Synergistic SGLT2 and GLP-1R targeting alleviates systemic inflammation-induced and M1 monocyte-driven endothelial dysfunction in coronary artery disease

Abstract

Background and objective

Systemic residual inflammation plays a pivotal role in the pathophysiology of coronary artery disease (CAD). Cardiovascular protection by SGLT2 inhibitors (SGLT2i) and GLP-1 receptor agonists (GLP-1Ra) is associated with reduced inflammatory burden but underlying cellular mechanisms remain incompletely defined. We investigated whether SGLT2i and GLP-1Ra synergistically suppress monocyte activation and prevent both systemic inflammatory mediator-induced and monocyte-driven endothelial dysfunction in CAD.

Methods and results

Plasma and circulating monocytes were analyzed in healthy individuals (n = 20), patients with cardiovascular disease without CAD (n = 20), and patients with stable CAD (n = 55), and their effects on endothelial cell responses were assessed. CAD plasma showed increased IL-1β, IL-6, TNF-α, MCP-1, soluble ICAM-1, and VCAM-1, and a proteomic profile enriched in complement, innate immune, and extracellular matrix remodeling pathways. CAD plasma induced oxidative stress in endothelial cells, reduced nitric oxide, increased leukocyte and platelet adhesion, and enhanced procoagulant activity, correlating with circulating TNF-α and sICAM-1. CAD monocytes exhibited a metabolically activated phenotype with increased oxidative stress, mitochondrial activity, glucose and cholesterol uptake, calcium signaling, procoagulant activity, and adhesion to endothelial cells. These changes correlated with circulating TNF-α, sICAM-1, and plasma-induced endothelial dysfunction. CAD monocytes showed increased NF-κB, NOX2, and NLRP3 signaling with reduced CREB/NRF2 pathways, produced elevated levels of pro-inflammatory cytokines, while CAD monocytes-conditioned medium induced endothelial oxidative stress and blunted nitric oxide production. GLP-1Ra or SGLT2i attenuated these effects, while combined treatment provided synergistic protection, reducing CAD plasma-induced endothelial oxidative stress (~ 80%) and restoring endothelial function, reducing CAD monocytes oxidative stress (~ 82%), metabolic activation and pro-thrombotic activity, reprogramming monocytes toward anti-inflammatory phenotype and preventing CAD monocytes-induced endothelial dysfunction.

Conclusion

CAD features systemic inflammation that drives monocyte activation and endothelial dysfunction. Combined SGLT2i and GLP-1Ra synergistically suppress monocyte pro-inflammatory and pro-thrombotic activity and subsequently driven endothelial dysfunction.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12933-026-03267-0.

Research insights

What is currently known about this topic?

1. Residual inflammation drives CAD progression and outcomes.

2. Monocytes promote endothelial dysfunction in atherosclerosis.

3. SGLT2i and GLP-1Ra reduce CV events and inflammation.

What is the key research question?

Do SGLT2i and GLP-1Ra synergistically suppress monocyte-driven endothelial dysfunction in CAD?

What is new?

1. CAD monocytes show metabolic, pro-thrombotic activation.

2. Monocyte-endothelial crosstalk drives vascular dysfunction.

3. SGLT2i+GLP-1Ra synergistically restore endothelial function.

How might this study influence clinical practice?

Supports combined SGLT2i/GLP-1Ra to target inflammation in CAD.

Introduction

Coronary artery disease (CAD) remains the leading cause of morbidity and mortality worldwide despite advances in preventive and interventional strategies [1]. Beyond traditional risk factors, a growing body of evidence supports a central role for chronic low-grade inflammation in driving atherosclerosis progression and its clinical complications [2, 3]. Large clinical trials targeting inflammatory pathways have further established inflammation as a causal component of cardiovascular disease, highlighting the need to better understand the cellular mediators linking systemic inflammation to vascular dysfunction [4].

Circulating monocytes are key effectors in this process. In CAD, monocytes exhibit a pro-inflammatory phenotype characterized by enhanced cytokine production, oxidative stress, and increased adhesive and migratory capacity [5, 6]. These activated monocytes contribute to endothelial dysfunction, a hallmark of early atherogenesis and plaque progression, by promoting oxidative stress, impairing nitric oxide (NO) bioavailability, and enhancing leukocyte recruitment and thrombogenicity [7, 8]. Importantly, emerging evidence suggests that monocyte-derived inflammatory signals may directly propagate endothelial activation, establishing a pathogenic monocyte-endothelial axis that amplifies vascular injury [9]. However, the mechanisms linking circulating inflammatory mediators, monocyte activation, and endothelial dysfunction in human CAD remain incompletely understood.

In parallel, certain cardiometabolic therapies have demonstrated substantial cardiovascular benefits beyond glycemic control. Both sodium-glucose cotransporter-2 inhibitors (SGLT2i) and glucagon-like peptide-1 receptor agonists (GLP-1Ra) have consistently reduced major adverse cardiovascular events in large-scale randomized trials [10–12]. While SGLT2i have been associated with improvements in heart failure outcomes and vascular function, GLP-1Ra have shown robust anti-atherosclerotic and anti-inflammatory effects [13, 14]. Experimental studies suggest that both drug classes modulate oxidative stress, inflammatory signaling, and endothelial function, yet their direct effects on circulating immune cells, particularly monocytes, remain incompletely characterized [15, 16].

Notably, the combined use of SGLT2i and GLP-1Ra is increasingly adopted in clinical practice for high-risk patients with cardiometabolic disease, given their complementary metabolic and cardiovascular benefits [17]. Despite this, whether these therapies exert synergistic effects on cardiovascular inflammation and immune cell activation has not been systematically investigated. Their potential to cooperatively modulate monocyte-driven inflammatory responses and endothelial dysfunction, a central mechanism in atherogenesis, remains unknown.

In this study, we sought to investigate whether SGLT2 inhibition and GLP-1R activation synergistically modulate circulating pro-inflammatory monocyte activation and prevent monocyte-driven endothelial dysfunction in patients with CAD. By integrating clinical samples, proteomic profiling, and functional cellular assays, we aimed to define a mechanistic link between systemic inflammation, monocyte activation, and endothelial injury, and to determine whether combined cardiometabolic therapy can effectively interrupt this pathogenic axis.

Methods

Detailed Materials and Methods are available through Online Supplemental Material.

Population

This prospective, cross-sectional study was done using residual blood samples from indwelling catheters following coronary angiography. It was approved by the local Ethics Committee of the University Hospital of Strasbourg (DC-2024—6378) and was conducted in accordance with the Declaration of Helsinki. Informed written consent was given prior to the inclusion of subjects in the study. A total of 75 patients undergoing coronary angiography were included, of which 55 were diagnosed with coronary artery disease (CAD) according to current guideline-recommended criteria, including ≥ 50% luminal stenosis in at least one major epicardial coronary artery as defined by the American College of Cardiology/American Heart Association (ACC/AHA) and European Society of Cardiology (ESC) guidelines. The remaining 20 cardiovascular disease patients (CVD) served as an additional reference besides 20 other healthy subjects.

Experimental methods

EDTA-treated blood samples were used to isolate plasma and peripheral blood mononuclear cells and monocytes were purified from the latter. Plasma levels of circulating cytokines and cyto-adhesions were quantified using commercially available ELISA kits (R&D Systems, Minneapolis, MN, USA). Mass spectrometry was performed in frozen plasma samples derived from CAD patients (20) and from healthy controls (20) using a liquid chromatography system (Evosep One) connected to a timsToF SCP mass spectrometer equipped with Captive spray (Bruker Daltonics). The raw data were processed by DIA-NN 2.1.0 software. Protein intensities were log₂-transformed where appropriate, filtered to retain proteins quantified in at least 70% of samples, and normalized using quantile normalization prior to downstream analyses. Conditioned medium (CM) of monocytes was obtained following their culture for 24 h at a density of 250*10³/mL and CM levels of circulating cytokines were quantified using commercially available ELISA kits. Human coronary artery endothelial cells (hCAECs, # CC-2858, LONZA) were used to evaluate plasma and CM induced endothelial effects. The activity of isolated monocytes and hCAECS were assessed using specific fluorescent/colorimetric probes to identify intracellular oxidative stress, glucose and cholesterol uptake, calcium signaling, mitochondrial metabolic activity, and NO formation . Tissue factor (TF) activity was studied by the generation of thrombin on the surface of ECs using the chromogenic substrate of thrombin (ß-Ala-Gly-Arg p-nitroanilide diacetate), human platelet and monocyte adhesion were assessed using fluorometric adhesion assays, nuclear translocation of NF-κB by immunofluorescence staining, and mRNA and protein expression were evaluated using RT-qPCR and western blot analysis. Where indicated, experiments were conducted in the presence of specific pharmacological inhibitors.

Statistical analysis

Statistical analyses were performed using GraphPad Prism version 9.0 (GraphPad Software, San Diego, CA). The normality of data was assessed using the Kolmogorov–Smirnov test. Groups were compared using one- or two-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test for multiple pairwise comparisons as appropriate. Synergy was assessed by comparing the expected additive effect with the observed combined effect using an unpaired t-test or multiple t-tests with Benjamini–Hochberg correction for false discovery rate. The equation used to determine the expected additive effect was as follows: % modulation with Empa 10 nM + % modulation with Sema 10 nM, while the synergistic effect was considered as the effect promoted by Empa 10 nM and Sema 10 nM when used together in combination. Correlations were analyzed using Pearson’s test, and results are reported as the correlation coefficient (r) and coefficient of determination (r²). The sample sizes available are indicated in each Figure legend, and a P value < 0.05 was considered statistically significant.

For patient characteristics, categorical variables are represented as frequencies and percentages, and continuous variables are expressed as median and IQR values. Clinical data were analyzed using the JMP Pro16 (SAS Institute Inc., Cary, NC, USA). The comparison between CVD and CAD patients was done using the chi-squared test for categorical variables and Mann-Whitney U test for the distribution of continuous variables reported as medians (IQR).

To control for baseline confounding, inverse probability of treatment weighting (IPTW) was applied using propensity scores derived from a logistic regression model including age, sex, body mass index, hypertension, dyslipidemia and type II diabetes. Covariate balance before and after IPTW adjustment was evaluated using standardized mean differences (SMDs). An SMD < 0.1 was considered acceptable in this study. Comparisons of different markers levels between CVD and CAD were performed in the IPTW-weighted population. Clinical factors associated with studied markers were evaluated using univariable and multivariable analyses in the non-weighted cohort. Variables with P < 0.10 in univariable analysis were included in the multivariable model. Because the healthy group had substantially different baseline characteristics compared with the other groups, it was excluded from these adjusted analyses. Further, to minimize selection bias between the patient groups, propensity score matching (PSM) was utilized to construct a well-balanced cohort comparing patients with CVD to those with CAD. Propensity scores were calculated for each patient using a logistic regression model. The model included 14 baseline covariates: demographics (age, male sex, and BMI), clinical comorbidities (hypertension, dyslipidemia, type II diabetes, smoking history, chronic kidney disease, heart failure, and LVEF < 60%), and baseline medication regimens (statin, β-blocker and ACE-I/ARB). Due to the study design, history of CAD was intentionally excluded from the propensity model to prevent over-matching. A 1:1 nearest-neighbor matching algorithm without replacement was applied to yield 10 optimally paired patients from each cohort (n = 20 total matched patients).

Results

Baseline characteristics of the study population

The study included healthy subjects (n = 20), and 75 subjects who were subjected to coronary angiography of which are patients with cardiovascular disease without coronary artery disease (CVD, n = 20), and patients with established coronary artery disease (CAD, n = 55) (Table 1). Healthy participants were younger (median 27 years, IQR 23–35) than CVD (75 years, IQR 66–78) and CAD (71 years, IQR 65–77) groups. Male sex was significantly lower in healthy subjects (25%) as compared to the two other groups (70% and 75%, respectively). Median BMI was similar across groups (24, 25, and 26 kg/m², respectively). Cardiometabolic comorbidities were absent in healthy participants but frequent in patients’ groups: hypertension (CVD 65%, CAD 66%), dyslipidemia (CVD 75%, CAD 66%), type II diabetes (CVD 20%, CAD 26%), chronic kidney disease (CVD 25%, CAD 24%) and heart failure (CVD 25%, CAD 16%). History of smoking was reported in 25%, 45%, and 51% of healthy, CVD, and CAD participants, respectively. By definition, CAD was present only in the CAD group; no participant had prior acute coronary event. Cardiovascular medication use was absent in healthy participants and more common in patients’ groups. Among CVD patients, 80% received statins, 55% β-blockers, and 70% ACE inhibitors/ARBs; corresponding values in CAD were 76%, 58%, and 60%. No participant received SGLT2 inhibitors or GLP-1R agonists. According to IPTW-adjusted analysis for CVD and CAD patients’ groups, the population was considered acceptably balanced for the main demographic and clinical covariates (Table S1). Further, a total of 20 patients were successfully matched using propensity score analysis, establishing two highly comparable cohorts of 10 patients each for CVD and CAD (Table S2). The demographic profiles were heavily balanced between the matched groups, demonstrating similar mean ages (62.8 ± 4.9 years vs. 62.9 ± 4.6 years) and identical gender distributions (70.0% male in both groups). Following the matching protocol, an absolute exact balance (100% concordance) was achieved across all baseline clinical comorbidities and cardiovascular medication categories. Rates of hypertension (80.0%), dyslipidemia (90.0%), type II diabetes (40.0%) and history of smoking (50.0%) were perfectly symmetrical between the cohorts. Similarly, there were no residual discrepancies in medication usage, including statins (90.0%), β-blockers (70.0%) and ACE-I/ARB (60%). These 20 matched patients were used for subsequent analysis where indicated.

Table 1.

Baseline characteristics of the study population

Baseline characteristic	Healthy (n = 20)	CVD (n = 20)	CAD (n = 55)	P value(CVD vs. CAD)	SMD(CVD vs. CAD)
Age, years, median (IQR)	27 (23–35)	75 (66–78)	71 (65–77)	0.84	0.043
Male sex, %	8 (40)	14 (70)	41 (75)	0.77	0.102
BMI, kg/m2, median (IQR)	24 (23–28)	25 (23–30)	26 (24–29)	0.66	0.021
Hypertension, %	0 (0)	13 (65)	36 (66)	1	0.01
Dyslipidemia, %	0 (0)	15 (75)	36 (66)	0.58	0.21
Type II diabetes, %	0 (0)	4 (20)	14 (26)	0.77	0.13
History of smoking, %	5 (25)	9 (45)	28 (51)	0.80	0.118
Chronic kidney disease, %	0 (0)	5 (25)	13 (24)	1.00	0.032
Heart failure, %	0 (0)	5 (25)	9 (16)	0.50	0.214
LVEF < 60%, %	0 (0)	11 (55)	15 (27)	0.032	0.587
Coronary artery disease, %	0 (0)	0 (0)	55 (100)	<0.001	NA
Statin use, %	0 (0)	16 (80)	42 (76)	1	0.088
β-blocker use, %	0 (0)	11 (55)	32 (58)	1	0.064
ACE-I/ARB use, %	0 (0)	14 (70)	33 (60)	0.59	0.21
SGLT2i / GLP-1 Ra use, %	0 (0)	0 (0)	0 (0)	NA	<0.001

Continuous variables are presented as median (interquartile range), and categorical variables as percentages. Body mass index (BMI) is calculated as weight in kilograms divided by height in meters squared. Left ventricular ejection fraction (LVEF) refers to the proportion of participants with LVEF < 60%. ACE-I indicates angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; GLP-1 Ra, glucagon-like peptide-1 receptor agonist; SGLT2i, sodium-glucose cotransporter-2 inhibitor

Plasma of CAD patients contain elevated levels of circulating pro-inflammatory cytokines and cyto-adhesins

Circulating cytokines and endothelial activation markers were quantified in plasma from 20 healthy individuals, 20 patients with CVD, and 55 patients with stable CAD. Plasma concentrations of IL‑1β, IL‑6, TNF‑α, MCP‑1, soluble ICAM‑1, and soluble VCAM‑1 were significantly increased in CAD compared with healthy subjects (2.1-, 4.9-, 6.6-, 2-, 1.9-, and 1.8- fold, respectively) and CVD patients (all but IL-1β) which in turn had significantly higher levels of IL-1β, IL-6, TNF-α and soluble ICAM-1 compared with healthy subjects (2.1-, 2.9-, 2.7-, and 1.4-fold, respectively) (Fig. 1a). Notably, these results remained significant following IPTW adjusted analysis (Table S1). Further, multivariable analysis confirmed independent positive association of CAD with the circulating levels of the above-mentioned markers (Table S3).

Fig. 1.

Elevated circulating pro-inflammatory cytokines and cyto-adhesions and plasma proteomic alterations in CAD. a Plasma levels of interleukin-1β (IL-1β), interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α), monocyte chemoattractant protein-1 (MCP-1), soluble intracellular adhesion molecule−1 (sICAM-1) and soluble vascular cell adhesion molecule−1 (sVCAM-1) in patients with coronary artery disease (CAD), cardiovascular disease (CVD) and healthy subjects (H) measured by ELISA. b Principal component analysis (PCA) of plasma proteomic profiles showing separation of CAD and healthy samples along PC1 (22.2%) and PC2 (10.3%) (left panel). Volcano plot depicting differential protein abundance in CAD versus healthy plasma (middle panel). Significantly upregulated proteins are shown in red, downregulated in blue, and nonsignificant proteins in grey (FDR < 0.05, |log₂ fold change|≥ 0.5). Selected proteins related to inflammation, platelet activation, coagulation, and endothelial pathways are annotated. Reactome pathway gene set enrichment analysis (GSEA) based on ranked differential protein abundance identifying enrichment of immune and complement-related pathways in CAD plasma (right panel). Dot size represents the number of proteins contributing to each pathway. c Heatmap of the top differentially abundant plasma proteins (10 upregulated, 10 downregulated; row Z-scores), demonstrating disease-specific alterations in extracellular matrix, immune, and endothelial pathways. Samples are displayed in fixed order (healthy subjects followed by CAD) without clustering. Data are presented as median [25%-75%]. a) n = 20 for healthy subjects and patients with CVD and 60 for patients with CAD. b) n = 20 for healthy subjects and patients with CAD. Statistical significance was determined using one-way analysis of variance (ANOVA) followed by Sidak’s post-hoc test (a). *P < 0.05 versus Healthy, ^§P < 0.05 versus CVD. Two repetitive signs indicate P < 0.01 and three indicate P < 0.001

Global plasma proteomic profiling reveals a pro-inflammatory and pro-thrombotic signature in CAD

Unbiased proteomic profiling of plasma samples from CAD patients and healthy subjects demonstrated marked differences between the two groups. Principal component analysis revealed a clear separation between CAD and healthy plasma samples along the first two principal components, indicating disease‑specific proteomic signatures (Fig. 1b). Differential abundance analysis identified multiple proteins significantly enriched in CAD plasma, including components of the complement cascade and innate immune system, such as complement factor D (CFD), C8G, and C9, as well as proteins involved in extracellular matrix remodeling and vascular biology, including MMP23B, COL4A4, and PCOLCE. In addition, chemerin (RARRES2), an adipokine implicated in immune cell recruitment and endothelial activation, was elevated in CAD plasma. Further, Gene set enrichment analysis further demonstrated significant enrichment of pathways related to complement cascade activation, innate immune responses, and broader inflammatory signaling networks. Consistent with these observations, heatmap visualization of the most differentially abundant proteins demonstrated a clear segregation of CAD and healthy samples and highlighted coordinated alterations in extracellular matrix remodeling, immune regulation, and endothelial signaling pathways (Fig. 1c).

Combined SGLT2 inhibition and GLP‑1r activation synergistically protect human coronary endothelial cells from CAD plasma-induced dysfunction

Human coronary artery endothelial cells (hCAECs) exposed to 20% v/v CAD plasma exhibited a significant increase of 1.6- and 1.2-fold of intracellular oxidative stress compared with cells exposed to same volume of plasma from healthy individuals or propensity score matched CVD patients (Fig. 2a and c). Importantly, CAD plasma-induced oxidative stress levels in hCAECs significantly and positively correlated with circulating TNF‑α and soluble ICAM-1 concentrations (r² = 0.717 and 0.617, respectively) (Fig. 2b). To further determine the contribution of inflammatory mediators, hCAECs were pretreated with a combination of neutralizing antibodies against IL-1β, IL-6, or TNF-α before plasma stimulation, which significantly attenuated oxidative response in hCAECs (Fig. 2c). Similarly, pharmacological activation of the GLP‑1 receptor with semaglutide or inhibition of SGLT2 with empagliflozin (10 or 100 nM each) significantly attenuated CAD plasma‑induced oxidative stress in endothelial cells in a concentration dependent manner (Fig. 2c). Notably, combined treatment with semaglutide and empagliflozin (10 nM each) produced a greater reduction than predicted from additive responses, demonstrating synergistic suppression of endothelial oxidative stress reaching ~79.7% inhibition (Fig. 2d).

Fig. 2.

SGLT2i and GLP-1Ra synergistically prevent CAD plasma-induced coronary endothelial cells dysfunction and thrombogenicity. a Oxidative stress in human coronary artery endothelial cells (hCAECs) following stimulation with plasma of healthy subjects (H), plasma from patients with cardiovascular diseases (CVD) and plasma from patients with coronary artery disease (CAD). b Correlation between plasma-induced oxidative stress in hCAECs and CAD plasma TNF-α levels (left panel) or sICAM-1 levels (right panel). Oxidative stress c, monocyte e and platelet adhesion g, thrombin generation i and nitric oxide (NO) formation by hCAECs j, following stimulation with plasma from healthy subjects or CAD patients in the presence or absence of GLP-1Ra semaglutide (Sema, 10 or 100 nM), SGLT2i empagliflozin (Empa, 10 or 100 nM), combination of Sema and Empa (E + S, 10 nM each), combination of neutralizing antibodies against IL-1β, IL-6, TNF-α (nAbs, 30 ng/mL each) or VAS-2870 (VAS). Comparison of expected additive effects of Sema (10 nM) and Empa (10 nM) to their combined effects on oxidative stress d, monocyte f and platelet adhesion h to hCAECs. Data are presented as mean ± SEM. a) n = 20 for healthy subjects and patients with CVD and 60 for patients with CAD. b—i) n = 10 for all groups. Statistical significance was determined using one-way (a, c, e, g) or two-way (i, j) analysis of variance (ANOVA) followed by Tukey’s post-hoc test, Pearson correlation where r indicates coefficient of correlation, and r² coefficient of determination (b) and synergy was assessed by comparing the expected additive effect with the observed combined effect using an unpaired t-test (d, f, h). *P < 0.05 versus Healthy plasma-stimulated hCAECs, ^§P < 0.05 versus CVD plasma-stimulated hCAECs, ^#P < 0.05 versus CAD plasma-stimulated hCAECs, ⁺P < 0.05 combined versus additive effects, ^†P < 0.05 versus bradykinin-stimulated hCAECs, ^$P < 0.05 versus bradykinin stimulated CVD plasma-treated hCAECs, ^‡P < 0.05 versus bradykinin stimulated CAD plasma-treated hCAECs. Two repetitive signs indicate P < 0.01 and three indicate P < 0.001

Similarly, CAD plasma also significantly increased endothelial adhesiveness, as evidenced by enhanced human monocyte (2.8-, and 1.7-fold) and platelet adhesion to hCAECs (1.9-, and 1.3-fold) as compared to healthy plasma or propensity score matched CVD plasma, respectively (Fig. 2e and g). Both drugs attenuated these pro‑adhesive responses, while their combination resulted in significantly greater inhibition (74.4 and 83.1%, respectively), consistent with synergistic activity (Fig. 2f and h).

In addition, CAD plasma promoted increased thrombin generation at the surface of endothelial cells leading to higher procoagulant activity as compared to healthy plasma or propensity score matched CVD plasma (Fig. 2i). This pro‑thrombotic response was attenuated by either drug, while was most effectively suppressed by combined therapy (63.4% reduction).

Consistent with these findings, CAD plasma significantly impaired basal and bradykinin-induced endothelial NO production by ~40% and 28% as compared to healthy plasma or propensity score matched CVD plasma, respectively, reflecting endothelial dysfunction. On the other hand, pre-treatment with 100 nM semaglutide or empagliflozin partially restored endothelial NO production, whereas combined therapy at 10 nM each produced the strongest protective effect (> 90%) (Fig. 2j).

CAD monocytes display a metabolically activated and pro‑thrombotic phenotype that is suppressed by combined SGLT2i and GLP‑1Ra therapy

Given the pronounced inflammatory milieu observed in CAD plasma, we next assessed the functional phenotype of circulating monocytes. Monocytes isolated from CAD patients exhibited significantly elevated intracellular oxidative stress compared with monocytes from healthy individuals or CVD patients (2.3-, and 1.3-fold, respectively), which maintained significance following IPTW adjusted analysis (Fig. 3a, and Table S4). Further, multivariable analysis confirmed independent positive association of CAD with the monocytes oxidative stress (Table S4). Treatment with semaglutide or empagliflozin at 10 or 100 nM promoted concentration-dependent reduction in oxidative stress in CAD monocytes, whereas their combination produced a significantly greater reduction than predicted by additive responses, indicating synergistic inhibition reaching 81.7% reduction in oxidative stress (Fig. 3b). Importantly, oxidative stress levels in CAD monocytes positively correlated with circulating TNF‑α and ICAM-1 concentrations as well as with CAD plasma-induced endothelial oxidative stress (r² = 0.779, 0.707 and 0.890, respectively) (Fig. 3c).

Fig. 3.

SGLT2i and GLP-1Ra synergistically reduce elevated oxidative stress, metabolic activity, and procoagulant and adhesive capacity in CAD monocytes. Oxidative stress (a), mitochondrial metabolic activity (d), glucose uptake (f), intracellular calcium (g), cholesterol uptake (h), procoagulant activity (j) and adhesion capacity (k) of healthy and CAD monocytes in the presence or absence of GLP-1Ra semaglutide (Sema, 10 or 100 nM), SGLT2i empagliflozin (Empa, 10 or 100 nM), combination of Sema and Empa (E + S, 10 nM each), combination of neutralizing antibodies against IL-1β, IL-6, TNF-α (nAbs, 30 ng/mL each). Comparison of expected additive effects of Sema (10 nM) and Empa (10 nM) to their combined effects on oxidative stress (b), mitochondrial activity (d) and cholesterol uptake (i) by CAD monocytes. c Correlation between oxidative stress levels in CAD monocytes and CAD plasma TNF-α levels (left panel) or sICAM-1 levels (middle panel) and CAD plasma-induced oxidative stress in hCAECs (right panel). Data are presented as mean ± SEM. a–g n = 20 for healthy subjects and 60 for all other groups. h–k) n = 10 for all groups. Statistical significance was determined using one-way (a, d, f, g, h) or two-way (j) analysis of variance (ANOVA) followed by Tukey’s post-hoc test, unpaired t-test (k) Pearson correlation where r indicates coefficient of correlation, and r² coefficient of determination (c), and synergy was assessed by comparing the expected additive effect with the observed combined effect using an unpaired t-test (B, E, I). *P < 0.05 vs. Healthy monocytes, ^§P < 0.05 vs. CVD monocytes, ^#P < 0.05 versus CAD monocytes, ⁺P < 0.05 combined versus additive effects, ^†P < 0.05 100 nM versus 10 nM of the same treatment. Two repetitive signs indicate P < 0.01 and three indicate P < 0.001

Adding on, CAD monocytes displayed a pronounced metabolic activation phenotype. Mitochondrial activity (2.2-, and 1.5-fold), glucose uptake (1.8-, and 1.4-fold) and intracellular calcium signaling (1.9-, and 1.3-fold) were all significantly increased compared with monocytes from healthy controls or CVD patients, respectively (Fig. 3d–h). Notably, these results remained significant following IPTW adjusted analysis (Table S4). Further, multivariable analysis confirmed independent positive association of CAD with the above -measured monocytes’ metabolic activity (Table S4). Similarly, cholesterol uptake by CAD monocytes was significantly increased by 2.7-, and 1.5-fold when compared to monocytes from healthy donors or propensity score matched CVD patients, respectively (Fig. 3i). Synergistic inhibition was observed for mitochondrial activity (Fig. 3e), while SGLT2i and GLP-1Ra induced concentration-dependent reduction in glucose uptake (Fig. 3f) and intracellular calcium via cAMP signaling (Fig. 3g), respectively, pointing out an on-target effects of the modulators. Notably, cholesterol uptake was only attenuated by the synergy of the combination treatment with semaglutide and empagliflozin (59.7% reduction) while single treatment remained inefficient (Fig. 3h and i).

Functionally, CAD monocytes exhibited enhanced procoagulant activity (~ 500%) and increased adhesive capacity (2-, and 1.4-fold) compared with monocytes from healthy controls or propensity score matched CVD patients (Fig. 3j and k), consistent with a pro‑thrombotic and pro‑inflammatory phenotype. Notably, procoagulant activity of CAD monocytes was reduced by 59.7% using combined SGLT2i and GLP‑1Ra pharmacological modulators.

Combined therapy synergistically modulates inflammatory signaling networks in CAD monocytes

To define the molecular mechanisms underlying monocyte activation in CAD, gene and protein expression profiles were examined. CAD monocytes exhibited significantly increased polarization toward a pro-inflammatory phenotype, as evidenced by increased mRNA expression of pro-inflammatory markers CD80, CD86, IL1B, IL6, TNFA, and CCL2, genes involved in oxidative stress (CYBA and CYBB), and endothelial activating pathways together with reduced expression of anti-inflammatory markers CD204 and CD206, and IL10, as well as increased SLC5A2 and GLP1R as compared to monocytes from healthy donors or propensity score matched CVD patients (Fig. 4a).

Fig. 4.

SGLT2i and GLP-1Ra synergistically reduce pro-inflammatory phenotype and altered gene and protein expression in CAD monocytes. a Heatmap of gene expression and b representative immunoblots and quantification of protein expression levels in healthy and CAD monocytes in the presence or absence of GLP-1Ra semaglutide (Sema, 10 or 100 nM), SGLT2i empagliflozin (Empa, 10 or 100 nM), combination of Sema and Empa (E + S, 10 nM each) for 24h. c Comparison of expected additive effects of Sema (10 nM) and Empa (10 nM) to their combined effects on protein expression in CAD monocytes. Data are presented as mean ± SEM. A) n = 10 and B, C) n = 5. Statistical significance was determined using one-way (a, b) analysis of variance (ANOVA) followed by Tukey’s post-hoc test, and synergy was assessed by comparing the expected additive effect with the observed combined effect using multiple t-tests with Benjamini–Hochberg correction for false discovery rate (c). *P < 0.05 versus Healthy monocytes, ^#P < 0.05 versus CAD monocytes, ⁺P < 0.05 combined vs. additive effects, ^†P < 0.05 versus Sema 100 nM, ^‡P < 0.05 versus Empa 100 nM. Two repetitive signs indicate P < 0.01 and three indicate P < 0.001

At the protein level, CAD monocytes displayed enhanced activation of key inflammatory signaling pathways. Immunoblot analysis revealed increased expression of phosphorylated NF‑κB p65, NOX2, CD86, and NLRP3 along with SGLT2 and GLP-1R compared with healthy controls (Fig. 4b). In contrast, expression of proteins associated with protective or anti‑inflammatory pathways, including phosphorylated CREB (Ser133), NRF2 and CD163, was comparatively reduced as compared to monocytes from healthy donors (Fig. 4b).

Alternatively, pharmacological treatment with semaglutide or empagliflozin significantly and differentially modulated these signaling pathways. For instance, SGLT2i promoted significantly greater effects on reducing pro-inflammatory genes (IL1B, IL6, TNFA, CCL2, CYBA, CYBB) and proteins (p-p65 and NOX2), while GLP-1Ra induced significantly higher effects on promoting anti-inflammatory genes (IL10) and protein (p-CREB^Ser133). Notably, SGLT2i and GLP-1Ra significantly reduced expression of both SGLT2 and GLP-1R. Importantly, quantitative comparison of additive versus combined responses to SGLT2i and GLP-1Ra demonstrated significant synergistic modulation of all the quantified proteins (Fig. 4c).

Monocyte‑derived Inflammatory mediators promote endothelial dysfunction which is inhibited by combined therapy

To determine whether activated CAD monocytes directly contribute to endothelial dysfunction, conditioned medium (CM) derived from healthy, CVD or CAD monocytes (250 * 10³ cell/mL, 24 h) was analyzed then applied to hCAECs. ELISA results of CM revealed higher concentrations of pro‑inflammatory cytokines, including IL‑1β (2.2-, and 1.3-fold), IL‑6 (2.8-, and 1.5-fold), TNF‑α (5.5-, and 1.9-fold), and MCP‑1 (4.4-, and 1.6-fold) in CM of CAD monocytes and lower levels of anti-inflammatory IL-10 (67.3%) compared to that of healthy and CVD, respectively (Fig. 5a). Notably, these results remained significant following IPTW adjusted analysis (Table S1). Further, multivariable analysis confirmed independent association of CAD with the secretion of above-mentioned markers by monocytes (Table S5). On the other hand, pretreatment of CAD monocytes with empagliflozin and to lesser extent semaglutide significantly reduced the production of pro-inflammatory cytokines, while production of IL-10 was significantly promoted by SGLT2i and to a greater extent by GLP-1Ra. Interestingly, the combination of both modulators exerted the most pertinent effects on all cytokines.

Fig. 5.

SGLT2i and GLP-1Ra synergistically reduce pro-inflammatory signaling in CAD monocytes and prevent monocyte mediators-induced endothelial dysfunction. a Levels of interleukin-1β (IL-1β), interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α), monocyte chemotactic protein-1 (MCP-1), interleukin-10 (IL-10) in healthy or CAD monocytes conditioned medium (CM) in the presence or absence of GLP-1Ra semaglutide, SGLT2i empagliflozin, combination of Sema and Empa (E + S) measured by ELISA. b Oxidative stress in human coronary artery endothelial cells (hCAECs) following stimulation with healthy or CAD monocytes CM. c Correlation between CM-induced oxidative stress in hCAECs and oxidative stress level in CAD monocytes (left panel) and CAD plasma-induced oxidative stress in hCAECs (right panel). Oxidative stress d p65 NF-κB nuclear translocation f and nitric oxide (NO) formation g by hCAECs following stimulation with CM of healthy or CAD monocytes pretreated or not with GLP-1Ra semaglutide (Sema, 10 or 100 nM), SGLT2i empagliflozin (Empa, 10 or 100 nM), combination of Sema and Empa (E + S, 10 nM each), combination of neutralizing antibodies against IL-1β, IL-6, TNF-α. e Comparison of expected additive pre-treatment effects of Sema (10 nM) and Empa (10 nM) to their combined pre-treatment effects on CAD monocytes CM-induced hCAECs oxidative stress. Data are presented as mean ± SEM. A-C) n = 20 for healthy subjects and 60 for all other groups. d–g) n = 10 for all groups. Statistical significance was determined using one-way (a, d, f) or two-way g analysis of variance (ANOVA) followed by Tukey’s post-hoc test, unpaired t-test (b) Pearson correlation where r indicates coefficient of correlation, and r² coefficient of determination (c), and synergy was assessed by comparing the expected additive effect with the observed combined effect using an unpaired t-test (g). *P < 0.05 versus Healthy monocytes CM (A) or Healthy CM-stimulated hCAECs (b–g), ^§P < 0.05 versus CVD monocytes CM (a) or CVD CM-stimulated hCAECs (B–G), ^#P < 0.05 versus CAD monocytes CM (a) or CAD CM-stimulated hCAECs (B-G), ⁺P < 0.05 combined versus additive effects, ^†P < 0.05 versus hCAECs stimulated with Sema-treated CAD monocytes CM (b) or bradykinin-stimulated hCAECs (g), ^‡P < 0.05 versus hCAECs stimulated with Empa-treated CAD monocytes CM (b) or bradykinin stimulated CAD CM-treated hCAECs (g), ^$P < 0.05 versus bradykinin stimulated CVD CM-treated hCAECs. Two repetitive signs indicate P < 0.01 and three indicate P < 0.001

Adding on, exposure of endothelial cells to CAD monocyte CM significantly increased endothelial oxidative stress (1.8-, and 1.3-fold) compared with CM from monocytes of healthy donors or propensity score matched CVD patients (Fig. 5b and d). Notably, CM‑induced endothelial oxidative stress strongly correlated with oxidative stress levels in CAD monocytes and with CAD plasma-induced endothelial oxidative stress (r² = 0.877 and 0.826, respectively) (Fig. 5c). Conversely, pretreatment of CAD monocytes with semaglutide or empagliflozin reduced the ability of their CM to induce endothelial oxidative stress, while their combination exerted the most inhibitory effect (58.3%), significantly greater than predicted from their additive responses, demonstrating synergistic suppression of monocyte‑driven endothelial oxidative stress (Fig. 5d and e).

Further, CM of CAD monocytes promoted nuclear translocation of NF‑κB in endothelial cells and impaired endothelial NO production (Fig. 5f and g), key features of endothelial activation. These responses were attenuated by cytokine neutralization and to a greater extent by the combination of SGLT2i and GLP-1Ra.

Discussion

CAD is characterized by a systemic inflammatory and pro-thrombotic milieu

In the present study, we identified a pronounced systemic inflammatory and pro-thrombotic signature in patients with CAD, characterized by elevated circulating cytokines and endothelial activation markers. These findings are consistent with the established role of chronic inflammation in atherosclerosis [2–4] and extend current knowledge through unbiased proteomic profiling, which revealed enrichment of complement cascade components, innate immune mediators, and extracellular matrix remodeling proteins. The identification of complement proteins and chemerin further supports CAD as a state of persistent immune activation and metabolic inflammation [18, 19]. Importantly, CAD plasma was functionally active and directly induced endothelial dysfunction. Exposure of endothelial cells to CAD plasma recapitulated key features of cardiovascular injury, including increased oxidative stress, impaired NO bioavailability, enhanced leukocyte and platelet adhesion, and increased procoagulant activity. These alterations, which are central mechanisms in atherothrombosis [7, 8], were significantly attenuated by cytokine neutralization, highlighting the causal role of inflammatory mediators. Together, these findings establish a direct link between systemic inflammation and endothelial dysfunction in human CAD.

Monocyte activation links systemic inflammation to endothelial dysfunction

A major finding of this study is the identification of circulating monocytes as key intermediaries linking systemic inflammation to endothelial dysfunction. CAD monocytes exhibited an activated phenotype characterized by increased oxidative stress, enhanced mitochondrial metabolism, elevated glucose and cholesterol uptake, augmented calcium signaling, and increased procoagulant and adhesive activity, consistent with a pro-inflammatory and pro-thrombotic state. Importantly, monocyte oxidative stress strongly correlated with circulating TNF-α levels and plasma-induced endothelial oxidative stress, supporting a functional link between systemic inflammation, monocyte activation, and vascular injury. These observations align with the concept of trained innate immunity in atherosclerosis, whereby circulating monocytes acquire a persistent activated phenotype that promotes systemic inflammation, endothelial activation, and disease progression despite optimal guideline-directed medical treatment [6, 21].

Mechanistically, CAD monocytes showed activation of key inflammatory pathways, including NF-κB, NOX2, and the NLRP3 inflammasome, alongside suppression of protective pathways such as CREB and NRF2. This imbalance between pro- and anti-inflammatory signaling provides a molecular basis for sustained monocyte activation and highlights potential therapeutic targets [22, 23]. Collectively, our findings position circulating monocytes as central drivers of endothelial dysfunction in CAD.

Direct and complementary effects of SGLT2i and GLP-1Ra on monocytes

An important mechanistic insight from our study is that both SGLT2 and GLP-1R are expressed in CAD monocytes, supporting direct on-target effects of these therapies on immune cells. This finding aligns with emerging evidence that SGLT2 expression extends beyond the kidney and that GLP-1R are present in circulating immune cells [24, 25].

Functionally, SGLT2 inhibition and GLP-1R activation exerted distinct but complementary effects on monocyte signaling and function. SGLT2i primarily reduced glucose uptake, oxidative stress, and pro-inflammatory signaling pathways, including NF-κB activation and NOX2 expression, consistent with reported effects on cellular metabolism and redox balance [26, 27]. In contrast, GLP-1Ra preferentially activated cAMP signaling and enhanced anti-inflammatory responses, including CREB phosphorylation and IL-10 expression, in line with its known immunomodulatory properties [28]. Both agents also reduced expression of their respective transporter/receptor, suggesting negative feedback regulatory mechanisms that may modulate cellular responsiveness. Together, these complementary effects support the potential of combined therapy to achieve broader modulation of monocyte activation.

A notable finding was that cholesterol uptake in CAD monocytes was significantly reduced only by combined SGLT2 inhibition and GLP-1R activation, whereas single-agent treatment was ineffective. This suggests that lipid handling in activated monocytes is controlled by convergent but non-redundant pathways requiring simultaneous modulation. Given the central role of cholesterol accumulation in monocyte-to-macrophage differentiation and foam cell formation during atherogenesis [29, 30], this observation provides mechanistic insight into how combined cardiometabolic therapy may directly influence plaque development. While SGLT2 inhibitors mainly modulate cellular metabolism and oxidative stress, and GLP-1R agonists enhance anti-inflammatory cAMP-dependent signaling [25, 28], their integration appears necessary to effectively reprogram lipid uptake pathways. This synergy may involve coordinated regulation of scavenger receptor expression, intracellular cholesterol trafficking, or metabolic reprogramming of activated monocytes. Collectively, these findings identify cholesterol handling as a key point of therapeutic convergence and support the concept that dual SGLT2i/GLP-1Ra therapy exerts vascular protective effects beyond those achieved with either agent alone.

In addition to suppressing pro-inflammatory activation, combined SGLT2 inhibition and GLP-1R activation promoted a coordinated shift of CAD monocytes toward an anti-inflammatory, M2-like phenotype. Although both agents individually modulated inflammatory signaling, their combination produced a more pronounced upregulation of anti-inflammatory markers, including IL10, CD204, and CD206, together with suppression of M1-associated genes such as IL1B, IL6, TNFA, CD80, and CD86. This phenotype shift is particularly relevant given the established role of M2-like macrophages in inflammation resolution, tissue repair, and plaque stabilization [31, 32]. Mechanistically, GLP-1R activation promotes cAMP-CREB signaling, whereas SGLT2 inhibition attenuates oxidative stress and NF-κB-dependent transcription [22, 24, 26]. Our findings suggest that integration of these pathways enables more complete immune reprogramming than either intervention alone. Importantly, this anti-inflammatory shift translated into reduced endothelial activation and oxidative stress, supporting the concept that dual therapy promotes resolution of vascular inflammation rather than merely suppressing it.

Synergistic attenuation of pathological monocyte-endothelial crosstalk

A novel finding of this study is that combined SGLT2 inhibition and GLP-1R activation synergistically prevent pro-inflammatory monocyte-endothelial crosstalk. Although both therapies individually attenuated endothelial oxidative stress and monocyte activation, their combination consistently produced effects exceeding additive responses across multiple functional and molecular endpoints. While SGLT2i and GLP-1Ra have independently been shown to improve endothelial function and reduce vascular inflammation [13–15], evidence for synergistic vascular or immunomodulatory effects in human CAD has remained limited.

This synergistic interaction was particularly evident in the suppression of oxidative stress, restoration of endothelial NO bioavailability, and reduction of monocyte procoagulant activity, all key processes in atherothrombosis [7, 8, 20]. At the cellular level, combined therapy coordinately inhibited pro-inflammatory pathways, including NF-κB and NADPH oxidase-dependent oxidative stress, while enhancing anti-inflammatory signaling, consistent with the established role of redox and inflammatory pathways in vascular dysfunction [20, 22]. These findings suggest that simultaneous targeting of metabolic and inflammatory pathways restores immune-vascular homeostasis more effectively than single-agent approaches.

Importantly, conditioned medium from CAD monocytes reproduced endothelial dysfunction, confirming that monocyte-derived factors are sufficient to drive vascular injury. This is consistent with previous studies showing that activated monocytes and macrophages promote endothelial activation through cytokine release and oxidative stress [9, 33]. Pretreatment of monocytes with SGLT2i and GLP-1Ra markedly reduced cytokine secretion and attenuated their deleterious endothelial effects, supporting a causal role for monocyte-derived inflammatory mediators in propagating endothelial dysfunction.

Together, these findings provide direct mechanistic evidence for a pathogenic monocyte-endothelial axis in CAD and identify synergistic pharmacological modulation as an effective strategy to disrupt this interaction. By integrating immune and endothelial effects, our data extend current understanding of cardiometabolic therapy and suggest that combined SGLT2i/GLP-1Ra treatment may confer cardiovascular protection through coordinated suppression of inflammation-driven endothelial injury.

Positioning dual SGLT2i/GLP-1Ra therapy within emerging anti-inflammatory strategies in cardiovascular disease

Beyond the mechanistic insights provided by our study, the observed anti-inflammatory effects should be interpreted within the broader landscape of emerging anti-inflammatory strategies in cardiovascular disease. Inhibition of the IL-1β pathway with canakinumab in the CANTOS trial provided the first definitive evidence that selective suppression of innate immunity reduces cardiovascular events, establishing inflammation as a causal driver of atherosclerosis [2]. Similarly, IL-6 inhibition has been shown to reduce inflammatory biomarkers and cardiovascular risk, underscoring the central role of the IL-1β-IL-6 axis in atherothrombosis [34]. More recently, modulation of adaptive immunity using low-dose interleukin-2 has emerged as a promising strategy, with studies in Nature and Nature Medicine demonstrating expansion of regulatory T cells and attenuation of arterial inflammation in acute coronary syndromes [35, 36], highlighting the therapeutic potential of promoting immune resolution.

Broader anti-inflammatory approaches such as colchicine have also shown efficacy in large-scale trials including COLCOT and LoDoCo2, with reductions in major adverse cardiovascular events attributed to inhibition of the NLRP3 inflammasome and downstream IL-1β signaling [37–39]. However, these strategies largely target individual inflammatory pathways or broadly suppress inflammation without directly addressing the cellular drivers of vascular injury.

In contrast, our findings indicate that combined SGLT2 inhibition and GLP-1R activation exert a more integrated anti-inflammatory effect by modulating multiple levels of the immune-vascular axis. Rather than targeting a single cytokine or pathway, dual therapy suppresses pro-inflammatory signaling, promotes anti-inflammatory polarization, and disrupts monocyte-endothelial crosstalk in a synergistic manner. The induction of an M2-like phenotype, together with reduced monocyte-derived inflammatory mediator release, suggests active immune reprogramming rather than simple cytokine inhibition. This coordinated modulation of inflammation may mechanistically contribute to the cardiovascular benefits observed with these agents in clinical trials and supports the concept that combination therapy could complement existing anti-inflammatory approaches. Consistent with this concept, recent real-world data show that combined SGLT2 inhibitor and GLP-1R agonist therapy is associated with greater reductions in all-cause mortality and cardiovascular outcomes than either monotherapy alone, although causality cannot be inferred from observational analyses [40].

Safety and tolerability further distinguish these strategies. Although targeted cytokine inhibition is mechanistically precise, IL-1β blockade in CANTOS was associated with increased fatal infection and sepsis [2], while IL-6 inhibition may impair host defense and has been linked to hematologic and hepatic adverse effects [34]. Low-dose interleukin-2 remains investigational, and its long-term safety requires further clarification [35, 36]. Colchicine, although effective, is often limited by gastrointestinal intolerance, drug-drug interactions in renal impairment, and possible signals of increased non-cardiovascular mortality [37, 38].

In contrast, SGLT2 inhibitors and GLP-1R agonists have well-established safety profiles from large cardiovascular outcome trials, together with additional metabolic and renal benefits [10–12]. Although SGLT2i may increase the risk of genital infections and, rarely, euglycemic ketoacidosis, and GLP-1Ra can cause transient gastrointestinal effects, both classes are generally well tolerated and widely used in clinical practice. Their complementary mechanisms and favorable safety profiles support the feasibility of combined use, positioning dual SGLT2i/GLP-1Ra therapy as a practical and potentially more integrative strategy for long-term modulation of cardiovascular inflammation.

Translational relevance and therapeutic implications

The present findings provide important translational insight into the mechanisms underlying cardiovascular protection by SGLT2i and GLP-1Ra. Although both drug classes reduce cardiovascular events in large clinical trials, their combined use has largely been driven by metabolic considerations rather than mechanistic understanding of cardiovascular effects. Our data identify circulating monocytes as potential therapeutic targets and demonstrate that dual SGLT2i/GLP-1Ra therapy synergistically suppresses monocyte-driven endothelial dysfunction, a key process in atherothrombosis. These findings suggest that, beyond glycemic control, combined cardiometabolic therapy may directly modulate low-grade systemic inflammation and thrombotic risk. The observed synergy provides a biological rationale for early combined use of these agents in high-risk CAD patients and supports targeting immune-vascular interactions as a novel therapeutic strategy. Importantly, the synergistic effects on monocyte lipid handling and immune polarization further suggest that combined therapy may influence plaque composition and stability, providing a mechanistic basis for enhanced cardiovascular risk reduction.

Study limitations

Several limitations should be considered when interpreting our findings. First, although the study integrates clinical samples with detailed mechanistic in vitro analyses, causal relationships cannot be definitively established without in vivo validation. Second, while we demonstrate expression of SGLT2 and GLP-1R in circulating monocytes, the downstream signaling pathways mediating their effects require further characterization. In addition, the functional contribution of several overexpressed markers in CAD monocytes was not directly investigated and conclusions were based on prior literature. Third, the study population consisted of patients with stable CAD, and whether similar mechanisms operate in acute coronary syndromes remain unclear. Fourth, the extent and complexity of atherosclerotic burden were not assessed (e.g., SYNTAX score). Fifth, HbA1c was not measured, precluding assessment of long-term glycemic status. Sixth, proteomic data were not available for the CVD patient group. Seventh, in the absence of intravascular imaging to assess plaque morphology and vulnerability, the relationship between inflammation and plaque instability could not be established.

In addition, the influence of concomitant medications and comorbidities cannot be fully excluded, although the consistency of findings across multiple experimental approaches supports their robustness. Furthermore, while our data strongly support synergistic cellular effects, formal in vivo pharmacological interaction studies are needed to quantify the extent of synergy in clinical settings. Finally, the concentrations of empagliflozin and semaglutide used experimentally were lower than typical clinical plasma levels, and the combined low-concentration treatment (10 nM each) was selected based on concentration–response curves for single agents on ROS inhibition and may not directly correspond to clinically achieved levels.

Conclusion

In conclusion, our study identifies a mechanistic link between systemic inflammation, circulating monocyte activation, and endothelial dysfunction in CAD. We demonstrate that SGLT2 inhibition and GLP-1R activation exert direct and complementary effects on monocytes and synergistically disrupt pathologic monocyte-endothelial crosstalk. These findings position dual SGLT2i/GLP-1Ra therapy as a strategy that not only suppresses cardiovascular inflammation but actively reprograms immune-metabolic pathways governing endothelial function, thrombogenicity, lipid handling and immune polarization in CAD. Such convergent effects collectively define a multi-level synergistic mechanism.

Abbreviations

ACE-I

Angiotensin-converting enzyme inhibitor

ANOVA

Analysis of variance

ARB

Angiotensin receptor blocker

BMI

Body mass index

CAD

Coronary artery disease

cAMP

Cyclic adenosine monophosphate

CM

Conditioned medium

CVD

Cardiovascular disease

eGFR

Estimated glomerular filtration rate

ELISA

Enzyme-linked immunosorbent assay

GLP-1R

Glucagon-like peptide-1 receptor

GLP-1Ra

Glucagon-like peptide-1 receptor agonist

hCAECs

Human coronary artery endothelial cells

ICAM-1

Intercellular adhesion molecule-1

IL

Interleukin

IQR

Interquartile range

LVEF

Left ventricular ejection fraction

MCP-1

Monocyte chemoattractant protein-1

NF-κB

Nuclear factor kappa B

NLRP3

NOD-, LRR- and pyrin domain-containing protein 3

NO

Nitric oxide

NOX2

NADPH oxidase 2

NRF2

Nuclear factor erythroid 2–related factor 2

PCA

Principal component analysis

RT-qPCR

Reverse transcription quantitative polymerase chain reaction

sICAM-1

Soluble intercellular adhesion molecule-1

SGLT2

Sodium-glucose cotransporter 2

SGLT2i

Sodium-glucose cotransporter 2 inhibitor

TF

Tissue factor

TNF-α

Tumor necrosis factor alpha

VCAM-1

Vascular cell adhesion molecule-1

Author contributions

AM contributed to the design of this study, acquisition, analysis and interpretation of the data and drafted the initial manuscript. WF contributed to samples preparation and data acquisition. SK performed the analysis of proteomic data. Sh. Ki. performed the statistical analysis of clinical variables. CA collected patients’ clinical characteristics. DSG, MC, AN and SF contributed to samples preparation. AT, AG and AC contributed to the collection of blood samples. PP performed the mass spectrometry. SA, MHO, LJ and AG provided intellectual input. OM and VSK contributed to the conception and design of this study, interpretation of the data and the final revision of the manuscript.

Funding

The authors received no financial support for the research, authorship, and publication of this article.

Data availability

All data supporting the findings of this study are included within the article and its supplementary materials. Additional raw data are available from the corresponding author upon reasonable request, subject to ethical and privacy considerations.

Declarations

Ethics approval and consent to participate

Not Applicable.

Competing interests

OM declared grants from AstraZeneca, Medtronic and Boehringer Ingelheim and VSK. from Boehringer Ingelheim, Schwabe, and Servier. DSG and MHO were supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (IRIS RS-2025–25436621). The other authors have nothing to disclose. The authors received no financial support for the research, authorship, and publication of this article.