TRPM7 is a crucial regulator of pancreatic endocrine development and high-fat-diet-induced β-cell proliferation

Although Mg²⁺ serves many essential roles and is the most abundant divalent cation in pancreatic β-cells, the mechanisms that modulate β-cell Mg²⁺ homeostasis have not been conclusively determined (Fang et al., 2013; Sahni and Scharenberg, 2008; Sahni et al., 2010; Sgambato et al., 1999; Yu et al., 2014). The transient receptor potential cation (TRPM) channel, TRPM7, is a regulator of cellular Mg²⁺ homeostasis and TRPM7 is one of the most abundant TRPM transcripts expressed in human β-cells (Li et al., 2016a; Ryazanova et al., 2010; Schmitz et al., 2003). TRPM7 channels are inactivated at physiological Mg²⁺ levels and are activated in response to a decrease in intracellular or extracellular Mg²⁺; thus, TRPM7 channels tightly control intracellular Mg²⁺ levels (Schmitz et al., 2003). For example, TRPM7 channels facilitate Mg²⁺ influx during the G1 phase of the cell cycle, which stimulates DNA and protein synthesis (Rubin, 2005b; Tani et al., 2007). TRPM7 is also essential for embryonic development, as silencing or deleting TRPM7 perturbs organogenesis (Jin et al., 2008, 2012). Furthermore, reduced TRPM7 expression results in decreased proliferation (Fang et al., 2013; Ryazanova et al., 2010; Sahni et al., 2010). Interestingly, Mg²⁺ supplementation restores cell cycle progression and organogenesis in the absence of TRPM7, which reinforces the importance of Mg²⁺ homeostasis during development (Ryazanova et al., 2010; Yee et al., 2005, 2011). Although it has been established that TRPM7 control of Mg²⁺ handling regulates cellular development and proliferation, the role of TRPM7 in β-cell function remains poorly understood.

TRPM7 control of Mg²⁺ regulates the expression and activity of cell cycle and signaling pathways that influence pancreatic-cell and β-cell proliferation (Fang et al., 2013; Sahni and Scharenberg, 2008; Sahni et al., 2010; Sgambato et al., 1999; Yee et al., 2011; Yu et al., 2014). TRPM7 knockdown (KD) in pancreatic adenocarcinoma cells (BxPC-3 and PANC-1) decreases proliferation owing to upregulation of p21^cip1 and suppressor of cytokine signaling 3a, which is restored by supplementation with high extracellular Mg²⁺ (Yee et al., 2011). In other cell types, TRPM7 ablation altered expression of key cell cycle regulatory proteins, many of which modulate β-cell development and proliferation (Fang et al., 2013; Sahni and Scharenberg, 2008; Sahni et al., 2010; Sgambato et al., 1999; Yu et al., 2014). For example, lymphocyte proliferation is decreased following TRPM7 ablation due in part to p27^Kip1 upregulation; similarly, overexpression of p27^Kip1 inhibits both neonatal and postnatal β-cell proliferation, whereas β-cell knockout (KO) of p27^Kip1 increases β-cell mass (Rachdi et al., 2006; Sahni et al., 2010). TRPM7 KD in hepatic cells reduces expression of G1 checkpoint regulators (cyclin D1 and Cdk4), both of which stimulate β-cell proliferation (Cozar-Castellano et al., 2004; Fang et al., 2013; Rane et al., 1999; Zhang et al., 2005). TRPM7 KO also resulted in decreased mTOR, AKT and ERK activity, which are crucial determinates of β-cell proliferation in response to a high fat diet (HFD), during pregnancy and in neonates (Alejandro et al., 2014; Balcazar et al., 2009; Bernal-Mizrachi et al., 2004, 2001; Chen et al., 2011; Fang et al., 2013; Gupta et al., 2007; Sahni and Scharenberg, 2008; Sahni et al., 2010; Sgambato et al., 1999; Yu et al., 2014). This suggests that TRPM7 channel modulation of pancreatic-cell Mg²⁺ handling may influence cell cycle progression and proliferative capacity.

Recent evidence has begun to reveal the importance of TRPM7 in maintaining Mg²⁺ homeostasis during pancreas development. A random mutagenesis screen in zebrafish found that TRPM7 inactivation reduces pancreas mass, a phenotype replicated by TRPM7 KD of zebrafish (Yee et al., 2005, 2011). Importantly, Mg²⁺ supplementation partially restores pancreatic size in TRPM7 KD zebrafish, which suggested that TRPM7 control of Mg²⁺ homeostasis is crucial for pancreas development (Yee et al., 2005, 2011). As TRPM7 is known to serve an important role in early organogenesis, it may impact pancreatic progenitor proliferation and/or function, which could affect the endocrine progenitor pool (Jin et al., 2012). Indeed, TRPM7 ablation in neural crest progenitors (NCPs) leads to loss of pigment cells and dorsal root ganglion neurons (Jin et al., 2012). Interestingly, ablation of NCP TRPM7 at embryonic day (E) 10.5, but not E14.5, disrupted development (Jin et al., 2012). Although this suggests that TRPM7 serves a function during early pancreas organogenesis, the role of TRPM7 in controlling development from either pancreatic or endocrine progenitors has not been established. Therefore, it is important to determine whether TRPM7 impacts pancreas organogenesis and endocrine specification during the specific progenitor developmental time points that determine endocrine mass.

In addition to its influence on pancreas development, Mg²⁺ flux through TRPM7 channels directly regulates β-cell function (Gommers et al., 2019). For example, glucose-stimulated insulin secretion (GSIS) from INS-1 rat insulinoma cells is enhanced following TRPM7 KD (Gommers et al., 2019). The changes in GSIS caused by TRPM7 are likely because of prolonged reductions in Mg²⁺ or mediated via its permeability to other divalent ions, such as Zn²⁺ or Ca²⁺ (Gommers et al., 2019). Zn²⁺ is crucial for insulin crystallization in granules; rodent β-cells expressing a loss-of-function Zn²⁺ transporter (ZnT8) display increased GSIS resembling INS-1 TRPM7 KD GSIS, whereas ZnT8 deletion results in decreased secretion (Hardy et al., 2012; Kleiner et al., 2018). Because TRPM7 is an outward-rectifying channel with only modest Ca²⁺ conductance at physiological membrane potentials (Runnels et al., 2001), it is unlikely that Ca²⁺ entry through TRPM7 channels would directly affect β-cell Ca²⁺ handling. However, TRPM7 channel control of intracellular Mg²⁺ could indirectly modulate β-cell Ca²⁺ handling by influencing voltage-dependent Ca²⁺ channel activity (Nguemo et al., 2014; Zierler et al., 2017). Thus, it is important to determine whether TRPM7 influences primary β-cell GSIS and whether this could be due to alterations in divalent cation flux.

Here, we show that TRPM7 serves important roles in pancreatic endocrine development and function. The data indicate that TRPM7 promotes early pancreatic endocrine development and β-cell proliferation. Furthermore, these findings identify that TRPM7 controls β-cell Mg²⁺ and Ca²⁺ handling, which could contribute to the enhanced GSIS observed in islets with TRPM7 ablation. Finally, the data reveal that TRPM7 contributes to elevating β-cell Mg²⁺ during proliferation.

TRPM7 is a key determinant of pancreas size in zebrafish and serves a crucial role during early organogenesis (Jin et al., 2012; Yee et al., 2005, 2011); therefore, we investigated the effect of TRPM7 on mouse pancreatic development. A mouse line with floxed TRPM7 exon 17 (TRPM7^fl/fl) and a Pdx1-Cre^TUV transgene was used to selectively knockout TRPM7 in pancreatic progenitor cells (termed TRPM7KO^Panc); loss of TRPM7 protein expression in TRPM7KO^Panc pancreata was confirmed following removal of TRPM7 exon17 in islets (Fig. S1A-C). TRPM7KO^Panc mice displayed pancreatic hypoplasia and reduced pancreatic mass at E16.5 compared with controls [35.6±8.5% reduction (mean±s.e.m.), Fig. 1A-C; Fig. S2A]. Total pancreatic area was also reduced in TRPM7KO^Panc mice compared with controls (59.4±20.1% decrease, Fig. 1F). These results indicate that TRPM7 channels are crucial mediators of mouse exocrine pancreatic development.

To determine the effect of TRPM7 ablation on pancreatic endocrine development, we compared β- and α-cell mass of control and TRPM7KO^Panc mice (Fig. 1D-J). TRPM7KO^Panc pancreata had fewer β-cells (31.2±10.8% decrease, Fig. 1G), reduced β-cell mass (55.8±9.9% decrease, Fig. 1H) and decreased total pancreatic insulin content (47.4±15.8% decrease, Fig. S2B) compared with controls. TRPM7KO^Panc pancreata also had fewer α-cells (39.6±11.7% decrease, Fig. 1I) and decreased total α-cell mass (61.2±9.2% decrease, Fig. 1J) compared with controls. Intraperitoneal glucose tolerance tests (IPGTTs) were performed to assess the impact of reduced endocrine mass on glucose homeostasis. Despite a significant decrease in β-cell number (Fig. 1G), β-cell mass (Fig. 1H) and total insulin content (Fig. S2B), the TRPM7KO^Panc mice were only modestly glucose intolerant compared with controls (Fig. 1K,L). Interestingly, isolated TRPM7KO^Panc islets showed enhanced GSIS compared with controls (40.0±10.7% increase, Fig. S2C,D) even though total islet insulin content was decreased, which could potentially limit glucose intolerance in TRPM7KO^Panc mice. Furthermore, although fasting plasma insulin was reduced in TRPM7KO^Panc mice, plasma insulin reached control levels at 30 min post intraperitoneal (IP) glucose challenge (Fig. S2E). Also, the Pdx1-Cre^TUV line used for this study not only expresses cre in pancreatic progenitors but also in a limited number of other tissues such as the hypothalamus (Wicksteed et al., 2010); therefore, TRPM7 loss in other tissues may also impact glucose tolerance tests (GTT) in TRPM7KO^Panc mice. Taken together, these data reveal that TRPM7 channels are key determinants of early pancreatic endocrine development.

Having identified a role for TRPM7 during early pancreatic organogenesis, we sought to determine whether TRPM7 is also required for endocrine progenitor differentiation. Although previous studies have found that TRPM7 is primarily a determinant of early organogenesis (Jin et al., 2012), it was important to confirm whether TRPM7 also serves a cell- or time-specific role during endocrine development. Therefore, an endocrine progenitor-selective TRPM7 ablation mouse was created by combiningTRPM7^fl/fl and Ngn3-Cre mice (TRPM7KO^Endo; Figs S1A and S3). There was no difference in β- and α-cell number, β- and α-cell area or pancreas area in TRPM7KO^Endo pancreata compared with controls (Fig. 2A-G). This indicates that TRPM7 is not required for pancreatic development after endocrine progenitor specification. Moreover, TRPM7KO^Endo mice showed equivalent glucose tolerance as controls (Fig. 2H,I males, Fig. S5A females). Although these findings suggest that the role of TRPM7 during pancreatic development is restricted to pancreatic progenitors, it is possible that TRPM7 has additional functions in mature islet cells (Gommers et al., 2019).

As TRPM7 regulates cell cycle and signaling pathways that influence β-cell proliferation (Chen et al., 2011; Fang et al., 2013; Sahni and Scharenberg, 2008; Sahni et al., 2010), TRPM7 control of β-cell proliferation was investigated. Mice with conditional β-cell ablation of TRPM7 were placed on a HFD to examine the role of TRPM7 during HFD-induced β-cell proliferation; this mouse line contains both TRPM7^fl/fl and Ins1^CreERT2(TRPM7KO^INSβ). Twelve-week-old TRPM7KO^INSβ mice or controls (Ins1^CreERT2) were treated with tamoxifen, allowed to recover for 2 weeks, placed on a HFD for 2 weeks, and then β-cell proliferation (Ki67⁺) or apoptosis (TUNEL) were assessed. We found that β-cell proliferation was significantly reduced in TRPM7KO^INSβ pancreatic sections compared with control sections (47.9±12.0% decrease, Fig. 3A-C); TRPM7 KO did not affect β-cell apoptosis (Fig. S4A,B). Interestingly, although β-cell proliferation was reduced in TRPM7KO^INSβ sections compared with controls, total pancreatic area (Fig. 3D), islets per section (Fig. 3E) and average islet area were indistinguishable (Fig. 3F). In addition, there was no difference in glucose tolerance pre- or post-HFD in TRPM7KO^INSβ animals compared with controls (Fig. 3G-J males, Fig. S5B females). Based on these findings, TRPM7 regulates HFD-induced β-cell proliferation but does not alter glucose tolerance.

TRPM7 control of Mg²⁺ has been shown to regulate pancreatic mass in zebrafish and proliferation in numerous cell types (Fang et al., 2013; Ryazanova et al., 2010; Sahni et al., 2010; Yee et al., 2005, 2011). Therefore, we explored whether TRPM7KO^Panc and/or TRPM7KO^INSβ β-cells have altered Mg²⁺ handling. Some studies also used another conditional β-cell TRPM7 KO mouse line (TRPM7^fl/fl crossed with MIP-Cre/ERT, termed TRPM7KO^MIPβ; Fig. S6A). First, TRPM7 channels were activated by removing extracellular divalent cations and β-cell TRPM7 currents measured in response to a voltage ramp from −120 mV to +60 mV (Fig. 4A; Fig. S7A,B). TRPM7 currents were significantly reduced under hyperpolarized (−120 to −65 mV) and depolarized (+25 to +60 mV) conditions in TRPM7KO^Panc β-cells compared with controls (Fig. 4A; Fig. S7A,B). Next, intracellular Mg²⁺ was monitored in TRPM7KO^Panc, TRPM7KO^INSβ and TRPM7KO^MIPβ β-cells; following removal of extracellular Mg²⁺, intracellular Mg²⁺ levels were elevated in TRPM7 KO β-cells relative to controls (Fig. 4B,C,D; Fig. S8A). As TRPM7 modulation of Mg²⁺ homeostasis regulates cell proliferation in other tissues (Tani et al., 2007), we sought to determine whether this is the case in β-cells. Thus, cytosolic Mg²⁺ was quantified in proliferating (squares, Ki67-positive) and non-proliferating (circles, Ki67-negative) β-cells (Fig. 4C,D). Intracellular Mg²⁺ was elevated in non-proliferating TRPM7KO^INSβ single β-cells and also in β-cells within clusters of islet cells (Fig. 4C,D; Fig. S9A,B). Proliferating β-cells from TRPM7KO^INSβ and controls have significantly elevated intracellular Mg²⁺ when normalized to non-proliferating control cells (21.4±3.9% and 29.4±3.4% increase, respectively, Fig. 4E; Fig. S9C). However, as the non-proliferating TRPM7KO^INSβ β-cells had greater basal Mg²⁺, the associated magnitude of elevation in Mg²⁺ during proliferation was reduced in TRPM7KO^INSβ compared with controls (48.5±17.9% decrease, Fig. 4F), presumably owing to loss of TRPM7. Furthermore, TRPM7 KO resulted in decreased β-cell proliferation (53.3±19.1% decrease, Fig. S9D), which is consistent with intact pancreatic sections (Fig. 3C). This suggests that TRPM7 limits non-proliferating β-cell Mg²⁺ levels; thus, during proliferation β-cells undergo a greater magnitude in elevation of intracellular Mg²⁺ when TRPM7 is active, which may promote proliferation.

In addition to Mg²⁺, TRPM7 channels are permeable to other divalent cations, thus we examined whether entry of Zn²⁺ or Ca²⁺ though TRPM7 channels may contribute to the enhanced GSIS observed in TRPM7 KD INS-1 cells (Gommers et al., 2019). First, we confirmed whether TRPM7KO^MIPβ islets show enhanced GSIS. Although insulin secretion from TRPM7KO^MIPβ islets was indistinguishable from control islets under euglycemic (7 mM glucose) conditions, GSIS (14 mM glucose) was increased compared with controls (73.7±16.0% increase, Fig. S6C). In the presence of elevated extracellular Zn²⁺, the rate of cytosolic Zn²⁺ accumulation in TRPM7KO^MIPβ β-cells was indistinguishable from controls (Fig. S8B). This indicates that the minimal Zn²⁺ conductance through TRPM7 channels in β-cells is likely not important to their function. However, glucose-stimulated (14 mM) Ca²⁺ influx was decreased in TRPM7KO^Panc islets compared with controls (24.1±0.6% decrease, Fig. S2F,G). To determine whether TRPM7 channel-mediated enhancement of β-cell intracellular Ca²⁺ concentration ([Ca²⁺]_i) was due to developmental changes in TRPM7KO^Panc β-cells, glucose-stimulated Ca²⁺ influx into TRPM7KO^MIPβ islets was also measured. Similar to total pancreatic TRPM7 channel ablation, glucose-stimulated Ca²⁺ influx was reduced in TRPM7KO^MIPβ islets compared with controls (Fig. S8C,D). As these data show that TRPM7 channels augment β-cell glucose-stimulated Ca²⁺ influx, it is unlikely that increased GSIS from TRPM7 KD INS-1 cells, or from TRPM7KO^MIPβ and TRPM7KO^Panc islets, is due to TRPM7 control of β-cell [Ca²⁺]_i (Gommers et al., 2019). However, as the frequency and amplitude of β-cell Ca²⁺ oscillations were not measured, TRPM7 modulation of GSIS by altering the kinetics of Ca²⁺ oscillations cannot be ruled out (Vierra et al., 2015). The change in GSIS is likely compensated for as TRPM7KO^MIPβ and control mice have equivalent glucose tolerance (Fig. S6B). Taken together, these data suggest that TRPM7 channels limit islet GSIS in addition to regulating β-cell development and proliferation.

Here, we show for the first time that TRPM7 regulates β-cell proliferation and GSIS, presumably by controlling β-cell Mg²⁺ homeostasis. Our findings indicate that β-cell intracellular Mg²⁺ levels increase during HFD-induced proliferation, as is the case in many other cell types (Fang et al., 2013; Rubin, 2005a; Sahni and Scharenberg, 2008; Sahni et al., 2010; Sgambato et al., 1999; Yee et al., 2011; Yu et al., 2014). Importantly, there was a significantly greater magnitude in elevation of intracellular Mg²⁺ during β-cell proliferation with functional TRPM7 channels, which is likely due to TRPM7 limiting intracellular Mg²⁺ levels in non-proliferating β-cells. This suggests that TRPM7-mediated control of Mg²⁺ homeostasis could play a role in β-cell proliferation (Tani et al., 2007). Moreover, reduced intracellular Mg²⁺ in non-proliferating β-cells with functional TRPM7 channels also indicates that TRPM7 regulates basal β-cell Mg²⁺ homeostasis. Although the exact mechanism(s) of how an elevation in cytosolic Mg²⁺ modulates β-cell proliferation remain(s) to be determined, in other cells Mg²⁺ influx can activate phosphoinositide 3-kinase-dependent signaling and AKT phosphorylation (Peng et al., 2020; Tuttle et al., 2001). Because AKT signaling is known to play a crucial role in β-cell proliferation (Peng et al., 2020; Tuttle et al., 2001), one possibility is that blunted AKT phosphorylation during proliferation in TRPM7-deficient β-cells limits proliferation in response to HFD. Perturbed Mg²⁺ homeostasis in TRPM7-deficient β-cells may impact other cell cycle pathways controlled by Mg²⁺ such as CDK2 kinase activity, expression of kinase inhibitor genes (e.g. p21^cip1 and p27^kip1), DNA synthesis and/or cytoskeletal remodeling (Harper et al., 1993; Polyak et al., 1994; Rubin, 1975). Finally, it has been shown that lymphocytes without TRPM7 channels enter a quiescent state, potentially owing to a large reduction in aerobic glycolysis necessary for sustained proliferation (Sahni et al., 2010). As Mg²⁺ is required for sequential enzymatic reactions within the glycolytic pathway, this may limit proliferation in TRPM7KO^Insβ β-cells under stressful conditions (Garfinkel and Garfinkel, 1985; Matsuda et al., 1999; Sahni and Scharenberg, 2008; Wimhurst and Manchester, 1972). Future studies are required to understand exactly how TRPM7 increases β-cell proliferation and how elevations in β-cell Mg²⁺ impact proliferation.

TRPM7 control of Mg²⁺ plays a crucial role in zebrafish pancreatic development and may also impact endocrine development. Indeed, the role of TRPM7 in endocrine cell development is striking, as TRPM7 KO in pancreatic progenitors resulted in >60% reduction of α- and β-cell mass (Fig. 1H,J). During early pancreatic development there is a large expansion of the multipotent progenitor cells, which are crucial determinants of adult pancreatic size (Stanger et al., 2007). Thus, TRPM7 may regulate the pool size of pancreatic progenitors, presumably through Mg²⁺ homeostasis that potentially regulates proliferation. Indeed, pancreatic development in TRPM7 KD zebrafish is rescued by Mg²⁺ supplementation (Yee et al., 2011). Therefore, future studies are needed to determine how Mg²⁺ alters pancreatic progenitor function as well as proliferation, and whether this is altered under conditions of aberrant Mg²⁺ handling during gestation. Conversely, there are no observed deleterious effects on endocrine mass or glucose tolerance following loss of TRPM7 during endocrine progenitor specification. Therefore, TRPM7 control of Mg²⁺ is likely not a crucial factor in endocrine differentiation (Castiglioni et al., 2013; Wang et al., 2010). This temporal developmental effect of TRPM7 is supported by studies on neural, kidney and pigment cells, in which TRPM7 is crucial for early organogenesis, but has a minimal effect on terminal differentiation (Jin et al., 2012). Taken together, our data indicate that TRPM7 plays an essential role in early pancreatic development as well as HFD-induced β-cell proliferation, which may be due in-part to TRPM7 control of Mg²⁺ handling. Therefore, these findings reveal that targeting TRPM7 or developing methods to control β-cell Mg²⁺ handling could provide therapeutic strategies for increasing β-cell mass.

All mice used in these studies were 12-19 weeks old or harvested at E16.5, age-matched males on a C57BL/6 background. Animals were handled in compliance with guidelines approved by the Vanderbilt University Animal Care and Use Committee protocols (protocol M1600063-01).

For spatiotemporal TRPM7 ablation, transgenic animals were used by crossing 129S4/SvJae mice containing LoxP sites inserted around exon 17 in the TRPM7 gene (TRPM7^fl/fl, The Jackson Laboratory, 018784) (Jin et al., 2008) with various Cre-recombinase lines. For pancreatic progenitor-specific ablation, TRPM7^fl/fl mice were crossed with C57BL/6 mice expressing a Cre-recombinase under the Pdx1 promoter (Pdx1-cre)Tuv/J (Jin et al., 2008). Endocrine-specific ablation mice were created by crossing TRPM7^fl/fl mice with Ngn3-cre mice (Gu et al., 2002). β-Cell-specific TRPM7 ablation mice were generated by crossing TRPM7^fl/fl and C57BL/6 mice expressing a tamoxifen-activated CreERT-recombinase in β-cells (MIP-CreERT, the Jackson Laboratory, 024709; Ins1^CreERT2, the Jackson Laboratory, 026802) (Thorens et al., 2015; Wicksteed et al., 2010). Translocation of the CreERT into the nucleus was induced by treating CreERT mice with tamoxifen (2 mg ml⁻¹; RPI) every other day five times as indicated. Age-matched TRPM7^fl−/fl− mice were used as controls and were treated with tamoxifen identically to the TRPM7^fl/fl MIP-CreERT and TRPM7^fl/fl-Ins1^CreERT2 mice.

Mice were placed either on a normal chow or HFD (60% kcal from fat, D12492 Research Diets) and monitored for glucose tolerance. The GTT was performed as described previously by injecting 2 mg/kg dextrose and monitoring blood glucose at the indicated time points post-glucose injection after an ∼5-6 h fast (Jacobson et al., 2007). GTT was performed between 12-19 weeks of age and also in a cohort of mice placed on HFD for 2 weeks at 15 weeks of age.

Mouse islets were isolated by digesting the pancreas with collagenase P (Roche) and performing density gradient centrifugation as previously described (Roe et al., 1994). Islets were plated or dissociated in 0.005% trypsin, placed on glass coverslips and cultured for 16 h in RPMI 1640 medium supplemented with 15% fetal bovine serum, specified glucose concentrations, 100 mg ml⁻¹ penicillin, and 100 mg ml⁻¹ streptomycin. Dissociated β-cells were used in all voltage clamp experiments recording Mg²⁺ currents. β-Cells on the periphery of intact islets were recorded in current clamp mode in all membrane potential recordings. Cells and islets were maintained in a humidified incubator at 37°C under an atmosphere of 95% air and 5% CO₂.

Insulin secretion from mouse islets was determined as previously described (Dadi et al., 2014). In summary, mouse islets were allowed to recover following isolation overnight at 5.6 mM glucose. For technical replicates, 20 equal-sized islets were incubated in 2 mM glucose in Krebs-Ringer buffer (KRB; 119 mM NaCl, 2 mM CaCl₂, 4.7 mM KCl, 10 mM HEPES, 1.2 mM MgSO₄, 1.2 mM KH₂PO₄, adjusted to pH 7.35 with NaOH) for 1 h, followed by the indicated treatments. Insulin concentrations were determined using the Insulin Rodent Chemiluminescence ELISA kit (ALPCO).

TRPM7KO^Panc islets were dispersed into single cells and cultured overnight at 37°C, 5% CO₂. Recording pipettes (8-10 MΩ) were backfilled with intracellular solution containing (mM): 140.0 CsCl, 1.0 MgCl₂, 10.0 EGTA and 10.0 HEPES (pH 7.2, adjusted with CsOH) supplemented with 3.7 Mg-ATP. Cells were patched in extracellular buffer containing (mM): 119.0 NaCl, 4.7 KCl, 25.0 HEPES, 1.2 MgSO₄, 1.2 KH₂PO₄, 2.0 CaCl₂, (pH 7.4, adjusted with NaOH) supplemented with 14 mM glucose. After a whole-cell configuration was established (seal resistance>1 GΩ) the bath solution was exchanged (3 min) for divalent cation-free buffer containing (mM): 119.0 NaCl, 4.7 KCl, 25.0 HEPES, 1.2 KH₂PO₄ and 10.0 tetraethylammonium chloride supplemented with 14 mM glucose (pH 7.4, adjusted with NaOH). A voltage-clamp protocol was employed to record whole-cell currents with and without extracellular divalent cations in response to a voltage ramp from −120 to 60 mV (1 s interval) using an Axopatch 200B amplifier with pCLAMP10 software (Molecular Devices) as previously described (Vierra et al., 2015, 2017). TRPM7 currents were calculated by subtracting currents without divalents from currents with divalents. Data were analyzed using Clampfit 10 (Molecular Devices) and Microsoft Excel software.

Islet cell clusters were incubated (25 min at 37°C) in RPMI supplemented with 2 μM Fura-2 acetoxymethyl ester (Ca²⁺) or 2 μM FluoZin-3 acetoxymethyl ester (Zn²⁺; Molecular Probes). Fluorescence imaging was performed using a Nikon Eclipse TE2000-U microscope equipped with an epifluorescence illuminator (SUTTER), a CCD camera (HQ2, Photometrics) and Nikon Elements software. Relative Ca²⁺ concentrations were recorded with perifusion of KRB containing the indicated glucose concentrations and quantified every 5 s by determining the ratio of emitted fluorescence intensities at excitation wavelengths of 340 and 380 nm (F₃₄₀/F₃₈₀). The relative Ca²⁺ is plotted as the mean FURA-2 ratio (F₃₄₀/F₃₈₀) of each experimental group±s.e.m. Zn²⁺ concentrations were recorded with perifusion of KRB containing the indicated glucose and Zn²⁺ concentrations and quantified every 5 s by measuring the emitted fluorescent intensity at excitation of 516 nm and normalized to the lowest intensity per replicate. The averaged area under the curve (AUC) measurement for FURA-2 ratios and normalized FluoZin-3 levels during the indicated time period (min) was plotted as a bar graph. Data were analyzed using Excel and GraphPad Prism software and compared using an unpaired two-tailed Student's t-test.

Islet cell clusters or dispersed β-cells were incubated (35 min at 37°C) in RPMI supplemented with ratiometric intracellular Mg²⁺ indicator Mag-fura-2 AM. This acetoxymethyl (AM) ester form was used for noninvasive intracellular loading (Invitrogen, M1292). Islet clusters or dispersed cells were washed three times with KRB-based solution without Mg²⁺ and then pre-incubated for 25-30 min before imaging. Fluorescence imaging and analysis of data were performed on a Nikon Eclipse scope and quantified every 5 s by measuring the emitted fluorescent intensity at excitation wavelengths of 340 and 380 nm (F₃₄₀/F₃₈₀) of each experimental group. Mg²⁺ monitored cells were then fixed with 4% paraformaldehyde and immunostained with primary antibodies: guinea pig anti-insulin (1:300; Millipore, 4011-01F; Hogh et al., 2013) and rabbit anti-Ki67 (1:300; Abcam, ab15580; Ji et al., 2019) and fluorescent isothiocyanate-conjugated secondary antibodies (1:300; Jackson ImmunoResearch) to identify Ki67⁺ β-cells.

Animals were starved for ∼4 h before glucose injections and then blood was collected for insulin measurements. Blood was collected in EDTA-Heparin coated VAT tubes (Sarstedt). VAT Tubes were spun down with centrifugation at 10,000 RPM (∼9,400 g) for 10 min at 4°C and serum supernatants were collected. Serum samples were determined by the Vanderbilt Hormone Assay and Analytical Services Core using an Insulin radioimmunoassay kit (Sigma-Aldrich).

Pancreatic tissue was dissected and fixed overnight at 4°C in 4% paraformaldehyde, paraffin embedded and cut into 5-μm sections using a microtome (HM 355S, Thermo Fisher Scientific). Sections were incubated overnight with primary antibodies at 4°C followed by 1 h incubation in fluorescent isothiocyanate-conjugated secondary antibodies (1:300; Jackson ImmunoResearch) and DAPI nuclear stain (Invitrogen). Primary antibodies used were: guinea pig anti-insulin (1:300; Millipore, 4011-01F), rabbit anti-glucagon (1:300; Sigma-Aldrich, SAB4501137; Li et al., 2016b), rabbit anti-Ki67 (1:300; Abcam, ab15580) and mouse anti-cre (1:300; Sigma-Aldrich, MAB3120; Jeanes et al., 2012).

Measurements were carried out according to the protocol described by Henley et al. (2012). Briefly, 10 immunostained sections ∼250 μm apart (10 per animal) were visualized on an Aperio ScanScope FL and a custom macro for the ImageScope software was used to measure total pancreatic, β-cell and α-cell area (Rodriguez-Calvo et al., 2017). β-Cell mass was calculated using the ratio of β-cell area to total pancreas area of all sections for each animal and multiplied by the tissue wet weight.

Whole pancreas lysates from adult C57BL/6J and TRPM7KO^Panc were extracted and homogenized with a manual Potter-Elvehjem Tissue Homogenizer in RIPA buffer. Equal concentrations of extracts, as determined by BCA Protein Assay (Pierce), were then incubated overnight with mouse anti-TRPM7 (1:10; NeuroMab, 73-114) at 4°C, followed by an incubation with Protein A/G magnetic beads (Pierce). Bound protein was eluted according to the manufacturer's protocol. Eluted protein was run on a 10% SDS-Page Gel (Bio-Rad), followed by transfer to a nitrocellulose membrane. The blot was probed with rabbit anti-TRPM7 (1:1000; Sigma-Aldrich, PA5-15302) antibody and visualized with anti-rabbit conjugated to HRP antibody (1:1000; Promega, W401B) using SuperSignal West PicoPlus (Pierce).

All data were analyzed using Excel and GraphPad Prism software (v. 8.01). Statistical analysis for AUC, endocrine mass, proliferation and TUNEL was analyzed using a two-tailed unpaired Student's t-test using GraphPad Prism. Each n represents an individual animal and n values for all experiments are listed in the figure legends. A value of P≤0.05 was considered statistically significant.

We thank Maureen Gannon, Karin Bosma and Peter Kropp (Vanderbilt University Medical Center) for providing discussion and helpful suggestions regarding analysis and quantification of endocrine development as well as the CRE antibody. We thank Sonia Srikanth (Vanderbilt University) for taking pictures of some of the pancreatic sections. We thank the Vanderbilt Hormone Assay Core (supported by National Institutes of Health Grants DK-059637 and DK-020593) for performing insulin secretion immunoassays. We also thank the Vanderbilt Cell Imaging Shared Resource (CISR) for the confocal microscopes used for some of the images included in this manuscript (supported by National Institutes of Health grants DK-58404, C-68485, EY-08126 and DK-59637).

The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.194928