Prolonged exposure to hyperglycemia induces β-cell dysregulation and ultimately cell death; phenomenon is known as glucotoxicity (1-5). In type 2 diabetes, insulin secretion progressively increases in an attempt to regulate glucose homeostasis during insulin resistance (6). However, chronic up-regulation of insulin secretion causes defects in insulin secretion, leading to constant hyperglycemia (7). Chronic hyperglycemia is toxic to pancreatic β-cells, impairing insulin synthesis and secretion, and inducing apoptosis, as has been observed in type 2 diabetes (4, 8).
In type 2 diabetes, the chronic toxicity of high glucose, oxidative stress and free fatty acids is involved in the increase of pro-apoptotic factors, such as caspases, and decreased anti-apoptotic factors, such as Bcl-2 in β-cells (5, 9). In several studies, glucotoxicity-induced β-cell apoptosis has been shown to be related to an imbalance between pro-apoptotic proteins and anti-apoptotic proteins, such as the Bax/Bcl-2 ratio, which was skewed towards apoptosis in insulinoma cell culture and isolated islets
Lin28a is an RNA-binding protein that selectively represses the expression of microRNAs that regulate glucose metabolism as well as cell proliferation and the differentiation of embryonic cells, stem cells and cancer (12-16). Lin28 consists of two homologues, Lin28a and Lin28b, with similar structural and functional properties, but which have a few differences (17). Recently, it was reported that Lin28a regulates the miRNA Let-7 which acts as a suppressor of genes related to the insulin signaling pathway (15). In Lin28a transgenic mice, Lin28a improved an insulin-sensitive state and prevented the development of diabetes, while Let-7 led to insulin resistance and an impairment of glucose tolerance in the skeletal muscle and brown adipose tissue of Let-7 transgenic mice (15). Impairment of insulin secretion and glucose intolerance were observed in mice with whole body and pancreas-specific overexpression of Let-7 (18). Although these observations suggest Lin28a has an important role in regulating glucose homeostasis, the question of whether Lin28a reduces glucotoxicity-induced β-cell dysfunction and apoptosis has yet to be considered.
We investigated whether glucotoxicity affects the expression in β-cells of Lin28a, which is associated with apoptosis. We also determined the role of Lin28a in glucotoxicity-induced β-cell dysfunction and apoptosis and investigated its mechanism of action.
Glucotoxicity is known to have harmful effects on β-cells causing dysfunction and ultimately leading to β-cell death (1-3, 5). We examined whether the level of Lin28a in pancreatic β-cells was altered by glucotoxicity. To confirm the induction of glucotoxicity in β-cells, the expression of insulin and its transcription factors was measured. When primary rat islet cells were cultured in high glucose (30 mM/L D-glucose) for 48h, the gene levels of insulin-1 and insulin-2 decreased over time (Fig. 1C, D). Incubation with high glucose for 48 h also impaired glucose-stimulated insulin secretion (GSIS) in rat islet cells (Fig. 1E). The mRNA and protein expressions of PDX-1 and BETA2 were reduced by high glucose levels in rat islet cells (Fig. 1 F-H). However, Bax, cleaved-caspase3 (c-caspase3), and the Bax/Bcl-2 ratio, all increased in response to high glucose in rat islet cells. Bcl-2 protein expression was decreased by high levels of glucose in rat islet cells (Fig. 1I). The mRNA and protein expression of Lin28a were reduced by exposure to high glucose in rat islet cells (Fig. 1A, B). These results suggest that glucotoxicity induces downregulation of Lin28a expression as well as β-cell dysfunction and apoptosis.
We tested whether the overexpression of Lin28a ameliorated glucotoxicity-stimulated β-cell dysfunction in primary rat islet cells. As shown in Fig. 2A, B, adenovirus-mediated overexpression of Lin28a in rat islets ameliorated the downregulation of insulin mRNA induced by high glucose. High glucose impaired GSIS; however, Ad-Lin28a ameliorated high glucose-induced impairment of GSIS (Fig. 2C). Ad-Lin28a restored high glucose-induced downregulation of mRNA (Fig. 2D) and protein (Fig. 2E) levels of PDX-1 and BETA2 in INS-1 cells. These results indicate that Lin28a decreased β-cell dysfunction via regulation of β-cell-enriched transcription factors.
We investigated whether Lin28a inhibits glucotoxicity-stimulated apoptosis in primary rat islet cells. As shown in Fig. 3A and Supplementary Fig. 1, Lin28a overexpression in rat islet cells caused by infection with Ad-Lin28a inhibited the high glucose-induced the mRNA and protein levels of Bax/Bcl-2 ratio induced by high glucose. Lin28a overexpression decreased the expression of c-caspase3 protein induced by high glucose (Fig. 3A). To investigate the impact of Lin28a overexpression on the survival of β-cells, we stained islets to detect dead cells using Ethidium-1 to stain dead cells. Dead cells stained red were increased in high glucose and Ad-GFP treated β-cells. The number of dead cells induced by high glucose were decreased when Lin28a was overexpressed (Fig. 3B). To investigate whether Lin28a regulates the apoptotic factors in β-cell, INS-1 cells were transfected with siLin28a. The protein levels of Bax and c-caspase3 were increased, while the Bcl-2 protein levels were decreased by Lin28a inhibition. The ratio of Bax/Bcl-2 was significantly up-regulated by siRNA administration (Fig. 3C). These results indicate that Lin28a regulates the expression of apoptotic factors.
The role of PI3 kinase/Akt/mTOR (PI3K/Akt/mTOR) axis in β-cell growth, survival, and function has been studied in detail (19, 20). We examined whether the protective effects of Lin28 in glucotoxicity-stimulated apoptosis were related to the PI3K/Akt/mTOR pathway. Treatment with high levels of glucose decreased the phosphorylation of Akt and mTOR in INS-1 cells. Adenovirus-mediated overexpression of Lin28a effectively recovered the decreased phosphorylation of Akt and mTOR caused by high glucose (Fig. 4, left). However, the levels of Akt and mTOR phosphorylation following the administration of Wortmannin, a specific inhibitor of PI3K, treated cells, were not decreased by high glucose and were not changed by Lin28a overexpression (Fig. 4, right). These results indicated that Lin28a regulated glucotoxicity-related genes via the PI3K/Akt/mTOR pathway in β-cells.
In this study, we established that chronic high glucose induced the downregulation of Lin28a expression, resulting in decreases in insulin mRNA levels and secretion, as well as increases in the Bax/Bcl-2 ratio and the levels of c-caspase3 proteins. The adenovirus-mediated overexpression of Lin28a ameliorated glucotoxicity-induced β-cell dysfunction. Ad-Lin28a inhibited the glucotoxicity-induced apoptotic factors, such as c-caspase3, and the Bax/Bcl-2 ratio, while inhibition of Lin28a, by transfection with siRNA-Lin28a, increased these apoptotic factors. Ad-Lin28a recovered the glucotoxicity-induced downregulation of Akt and mTOR phosphorylation. These observations indicate that Lin28a ameliorates glucotoxicity-stimulated β-cell dysfunction and apoptosis via regulation of the PI3K/Akt/mTOR pathway as well as β-cell enriched transcription factors.
Recently, Lin28a was shown to regulate glucose metabolism and insulin sensitivity (15). Hypothalamic ventromedial nucleus specific overexpression of Lin28a induced a significant amelioration in glucose metabolism, while downregulation of Lin28a due to a high fat diet was detrimental (21). Chronic increase of glucose contributed to β-cell dysfunction and apoptosis in type 2 diabetes, as it inhibits insulin gene expression and elevates apoptotic factors (2-4). We demonstrated that Lin28a plays a critical role in the glucotoxicity induced by a prolonged high level of glucose in β-cells. In this study, we demonstrate that, under glucotoxicity conditions, the gene and protein expressions of Lin28a were significantly reduced in primary rat islet cells, while β-cell dysfunction and the expressions of apoptotic factors were increased.
Lin28a overexpression inhibited high glucose-induced down-regulation of the expression of the insulin gene and GSIS in rat islet cells. PDX-1 and BETA2/NeuroD play critical roles in regulating insulin gene expression as pancreatic β-cell specific transcription factors (22). Lin28a overexpression recovered the high glucose-induced reduction of PDX-1 and BETA2. These data strongly suggest that Lin28a ameliorates glucotoxicity-induced β-cell dysfunction via the regulation of PDX-1 and BETA2 expression.
The up-regulation of pro-apoptotic factors, such as c-caspase3 activates the apoptotic process (4, 23). The Bcl-2 family is a key mediator of the process of apoptosis. The relative ratio of pro-apoptotic factors and anti-apoptotic factors, such as the Bax/Bcl-2 ratio, plays an important role in determination of cell survival or death (11, 24). An increase in the Bax/Bcl-2 ratio has found to be associated with a high risk of apoptotic activation (25, 26). Our data indicates that overexpression of Lin28a inhibited the levels of Bax and c-caspase3 induced by high levels of glucose, decreasing the Bax/Bcl-2 ratio in rat islet cells. When endogenous Lin28a expression in INS-1 cells was inhibited by Lin28a-specific siRNA, Bcl-2 expression was decreased while the levels of Bax and c-caspase3 were increased, resulting in an increase of the Bax/Bcl-2 ratio. Collectively, these data suggest that Lin28a can modulate anti-apoptotic and pro-apoptotic factors, and possibly improve the β-cell dysregulation that results from metabolic stress caused by chronically elevated glucose levels.
Several studies have shown that constitutively active Akt inhibits fatty acid-stimulated apoptosis in β-cells (27, 28), and β-cell specific overexpression of Akt has also been shown to prolong cell survival and protect against streptozotocin-induced diabetes (29). In several cells, pancreatic β-cells, muscle cell and cardiomyocytes Lin28a has been reported to prevent apoptosis and regulate the PI3K/Akt/mTOR pathway through downregulation of micro RNA Let-7 (15, 30, 31). We found that INS-1 cells treated with high glucose had decreased phosphorylation of Akt and mTOR while Lin28a overexpression recovered the high glucose-induced downregulation of p-Akt and p-mTOR. These data suggest that Lin28a regulates glucotoxicity-induced downregulation of the PI3K/Akt/mTOR pathway in β-cells.
It has been reported that high glucose suppresses the expression of the insulin gene, and its transcription factors such as PDX-1 and BETA 2, and insulin secretion independently of apoptosis (3, 10). Our data showed that Lin28a expression in β-cells was decreased by glucotoxicity and previous reports determined the suppression of Lin28a expression in cardiomyocytes by diabetes or ischemic insults (31). However, we have not identified the relationship between of high glucose-induced apoptosis and the high glucose-related regulatory mechanisms of Lin28a expression, and further research is expected in the future.
In summary, the present study shows that a chronic increase of glucose induces β-cell dysfunction and apoptosis, resulting in the downregulation of Lin28a expression. Under glucotoxic conditions, Lin28a up-regulates the expression and secretion of the insulin gene, while decreasing the Bax/Bcl-2 ratio and c-caspase3 protein expression via the PI3K/Akt/mTOR pathway. These results suggest that Lin28a has a crucial role in anti-apoptosis, making it a promising therapeutic target for addressing pancreatic β-cell dysfunction caused by glucotoxicity.
The INS-1 rat insulinoma cell line was cultured as previously described (22). To determine the effect of high glucose, cells were treated with 11.2 mM or 30 mM D-glucose for the indicated time. To investigate the effect of overexpression of Lin28a, cells were infected with Ad-Lin28a for 2 h in serum-free medium, and then incubated with fresh medium containing 11.2 mM or 30 mM D-glucose for 48 h.
Primary rat islets were isolated by collagenase digestion, and cultured as described previously (22). We performed all animal experiments in accordance with national guidelines and applicable national laws regarding animal protection. DGIST’s Animal Care and Use Committee approved the animal protocols (DGIST-IACUC-19052103-03). The islets were treated with Ad-Lin28a for 2 h in serum free medium, and then incubated with fresh medium containing 11.2 mM or 30 mM D-glucose for 48 h.
Recombinant adenovirus was prepared in a previously reported method (32).
RNA was obtained from conditioned cells using RNA Extraction Kits (Takara, Japan). cDNAs were generated using Reverse Transcription Kits (Applied Biosystems, USA). Gene expression levels were evaluated using Applied Biosystems SYBR Green PCR master mix kit in an ABI 7500 qPCR System. The qPCR conditions and the primers designed by Primer Express Software ver. 3.0 are described in Supplementary Table 1.
The cells or islets were lysed using RIPA buffer (Sigma, USA) with protease inhibitor and phosphatase inhibitor, and the cell lysate was harvested by centrifugation at 13000 rpm for 20 min. The protein samples were separated by Mini-PROTEANⓇ TGXTM Gels (Bio-Rad, USA) and electrophoretically transferred onto membranes. The membranes were incubated with blocking solution containing 5% BSA (Sigma) for 1 h, and further incubated with anti-Lin28a antibody, anti-Bax antibody, anti-cleaved-caspase3 antibody, anti-pro-caspase3 antibody, anti-phospho-Akt antibody, anti-Akt antibody, anti-phospho-mTOR antibody and anti-mTOR antibody (Cell Signaling, USA), anti-Bcl-2 antibody and anti-PDX-1 antibody (Santacruz, USA), or anti-BETA2 antibody (Abcam, UK). Antibodies were detected using a horseradish peroxidase-conjugated secondary antibody (Santacruz) and Clarity MaxTM Western ECL Substrate (Bio-Rad). The membranes were re-blotted with an anti-actin antibody (Santacruz) and immunoblot images were analyzed by ChemiDocTMXRS+ (Bio-Rad), and the intensity of the bands was quantified using Image LabTM Software (Bio-Rad).
Dead cell staining was performed using LIVE/DEAD Viability/Cytotoxicity Assay Kit (Invitrogen, USA) according to the manu-facturer’s protocol, with some modifications. INS-1 cells were incubated for 2 days in 36 mm dishes and then washed twice with Dulbecco’s phosphate-buffered saline. To stain dead cells, the cells were incubated with 4 μM EthD-1 solution for 10 min. Fluorescein labeling images were obtained at wave-lengths ranging from ∼495 nm/∼635 nm (Red:dead cell) using a fluorescence microscope (Leica, Germany) and a MataVue Imaging System Version 220.127.116.11 (Molecular Devices, USA).
Predesigned siRNAs for the control and Lin28a were purchased from BIONEER (South Korea). The Lin28a and control siRNA (C-siRNA) sequences are as follows: rat siLin28a, GAC GUC UUU GUG CAC CAG A tt; C-siRNA, CCU ACG CCA CCA AUU UGG U tt. 100 nM siRNA was transfected into INS-1 cells using transfection reagent (Lipofectamine 2000, Invitrogen, USA) in transfection media (Opti-MEM, Gibco, USA) for 4 h; then they were moved to a culture medium containing 2% FBS and collected ∼24 h after transfection. The effect of Lin28a siRNA was confirmed by western blot analysis.
Ten islets (triplets/condition) were infected with adenovirus Lin28a with a MOI of 20 in serum-free medium; adenovirus GFP was used for the control group. After 2 h, the islets were transferred to fresh medium with 2% FBS and 5 mM or 30 mM D-glucose for 48 h. The islets were then starved in 2% FBS and 5 mM D-glucose medium for 5 h and washed with 5 mM D-glucose KRBB. The islets were then replaced with a 5 mM or 25 mM KRBB and incubated for 1 h. The supernatant was gently collected and stored at −72°C until measurement. Changes in insulin secretion were analyzed using Rat Ultrasensitive Insulin ELISA (ALPCO, USA).
Data were expressed as the mean ± SEM. ANOVA was used to evaluate the variance, followed by a post- hoc least significant difference test. Statistical significance is indicated by P values < 0.05.
This work was supported by DGIST projects 21-BT-06, NRF-2021R1A2C1011314 and NRF-2018R1C1B6008955 from the National Research Foundation of Korea funded by the Ministry of Science, ICT of the Republic of Korea.
The authors have no conflicting interests.