Selenomethionine modulates insulin secretion in the MIN6-K8 mouse insulinoma cell line

Abstract

Selenium is an essential trace element of interest for its potential role in glucose homeostasis. The present study investigated the impact of selenium supplementation as selenomethionine (SeMet) on insulin secretion in MIN6-K8 cells, a pancreatic β-cell model. We found that SeMet enhanced percent glucose-induced insulin secretion, while also increasing tolbutamide- and KCl-induced percent insulin secretion. RNA-sequencing showed that SeMet supplementation altered expression of several selenoproteins, including glutathione peroxidase 3 (Gpx3) and selenoprotein P (SelP). Targeted knockdown of Gpx3 increased both percent and total insulin release, while SelP knockdown increased insulin content and insulin release. Collectively, these studies support a putative role for selenium and selenoproteins in the regulation of insulin secretion, glucose homeostasis, and diabetes risk.

Diabetes mellitus is a multiorgan system disorder defined by elevated levels of blood glucose resulting from insulin insufficiency, insulin resistance, or both. It is currently estimated that 34.2 million people in the United States have diabetes, which is 10.5% of the population [1]. Globally, it is estimated that 463 million adults between the ages of 20 and 79 years have diabetes, with this number projected to rise to 700 million by the year 2045 [2]. As the leading cause of adult blindness, kidney failure, and non-traumatic amputations in the United States and a potent driver of cardiovascular disease, diabetes contributes significantly to individual morbidity and mortality [3]; moreover, the disease exerts a significant economic burden, with diagnosed diabetes alone costing the US economy $327 billion annually [4]. Thus, understanding and addressing the causes and consequences of diabetes is critical for mitigating the devastating toll of this disease. Intriguingly, while the mechanisms of insulin insufficiency and insulin resistance that promote diabetes pathogenesis are incompletely understood, growing evidence suggests that oxidative stress may play a role in diabetes pathogenesis [5], raising the possibility that factors modulating redox state may play an important role in addressing the disorder.

Selenium (Se) is an essential trace element with important functions in antioxidant defense, thyroid hormone metabolism, and immune function that are at least partially mediated through the incorporation of Se into selenoproteins [6]. The human selenoproteome is encoded by 25 genes, while the mouse genome contains 24 [7]. The most well-characterized of the selenoproteins include thioredoxin reductases and glutathione peroxidases, which regulate redox homeostasis, and deiodinases, which control thyroid hormone metabolism [7]. Some of these selenoproteins have been implicated in glucose homeostasis. Indeed, levels of plasma glutathione peroxidase are significantly lower in patients with diabetes [8]. Conversely, overexpression of glutathione peroxidase 1 (GPx1) increases pancreatic insulin content and insulin secretion in mouse islets [9]. Selenoprotein P is the major carrier of Se in the plasma and supplies selenium to peripheral tissues [10]. Serum selenoprotein P levels have been shown to be increased in type 2 diabetes and prediabetes [11]. Intriguingly, Se has also been suggested to have insulin-mimetic properties [12,13]. Importantly, however, there is still controversy regarding the specific role Se plays in the regulation of glucose homeostasis. Some human epidemiological and clinical studies have shown that serum and toenail Se concentrations are negatively correlated with the prevalence of type 2 diabetes mellitus (T2DM), suggesting that higher Se status correlates with lower diabetic risk [14-16]. In animal studies, blood Se levels were shown to decline in diabetic mice [17]. In contrast to these studies, others have documented a positive association between plasma Se levels or dietary Se intake and the prevalence of type 2 diabetes (T2DM) [18-21]. Furthermore, diverse animal models, including mice, rats, and pigs, have shown that elevated dietary Se intake (ranging from 0.4 to 3.0 mg·kg⁻¹ of diet) induces insulin resistance and diabetes-like phenotypes [22]. The Selenium and Vitamin E Cancer Prevention Trial (SELECT) suggested a possible association between long-term Se supplementation and heightened risk of diabetes in men; however, this association did not reach statistical significance (relative risk: 1.07; 99% confidence interval: 0.94–1.22; P = 0.16) [23]. Thus, the relationship between Se and glucose homeostasis remains somewhat inconclusive and additional studies are required to both clarify the relationship between Se and glucose homeostasis as well as to identify the mechanisms by which Se modulates metabolic physiology.

Se naturally exists in both organic and inorganic forms, which exhibit dramatic differences in absorption rates and metabolic pathways [24,25]. Previous studies have shown that organic and inorganic forms of Se have different potencies in their roles regulating cellular function [26]. Half of dietary Se is in the form of selenomethoinine (SeMet), and humans absorbs more than 90% of SeMet (an organic form of Se), while the absorption of Se as inorganic selenite varies, but is generally greater than 50% [25,27]. Interestingly, inorganic forms of Se have been found to induce insulin secretion. For example, Campbell et al. found that sodium selenite increases insulin content and secretion from rat islets while stimulating insulin gene expression in the MIN6 β-cell line [28]. Consistent with this result, 24-h treatment with micromolar concentrations of sodium selenite was also found to augment insulin secretion in MIN6 β-cells [29]. However, there have been no reports on the effects of organic forms of Se on insulin secretion in β-cells. The objective of this study was to investigate the effects of selenomethoinine on β-cell toxicity and function in order to bring us closer to understanding Se’s role and mechanisms of action in the context of glucose homeostasis.

Materials and methods

Cell culture

MIN6-K8 cells were cultured in Dulbecco’s Modified Eagle’s Medium (Sigma-Aldrich, St. Louis, MO #D5796) containing 10% heat-inactivated FBS (Gibco, Grand Island, NY #16000-069), 1 mm sodium pyruvate (Gibco #S8636), and 55 μm β-mercaptoethanol (Sigma-Aldrich #O3446I) at 37 °C with 5% CO₂. Cells were plated in 96-well plates for cell viability assays, while 24-well plates were used for glucose-induced insulin secretion (GIIS) assays as well as RNA extractions. Twenty-four hours after plating, cells (passage 26–30) were incubated in 10% FBS supplemented with 0, 0.4, 4, 20, 40, or 80 μm selenomethoinine (SeMet, Sigma-Aldrich) for 6 days. The cell line was a generous gift from Jun-Ichi Miyazaki of Osaka University. Its derivation, clonal isolation, and basic phenotyping have been described previously [30].

Cell viability and cytotoxicity assay

Live and dead cells were stained using the LIVE/DEAD^™ Viability/Cytotoxicity Kit (Invitrogen, Carlsbad, CA #L3224) according to the manufacturer’s instructions. Briefly, 80,000 cells were plated into each well of 96-cell black-walled, clear-bottomed plates. After cells were treated with SeMet for 6 days, the culture medium was completely removed, and the cells were washed twice with HBSS-HEPES buffer after which 100 μL of HBSS-HEPES containing 1 μm Calcein-AM and 3 μm EthD-1 was added to each well. Cells were incubated for 30 min at 37 °C, and fluorescence was quantified using a BioTek Synergy H1 plate reader. Live cells were quantified by measuring Calcein-AM fluorescence at an excitation wavelength of 485 nm and an emission wavelength of 517 nm; dead cells were quantified by measuring EthD-1 fluorescence at an excitation of 528 nm and an emission wavelength of 617 nm. Live and dead cells were also assessed by microscopy at 100× magnification.

Insulin secretion assay

Insulin secretion assays were performed as previously described [31]. Briefly, MIN6-K8 cells were washed with Krebs–Ringer–bicarbonate HEPES (KRBH) buffer containing 2.8 mm glucose (low glucose) twice, then preincubated for 30 min in KRBH containing low glucose. After preincubation, the cells were stimulated with KRBH containing either different concentrations of glucose (2.8 mm, 11.2 mm, and 16.7 mm), 60 mm KCl, or 100 μm tolbutamide (Sigma Aldrich #T0891) for 30 min. Following incubation, the supernatants were collected for measurement of released insulin. The cells were lysed in 0.1% Triton in KRBH and sonicated for 5 s for determination of cellular insulin content. Both insulin release and content were measured using the CisBio Ultra-sensitive HTRF assay kit (PerkinElmer, Shelton, CT #62IN2PEH). Total DNA was measured using Quant-iT^™ PicoGreen^™ dsDNA Assay Kit (Invitrogen # P11495).

Cytosolic calcium assay

Cytosolic calcium influx was measured using the Fura-2 AM Ca²⁺-sensitive dye (Invitrogen #F1225). Cells were plated in black-walled, clear-bottomed 96-well plates and treated with SeMet for 6 days. Cells were then washed and incubated with 1 μm Fura-2 AM for 30 min at 37 °C. After incubation cells were washed and stimulated with 11.2 mm or 16.7 mm glucose. The fluorescence was measured at 15-s intervals using a BioTek Synergy H1 plate reader. The ratio of emission fluorescence at 510 nm resulting from excitation wavelengths of 340 and 380 nm were used to indicate relative calcium levels.

RNA-seq

Total RNA was extracted using the Qiagen RNeasy kit (Qiagen, Hilden, Germany). RNA-seq and genome-wide transcriptome analyses were performed by Novogene Inc. (Sacramento, CA). Briefly, sequencing libraries were generated using NEBNext Ultra^™ RNA Library Prep Kit for Illumina Inc. (NEB, USA). The clustering of the index-coded samples was performed on a cBot Cluster Generation System using TruSeq PE Cluster Kit v3-cBot-HS (Illumina Inc., San Diego, CA). The prepared libraries were then sequenced on an Illumina HiSeq 4000 platform, and 150 bp paired-end reads were generated; low-quality reads were removed from the sequenced reads. All analyses were based on the clean data. The reads were aligned to the mouse reference genome mm10 using STAR software. The gene expression level was calculated using fragments per kilobase of transcript sequence per millions base pairs sequenced (FPKM). Differential expression analysis was performed using the DEGSeq R package. Genes with an adjusted P value (using q value) <0.05 were considered as differentially expressed (DEG). The RNA-seq data is available in NCBI’s Gene Expression Omnibus (GEO) series (accession number GSE183775).

Real-time PCR

Total RNA was extracted using the Qiagen RNeasy kit (Qiagen). cDNA was synthesized using the qScript cDNA synthesis kit (Quantabio, Beverly, MA) according to the manufacturer’s instructions. Realtime PCR was performed using SYBR Green master mix (Bio-Rad, Hercules, CA). Data were analyzed using the 2-^ΔΔCt method [32]. Data were normalized to the geometric mean of two housekeeping genes (HPRT1 and β-actin). The primers used in this study are listed in Table 1.

Table 1.

Primer sequences for quantitative real-time PCR (qPCR).

Gene	Forward	Reverse
Gpx3	AACGTAGCCAGCTACTGAGGTCTGA	CTGTTTGCCAAATTGGTTGGAAG
Gpx4	GCCTGGATAAGTACAGGGGTT	CATGCAGATCGACTAGCTGAG
SelH	GTGGACAAGCGCGAGAAACT	GTTTGGACGGGTTCACTTGC
SelO	CACTATCGACTATGGACCCTTTG	GGGCTGCTTACTGTATGTGTAG
SelP	TGTTGAAGAAGCCATTAAGATCG	CACAGTTTTACAGAAGTCTTCATCTTC
SelT	GGCCAAGAAAATAAGGTTTAT	ATGGATGGAAGATGTCC
Txnrd1	ATTCAGCAGAGCGGTTCCTC	GCGACATAGGATGCACCAACT
Txnrd2	CATCTTCTGGCTGAAGGAGTC	ACAGTGGTATCCAGTCCAATTC
HPRT1	GGCCAGACTTTGTTGGATTTG	CGCTCATCTTAGGCTTTGTATTTG
β-actin	AGAGGGAAATCGTGCGTGAC	CAATAGTGATGACCTGGCCG

Cell transfection

MIN6-K8 cells were plated at 10⁵ cells-well⁻¹ in 24 well plates, then transfected with 20 nm Gpx3 (Dharmcon, Lafayette, CO; #L-042578-01-0005) or SelP (Dharmcon; #L-041618-01-0005) ON-TARGETplus siRNA containing four specific target sequences for each gene or non-targeting control siRNA at the time of plating. The target sequences for Gpx3 were CCAGAUGGCAUACCGGUUA, AAGACAACUGUGAGCGCGGA, UCUCAAGUAUGUUCGACCA, and CUUAGUGCAUUCAGGCUUA. The target sequences for SelP were UUGGAAGACCUGCGCAUAA, GUGAGGAGAGGUGCGGAAA, CGGAGUGGUACAUAGGAGA, and AGGAAAGCGUGGUGUUAAU. The transfection reagent DharmaFECT2 (Dharmcon; #T-2002-03) was used according to the manufactures’ instruction. GIIS was measured 3 days after transfection.

Statistical analysis

All statistical analyses were performed using GRAPHPAD Prism (version 9.0). An unpaired two-tailed Student’s t-test was used for comparisons between two groups. One-way ANOVA with Dunnett’s multiple comparison tests were used to compare more than two groups. A P-value less than 0.05 was considered statistically significant. For non-monotonic dose-response analysis, SeMet concentrations were log transformed, and data were analyzed using the lm function in R (Vienna, Austria). Data are represented as means with the error bars representing the standard error of the mean (SEM).

Results

SeMet does not induce cytotoxicity in MIN6-K8 cells at the concentrations tested

SeMet cytotoxicity was assessed across a concentration range of 0.4 to 80 μm. The live cells were stained with green fluorescence and dead cells were stained with red fluorescence as shown in Fig. 1A. The quantification of fluorescence intensity for live and dead cells are shown in Fig. 1B,C, respectively. At the concentrations tested, SeMet treatment for 6 days did not induce cytotoxicity as assessed by a decrease in live cells or an increase in dead cells.

Fig. 1.

Effects of selenomethoinine supplementation (SeMet) on cell viability. MIN6-K8 cells were exposed to SeMet at the indicated concentrations for 6 days. (A) Representative microscopy images following staining with either Calcein-M (green, live cells) or EthD-1 (red, dead cells) (100× magnification). Quantitative results of (B) live cell staining and (C) dead cell staining. The results are represented as means ± SEM, n = 18.

Higher concentrations of SeMet increased insulin secretion in MIN6-K8 cells

To study the impact of selenium supplementation on β-cell function, cells were treated with SeMet across a wide concentration range for 6 days. Insulin secretion was measured at 2.8 mm glucose, 11.2 mm glucose, and 16.7 mm glucose concentrations. At 40 and 80 μm SeMet, %insulin release was significantly increased at basal glucose levels (2.8 mm glucose) and upon stimulation with 11.2 mm glucose (Fig. 2A,B, respectively). With high glucose stimulation (16.7 mm glucose), %insulin secretion was decreased by 21% at 4 μm SeMet, but again increased as SeMet concentrations increased further. Specifically, 40 μm and 80 μm SeMet significantly enhanced %insulin secretion by 38% and 77%, respectively (Fig. 2C). Statistical analyses showed that SeMet increased %insulin release with a non-monotonic, dose-response relationship. Insulin release normalized to total DNA is shown in Figs. 2D,E,F. At 80 μm, SeMet significantly increased absolute insulin release at both 2.8 mm and 16.7 mm glucose stimuli. No changes were observed at 11.2 mm glucose stimulation. The total insulin content normalized to total DNA content was decreased by 80 μm SeMet at all glucose conditions; however, lower concentrations of SeMet did not impact insulin content.

Fig. 2.

Effects of SeMet on insulin secretion. MIN6-K8 ells were treated with selenomethionine (SeMet) at the indicated concentrations for 6 days. Insulin secretion was measured at (A) 2.8 mm glucose, (B) 11.2 mm glucose, and (C) 16.7 mm glucose. (Top) percentage of insulin released, (Middle) insulin released per μg of total DNA, and (Bottom) total insulin content per μg of DNA. The results are means ± SEM, n = 16, *P < 0.05, **P < 0.01, ***P < 0.001 versus 0 μm SeMet controls.

SeMet reduces Ca²⁺ influx at high concentrations

To determine if SeMet modulates insulin secretion through an effect on the ATP-sensitive K⁺ (K_ATP) channel, the effects of SeMet on insulin secretion were assessed after K_ATP channel closure by treatment with tolbutamide or by direct membrane depolarization with potassium under low glucose conditions. Similar to the effects observed with glucose stimulation, the % insulin secretion induced by tolbutamide was enhanced by treatment with 40 μm and 80 μm SeMet (Fig. 3A, P < 0.01). Normalization to total DNA content revealed that 80 μm SeMet increased absolute insulin secretion upon tolbutamide stimulation (Fig. 3C, P < 0.01). The %insulin secretion by potassium was increased by 80 μm SeMet (Fig. 3B, P < 0.01), while absolute insulin secretion was decreased by 4 and 20 μm SeMet and increased by 80 μm SeMet (Fig. 3D, P < 0.01). Insulin content was unchanged (Fig. 3E,F, respectively).

Fig. 3.

Effects of SeMet on tolbutamide and potassium responsiveness. Cells were treated with SeMet at the indicated concentrations for 6 days. Insulin secretion and content were measured at (A, C, E) 100 μm tolbutamide, and (B, D, F) 60 mm KCl. (Top) percentage of insulin released, (Middle) insulin released per μg of total DNA, (Bottom) total insulin content per μg of DNA. The results are mean ± SEM, n = 16, *P < 0.05, **P < 0.01, ***P < 0.001 as compared with the 0 μm SeMet.

To ascertain whether SeMet affected insulin secretion by modulating calcium influx, the FURA-2AM calcium-sensitive dye was loaded into SeMet-treated MIN6-K8 cells, and calcium flux was quantified. As expected, addition of 11.2 mm and 16.7 mm glucose significantly increased intracellular calcium; however, SeMet concentrations at or below 40 μm had no effect on calcium influx (Fig. 4A,B), whereas 80 μm SeMet actually inhibited calcium influx and delayed the time to reach the peak of calcium at 11.2 mm glucose (Fig. 4C,E, respectively). There were no significant differences in calcium dynamics at 2.8 mm glucose (data not shown). These data suggest that the mechanism(s) by which SeMet augments insulin release are either calcium-independent or downstream of calcium influx.

Fig. 4.

Effects of SeMet on Ca²⁺ influx. MIN6-K8 cells were treated with SeMet for 6 days. Calcium changes were monitored during stimulation with 11.2 mm (A) or 16.7 mm glucose (B). Area under the curve at 11.2 mm (C) and 16.7 mm glucose (D). Time to peak calcium signal at 11.2 mm (E) and 16.7 mm glucose (F). The results are means ± SEM, n = 12, *P < 0.05 as compared with the 0 μm SeMet.

SeMet altered selenoprotein gene expression

To better understand how SeMet promoted insulin secretion, RNA-seq was performed on control and 40 μm SeMet-treated MIN6-K8 cells. A total of 3278 genes were differentially expressed between control and SeMet-treated cells. 1541 genes were up-regulated and 1737 genes were down-regulated with an adjusted P < 0.05 (data not shown). Based on the fact that Se exerts its biological effects on cell biology through its incorporation into selenoproteins, we specifically examined the impact of SeMet supplementation on the transcription of these proteins. Transciptomics revealed that SeMet altered expression of eight selenoprotein genes (Fig. 5A). The log-fold changes and P-value of each gene are listed in Table 2. Using qRT-PCR, we confirmed that selenoprotein H (SelH) and thioredoxin reductase 1 (Txnrd 1) were upregulated by SeMet supplementation, while glutathione peroxidase 3 (Gpx3), selenoprotein O (SelO), and selenoprotein P (SelP) were down-regulated by SeMet treatment (Fig. 5B). Importantly, mRNA expression of both Gpx3 and SelP were decreased by SeMet in a dose-dependent manner (Fig. 5C,D, respectively).

Fig. 5.

Selenoprotein genes changed by SeMet. Heat map of selenoprotein mRNA expression measured by RNA-seq (A). Gene expression qRT-PCR verification of selenoprotein transcription (B). Dose-dependent mRNA expression changes of Gpx3 (C) and SelP by SeMet (D). The dose dependent mRNA results are means ± SEM, n = 8. **P < 0.01, ***P < 0.001 compared with 0 μm SeMet.

Table 2.

mRNA expression of selenoprotein genes changed by SeMet measured using RNA sequencing.

	Gene_ID	GeneName	log2FoldChange	P-adj
Down-regulated	ENSMUSG00000018339	Gpx3	−1.28	2.69E-12
	ENSMUSG00000035757	SelO	−0.44	2.99E-04
	ENSMUSG00000075704	Txnrd2	−0.24	2.87E-04
	ENSMUSG00000064373	SelP	−0.23	3.65E-06
	ENSMUSG00000075706	Gpx4	−0.18	5.63E-03
	ENSMUSG00000075700	SelT	−0.14	1.77E-02
Up-regulated	ENSMUSG00000076437	SelH	0.17	3.84E-03
	ENSMUSG00000020250	Txnrd1	0.29	2.15E-13

Knocking down Gpx3 increased insulin release

To ascertain whether Gpx3 and SelP mediated, at least in part, the SeMet-induced changes in GIIS, both Gpx3 and SelP were knocked down in MIN6-K8 cells using a siRNA-based approach. Knocking down Gpx3 expression by 80% (Fig. 6A) did not change insulin secretion or content at 2.8 mm glucose condition (Fig. 6C,E, and G); however, Gpx3 knockdown increased both percent and absolute insulin release at 16.7 mm glucose (Fig. 6D,F). In contrast, knocking down SelP expression by 88% (Fig. 6B) decreased % insulin secretion (Fig. 6C) under basal glucose conditions (2.8 mm), an effect that was due to increased insulin content (Fig. 6G). At 16.7 mm glucose, both absolute insulin release and insulin content were increased by SelP knockdown (Fig. 6F,H, respectively). Because of the concordant increase in insulin content, percent insulin release was unchanged (Fig. 6D).

Fig. 6.

Insulin secretion following manipulation of selenoprotein expression. Cells were transfected with either siRNA targeting Gpx3 or SelP, or with non-targeting siRNA (NT) and allowed to incubate for 3 days. qPCR verification of Gpx3 and Selp mRNA knockdown are shown in A and B. The percentage insulin release, insulin release, and insulin content normalized to DNA were measured at 2.8 mm glucose (C, E, G) and 16.7 mm glucose (B, F, H). The percent insulin released is shown in C and D. Absolute insulin release per μg of total DNA is shown in E and F. Total insulin content per μg of DNA is shown in G and H. The results are means ± SEM, n = 12, *P < 0.05, **P < 0.01, ****P < 0.0001 versus the NT group.

Discussion

Se is a trace element that plays critical roles in antioxidant defense, thyroid hormone metabolism, and immune function; it may also have anti-cancer activity [33]. Intriguingly, Se status has been linked to metabolic function; however, the precise nature of its role in glucose homeostasis remains incompletely understood. Despite this, it is estimated that 28% of patients with diabetes in the United States use selenium supplements, a prevalence that has not significantly changed from 1999 to 2014 [34]. Although selenium supplementation has been widely used, the effect of selenium on diabetes is controversial. The present study was intended to extend our knowledge regarding the impact of selenium on insulin secretion in pancreatic β-cells. In this work, we showed that SeMet supplementation augmented both basal and stimulated insulin secretion. Moreover, this effect was associated with changes in the expression of specific selenoproteins. Indeed, targeted knockdown of Gpx3 augmented glucose-induced insulin secretion at high glucose, while knockdown of SelP increased total insulin release and content. Collectively, these data suggest that selenium and selenoproteins may be important modulators of β-cell function and thus may play a role in mediating diabetes risk.

Our data are largely consistent with the available literature. Steinbrenner et al. [35] examined the cytotoxicity of SeMet on the rat INS-1 insulinoma cell line using the MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) assay and found that SeMet did not impair metabolic activity up to concentrations of 100 μm during a 24-h incubation. In our study, more prolonged exposure (6 days) of MIN6-K8 cells to SeMet revealed that cell viability was not significantly changed up to 80 μm SeMet. Collectively, our study alongside the published literature suggests that SeMet does not appreciably disrupt cell viability well into the micro-molar range.

In this study, we have shown that 4 μm SeMet decreased percent insulin secretion in the presence of 16.7 mm glucose, while higher SeMet concentrations (40 and 80 μm) increased the percentage of insulin secretion at both basal and high glucose concentrations. At the highest SeMet concentration, the increase in percent insulin release was due to both increases in insulin release and decreases in insulin content. This suggests that at high SeMet concentrations, changes in insulin secretory dynamics may be dependent upon multiple mechanisms. Because diabetes can be defined as a state of relative or absolute insulin deficiency, our data indicating that high SeMet levels augment insulin release suggest that such treatment may have potentially beneficial effects on β-cell function in those at risk for diabetes. Indeed, there are studies demonstrating protective effects of selenium on diabetes both in vitro and in vivo. For example, supranutritional doses of selenate given to db/db mice have been shown to exert anti-diabetic effects by improving whole body insulin sensitivity and reducing the activity of liver cytosolic protein tyrosine phosphatases (PTPs), negative regulators of insulin signaling [36]. While another study showed that dietary supplementation with SeMet lowered blood glucose and improved insulin sensitivity in KKAy mice [37].

Despite this evidence, however, the impact of Se on glucose homeostasis remains controversial. In our own studies, dietary selenium deficiency resulted in a reduction in fasting glucose and insulin with a trend toward improved glucose tolerance [38]. Stranges et al. [39] found that long-term selenium supplementation did not reduce and may have increased the risk of type 2 diabetes. The SELECT also suggested a possible association between long-term Se supplementation and heightened risk of diabetes in men, although this association did not reach statistical significance [23]. The duration, timing, and concentration of exposure to Se may all contribute to the disparate results from studies examining selenium’s role in glucose homeostasis. Our data also showed that high concentrations of SeMet increased insulin secretion in MIN6-K8 cells at low glucose levels (2.8 mm), an effect akin to increasing basal insulin levels. Indeed, this may be problematic at the whole organism level as higher basal insulin levels can result in a downregulation of insulin receptor expression and signal transduction, effectively inducing insulin resistance in peripheral tissues and thereby increasing the risk of prediabetes and type 2 diabetes [40-42]. Consistent with this concept, one study in rats showed that supra-nutritional selenium supplementation induced hepatic insulin resistance through a reduction in reactive oxygen species signaling [43], while another revealed that dietary supplementation with selenium can induce insulin resistance [44]. Induction of insulin resistance by insulin-induced insulin resistance resulting from Se-mediated augmentation of insulin secretion is one potential explanation that could link these data. Finally, it has been suggested that the impact of selenium on glucose homeostasis and diabetes risk is likely best characterized by a U-shaped relationship [45]. Clearly, further studies are required to better understand the effects of selenium on insulin secretion and insulin resistance in vivo as well as their impact on diabetes risk.

Campbell et al. reported that sodium selenite, an inorganic form of selenium, increased insulin secretion in rat islets, while also increasing insulin gene expression and protein content [28]. While we similarly observed an increase in insulin secretion, insulin content was unchanged or even decreased at the highest levels of SeMet supplementation in the present study. This difference may be explained by the form of the selenium (SeMet versus selenite), the length of the treatment (6 days vs. 24 hours), or unique aspects of the different model systems (MIN6-K8 vs. INS-1 cells). Regardless, the available evidence suggests a role for selenium in modulating β-cell physiology.

The ATP-sensitive K⁺ (K_ATP) channel is central to the regulated release of insulin from pancreatic β-cells. Physiological closure of the K_ATP channel by rising ATP levels arising from glucose metabolism leads to β-cell depolarization and subsequent insulin release. Treatment of MIN6-K8 cells with tolbutamide completely closes the K_ATP channel, allowing us to evaluate whether the SeMet effect results from augmentation of K_ATP channel closure or as part of the secretory phase of the stimulus-coupling mechanism. Closure of the K_ATP channel by tolbutamide treatment did not prevent the SeMet-mediated augmentation of insulin release, indicating that SeMet effects are mediated downstream of K_ATP channel closure. Membrane depolarization opens voltage gated Ca²⁺channels, allowing Ca²⁺ influx into the cells with consequential release of insulin granules. To ascertain whether Ca²⁺ flux was altered by SeMet, we first investigated the impact of SeMet on potassium-induced insulin secretion. The result showed that SeMet-mediated increases in insulin secretion persisted at 80 μm but disappeared at lower SeMet concentrations. We further measured Ca²⁺ flux specifically, and we found that Ca²⁺influx was actually reduced at 80 μm and unchanged at lower SeMet concentrations. Collectively, these data suggest that SeMet likely modulates insulin secretion independent of calcium or downstream of calcium influx.

To further study the mechanisms and potential mediators by which higher concentrations of SeMet increase insulin secretion, RNA-sequencing was performed in control and 40 μm SeMet-treated MIN6-K8 cells. Since selenium functions primarily through its incorporation into selenoproteins in mammalian systems [46], targeted analyses were performed to identify changes in the expression of selenoproteins among those 3,000 genes. As shown, eight selenoproteins were shown to exhibit altered expression by SeMet treatment, and five of them were confirmed by qRT-PCR. Gpx3 was the gene whose expression was most significantly changed. In vivo, Gpx3 is the only extracellular subtype of the Gpx family, and it is most highly expressed by kidney [47]; however, it is also highly expressed in the pancreas, lung, heart, liver, and skeletal muscle [47]. In a gene expression analysis of mouse islets, Gpx3 was the most highly expressed selenoprotein transcript in both males and females [48]. Several clinical trials had reported single-nucleotide polymorphisms (SNP) in the Gpx3 gene promoter region are associated with cardiovascular disease [49-51]. Serum levels of Gpx3 were increased in patients with metabolic syndrome, and the concentration of Gpx3 was correlated with insulin sensitivity [51]. The role of Gpx3 in insulin secretion, however, has been unknown. Intriguingly, despite being a selenoprotein, in this study we observed a pronounced dose-dependent decrease in Gpx3 gene expression with increasing SeMet supplementation. Furthermore, knocking down Gpx3 in MIN6-K8 cells increased insulin secretion under high glucose conditions, mimicking the effects of SeMet supplementation. This suggests that SeMet may exert its effects at least partially by down-regulating Gpx3 expression.

SelP is another secreted selenoprotein that is enriched in the circulation in vivo. Intriguingly, circulating SelP has been found to be positively associated with fasting glucose [52]. Specifically, serum SelP levels are higher in patients with type 2 diabetes or prediabetes [11]. While liver is the primary source of plasma SelP, it is also widely expressed in other murine tissues [53]. In rodent pancreas, SelP expression is restricted to the islets [35]. Mita et al. reported that insulin secretion was significantly decreased by excess exogenous human SelP in MIN6 cells, and the addition of SelP neutralizing antibody improved insulin secretion induced by high levels of glucose [54]. In the present study, SeMet supplementation decreased SelP gene expression in MIN6-K8 cells, raising the intriguing question of whether SelP downregulation is directly involved in SeMet-mediated augmentation of GIIS. By knocking down SelP using siRNA, we observed increases in both total insulin release and insulin content. In the context of the available literature, these data indicate that SelP modulates β-cell physiology, potentially in an autocrine or paracrine fashion.

Collectively, our data suggest that SeMet supplementation augments insulin release in the MIN6-K8 cell line and that these effects are likely mediated through multiple selenoproteins, including Gpx3 and SelP. Furthermore, these data indicate a role for selenium and selenoproteins in the regulation of glucose homeostasis. Despite these intriguing findings, the present studies have several limitations. The central limitation of the present study is that all the experiments were performed in MIN6-K8 cells, which may not fully model SeMet effects in vivo. More studies are clearly required to better understand the effects of selenium on insulin secretion and insulin resistance in vivo. Additionally, the SeMet concentrations studied herein are higher than physiological Se concentrations in plasma; however, it is known that some tissues can concentrate Se at levels up to 10-fold higher than plasma concentrations [55]. Further work is required for generating a cell culture model that fully recapitulates the in vivo state. Another limitation of this work is that we focused on the effects of knocking down Gpx3 and SelP on insulin secretion. The effects of overexpression of both selenoproteins with SeMet treatment will need to be assessed to better understand the mechanisms. Additionally, the role of SelO, SelH, and Txnrd1 in regulating insulin secretion also merit further evaluation and are actively under study.

In conclusion, high SeMet supplementation increased insulin secretion. This effect was likely not induced by alterations in K_ATP channel activity nor by changes in Ca²⁺ flux. Decreased Gpx3 and SelP expression caused by SeMet supplementation may mediate some of these observed effects on insulin secretion. Further studies are required to better characterize the complex relationships among selenium, selenoproteins, and glucose homeostasis to better understand the impact of selenium supplementation and status as well as selenoprotein variants on glucose homeostasis and diabetes risk.

Abbreviations

GIIS

glucose-induced insulin secretion

Gpx1

glutathione peroxidase 1

Gpx3

glutathione peroxidase 3

Gpx4

glutathione peroxidase 4

HPRT1

hypoxanthine phosphoribosyltransferase 1

KRBH

Krebs-Ringer-bicarbonate HEPES buffer

Se

selenium

SelH

selenoprotein H

SelO

selenoprotein O

SelP

selenoprotein P

SelT

selenoprotein T

SeMet

selenomethoinine

T2DM

type 2 diabetes mellitus

Txnrd1

thioredoxin reductase 1

Txnrd2

thioredoxin reductase 2

Data accessibility

The data that support the findings of this study are available from the corresponding author (rsargis@uic.edu) upon reasonable request. The RNAseq data are available in NCBI’s Gene Expression Omnibus (GEO), accession number: GSE183775.