Absence of Slc39a14/Zip14 in mouse pancreatic beta cells results in hyperinsulinemia

Abstract

Zinc is an essential component of the insulin protein complex synthesized in β cells. The intracellular compartmentalization and distribution of zinc are controlled by 24 transmembrane zinc transporters belonging to the ZnT or Zrt/Irt-like protein (ZIP) family. Downregulation of SLC39A14/ZIP14 has been reported in pancreatic islets of patients with type 2 diabetes (T2D) as well as mouse models of high-fat diet (HFD)- or db/db-induced obesity. Our previous studies observed mild hyperinsulinemia in mice with whole body knockout of Slc39a14 (Zip14 KO). Based on our current secondary data analysis from an integrative single-cell RNA-seq dataset of human whole pancreatic tissue, SLC39A14 (coding ZIP14) is the only other zinc transporter expressed abundantly in human β cells besides well-known zinc transporter SLC30A8 (coding ZnT8). In the present work, using pancreatic β cell-specific knockout of Slc39a14 (β-Zip14 KO), we investigated the role of SLC39A14/ZIP14-mediated intracellular zinc trafficking in glucose-stimulated insulin secretion and subsequent metabolic responses. Glucose-stimulated insulin secretion, zinc concentrations, and cellular localization of ZIP14 were assessed using in vivo, ex vivo, and in vitro assays using β-Zip14 KO, isolated islets, and murine cell line MIN6. Metabolic evaluations were done on both chow- and HFD-fed mice using time-domain nuclear magnetic resonance and a comprehensive laboratory animal monitoring system. ZIP14 localizes on the endoplasmic reticulum regulating intracellular zinc trafficking in β cells and serves as a negative regulator of glucose-stimulated insulin secretion. Deletion of Zip14 resulted in greater glucose-stimulated insulin secretion, increased energy expenditure, and shifted energy metabolism toward fatty acid utilization. HFD caused β-Zip14 KO mice to develop greater islet hyperplasia, compensatory hyperinsulinemia, and mild insulin resistance and hyperglycemia. This study provided new insights into the contribution of metal transporter ZIP14-mediated intracellular zinc trafficking in glucose-stimulated insulin secretion and subsequent metabolic responses.

NEW & NOTEWORTHY Metal transporter SLC39A14/ZIP14 is downregulated in pancreatic islets of patients with T2D and mouse models of HFD- or db/db-induced obesity. However, the function of ZIP14-mediated intracellular zinc trafficking in β cells is unknown. Our analyses revealed that SLC39A14 is the only Zn transporter expressed abundantly in human β cells besides SLC30A8. Within the β cells, ZIP14 is localized on the endoplasmic reticulum and serves as a negative regulator of insulin secretion, providing a potential therapeutic target for T2D.

INTRODUCTION

Insulin resistance (IR) represents one of the common complications of obesity and leads to the development of type 2 diabetes (T2D) (1–3). IR is associated with glucose intolerance by the peripheral tissues and hyperinsulinemia, the compensatory increase in insulin production by pancreatic β cells (4). Uncontrolled IR can proceed to T2D, where the chronic overload of insulin synthesis and secretion eventually causes pancreatic β-cell dysfunction and death (4). Metabolism and intracellular trafficking of zinc are important for pancreatic β-cell function. The pancreas has the highest concentration of zinc (140 µg/g) compared with other tissues in the human body (5), in part due to the fact that zinc is an essential component of the insulin protein complex synthesized in β cells (6). A better understanding of zinc metabolism in the context of β-cell biology may advance our knowledge of T2D pathogenesis and potentially help the development of novel treatment strategies against T2D.

The intracellular compartmentalization and distribution of zinc is controlled by intracellular zinc-binding metallothionein proteins, as well as 24 transmembrane zinc transporters that belong to either the ZnT or Zrt/Irt-like protein (ZIP) family (7, 8). Several studies have uncovered functional impacts of specific ZnT members, namely, ZnT3 (9), ZnT5 (10), ZnT7 (11), and ZnT8 (12–14), on β-cell health and function. For example, mice β cell-specific knockout of Slc30a8 (coding ZnT8) display defects in zinc contents in β cells, insulin crystallization failure, and the presence of atypical insulin granules (13, 14). In addition, a genetic variant at a SLC30A8 locus (R325W) identified by human genome-wide association studies (GWAS) has a strong linkage to T2D (15–18). On the other hand, several ZIP family members were also studied in the context of β cell biology and/or systemic glucose homeostasis. Mice with β cell-specific knockout of Slc39a5 (ZIP5) demonstrated a significant reduction in glucose-stimulated insulin secretion (GSIS) and glucose intolerance compared with control mice under both chow and HFD-fed conditions (19). Although ZIP4 was reported to positively regulate insulin secretion in vitro (20), mice with β cell-specific knockout of Slc39a4 (ZIP4) paradoxically displayed unaltered insulin secretion but improved glucose tolerance compared with control (20). In addition, the roles of other ZIP protein family members, including ZIP6 (21), ZIP7 (21) and ZIP14 (22), have been implicated in pancreatic β cell lines. However, whether these ZIP proteins in β cells exert significant impacts at the physiological level or contribute to IR/T2D pathogenesis remains unclear.

In the present study, we aimed to better understand the role of ZIP14 in pancreatic β cell function and its contribution to IR/T2D pathogenesis for several reasons. First, studies have shown that SLC39A14 (coding ZIP14) is enriched in pancreatic tissues compared with other peripheral tissues in humans and sorted β cells compared with whole islets in mice (23), suggesting an important role of ZIP14 in β cells. Second, there are translational implications of β cell ZIP14 in the disease context of obesity and T2D. Downregulation of SLC39A14 has been reported in pancreatic islets of patients with T2D (23) as well as mouse models of HFD-induced or db/db-induced obesity (19). Third, our previous studies observed hyperinsulinemia in mice with whole body knockout of Slc39a14 (Zip14 KO) (24–26). However, whether the hyperinsulinemia phenotype of whole body Zip14 KO mice was due to β cell-specific effects or secondary effects of whole body KO of Zip14 remains unclear.

To investigate the role of ZIP14 in pancreatic β cell function, we generated a mouse model with pancreatic β cell-specific knockout of Slc39a14 (β-Zip14 KO). By performing in vivo, ex vivo, and in vitro assays using β-Zip14 KO and murine cell line MIN6 in this study, we showed that ZIP14 regulates intracellular zinc trafficking in β cells and serves as a negative regulator of insulin secretion. Furthermore, we demonstrated that β-Zip14 KO mice developed diet-induced metabolic dysfunctions such as mild IR and hyperglycemia, which are not present in whole body Zip14 KO mice. Taken together, this study provided new insights into β cell ZIP14 regulation of insulin secretion and its role in diet-induced IR/T2D pathogenesis.

METHODS

Animals

To generate whole body ZIP14 knockout mice, heterozygous (ZIP14^+/−) mice of the C57BL/6;129S5 strain were originally obtained from the Mutant Mouse Research Resource Consortium at The University of California-Davis, and a breeding colony was established at Cornell University (25, 27). Transposagen Biopharmaceuticals generated floxed ZIP14 mice. Founder ZIP14^flox/+ (LoxP in introns 4 and 8) mice of the C57BL6 strain were bred to obtain ZIP14^flox/flox (ZIP14^fl/fl) mice (28, 29). Following genotype confirmation, ZIP14^fl/fl mice were crossed with Ins1^cre [(B6(Cg)-Ins1^{tm1.1(cre)Thor}/J-Jackson Lab stock # 026801)] mice to create β cell-specific Zip14 KO mice (β-Zip14 KO). Genotyping was performed on genomic DNA extracted from tail tissues, using forward primer 5′-GTC AAA CAG CAT CTT TGT GGT C-3′ and reverse primer 5′-GGA AGC AGA ATT CCA GAT ACT TG-3′ amplifying a 488-bp product in ZIP14^fl/fl mice and forward primer 5′-GCT GGA AGA TGG CGA TTA GC-3′ and reverse primer 5′-GGA AGC AGA ATT CCA GAT ACT TG-3′ amplifying a 675-bp product in β-Zip14 KO mice. ZIP14^fl/fl mice were used as control. Mice from both sexes were used in this study. Mice were maintained using standard rodent husbandry and received a commercial, irradiated chow diet (Envigo 7912; containing 60 mg Zn/kg). In high-fat diet (HFD) studies, mice were randomized to be fed with either HFD (Research Diets, #D12492-60 kcal% Fat) or chow diet for 16 wk. Except for dietary studies, mice 8–16 wk of age were used for experiments. Mice were anesthetized by isoflurane inhalation for injections or gavage procedures and euthanized by cardiac puncture. Tissue samples were immediately flash-frozen in liquid nitrogen at the collection and stored at −80 C. All animal procedures were performed with the approval and authorization of the Institutional Animal Care and Use Committee at Cornell University. Cornell University operates its Animal Care and Use program under the Animal Welfare Assurance A3347-01 on file with the Office of Laboratory Animal Welfare (OLAW). Mice were monitored daily for signs of animal pain and distress and none were observed.

Metabolic Phenotyping

Metabolic parameters of the mice were measured by the Comprehensive Laboratory Animal Monitoring System (CLAMS) using Promethion (Sable Systems) as previously described (30). Mice were transferred to the CLAMS metabolic cages available under a CC-BY-NC-ND 4.0 International license and allowed to acclimate for 2–3 days with free access to food and water. Following the acclimation, measurements of V̇o₂ and V̇co₂ gas exchanges and food and water consumption were collected for 5 days. The data were analyzed using the Expedata software system (v1.9.27), Macro 264 interpreter (v2.40), and CalR software (v1.2).

Body Composition

Measurements were conducted in awake mice by time-domain nuclear magnetic resonance using the Minispec LF65 Body Composition Mice Analyzer.

Oral Glucose Tolerance and Insulin Tolerance Test

Following overnight fasting, blood glucose levels were measured from tail bleeding using OneTouch Ultra Blood Glucose Monitoring System immediately before and at 5, 15, 30, 90, and 180 min after oral gavage of glucose (3 mg/kg). For the insulin tolerance test (ITT), mice were fasted for 4 h and blood glucose levels were measured immediately before and after an intraperitoneal injection of insulin (0.75 U/kg).

Pancreatic and Serum Insulin Measurements

The whole pancreas samples were weighed and minced in the presence of 0.2 M acetic acid in 75% (vol/vol) ethanol. Blood was collected at specific time points by cardiac puncture and serum was obtained by centrifugation at 3,000 g for 15 min. Insulin measurements using insulin ELISA (Mercodia, #10–1113-01) or Ultrasensitive Insulin ELISA (Mercodia, #10–1132-01). ELISA results from the pancreas were normalized to tissue weight or DNA content. Serum IL-6 levels were determined using ELISA (ThermoFisher Scientific, #BMS603HS) according to the manufacturer’s protocol.

Mouse Islet Isolation

Mouse islets were isolated following the procedures described previously with minor modifications (31, 32). Briefly, mice were fasted 4 h before the isolation. For islet isolation, the pancreas was perfused through the common bile duct with Hank’s balanced salt solution (HBSS) containing 0.02% BSA and 0.75 mg/mL collagenase P (Sigma-Aldrich, # 11213857001). The inflated pancreas was carefully removed and incubated at 37°C for 13–17 min. The islets were separated and purified by Ficoll density gradient solution (Cytiva, # 17030010). The islets were hand-picked under a dissecting microscope. The islets were recovered overnight in an RPMI 1640 medium (Gibco, # 11879020) in the presence of 10% FBS (Gibco, #10437-028), 100 U/mL penicillin, and 100 μg/mL streptomycin (Gibco, # 15140122) in a 5% CO₂ humidified incubator.

MIN6 Cell Culture

The adherent insulinoma β-cell line MIN6 cells (33) (Addexbio Technologies, San Diego, CA) were cultured in high glucose Dulbecco’s modified Eagle’s medium (Cytiva, #SH30243.01) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, 100 μg/mL streptomycin, 25 mM HEPES, and 0.05 mM β-mercaptoethanol. The cells were maintained at 37°C in a humidified atmosphere of 95% air and 5% CO₂. All assays used MIN6 cells grown to 70–80% confluence and under 20 passages to ensure robust GSIS responses (34). To overexpress ZIP14 in MIN6 cells, cells were transfected with pcDNA3.1+/C-(K)-DYK vector carrying Slc39a14 sequence (XM_006518789.3) or empty vector (as control) (GenScript, Piscataway, NJ). The transfection experiments were performed using Effectene Transfection Reagent (Qiagen, # 301425) according to the manufacturer’s protocol. To knock down ZIP14 in MIN6 cells, cells were transfected with either control or the siRNA targeting Zip14 (ThermoFisher Scientific #s102780) using Lipofectamine RNAiMAX Transfection Reagent (ThermoFisher Scientific # 13778075) according to the manufacturer’s protocol. To isolate crude endoplasmic reticulum (ER) fraction, the protocols were modified (35, 36). MIN6 cells were scraped into PBS and collected by centrifugation at 200 g for 5 min. The pellets were gently resuspended in buffer 1 (10 mM NaCl, 1.5 mM MgCl₂, 10 mM Tris-HCl, pH 7.5) and centrifuged at 370 g (repeated 3 times). The pellets were homogenized in buffer 2 (10 mM Tris-HCl, pH 6.7, 10 mM KCl, 0.15 mM MgCl₂, 1 mM PMSF, 1 mM DTT) with a Teflon pestle (glass/Teflon Potter- Elvehjem) by using 10 strokes. The homogenate was transferred to a tube containing 2 M sucrose solution. The resulting supernatant from centrifugation at 1,200 g three times was centrifuged at 7,000 g for 10 min and 20,000 g for 30 min. The supernatant was centrifuged at 100,000 g for an hour to obtain crude ER fraction in the pellet and crude cytosolic fraction in the supernatant.

Glucose‐Stimulated Insulin Secretion

To perform GSIS in MIN6 cells, cells were seeded in 96-well plates (5 × 10⁴ cells/well) until 80% confluence for GSIS assay. Cells were washed with PBS and incubated basal serum-free media [DMEM (Gibco, #A14420) containing 2 mM glucose, 100 U/mL penicillin, 100 μg/mL streptomycin, 2 mM l-glutamine, 1 mM sodium pyruvate, 25 mM HEPES, and 0.05 mM β-mercaptoethanol] overnight. Cells were incubated with basal serum-free media containing 5 μM FluoZin-3 AM (ThermoFisher Scientific, #F24195) for an hour and then treated with 20 mM glucose and/or 40 mM KCl to induce cell polarization and insulin secretion. To enhance live FluoZin-3 AM signal 10 μM ZnCl₂ was included in all treatment conditions. FluoZin-3 AM intensity was measured by SpectraMax M3 reader (Molecular Devices, San Jose, CA). Protein concentration (determined by Pierce B.C.A. assay, ThermoFisher Scientific # 23227) or nuclei staining (determined by Hoechst 33348, ThermoFisher, Waltham, MA) was used to normalize FluoZin-3 AM readout of MIN6 experiments. In parallel, for ex vivo islet experiments, 10 islets per time point were used for GSIS and 40 islets for FlouZin-3 AM experiments. DNA content extracted from islets was used to normalize FluoZin-3 AM readout of ex vivo islet studies. To measure insulin concentration per well, media was taken from specific timepoints during GSIS, and insulin concentration was determined by insulin ELISA (Mercodia, #10–1113-01).

Zinc Measurements by Microwave Plasma Atomic Emission Spectrometer

Pancreas tissue was collected and digested (∼50 mg) in HNO₃ overnight at 80°C and diluted in deionized water. Cell pellets from MIN6 cells were digested in HNO₃ overnight at 80°C and diluted in deionized water. Zinc concentrations were measured by microwave plasma atomic emission spectrometer (MP-AES) at 213.857 nm. Absorbance values for tissue and cellular zinc were normalized to the wet tissue weight and total protein concentrations, respectively.

⁶⁵Zn Efflux Assay

To measure ⁶⁵Zn efflux, we incubated cells with ⁶⁵Zn (0.18 µci/mL) (Eckert&Ziegler, Valencia) in [DMEM (Gibco, #A14420) serum-free basal media containing 10 mM ZnCl₂, 2 mM glucose, 100 U/mL penicillin, 100 μg/mL streptomycin, 2 mM l-glutamine, 1 mM sodium pyruvate, 25 mM HEPES, and 0.05 mM β-mercaptoethanol] overnight. The next day, we removed the media and washed the cells with chelating buffer (10 mM EDTA, 10 mM HEPES, and 0.9% NaCl₂) to remove any remaining ⁶⁵Zn. Then, we treated cells with either glucose alone or combined with KCl for 20 min. Radioactivity was measured both in the media and the cells by gamma counter, and the amount of radioactivity in media is normalized to the amount of radioactivity in media.

Western Blotting

All the buffers were supplemented with protease inhibitors (AGScientific) during isolation. The final lysates were supplemented with cocktails of protease ThermoFisher Scientific # 1860932) and phosphatase (ThermoFisher Scientific # 78428) inhibitors and PMSF (Boston Bioproducts # BP481). Solubilized proteins were separated by 10 and 12% SDS-PAGE. Visualization was by chemiluminescence (SuperSignal, Thermo Fisher, 34580) and digital imaging (Protein Simple, San Jose, CA). Rabbit anti-mouse ZIP14 antibody (SSAGLPPLSATSFLEDLMDRYGKNDSLTLTQLKSLLDHLHVGVGRDNVSQPKEGPRNLSTCFSSGDLFAAHNLSERSQIGASEFQEFCPTILQQLDSQACTSENQKSEENEQTEEGKPSAIE) was custom-made by Genscript (Piscataway, NJ). The specificity of the custom-made ZIP14 antibody was tested in multiple tissues using whole body Zip14 KO and WT mice (28, 29). Furthermore, we tested the specificity of the antibody using multiple tissues from whole body Zip14 KO mice to the pancreatic islets from floxed control and β Zip14 KO (Supplemental Fig. S1). Other antibodies were GAPDH (Cell Signaling Technology, #2118), GLUT2 (Proteintech, #20436-1-AP), and Calnexin (Novus Biologicals, #NBP100-1974).

Immunofluorescent Staining

To examine the presence of ZIP14 in pancreatic β cells of wild-type mice and to validate the absence of ZIP14 in β cells of β-ZIP14 knockout mice, mouse pancreatic tissue sections were probed by Insulin rabbit mAb (Cell Signaling, #C27C9) and custom-made ZIP14 rabbit Ab (Genscript). The immunofluorescent staining of mouse pancreatic tissue sections was performed by iHisto. To examine the presence of ZIP14 on ER membrane in β cells, MIN6 cells were seeded on d-lysine (MP Biomedicals Inc, # 0210269480)-coated glass coverslips in 24-well plates, fixed by 2% paraformaldehyde, and stained for ZIP14 (at a final concentration 8 μg/mL, custom made by GenScript; Piscataway, NJ) and ER marker KDEL (Novus Biologicals, #NBP1-97469 goat-anti-mouse 10C3, 1:100). The staining was visualized by fluorescent-conjugated secondary antibodies, Alexa Fluor 594 (ThermoFisher Scientific, #A11012, 1:1000) and Alexa Fluor 488 (ThermoFisher Scientitic, #A11001, 1:1,000). Nuclei were stained with DAPI (1:10,000 dilution from 5 mg/mL stock).

Image Acquisition and Processing

The immunofluorescent images were acquired using Zeiss LSM880 confocal inverted microscope with a ×40oil immersion objective. Images were processed and analyzed by the software Fiji ImageJ (v2.3.0). Briefly, 9–11 Z-stacked images were taken per view, and each of the images was processed through a colocalization analysis using JACoP plugins (using Otsu expression thresholds).

Histological Analyses

Pancreatic tissues collected from multiple mice within the same experimental group were pooled together, fixed in 10% (vol/vol) formalin, embedded in paraffin, and cut into 5-μm sections for hematoxylin and eosin (H&E) staining. Tissue embedding, sectioning, and H&E staining were performed by iHisto (Salem, MA) or the Histopathology Core at Cornell Animal Health Diagnostic Center. Sizes of pancreatic islets were measured using the software QuPath (37) or Fiji ImageJ (v2.3.0).

RNA Isolation and qPCR

The total RNA of pancreatic tissues or cell culture materials was isolated using TRI Reagent (Molecular Research Center Inc. # TR118). cDNA synthesis was performed using M-MLV reverse transcriptase (Invitrogen, #28025-013). Gene expression was measured using PowerTrack SYBR Green Master Mix (Invitrogen, #A46012). The quantitative PCR (qPCR) reactions were run in the Roche Lightcycler 480-II RT-PCR and the Applied Biosystems QuantStudio 3 systems. Primers used in this study were purchased from Integrated DNA Technologies, Coralville, IA. Primer sequences used in this study include: Gapdh forward 5′- AGGTCGGTGTGAACGGATTTG-3′, Gapdh reverse 5′- GGGGTCGTTGATGGCAACA-3′, Ins1 forward 5′ CTTCTTCTACACACCCAAGTCC-3′, Ins1 reverse 5′- CAGCTCCAGTTGTTCCACTT-3′, Ins2 forward 5′- GAAGTGGAGGACCCACAAG-3′, Ins2 reverse 5′- GTCTGAAGGTCACCTGCTC-3′, Slc39a14 forward 5′- GTAAACCTTGAGCTGCACATTAGC-3′, Slc39a14 reverse 5′- TGAAGCCGCTTCATGGT-3′. Data were normalized with Gapdh measurements for relative quantitation.

Secondary Data Analysis

Data were obtained from a meta-analysis of scRNA-seq study of human whole pancreatic tissues (38). UMAPs for the zinc transporter genes were generated using the web-based, publicly available platform provided by the original authors of Mawla et al. (38): https://www.huisinglab.com/diabetes_2019/index.html.

Statistics

Data are presented as means ± SE. Significance was assessed by two-tailed Student’s t test and one-way ANOVA with the Tukey–Kramer test unless otherwise noted. Analyses were performed using R (v4.0.4) or GraphPad Prism (v9.1.2). The number of replicates for each experiment is included in the figure legends.

RESULTS

Islet ZIP14 Colocalizes with Insulin-Expressing Cells and Responds to a Glucose Stimulus

The presence of ZIP14 in both the endocrine and exocrine pancreas was shown in the human pancreas by immunofluorescent staining (39). To examine whether ZIP14 is present in pancreatic β cells in mice, we performed immunofluorescent staining of ZIP14 and insulin (INS) in pancreatic tissue sections of wild-type mice. We were able to show that ZIP14 was highly expressed in pancreatic islets and largely colocalized with INS+ cells (Fig. 1A). This pattern of staining in the mouse pancreas is in agreement with our previous findings (25). Next, we gave wild-type mice a glucose challenge by oral gavage to examine whether pancreatic ZIP14 responds to glucose stimuli. We showed that glucose administration via oral gavage upregulated pancreatic expression of Ins1 and Ins2 (Fig. 1B), likely to meet the demands of zinc utilization in β cells during insulin synthesis and secretion. Interestingly, we observed that Slc39a14 was robustly decreased in the mouse pancreas upon glucose stimuli (Fig. 1B). By Western blotting, we also demonstrated a reduced protein level of ZIP14 in isolated islets of mice receiving glucose administration (Fig. 1C).

Figure 1.

Islet ZIP14 colocalizes with insulin-expressing cells and responds to a glucose stimulus. A: immunofluorescent staining of INS1 (red; left), ZIP14 (green; middle), and merged image (right) in mouse pancreatic tissue slices. Nuclei were stained with DAPI (blue). Scale bar = 400 μm. B: pancreatic mRNA levels of Slc39a14, Ins1, and Ins2 in mice receiving glucose administration through oral gavage. n = 3 or 4 mice/timepoint. C: Western blot probing ZIP14 in isolated islets (islets were combined from 3 mice per timepoint) at the indicated timepoints after glucose administration. D: a meta-analysis scRNA-seq study of human whole pancreatic tissues [Mawla et al. (38) showed robust expression of SLC39A14 (coding ZIP14) and SLC30A8 (coding ZnT8) in human β cells]. Left: UMAP overlain with cell cluster annotation. Middle: UMAP overlain with SLC39A14 expression. Right: UMAP overlain with SLC30A8 expression. Color shade in the middle and right panels indicates expression level.

To investigate whether ZIP14 is relevant to human β-cell biology, we checked expression levels of ZIP14 using an integrative single-cell RNA-seq (scRNA-seq) dataset of whole pancreatic tissues from healthy human subjects (38) (Fig. 1D). We found that SLC30A8 (encoding ZNT8) and SLC39A14 (encoding ZIP14) were the only two zinc transporters that were expressed abundantly in human β cells (Fig. 1D; Supplemental Fig. S2), the former of which has been shown to play a critical role in regulating insulin secretion (12–14). SLC39A14 is enriched in both islets and acinar cells, consistent with earlier studies with immunofluorescent staining (39). Importantly, in contrast to SLC30A8 expressed abundantly in both α and β cells of islets, SLC39A14 is particularly enriched in β cells among all major cell types in the islets (Fig. 1D). Taken together, the presence of Zip14/ZIP14 in both mouse and human β cells, as well as the glucose-induced change in murine islet ZIP14, strongly indicated a functional role of ZIP14 in regulating β-cell function.

β-Zip14 KO Mice on Chow Display Higher Circulating Insulin in Response to Glucose

We previously reported that mice with whole body knockout of Slc39a14 (Zip14 KO) developed mild hyperinsulinemia (24–26). Consistent with these previous reports, we were able to recapitulate many key metabolic phenotypes of whole body Zip14 KO mice, including hyperinsulinemia, either on chow or on HFD for 16 wk (Supplemental Fig. S3, A–E). However, whole body Zip14 KO mice are also known to develop chronic, low-grade systemic inflammation (24, 25), an impairment of intestinal epithelial permeability (40, 41), as well as altered glucose metabolism in the liver and adipose tissue (24, 40). Therefore, the hyperinsulinemia phenotype seen in whole body Zip14 KO mice could be attributed to the secondary effects of chronic systemic inflammation and changes in other peripheral tissues. Indeed, in our mouse cohort, there seemed to be a positive correlation between circulating IL-6 (a marker of systemic inflammation) and circulating insulin (Supplemental Fig. S3, B and E). The specific effects of β cell ZIP14 remain unclear.

To investigate the role of β-cell ZIP14 in regulating β-cell function and its impacts at the physiological level, we generated the mouse model with β-cell-specific knockout of Slc39a14 (β-Zip14 KO) (Fig. 2A). First, we show the mRNA expression of cre in islets from β-Zip14 KO. Accordingly, Zip14 mRNA expression was significantly downregulated in the islets from β-Zip14 KO. At the same time, there was no change in the Zip14 expression in the liver and small intestine, where Zip14 is known to be highly expressed. Our Western blot from isolated islets showed the depletion of ZIP14 in β-Zip14 KO mice (Fig. 2A; Supplemental Fig. 1B), and immunofluorescent staining of pancreatic sections barely detected colocalization between ZIP14 and INS in the islets of β-Zip14 KO mice (Fig. 2B). We have detected the ZIP14 band in our Western blots at ∼250 kDa as the protein band disappeared in the KO at this size. This band size agrees with previous studies conducted with multiple tissue types from Zip14 KO mice and antibodies raised against two peptide sequences (24–29, 40, 42–45). Others reported different sizes of ZIP14 in tissues from Zip14 KO mice. The observed differences in band size of ZIP14 could be because of using the antibodies raised against different peptide sequences, denaturation conditions for the protein lysates, and posttranslational modifications within the different tissues (46–48).

Figure 2.

β-Zip14 KO mice on chow have higher circulating insulin and pancreatic zinc depletion in response to glucose stimuli. A: establishment of β-Zip14 KO mouse model. Left: mRNA expression of cre and Zip14 in islets, liver, and small intestine from floxed control and β-Zip14 KO mice. n = 3 or 4 per genotype (combined islets from 2 mice per sample). Right: Western blotting of ZIP14 of islets isolated from β-Zip14 KO and floxed mice. B: representative images of immunofluorescent staining for INS (red; left), ZIP14 (green; middle), and merged image (right) in pancreatic tissue slices of β-Zip14 KO and floxed mice. Nuclei were stained with DAPI (blue). Body weight (C), fed blood glucose (D), and fed blood insulin (E) of chow-fed β-Zip14 KO and floxed mice. n = 5–13 per genotype. F: representative H&E images of pancreatic tissue sections and quantification of islet size from chow-fed β-Zip14 KO and floxed mice. n = 85 islets from 5 floxed mice and n = 93 islets from 5 β-Zip14 KO mice were quantified. G: levels of serum insulin at the indicated timepoints during OGTT assays. n = 10 floxed mice and n =10 β-Zip14 KO across two independent experiments. H: levels of pancreatic insulin at the indicated timepoints during OGTT assays. n = 7 floxed mice and n = 11 β-Zip14 KO across two independent experiments. I: pancreatic zinc content at the indicated timepoints during OGTT assays. n = 4 or 5 mice per genotype per timepoint. Percentage changes in insulin secretion (J) and Fluozin3-AM readout (K) from islets isolated from β-Zip14 KO and floxed mice during GSIS assays. n = 7 wells of islets (40 islets per well) per genotype across two independent experiments. *P < 0.05 by two-tailed Student’s t test. Data are expressed as means (SE). β-Zip14 KO, β cell-specific knockout of Slc39a14; GSIS, glucose-stimulated insulin secretion; H&E, hematoxylin and eosin; OGTT, oral glucose tolerance test.

KO mice were, in general, healthy, with similar body weight, blood glucose (fed), and blood insulin (fed) when compared with floxed control mice (Fig. 2, C–E). Next, we analyzed pancreatic tissue sections from β-Zip14 KO and floxed mice to assess changes in pancreatic morphology. Quantification revealed similar islet size between β-Zip14 KO and floxed mice on a chow diet (Fig. 2F). We measured mRNA expression of selected zinc transporters from islets of floxed and β-Zip14 KO mice and found no significant alteration (Supplemental Fig. S4). To assess changes in β cell function between β-Zip14 KO and floxed mice, we measured levels of pancreatic and circulating insulin of β-Zip14 KO and floxed mice in response to oral gavage of glucose. We observed that β-Zip14 KO mice had significantly higher circulating insulin at 15 min after glucose stimulus compared with floxed control (Fig. 2G). The stronger spike of circulating insulin in β-Zip14 KO mice was likely due to a greater rate of insulin secretion, as insulin content in the pancreas tended to be depleted earlier in β-Zip14 KO mice in response to a glucose challenge (Fig. 2H). Furthermore, prolonged stimulation of insulin secretion is known to deplete intracellular zinc in β cells (49). β-Zip14 KO mice had a significant decrease in pancreatic zinc at 30 min after glucose stimulus compared with the control (Fig. 2I).

The Absence of ZIP14 in β Cells Alters Intracellular Zinc Flux and Enhances Insulin Secretion

To test the hypothesis that the absence of ZIP14 in β cells enhanced insulin secretion, we performed assays of glucose-stimulated insulin secretion (GSIS) with islets isolated from β-Zip14 KO and floxed mice. Our GSIS assays detected greater levels of insulin secreted from the islets of β-Zip14 KO mice compared with the floxed mice (Fig. 2J). To examine further how the absence of ZIP14 affected intracellular zinc dynamics during insulin synthesis and secretion, we used Fluozin-3 AM (19, 21), a fluorescent zinc probe, to assess intracellular zinc content in the islets of β-Zip14 KO and floxed mice. We observed that β-Zip14 KO mice exhibited lower intracellular zinc content in the islets compared with floxed mice in response to treatments of 20 mM glucose followed by 40 mM KCl, a chemical that induces depolarization and thereby maximizes insulin secretion (Fig. 2K). The lower Fluozin-3 AM signal detected in β-Zip14 KO islets in vitro (Fig. 2K) is consistent with the pancreatic zinc depletion observed in β-Zip14 KO mice in vivo (Fig. 2I).

During insulin synthesis and secretion, the pool of intracellular zinc in β cells is determined by both zinc uptake (for insulin synthesis) and secretion (zinc being part of the insulin protein complex). To be able to conduct transport studies to better understand the zinc dynamics during insulin secretion, we switched to the mouse β cell line MIN6 cells. First, we used Fluozin-3 AM to record intracellular zinc dynamics in MIN6 cells with the same experimental design used in islets upon glucose treatment (stimulating insulin synthesis) followed by KCl (specifically promoting insulin secretion). We observed the same pattern of increase in Fluozin-3 AM from islets (Supplemental Fig. S5), indicating similar zinc dynamics between islets and MIN6 cells. In the following experiments with MIN6 cells, we had three groups to better compare glucose only and the combination of glucose and KCl treatments. We detected increased Fluozin-3 AM signals in response to glucose only or glucose and KCl combined treatment. Notably, the combination of glucose and KCl had significantly lower Fluozin-3 AM signals than the glucose-only condition (Fig. 3A) due to greater insulin secretion (Fig. 3C). In agreement with Fluozin-3 AM data, we found significantly lower zinc concentrations in glucose and KCl treatment compared with glucose only at 20 min after treatment as measured by microwave plasma-atomic emission spectrometer (Fig. 3B). Importantly, we found the highest level of ⁶⁵Zn efflux following the combined glucose and KCl treatment (Fig. 3D). Based on these data, we proposed a putative model where pancreatic β cells lose more zinc due to KCl-induced insulin secretion (Fig. 3E). Furthermore, congruent with our in vivo data (Fig. 1, B and C), Zip14 was downregulated in response to glucose and KCl treatment (Fig. 3F). To test the direct link between Zip14 downregulation and insulin secretion, we knocked down (KD) Zip14 in MIN6 cells using siRNA transfection. The level of reduction in ZIP14 protein was shown by Western blotting (Fig. 3G; also see Supplemental Fig. S6A). We found greater ⁶⁵Zn efflux (Fig. 3H) and insulin secretion (Fig. 3I) in the Zip14 KD cells than in the control following the combined glucose and KCl treatment. Furthermore, we found greater expression of ZnT8 in Zip14 KD cells in response cotreatment of glucose (Supplemental Fig. S7).

Figure 3.

The absence of ZIP14 in β cells alters intracellular zinc flux and enhances insulin secretion. A: Fluozin 3-AM readout from MIN6 cells treated with either glucose alone or glucose combined with KCl for 20 min. B: zinc concentrations in MIN6 cells at 20 min after treatment were measured by microwave-plasma atomic emission spectrometry (MP-AES). n = 3 wells per condition. C: percent change in secreted insulin from MIN6 cells at 20 min after treatment. n = 6 wells per treatment. D: percent change in ⁶⁵Zn efflux at 20 min after treatment. n = 6 wells per treatment. E: working model showing the effects of glucose and KCl on regulating intracellular zinc in β cells. F: fold change in relative Zip14 mRNA expression. n = 11 or 12 wells per condition across three independent qPCR experiments. G: representative Western blot showing lower ZIP14 abundance in siRNA-transfected MIN6 cells. H: percent change in ⁶⁵Zn efflux at 20 min after treatment following Zip14 knockdown. n = 3 per group (representative of two independent experiments). I: percent change in secreted insulin from MIN6 cells at 20 min after treatment following Zip14 knockdown. n = 5–8 wells per condition across two independent experiments. One-way ANOVA for A, B, C, and D and two-tailed Student’s t test for F, H, and I. Data are expressed as means (SE).

As deletion of ZIP14 (both in vivo and in vitro) led to a greater insulin secretion during GSIS assays, we tested whether a gain-of-function of ZIP14 might exert the opposite effect on zinc and insulin dynamics by overexpressing ZIP14 in MIN6 cells. We showed that MIN6 cells with ZIP14 overexpression, though with a low transfection efficacy, displayed a significant elevation of intracellular zinc and a mild decrease in secreted insulin at 20 min after treatment during a GSIS assay compared with cells transfected with control vectors (Supplemental Fig. S8, A–C). In sum, the data suggested an inverse correlation between the level of intracellular zinc, Zip14 expression, and the level of secreted insulin. The absence of ZIP14 in β cells facilitates intracellular zinc flux toward insulin secretion, resulting in the phenotype of hyperinsulinemia.

β Cells ZIP14 Contributes to Zinc Efflux from the ER Membrane during Insulin Secretion

Based on our previous studies in other tissue types, ZIP14 is located on the cell membrane and/or membrane of multiple cellular organelles, including the endoplasmic reticulum (ER) (24, 43, 50). ZIP14 on the cell membrane is known to transport zinc into cells, whereas ZIP14 on the organelle membrane transports zinc within membrane systems to the cytoplasm. As β-Zip14 KO mice did not display zinc deficiency in the pancreas at baseline (Fig. 2I) and instead had facilitated utilization of intracellular zinc for insulin secretion (Fig. 2, G–K), the absence of ZIP14 in the β-Zip14 KO on the cell membrane did not seem to compromise zinc uptake for insulin synthesis and secretion. Therefore, we hypothesized that the absence of ZIP14 on the organelle membrane decreased the liberation of zinc within the organelle membrane to cytoplasm and thus retained more zinc within the ER/Golgi system for insulin synthesis and secretion.

We performed several experiments in MIN6 cells to support this hypothesis. First, we leveraged confocal microscopic imaging to determine the cellular location of ZIP14 in β cells. The Z-stack images revealed that the immunofluorescent signal of ZIP14 was present on the edge of the cell as well as within the intracellular space (Figure 4A, i and ii; also see Supplemental Fig. S9A), suggesting the presence of ZIP14 on the cell surface and organelle membranes in β cells. We also overlayed the immunofluorescent signal of ZIP14 with KDEL (ER marker) in MIN6 cells. The colocalization between ZIP14 and KDEL again suggested the presence of ZIP14 on the intracellular membrane (Figure 4Aiii). Second, to examine how ZIP14 behaved upon activation of insulin synthesis and secretion, we compared differences between MIN6 cells with and without cotreatment of 20 mM glucose and 40 mM KCl. We observed a significant increase in secreted insulin (Fig. 4B) in MIN6 cells at 20 min after cotreatment of glucose and KCl, which was consistent with our observations made in vivo studies (Fig. 1, C and D). Furthermore, quantification of colocalization between the immunofluorescent signal of ZIP14 and KDEL revealed that the percentage of ER-positive for ZIP14 was significantly reduced at 20 min after cotreatment of glucose and KCl (Fig. 4C). In addition, a reduction in ZIP14 abundance in response to the cotreatment of glucose and KCl was shown by Western blot in crude endoplasmic reticulum fractions (Fig. 4D; Supplemental Fig. S6B). Therefore, the transient reduction in ZIP14 (in vitro MIN6 model) or absence of ZIP14 (in vivo β-Zip14 KO mouse model) on the intracellular membrane likely served as a mechanism to retain zinc within the ER/Golgi system for utilization in insulin synthesis and secretion.

Figure 4.

β Cells ZIP14 contributes to zinc efflux from the endoplasmic reticulum (ER) membrane during insulin synthesis and secretion. A: representative immunofluorescent images show the presence of ZIP14 on the KDEL+ membrane in MIN6 cells. ZIP14 is present not only on the cell membrane (i) but also in the intracellular membrane positive for KDEL, an ER marker (ii and iii). Red: ZIP14; green: KDEL; blue: DAPI. Arrows highlight the colocalization of ZIP14 and KDEL. Scale bar = 10 μm. See also Supplemental Fig. S9 for additional images. B: level of secreted insulin of MIN6 cells at 20 min after a cotreatment of 20 mM glucose and 40 mM KCl. n = 8 or 9 wells per condition across two independent experiments of insulin ELISA measurements. RQV = relative quantitative value. C: quantification of ER membrane positive for ZIP14 in MIN6 cells with and without a cotreatment of 20 mM glucose and 40 mM KCl. Colocalization of immunofluorescent staining for ER membrane (KDEL) and ZIP14 was performed. n = 5 or 6 views per condition and n = 8–10 z-stack images per view were examined by confocal microscopic imaging. Two-tailed Student’s t test. Data are expressed as means (SE). D: representative Western blot image showing ZIP14 abundance in crude endoplasmic reticulum (ER) and cytoplasmic (Cyto) fractions from MIN6 cells with and without a cotreatment of 20 mM glucose and 40 mM KCl; also see Supplemental Fig. S6B.

β-Zip14 KO Mice on Chow Have Impaired Glucose Homeostasis and a Shifted Fuel Source Preference toward Fatty Acids

Given that β-Zip14 KO mice had a greater capacity for insulin secretion in response to glucose stimulus (Figs. 2 and 3), we examined whether this phenotype impacted whole body metabolism. First, we performed an oral glucose tolerance test (OGTT) and insulin tolerance test (ITT) in β-Zip14 KO and floxed mice to characterize their profiles of glucose homeostasis. The results showed that β-Zip14 KO mice had no difference in glucose clearance in OGTT assays and remained similar in insulin sensitivity in ITT assays compared with floxed control (Fig. 5, A and B). Second, we determined food consumption and energy expenditure in β-Zip14 KO and floxed mice using a Comprehensive Laboratory Animal Monitoring System (CLAMS). We found that β-Zip14 KO mice had unchanged food intake (Fig. 5C) but a significant increase in locomotor activity in both dark and light cycles (Fig. 5D), resulting in significant increases in energy expenditure, oxygen consumption, and carbon dioxide production (Fig. 5, E–G). Notably, the greater energy expenditure of β-Zip14 KO mice did not affect their body weight compared with floxed mice (Fig. 2C). Lastly, we measured the respiratory exchange ratio (RER) of β-Zip14 KO and floxed mice to examine changes in fuel source preference. The data revealed a significant reduction in RER in β-Zip14 KO compared to floxed mice (Fig. 5, H and I), suggesting a metabolic shift toward fatty acids as a preferred fuel source. Taken together, these data suggested that the absence of ZIP14 in β cells dysregulated not only β cell function locally but also altered systemic energy metabolism.

Figure 5.

β-Zip14 KO mice on chow shifted fuel source preference toward fatty acids. A: blood glucose levels of β-Zip14 KO and floxed mice during OGTT assays. The area under the curve (AUC) was calculated. n = 9 or 10 mice per genotype across two independent experiments. B: glucose levels of β-Zip14 KO and floxed mice during ITT assays. AUC was calculated. n = 9–11 mice per genotype across two independent experiments. Food intake (C), locomotor activity (D), energy expenditure (E), oxygen consumption (F), carbon dioxide consumption (G), and respiratory exchange ratio of β-Zip14 KO and floxed control mice (H and I). Data from C–I were collected over 5 days by the Comprehensive Laboratory Animal Monitoring System (CLAMS). n = 4 mice per genotype for assays determined by CLAMS. *P < 0.05, **P < 0.01, ****P < 0.0001 by two-tailed Student’s t test. Data are expressed as means (SE). ITT, insulin tolerance test; OGTT, oral glucose tolerance test.

β-Zip14 KO Mice on an HFD Develop More Severe Islet Hyperplasia and Compensatory Hyperinsulinemia

Given that β-Zip14 KO mice on chow exhibited higher insulin secretion in response to glucose and shifted energy metabolism toward fatty acid utilization compared with floxed control (Figs. 2 and 5), we hypothesize that β-Zip14 KO mice developed more severe metabolic dysregulations upon chronic HFD feeding. We performed a chronic HFD intervention (60% fat) on β-Zip14 KO and floxed mice to test this hypothesis. At 16 wk of HFD intervention, β-Zip14 KO mice had significantly higher body weight and lowered lean mass than floxed mice, suggesting that the absence of ZIP14 in β cells led to a more severe diet-induced obesity (Fig. 6, A and B).

Figure 6.

β-Zip14 KO mice on high-fat diet (HFD) develop more severe obesity, islet hyperplasia, and hyperinsulinemia. Body weight (A) and body composition (lean/fat mass; B) of β-Zip14 KO and floxed mice in week 16 of chronic HFD intervention. n = 5 or 6 mice per genotype. C: blood glucose levels of β-Zip14 KO and floxed mice (week 16 of HFD) during OGTT assays. n = 4–6 mice per genotype. D: glucose levels of β-Zip14 KO and floxed mice (week 16 of HFD) during ITT assays. n = 4–6 mice per genotype. E: representative images of H&E staining in pancreas tissues and islet size quantification of β-Zip14 KO and floxed control mice in week 16 of HFD intervention. n = 28 islets across five floxed mice and 61 islets across five β-Zip14 KO mice were used for quantification. F: blood insulin of HFD-fed β-Zip14 KO and floxed control mice at 15 min after receiving an oral gavage of glucose. n = 9 floxed mice (4 mice from week 16 and 5 mice from week 24 of HFD) and n = 11 β-Zip14 KO (6 mice from week 16 and 5 mice from week 24 of HFD). *P < 0.05, **P < 0.01 by two-tailed Student’s t test. Data are expressed as means (SE). H&E, hematoxylin and eosin; OGTT, ITT, insulin tolerance test; oral glucose tolerance test.

Additional analyses were performed to characterize β-Zip14 KO mice in response to chronic HFD exposure. First, OGTT and ITT metabolic assays showed slight but not statistically significance difference between floxed control and β-Zip14 KO mice (Fig. 6, C and D). Second, islet morphology assessment showed a remarkable increase (∼threefold) in islet size of HFD-fed β-Zip14 KO compared with HFD-fed floxed mice (Fig. 6E), suggesting a greater diet-induced islet hyperplasia in β cells lacking ZIP14. Lastly, HFD-fed β-Zip14 KO mice developed more severe hyperinsulinemia compared with HFD-fed floxed mice (Fig. 6F). Together, the data demonstrated that the absence of ZIP14 in β cells resulted in exacerbated HFD-induced islet hyperplasia and compensatory hyperinsulinemia.

DISCUSSION

The in vivo and in vitro data presented in this study have significantly expanded our current understanding of ZIP14 in regulating β-cell function. We demonstrated that 1) β-cell ZIP14 is sensitive to glucose stimuli and serves as a negative regulator (a “brake”) for the process of insulin synthesis and secretion, 2) absence of ZIP14 in β cells leads to hyperinsulinemia due to facilitated, uncontrolled insulin secretion from β cells, and 3) the increased insulin secretion due to the absence of ZIP14 in β cells alters whole body energy metabolism and exacerbates diet-induced obesity and metabolic dysfunctions. In sum, the present study revealed that β-cell ZIP14 exerted local effects on regulating β-cell function and a profound impact on whole body metabolism.

The novel knowledge of β-cell ZIP14 we provided in our studies was built upon pioneering efforts made by Maxel et al. (51), an in vitro study investigating the role of ZIP14 in β cells. Using rat β cell line INS-1E, it was demonstrated that silencing of ZIP14 altered β-cell survival and the ratio of insulin synthesis to secretion after a chronic (24-h) high-dose exposure to glucose (22). In contrast, the scope of our study focused on the impacts of ZIP14 on cellular zinc flux and potential downstream signaling events that are critical to initiating insulin synthesis and secretion. This key difference in experimental design between the two studies might explain the discrepancies regarding the effects of ZIP14 on β-cell function. Nonetheless, the phenotype of islet hyperplasia in HFD-fed β-Zip14 KO mice is consistent with key findings of decreased β-cell apoptosis by silencing ZIP14 (22).

Our secondary data analysis from a meta-analysis of scRNA-sequences of human pancreatic tissues from healthy subjects revealed that SLC39A14 (encoding ZIP14) was expressed abundantly in human β cells (Fig. 1D; Supplemental Fig. S2). SLC39A14 is enriched in both islets and acinar cells, consistent with earlier studies with immunofluorescent staining (39). Furthermore, this expression pattern in β cells and acinar cells agreed with the recent work by Elgamal et al. (52), in which ZIP14 is expressed in both the β and acinar cells. However, there is a greater expression of other transporters in the β cells, whereas SLC39A14 was the only other zinc transporter besides SLC30A8 found to be highly expressed in β cells from the meta-analysis by Mawla et al. The discrepancies between these two studies could stem from multiple reasons. First, and possibly most importantly, Mawla et al. is a meta-analysis of the human pancreatic tissues only from healthy individuals. In contrast, Elgamal et al. included islets from 65 individuals, including both healthy and diabetic. Our findings in the mice are consistent with the ZIP14 expression in the human β-cells data from Mawla et al. However, the scarcity of scRNA sequencing from human samples makes it difficult to draw absolute conclusions. Future studies are needed for more cumulative and conclusive data from human samples.

Our β-Zip14 KO data suggested that the hyperinsulinemia phenotype seen in whole body Zip14 KO mice (25) was, in part, contributed by β-cell dysregulation. However, we also observed distinct phenotypes between β-Zip14 KO and whole body Zip14 KO mice. First, β-Zip14 KO mice had similar islet sizes to control mice on chow, whereas whole body Zip14 KO mice were reported to have larger islet sizes on chow (25). Second, β-Zip14 KO mice developed exacerbated HFD-induced obesity and metabolic dysregulation. In contrast, whole body Zip14 KO mice on chronic HFD had better glucose homeostasis and insulin sensitivity primarily due to effects on hepatic glucose metabolism (24). Lastly, β-Zip14 KO mice had shifted fuel source preference toward fatty acids, a phenotype associated with prediabetic condition (53), whereas whole body Zip14 KO mice conversely shifted their fuel source preference toward glucose (30). The striking differences between these two models suggested that the β-cell-specific effects of ZIP14 were possibly masked by the effects of ZIP14 in other tissue/cell types in the whole body KO system. In this study, the generation of β-Zip14 KO mice was able to uncover unique physiological changes due to β-cell ZIP14. In addition, our present study and previous studies in ZIP14 (22, 24, 25, 30, 40, 43) strongly demonstrated tissue/cell type-specific effects of ZIP14 and emphasized the need to generate cell-type-specific knockout systems for a better understanding of the functional role of ZIP14.

We used the MIN6 cell model to provide mechanistic insights into ZIP14 in β-cell biology. In this study, MIN6 cells displayed a reduction of ZIP14 when insulin synthesis and secretion were stimulated, which resembled ZIP14 behaviors in islets of wild-type mice receiving oral gavage of glucose. In addition, MIN6 cells exhibited less accumulation of cellular zinc when insulin secretion was enhanced, which mirrored the phenotypes of islets of β-Zip14 KO compared with floxed mice. Using this cell model, we also showed that the reduction of ZIP14 on ER membrane serves as a mechanism to limit zinc efflux into the cytoplasm, thereby retaining more zinc in the ER for insulin synthesis and secretion. One potential outcome of the reduction of ZIP14 on organelle membranes is a decrease in cytoplasmic zinc, a known player in driving the phosphorylation of AKT in several biological contexts (54, 55). The effect of inhibition of AKT phosphorylation on potentiating insulin secretion from β cells was shown previously (56). Whether ZIP14 plays an essential role in modulating AKT activity and its downstream cellular events in β cells warrants future investigations.

As implicated by our findings, β-cell ZIP14 may play a role in driving obesity and its complications. ZIP14 is enriched in β cells compared with other cell types in islets, and its downregulation in islets is associated with obesity and T2D in both human and mouse models (19, 21, 23). Consistent with these reports, β-cell Zip14 KO mice develop more severe obesity and metabolic dysregulation on chronic HFD. Our data suggested that the severe phenotypes of β-Zip14 KO mice are likely driven by their greater capacity for insulin secretion while on a chow diet. This is significant as emerging evidence suggests that hyperinsulinemia and hyperinsulin secretion are the driving factors of obesity and metabolic diseases (57, 58). It has been shown that hyperinsulinemia (postprandial or hypersecretion) promotes significant fat storage, resulting in obesity and consequent metabolic disorders. Prevention of chronic hyperinsulinemia through genetic manipulation in mice reduced high-fat diet-induced obesity and insulin resistance (59). Importantly, in humans, hyperinsulin secretors have been shown to have an increased incidence of impaired glucose tolerance and type 2 diabetes (60). In conclusion, the phenotypes of β-cell Zip14 KO mice support an increasing notion that hyperinsulinemia and hyperinsulin secretion are the driving factors of obesity and metabolic diseases (57–60).

DATA AVAILABILITY

The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Figs. S1–S9: https://doi.org/10.6084/m9.figshare.24050517.

GRANTS

This project was supported by Cornell University Division of Nutritional Sciences funds to T. B. Aydemir and the National Institutes of Health under award T32-DK007158 to S. B. Mitchell.

DISCLAIMERS

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) or the National Institutes of Health. NYSTEM C029155 and NIH S10OD018516 for the Zeiss LSM880 microscopes (i880 and u880).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Y-H.H. and T.B.A. conceived and designed research; Y-H.H., Y.K., S.B.M., T.L.T., and T.B.A. performed experiments; Y-H.H., Y.K., and T.B.A. analyzed data; Y-H.H. and T.B.A. interpreted results of experiments; Y-H.H. and T.B.A. prepared figures; Y-H.H. and T.B.A. drafted manuscript; S.B.M. edited and revised manuscript; Y-H.H., Y.K., S.B.M., T.L.T., and T.B.A. approved final version of manuscript.