Liraglutide protects against glucolipotoxicity‐induced RIN‐m5F β‐cell apoptosis through restoration of PDX1 expression

Abstract Prolonged exposure to high levels of glucose and fatty acid (FFA) can induce tissue damage commonly referred to as glucolipotoxicity and is particularly harmful to pancreatic β‐cells. Glucolipotoxicity‐mediated β‐cell failure is a critical causal factor in the late stages of diabetes, which suggests that mechanisms that prevent or reverse β‐cell death may play a critical role in the treatment of the disease. Transcription factor PDX1 was recently reported to play a key role in maintaining β‐cell function and survival, and glucolipotoxicity can activate mammalian sterile 20‐like kinase 1 (Mst1), which, in turn, stimulates PDX1 degradation and causes dysfunction and apoptosis of β‐cells. Interestingly, previous research has demonstrated that increased glucagon‐like peptide‐1 (GLP‐1) signalling effectively protects β cells from glucolipotoxicity‐induced apoptosis. Unfortunately, few studies have examined the related mechanism in detail, especially the role in Mst1 and PDX1 regulation. In the present study, we investigate the toxic effect of high glucose and FFA levels on rat pancreatic RINm5F β‐cells and demonstrate that the GLP‐1 analogue liraglutide restores the expression of PDX1 by inactivating Mst1, thus ameliorating β‐cell impairments. In addition, liraglutide also upregulates mitophagy, which may help restore mitochondrial function and protect β‐cells from oxidative stress damage. Our study suggests that liraglutide may serve as a potential agent for developing new therapies to reduce glucolipotoxicity.

resistance and results in constantly high blood sugar levels. 5 Thus, hyperglycaemia seems to work synergistically with hyperlipidemia via a process referred to as glucolipotoxicity, which contributes to various cellular dysfunctions, including endoplasmic reticulum (ER) stress, oxidative stress, mitochondrial dysfunction and chronic low-grade inflammation. 6 Evidence has emerged that elevated serum FFA is implicated in stimulating cell dysfunction or cell death events, such as insulin resistance in skeletal muscle cells, fatty liver and steatosis in hepatocytes, and dysregulated insulin secretion and apoptosis in β-cells. 7 Although some studies demonstrate that lipotoxicity-induced apoptosis is a specific effect of saturated FFA, unsaturated FFA is also toxic in an alternative pathway, which suggests that both saturated and unsaturated FFA markedly differ in their contributions to lipotoxicity. 8 Therefore, the combination of both types of FFAs may best reflect the physiological or pathological situation under glucolipotoxicity. 9 Research has indicated that β-cells are particularly sensitive to glucolipotoxicity. 10 In fact, pancreatic β-cell dysfunction and death are two core features of the late stages of T2D. 11 In diabetics, β-cells secrete larger-than-normal amounts of insulin to overcome insulin resistance, causing overwork and overproduction of insulin for extended periods of time. Glucolipotoxicity also damages β-cells, leading to little or no insulin production in a vicious cycle that ultimately promotes cell dysfunction and death. 12 Thus, glucolipotoxicitymediated β-cell loss is a critical causal factor of the late stages of diabetes. Although the exact mechanisms have not been delineated, studies have indicated that high glucose and FFA-induced metabolic stress appears to regulate β-cell identity and fate by perturbation of some specific transcription factors. 13 In particular, pancreatic duodenal homeobox 1 (PDX1) has been reported to play a key role in maintaining β-cell function and survival. 14 Decreased expression of PDX1 can be found in β-cells isolated from T2D patients, which suggests that PDX1 is important in islet compensation for glucolipotoxicity-induced insulin resistance. 15 Similarly, prolonged exposure of β-cells to high glucose and FFA levels has been demonstrated to stimulate PDX1 nuclear exclusion and degradation, resulting in decreased insulin gene transcription and cell survival. 13 On the contrary, recent studies reveal that administration of a glucagon-like peptide 1 (GLP-1) analogue, liraglutide, ameliorates the impairments to β-cells by upregulation of PDX1 in high-fat diet-induced diabetic mice. 16 Therefore, upregulation of PDX1 may prevent the progression of the glucolipotoxicity-induced β-cell dysfunction found in T2D. These previous findings highlight the importance of understanding how PDX1 is regulated.
Ardestani et al showed that mammalian sterile 20-like kinase 1 (Mst1) can stimulate ubiquitin-proteasome degradation of PDX1 and prohibit its function as a transcription factor in the nucleus. 17 This result indicates that regulation of Mst1 and PDX1 may be involved in liraglutide-mediated β-cell protection. However, few studies have examined this mechanism in detail. Therefore, in the present study, we investigate the effect of high glucose and FFA treatment on rat pancreatic RINm5F β-cells. Our results demonstrate that liraglutide restores the expression of PDX1 by inactivating Mst1, thereby ameliorating the impairments to RINm5F cells due to glucolipotoxicity. In addition, liraglutide also upregulates mitophagy to restore mitochondrial function and protect β-cells from glucolipotoxicity. Our findings suggest that activation of GLP-1 signalling by liraglutide may serve as an agent for the development of new therapeutic strategies against metabolic syndrome and T2D.
Slides were incubated with an FITC-labeled second antibody (Santa Cruz) in accordance with the origin of the primary antibody. Cells were stained with 1 μg/mL AO for 15 minutes and then washed with RPMI-1640 medium. Thereafter, images were acquired by using a fluorescence microscope (DP80/BX53; Olympus) and cellSense V 1.9 digital imaging software.
2.9 | Immunocytochemistry staining and AO staining After treatment, cells were fixed with 2% buffered paraformaldehyde, permeabilized in 0.25% Triton X-100 (Sigma-Aldrich) for KORNELIUS ET AL.  To further elucidate whether high glucose + FFA treatment interferes with RINm5f β-cell function during insulin synthesis and secretion, we performed relative expression qPCR assays to measure the levels of proinsulin mRNA transcripts. As shown in Figure 2B, no significant changes were found in the expression of proinsulin mRNA in the high-glucose group compared with that in the control group.
However, high FFA significantly suppressed mRNA levels of proinsulin, and this inhibition was enhanced by co-treatment with high FFA. The amount of insulin in the medium was determined by ELISA, and results revealed that high glucose + FFA treatment strongly suppresses insulin secretion ( Figure 2C).
To determine whether the observed glucolipotoxicity-mediated β-cell dysfunction is due to Mst1, we conducted knockdown experiments using Mst1 siRNA. As expected, our results showed a F I G U R E 1 High glucose and high FFA induced apoptosis of rat RINm5f β-cells. A, Phase-contrast microscopic images of cells taken after 24 hours of treatment. Treatment of RINm5f cells with high glucose did not induce morphological changes and toxicity. However, high FFA or high glucose + FFA treatment markedly inhibited cell viability. B, MTT assay results indicated 38% cell death in the high FFA group and 60% cell death in the glucose + FFA group after 24 h of treatment relative to those in the control and high glucose groups. C, Results of Western blot analysis demonstrated that 24 h of treatment of glucose + FFA stimulates caspase 3 and PARP activation in RINm5f cells. D, Nuclear condensation and fragmentation were markedly increased in both high FFA-and high glucose + FFA-treated cells compared with those in the control and high glucose-treated cells after 24 h. Results were determined on the basis of condensed and fragmented nuclear morphology through DAPI fluorescence. All data were collected from at least three independent experiments, and values are presented as mean ± SEM. Significant differences were determined through multiple comparisons with Dunnett's posthoc test at *P < 0.05 and **P < 0.01 compared with the control groups. Scale bar = 20 μm KORNELIUS ET AL.   Figure 3B). Furthermore, the results in Figure 3C show that liraglutide markedly up-regulates PDX1 protein levels in high glucose and high FFA-treated cells. Moreover, cleaved and activated Mst1 protein was reduced by liraglutide. This result was also confirmed through MTT assays, which showed that liraglutide effectively alleviates high glucose and high FFA-induced cytotoxicity (Figure 3D). Liraglutide further reduced the cleavage of caspase 3 and PARP ( Figure 3E), thus indicating that it can protect against glucolipotoxicity-induced β-cell apoptosis by restoring Mst1-suppressed PDX1 expression levels.
F I G U R E 2 High glucose and FFA causes RINm5f β-cell dysfunction through reduction of PDX1 expression. A, Immunoblotting revealed that the expression of PDX1 is clearly down-regulated when cells are treated with high glucose and FFA for 24 hours. Conversely, the level of cleaved Mst1 markedly increased in the high glucose and FFA groups. B, Real-time qPCR was used to measure proinsulin mRNA transcript levels. C, Treatment of cells with high glucose and high FFA for 24 h showed significantly decreased insulin secretion levels in the culture medium. D, High glucose and high FFA induced marked PDX1 reduction. However, knockdown of Mst1 by siRNA partially restored this inhibition. E, Cell viability was determined through MTT assay. Mst1 knockdown led to a significant restoration of cell survival. At least three independent experiments were performed, and values are presented as mean ± SEM. Significant differences were determined through multiple comparisons with Dunnett's posthoc test at *P < 0.05 and **P < 0.01 compared with the control groups ER stress-induced apoptosis plays a key role in PDX1-deficient β-cell during glucolipotoxicity. 20,21 To determine whether liraglutide protects cells from ER stress-induced cell death events, we measured two ER stress downstream typical markers, including PKR-like endoplasmic reticulum kinase (PERK) and eukaryotic translation initiation factor 2α (eIF2α), which activates a signalling network called the unfolded protein response to trigger C/EBP homologous proteinmediated apoptosis. 22 As shown in Figure 3F, treatment with high glucose and high FFA for 24 hours significantly increased the phosphorylation of PERK on threonine 980 and eIF2α on serine 51. However, liraglutide inhibited levels of p-Thr 980 PERK and p-Ser 51 eIF2α in high glucose and high FFA co-treated cells, which means glucolipotoxicityinduced ER stress is returned by addition of liraglutide.

| Liraglutide alleviates glucolipotoxicity-induced oxidative stress and cellular senescence
Previous studies have suggested that oxidative stress occurs in β-cells as a consequence of glucolipotoxicity. 23 Therefore, we investigated whether liraglutide protects cells from glucolipotoxicityinduced oxidative stress damage by using DHE probe, an indicator of superoxide. As shown in Figure 3A Some evidence suggests that glucolipotoxicity-induced oxidative stress mediates cellular senescence, which contributes to β-cell dysfunction. 26 To investigate whether liraglutide can attenuate high glucose and high FFA-induced senescence, we performed cytochemical SA-β-gal staining, a commonly used biomarker of senescent cells in culture. The results revealed a significant increase in the percentage of SA-β-galactosidase positive cells in high glucose and high FFA-F I G U R E 3 Liraglutide alleviates glucolipotoxicity-induced cell death by restoring PDX1 protein levels. A, Immunofluorescence staining revealed that nuclear PDX1 protein levels were markedly reduced by treatment with high glucose and FFA for 24 h. Co-treatment with liraglutide (0.1 μmol L −1 ) effectively restored the amount of nuclear PDX1. B, Western blot analysis of PDX1 from nuclear and cytosolic fractions from RINm5f β-cell extracts. Results showed that nuclear fraction of PDX1 is significantly restored by liraglutide. C, Western blot analysis showed up-regulation of PDX1 and downregulation of cleaved Mst1 by liraglutide after 24 h of treatment with high glucose and FFA. D, Cell viability was determined through MTT assay. Liraglutide effectively alleviated glucolipotoxicity-induced β-cell death. E, Western blot analysis demonstrated that liraglutide inhibits caspase 3 and PARP activation during glucolipotoxicity. F, Immunoblotting revealed that phosphorylation of p-Thr 980 PERK and p-Ser 51 eIF2α are downregulated when cells are treated with 0.1 μmol L −1 liraglutide for 24 h. All data were collected from at least three independent experiments, and values are presented as mean ± SEM. Significant differences were determined through multiple comparisons with Dunnett's posthoc test at *P < 0.05 and **P < 0.01 compared with the control groups treated cells. Liraglutide attenuated these effects caused by glucolipotoxicity, thus suggesting that decreased ROS content by liraglutide may be responsible for protection against glucolipotoxicity-induced cellular senescence.

| Liraglutide restores glucolipotoxicity-impaired mitophagy and mitochondrial membrane potential
Glucolipotoxicity-induced ROS accumulation and oxidative damage is caused by dysfunctional mitochondria. 27 Thus, the role of liraglutide in preventing glucolipotoxicity-mediated impairment of mitochondrial membrane potential was also investigated in this work. As shown in Figure 5A, exposure of cells to high glucose + FFA resulted in an increase in green fluorescence, indicating a great loss of mitochondrial membrane potential. However, co-treatment with liraglutide reduced the effects of high glucose + FFA on mitochondrial membrane potential significantly, thus suggesting that liraglutide preserves mitochondrial function against glucolipotoxicity.
Proliferator-activated receptor gamma coactivator 1 α (PGC1α) has been reported to be a positive regulator of mitochondrial function and renewal. 28 Therefore, we investigated whether glucolipotoxicity down-regulates PGC1α expression. As shown Figure 5B, high glucose + FFA caused a marked decrease in the expression of PGC1α at 24 hours post-treatment, and this inhibition was effectively restored by co-treatment of liraglutide.
Activation of PGC1α a and AMPK has been shown to enhance mitochondrial function and biogenesis by facilitating mitophagy. 29 We therefore investigated whether inhibition of mitophagy by high glucose + FFA results in impairment of mitochondrial function. The results presented in Figure 5B show that treatment with high glucose + FFA for 24 hours causes markedly down-regulated LC3-II, a positive regulator in type 1 and type 2 mitophagy stimulation. 30 Conversely, liraglutide co-treatment significantly restored this down-regulation, thus suggesting that impaired mitophagy may be recovered by liraglutide. The AO staining results also provided further evidence that high glucose + FFA significantly decreases the number of AVOs, a marker Many studies have demonstrated that long-term exposure of β-cells to high glucose and high FFA induces dysregulated insulin synthesis and secretion, eventually resulting in apoptosis. Thus, identification of the underlying mechanisms by which glucolipotoxicity contributes to β-cell dysfunction and death is critical to develop novel therapeutic strategies aimed at delaying or preventing β-cell exhaustion. Although the exact mechanism involved in β-cell glucolipotoxicity has yet to be elucidated, several hypotheses have been proposed, including mitochondrial dysfunction, oxidative stress, ER stress and islet inflammation. 23 However, the specific apoptotic signalling pathways involved in these mechanisms have not been comprehensively studied.
F I G U R E 5 Liraglutide restores glucolipotoxicity-impaired mitophagy and mitochondrial membrane potential. A, JC-1 immunofluorescent staining. Green fluorescence represents the dissipation of mitochondrial membrane potential in high glucose + FFA-treated RINm5f β-cells after 24 h. Red fluorescence indicates that co-treatment with liraglutide effectively preserves the mitochondrial membrane potential. B, Levels of PGC1α and LC3-I/II protein in high glucose + FFA-treated RINm5f β-cells. Liraglutide markedly restored LC3-II levels, thus suggesting that impaired mitophagy may be recovered. C, Representative images of AVO-positive cells treated through AO staining. The percentage of red AVO-positive cells was calculated from five images of each treatment. All data were collected from at least three independent experiments, and values are presented as mean ± SEM. The significance of differences was determined through multiple comparisons with Dunnett's posthoc test at *P < 0.05 and **P < 0.01 compared with the control groups. Scale bar = 20 μm Ardestani et al recently demonstrated that Mst1 is a key regulator of β-cell apoptosis and dysfunction in diabetes 17 ; the group found that activated Mst1 directly phosphorylates the β-cell transcription factor PDX1, resulting in ubiquitination and degradation of PDX1 leading to impairment of β-cells. These observations confirm that inhibition of Mst1 is a potent target for β-cell rescue. Interestingly, there are some reports showing that overexpression of PDX1 may stimulate β-cell neogenesis. For example, it has been demonstrated that the up-regulation of PDX1 is involved in β-cell reprogramming and neogenesis in adult mice. 32 Similarly, transgenic overexpression of PDX1 in STZ-treated mice can generate new insulin-producing islet cells. 33 All these supported that the up-regulation of PDX1 may act as a key factor for preservation of β-cell in dia-  35 Therefore, inhibition of caspase 3 by liraglutide may silence Mst1-dependent PDX1 suppression and protecting β-cells against glucolipotoxicity. Although our results can be partially rationalized as explained above, the precise mechanism of liraglutide-mediated Mst1 inhibition requires further elucidation.
Mitophagy is the selective degradation of mitochondria by autophagy and has been the subject of intense research in biological processes ranging from cellular metabolism to senescence. Increased mitophagy is thought to accelerate mitochondrial turnover, which, in turn, appears to maintain mitochondrial function and quality. 36 In diabetes, dysfunctional mitochondria have been linked to increased levels of ROS accumulation, which ultimately underlies β-cell failure and apoptosis. 37 Moreover, mitochondrial dysfunction and oxidative stress are largely involved in cellular senescence, thus suggesting that protecting and improving mitochondrial function may be a useful strategy for treating some oxidative stress-associated diseases, including neurodegeneration, metabolic syndrome and diabetes. In our results, we demonstrated that, under glucolipotoxicity, liraglutide up-regulates the AMPK/LC3 autophagy pathway, which drives the turnover of dysfunctional mitochondria in β-cells. Interestingly, previous studies indicate that binding of GLP-1 to its β-cell receptor elevates PGC1α expression, which means activation of GLP-1 signalling may improve β-cell functions via enhanced mitochondrial performance. 38 Our findings revealed that co-treatment of liraglutide effectively restore glucolipotoxicity-induced PGC1α downregulation.
Recent evidence also suggests the diminished expression of PGC1α in humans with insulin resistance and diabetes. 39 As PGC1α is the master regulator of mitochondrial turnover, restoration of mitophagy by liraglutide can prevent glucolipotoxicity toxicity. 40 In the present study, we provide evidence of the ability of liraglutide to inhibit high glucose and high FFA-induced β-cell toxicity. This protective effect appears to be associated with Mst1 and PDX1-related regulation and mitophagy activation. To the best of our knowledge, our group is the first to demonstrate the molecular mechanism of liraglutide against glucolipotoxicity-induced β-cell damage by the Mst1 and PDX1 pathway. Our results provide new insights into the potential use of incretin-based agents, such as liraglutide, in the preservation of β-cell function in metabolic syndrome and diabetes.