Characterization of a new specific mPGES-1 inhibitor in cardiovascular system

Abstract

This study investigates the effects of BI 1,029,539 (GS-248 or Vipoglanstat), a novel selective inhibitor of human microsomal prostaglandin E synthase-1 (mPGES-1), in experimental models of myocardial infarction and vascular injury using transgenic mice constitutively expressing the humanized mPGES-1 (Ptges) allele. Coronary artery occlusion was induced in mice randomized to receive vehicle, BI 1,029,539, or celecoxib either one week prior to (pre-treatment) or one week after (post-treatment) coronary artery occlusion. Wire-induced vascular injury was performed in mice randomized to receive vehicle or BI 1,029,539, and carotid arterial smooth muscle cell outgrowth was assessed ex vivo using explant cultures. Pre- and post-treatment with BI 1,029,539 did not affect mortality but significantly attenuated myocardial hypertrophy and improved left ventricular function compared to vehicle controls. In contrast, treatment with the COX-2 inhibitor, celecoxib, during the post-infarction period was associated with increased mortality. Additionally, BI 1,029,539 inhibited carotid artery smooth muscle cell migration and proliferation in explant cultures and reduced neointima formation in response to wire-induced vascular injury. These findings demonstrate a favorable cardiovascular profile for BI 1,029,539, a selective inhibitor of human mPGES-1, in experimental models of myocardial infarction and vascular injury. The results suggest its potential therapeutic benefits in mitigating cardiovascular risks.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-33453-1.

Introduction

Nonsteroidal anti-inflammatory drugs (NSAIDs), which inhibit cyclooxygenase (COX)−1 and COX-2, are widely used for their analgesic and anti-inflammatory properties. However, their use has been associated with an increased risk of cardiovascular adverse events, including myocardial infarction, stroke, hypertension, heart failure, and sudden cardiac death^1–3. These adverse effects are attributed to the broad inhibition of prostaglandin synthesis, as NSAIDs block an early step in the production of multiple prostaglandins, including prostaglandin E2 (PGE2), a key proinflammatory mediator, and prostacyclin (PGI2), an antithrombotic vasodilator^1–4.

PGE2 is the primary product of COX-2 and microsomal PGE synthase-1 (mPGES-1), both of which are upregulated in response to injury or inflammation^4,5. PGE2 synthase, a terminal prostanoid synthase, catalyzes the conversion of COX-derived PGH2 to PGE2^5–7. Three isoforms of PGES have been identified: cytosolic PGES (cPGES), mPGES-1, and mPGES-2^5–7. While cPGES and mPGES-2 are constitutively expressed in various cells and tissues, mPGES-1 is induced by proinflammatory stimuli and is functionally coupled to COX-2, playing a dominant role in inflammatory PGE2 production^5–8.

Studies using mPGES-1 knockout mice suggest that selective inhibition of mPGES-1 may offer a more favorable cardiovascular safety profile compared to COX-2 inhibitors^9–13. Deletion of mPGES-1 has been shown to enhance PGI2 production, reduce atherogenesis, attenuate ischemic-reperfusion brain injury, suppress neointimal hyperplasia following vascular injury, and mitigate oxidative stress and angiotensin II-induced abdominal aortic aneurysm formation^9–13. These findings position mPGES-1 as a promising target for the development of next-generation anti-inflammatory therapies^14,15.

Although several mPGES-1 inhibitors have been developed, only a few have demonstrated biological activity in vivo, and no mPGES-1 inhibitor is currently available for clinical use^3,14,15. A significant challenge in the development of these inhibitors lies in the differences in the amino acid sequences of mPGES-1 between humans, mice, and rats, which may have hindered translational research efforts¹⁶.

BI 1,029,539 (also known as OX-MPI or GS-248 or vipoglanstat in clinical trials) is a selective, small-molecule, non-peptide, and orally active inhibitor of human mPGES-1 with potent anti-inflammatory properties^17–19. This compound exhibits no affinity for murine or rat mPGES-1. To address this limitation, knock-in mice expressing human mPGES-1 were utilized in this study to evaluate the cardiovascular effects of BI 1,029,539 in experimental models of myocardial infarction and vascular injury. BI 1,029,539 improved left ventricular function, attenuated myocardial hypertrophy, and reduced neointima formation following vascular injury. Additionally, it inhibited smooth muscle cell migration and proliferation in ex vivo explant cultures, demonstrating its potential as a therapeutic candidate for cardiovascular protection.

Materials and methods

Animals

Animal studies were approved by the Institutional Animal Care and Use Committee at Mount Sinai Medical Center (2229A10) and complied with the US Animal Welfare Act.

Knock-in mice expressing the mPGES1 (Ptges) humanized allele were generated by Boehringer Ingelheim as previously reported^17,19 and outlined in Supplementary Material. In total, 250 homozygous humanized mPGES-1 C57Bl/6 mice (8–12 weeks of age, both genders) were used. Age and gender-matched mice were equally distributed among all study groups. Mice were group-housed under controlled conditions (21 ± 1 °C, 12-hour light/dark cycle) with free access to water and chow. All animals were observed daily for general health, and all invasive procedures were performed under aseptic conditions.

Coronary artery occlusion-induced myocardial infarction

Coronary artery occlusion in mice was induced as previously described^20,21. Briefly, mice were anaesthetized with ketamine (60 mg/kg, i.m.) plus xylazine (10 mg/kg, i.m.). The animals were intubated and ventilated with room air (tidal volume 0.4 ml, 120 breath/min) with a rodent respirator (Columbus Instruments, OH). A left sided thoracotomy was performed at the level of the fourth intercostal space. With the aid of a surgical microscope, the left anterior descending coronary artery was ligated approximately 1.0–1.5 mm from its origin with a 8 − 0 silk suture.

The successful performance of LAD occlusion was verified by visual inspection of the LV wall color change. Sham operations included thoracotomy and pericardial manipulation without LAD ligation. The incisions were then closed in layers with 5.0 Vicryl suture, and the animals were gradually weaned from the respirator, returned to their cages, and allowed access to food and water.

To evaluate the effects of BI 1,029,539, mPGES-1 knock-in mice were divided into two treatment protocols: pretreatment and post-treatment. In the pretreatment protocol, mice received vehicle (0.5% Natrosol + 0.01% TWEEN80), the mPGES-1 inhibitor BI 1,029,539 (30 mg/kg, orally, twice daily), or the COX-2 inhibitor celecoxib (30 mg/kg, orally, twice daily) starting one week prior to coronary artery occlusion and continuing for four weeks post-occlusion. In the post-treatment protocol, mice received the same treatments beginning one week after coronary artery occlusion and continuing for four weeks. The doses of the inhibitors were selected based on their established efficacy as analgesic agents in clinical and preclinical studies^17–19.

Each treatment protocol included four study groups. Mice were randomly assigned to one of four study groups: Group 1 (Sham control, n = 10), Group 2 (myocardial infarction [MI] + vehicle, negative control), Group 3 (MI + BI 1029539, test group), and Group 4 (MI + celecoxib, positive control).

Hemodynamic assessment

Hemodynamics was measured as previously described^22,23. Briefly, at 4 weeks post myocardial infarction, mice were anesthetized with ketamine (60 mg/kg, i.m.) plus xylazine (10 mg/kg, i.m.). An ultraminiature pressure catheter transducer (model SPR-1000, 1-F, Millar Instruments; Houston, TX) was inserted into the right common carotid artery for the measurement of arterial blood pressure with a Powerlab data acquisition system (ADInstruments Inc., CO). Heart rate was derived from the blood pressure signal. The mice were allowed to spontaneously ventilate. Body temperature was maintained between 37 °C and 39 °C using a heating pad. After arterial blood pressure measurement was obtained, the catheter was introduced into the left ventricle through the right carotid artery to monitor left ventricular pressure (LVP) and its first derivative (± dp/dtmax).

Assessment of infarct size and tissue harvest

At 4 weeks after myocardial infarction, the mice (28.5 ± 2.3 g) were anaesthetized and weighted. The animals were sacrificed by cervical dislocation. The hearts and lungs were removed, trimmed, and weighted. The heart/body weight ratio (HW/BW), lung/body weight ratio (LW/BW) were obtained. The hearts were then cut from the apex to base into four equal sections. The cut surfaces of each section were scanned by a color image scanner and infarct size on the surfaces of each slice were determined by Image J program by using a manual tracing method to measure white necrotic tissue (infarct area) from red tissue area (non-infarct tissues).

The measurement was repeated by two different researchers blinded to treatment. The percentage of infarction was obtained by dividing the surface area of the scar by the total surface area of the LV transverse sections. The infarct and non-infarct area of left ventricle were then separated, and tissues were snap-frozen in liquid nitrogen.

Biochemical assay

Heart tissue levels of PGE₂ and metabolite of PGI₂, 6-keto prostaglandin F_1α were determined by EIA Kit (Cayman Chemical, MI). Assays were performed according to the manufacturer’s recommendation.

Vascular injury-induced neointima formation

Mice were anaesthetized with ketamine (80 mg/kg, i.p.) plus xylazine (10 mg/kg, i.p.). The wire-induced right common carotid artery injury model was performed as previously described²⁴. Briefly, under a dissecting microscope, the right common carotid artery was exposed and separated from external carotid artery by placing two ligatures (black silk suture 6 − 0) around it. Distal ligature was tied off. Through the incision hole between two ligatures of right external carotid artery, 0.38 mm flexible angioplasty guide wire was advanced into the common carotid artery, and endothelial denudation was achieved by passing the wire three times. Then, the external carotid artery was tied off proximal to the incision hole with the proximal ligature and the skin incision was closed. Following carotid artery injury, mice were randomly assigned to receive vehicle (0.5% Natrosol + 0.01% TWEEN 80, n = 8) or BI 1026539 (30 mg/kg, p.o., b.i.d., n = 8 for 4 weeks. Sham control mice received vehicle (0.5% Natrosol + 0.01% TWEEN 80, n = 6) as well. At 4 weeks after carotid artery injury, Mice were anaesthetized as above described. Blood samples were collected before perfusion fixation, the carotid arteries were collected and embedded in paraffin.

Quantitative histomorphometry

Serial tissue sections (5 μm) were obtained from the right common carotid artery, starting at the bifurcation. Following deparaffinization, H&E staining was performed (H&, E, Sigma, St. Louis, MO, USA) and images were taken using Nikon LWD 0.52 microscope. Areas within lumen, internal and external elastic laminae were measured by using image J software to determine neointimal and medial areas. Briefly, the percentage of luminal narrowing and the intima-to-media (I/M) ratio were calculated as described elsewhere²². The following parameters were measured: LA = luminal area, area of arterial lumen; MA = medial area, original medial layer encircled by the internal elastic lamina (IEL) and external elastic lamina (EEL), VA = vessel area, total arterial area encircled by the EEL and IA = intimal area, area occupied by the neointima. From these measurements, the vessel wall thickness was expressed as follows: % Luminal Narrowing = 100*IA/(LA + IA). All assays were performed in a blinded fashion.

Biochemical analysis

Plasma levels of tumor necrosis factor α (TNF-α; R&D Systems, Minneapolis, MN), and interleukin 1β (IL-1β; BioLegend, San Diego, CA) were determined using ELISA kits by following manufacturer’s guide.

Carotid arterial smooth muscle cell (CASMC) outgrowth & immunostaining

Explant cultures for CASMC outgrowth experiment was performed as previously described²². Briefly, carotid arteries from two mPGES-1 knock-in mice were carefully removed, suspended in sterile phosphate-buffered saline. Surrounding connective tissues and adventitia were carefully removed, and the endothelial layer was removed by gently rubbing under a clean bench. The cleaned arteries were then cut into 2-mm² pieces and were planted in 12 well culture plates. The explants were cultured with Smooth Muscle Cell Growth Medium 2 (PromoCell GmbH) containing 10% fetal bovine serum and were kept in a humidified atmosphere of 5% CO2 at 37 °C.

The effects of BI 1,029,539 (0, 0.001, 0.01, 0.1, 1, and 10 µM) on CASMCs migration and proliferation was determined in cultures stimulated by growth factors: Smooth Muscle Cell Growth Medium 2 supplemented with 10% fetal bovine serum, 0.5 ng/ml human recombinant epidermal growth factor, 2 ng/ml human recombinant fibroblast growth factor, and 5 µg/ml recombinant human insulin. The medium was changed every two days. CASMCs were counted under a microscope using hemocytometer after 5 days of culture. Triplicate wells were used for each concentration. All experiments were repeated three times (n = 6).

Immunolabeling of CASMCs with α-smooth muscle actin: the cleaned mice carotid arteries were then cut into approximately 2-mm² pieces and cultured in glass-bottom confocal dishes as above described. Mice CASMCs were washed with PBS, and then fixed in acetone (− 20 °C) for 10 min. Fixed cells were subjected to immunostaining with α-smooth muscle actin as previously described^22,23. Briefly, the acetone fixed CASMCs were incubated overnight with a primary antibody to α-smooth muscle actin (α-SMA, Abcam, ab5694, Cambridge, MA). After three washes with PBST, the samples were incubated for 1 h with the FITC-labeled goat anti-rabbit IgG secondary antibody (sc-2012, Santa Cruz Biotechnology). Samples were then washed three additional times with TBST. Glass bottoms were carefully detached and mounted onto glass slides counterstained with Ultra Cruz Mounting Medium with 4’, 6-diamidino-2-phenylindole (DAPI; sc-24941, Santa Cruz Biotechnology). Fluorescence images were obtained by a Zeiss LSM 880 Confocal microscope, (Excitation wavelength:488; Confocal Magnification:63X-objective).

Statistical analysis

The results are presented as the mean ± standard error (S.E.M). Data were assessed for normality using the Shapiro–Wilk test. For datasets that passed the normality test, parametric tests (e.g., unpaired or paired Student’s t-test, one-way ANOVA with Tukey’s post-hoc test) were used. For datasets that did not meet normality, non-parametric tests (e.g., Mann–Whitney U test, Kruskal–Wallis test with Dunn’s post-hoc test) were performed. All analyses were performed using GraphPad Prism (GraphPad Software Inc., San Diego, CA, USA) and IBM SPSS Statistics for Windows, version (IBM Corp., Armonk, NY). P values < 0.05 were considered statistically significant.

Sample sizes were calculated using prior data with the same animal models, effect sizes reported in earlier studies involving mPGES-1 inhibitors, as well as feasibility constraints for long-term survival studies. The chosen sample sizes are consistent with published standards for these models. No sex-dependent differences were observed in preliminary analyses; therefore, data from males and females were pooled.

Results

Effects of BI 1,029,539 or celecoxib on myocardial infarction induced by coronary artery occlusion

In mice subjected to left coronary artery occlusion, 14 of 32 vehicle-treated, 10 of 30 BI 1,029,539-treated, and 11 of 32 celecoxib-treated animals died within the first two hours following the acute ischemic insult. The surviving mice were weaned off the ventilator and monitored for four weeks post-myocardial infarction. During this period, all vehicle-treated (18/18) and BI 1,029,539-treated (20/20) mice survived the entire experimental protocol. In contrast, 5 of 21 celecoxib-treated mice died during the four-week post-infarction period (1 on day 4, 2 on day 5, 1 on day 16, and 1 on day 18) (Table S1). Similarly, in the post-treatment protocol, 62 of 97 mice survived the initial ischemic insult and were weaned off the ventilator. Over the subsequent four weeks, all BI 1,029,539-treated mice (20/20) survived, compared to 20 of 22 vehicle-treated and 16 of 20 celecoxib-treated mice.

At four weeks post-myocardial infarction, infarct sizes were not significantly different among the study groups in either the pretreatment or post-treatment protocols (Fig. 1A and D). However, vehicle-treated animals exhibited significant increases in HW/BW and LW/BW ratios, indicative of pathological remodeling and myocardial fibrosis (Fig. 1B, C, E and F). In the pretreatment protocol, treatment with BI 1,029,539 significantly reduced HW/BW (Fig. 1B) and LW/BW (Fig. 1C), indicating attenuation of post-infarction heart and lung hypertrophy and fibrosis. Celecoxib treatment, however, did not significantly alter HW/BW or LW/BW ratios compared to vehicle-treated animals (Fig. 1B and C). In the post-treatment protocol, BI 1,029,539 similarly prevented significant increases in HW/BW and LW/BW ratios, maintaining values comparable to sham animals (Fig. 1E and F). In contrast, celecoxib treatment was associated with a significant increase in HW/BW compared to sham animals (Fig. 1F).

Fig. 1.

Infarct size, heart weight to body weight ratio (HW/BW), lung weight to body weight ratio (LW/BW) at 4 weeks following coronary artery occlusion in mPGES-1 knock-in mice. A-C pretreatment protocol; D-F therapeutic -treatment setting. All values are the mean ± SEM. N = 10–20. #p < 0.05, vs. sham group, *p < 0.05, vs. vehicle control. P-values were derived from one-way ANOVA followed by Tukey’s multiple comparison test.

Hemodynamic assessments at four weeks post-myocardial infarction revealed that mean blood pressure (MBP) was significantly reduced in vehicle-treated animals compared to sham controls (Fig. 2B and G). Treatment with BI 1,029,539 significantly restored MBP in both the pretreatment (Fig. 2B) and post-treatment (Fig. 2G) protocols. Celecoxib also improved MBP in the pretreatment protocol (Fig. 2B) but showed only modest effects in the post-treatment protocol (Fig. 2G). Heart rate remained consistent across all groups (Fig. 2A and F). In vehicle-treated animals, myocardial infarction led to a significant reduction in left ventricular peak systolic pressure (LVP) (Fig. 2C and H) and ± LV dP/dt max (Fig. 2E and J), as well as an increase in left ventricular end-diastolic pressure (LVEDP) (Fig. 2D and I). BI-1,029,539 treatment significantly improved LVP (Fig. 2C and H) and ± LV dP/dt max (Fig. 2E and J) in both protocols, while celecoxib treatment showed improvements in LVP (Fig. 2C and H) but only modest effects on ± LV dP/dt max (Fig. 2E and J). LVEDP was not significantly different across all groups in either protocol (Fig. 2D and I).

Fig. 2.

Hemodynamic parameters at 4 weeks after myocardial infarction in mPGES-1 mice treated with either vehicle, mPGES-1 inhibitor (BI 1029539) or COX-2 inhibitor (celecoxib). A-E pretreatment protocol; F-J therapeutic treatment setting. All values are the mean ± SEM. N = 10–18. #p < 0.05, ##p < 0.01, ###p < 0.001, vs. sham group, *p < 0.05, *p < 0.01, ***p < 0.001, vs. vehicle control. P-values were derived from one-way ANOVA followed by Tukey’s multiple comparison test, or Kruskal-Wallis test followed by Dunn’s post hoc tests.

Prostaglandin synthesis was analyzed to assess the effects of the selective mPGES-1 inhibitor on PGE2 and PGI2 levels in the infarct and non-infarct areas of the left ventricle. In vehicle-treated animals, there was a twofold increase in PGE2 levels in the non-infarct area and a fourfold increase in the infarct area (Fig. 3A and B). Levels of PGI2, measured as 6-keto-PGF1α, were also elevated in both the infarct and non-infarct areas of heart tissue (Fig. 3C and D). BI 1,029,539 significantly inhibited the increased synthesis of PGE2 in both infarct and non-infarct areas (Fig. 3A and B) but did not inhibit PGI2 synthesis (Fig. 3C and D). In contrast, celecoxib inhibited the synthesis of both PGE2 and PGI2 in heart tissues (Fig. 3A and D).

Fig. 3.

Levels of PGE₂ and metabolite of PGI₂, 6-keto prostaglandin F_1α, in left ventricle (pretreatment protocol) at 4 weeks following coronary artery occlusion in mPGES-1 knock-in mice. All values are the mean ± SEM. N = 6–8. #p < 0.05, ##p < 0.01, ###p < 0.001, vs. sham group; *p < 0.05, *p < 0.01, ***p < 0.001, vs. vehicle control; ‡p < 0.05, ‡‡p < 0.01, ‡‡‡p < 0.001 vs. the BI 1,029,539-treated group. P-values were derived from one-way ANOVA followed by Tukey’s multiple comparison test.

These findings demonstrate that BI 1,029,539 improves survival, attenuates pathological remodeling, and preserves cardiac function following myocardial infarction, with a distinct and favorable cardiovascular profile compared to celecoxib.

Inhibition of vascular injury-induced neointima formation in mPGES-1 knock-in mice

Wire-induced carotid arterial injury in mPGES-1 knock-in mice resulted in an extensive vascular stenosis at 4 weeks following the wire injury (Fig. 4A). The wire-injured carotid arteries from vehicle-treated animals showed a significant increase in intima-to-media ratio, as well as higher percentage of luminal narrowing, compared to arteries from sham-operated mice (Fig. 4B and C). In contrast, injured vessels from BI 1,029,539-treated mice show reduced neointima formation, with a reduction of intima-to-media ratio by 59%, and vascular stenosis by 56% compared to vehicles treated mice (Fig. 4B and C). Furthermore, BI 1,029,539-treated mice exhibited lower plasma level of cytokines TNF-α, and IL-1β, compared to vehicle-treated mice (Fig. 4D and E).

Fig. 4.

Treatment with BI 1,029,539 reduced carotid artery wire injury-induced neointimal generation, and plasma of TNF-α and IL-1β compared with vehicle treated group in mPGES-1 knock-in mice. (A) representative pictures from tissue H&E staining, (B) percentage of luminal narrowing, (C) quantification of intimal/medial ratio; (D) plasma levels of TNF-α, (E) plasma levels of IL-1β. All values are mean ± SEM; n = 6–8. # p < 0.05, ###p < 0.001, versus the control group and *p < 0.05, ***p < 0.001, versus the vehicle group. P-values were derived from one-way ANOVA followed by Tukey’s multiple comparison test.

BI 1,029,539 concentration-dependently inhibits CASMC outgrowth

In mPGES-1 knock-in mice carotid artery explant cultures, CASMCs began to migrate and proliferate from explants after 2 days in culture (Fig. 5A and B). Cells were counted after 5 days in culture. BI 1,029,539 (0.1–10 µM) significantly inhibited growth factor stimulated migration and proliferation of mice CASMCs (Fig. 5C).

Fig. 5.

Effect of BI 1,029,539 on mPGES-1 knock-in mice CASMC migration and proliferation. (A) Representative photographs of rat CASMC outgrowth at day 5. (B) Immunohistochemical staining for αSMA (as a marker for the smooth muscle cells). (C) BI 1,029,539 concentration-dependently inhibited carotid artery SMC migration and proliferation stimulated by growth factors. All values are mean ± SEM, n = 6. *p < 0.05, **p < 0.01 vs. Control. P-values were derived from one-way analysis of variance, followed by Dunnett’s multiple comparison test.

Discussion

The well-documented cardiovascular risks associated with NSAIDs have spurred significant interest in mPGES-1, the primary enzyme responsible for PGE2 biosynthesis downstream of COX-2, as a promising therapeutic target for analgesia and anti-inflammatory interventions^12–15. Despite considerable efforts over the past decades, no mPGES-1 inhibitors have yet reached market authorization.

Our previous studies demonstrated that BI 1,029,539 (also known as GS-248 or vipoglanstat), a novel and selective inhibitor of human mPGES-1, effectively reduced leukocyte infiltration and mitigated lung injury in preclinical models of endotoxin- and sepsis-induced lung injury¹⁹. Additionally, BI 1,029,539 was shown to suppress PGE2 biosynthesis while simultaneously increasing PGI2 levels in human subjects¹⁸. In the present study, we investigated the effects of BI 1,029,539 in experimental models of myocardial infarction (MI) and vascular injury, utilizing knock-in mice expressing human mPGES-1.

In a model of ischemia-induced prolonged myocardial infarction, both pre- and post-treatment with BI 1,029,539 significantly attenuated myocardial hypertrophy and improved left ventricular function without affecting mortality, compared to vehicle controls. In contrast, treatment with the COX-2 inhibitor celecoxib was associated with increased mortality during the post-infarction period. Furthermore, BI 1,029,539 inhibited smooth muscle cell migration and proliferation in carotid artery explant cultures and reduced neointima formation following wire-induced vascular injury in mice.

Distinct isoforms of PGES play differential roles in cardiovascular physiology and pathology^4,5. Cytosolic PGES, which is constitutively expressed and functionally linked to COX-1, mediates the rapid biosynthesis of PGE2 during acute myocardial ischemia and may exert cardioprotective effects²⁵. Conversely, mPGES-2, which is also constitutively expressed and not regulated by pro-inflammatory cytokines, appears to play a general role in basal PGE2 production critical for tissue homeostasis^26,27. In contrast, mPGES-1 is inducible by pro-inflammatory stimuli and functionally coupled to COX-2⁵.

Previous studies, including our own, have shown that genetic deletion of mPGES-1 enhances PGI2 formation, reduces atherogenesis, and attenuates ischemic-reperfusion brain injury without exacerbating acute cardiac ischemic injury^10,11,20. For example, Chen et al. reported improved survival and no adverse cardiac remodeling in mice with myeloid-specific mPGES-1 knockout following MI, while global deletion of mPGES-1 did not worsen post-MI cardiac dysfunction²⁸. Similarly, we previously demonstrated that genetic loss of mPGES-1, in contrast to COX-2 inhibition, did not reduce post-MI survival or exacerbate myocardial damage, as evidenced by lower levels of cardiac biomarkers such as creatine phosphokinase (CPK) and cardiac troponin-I²¹. These findings underscore the potential of mPGES-1 inhibitors to mitigate inflammation without the cardiovascular risks associated with COX-2 inhibitors.

However, conflicting reports suggest that mPGES-1 may be required for left ventricular function following MI²⁹. This highlights the need for caution when extrapolating findings from genetic knockout models to pharmacological inhibition³⁰. Notably, Zhang et al. recently demonstrated that an mPGES-1 inhibitor (CIII) protected against heart failure and adverse cardiac remodeling post-MI by augmenting the PGI2/PGE2 metabolite ratio³⁰. Consistent with these findings, our study shows that BI 1,029,539 attenuated myocardial hypertrophy and improved cardiac function without increasing mortality, unlike celecoxib, which was associated with higher post-MI mortality.

mPGES-1 inhibition reduces pro-inflammatory PGE2 while diverting the PGH2 substrate toward vasoprotective PGI2, a potent antithrombotic and vasodilatory mediator critical for vascular tone regulation^17,18,31,32. In a Phase 2 clinical trial, oral administration of BI 1,029,539 to patients with systemic sclerosis resulted in a 57% reduction in urinary PGE2 metabolites and a 50% increase in PGI2 metabolites, with a favorable safety and tolerability profile³³. In the present study, BI 1,029,539 selectively inhibited PGE2 synthesis in cardiac tissues without suppressing PGI2 production, further supporting its potential to mitigate inflammation while avoiding the cardiovascular risks associated with COX-2 inhibitors.

Restenosis following percutaneous vascular interventions is characterized by a cascade of pathological processes, including platelet aggregation, inflammatory cell infiltration, cytokine production (e.g., TNF-α, IL-1β, IL-6), smooth muscle cell proliferation, neointima formation, and vascular remodeling^34–36. Inflammation, triggered by vascular injury and mediated through autocrine and paracrine signaling, is central to this pathophysiological process^36,37.

One hallmark in vascular injury is that vascular SMCs stimulated by inflammatory mediators migrate into the media, proliferate and synthesize collagen, contributing to neointimal growth and restenosis³⁸. Prostaglandins and COX enzymes differentially modulate vascular injury responses^12,39–42. While COX-2 inhibitors have been implicated in adverse cardiovascular events^1–3, celecoxib has been shown to reduce neointimal hyperplasia in animal models and late luminal loss in patients with coronary artery disease^39–41. Conversely, deletion of the PGI2 receptor enhances platelet activation and vascular proliferation following vascular injury⁴². Deletion of mPGES-1 in mice attenuates neointimal hyperplasia by suppressing PGE2 production and redirecting PGH2 toward PGI2, impairing SMC proliferation and migration¹². These findings suggest that targeting mPGES-1 could offer therapeutic benefits in percutaneous coronary interventions¹².

In a Phase 2 clinical trial, BI 1,029,539 inhibited mPGES-1, reducing PGE2 metabolites by 57% and increasing PGI2 metabolites by 50%, with a favorable safety profile³³. Consistent with these findings, the present study demonstrates that BI 1,029,539 significantly inhibited growth factor-stimulated migration and proliferation of carotid artery SMCs in explant cultures. Furthermore, BI 1,029,539 inhibited cytokine production (TNF-α, IL-1β) and reduced neointima formation in response to wire-induced vascular injury in humanized mPGES-1 knock-in mice. Our findings support the potential utility of targeting mPGES-1 in PCI.

Study limitation

Several constraints should be acknowledged when interpreting these findings First, results are derived from a single mPGES-1 inhibitor (BI 1029539 alias GS-248 or vipoglanstat), restricting generalizability to other mPGES-1 inhibitors. Second, knock-in mice expressing human mPGES-1 may not fully recapitulate human cardiovascular physiology, limiting translational relevance. Third, studies were conducted exclusively in young adult mice without sex stratification, which may overlook age- and sex-dependent variability.

Finally, although safety and off-target effects were not comprehensively assessed in these preclinical models, BI 1,029,539 has advanced to Phase II clinical trials, indicating a favorable safety profile and enabling clinical evaluation of potential sex-related differences in drug response. These factors underscore the need for additional pharmacological studies and rigorous clinical assessment in diverse populations.

Conclusion

In summary, our findings reveal a favorable cardiovascular profile for the selective human mPGES-1 inhibitor BI 1,029,539 in preclinical models of myocardial infarction and vascular injury. These results support the hypothesis that mPGES-1 inhibitors may serve as a safer alternative to NSAIDs for treating inflammatory diseases, with the added benefit of mitigating cardiovascular risks.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

A.P. and D.W., conceived the study concept and experimental design, provided intellectual input and supervision. X.L; M.G., S.N. performed experiments, collected data and performed analysis. HF and JRZ performed the biochemical analyzes and statistical evaluation. A.P. and D.W. drafted the manuscript. All authors reviewed and edited the manuscript.

Funding

This work was supported by Boehringer Ingelheim.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. We confirm that the study is reported in accordance with ARRIVE guidelines.

Declarations

Competing interests

D.W. was a former employee (1996–1999) and consultant of BI and has continued to collaborate with Boehringer Ingelheim. A.P. is an employee of Boehringer Ingelheim. This does not alter the authors’ adherence to all the Journal policies on sharing data and materials. The remaining authors have disclosed that they do not have any conflicts of interest.

Consent for publication

not applicable.

Ethics approval and consent to participate

Animal studies were performed according to national and institutional animal care and ethical guidelines and was approved by the local board.