Over the past four decades, the number of people living with diabetes has increased from 108 million in 1980 to 537 million in 2021, of which the overwhelming majority (over 90%) were diagnosed as type 2 diabetes mellitus (T2DM). In 2021, 6.7 million deaths were caused by diabetes or its complications (International Diabetes Federation, 2022). Among the extensive T2DM-related complications, acute myocardial infarction is a life-threatening and severe complication (Rosenblit, 2019). More than one-third of T2DM patients with myocardial infarction (MI) die within 10 years, and long-term all-cause mortality and cardiovascular mortality are even higher in younger patients than in elderly patients (Singh et al., 2020; Zheng et al., 2021). Numerous studies have shown that strict glycemic control promotes a decrease in non-fatal MI (Rodriguez-Gutierrez et al., 2019). However, intensive controls are followed by severe side effects, such as hypoglycemia; therefore, effective and safe methods for controlling glycemic levels, while simultaneously reducing risk factors for MI, act as necessary interventions in treating patients with T2DM.

Glucagon-like peptide-1 receptor agonists (GLP-1RAs), such as liraglutide and dulaglutide, are widely used to treat patients with T2DM and obesity. They exert beneficial effects, including inhibition of glucagon secretion, delayed gastric emptying, decreased appetite, rare occurrence of hypoglycemia, and controlled weight gain (Helmstadter et al., 2022). In recent years, four clinical trials have shown that dulaglutide (REWIND trial) (Gerstein et al., 2019), albiglutide (HARMONY trial) (Hernandez et al., 2018), semaglutide (SUSTAIN-6 trial) (Marso et al., 2016a), and liraglutide (LEADER trial) (Marso et al., 2016b) have cardiovascular benefits in patients with T2DM, including reducing the occurrence of non-fatal MI. A meta-analysis reported that patients with T2DM benefited from different GLP-1RAs in terms of major adverse cardiac events, all-cause mortality, hospital admission for heart failure, and renal function (Sattar et al., 2021). However, although GLP-1RA therapies are approved and considered safe for treating patients with T2DM, the exerted cardiovascular protection mechanism is still not fully clear. An increasing number of studies have demonstrated that the GLP-1 receptor is expressed in numerous types of cells, including those in the cardiovascular tissues, such as endothelial cells of the left ventricle (GTExPortal, 2021). Theoretically, GLP-1 binds to its receptor, stimulating the adenylyl cyclase pathway, and leading to insulin synthesis and release. As the treatment for T2DM may not fully explain the cardiovascular protective effects of GLP-1RAs, these still must be comprehensively investigated.

Network pharmacology is a big data integration method based on numerous databases and statistical algorithms (Hong et al., 2021). It aims to investigate diseases at the systemic level and define the interaction between drugs and the body based on the equilibrium theory of biological networks (Zhang, 2016). Chronic diseases are generally caused by a complicated dysfunction of a related regulatory network instead of a single protein or gene (Nogales et al., 2022). Based on an integrated research strategy, the network pharmacology method provides a more efficient and convenient system for determining the relationship between drugs and diseases. In this study, we applied an integrated research strategy to investigate the mechanism of specific GLP-1RAs in T2DM and MI, which may provide a comprehensive interpretation of the cardiovascular protective effect of GLP-1RAs. A flow chart of the study process is shown in Figure 1.

With the keywords “type 2 diabetes,” “type 2 diabetes mellitus,” “myocardial infarction,” “non-fatal myocardial infarction,” “acute myocardial infarction,” “ST-segment elevation myocardial infarction,” and “non-ST-segment elevation myocardial infarction,” the target genes associated with T2DM and MI were retrieved from the PharmGkb (https://www.pharmgkb.org/), TTD (http://db.idrblab.net/ttd/), GeneCards (https://www.genecards.org), DrugBank (https://go.drugbank.com/), and OMIM (https://www.omim.org) databases. The PharmGKB database is a pharmacogenomic knowledge resource containing clinical information (Whirl-Carrillo et al., 2021). The TTD database provides information about the known and explored therapeutic protein and nucleic acid targets, the targeted disease, pathway information, and the corresponding drugs directed at each target (Wang et al., 2020). The GeneCards database integrates gene-centric data from more than 150 web sources, including genomic, transcriptomic, proteomic, genetic, clinical, and functional information (Stelzer et al., 2016). The DrugBank database contains information regarding drugs and drug targets (Wishart et al., 2018). The OMIM database is a comprehensive and authoritative compendium of human genes and genetic phenotypes. The five databases provided comprehensive and complementary resources for obtaining targets for the diseases. All T2DM and MI targets from the five databases were transformed into the official gene symbol format.

Complications related to diabetes result in 1.5 million deaths per year, and cardiovascular events are the primary cause of death (Collaborators, 2018). Several trials have demonstrated that GLP-1RAs protect diabetic patients from the occurrence of MI, but the underlying mechanism remains unclear. The present study provides novel and systematic explanations for how GLP-1RAs decrease the occurrence of MI in T2DM patients. We applied network pharmacology to predict 51 related targets between GLP-1RAs and diseases and filtered out seven core targets: AGT, TGFB1, STAT3, TIMP1, MMP9, MMP1, and MMP2. MAFB is an essential transcription factor that regulates the expression of all seven core targets. The GO enrichment terms mainly involved angiotensin, ECM, and platelets. For KEGG enrichment analysis, the related targets were principally enriched in RAS, complement and coagulation cascades, and the AGE-RAGE signaling pathway in diabetic complications.

Researchers have proposed various hypotheses to elucidate the therapeutic mechanism of GLP-1RAs in cardiovascular diseases. Many studies have consistently reported that several GLP-1RAs exert inhibitory effects on plaque formation, development, and rupture (Tashiro et al., 2014; Sudo et al., 2017; Wu et al., 2019; Li et al., 2020). However, evidence is insufficient to comprehensively explain the signaling pathways involved in the prophylactic effects of GLP-1RAs on MI. Our results are aligned with previous studies and hypotheses while also unveiling novel perspectives. After applying the CytoHubba plugin using a multi-algorithm, we identified seven core targets. Four targets (MMP1, MMP2, MMP9, and TIMP1) are associated with metalloproteinases. MMP1, MMP2, and MMP9 participate in ECM proteolysis, whereas TIMP1 acts as a metalloproteinase inhibitor that inhibits the function of MMP 1, MMP2, and MMP9 (Moore et al., 2012). The dynamic balance between MMPs and TIMP1 maintains myocardial ECM stability. Several reports have demonstrated that diabetes disrupts the balance of MMPs/TIMPs in the serum and related tissues (Li et al., 2013; Bastos et al., 2017; Zhou et al., 2021), which may significantly enhance the activities of MMPs and pathological remodeling of the vessel wall (Wang and Khalil, 2018), subsequently resulting in obstruction and ischemia. Several reports have shown that GLP-1, exenatide, and semaglutide reduce MMP expression (such as MMP1, MMP2, MMP9, and MMP13), which maintains intact fibrous caps and protects atheromatous plaque from rupture (Burgmaier et al., 2013; Garczorz et al., 2018; Rakipovski et al., 2018). GLP-1RAs may contribute to the reduction of atherosclerotic plaque instability and cardiac ECM degradation by maintaining the balance between MMPs and TIMPs.

The other three core targets identified were AGT, STAT3, and TGFB1. AGT-encoding angiotensinogen is an essential component of the RAS and participates in the regulation of blood pressure, body fluids, and electrolyte balance. Angiotensinogen undergoes two cleavages to form angiotensin II (Ang II), which has well-known adverse effects on the myocardium. A recent study showed that the mRNA expression of GLP-1R was considerably associated with the components of the renin-angiotensin-aldosterone system (RAAS) detected in epicardial and pericardial fat in patients with severe coronary artery disease (Haberka et al., 2021). However, the interactive regulation between GLP-1R and the RAAS is still unclear. TGFB1 mediates Ang II-induced myocardial fibrosis (Frangogiannis, 2019). Limited data have shown an inconsistent relationship between GLP-1RAs and TGFB1. Long-acting semaglutide decreased hepatic TGFB1 expression (McLean et al., 2021), whereas exendin-4 and liraglutide did not reduce TGFB1 levels in adipose tissue (Pastel et al., 2016; Pastel et al., 2017). Although GLP-1 has a beneficial effect on myocardial ECM remodeling (Robinson et al., 2015), the relationship between GLP-1RAs and TGFB1 in the heart is unclear and requires further investigation.

STAT3 responds to cytokines and growth factors. Shiraishi et al. demonstrated that GLP-1 induces macrophage transformation into the M2 phenotype, contributing to the beneficial effects of GLP-1 against diabetes (Shiraishi et al., 2012). A later study also provided consistent evidence that in the process of myocardial repair, STAT3 activation was a prerequisite for macrophage transformation to the reparative M2 phenotype (Shirakawa et al., 2018). Thus, the changes induced by GLP-1RAs contribute to a reduction in the size and instability of atherosclerotic plaques (Vinué et al., 2017), which could, to some extent, explain the lower MI mortality and incidence in patients using GLP-1RA.

The Iregulon plugin used in this study showed that the transcription factor MAFB may regulate all seven core targets, suggesting that it exerts crucial effects on GLP-RAs by interfering with T2DM and MI. In patients with T2DM, MAFB expression is significantly reduced (Guo et al., 2013). In contrast, overexpressed MAFB can upregulate some cell cycle regulators and subsequently promote human β cell proliferation (Lu et al., 2012). Many previous studies have considered MAFA to be a characteristic of human β-cell function, whereas this view is increasingly being challenged (Velazco-Cruz et al., 2019). Recently, it was suggested that MAFB could be regarded as an essential regulator of the human β-cell signature (Russell et al., 2020). However, only a few studies have tried to reveal the potential relationship between GLP-1 and MAFB. A recent study showed that exendin (9-39) accelerated the transdifferentiation from α cells to β cells by reducing MAFB expression in α cells (Zhang et al., 2019). Therefore, the relationship between GLP-1 and MAFB in β-cells warrants further investigation.

The MCODE plugin provides a practical clustering algorithm to identify the potential functional modules behind these targets. Three modules were obtained in this study and were subsequently subjected to KEGG analysis. The enrichment results mainly focused on the AGE-RAGE signaling pathway, RAS, and complement and coagulation cascades. These biological processes have been demonstrated to have a significant pathogenic relationship with T2DM and MI (Beckman et al., 2002; Husain et al., 2015; Yamagishi, 2019). Some researchers have provided evidence regarding the association between GLP-1RAs and the abovementioned results. GLP-1RAs, such as liraglutide and exenatide, attenuate RAGE expression in several cell types, especially under diabetic conditions (Zhang et al., 2016; Zhang et al., 2017; Zhang et al., 2020), suggesting that the downregulation of RAGE represents a potential mechanism of GLP-1RAs against T2DM. In module two, RAS was a significant KEGG-enriched item. GLP-1RAs play a competitive role in regulating RAS by inhibiting renin synthesis and increasing the inactive form of renin in blood circulation (Puglisi et al., 2021). The final module included three coagulation-related targets. To date, there is rare information on the direct influence of GLP-1RAs on the processes of coagulation cascades. Furthermore, hyperglycemia facilitates coagulation activation and invalidation of fibrinolytic activity in diabetic patients (Sechterberger et al., 2015). Thus, we hypothesize that the effect of GLP-1RAs on blood coagulation is mediated primarily through controlling blood glucose levels.

Several GLP-1RAs trials have demonstrated that these agonists reduce cardiovascular risk factors, including glycated hemoglobin (HbAc1) values, systolic blood pressure, and body weight (Marso et al., 2016a; Marso et al., 2016b; Hernandez et al., 2018). The median duration of these trials ranged from 1.6 to 3.8 years. Cumulative beneficial changes induced by GLP-1RAs contributed to a decrease in MI prevalence. However, the exact mechanisms underlying the contribution of GLP-1RAs remain unclear, as no cardiac tissue was detected in the trials. The GO and KEGG analyses showed that the enrichment terms mainly focused on ECM, coagulation, RAS, and endopeptidase. As most matrix metalloproteinases are elastase-type endopeptidases, mainly MMP2 and MMP9 (Shapiro, 1998), the dynamic balance of MMPs/TIMPs maintains the stabilization of the ECM; however, diabetes disturbs this balance and causes atherosclerotic plaque disruption, myocardial fibrosis, and remodeling. Four of the seven core targets were involved in this balance, and AGT and TGFB1 directly influenced fibrosis and remodeling. These pathological processes considerably induce cardiac death in diabetic cardiomyopathy and acute MI. Therefore, our results suggest that GLP-1RAs play a crucial role in regulating plaque stability, myocardial fibrosis, and remodeling. Other significant enrichment sets included coagulation, complement, and platelets. The burden of thrombus formation induced by coagulation and the complement system in the coronary artery mainly determines the MI area and clinical outcomes (Sianos et al., 2007; Sianos et al., 2010). Thus, inhibition of the coagulation cascades is an effective measure when plaque ruptures occur. To a certain extent, the related targets and enrichment terms involved in coagulation and complement provide a plausible mechanistic explanation for the fact that some GLP-1RAs reduce non-fatal MI in patients with T2DM.

Our study provides a comprehensive investigation and analysis of the multi-dimensional effects of GLP-1RAs on preventing MI in patients with T2DM, which may be mainly mediated by interfering with specific targets, biological processes, and cellular signaling pathways related to atheromatous plaque, myocardial remodeling, and thrombosis. However, this study has some limitations as it lacks a series of experiments to prove the proposed hypothesis. Accordingly, further experiments and multi-omics studies are warranted to understand the comprehensive mechanism of GLP-1RAs.