English

• Type 2 diabetes mellitus (T2DM) is a chronic metabolic disorder characterized by insulin resistance and relative insulin deficiency. It is the most common form of diabetes, accounting for approximately 90%–95% among patients with diabetes [1]. Lipid metabolism plays a critical role in the development and progression of T2DM. Lipid accumulation can lead to insulin resistance. Dyslipidemia can also trigger inflammatory responses in adipose tissue, releasing pro‐inflammatory cytokines that contribute to insulin resistance and the pathogenesis of T2DM [2]. In addition, various kinds of lipids can also act as signaling mediator. Certain lipid metabolites, including ceramides and diacylglycerols, can impair insulin signaling pathways, leading to increased insulin resistance [3].

  Peroxisome proliferator‐activated receptors (PPARs), play significant roles in regulating lipid metabolism and insulin sensitivity. They are a family of ligand-dependent nuclear transcription factors, which are important in keeping metabolic homeostasis [4]. Peroxisome proliferator-activated receptor delta (PPARδ) is one of the three PPARs and universally expressed in multiple organs. Recent research reported that PPARδ involved in many biological activities, such as mitochondrial respiration [5], vascular function [6], exercise performance [7], and lipid and lipoprotein metabolism [8]. Several studies have demonstrated that PPARδ is involved in the regulation of fatty acid metabolism, insulin sensitivity, and energy expenditure. Activation of PPARδ can enhance fatty acid oxidation, which can help to improve metabolic profiles in individuals with T2DM [9]. Additionally, PPARδ activation has been shown to improve insulin sensitivity [10], a critical factor in managing T2DM. Improved insulin sensitivity helps the body utilize glucose more effectively, lowering blood sugar levels. Despite the various benefits of PPARδ activation in treating T2DM, there are notable disadvantages and concerns, particularly regarding its regulation of lipid metabolism in this context. A comprehensive study of the regulatory effects of PPARδ activation on lipid metabolism in T2DM is essential.

  GW501516, or cardarine or endurobol, is a synthetic agonist of PPARδ. This compound is originated by structure-based drug design produced by collaboration between GlaxoSmithKline and Ligand Pharmaceuticals in 1990s [11], specifically targeting PPARδ for the treatment of hyperlipidemia. A few studies have approved that GW501516 can reduce adiposity [12], enhance fatty acids β-oxidation in muscle [4], improve physical performance [13,14], and enhance insulin sensitivity in T2DM mice [9,10]. However, despite its promising therapeutic potential, some studies have indicated that this drug could induce cancer and lead to rapid tumor development in certain organs [15]. Considering the promising therapeutic potential of PPARδ agonist GW501516 for T2DM, this study aims to comprehensively understand its regulatory effects on lipid metabolism, exploring the underlying mechanisms of PPARδ activation in treating T2DM, as well as potential side effects. To achieve this, both non-targeted and targeted lipidomics analyses were conducted in the plasma of db/db diabetic mice. The results provide valuable insights into the effects of GW501516-induced PPARδ activation on lipid metabolism in T2DM mice.

  In this study, 7–8 weeks old C57BL/KsJ leptin deficiency db/db mice was chosen as diabetic model mice, and age matched db/m + mice was selected as control mice (non-diabetic). The procedure of animal experiment was the same as described previously [9,10]. All animals were purchased from the Animal Facility of the Chinese University of Hong Kong (CUHK), and all animal experimental procedure were approved by Department of Health, Hong Kong SAR Government and the CUHK Animal Research Ethics Committee.

  For lipidomics sample preparation, MTBE/Methanol/H₂O method was used. The sample preparation for targeted measurement of sterol metabolites (methods and Table S1 in Supporting information) including extraction, hydrolysis, derivatization and cleaning referred to the method by Jiaqian Qiu et al. [16]. Non-targeted lipidomics analysis was performed on liquid chromatography coupled with Orbitrap Exploris™ 120 Mass Spectrometer (LC-120-Orbitrap-MS) (methods and Table S2 in Supporting information). Targeted measurement of sterol metabolites was performed on liquid chromatography combined with triple quadrupole mass spectrometry (LC-QqQ-MS (methods and Tables S3 and S4 in Supporting information). As described previously [17], Non-targeted lipidomics data was extracted and identified by LipidSearchTM 3.0 software (Thermo Fisher, San Jose, USA) (methods and Table S5 in Supporting information). Each class of lipids were normalized to the intensity of relevant internal standards (Table S5). The total carbon numbers and double bond numbers of the combined fatty acyls are presented in this study (methods and Table S6 in Supporting information). Stereospecific designation is not assigned to the two or three fatty acyl chains. There are limitations in lipid identification. Targeted extraction of sterol metabolites were processed on Thermo Xcalibur Processing Setup-Quan-Identification software (Fig. S1 in Supporting information). The peak area ratio of each sterol to internal standard was used for statistical analysis. Each peak was manually checked.

  Multivariate statistical method principle component analysis (PCA) was carried out on MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/) [18]. Statistical analysis was carried out on Prism-GraphPad software by using analysis of variance (ANOVA) method with Tukey's multiple comparison test or student's t-test. Heatmaps were performed by Heml software [19]. Significance was considered when P-value < 0.05. All the results were presented as means ± standard error of the mean (SEM).

  PPARδ is one of three PPAR isotypes and universally expressed in the body. Activation of PPARδ has been investigated as a therapeutic strategy for metabolic disorders, including T2DM. A bunch of studies have reported that PPARδ activation can improve insulin sensitivity, regulate lipid metabolism, increase energy expenditure and protect against inflammation [20,21], etc. Pharmacological agents that activate PPARδ, known as PPARδ agonists, have been studied for their potential benefits in treating metabolic diseases. GW501516 is a PPARδ-specific agonist that has long been investigated in treating T2DM. It is proved to influence the expression of various genes involved in lipid metabolism. The underlying mechanism of GW501516 may involve the activation of PPARδ, which leads to the upregulation of genes associated with fatty acid oxidation (FAO), including the rate-limiting enzyme carnitine palmitoyltransferase Ⅰ (CPT1) in the FAO pathway [22,9]. Additionally, activation of PPARδ was also reported in inducing cholesterol flux [22]. However, despite the potential benefits observed in preclinical models, the development of GW501516 was discontinued for clinical use due to safety concerns. Animal studies revealed that the compound caused various serious side effects, including the development of cancer with long-term use [23]. However, it is important to note that PPARδ activation presents a promising therapeutic avenue, especially in improving insulin sensitivity and lipid profiles. More studies should be carried out to ensure that this treatment are effective and safe for widespread use.

  In our previous study [10,9], 6-week treatment of PPARδ agonist GW501516 significantly alleviated insulin resistance, which is the key parameter in the indication of T2DM. Additionally, the blood total triglyceride (TG) was also significantly reduced by GW501516. However, we found that the total cholesterol (Cho) levels was significantly up-regulated in the plasma of db/db mice in response to 6-week treatment of GW501516. Therefore, in order to investigate the underlying mechanism of this bidirectional phenomenon regulatory effect of GW501516 in db/db mice, a lipid profiling analysis was carried out.

  In this study, non-targeted lipidomics study was carried out to obtain an overview of lipid changes in response to T2DM and PPARδ agonist GW501516 treatment. In total, 19 subclass of lipids were detected, including phosphatidylcholines (PC), lysophosphatidylcholines (LPC), phosphatidylethanolamines (PE), lysophosphatidylethanolamines (LPE), phosphatidylinositols (PI), phosphatidylglycerols (PG), phosphatidylserines (PS), phosphatidic acid (PA), fatty acyl carnitines (AcCa), fatty acids (FA), diacylglycerides (DAG), TG, ceramides (Cer), hexosylceramides (Hex1Cer), sphingomyelin (SM), Cho, cholesterol ester (ChoE), cardiolipin (CL), and anandamide (AEA). Based on the identified lipids, the PCA score plots demonstrated a clear separation between db/m + and db/db mice (Fig. 1A), indicating lipids changed a lot between the two types of mice (Fig. 1B). There is also a clear separation between control and GW501516 treated groups both in db/m + and db/db mice, indicating activation of PPARδ may cause lipid shift regardless of the type of mice. For instance, the lipids that showed the most significant changes in response to T2DM, such as PC(20:1_18:2) and SM(d44:6), were notably reversed by GW501516, especially on db/db mice (Fig. 1C). When taking a depth analysis of each lipid class, we found that most class of lipids were significantly increased in the plasma of db/db mice, especially total TG, PG, PE and AcCa. In addition, a few classes of lipids, such as Cer and Cho were reversely regulated only in db/db mice by GW501516 (Fig. 1D). More detailed analysis was carried out to explain the underlying mechanism of these phenomenon.

  Figure 1

  

  Figure 1.  Lipidomics analysis of the regulatory effect of PPARδ agonist GW501516 on db/db mice in plasma. (A) PCA score plot. Volcano plots of db/db control group vs. db/m + control groups (B) and GW501516 treated group vs. control group on db/db mice (C). (D) Heat map of different classes of lipids in response to GW501516 treatment in db/m + and db/db mice. The total level of each lipid class was calculated by summing the detected lipids (Table S6) within the same class. The color bar stands for the percentage of each lipid class in the four groups. Ctrl, control group. GW, GW501516 treated group. Data were analyzed by one-way ANOVA in (D) with Tukey's post-hoc tests. ^*P < 0.05 vs. db/m + control group. ^#P < 0.05 vs. db/db control group, n = 9 in each group.

  DownLoad: Full-Size Img PowerPoint

  The accumulation of free FA in plasma has long been reported associated with T2DM and could elevate the risk of cardiovascular disease [24]. In this study, as consistent with previous studies, the levels of total FA were significantly increased in the plasma of db/db mice (Fig. 1D). In addition to FA, long-chain AcCa, a class of intermediate metabolites of FAO pathway, was also significantly increased (Fig. 2A). Although the total FA was not significantly reduced by the treatment of GW501516, a reduction trend can also be observed, and the levels of individual FA were substantially decreased in our previous study in response to GW501516 treatment in plasma [10]. Moreover, the total level of AcCa was markedly down-regulated, especially the most high abundant AcCa steroyl-carnitine [AcCa(C18:0)], oleoyl-carnitine [AcCa(C18:1)] and linoleyl-carnitine [AcCa(C18:2)], were significantly reduced by activation of PPARδ (Fig. 2B). The results may indicate that overall FAO efficiency has improved, likely due to enhanced protein activity, such as that of the rate-limiting enzyme CPT1, as a result of PPARδ activation in the FAO pathway. Considering the accumulation of FA contribute to insulin resistance [25], the current results revealed that activation of PPARδ by GW501516 plays positive role in reducing circulating levels of toxic free fatty acids, which may contribute to the improvement in insulin resistance.

  Figure 2

  

  Figure 2.  Regulatory effect of GW501516 on fatty acid metabolism pathway, PGs and TGs in plasma of db/m + and db/db mice. (A) Total acylcarnitines, (B) individual acylcarnitines. (C) The effect of GW501516 on total PG. (D) Effects of GW501516 on the unsaturation of PGs in plasma of db/m + and db/db mice. (E) Regulatory effect of GW501516 on total TGs. (F) Volcano plot of plasma TGs between db/db and db/m + control mice. (G) Volcano plot of plasma TGs in response to GW501516 treatment on db/db mice. (H) Regulatory effect of GW501516 on fatty acyl length of TGs. (I) Regulatory effect of GW501516 on the saturation of TGs in plasma of db/m + and db/db mice. AcCa, acylcarnitines. Ctrl, control group. GW, GW501516 treated group. Results are means ± SEM, n = 9. Data were analyzed by one-way ANOVA with Tukey's post-hoc tests (AA-E, H-I): ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. ^#P: student's t-test.

  DownLoad: Full-Size Img PowerPoint

  PG is a glycerophospholipid in the plasma membrane and inner mitochondrial membrane. It is a precursor for the synthesis of CL. Many studies reported that the accumulation of PG was positively associated with both prediabetes and T2DM [26]. In this study, we found that the total PG in plasma of db/db mice was substantially increased (Fig. 2C). In addition, all PGs, including saturated fatty acids (SFA), monounsaturated fatty acids (MUFA) and polyunsaturated fatty acids (PUFA) were all significantly upregulated in plasma of db/db mice (Fig. 2D). Interestingly, the total levels of PG (Fig. 2C), PUFA PGs (Fig. 2D) and the most highly abundant PG such as PG (O-30:0_20:3), PG (O-30:1_20:0) and PG (28:0_18:2)c were significantly reduced by GW501516 (Fig. S2 in Supporting information). Since PG is a precursor of CL in cardiolipin synthase (cls) (Fig. S3 in Supporting information), consequently, the level of total CL was markedly reduced in the plasma of db/db mice (Fig. 1D). As CL is important building blocks of the inner membrane of mitochondrial, and plays crucial roles in the maintenance of mitochondrial structure and function [27]. The observed deficiency in circulating levels of CL in the plasma of db/db mice could potentially lead to impairments in several mitochondrial functions, including aggregation of structures [28], facilitation of quaternary structure formation, and the initiation of apoptosis [29]. In this study, although treatment with GW501516 did not result in statistically significant changes in the plasma CL levels of db/db mice, the modulatory impact of GW501516 on PG implies that PPARδ activation could confer protective effects against the dysregulation of PGs caused by T2DM. Considering that FAO occurs in the mitochondria, mitochondrial dysfunction may lead to an imbalance in FAO, which could further contribute to insulin resistance in db/db mice. The regulate effects of PGs suggest a potential therapeutic benefit of GW501516 in ameliorating T2DM induced insulin resistance by correcting mitochondrial dysfunction through positively influencing the balance of these essential lipid components.

  TG in the blood are a type of lipid, serving as a form of fat that is crucial for energy storage and various other functions in the body. TG has long been regarded as a risk factor for diabetes [30,31]. High levels of TG in the blood have been associated with an increased risk of developing T2DM. This is because high TG are often a part of a cluster of metabolic abnormalities known as metabolic syndrome of T2DM, which also includes insulin resistance, obesity, high blood pressure, and abnormal cholesterol levels. In alignment with previous studies [10], the total level of TG and most of the individual TG was significantly accumulated in the plasma of db/db mice (Figs. 2E and F), as well as its precursor lipid classes including PA and DG, and downstream lipid classes including PE, PC, LPE and LPC (Fig. 1D and Fig. S3). In this study, six-week treatment of GW501516 could almost significantly reduce the total levels of TG (Fig. 2E), and a number of individual TGs were significantly down-regulated (Fig. 2G), especially long-chain TG (Fig. 2H) but without showing a particular preference for the saturation level of the TGs (Fig. 2I). Given that elevated TG levels are a risk factor for T2DM, activation of PPARδ by GW501516 may play positive effect in T2DM induced dysfunction of hyperlipidemia.

  Cholesterol is a fat-like substance that found in all the cells in the body. It is crucial for the formation and maintenance of cell membranes [32], providing them with structural integrity and fluidity. Moreover, Cholesterol is essential precursor for the production of many important metabolites, such as vitamin D, steroid hormones, bile acids, cholesterol esters, oxysterols (Fig. 3A). High intracellular cholesterol is toxic to the cells, and high serum cholesterol built up in the arterial walls which will lead to the plaque formation, one of the initial steps in atherosclerosis development [33]. Therefore, T2DM patients with high serum cholesterol increases the risk of cardiovascular disease [34]. In this study, lipidomics analysis revealed that the level of cholesterol was significantly increased in response to 6-week treatment of PPARδ agonist GW501516 in plasma of both db/m + and db/db mice. These findings were further validated using a cholesterol assay kit (Fig. 3B). A similar increase in cholesterol has been documented in other studies upon activation by different members of the PPAR family, such as PPARα. Administration of the PPARα agonist fenofibrate (Tricor) for two months resulted in raised serum total cholesterol levels in both db/m + and db/db mice [35]. In addition, in this study, a marked elevation in high-density lipoprotein (HDL) cholesterol levels (Fig. 3C) and HDL/total cholesterol ratio (Fig. 3D) in db/db mice was also observed, which is consistent with the previous findings [11]. HDL is often termed as “good” cholesterol because it helps remove excess cholesterol from the body and reduces the risk of atherosclerosis, heart disease and other cardiovascular events [36]. The upregulation of HDL and HDL/total cholesterol ratio by GW501516 may indicate a beneficial effect on the regulation of circulating cholesterol levels in db/db mice.

  Figure 3

  

  Figure 3.  Regulatory effect of GW501516 in cholesterol synthesis and metabolism pathway. (A) Schematic diagram of cholesterol synthesis and metabolism pathway. (B) Total cholesterol level in response to GW501516 in plasma of db/m + and db/db mice. (C) HDL levels in response to GW501516 in db/db mice. (D) HDL/total cholesterol ratio in db/db mice. (E) Effects of GW501516 on precursors of cholesterol in plasma of db/m + and db/db mice. (F) Total cholesterol ester (ChE) levels in plasma. (G) Individual cholesterol esters in plasma in response to GW501516 treatment. Ctrl, control group. GW, GW501516 treated group. Results are means ± SEM, n = 9. Data were analyzed by one-way ANOVA in (B, E-G) with Tukey's post-hoc tests or student's t-test (C, D, G): ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. ^#P: student's t-test.

  DownLoad: Full-Size Img PowerPoint

  To address the underlying mechanism of the increased cholesterol flux in blood, we specifically measured the upstream precursors of cholesterol. We can see from Fig. 3E that, there was a significant increase in nearly all cholesterol precursor metabolites in the plasma of db/db mice compared to db/m + mice, with the exception of desmosterol. Downstream in the cholesterol synthesis pathway, particularly desmosterol and 7-dehydrocholesterol, showed a pronounced increase following GW501516 treatment, especially in db/db mice. These findings suggest that cholesterol biosynthesis is augmented in db/db mice and further enhanced by GW501516 treatment, which may partially explain the observed elevation in cholesterol levels under both conditions.

  In addition, we also analyzed the downstream metabolites of cholesterol, cholesterol esters, which are essential for the proper storage and transport of cholesterol participate in the regulation of lipid metabolism and cell membrane properties, and can influence cellular energy balance and signaling pathways [37]. Cells can convert free cholesterol into cholesteryl esters (ChE) for storage and then hydrolyze them back into free cholesterol when needed. This dynamic balance helps maintain cellular cholesterol homeostasis. In this study, we found that the level of total ChE was significantly reduced in the plasma of db/db mice, indicating the failure of conversion of cholesterol into ChE under T2DM condition, which may contribute to increased total circulating level of cholesterol in db/db mice. Treatment of GW501516 could significantly increase the level of the total ChE in db/m + mice (Fig. 3F), and partially in several individual ChE in db/db mice, such as palmitoyl cholesterol [ChE (C16:0)], eicosatrienoyl cholesterol [ChE (C20:3)], docosapentaenoyl cholesterol [ChE (C22:5)], and docosahexaenoyl cholesterol [ChE (C22:6)] (Fig. 3G). This implies that GW501516 could have a beneficial role in promoting ChE production, which in turn may help reduce the buildup of cholesterol in the plasma of db/db mice. Building on previously reported findings that activation of PPARδ by GW501516 significantly enhances the expression of ATP-binding cassette sub-family A member 1 (ABCA1) [11], a vital membrane protein involved in the transport of cholesterol and phospholipids. These results suggest that cholesterol influx is increased through PPARδ activation. Collectively, all the results indicate that GW501516 may exert positive effect through elevating circulating cholesterol flux and improving cholesterol homeostasis, exhibiting a potential therapeutic benefit in managing cholesterol levels in T2DM.

  Cer are the central molecules in the sphingolipid metabolism, and are considered the building blocks for other complex sphingolipids. Cer are not only structural components of cell membranes but also act as signaling molecules involved in processes such as apoptosis, cell growth, differentiation, and inflammation [38]. The relationship between Cer and diabetes, particularly T2DM, is an area of active research and interest due to the role of ceramides in metabolic pathways. High levels of Cer in tissues such as muscle and liver may disrupt insulin signaling pathways, reducing the effectiveness of insulin and thus contributing to the development of insulin resistance, which is a hallmark of T2DM [39]. In this study, aligning with findings from other research, we observed a significant elevation in the total Cer levels in plasma of db/db mice when compared with control lean mice (Fig. 4A). Among the various Cer species, those with 34 and 42 carbon fatty acyl chains exhibited the most significant increase (Fig. S4A in Supporting information). Furthermore, Cer featuring mono-unsaturated fatty acyl chains displayed the greatest elevation in levels (Fig. 4B). Intriguingly, following treatment with the PPARδ agonist GW501516, there was an even more pronounced increase in Cer levels in plasma of both db/m + and db/db mice, especially Cer(42:1) and Cer(42:2) with high abundance (Figs. S4B and C in Supporting information).

  Figure 4

  

  Figure 4.  Effect of GW501516 in ceramide metabolism pathway. (A) Total ceramide level in response to GW501516 in plasma of db/m + and db/db mice. (B) Effect of GW501516 on the saturation of Cer in plasma of db/m + and db/db mice. (C) Total Hex1Cer level in response to GW501516 in plasma of db/m + and db/db mice. (D) Effect of GW501516 on the saturation of Hex1Cer in plasma of db/m + and db/db mice. (E) Effect of GW501516 on individual Hex1Cer. (F) Total SM level in response to GW501516 in plasma of db/m + and db/db mice. (G) Effect of GW501516 on individual SM. Results are means ± SEM, n = 9. Data were analyzed by one-way ANOVA in (A)-(G) with Tukey's post-hoc tests or student's t-test (A)-(E) and (G): ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001, analyzed by one-way ANOVA. |P: student's t-test.

  DownLoad: Full-Size Img PowerPoint

  We also investigated the downstream lipids of Cer, Hex1Cer and SM. Hex1Cer are formed by the attachment of a single sugar molecule, either glucose or galactose, to the ceramide structure catalyzed by glucosylceramide synthase (gcs) (Fig. S3). In the present study, we observed a significant decrease in the concentration of total Hex1Cer in the plasma of db/db mice relative to that of the healthy controls. Notably, administration of the PPARδ agonist GW501516 for six weeks elicited a significant elevation in the total Hex1Cer levels (Fig. 4C), affecting both monounsaturated and polyunsaturated Hex1Cer species in the plasma of db/m + and db/db mice (Fig. 4D). Further examination of the Hex1Cer fatty acyl chain composition revealed that C40 and C42 fatty acyl chains were predominantly present in plasma Hex1Cer. The concentrations of these Hex1Cer were profoundly diminished in the db/db mice, a trend which was notably counteracted by treatment with GW501516 (Fig. 4E). Recent studies analyzing human plasma samples have revealed that individuals with T2DM exhibit reduced levels of Hex1Cer [40]. This finding suggests that the observed upregulation of Hex1Cer could reflect a beneficial impact of GW501516 on T2DM management.

  Additionally, the concentration of SM, a lipid species synthesized by attaching a phosphocholine group to a ceramide backbone, was found to be considerably lower in the plasma of db/db mice (Fig. 4F). This reduction was particularly pronounced in SM species containing MUFA and PUFA fatty acyl chains (Fig. S4D in Supporting information). The reason might be due to decreased SM synthesis from Cer. While the overall level of SM was not notably altered by GW501516 treatment, the concentrations of specific SM species, including SM(d44:6), SM(44:4), SM(42:1), and SM(42:2), were significantly increased in the plasma upon GW501516 administration, especially on db/db mice (Fig. 4G). The results indicated that PPARδ activation by GW501516 may help to restore the imbalance of SM on db/db mice.

  In this study, the results of three sphingolipids including Cer, Hex1Cer and SM indicated that Cer was accumulated in the plasma of db/db mice. Consequently, its down-stream lipid species Hex1Cer and SM were significantly reduced in response to T2DM. The findings indicated that the catabolism rate of Cer might be reduced in response to diabetes. Although part of the levels of Hex1Cer and SM were significantly reversed by the treatment of PPARδ agonist GW501516, the levels of Cer remained substantially up-regulated. The results demonstrated that Cer generation might be substantially enhanced by GW501516. Considering that the accumulation of Cer is negatively correlated with insulin sensitivity, the findings may indicate that GW501516 could exert a negative effect on db/db mice through up-regulating circulating Cer level. Therefore, regulatory effect of GW501516 on the ceramide metabolism pathway is intricate, more studies are needed to explore the mechanism of PPARδ agonists GW501516 in ceramides regulation in the context of T2DM.

  In summary, in this study, PPARδ activation by GW501516 was investigated as a treatment for metabolic disorders like T2DM, and has shown benefits such as improved insulin sensitivity and lipid regulation. The study's lipid profiling analysis revealed significant alterations in 19 lipid subclasses, notably increasing TG, PG, and free FA in plasma of db/db mice, while GW501516 treatment reduced toxic FA levels, suggesting improved insulin resistance. Additionally, the regulatory effect of GW501516 on improving cholesterol homeostasis also highlights its beneficial effects in treating T2DM. However, the treatment also elevated Cer levels, potentially worsening insulin sensitivity. Despite these adverse effects, GW501516 partially restored Hex1Cer and SM levels, indicating a complex impact on homeostasis of Cer metabolism pathway in db/db diabetic mice, which may demonstrate both therapeutic potential and safety concerns. Therefore, these findings may underscore the need for further research to balance the therapeutic benefits and risks of PPARδ agonists in treating T2DM. However, this study has limitations regarding lipid identification. Some research has indicated that changes in lipid isomers at the C=C position may not align with the overall trends in the total amount of isomers. Therefore, accurately identifying lipid isomers at the C=C position is essential for fully understanding the regulatory effects of GW501516 on lipid changes in db/db mice. To address this issue, future studies will need to incorporate additional analyses, such as chemical derivatization methods [41-43], to distinguish the C=C bond location or sn-position of lipid isomers.

  In this study, a comprehensive lipid profiling study was conducted to investigate the regulatory effect of PPARδ activation by GW501516 in plasma of db/db mice. The results revealed complex regulatory effects on lipid metabolism in response to GW501516 on T2DM mice. Significant regulations were observed in various lipid subclasses following treatment by GW501516, with notable effect in alleviating the dysfunction of toxic fatty acid metabolism pathway and in improving circulating cholesterol homeostasis on db/db mice. Although GW501516 significantly restored the down-stream lipids of ceramides hexosylceramide and sphingomyelin in plasma of db/db mice, the treatment also elevated ceramide levels. Considering that the accumulation of ceramides is negatively correlated with insulin sensitivity, these findings may indicate an adverse effect. Despite these mixed outcomes, the study underscores the promising therapeutic potential of PPARδ activation in metabolic disorders, provided that safety concerns, particularly regarding long-term use, are adequately addressed. Further research is essential to balance the therapeutic benefits against the potential risks associated with GW501516 treatment.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Li Xiang: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Jingchun Shi: Writing – original draft, Visualization, Software, Funding acquisition, Formal analysis, Data curation. Zifan Zhu: Methodology, Investigation. Yi Ru: Writing – review & editing, Validation. Lin Peng: Methodology. Wei Wang: Methodology. Hongzhi Zhao: Writing – review & editing, Validation. Runhui Liu: Writing – review & editing. Yujie Pu: Validation, Methodology. Lei He: Writing – review & editing, Validation. Li Wang: Writing – review & editing, Validation, Supervision, Project administration, Formal analysis. Zongwei Cai: Writing – review & editing, Supervision, Resources, Funding acquisition.

  Acknowledgments

  We thank the financial support by Hong Kong Research Grants Council (No. T12–101/23-N) and National Natural Science Foundation of China (No. 22206160).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.110910.

  
  
• 1. [1]

     R.J. Henning, Future Cardiol. 14 (2018) 491–509. doi: 10.2217/fca-2018-0045

  2. [2]

     S. Tangvarasittichai, World J. Diabetes 6 (2015) 456.

  3. [3]

     M.C. Petersen, G.I. Shulman, Trends Pharmacol. Sci. 38 (2017) 649–665. doi: 10.1016/j.tips.2017.04.004

  4. [4]

     T. Tanaka, J. Yamamoto, S. Iwasaki, et al., Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 15924–15929. doi: 10.1073/pnas.0306981100

  5. [5]

     W. Fan, W. Waizenegger, C.S. Lin, et al., Cell Metab. 25 (2017) 1186–1193 e4. doi: 10.1016/j.cmet.2017.04.006

  6. [6]

     X.Y. Tian, W.T. Wong, N. Wang, et al., Diabetes 61 (2012) 3285–3293. doi: 10.2337/db12-0117

  7. [7]

     V.A. Narkar, M. Downes, R.T. Yu, et al., Cell 134 (2008) 405–415. doi: 10.1016/j.cell.2008.06.051

  8. [8]

     S.A. Ham, H.J. Kim, H.J. Kim, et al., Cell. Mol. Med. 14 (2010) 1747–1759. doi: 10.1111/j.1582-4934.2009.00816.x

  9. [9]

     L. Xiang, L. Wang, Y. Xia, et al., Innov. Life 2 (2024) 100065. doi: 10.59717/j.xinn-life.2024.100065

  10. [10]

      L. Xiang, J. Nie, L. Wang, et al., Chin. Chem. Lett. 32 (2021) 2197–2202. doi: 10.1016/j.cclet.2020.12.042

  11. [11]

      W.R. Oliver, J.L. Shenk, M.R. Snaith, et al., Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 5306–5311. doi: 10.1073/pnas.091021198

  12. [12]

      D.C. Magliano, A. Penna-de-Carvalho, M. Vazquez-Carrera, C.A. Mandarim-de-Lacerda, M.B. Aguila, Endocrine 50 (2015) 355–367. doi: 10.1007/s12020-015-0590-1

  13. [13]

      W. Chen, R. Gao, X. Xie, et al., Sci. Rep. 5 (2015) 1–13.

  14. [14]

      C. Zizola, P.J. Kennel, H. Akashi, et al., Am. J. Physiol. Circ. Physiol. 308 (2015) H1078–H1085. doi: 10.1152/ajpheart.00679.2014

  15. [15]

      D. Zhou, J. Jin, Q. Liu, J. Shi, Y. Hou, Eur. J. Pharmacol. 842 (2019) 248–254. doi: 10.1016/j.ejphar.2018.10.050

  16. [16]

      J. Qiu, T. Li, Z.J. Zhu, Anal. Chim. Acta 1142 (2021) 108–117. doi: 10.1016/j.aca.2020.10.058

  17. [17]

      J. Wei, X. Li, L. Xiang, et al., J. Hazard. Mater. 390 (2020) 121451. doi: 10.1016/j.jhazmat.2019.121451

  18. [18]

      Z. Pang, J. Chong, G. Zhou, et al., Nucleic Acids Res. 49 (2021) W388–W396. doi: 10.1093/nar/gkab382

  19. [19]

      F. Lyu, F. Han, C. Ge, et al., iMeta 2 (2023) e85. doi: 10.1002/imt2.85

  20. [20]

      S. Luquet, C. Gaudel, D. Holst, et al., Biochim. Biophys. Acta Mol. Basis Dis. 1740 (2005) 313–317. doi: 10.1016/j.bbadis.2004.11.011

  21. [21]

      L. Salvadó, L. Serrano-Marco, E. Barroso, X. Palomer, M. Vázquez-Carrera, Expert Opin. Ther. Targets 16 (2012) 209–223. doi: 10.1517/14728222.2012.658370

  22. [22]

      U. Dressel, T.L. Allen, J.B. Pippal, et al., Mol. Endocrinol. 17 (2003) 2477–2493. doi: 10.1210/me.2003-0151

  23. [23]

      R.L. Stephen, M.C.U. Gustafsson, M. Jarvis, et al., Cancer Res. 64 (2004) 3162–3170. doi: 10.1158/0008-5472.CAN-03-2760

  24. [24]

      A.I.S. Sobczak, C.A. Blindauer, A.J. Stewart, Nutrients 11 (2019) 2022. doi: 10.3390/nu11092022

  25. [25]

      G. Boden, Curr. Diab. Rep. 6 (2006) 177–181. doi: 10.1007/s11892-006-0031-x

  26. [26]

      P.J. Meikle, G. Wong, C.K. Barlow, et al., PLoS One 8 (2013) e74341. doi: 10.1371/journal.pone.0074341

  27. [27]

      M. Schlame, L. Horvàth, L. Vìgh, Biochem. J. 265 (1990) 79–85. doi: 10.1042/bj2650079

  28. [28]

      E. Sedlák, T. Kožár, A. Musatov, Cells 9 (2020) 2588. doi: 10.3390/cells9122588

  29. [29]

      F. Gonzalvez, E. Gottlieb, Apoptosis 12 (2007) 877–885. doi: 10.1007/s10495-007-0718-8

  30. [30]

      H.N. Ginsberg, Y.L. Zhang, A. Hernandez-Ono, Arch. Med. Res. 36 (2005) 232–240. doi: 10.1016/j.arcmed.2005.01.005

  31. [31]

      X. Ye, W. Kong, M.I. Zafar, L.L. Chen, Cardiovasc. Diabetol. 18 (2019) 48. doi: 10.1007/978-3-030-27535-8_5

  32. [32]

      P.L. Yeagle, Biochim. Biophys. Acta - Rev. Biomembr. 822 (1985) 267–287. doi: 10.1016/0304-4157(85)90011-5

  33. [33]

      C.K. Glass, J.L. Witztum, Cell 104 (2001) 503–516. doi: 10.1016/S0092-8674(01)00238-0

  34. [34]

      W. Lu, H.E. Resnick, K.A. Jablonski, et al., Diabetes Care 26 (2003) 16–23. doi: 10.2337/diacare.26.1.16

  35. [35]

      C. Park, Y. Zhang, X. Zhang, J, et al., Kidney Int. 69 (2006) 1511–1517. doi: 10.1038/sj.ki.5000209

  36. [36]

      D.J. Rader, G.K. Hovingh, Lancet 384 (2014) 618–625. doi: 10.1016/S0140-6736(14)61217-4

  37. [37]

      J.D. Horton, J.L. Goldstein, M.S. Brown, J. Clin. Invest. 109 (2002) 1125–1131. doi: 10.1172/JCI0215593

  38. [38]

      S. Galadari, A. Rahman, S. Pallichankandy, A. Galadari, F. Thayyullathil, Lipids Health Dis. 12 (2013) 98. doi: 10.1186/1476-511X-12-98

  39. [39]

      J.A. Chavez, M.M. Siddique, S.T. Wang, et al., J. Biol. Chem. 289 (2014) 723–734. doi: 10.1074/jbc.M113.522847

  40. [40]

      P. Düsing, N.N. Heinrich, B. Al-Kassou, et al., BMC Cardiovasc. Disord. 23 (2023) 454. doi: 10.1186/s12872-023-03454-x

  41. [41]

      J. Luo, Q. Wan, S. Chen, Chin. Chem. Lett. 36 (2025) 109836. doi: 10.1016/j.cclet.2024.109836

  42. [42]

      Q. Wan, Y. Xiao, G. Feng, et al., Chin. Chem. Lett. 35 (2024) 108775. doi: 10.1016/j.cclet.2023.108775

  43. [43]

      G. Feng, M. Gao, H. Chen, et al., Anal. Chem. 96 (2024) 2524–2533. doi: 10.1021/acs.analchem.3c04824

• 

  Figure 1  Lipidomics analysis of the regulatory effect of PPARδ agonist GW501516 on db/db mice in plasma. (A) PCA score plot. Volcano plots of db/db control group vs. db/m + control groups (B) and GW501516 treated group vs. control group on db/db mice (C). (D) Heat map of different classes of lipids in response to GW501516 treatment in db/m + and db/db mice. The total level of each lipid class was calculated by summing the detected lipids (Table S6) within the same class. The color bar stands for the percentage of each lipid class in the four groups. Ctrl, control group. GW, GW501516 treated group. Data were analyzed by one-way ANOVA in (D) with Tukey's post-hoc tests. ^*P < 0.05 vs. db/m + control group. ^#P < 0.05 vs. db/db control group, n = 9 in each group.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Regulatory effect of GW501516 on fatty acid metabolism pathway, PGs and TGs in plasma of db/m + and db/db mice. (A) Total acylcarnitines, (B) individual acylcarnitines. (C) The effect of GW501516 on total PG. (D) Effects of GW501516 on the unsaturation of PGs in plasma of db/m + and db/db mice. (E) Regulatory effect of GW501516 on total TGs. (F) Volcano plot of plasma TGs between db/db and db/m + control mice. (G) Volcano plot of plasma TGs in response to GW501516 treatment on db/db mice. (H) Regulatory effect of GW501516 on fatty acyl length of TGs. (I) Regulatory effect of GW501516 on the saturation of TGs in plasma of db/m + and db/db mice. AcCa, acylcarnitines. Ctrl, control group. GW, GW501516 treated group. Results are means ± SEM, n = 9. Data were analyzed by one-way ANOVA with Tukey's post-hoc tests (AA-E, H-I): ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. ^#P: student's t-test.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Regulatory effect of GW501516 in cholesterol synthesis and metabolism pathway. (A) Schematic diagram of cholesterol synthesis and metabolism pathway. (B) Total cholesterol level in response to GW501516 in plasma of db/m + and db/db mice. (C) HDL levels in response to GW501516 in db/db mice. (D) HDL/total cholesterol ratio in db/db mice. (E) Effects of GW501516 on precursors of cholesterol in plasma of db/m + and db/db mice. (F) Total cholesterol ester (ChE) levels in plasma. (G) Individual cholesterol esters in plasma in response to GW501516 treatment. Ctrl, control group. GW, GW501516 treated group. Results are means ± SEM, n = 9. Data were analyzed by one-way ANOVA in (B, E-G) with Tukey's post-hoc tests or student's t-test (C, D, G): ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. ^#P: student's t-test.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Effect of GW501516 in ceramide metabolism pathway. (A) Total ceramide level in response to GW501516 in plasma of db/m + and db/db mice. (B) Effect of GW501516 on the saturation of Cer in plasma of db/m + and db/db mice. (C) Total Hex1Cer level in response to GW501516 in plasma of db/m + and db/db mice. (D) Effect of GW501516 on the saturation of Hex1Cer in plasma of db/m + and db/db mice. (E) Effect of GW501516 on individual Hex1Cer. (F) Total SM level in response to GW501516 in plasma of db/m + and db/db mice. (G) Effect of GW501516 on individual SM. Results are means ± SEM, n = 9. Data were analyzed by one-way ANOVA in (A)-(G) with Tukey's post-hoc tests or student's t-test (A)-(E) and (G): ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001, analyzed by one-way ANOVA. |P: student's t-test.

  下载: 全尺寸图片 幻灯片
•