islets of Langerhans (α-cells, δ-cells, and PP cells) to enhance the physiological relevance of the assay, or by the use of patient-specific stem cell-derived organoids for personalized medicine. The uniform shape and positioning of individual 3D spheroids in the multi-well plate format may also find applications in other fields of life sciences where the use of 3D structures for screening purposes is of importance.

A current limitation of the platform is its optimization for the MIN6 insulinoma cell line, which, despite its widespread use in β-cell research, differs from primary pancreatic β-cells in several important respects, including glucose sensitivity and secretory capacity. Initial attempts to adapt the protocol to hiPS-derived β-cells were unsuccessful, likely due to insufficient cellular viability following transfer to the micropatterned plate format. Optimizing the platform for use with primary β-cells or hiPS-derived β-cells therefore represents a priority for future development and would substantially enhance the physiological relevance of the assay. An additional avenue for optimization is the introduction of perfusion-induced flow, which has been shown to improve viability, insulin secretory function, and morphological integrity of 3D β-cell tissues under dynamic culture conditions (Savitzky and Golay, 1964; Jun et al., 2019). Finally, while the current platform operates in a 96-well plate format providing up to 1824 individual spheroid measurements per plate, adaptation to higher density formats such as 384-well or 1536-well plates would represent a critical step toward full compatibility with industrial high-throughput screening pipelines, and is therefore identified as an important direction for future development of the platform.

Materials and methods

Key resources table.

Reagent type (species) or resource	Designation	Source or reference	Identifiers
Cell line (Mus musculus)	MIN6	Merck	SCC623; RRID:CVCL_0431
Commercial assay or kit	Mercodia Ultrasensitive Mouse Insulin ELISA kit	Mercodia AB	10-1249-01 RRID:AB_3101981
Chemical compound, drug	Cal-520 AM	AAT Bioquest, USA	21130 RRID:SCR_021130
Chemical compound, drug	Diazoxide	Sigma-Aldrich	D9035 RRID:SCR_025590
Chemical compound, drug	Glimepiride	Sigma-Aldrich	PHR1617 RRID:SCR_025556
Chemical compound, drug	Nateglinide	Sigma-Aldrich	N3538 RRID:SCR_008988
Chemical compound, drug	Tolbutamide	Sigma-Aldrich	T0891
Chemical compound, drug	Pregnenolone sulfate	Sigma-Aldrich	P162
Chemical compound, drug	Isosakuranetin	Sigma-Aldrich	PHL82569 RRID:SCR_008988
Software, algorithm	Fiji	National Institutes of Health, USA	ImageJ 1.52i; RRID:SCR_002285

Open in a new tab

Mold design and characterization

Micromachined 316L stainless steel molds were obtained from high-resolution CNC machining according to our computer-aided designs (CADs). The designs, consisting of molds slightly smaller than the footprint of a 96-well plate, contain 19 micro-pillar structures.

Hydrogel substrate preparation

The wells of the glass-bottom 96-well plate (P96-1.5H-N, Cellvis Inc) were silanized to ensure adequate bonding of the PEG to the glass. 20 µL of gel solution was placed within the well and 10 μL was added on top of the mold, which was then placed in the well. The gel solution was briefly exposed to 365 nm UV light to induce polymerization. The mold was unmolded, revealing a micropatterned hydrogel, and the well was immediately rinsed with distilled water at least three times before being submerged in water. Prior to cell seeding, the plate was sterilized using UV radiation for at least 20 min and sealed with a sterilized PCR film (Thermo Fisher, USA). The protocol and associated hydrogel plates were developed in-house at 4Dcell (SmartSphero Plates).

Cell culture

MIN6 cells were obtained from Merck Life Science (Hoeilaart, Belgium) – no further verification of cell identity was performed. The cells were grown in high-glucose (25 mM) Dulbecco’s modified Eagle’s medium (DMEM) (11965092, Thermo Fisher, USA) supplemented with 10% fetal bovine serum, 2% penicillin-streptomycin, 2% GlutaMAX, 1% sodium pyruvate, and 0.001% 2-mercaptoethanol. MIN6 cells seeded into the 96-well plates were harvested from a T-25 flask at passages 10–20 with an 80% confluency. The cells were tested monthly for mycoplasma contamination, yielding only negative results in the course of the project.

β-Cell spheroid formation

The PCR film was removed, and the storage solution was aspirated from each well. The wells were then washed with PBS twice and incubated at 37°C and 5% CO₂ with the gels submerged in 200 µL of cell culture medium per well. MIN6 cells were dissociated with TrypLE Express Enzyme (Thermo Fisher, USA). After cell counting, a cell suspension was prepared in cell culture medium with a cell density of 570,000 cells per mL. An aliquot of 100 µL of cell suspension was added to each well of the 96-well plate to achieve a density of 3000 cells per microwell. The plate was then incubated at 37°C and 5% CO₂ for 30 min for the cells to sediment into the microwells. An additional 150 µL of medium was added to each well before leaving the plate incubated at 37°C and 5% CO₂ overnight. The next day, the spheroids were evaluated for correct morphology and size uniformity using brightfield optical microscopy.

Compound plate preparation

Test compounds were prepared to the correct concentration and aliquoted into a clear 96-well plate with conical bottom (651101, Greiner Bio-One International GmbH, Austria) in accordance with the protocol for the µCELL FDSS fluorescence plate imager. All test compounds were of analytical standard or certified reference material purity (>96%) and obtained from Sigma-Aldrich (St. Louis, MO, USA). Stock solutions were prepared in dimethyl sulfoxide and stored at –20°C. Appropriate volumes of the stock solutions were added to the same KREBS solution used for the baseline measurement (NaCl 150 mM, KCl 6 mM, CaCl₂ 2 mM, MgCl₂ 1.5 mM, HEPES 10 mM, glucose 3 mM, pH 7.4) to achieve the desired final concentrations in the compound plate.

Ca²⁺ imaging

Medium was removed from the wells 18–24 hr after seeding and 200 µL of a 10 μM Cal-520 AM (AAT Bioquest, USA) solution with 0.04% Pluronic F-127 (AAT Bioquest, USA) was added to each well and the plate was incubated at 37°C and 5% CO₂ for 1 hr. The wells were then washed with KREBS solution (NaCl 150 mM, KCl 6 mM, CaCl₂ 2 mM, MgCl₂ 1.5 mM, HEPES 10 mM, glucose 3 mM, pH 7.4) twice and incubated in 200 µL of the same KREBS solution for 30 min at room temperature to allow esterases to cleave the lipophilic blocking groups. The KREBS solution was removed and a final volume of 120 µL 3 mM glucose KREBS solution was placed in each well. The plate was then placed in a functional drug screening system (µCELL FDSS, Hamamatsu Photonics) fluorescence plate imager with integrated dispensing head. The automated instrument dispensed 15 µL per well of a compound plate into the assay plate according to a predetermined protocol and detects fluorescence across the plate simultaneously at a resolution sufficient to yield 4 pixels per spheroid. Ca²⁺ flux signal was captured continuously, at a sampling rate of 1 Hz, by a cooled CCD camera during the entire protocol. The excitation wavelength was set to 470 nm and the emitted light passed through a 530 nm bandpass filter. Prior to the addition of the solutions from the first compound plate, a baseline was recorded for a minimum of 5 min. All measurements were performed at 37°C. The acquired image sequences of the fluorescence signal were exported in TIFF format for further analysis.

Insulin detection

After 1 hr of exposure to increased levels of extracellular glucose and/or compounds, between 15 µL and 100 µL of supernatant was aspirated from the assay plate and deposited into a separate 96-well plate using the dispensing head of the µCELL FDSS. This plate was sealed with PCR film and stored at –20°C until analysis. Insulin levels in the supernatant were quantified using the Mercodia Ultrasensitive Mouse Insulin ELISA kit (Mercodia AB, Sweden), according to the manufacturer’s protocol. Fluorescence values of the plate were recorded using a Flexstation 3 Microplate Reader (Molecular Devices, USA), and insulin concentrations were calculated via cubic spline regression based on the fluorescence of the calibrator solutions.

Data analysis

Regions of interest (ROIs) were selected to contain the fluorescence signal of one spheroid per ROI using Fiji ImageJ 1.52i. The resulting signal for each individual spheroid was then analyzed using a Python script (v3.10.11). The background signal was defined as the average of the pixels directly bordering the ROI and was subtracted from the ROI signal. The resulting signal was then smoothed by applying a Savitzky-Golay filter with a polynomial order of 10 and a window length of 60 (Sankar et al., 2011). Next, a multi-polynomial fit was applied to approximate the baseline of the signal using the IModPoly method from the BaseLineRemoval Python-module, utilizing a polynomial degree of 10 (Haque, 2022). The baseline was subtracted from the original fluorescence signal, and peaks were identified with the find_peaks function from the SciPy.Signal data processing Python library (Virtanen et al., 2020). Minimum peak height was set to a multiple of the standard deviation of the signal during basal levels of extracellular glucose (specified in-text), minimum width was set at 10 s, and a minimal horizontal distance between peak centers of 60 s. The normalized AuP was defined as the sum of the fluorescent signal for each timepoint after baseline removal within the given time window, normalized over the amplitude of the control stimulus (50 mM KCl).

Statistical analysis

Analysis was performed using Origin 9.0 (OriginLab Corporation) and GraphPad Prism 10.0 (GraphPad Software). All numerical results are expressed as mean ± standard error of the mean (SEM). EC₅₀ and IC₅₀ values were determined using a nonlinear regression based on the Hill equation. Assay quality was assessed by calculating the Z-factor using the Ca²⁺ oscillation frequency at 3 mM glucose as the negative control and at 20 mM glucose supplemented with 50 µM PS as the positive control.

Acknowledgements

We kindly thank M Benoit and A Janssens for their technical support and expertise at KU Leuven. This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Sklodowska-Curie grant agreement No. 955643. This work was further supported by grants from the Research Foundation – Flanders (FWO; G0B9520N and G0B7620N to TV), the Queen Elisabeth Medical Foundation for Neurosciences (to TV), and the VIB to (TV). Research infrastructure was funded by the KU Leuven Research Council (AKUL/19/34).

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

Thomas Voets, Email: thomas.voets@kuleuven.be.

Leon D Islas, Universidad Nacional Autónoma de México, Mexico.

Merritt Maduke, Stanford University, United States.

Funding Information

This paper was supported by the following grants:

Horizon 2020 Framework Programme 10.3030/955643 to Patricia Davidson, Thomas Voets.
Fonds Wetenschappelijk Onderzoek G0B9520N to Thomas Voets.
Fonds Wetenschappelijk Onderzoek G0B7620N to Thomas Voets.
KU Leuven AKUL/19/34 to Thomas Voets.
Vlaams Instituut voor Biotechnologie - to Thomas Voets.

Additional information

Competing interests

Affiliated with 4Dcell, Montreuil; no other competing interests to declare.

Inventor on patents related to the role of TRPM3 in pain; no other competing interests to declare.

Author contributions

Conceptualization, Formal analysis, Investigation, Methodology, Writing – original draft.

Resources, Supervision, Funding acquisition, Writing – review and editing.

Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Investigation, Visualization, Methodology, Writing – original draft, Project administration.

Additional files

MDAR checklist

elife-110428-mdarchecklist1.pdf^{(186.1KB, pdf)}

Data availability

All source data needed to generate the figures has been included.

References

American Diabetes Association Diagnosis and classification of diabetes mellitus. Diabetes Care. 2010;33:S62–S69. doi: 10.2337/dc10-S062. [DOI] [PMC free article] [PubMed] [Google Scholar]
Amin J, Ramachandran K, Williams SJ, Lee A, Novikova L, Stehno-Bittel L. A simple, reliable method for high-throughput screening for diabetes drugs using 3D β-cell spheroids. Journal of Pharmacological and Toxicological Methods. 2016;82:83–89. doi: 10.1016/j.vascn.2016.08.005. [DOI] [PubMed] [Google Scholar]
Antunes CM, Salgado AP, Rosário LM, Santos RM. Differential patterns of glucose-induced electrical activity and intracellular calcium responses in single mouse and rat pancreatic islets. Diabetes. 2000;49:2028–2038. doi: 10.2337/diabetes.49.12.2028. [DOI] [PubMed] [Google Scholar]
Becker A, Mannebach S, Mathar I, Weissgerber P, Freichel M, Loodin AP, Fecher-Trost C, Belkacemi A, Beck A, Philipp SE. Control of insulin release by Transient Receptor Potential Melastatin 3 (TRPM3) ion channels. Cellular Physiology and Biochemistry. 2020;54:1115–1131. doi: 10.33594/000000304. [DOI] [PubMed] [Google Scholar]
Białkowska K, Komorowski P, Bryszewska M, Miłowska K. Spheroids as a type of three-dimensional cell cultures-examples of methods of preparation and the most important application. International Journal of Molecular Sciences. 2020;21:6225. doi: 10.3390/ijms21176225. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bommer C, Heesemann E, Sagalova V, Manne-Goehler J, Atun R, Bärnighausen T, Vollmer S. The global economic burden of diabetes in adults aged 20–79 years: a cost-of-illness study. The Lancet Diabetes & Endocrinology. 2017;5:423–430. doi: 10.1016/S2213-8587(17)30097-9. [DOI] [PubMed] [Google Scholar]
Broichhagen J, Schönberger M, Cork SC, Frank JA, Marchetti P, Bugliani M, Shapiro AMJ, Trapp S, Rutter GA, Hodson DJ, Trauner D. Optical control of insulin release using a photoswitchable sulfonylurea. Nature Communications. 2014;5:5116. doi: 10.1038/ncomms6116. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cho NH, Shaw JE, Karuranga S, Huang Y, da Rocha Fernandes JD, Ohlrogge AW, Malanda B. IDF diabetes atlas: global estimates of diabetes prevalence for 2017 and projections for 2045. Diabetes Research and Clinical Practice. 2018;138:271–281. doi: 10.1016/j.diabres.2018.02.023. [DOI] [PubMed] [Google Scholar]
Cole JB, Florez JC. Genetics of diabetes mellitus and diabetes complications. Nature Reviews. Nephrology. 2020;16:377–390. doi: 10.1038/s41581-020-0278-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
DeFronzo RA, Tripathy D. Skeletal muscle insulin resistance is the primary defect in type 2 diabetes. Diabetes Care. 2009;32 Suppl 2:S157–S163. doi: 10.2337/dc09-S302. [DOI] [PMC free article] [PubMed] [Google Scholar]
Diane A, Mohammed LI, Al-Siddiqi HH. Islets in the body are never flat: transitioning from two-dimensional (2D) monolayer culture to three-dimensional (3D) spheroid for better efficiency in the generation of functional hPSC-derived pancreatic β cells in vitro. Cell Communication and Signaling. 2023;21:151. doi: 10.1186/s12964-023-01171-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Essaouiba A, Jellali R, Shinohara M, Scheidecker B, Legallais C, Sakai Y, Leclerc E. Analysis of the behavior of 2D monolayers and 3D spheroid human pancreatic beta cells derived from induced pluripotent stem cells in a microfluidic environment. Journal of Biotechnology. 2021;330:45–56. doi: 10.1016/j.jbiotec.2021.02.009. [DOI] [PubMed] [Google Scholar]
Gilon P, Ravier MA, Jonas JC, Henquin JC. Control mechanisms of the oscillations of insulin secretion in vitro and in vivo. Diabetes. 2002;51 Suppl 1:S144–S151. doi: 10.2337/diabetes.51.2007.s144. [DOI] [PubMed] [Google Scholar]
Haque MA. Feature Engineering and Selection for Explainable Models: A Second Course for Data Scientists. LULU Press, Inc; 2022. [Google Scholar]
Islam MS. Calcium signaling in the islets. Advances in Experimental Medicine and Biology. 2010;654:235–259. doi: 10.1007/978-90-481-3271-3_11. [DOI] [PubMed] [Google Scholar]
Jun Y, Lee J, Choi S, Yang JH, Sander M, Chung S, Lee S-H. In vivo-mimicking microfluidic perfusion culture of pancreatic islet spheroids. Science Advances. 2019;5:eaax4520. doi: 10.1126/sciadv.aax4520. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lee SH, Hong S, Song J, Cho B, Han EJ, Kondapavulur S, Kim D, Lee LP. Microphysiological analysis platform of pancreatic Islet β-cell spheroids. Advanced Healthcare Materials. 2018;7:1701111. doi: 10.1002/adhm.201701111. [DOI] [PubMed] [Google Scholar]
Mohandas S, Gayatri V, Kumaran K, Gopinath V, Paulmurugan R, Ramkumar KM. New frontiers in three-dimensional culture platforms to improve diabetes research. Pharmaceutics. 2023;15:725. doi: 10.3390/pharmaceutics15030725. [DOI] [PMC free article] [PubMed] [Google Scholar]
Nielsen MS, Axelsen LN, Sorgen PL, Verma V, Delmar M, Holstein-Rathlou NH. Gap junctions. Comprehensive Physiology. 2012;2:1981–2035. doi: 10.1002/cphy.c110051. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rupnik M. The physiology of rodent beta-cells in pancreas slices. Acta Physiologica. 2009;195:123–138. doi: 10.1111/j.1748-1716.2008.01927.x. [DOI] [PubMed] [Google Scholar]
Saeedi P, Salpea P, Karuranga S, Petersohn I, Malanda B, Gregg EW, Unwin N, Wild SH, Williams R. Mortality attributable to diabetes in 20-79 years old adults, 2019 estimates: results from the international diabetes federation diabetes Atlas, 9th edition. Diabetes Research and Clinical Practice. 2020;162:108086. doi: 10.1016/j.diabres.2020.108086. [DOI] [PubMed] [Google Scholar]
Sankar KS, Green BJ, Crocker AR, Verity JE, Altamentova SM, Rocheleau JV. Culturing pancreatic islets in microfluidic flow enhances morphology of the associated endothelial cells. PLOS ONE. 2011;6:e24904. doi: 10.1371/journal.pone.0024904. [DOI] [PMC free article] [PubMed] [Google Scholar]
Savitzky A, Golay MJE. Smoothing and differentiation of data by simplified least squares procedures. Analytical Chemistry. 1964;36:1627–1639. doi: 10.1021/ac60214a047. [DOI] [PubMed] [Google Scholar]
Shigeto M, Katsura M, Matsuda M, Ohkuma S, Kaku K. First phase of glucose-stimulated insulin secretion from MIN 6 cells does not always require extracellular calcium influx. Journal of Pharmacological Sciences. 2006;101:293–302. doi: 10.1254/jphs.fp0060057. [DOI] [PubMed] [Google Scholar]
Shigeto M, Katsura M, Matsuda M, Ohkuma S, Kaku K. Nateglinide and mitiglinide, but not sulfonylureas, induce insulin secretion through a mechanism mediated by calcium release from endoplasmic reticulum. The Journal of Pharmacology and Experimental Therapeutics. 2007;322:1–7. doi: 10.1124/jpet.107.120592. [DOI] [PubMed] [Google Scholar]
Sun D, Gao W, Hu H, Zhou S. Why 90% of clinical drug development fails and how to improve it? Acta Pharmaceutica Sinica B. 2022;12:3049–3062. doi: 10.1016/j.apsb.2022.02.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
Vhora N, Naskar U, Hiray A, Kate AS, Jain A. Recent advances in In-Vitro assays for type 2 diabetes mellitus: an overview. Review of Diabetic Studies. 2020;16:13–23. doi: 10.1900/RDS.2020.16.13. [DOI] [PMC free article] [PubMed] [Google Scholar]
Virtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D, Burovski E, Peterson P, Weckesser W, Bright J, van der Walt SJ, Brett M, Wilson J, Millman KJ, Mayorov N, Nelson ARJ, Jones E, Kern R, Larson E, Carey CJ, Polat İ, Feng Y, Moore EW, VanderPlas J, Laxalde D, Perktold J, Cimrman R, Henriksen I, Quintero EA, Harris CR, Archibald AM, Ribeiro AH, Pedregosa F, van Mulbregt P, SciPy 1.0 Contributors SciPy 1.0: fundamental algorithms for scientific computing in Python. Nature Methods. 2020;17:261–272. doi: 10.1038/s41592-019-0686-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wagner TFJ, Loch S, Lambert S, Straub I, Mannebach S, Mathar I, Düfer M, Lis A, Flockerzi V, Philipp SE, Oberwinkler J. Transient receptor potential M3 channels are ionotropic steroid receptors in pancreatic beta cells. Nature Cell Biology. 2008;10:1421–1430. doi: 10.1038/ncb1801. [DOI] [PubMed] [Google Scholar]
Wassmer C-H, Bellofatto K, Perez L, Lavallard V, Cottet-Dumoulin D, Ljubicic S, Parnaud G, Bosco D, Berishvili E, Lebreton F. Engineering of primary pancreatic islet cell spheroids for three-dimensional culture or transplantation: a methodological comparative study. Cell Transplantation. 2020;29:963689720937292. doi: 10.1177/0963689720937292. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wikstrom JD, Sereda SB, Stiles L, Elorza A, Allister EM, Neilson A, Ferrick DA, Wheeler MB, Shirihai OS. A novel high-throughput assay for islet respiration reveals uncoupling of rodent and human islets. PLOS ONE. 2012;7:e33023. doi: 10.1371/journal.pone.0033023. [DOI] [PMC free article] [PubMed] [Google Scholar]

eLife. doi: 10.7554/eLife.110428.3.sa0

eLife Assessment

Leon D Islas ¹

In this valuable study, Robben et al. describe a 3D beta-cell spheroid culture platform, that allows high-throughput monitoring of cytoplasmic calcium concentrations and insulin secretion. The authors demonstrate measurements of calcium signals comparable to those recorded in primary pancreatic islets. The authors validate the method by culturing MIN6 cells in a 3D-culture system, and show solid evidence of its utility by recording calcium signals in a high-throughput format and characterizing these calcium signals using pharmacological tools. This methodology demonstrates the utility of the 3D beta-cell spheroids for screening pharmacological modulators of pancreatic beta-cell function.

eLife. doi: 10.7554/eLife.110428.3.sa1

Reviewer #1 (Public review):

Anonymous

[Editors' note: this version has been assessed by the Reviewing Editor without further input from the original reviewers and the authors have satisfactorily answered the previous reviewer's comments.]

Summary:

They use cultures of insulinoma MIN6 cells that form spheroids in a micro-patterned PEG-hydrogel to measure Ca2+ oscillations in multiple cells simultaneously.

Strengths:

They demonstrate that insulinoma spheroids are formed in multi-well plates and that Ca2+ imaging can be performed on them.

eLife. doi: 10.7554/eLife.110428.3.sa2

Reviewer #2 (Public review):

Anonymous

Summary:

The study by Robben et al., show 3D beta-cell spheroid platform, a valuable tool allowing high-throughput monitoring of cytoplasmic Ca concentrations and insulin secretion, with Ca signals comparable to those recorded in primary islets. The authors demonstrate a solid method to culturing MIN6 cells in a 3D culture system, recording Ca signals in a high-throughput format and characterizing these Ca signals using pharmacological tools, including TRPM3 channel and K-ATP channel modulators. This highlights the utility of the 3D beta-cell spheroid for screening new ion channel modulators in beta-cells of the pancreas.

Strengths:

- The study shows that the MIN-6-based 3D beta-cell model is better to study Ca-signaling and insulin secretion compared to 2D culture of single MIN-6 cells.

- The method allows imaging of Ca signaling in many spheroids in parallel followed by collecting medium to measure insulin release and correlate both effects.

- The authors demonstrate that this system is suitable for screening new pharmacological modulators and used as an agonist of the ATP-sensitive potassium channel (diazoxide) and the agonist and antagonist of the TRPM3 channel.

eLife. doi: 10.7554/eLife.110428.3.sa3

Reviewer #3 (Public review):

Anonymous

Summary:

The primary objective of this study is to develop high-throughput screening assays utilizing homogeneous 3D cell cultures that more accurately replicate the intricate architecture and cellular communication found in tissues. The authors have chosen pancreatic islet β-cells as a model system to evaluate agents that modulate insulin release, which is particularly relevant given the increasing prevalence of diabetes mellitus-a significant global health concern. Moreover, the incorporation of human-based 3D spheroids, organoids, or organ-on-chip technologies into drug discovery protocols is essential for enhancing clinical translation, as candidate compounds identified using animal models have often demonstrated limited success in clinical settings.

Strengths:

This study was thoughtfully planned and skillfully carried out. The use of micropatterned hydrogels to observe 19 spheroids at once is an ingenious aspect, which has been effectively validated with Ca microfluorography. Overall, I found this investigation to be exceptionally well-executed and free from notable flaws, as the results clearly back up the conclusions. Additionally, the developed method achieved the proposed aims, providing a high-throughput format with 3D cultures. I believe this study deserves publication.

eLife. 2026 Oct 5;15:RP110428. doi: 10.7554/eLife.110428.3.sa4

Author response

Stijn Robben ¹, Patricia Davidson ², Rita S Rodrigues Ribeiro ³, Thomas Voets ⁴

The following is the authors’ response to the original reviews.

Reviewer #1 (Public review):

Summary:

They use cultures of insulinoma MIN6 cells that form spheroids in a micro-patterned PEG-hydrogel to measure Ca ²⁺ oscillations in multiple cells simultaneously.

Strengths:

They demonstrate that insulinoma spheroids are formed in multi-well plates and that Ca ²⁺ imaging can be performed on them.

Weaknesses:

The type of equipment and multi-wells used for the experiments are very specialized to be used as a common tool. Insulinoma cells are tumoral cell lines that divide, unlike primary beta cells. Pancreatic islets are very different from this preparation, as they are highly heterogeneous, whereas these cells all respond equally. It would be good to see the same technique applied to primary cells.

MIN6 cells do not respond to glucose and other secretagogues in the same way as primary cells, and they cycle, depending on the phase of the cycle to which they are exposed.

The authors should report the number of cells per spheroid and the number of cells that are alive and dead.

I would like to examine the effects of calcium channel blockers on calcium transients, and the use of pregnenolone is already described in the literature, but remains less well known.

MIN6 cells secrete much insulin, because detecting the hormone in ELISAs requires too many primary cells. The authors should discuss the model in greater detail and compare it with primary beta cells. Also, they take 3 mM glucose as the basal concentration, which is low.

We thank the reviewer for their valuable comments, which have helped us to significantly enhance the quality of our manuscript. We have carefully considered these comments and have revised the manuscript accordingly. A point-by-point rebuttal is provided below.

Reviewer #1 (Recommendations for the authors):

(1) The manuscript contains numerous typos, including the combination of numbers and units without a space.

All spacing inconsistencies between numerical values and unit symbols (e.g. mM, μM, µL, and Hz) have been corrected throughout the text and images. In addition, the following typographical and grammatical errors have been addressed:

- Abstract: "the frequency of Ca ²⁺ oscillations correlate" → correlates

- Figure 2A caption: "200uL" → 200 µL

- Figure 5A caption: "glimepirde" → glimepiride

- Figure 5 title: "KATP-antagonists induces" → induce

- Figure 7D caption: "concentrations" → concentration

- Figure 7 caption: "Tukey’s post-hoc testm" → Tukey’s post-hoc test

- Results section header: "increases insulin secretions" → insulin secretion

- Discussion: "we obtained an EC50 values 7.4 ± 0.4 mM" → "we obtained EC50 values of 7.4 ± 0.4 mM"

- Discussion: "a EC50 value" → an EC50 value

- Discussion: "concentration dependence profiles that matches" → match

- Discussion: "PS-induced activation TRPM3" → activation of TRPM3

- Conclusion: "found the Islets of Langerhans" → found in the Islets of Langerhans

- Acknowledgements: "grant agreement No. 955643" → agreement No. 955643

(2) I suggest showing the experiments in primary beta cells because of the many differences from insulinoma cells, as the most important result is a better culture technique for calcium imaging.

We thank the reviewer for this valuable suggestion. We did indeed attempt to perform comparable experiments using the Cellartis hiPS Beta Cell Media Kit (Takara, cat. no. Y10108) as a more physiologically relevant cell model. To minimize cellular stress during the transition, we adjusted our protocol by seeding the differentiated cells into the micropatterned plates 24 hours prior to imaging, thereby maintaining the recommended culture conditions for as long as feasible. However, during calcium imaging, the cells failed to respond to either the elevated glucose stimulus or the positive control, suggesting that the cells did not survive the transfer to our plate format with sufficient viability to mount a functional response.

We acknowledge that the use of primary beta cells or hiPS-derived beta cells would strengthen the physiological relevance of the platform. Nevertheless, we would like to emphasize that the primary objective of this study is to demonstrate the feasibility of high-throughput calcium imaging in 3D cell culture using our optimized protocol: a proof-of-concept that is, by design, independent of the specific cell model employed. Adapting the protocol to accommodate more sensitive or terminally differentiated cell types is a meaningful avenue for future work, but falls outside the scope of the current manuscript. We have added a brief note to the Discussion section to explicitly acknowledge this limitation and to identify hiPS-derived beta cell compatibility as a priority for subsequent optimization.

(3) These kinds of cultures in three dimensions are interesting, but it has been shown that it is even better to have the liquid flow, simulating blood flow; this can at least be discussed.

We thank the reviewer for this insightful comment. We agree that the introduction of perfusion-based flow represents a meaningful improvement over static 3D culture systems. We have incorporated this point into the Discussion section, where we now explicitly acknowledge the potential benefits of dynamic culture conditions, including improved viability, insulin secretory function, and morphological integrity of 3D β-cell tissues, and identify the integration of perfusion-induced flow as a potentially meaningful path to explore for future development of the platform.

Reviewer #2 (Public review):

Summary:

The study by Robben et al., show 3D beta-cell spheroid platform, a valuable tool allowing high-throughput monitoring of cytoplasmic Ca concentrations and insulin secretion, with Ca signals comparable to those recorded in primary islets. The authors demonstrate a solid method to culturing MIN6 cells in a 3D culture system, recording Ca signals in a high-throughput format and characterizing these Ca signals using pharmacological tools, including TRPM3 channel and K-ATP channel modulators. This highlights the utility of the 3D beta-cell spheroid for screening new ion channel modulators in beta-cells of the pancreas.

Strengths:

- The study shows that the MIN-6-based 3D beta-cell model is better to study Ca-signaling and insulin secretion compared to 2D culture of single MIN-6 cells.

- The method allows imaging of Ca signaling in many spheroids in parallel followed by collecting medium to measure insulin release and correlate both effects.

- The authors demonstrate that this system is suitable for screening new pharmacological modulators and used as an agonist of the ATP-sensitive potassium channel (diazoxide) and the agonist and antagonist of the TRPM3 channel.

Weaknesses:

- The study is based on only one cell line, the MIN6 insulinoma cells, which may not fully mimic the pancreatic beta-cells within the islet.

- The authors show only spheroids cultured overnight. A long-term culture is missing to assess beta-cell viability long term function.

- The authors tested their platform using only two compounds. Testing a larger compound library is necessary to make a clear conclusion about the suitability of the platform for high-throughput screening.

Reviewer #2 (Recommendations for the authors):

Major Points

(1) In this study, only 2 pharmacological compounds (for TRPM3, and K channel) were tested. Testing a larger compound library would be necessary to fully demonstrate its suitability for high-throughput screening applications. If this is not possible at the moment, including data on additional pharmacological compounds, e.g., modulators of voltage-gated Ca channels, which are key regulators of Ca signaling in beta cells of the pancreas would strengthen the study. (e.g., use voltage gated Ca channel blocker such as verapamil or nimodipine).

We thank the reviewer for this constructive suggestion. We would like to clarify that our pharmacological characterization was not limited to two compounds. In total, six compounds spanning two distinct ion channel targets were evaluated: the K-ATP channel modulators diazoxide, glimepiride, tolbutamide and nateglinide, and the TRPM3 modulators pregnenolone sulphate and isosakuranetin. This panel includes both agonists and antagonists across two mechanistically distinct targets, and we believe this is sufficient to demonstrate the platform's suitability for high-throughput compound screening in the context of a proof-of-concept study.

We nonetheless agree with the reviewer that extending the compound panel to include modulators of additional ion channel classes, such as voltage-gated Ca ²⁺ channel blockers like verapamil or nimodipine, would further demonstrate the versatility of the platform. We have added a statement to the Discussion explicitly identifying this as a valuable direction for future work.

(2) Testing another beta-cell line (e.g., INS-1 cells) would strengthen the manuscript.

We thank the reviewer for this suggestion. We agree that validating the platform using an additional β-cell line, such as INS-1 cells, would further broaden the applicability of the approach. We did indeed attempt experiments with INS-1 cells; however, the results were inconclusive due to cell quality issues at the time of testing, and we were unable to generate reliable data suitable for inclusion in the manuscript.

We would also like to emphasize that the primary aim of this study was to demonstrate the methodology of the high-throughput screening platform, rather than to provide a comprehensive cross-cell-line validation. In this context, the use of the well-established MIN6 β-cell line is sufficient to serve as a proof-of-principle demonstration of the platform's capabilities.

Nonetheless, we consider a systematic evaluation of INS-1 cells on this platform as an important and natural next step and have included this explicitly as a future perspective in the Discussion.

(3) I find the presentation of the results and analysis of Ca oscillation frequency and area under the curve excellent. Could you please provide more details on the analysis method used to quantify the frequency of glucose-induced Ca oscillation. If a custom script was used, sharing this information with the scientific community would be great.

We thank the reviewer for this positive feedback. We confirm that the Ca ²⁺ oscillation analysis was performed using a custom Python script (v3.10.11), the key steps of which are described in the Data Analysis section of the experimental procedures.

In line with our commitment to open and reproducible science, the script will be made publicly available upon acceptance of the manuscript, allowing the broader scientific community to apply, adapt, and build upon the analysis pipeline.

(4) Please move the Supplementary Figure to the main Figure 3. This will allow a direct comparison between Ca signals in spheroids and in single cell (2D cultures) under identical conditions.

We thank the reviewer for this suggestion. We have partially incorporated the supplementary figure into the main manuscript. The mean Ca ²⁺ response of 2D-cultured MIN6 cells at 20 mM glucose has been added as Figure 3D, enabling direct visual comparison with the spheroid data under identical stimulation conditions. The remaining panels of the supplementary figure — showing representative Ca ²⁺ traces across multiple glucose concentrations and the corresponding dose-response curves for peak frequency and area under the peaks in 2D monolayers — have been retained in the supplementary information, as their inclusion in the main figure would substantially increase its complexity. The figure legend has been updated accordingly.

(5) A direct comparison of insulin secretion between 3D cultured spheroids and 2D cultures should also be shown.

We thank the reviewer for this suggestion. We attempted to include a direct comparison of insulin secretion between 3D spheroids and 2D MIN6 monolayers; however, the 2D measurements proved unreliable for quantitative comparison. Insulin values in the 2D condition consistently exceeded the upper detection limit of the ELISA, and inter-well variability was too high to draw meaningful conclusions. We therefore chose not to include this comparison and instead present the Ca ²⁺ imaging data in Figure 3D as a functional readout enabling direct comparison between the two culture formats under identical stimulation conditions.

(6) The authors should further discuss the remaining effects of pregnenolone sulphate on insulin secretion.

We thank the reviewer for this comment. We would like to clarify that in our experimental setup, pregnenolone sulphate and isosakuranetin were applied simultaneously rather than sequentially. As a result, the incomplete inhibition of PS-induced Ca ²⁺ oscillations and insulin secretion observed in the presence of isosakuranetin may in part reflect a kinetic offset between the two compounds, whereby PS-induced TRPM3 activation and downstream signalling may have been initiated prior to the establishment of effective TRPM3 blockade by isosakuranetin. In addition, as noted in the Discussion, TRPM3 may not be the only molecular target of PS in these spheroids, and alternative signalling pathways may contribute to the residual insulin secretion observed in the presence of the antagonist. We have added a brief clarification to the Discussion to explicitly acknowledge the potential influence of this kinetic limitation on the interpretation of these results.

(7) Is it possible to collect 3D cultured spheroids after each experiment to measure for example intracellular insulin content or protein levels by Western blot.

Physical recovery of spheroids from the PEG hydrogel plates for downstream biochemical analysis, such as intracellular insulin content measurements or Western blot, would indeed be a meaningful addition to the platform's capabilities. We would like to note that the firm attachment of spheroids to the glass substrate, while essential for maintaining spheroid positioning during the extensive washing and liquid handling steps, does present a practical challenge for post-experimental recovery. Although we have successfully extracted spheroids of other cell types from comparable plate formats, reliable recovery of the MIN6 β-cell spheroids without compromising their structural integrity has not yet been achieved. We therefore identify the optimization of spheroid recovery as a valuable direction for future development of the platform.

(8) Ca signals in response to the application of glucose appears more robust in spheroids compared to single MIN6 cells. What are the possible mechanisms underlying this difference. Whole RNA-seq experiments would be one approach to identify differentially expressed genes in 2D versus 3D culture (this is maybe a whole project by itself). An alternative is to look by RT-qPCR analysis for key β-cell markers and genes encoding ion channel and ion channel subunits.

We thank the reviewer for this thoughtful comment. The more robust Ca ²⁺ signals observed in 3D spheroids compared to 2D monolayer cultures likely reflect several interconnected factors. First, and importantly, it should be noted that Ca ²⁺ measurements in 2D monolayer cultures typically represent an averaged signal across a large population of cells, which tends to obscure individual oscillatory events and reduce the apparent amplitude and regularity of Ca ²⁺ responses. In contrast, our 3D spheroid platform enables Ca ²⁺ measurements at the level of individual spheroids, allowing discrete oscillatory peaks to be resolved with much greater fidelity. Beyond this methodological distinction, the 3D architecture also promotes enhanced cell-to-cell communication, better recapitulation of in vivo β-cell coupling, and a more physiologically relevant microenvironment, all of which are likely to contribute to the improved oscillatory Ca ²⁺ dynamics observed.

We agree with the reviewer that elucidating the transcriptional underpinnings of these differences, through whole RNA-seq or targeted RT-qPCR analysis of key β-cell markers and genes encoding ion channel subunits, would be highly informative and represents an elegant approach to understanding the molecular basis of the observed functional improvements. As the reviewer rightly acknowledges, however, such experiments constitute a substantial research effort in their own right. We have added a statement to the Discussion identifying this as a valuable direction for future investigation, alongside the other platform development priorities already outlined.

(9) A more detailed discussion of the limitations of the platform and potential strategies to further improve this system would strengthen the manuscript.

We thank the reviewer for this constructive suggestion. In response, we have expanded the Discussion to provide a more comprehensive overview of the current limitations of the platform and the strategies we envision for future development. Specifically, the revised Discussion now addresses the following points:

First, the platform is currently optimized for the MIN6 insulinoma cell line, which differs from primary pancreatic beta cells in several important respects, including glucose sensitivity and secretory capacity. Initial attempts to adapt the protocol to hiPS-derived beta cells were unsuccessful, likely due to insufficient cellular viability following transfer to the micropatterned plate format. Optimizing the platform for use with primary beta cells or hiPS-derived beta cells is therefore identified as a priority for future development.

Second, while the current study demonstrates proof-of-concept pharmacological characterization using six compounds across two mechanistically distinct ion channel targets, extending the compound panel to include modulators of additional ion channel classes, such as voltage-gated Ca ²⁺ channel blockers, as well as validation using alternative insulinoma cell lines such as INS-1, would further demonstrate the versatility and generalizability of the platform.

Third, the introduction of perfusion-induced flow, which has been shown to improve viability, insulin secretory function, and morphological integrity of 3D beta-cell tissues under dynamic culture conditions, is identified as an additional avenue for optimization.

Fourth, while the current platform operates in a 96-well plate format, which already provides a substantial throughput of up to 1824 individual spheroid measurements per plate, adaptation to higher density plate formats such as 384-well or 1536-well plates would be a necessary step towards true high-throughput screening compatible with industrial drug discovery pipelines. Miniaturization of the hydrogel design and adaptation of the molding procedure to accommodate these formats therefore represents an important direction for future development.

Minor points:

(10) Please clarify the glucose concentration at which the MIN6 cells the cultured 3D beta cell spheroids were maintained overnight prior to the experiments.

Prior to the experiments, both 2D MIN6 cells and 3D MIN6 β-cell spheroids were maintained overnight in standard high-glucose DMEM (25 mM glucose), consistent with widely used MIN6 culture protocols. This information has been added to the Methods section.

(11) In Figure 3A, is the presented Ca trace derived from one single spheroid? Showing representative Ca traces (e.g., 5 i traces per condition) would show the reproducibility of the recordings.

We thank the reviewer for this suggestion. The trace shown in Figure 3A is indeed derived from a single representative spheroid. To address the concern regarding reproducibility, we refer the reviewer to the updated Figure 3C, which now displays the individual Ca ²⁺ response traces of all spheroids within a single well in response to 20 mM glucose stimulation. Grey lines represent individual spheroids, while the black line denotes the mean response. This panel illustrates not only the reproducibility of the oscillatory response across the imaged population but also highlights an important consequence of inter-spheroid variability: because individual spheroids oscillate asynchronously, their peaks cancel out when averaged, resulting in a mean trace that appears relatively flat and lacks the oscillatory features visible in individual recordings. We note that this same cancellation effect likely underlies the comparably flat mean response observed for 2D-cultured MIN6 cells in Figure 3D. Additionally, the difference in the initial response profile between 2D and 3D cultures may in part reflect the geometry of the hydrogel microenvironment. In 2D cultures, the glucose stimulus equilibrates rapidly and near-uniformly across the culture plane, potentially driving a more synchronized initial response and the early peak visible in Figure 3D. In contrast, the PEG hydrogel surrounding the 3D spheroids may act as a diffusion barrier, causing the stimulus to reach individual spheroids with variable delay and thereby further desynchronizing response onsets across the population. The figure legend has been updated accordingly.

(12) In Figure 3A, the potassium application bar looks a bit shifted. Please check and correct if necessary.

We thank the reviewer for carefully examining the figure. The apparent shift in the potassium application bar is not an error. In these experiments, glucose was added after an initial 10-minute baseline period, and the observed delay in the Ca ²⁺ response reflects two contributing factors. First, the glucose solution was pipetted at the top of the well, and diffusion to the level of the spheroids introduces a short lag before the stimulus reaches the cells. Second, the spheroid shown in Figure 3A was located at the outer edge of the well, where mixing is slower and the stimulus arrives with additional delay compared to centrally positioned spheroids. Together, these factors account for the offset between the start of the glucose application bar and the onset of the visible Ca ²⁺ response.

(13) Is the system also compatible with the ratiometric Ca imaging dye Fura-2?

The platform is indeed compatible with ratiometric Ca ²⁺ imaging using Fura-2. The glass-bottom plate format is a prerequisite for Fura-2 imaging due to the requirement for UV excitation at 340/380 nm, and the transparency of the PEG-based hydrogel ensures that the optical properties of the platform are fully compatible with this approach. Furthermore, the µCELL FDSS fluorescence plate imager used in this study supports dual-excitation ratiometric imaging, making it instrumentally compatible with Fura-2 without any additional hardware modifications.

(14) In Figure 4 legend: 100 µm diazoxide should be corrected to 100 µM diazoxide.

This has been addressed in the revised manuscript

(15) Please comment on the cost of spheroid generation compared with conventional 2D MIN6 cultures.

We thank the reviewer for this relevant question. The cost of spheroid generation using our platform is largely comparable to conventional 2D MIN6 cultures, with the primary additional expense being the specialized micropatterned PEG-based hydrogel plates. All other aspects of the workflow, including cell culture reagents, imaging consumables, and instrumentation, remain identical. It should be noted that providing a precise cost comparison is difficult, as the price of the hydrogel plates represents the dominant variable cost and is subject to change depending on production scale and supplier agreements. Nevertheless, we consider the platform to be cost-effective relative to alternative 3D culture systems, which often require more complex fabrication procedures or proprietary consumables.

Reviewer #3 (Public review):

Summary:

The primary objective of this study is to develop high-throughput screening assays utilizing homogeneous 3D cell cultures that more accurately replicate the intricate architecture and cellular communication found in tissues. The authors have chosen pancreatic islet β-cells as a model system to evaluate agents that modulate insulin release, which is particularly relevant given the increasing prevalence of diabetes mellitus-a significant global health concern. Moreover, the incorporation of human-based 3D spheroids, organoids, or organ-on-chip technologies into drug discovery protocols is essential for enhancing clinical translation, as candidate compounds identified using animal models have often demonstrated limited success in clinical settings.

Strengths:

This study was thoughtfully planned and skillfully carried out. The use of micropatterned hydrogels to observe 19 spheroids at once is an ingenious aspect, which has been effectively validated with Ca microfluorography. Overall, I found this investigation to be exceptionally well-executed and free from notable flaws, as the results clearly back up the conclusions. Additionally, the developed method achieved the proposed aims, providing a high-throughput format with 3D cultures. I believe this study deserves publication.

Weaknesses:

For an HTS assay, authors should incorporate the Z-factor.

We thank the reviewer for their valuable comment, which we have directly addressed in the revised version, as outlined below.

Reviewer #3 (Recommendations for the authors):

(1) The study is very well performed, but to support the claim of suitability for HTS, the Z-factor of the method should be reported.

We thank the reviewer for this helpful suggestion. In response, we have now included the Z′-factor in the manuscript to explicitly quantify assay performance and robustness. Specifically, the Z′-factor (0.6) has been added and discussed in the Results section, incorporated into the Discussion to contextualize assay suitability for high-throughput screening, and included in the Methods under “Statistical Analysis,” where its calculation is described.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure 1—source data 1. Contact point diameters per sample.

elife-110428-fig1-data1.xlsx^{(21.3KB, xlsx)}

Figure 2—source data 1. Number of spheroids still present per well at the time of the control stimulus.

elife-110428-fig2-data1.xlsx^{(15.7KB, xlsx)}

Figure 3—source data 1. Source data for Figure 3A, B, C and D.

elife-110428-fig3-data1.xlsx^{(682.8KB, xlsx)}

Figure 4—source data 1. Source data for Figure 4 A & B.

elife-110428-fig4-data1.xlsx^{(139.5KB, xlsx)}

Figure 5—source data 1. Source data for Figure 5 A & B.

elife-110428-fig5-data1.xlsx^{(181.7KB, xlsx)}

Figure 6—source data 1. Source data for Figure 4 A, B & C.

elife-110428-fig6-data1.xlsx^{(200.3KB, xlsx)}

Figure 7—source data 1. Source data for Figure 7 A, B, C, D, E & F.

elife-110428-fig7-data1.xlsx^{(320.3KB, xlsx)}

MDAR checklist

elife-110428-mdarchecklist1.pdf^{(186.1KB, pdf)}

Data Availability Statement

All source data needed to generate the figures has been included.

[bib1] American Diabetes Association Diagnosis and classification of diabetes mellitus. Diabetes Care. 2010;33:S62–S69. doi: 10.2337/dc10-S062. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib2] Amin J, Ramachandran K, Williams SJ, Lee A, Novikova L, Stehno-Bittel L. A simple, reliable method for high-throughput screening for diabetes drugs using 3D β-cell spheroids. Journal of Pharmacological and Toxicological Methods. 2016;82:83–89. doi: 10.1016/j.vascn.2016.08.005. [DOI] [PubMed] [Google Scholar]

[bib3] Antunes CM, Salgado AP, Rosário LM, Santos RM. Differential patterns of glucose-induced electrical activity and intracellular calcium responses in single mouse and rat pancreatic islets. Diabetes. 2000;49:2028–2038. doi: 10.2337/diabetes.49.12.2028. [DOI] [PubMed] [Google Scholar]

[bib4] Becker A, Mannebach S, Mathar I, Weissgerber P, Freichel M, Loodin AP, Fecher-Trost C, Belkacemi A, Beck A, Philipp SE. Control of insulin release by Transient Receptor Potential Melastatin 3 (TRPM3) ion channels. Cellular Physiology and Biochemistry. 2020;54:1115–1131. doi: 10.33594/000000304. [DOI] [PubMed] [Google Scholar]

[bib5] Białkowska K, Komorowski P, Bryszewska M, Miłowska K. Spheroids as a type of three-dimensional cell cultures-examples of methods of preparation and the most important application. International Journal of Molecular Sciences. 2020;21:6225. doi: 10.3390/ijms21176225. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] Bommer C, Heesemann E, Sagalova V, Manne-Goehler J, Atun R, Bärnighausen T, Vollmer S. The global economic burden of diabetes in adults aged 20–79 years: a cost-of-illness study. The Lancet Diabetes & Endocrinology. 2017;5:423–430. doi: 10.1016/S2213-8587(17)30097-9. [DOI] [PubMed] [Google Scholar]

[bib7] Broichhagen J, Schönberger M, Cork SC, Frank JA, Marchetti P, Bugliani M, Shapiro AMJ, Trapp S, Rutter GA, Hodson DJ, Trauner D. Optical control of insulin release using a photoswitchable sulfonylurea. Nature Communications. 2014;5:5116. doi: 10.1038/ncomms6116. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib8] Cho NH, Shaw JE, Karuranga S, Huang Y, da Rocha Fernandes JD, Ohlrogge AW, Malanda B. IDF diabetes atlas: global estimates of diabetes prevalence for 2017 and projections for 2045. Diabetes Research and Clinical Practice. 2018;138:271–281. doi: 10.1016/j.diabres.2018.02.023. [DOI] [PubMed] [Google Scholar]

[bib9] Cole JB, Florez JC. Genetics of diabetes mellitus and diabetes complications. Nature Reviews. Nephrology. 2020;16:377–390. doi: 10.1038/s41581-020-0278-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] DeFronzo RA, Tripathy D. Skeletal muscle insulin resistance is the primary defect in type 2 diabetes. Diabetes Care. 2009;32 Suppl 2:S157–S163. doi: 10.2337/dc09-S302. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] Diane A, Mohammed LI, Al-Siddiqi HH. Islets in the body are never flat: transitioning from two-dimensional (2D) monolayer culture to three-dimensional (3D) spheroid for better efficiency in the generation of functional hPSC-derived pancreatic β cells in vitro. Cell Communication and Signaling. 2023;21:151. doi: 10.1186/s12964-023-01171-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib12] Essaouiba A, Jellali R, Shinohara M, Scheidecker B, Legallais C, Sakai Y, Leclerc E. Analysis of the behavior of 2D monolayers and 3D spheroid human pancreatic beta cells derived from induced pluripotent stem cells in a microfluidic environment. Journal of Biotechnology. 2021;330:45–56. doi: 10.1016/j.jbiotec.2021.02.009. [DOI] [PubMed] [Google Scholar]

[bib13] Gilon P, Ravier MA, Jonas JC, Henquin JC. Control mechanisms of the oscillations of insulin secretion in vitro and in vivo. Diabetes. 2002;51 Suppl 1:S144–S151. doi: 10.2337/diabetes.51.2007.s144. [DOI] [PubMed] [Google Scholar]

[bib14] Haque MA. Feature Engineering and Selection for Explainable Models: A Second Course for Data Scientists. LULU Press, Inc; 2022. [Google Scholar]

[bib15] Islam MS. Calcium signaling in the islets. Advances in Experimental Medicine and Biology. 2010;654:235–259. doi: 10.1007/978-90-481-3271-3_11. [DOI] [PubMed] [Google Scholar]

[bib16] Jun Y, Lee J, Choi S, Yang JH, Sander M, Chung S, Lee S-H. In vivo-mimicking microfluidic perfusion culture of pancreatic islet spheroids. Science Advances. 2019;5:eaax4520. doi: 10.1126/sciadv.aax4520. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] Lee SH, Hong S, Song J, Cho B, Han EJ, Kondapavulur S, Kim D, Lee LP. Microphysiological analysis platform of pancreatic Islet β-cell spheroids. Advanced Healthcare Materials. 2018;7:1701111. doi: 10.1002/adhm.201701111. [DOI] [PubMed] [Google Scholar]

[bib18] Mohandas S, Gayatri V, Kumaran K, Gopinath V, Paulmurugan R, Ramkumar KM. New frontiers in three-dimensional culture platforms to improve diabetes research. Pharmaceutics. 2023;15:725. doi: 10.3390/pharmaceutics15030725. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib19] Nielsen MS, Axelsen LN, Sorgen PL, Verma V, Delmar M, Holstein-Rathlou NH. Gap junctions. Comprehensive Physiology. 2012;2:1981–2035. doi: 10.1002/cphy.c110051. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib20] Rupnik M. The physiology of rodent beta-cells in pancreas slices. Acta Physiologica. 2009;195:123–138. doi: 10.1111/j.1748-1716.2008.01927.x. [DOI] [PubMed] [Google Scholar]

[bib21] Saeedi P, Salpea P, Karuranga S, Petersohn I, Malanda B, Gregg EW, Unwin N, Wild SH, Williams R. Mortality attributable to diabetes in 20-79 years old adults, 2019 estimates: results from the international diabetes federation diabetes Atlas, 9th edition. Diabetes Research and Clinical Practice. 2020;162:108086. doi: 10.1016/j.diabres.2020.108086. [DOI] [PubMed] [Google Scholar]

[bib22] Sankar KS, Green BJ, Crocker AR, Verity JE, Altamentova SM, Rocheleau JV. Culturing pancreatic islets in microfluidic flow enhances morphology of the associated endothelial cells. PLOS ONE. 2011;6:e24904. doi: 10.1371/journal.pone.0024904. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] Savitzky A, Golay MJE. Smoothing and differentiation of data by simplified least squares procedures. Analytical Chemistry. 1964;36:1627–1639. doi: 10.1021/ac60214a047. [DOI] [PubMed] [Google Scholar]

[bib24] Shigeto M, Katsura M, Matsuda M, Ohkuma S, Kaku K. First phase of glucose-stimulated insulin secretion from MIN 6 cells does not always require extracellular calcium influx. Journal of Pharmacological Sciences. 2006;101:293–302. doi: 10.1254/jphs.fp0060057. [DOI] [PubMed] [Google Scholar]

[bib25] Shigeto M, Katsura M, Matsuda M, Ohkuma S, Kaku K. Nateglinide and mitiglinide, but not sulfonylureas, induce insulin secretion through a mechanism mediated by calcium release from endoplasmic reticulum. The Journal of Pharmacology and Experimental Therapeutics. 2007;322:1–7. doi: 10.1124/jpet.107.120592. [DOI] [PubMed] [Google Scholar]

[bib26] Sun D, Gao W, Hu H, Zhou S. Why 90% of clinical drug development fails and how to improve it? Acta Pharmaceutica Sinica B. 2022;12:3049–3062. doi: 10.1016/j.apsb.2022.02.002. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib27] Vhora N, Naskar U, Hiray A, Kate AS, Jain A. Recent advances in In-Vitro assays for type 2 diabetes mellitus: an overview. Review of Diabetic Studies. 2020;16:13–23. doi: 10.1900/RDS.2020.16.13. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] Virtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D, Burovski E, Peterson P, Weckesser W, Bright J, van der Walt SJ, Brett M, Wilson J, Millman KJ, Mayorov N, Nelson ARJ, Jones E, Kern R, Larson E, Carey CJ, Polat İ, Feng Y, Moore EW, VanderPlas J, Laxalde D, Perktold J, Cimrman R, Henriksen I, Quintero EA, Harris CR, Archibald AM, Ribeiro AH, Pedregosa F, van Mulbregt P, SciPy 1.0 Contributors SciPy 1.0: fundamental algorithms for scientific computing in Python. Nature Methods. 2020;17:261–272. doi: 10.1038/s41592-019-0686-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] Wagner TFJ, Loch S, Lambert S, Straub I, Mannebach S, Mathar I, Düfer M, Lis A, Flockerzi V, Philipp SE, Oberwinkler J. Transient receptor potential M3 channels are ionotropic steroid receptors in pancreatic beta cells. Nature Cell Biology. 2008;10:1421–1430. doi: 10.1038/ncb1801. [DOI] [PubMed] [Google Scholar]

[bib30] Wassmer C-H, Bellofatto K, Perez L, Lavallard V, Cottet-Dumoulin D, Ljubicic S, Parnaud G, Bosco D, Berishvili E, Lebreton F. Engineering of primary pancreatic islet cell spheroids for three-dimensional culture or transplantation: a methodological comparative study. Cell Transplantation. 2020;29:963689720937292. doi: 10.1177/0963689720937292. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib31] Wikstrom JD, Sereda SB, Stiles L, Elorza A, Allister EM, Neilson A, Ferrick DA, Wheeler MB, Shirihai OS. A novel high-throughput assay for islet respiration reveals uncoupling of rodent and human islets. PLOS ONE. 2012;7:e33023. doi: 10.1371/journal.pone.0033023. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

A high-throughput assay for the measurement of Ca2+ oscillations and insulin release from uniformly sized mouse β-cell (MIN6) spheroids

Stijn Robben

Patricia Davidson

Rita S Rodrigues Ribeiro

Thomas Voets

Roles

Abstract

Introduction

Results

Hydrogel design

Figure 1. Hydrogel design.

Protocol

Figure 2. Cell seeding and data acquisition.

Increased levels of extracellular glucose induce Ca2+-oscillatory behavior

Figure 3. Oscillatory behavior of Ca2+ signal is modulated by extracellular glucose concentrations.

Figure 3—figure supplement 1. Ca2+ response to glucose of mouse insulinoma 6 (MIN6) cells in 2D culture.

KATP-agonist diazoxide inhibits Ca2+-oscillatory response to extracellular glucose

Figure 4. KATP-agonist diazoxide inhibits oscillatory response to extracellular glucose.

KATP-antagonists induce Ca2+-oscillatory behavior

Figure 5. KATP-antagonists induce Ca2+ oscillatory behavior with distinct potencies.

KATP-channel modulators impact insulin secretion by MIN6 spheroids

Figure 6. Glimepiride and diazoxide impact glucose-dependent insulin release by mouse insulinoma 6 (MIN6) spheroids.

PS induces Ca2+-oscillatory behavior and increases insulin secretion

Figure 7. Pregnenolone sulfate (PS) induces Ca2+-oscillatory behavior and increases insulin secretion.

Discussion

Materials and methods

Key resources table.

Mold design and characterization

Hydrogel substrate preparation

Cell culture

β-Cell spheroid formation

Compound plate preparation

Ca2+ imaging

Insulin detection

Data analysis

Statistical analysis

Acknowledgements

Funding Statement

Contributor Information

Funding Information

Additional information

Competing interests

Author contributions

Additional files

Data availability

References

eLife Assessment

Leon D Islas

Roles

Reviewer #1 (Public review):

Anonymous

Roles

Reviewer #2 (Public review):

Anonymous

Roles

Reviewer #3 (Public review):

Anonymous

Roles

Author response

Stijn Robben

Patricia Davidson

Rita S Rodrigues Ribeiro

Thomas Voets

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

A high-throughput assay for the measurement of Ca²⁺ oscillations and insulin release from uniformly sized mouse β-cell (MIN6) spheroids

Increased levels of extracellular glucose induce Ca²⁺-oscillatory behavior

Figure 3. Oscillatory behavior of Ca²⁺ signal is modulated by extracellular glucose concentrations.

Figure 3—figure supplement 1. Ca²⁺ response to glucose of mouse insulinoma 6 (MIN6) cells in 2D culture.

K_ATP-agonist diazoxide inhibits Ca²⁺-oscillatory response to extracellular glucose

Figure 4. K_ATP-agonist diazoxide inhibits oscillatory response to extracellular glucose.

K_ATP-antagonists induce Ca²⁺-oscillatory behavior

Figure 5. K_ATP-antagonists induce Ca²⁺ oscillatory behavior with distinct potencies.

K_ATP-channel modulators impact insulin secretion by MIN6 spheroids

PS induces Ca²⁺-oscillatory behavior and increases insulin secretion

Figure 7. Pregnenolone sulfate (PS) induces Ca²⁺-oscillatory behavior and increases insulin secretion.

Ca²⁺ imaging