How Semaglutide, Tirzepatide, and Metformin Alone or Together Affect Insulin Cell Health and Function in High Sugar and Fat Conditions

All three medications at the defined doses used in this experiment, as well as the combinations of 1 mM metformin with both 10 nM semaglutide and 10 nM tirzepatide, mitigate HG-HL-induced apoptosis in INS-1 β-cells (Figure 1A). Although 1 mM metformin demonstrates a greater modulatory effect on HG-HL-induced apoptosis than either 10 nM semaglutide or 10 nM tirzepatide, the combinations of semaglutide or tirzepatide with metformin achieve a similar level of apoptosis mitigation as 1 mM metformin, suggesting the absence of any potentiation or additive effects. The modulatory impact of both 10 nM semaglutide and 10 nM tirzepatide on HG-HL-induced apoptosis was comparable (Figure 1A). Consistently, Western blot analysis of apoptotic biomarkers, including Cleaved PARP (C-PARP), Cleaved Caspase-3 (C-Caspase-3), Cleaved Caspase-9 (C-Caspase-9), Bax, Bim_EL, and Bim_L, supports the results of the apoptosis assay (Figure 1B). Interestingly, treatment with 10 nM tirzepatide increases the expression of certain apoptotic markers, such as C-Caspase-3, C-Caspase-9, Bax, and Bim_EL (Figure 1B(c,d,f)). In most cases, 1 mM metformin combination with 10 nM semaglutide performed better than with 10 nM tirzepatide. Western blot analysis of C-Caspase-3 reveals that the MS and MT treatments exert a greater modulatory effect on its levels compared to treatments with 1 mM metformin, 10 nM semaglutide, or 10 nM tirzepatide (Figure 1B(f)).

Compared to the control, HG-HL conditions significantly increase S-phase entry and reduce G2-phase entry in INS-1 β-cells, likely due to the stress imposed by HG-HL levels. Treatments with 1 mM metformin, as well as in combination with both 10 nM semaglutide and 10 nM tirzepatide, modulate the HG-HL-induced entry into the S-phase. Although 1 mM metformin and the MS combination show similar modulatory effects on induced S-phase entry, the MT combination demonstrates a slightly less pronounced effect than 1 mM metformin. Notably, 10 nM semaglutide and 10 nM tirzepatide actually increase S-phase entry under HG-HL conditions, indicating a distinct effect from metformin. In contrast, 1 mM metformin, MS, and MT treatments increase the proportion of cells entering the G2-phase. The 10 nM semaglutide has no notable effect on G2-phase entry, while 10 nM tirzepatide reduces G2-phase entry under HG-HL conditions (Figure 2).

HG-HL induction impairs the secretory function of β-cells, affecting basal glucose levels, KSIS, and GSIS. Treatment with 1 mM metformin improved all three parameters in HG-HL-induced INS-1 β-cells. However, both 10 nM semaglutide and 10 nM tirzepatide, at the specific doses used in this experiment, had a negative impact on impaired GSIS (Figure 3). While 10 nM semaglutide, 10 nM tirzepatide, and their combinations with metformin had no significant effect on basal glucose levels and KSIS, combining 1 mM metformin with either 10 nM semaglutide or 10 nM tirzepatide significantly enhanced GSIS in HG-HL-induced INS-1 β-cells. Notably, these combined treatments yielded a greater improvement in GSIS than 1 mM metformin (Figure 3a,b). Although all treatments positively impacted total insulin production in HG-HL-induced INS-1 β-cells, 1 mM metformin showed the greatest improvement compared to other treatments, including 10 nM semaglutide, 10 nM tirzepatide, MS, and MT. The 10 nM semaglutide had a slightly better effect on total insulin production than 10 nM tirzepatide (Figure 3c). The qRT-PCR results of insulin genes confirmed that 1 mM metformin, as well as the combinations of MS and MT, were the most effective treatments in upregulating insulin gene expression. In contrast, neither 10 nM semaglutide nor 10 nM tirzepatide showed any restorative effect on Ins mRNA levels (Figure 3d,e).

The Western blotting results show that all treatments reduced the elevated levels of TXNIP and P-STAT1 in HG-HL-induced INS-1 β-cells. Although there were no significant differences between 1 mM metformin and 10 nM semaglutide in reducing TXNIP levels, 10 nM tirzepatide resulted in a lower inhibitory effect on the HG-HL-induced TXNIP expression. Compared to 1 mM metformin and 10 nM semaglutide treatments, the combination of 1 mM metformin and 10 nM semaglutide (MS) demonstrated a more effective modulatory impact (Figure 4).

Based on our results with 10 nM semaglutide and MS on HG-HL-induced apoptosis and S-phase entry, this part of the experiment was only conducted with 1 mM metformin, 10 nM semaglutide, and MS. Experiments were performed under two conditions: co-treatment and post-treatment. In the co-treatment setup, 1 mM metformin prevented the change in the expression of all studied genes in HG-HL-induced INS-1 β-cells, including Ins1, Ins2, Pdx1, Slc2a2, NeuroD1, MafA, and Foxo1 (Figure 5A), along with the protein levels of PDX-1 and FOXO1. However, the level of NKX6.1 protein, which was significantly reduced under HG-HL-induced conditions, could not be restored by 1 mM metformin or any other treatment (Figure 5A(b)). The MS treatment showed similar restorative effects to 1 mM metformin on most genes, including Ins1, Ins2, Slc2a2, Pdx1, and MafA, but had additional outcomes (Figure 5A(a)). MS reduced the expression of NeuroD1 more significantly than in HG-HL-induced β-cells and failed to restore Foxo1 mRNA or the protein levels of PDX-1 and FOXO1 (Figure 5B). The 10 nM semaglutide treatment restored the expression of only Ins2 and Pdx1 genes, and these effects were weaker than those observed with 1 mM metformin (Figure 5A(a)).

In the post-treatment experiment, we observed notable differences from the co-treatment findings. Specifically, 10 nM semaglutide demonstrated a stronger restorative effect on the reduced levels of Ins1, Ins2, and Slc2a2 mRNA compared to both 1 mM metformin and MS. Interestingly, 10 mM metformin and its combination with 10 nM semaglutide were less effective at restoring the levels of these genes in HG-HL-induced β-cells than 1 mM metformin. In contrast to the above-mentioned genes, all treatments positively influenced the reduced levels of MafA, with all groups showing successful restoration of its expression in HG-HL-induced β-cells. Interestingly, no significant differences were observed in the qPCR results of Pdx-1, Foxo1, and NeuroD1 across all treatments (Figure 5B(a)). Under post-treatment conditions, the HG-HL+Med group (exposed to HG-HL conditions for 48 h, followed by 48 h of incubation in a normal medium) showed the highest NKX6.1 protein levels, surpassing even the control group (Ct) (Figure 5B(b)). The 10 nM semaglutide positively affected NKX6.1 levels, though less so than the HG-HL+Med group but significantly higher than the Ct group. In contrast, 1 mM metformin and the MS combination had similar effects on NKX6.1 protein levels, which were comparable to the Ct group and lower than both HG-HL+Med and 10 nM semaglutide treatments (Figure 5B(b)).

Although all five treatments modulated HG-HL-induced apoptosis in INS-1 β-cells, 1 mM metformin, MS, and MT exhibited greater anti-apoptotic effects compared to 10 nM semaglutide and 10 nM tirzepatide (Figure 1A). The anti-apoptotic properties of 1 mM metformin are well documented [33,34,35,36,37], demonstrating its ability to reduce apoptosis through various mechanisms. These include the activation of AMPK and the inhibition of mTOR signaling [38], the suppression of CD36 expression [16], the activation of NF-κB signaling [39], the reduction in TXNIP levels [39,40], and the mitigation of oxidative stress [41,42]. These effects are associated with metformin's ability to preserve mitochondrial function, lower the Bax/Bcl-2 ratio, and inhibit Caspase-3 activation, collectively contributing to its protective role in promoting cell survival [33]. Regarding the role of the modulatory impact of metformin and TXNIP, it has been found that TXNIP activation, induced by HG, leads to oxidative stress and inflammation through ROS production and NLRP3 inflammasome activation. Metformin treatment mitigates these effects by promoting TXNIP degradation, enhancing thioredoxin (Trx) activity, reducing ROS levels, and inhibiting NLRP3-mediated inflammation [43].

Consistent with these findings, our study suggests that 1 mM metformin reduces apoptosis, which may be related to downregulating TXNIP (Figure 4), leading to decreased ROS levels and subsequent reductions in apoptotic biomarkers, such as Bax, Bim, C-Caspase 9, C-Caspase-3 proteins, and C-PARP, and thereby apoptosis (Figure 1B).

Earlier studies suggested that TXNIP transcription regulation may follow a tissue-specific pattern. For example, the transcription factor FOXO1 promotes TXNIP expression in neurons and endothelial cells but suppresses it in the liver and pancreatic β-cells [44]. FOXO1 is a key regulator of TXNIP expression in pancreatic beta cells, where it binds to the TXNIP promoter and inhibits its transcription. This inhibition occurs through a competitive interaction with ChREBP, a transcription factor responsible for glucose-induced TXNIP expression. FOXO1 prevents ChREBP from binding to the TXNIP promoter, without affecting ChREBP expression or localization. Additionally, FOXO1 also inhibits transcription of other ChREBP target genes, such as liver pyruvate kinase [45]. This aligns with our findings that 1 mM metformin treatment increases the level of FOXO1 in HG-HL-induced β-cells, which, in turn, mitigates the levels of TXNIP, ROS, apoptotic biomarkers, and apoptosis (Figure 5A(b)).

We observed that similar to 1 mM metformin, 10 nM semaglutide effectively mitigated HG-HL-induced apoptosis in INS-1 β-cells. However, its modulatory effects on apoptosis and apoptotic biomarkers, including Bax, C-caspase-3, C-Caspase-9, C-PARP, and Bim proteins, were less pronounced compared to metformin (Figure 1A and Figure 2a). Consistently, previous studies have demonstrated that semaglutide, along with liraglutide (another GLP-1R agonist), reduces apoptosis in diabetic contexts [46,47]. This protective effect is attributed to its cytoprotective actions through cAMP-dependent survival pathways [48], its ability to enhance GSIS, and its role in reducing oxidative stress [49,50]. In alignment with these findings, our study revealed that 10 nM semaglutide mitigates HG-HL-induced TXNIP expression (Figure 4), which in turn reduces ROS levels. These results provide further evidence that TXNIP may serve as an upstream target through which semaglutide exerts its modulatory effects on HG-HL-induced apoptosis. Unlike 1 mM metformin, the modulatory impact of 10 nM semaglutide on HG-HL-induced TXNIP is not mediated through increasing FOXO1 protein.

Our results indicate that similar to 1 mM metformin and 10 nM semaglutide, 10 nM tirzepatide modulated HG-HL-induced apoptosis and the apoptotic biomarkers BimL and C-PARP in INS-1 β-cells (Figure 1A), which may be mediated through its impact on HG-HL-induced TXNIP production (Figure 4). However, unlike 1 mM metformin and 10 nM semaglutide, 10 nM tirzepatide also increased the levels of additional apoptotic biomarkers such as C-Caspase-3, C-Caspase-9, and Bim EL (Figure 1B). The response pattern of 10 nM tirzepatide on these biomarkers closely aligns with the outcomes of the apoptosis assay (Figure 1B) and the behavior of TXNIP protein (Figure 4) under HG-HL conditions. This observation may be linked to prolonged exposure to GIP or GIPR agonists, which have been shown to lead to receptor desensitization, reducing their effectiveness. Studies indicate that extended GIPR activation can result in receptor desensitization, potentially explaining the stimulatory effects of tirzepatide on some apoptotic biomarkers [51,52,53]. In this regard, it has been observed that similar to the remarkable therapeutic effectiveness of dual GIPR/GLP-1R agonists, the combination of GLP-1R agonism with GIPR antagonism also demonstrates significant therapeutic benefits [53,54,55]. These findings underscore the complexity of GIPR mechanisms of action and the synergistic interplay between GIPR and GLP-1R, emphasizing the need for further research to fully elucidate their dual metabolic and apoptotic effects.

One of the main goals of this study was to investigate the combined effects of metformin with semaglutide and tirzepatide on the maintenance and function of β-cells under HG-HL conditions. Our results show that neither MS nor MT had a greater effect on HG-HL-induced apoptosis compared to 1 mM metformin. However, the combined treatments had a stronger modulatory effect on apoptosis than either 10 nM semaglutide or 10 nM tirzepatide. This suggests that combining 10 nM tirzepatide and 10 nM semaglutide with 1 mM metformin enhances their ability to modulate apoptosis in the context of T2D.

C-PARP is a key marker and mediator of apoptosis. During apoptosis, Caspase-3 cleaves full-length PARP-1 into 89 kDa and 24 kDa fragments, inactivating PARP's DNA repair function and preventing NAD+ and ATP depletion. The 24 kDa fragment may inhibit uncleaved PARP, while the 89 kDa fragment contributes to apoptosis despite reduced catalytic activity. PARP cleavage occurs downstream of caspase activation in both the extrinsic and intrinsic apoptosis pathways [56,57,58].

Based on our results (Figure 1B), since all treatments modulated the levels of C-PARP in HG-HL-induced INS-1 β-cells and the response pattern of C-PARP was similar to other apoptotic biomarkers, we suggest that PARP cleavage acts as a final target for all treatments in their regulation of HG-HL-induced apoptosis.

HG-HL conditions induce stress that triggers cellular adaptations to enhance survival, including a shift in the cell cycle that increases the proportion of cells in the S-phase to support DNA replication necessary for proliferation or survival responses. Glucolipotoxicity is known to generate ROS, cause ER stress, and disrupt mitochondrial function, leading to impaired cellular signaling and the dysregulation of normal cell cycle checkpoints [59]. This dysregulation could be associated with triggering compensatory adaptations for survival. These adaptations often include promoting entry into the S-phase (DNA synthesis phase) and reducing entry into the G2-phase, likely due to the inability to move past DNA replication (Figure 2).

Our results indicate that treatments with 1 mM metformin and MS or MT modulate HG-HL-induced metabolic stress and partially alleviate the dysregulation of the cell cycle in the S-phase and G2-phase (Figure 2). Specifically, 1 mM metformin in combination with both 10 nM semaglutide and 10 nM tirzepatide, appears to mitigate the HG-HL-induced dysregulation of the cell cycle by reducing excessive S-phase entry and restoring G2-phase entry. Interestingly, 10 nM semaglutide and 10 nM tirzepatide tend to exacerbate stress, promoting increased S-phase entry and decreased G2-phase entry. However, when combined with 1 mM metformin, these effects are alleviated (Figure 2), suggesting that the combination of these medications with metformin may help reduce their potential adverse impact on cell cycle regulation.

Since the patterns of S-phase and G2-phase entry responses to the treatments align with the patterns of C-PARP responses, it is likely that C-PARP activity plays a crucial role in mediating the treatments' effects on cell cycle dysregulation. In this context, it has been observed that the inhibition of PARP activity through its cleavage can destabilize replication forks, leading to replication stress and potentially excessive entry into the S-phase as cells attempt to compensate for incomplete DNA synthesis [60].

Based on our results and considering that the pattern of HG-HL-induced TXNIP responses to the treatments aligns with C-PARP responses, we propose that TXNIP may function as an upstream target for the treatments, mediating their effects on C-PARP and, consequently, on the cell cycle (Figure 4). The treatments likely exert their effects through the modulation of TXNIP, ROS, ER stress, and inflammatory responses.

To study the impact of treatments on β-cell function, we examined the responses of GSIS, KSIS, total insulin production, and expression of insulin-coding genes in HG-HL-induced INS-1 β-cells. As our results indicate, 1 mM metformin, MS, and MT all improved the reduced GSIS levels in HG-HL-induced INS-1 β-cells (Figure 3). These findings align with previous studies, which demonstrated that metformin significantly enhances impaired GSIS in INS-1β-cells [17,61]. The protective effects of metformin on GSIS are linked to its activation of the AMPK pathway ¹⁷. Interestingly, the combined treatments of MS and MT were more effective than 1 mM metformin in improving GSIS, suggesting that these combinations might offer more potent therapeutic options for impaired GSIS in HG-HL-induced INS-1 β-cells (Figure 4). In addition to GSIS, HG-HL conditions can impair the basal secretory ability of INS-1 β-cells. Remarkably, we observed that 1 mM metformin could, at least in part, restore this function (Figure 3). This restoration may be attributed to the effects of metformin on total insulin expression and production, as well as its ability to inhibit HG-HL-induced apoptosis in INS-1 β-cells. Neither 10 nM semaglutide nor 10 nM tirzepatide improved the impaired level of GSIS under HG-HL stress in our model.

A previous study indicates that semaglutide significantly enhances GSIS in individuals with T2D, improving both first- and second-phase insulin secretion. It also increases insulin secretion rates to levels similar to those in healthy participants and boosts maximal insulin secretion capacity [62,63]. Using an obese T2D mouse model, it has been found that while no significant difference in GSIS was observed between the control and semaglutide groups, GSIS was significantly greater in the tirzepatide group compared to the control group [24]. In contrast to these results, we found that both 10 nM semaglutide and 10 nM tirzepatide had a negative impact on the impaired GSIS in HG-HL-induced β-cells. It is worth mentioning that the previous findings were derived from in vivo studies [29], which suggests that the effects of these two medications may not be directly related to their impact on β-cells. For instance, both semaglutide and tirzepatide have well-established effects on weight loss [64,65,66,67,68]. It is possible that the beneficial effects of these medications on GSIS in in vivo systems result from their ability to reduce obesity, which in turn improves impaired GSIS in HG-HL-induced β-cells. In other studies, where the positive impact of semaglutide and tirzepatide was observed, the effects of higher doses on GSIS were assessed in uninduced INS-1 β-cells [63,69,70]. However, in our study, we investigated the effect of a single dose of semaglutide and tirzepatide (10 nM) on impaired GSIS in HG-HL-induced INS-1 β-cells. The concentration we selected was based on pharmacokinetic data from clinical trials, as it approximates the therapeutic drug levels observed in human plasma [71,72] while also considering the combination of these treatments with metformin.

Finally, we also assessed the response of KSIS to all treatments, adding another layer to our understanding of how metformin and its combinations impact β-cell function under HG-HL-induced stress.

The primary difference between GSIS and KSIS lies in their mechanisms. GSIS follows the full physiological insulin secretion pathway, including glucose metabolism, ATP/ADP ratio increase, closure of ATP-sensitive potassium channels, membrane depolarization, calcium influx, and insulin granule exocytosis. In contrast, KSIS bypasses glucose metabolism, directly triggering insulin release through cell membrane depolarization. KSIS isolates late-stage mechanisms, aiding in assessing secretory function and pinpointing defects in glucose metabolism or sensing [73,74,75,76]. Comparing GSIS and KSIS provides insights into β-cell function, highlighting specific disruptions in stimulus–secretion coupling under physiological and pathological conditions. Our results show that only 1 mM metformin treatment significantly increased the reduced KSIS levels in HG-HL-induced β-cells. While both MS and MT also elevated this response, the effect was not statistically significant. This suggests that 1 mM metformin treatment improves the secretory ability of HG-HL-induced INS-1 β-cells. However, when combined with either 10 nM semaglutide or 10 nM tirzepatide, this effect was neutralized (Figure 3b). The positive impact of metformin on impaired KSIS has been supported by other studies. Accordingly, it has been found that metformin helps preserve pancreatic β-cell function and maintain insulin secretion under certain conditions, possibly through its effects on AQP7 expression and the MAPK signaling pathway [18,77].

In line with previous studies [78], at least part of the modulatory effects of 1 mM metformin, MS, and MT on GSIS can be explained by their stimulatory effects on total insulin expression and production in HG-HL-induced β-cells (Figure 3). This is further supported by the observation that the patterns of 1 mM metformin's effects on total insulin content and GSIS are similar.

In the second part of this study, we investigated the responses of genes and proteins associated with β-cell identity, including Ins1, Ins2, Pdx1, Slc2a2, MafA, NeuroD1, and Foxo1 mRNA, as well as PDX-1, FOXO1, and NKX6.1 proteins, to 1 mM metformin, 10 nM semaglutide, and the combination of 1 mM metformin and 10 nM semaglutide (MS). Given the superior effects of 10 nM semaglutide and MS compared to 10 nM tirzepatide and MT observed in earlier experiments, we focused on 1 mM metformin, 10 nM semaglutide, and MS for these analyses.

Experiments were performed under two conditions: co-treatment and post-treatment. In the co-treatment setup, 1 mM metformin restored the expression of all studied genes in HG-HL-induced β-cells, including Ins1, Ins2, Pdx1, Slc2a2, MafA, NeuroD1, and Foxo1, along with the protein levels of PDX-1 and FOXO1. This suggests that 1 mM metformin may play a critical role in preserving β-cell identity by counteracting HG-HL-induced dedifferentiation. In line with our results, it has been found that metformin inhibits islet β-cell dedifferentiation by decreasing C3, Nga3, and Oct4 while increasing Pdx1 and MafA expression, with C3 promoting dedifferentiation through the Wnt/β-catenin pathway [79].

However, the level of NKX6.1 protein, which was significantly reduced under HG-HL-induced conditions, could not be restored by 1 mM metformin or any other treatment, suggesting the complex role of the Nkx6.1 gene in the maintenance and function of β-cells.

The MS combination treatment had similar effects to 1 mM metformin on most genes, such as Ins1, Ins2, Slc2a2, Pdx1, and MafA, but also showed additional results. MS more significantly reduced the expression of NeuroD1 compared to HG-HL-induced β-cells and did not restore Foxo1 mRNA or the protein levels of PDX-1 and FOXO1 (Figure 5A). These findings suggest that MS is not superior to 1 mM metformin in maintaining β-cell identity.

The impact of GLP-1 on the mitigation of β-cell dedifferentiation is well established [7]. Accordingly, we found that 10 nM semaglutide treatment partially restored the expression of Ins2 and Pdx1 genes, with these effects being less pronounced than those seen with 1 mM metformin. This suggests that 10 nM semaglutide is less effective than 1 mM metformin in maintaining β-cell identity under HG-HL-induced conditions (Figure 5A). However, combining 10 nM semaglutide with 1 mM metformin partially enhanced its effect, suggesting a greater potential for improving β-cell identity maintenance.

This analysis highlights that 1 mM metformin remains the most effective treatment among those tested for counteracting β-cell dedifferentiation and preserving β-cell identity under HG-HL stress conditions.

In the post-treatment experiment, we observed notable differences from the co-treatment findings. Specifically, 10 nM semaglutide demonstrated a stronger restorative effect on the reduced levels of Ins1, Ins2, and Slc2a2 mRNA compared to both 1 mM metformin and MS. Interestingly, 1 mM metformin and its combination with 10 nM semaglutide (MS) were less effective at restoring the levels of these genes in HG-HL-induced β-cells. This suggests that the protective effect of 1 mM metformin is stress dependent and that, in the absence of stress, it may have adverse effects on the normal function of β-cells.

In contrast to the above genes, all treatments positively influenced the reduced levels of MafA, with all groups showing successful restoration of its expression in HG-HL-induced β-cells.

The most intriguing result was observed at the protein level with NKX6.1. Under post-treatment conditions, the HG-HL+Med group exhibited the highest levels of NKX6.1 protein, surpassing even the control group (Ct). While 10 nM semaglutide showed a positive effect on NKX6.1 levels, its impact was lower than the HG-HL+Med group but significantly higher than the Ct group. Conversely, the effects of 1 mM metformin and MS on NKX6.1 protein levels were comparable to the Ct group and notably lower than the HG-HL+Med- and 10 nM semaglutide-treated groups.

These results further suggest that NKX6.1 is regulated by a complex mechanism at the cellular level. This highlights the need for additional research to better understand the specific role and function of this protein in β-cells.

Semaglutide (CAS No. 910463-68-2) and tirzepatide (CAS No. 2023788-19-2) were from Adipogen Corp. (San Diego, CA, USA). Metformin hydrochloride (CAS No. 1115-70-4), L-Glutamine (TMS-002-C), sodium pyruvate (S8636), HEPES (TMS-003-C), β-mercaptoethanol (ES-007), and INS-1 832/13 Rat Insulinoma Cell (SCC207) were purchased from EMD Millipore Corporation (Temecula, CA, USA). Roswell Park Memorial Institute Medium (RPMI-1640) (350-060-CL) and D-PBS, 1X(311-010-CL) were acquired from Wisent Inc. (Saint-Jean-Baptiste, QC, Canada).

INS-1 832/13 Rat Insulinoma Cells were cultured in RPMI 1640 medium supplemented with 2 mM L-Glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 0.05 mM β-mercaptoethanol, and 10% fetal bovine serum (FBS) ((10082147) acquired from Fisher Scientific Company (Ottawa, ON, Canada), along with 2.5 mM glucose. Cells within passages 5-10 were used for all experiments. After being cultured in a medium containing 2.5 mM glucose for two days, cells were exposed to a specified dosage of individual treatments, their combinations, and/or an equal quantity of the vehicle for a duration of 2 h. All three treatments used in this study were dissolved in D-PBS. We used a single dose of each treatment: 1 mM metformin [37,80], 10 nM semaglutide, and 10 nM tirzepatide [81,82,83]. The concentrations of semaglutide and tirzepatide commonly used in in vitro studies typically range from 10 nM to 100 nM. These concentrations are selected to approximate therapeutic doses while minimizing potential cytotoxic effects [63,69,70,81,82,83,84,85]. Subsequently, high-glucose (HG) (25 mM glucose) and high-lipid (HL) (400 μM palmitic acid) conditions were imposed for the next 48 h [30]. Following the experimental treatments, the cells were used for various assays, including MTT, apoptosis, cell cycles, GSIS, KSIS, Western blot, and q-RT-PCR. For the post-treatment experiment, cells were first exposed to HG-HL conditions for 48 h. Afterward, the HG-HL medium was replaced with normal culture media containing the specified doses of each treatment, and the cells were incubated for an additional 48 h. This approach allows for the evaluation of the treatments' effects on β-cell function and dedifferentiation after the stress condition is removed, simulating a recovery or therapeutic window following stress-induced damage.

The BD Pharmingen™ FITC Annexin V Apoptosis Detection Kit II (Cat No. BDB556570; BD Biosciences, Mississauga, ON, Canada) was used to quantify apoptotic cells by detecting phosphatidylserine (PS) translocation to the cell membrane's outer leaflet. FITC Annexin V binds to exposed PS on apoptotic cells, while propidium iodide (PI) distinguishes viable cells from non-viable ones. Cells positive for FITC Annexin V alone were classified as apoptotic, propidium iodide (PI)-positive cells as necrotic or late-apoptotic, and cells negative for both as viable. Cells were washed twice with cold D-PBS and then resuspended in 1X Annexin V Binding Buffer at a concentration of 1 × 10⁶ cells/mL. A 100 µL aliquot (1 × 10⁵ cells) was transferred to a 5 mL culture tube, and 5 µL of FITC Annexin V and 5 µL of PI were added. The cell suspension was gently vortexed and incubated in the dark at room temperature (25 °C) for 15 min. After incubation, 400 µL of 1X Binding Buffer was added, and the samples were analyzed by flow cytometry within one hour using a BD Biosciences (Mississauga, ON, Canada) BD FACSAria™ Fusion Flow Cytometer.

Cells were harvested and washed in PBS and then fixed by adding cold 70% ethanol dropwise to the pellet while vortexing to prevent clumping, followed by a 30 min fixation at 4 °C. After fixation, cells were washed twice in PBS by centrifuging at 850 g, carefully discarding the supernatant to minimize cell loss. To remove RNA, 50 µL of RNase (100 µg/mL) was added to each sample, followed by 200 µL of PI from a 50 µg/mL stock solution to stain DNA. After incubating for 30 min at room temperature in the dark, PI emission was detected using the BD FACSAria™ Fusion Flow Cytometer (BD Biosciences, Mississauga, ON, Canada). This system allows for high-resolution analysis of cell fluorescence, ensuring precise detection of propidium iodide (PI) signals during flow cytometric analysis.

The cells were rinsed twice with cold PBS and then lysed using a RIPA buffer. After centrifugation at 13,000 rpm for 15 min, the supernatant was collected and transferred to a new microtube. The protein content was determined using the Bradford assay, and the lysates were used for Western blotting. Proteins were separated on polyacrylamide gels of varying concentrations (8%, 10%, and 12%) and transferred to polyvinylidene difluoride (PVDF) Amersham Hybond^® P membranes (RPN2020F) acquired from GE Healthcare, Oakville, ON, Canada. The membranes were blocked with a PBS containing 1% tween 20 (PBST) solution and 5% milk before being incubated with primary antibodies overnight at 4 °C. After three washes with PBST, the membranes were incubated with secondary antibodies for two hours at room temperature and washed again with PBST. Immunoreactivity was identified using peroxidase-conjugated antibodies, and the ECL Plus Western Blotting Detection System (GE Healthcare, Oakville, ON, Canada) was utilized to render the reaction visible. The band intensities were quantified through analysis with ImageJ 1.54 K and normalized relative to the housekeeping protein's intensity. Primary antibodies used were Cleaved Caspase-3 (Asp175) (Product No. 9664), Caspase 7 (Product No. 12827), Cleaved Caspase-7 (Asp198) (Product No. 8438), Cleaved PARP (Asp214) (Product No. 5625), Bim (Product No. 2933), PDX-1 (Product No. D59H3), FoxO1 (Product No. 2880S), NKX6.1 (Product No. 54551), and TXNIP (Product No. 14716S), which were sourced from Cell Signal Technology (Whitby, ON, Canada). Anti-P-STAT1 (phospho S727) (Cat No. ab109461) and Anti-β Actin (Cat No. ab8227) were purchased from Abcam (Toronto, ON, Canada). The secondary antibodies were obtained from Santa Cruz Biotechnology (Dallas, TX, USA).

Total RNA was isolated using TRIzol^TM (15596018, Invitrogen, Life Technologies Inc., Burlington, ON, Canada), and its concentration was subsequently determined using a nanodrop instrument (NanoDrop 2000c, ThermoFisher Scientific, Waltham, MA, USA). A portion of the total RNA was used for cDNA synthesis, utilizing 1 μg of total RNA and a cDNA synthesis kit, iScript™ Reverse Transcription Supermix (1708897, BioRad Laboratories, Saint-Laurent, QC, Canada). The resulting cDNA was then used as a template for q-PCR, with 1 μL of cDNA being used per reaction. The qRT-PCR reactions were performed using a SsAdvancedTM Universal Inhibitor-Tolerant SYBR Green Supermix (1725017, Bio-Rad Laboratories (Saint-Laurent, QC, Canada). The primers were designed based on sequences reported in our previous publications [30].

GSIS and KSIS assays were conducted to assess β-cell function under HG-HL conditions following treatment. Cells were seeded at a density of 0.5 × 10⁴ per well in a 24-well plate, incubated for 2 days, and then pretreated with specified treatment concentrations 2 h before exposure to HG-HL conditions for 48 h. For GSIS and KSIS assays, cells were washed with HEPES Balanced Salt Solution (HBSS) containing 2.5 mM glucose, followed by a 1 h wash in the same solution. Three wells per treatment were designated for testing: one with HBSS containing 2.5 mM glucose (normal glucose), another with 16.5 mM glucose (high glucose), and a third with 2.5 mM glucose plus 50 mM KCl. After a 2 h incubation, samples were collected for insulin ELISA analysis. The Rat/Mouse Insulin ELISA Kit (EZRMI-13K) from Sigma Aldrich (Oakville, ON, Canada) was used. Each well was washed, loaded with 10 μL of the sample and 80 μL of Detection Antibody, incubated at room temperature for 2 h, washed again, and then treated with 100 μL of Enzyme Solution for 30 min. Wells were washed again, followed by the addition of 100 μL of Substrate Solution, and incubated for 5–20 min before stopping the reaction. Absorbance was measured at 450 nm and 590 nm using a SpectraMax i3x Multi-Mode Microplate Reader (Molecular Devices, San Jose, CA, USA). For the GSIS and KSIS assays, the samples were diluted 150-fold, while for total insulin content, the extracted proteins were diluted 3500-fold to ensure that the insulin levels were within the detectable range of the assay kit.

The collected data were analyzed statistically using a one-way analysis of variance (ANOVA) to determine overall significance among groups. Following the ANOVA, Tukey's post hoc test was performed for multiple comparisons to assess the differences between group means. Statistical analyses were conducted using GraphPad Prism, Version 9.5.1 (GraphPad Software, San Diego, CA, USA). Data are presented as mean ± standard deviation (SD) unless otherwise specified, and a p-value of less than 0.05 was considered statistically significant.