Preemptive cardioprotection with a small molecule in rodents that suppresses genes predictive of heart failure

Abstract

Ischemic heart disease is a leading cause of death worldwide. While percutaneous coronary intervention (PCI) restores blood flow in acute coronary syndrome (ACS), reperfusion injury exacerbates myocardial damage, contributing to heart failure (HF). Preemptive administration of a cardioprotective agent could help counter the imminent proinflammatory insult of PCI and reperfusion. Administering BT2, a small-molecule MAPK kinase/extracellular signal–regulated kinase inhibitor, 24 hours before and during ischemia in rats before reperfusion reduced infarct size by ~70% and preserved cardiac function 24 hours and 2 weeks postinjury. BT2 prevented adverse left ventricular remodeling and scarring. Single nucleus RNA sequencing (snRNA-seq) and bulk RNA-seq revealed that BT2 modulated genes associated with inflammation, fibrosis, and matrix production, especially within macrophages and myofibroblasts. BT2 suppressed macrophage and neutrophil infiltration. BT2 reduced the expression of genes in rodent hearts predictive of HF in patients with ACS, including many encoding cytokines, inflammasome components, and damage-associated molecular patterns. BT2 is a small molecule that can prevent myocardial ischemia-reperfusion injury, improve heart function, reduce cardiac fibrosis, and favorably modulate multiple key genes and biological processes in rats prognostic of HF when delivered before reperfusion. This strategy could be evaluated with high-risk unstable angina/non-ST-segment elevation myocardial infarction patients or those having an elective PCI.

Preemptive delivery of an anti-inflammatory small molecule confers cardioprotection after myocardial ischemia-reperfusion injury.

INTRODUCTION

Inflammation plays a central role in all stages of atherosclerosis, from initiation and progression to plaque formation, rupture, thrombosis, and myocardial infarction (MI) (1). Ischemic heart disease, a consequence of atherosclerosis, remains the world’s leading cause of death (2). Reducing infarct (IF) size and protecting heart function to prevent heart failure (HF) are critical goals in the treatment of acute MI (AMI) (3), particularly given the heart’s limited regenerative capability (4). Prompt primary percutaneous coronary intervention (PCI) is the standard approach in ST-segment elevation acute coronary syndrome (ACS) to restore flow, but paradoxically, reperfusion injury exacerbates inflammation (5, 6) and can account for up to 50% of the final IF size (7). Inflammation also contributes to left ventricular (LV) remodeling and fibrosis, hallmarks of postinfarction HF (8). LV remodeling occurs in approximately one-third of patients with ST-segment elevation myocardial infarction (STEMI) within 6 months of PCI (9–11), and HF develops in 20 to 30% of patients within 1 year (12). Globally, 26 million people live with HF, including 5.6 million in the United States, where the projected annual costs attributable to HF exceed US $70 billion by 2030 (13–15). Despite our growing understanding of signal transduction in cardioprotection (16–18), currently, there is no Food and Drug Administration (FDA)–approved cardioprotective drug to reduce IF size in the setting of MI (19). While not a pharmacologic, supersaturated oxygen therapy may reduce IF size in MI post-PCI. However, modest benefit, intracoronary administration, and the paucity of trials powered for HF hospitalization, long-term adverse remodeling, or mortality have limited its wider uptake (20, 21). In addition, while not FDA-approved for IF size limitation in ACS, several medications have been proposed for this purpose, including β-adrenergic blocking agents, P2Y12 antagonists, statins, and angiotensin-converting enzyme inhibitors, with varying levels of efficacy (22). While emerging anti-inflammatory drugs such as canakinumab (23), tocilizumab (24), and ziltivekimab (25), hold promise, they may leave patients vulnerable to infection. Moreover, low-dose colchicine (26) did not reduce major adverse cardiovascular events (MACE) in patients with ACS undergoing PCI and stenting after 3 years (27), and thus has controversial efficacy in the hyperacute phase of ACS and requires care when administered to patients with renal dysfunction (28). There remains a critical need for a safe and effective anti-inflammatory strategy to improve outcomes for patients with ACS undergoing revascularization.

Myocardial ischemia-reperfusion (M/IR) injury causes myocardial cell death and cardiac dysfunction, triggering ventricular dilation, cardiac fibrosis, impairment of contractile function, and HF (29). Complex molecular and cellular processes underpin these events. These involve a range of inducible proinflammatory responses including transcription factor (e.g., EGR1, KLF5), matrix metalloproteinase (e.g., MMP2, MMP3), and cytokine [e.g., interleukin-1β (IL-1β), interleukin-6 (IL-6), tumor necrosis factor–α (TNFα)] expression. The NLRP3 inflammasome is activated in cardiomyocytes and circulating inflammatory cells, causing activation of caspase-1 and consequent maturation of IL-1β (30). Release of damage-associated molecular patterns (DAMPs; e.g., S100A8, HMGB1, HSP) causes cardiomyocyte death through DAMP-sensing receptors (e.g., Toll-like receptors, TLR2, TLR4) (31) stimulating leukocyte recruitment to the IF zone. A single agent that inhibits these processes could offer value as a cardioprotective agent to prevent M/IR injury and its sequelae.

In the setting of PCI, M/IR injury triggers an immediate inflammatory response. Sudden trauma caused by PCI and subsequent reperfusion initiates a local cellular maelstrom involving injured and desquamated endothelium, platelet activation, and neutrophil and monocyte adhesion and infiltration, aggravating ischemic injury (32). It follows then that anti-inflammation therapy that only starts after PCI may not suffice to quell this rampant inflammation. The same limitation applies to preemptive anti-platelet therapy to counteract the prothrombotic insult triggered by PCI. Current STEMI guidelines require P2Y12 inhibitor [ticagrelor and/or clopidogrel, post–PCI-CLARITY (33), PCI-CURE (34), PLATO (35–37)] before PCI (38). Other examples of anti-inflammatory prophylaxis before planned clinical trauma include usage of preoperative low-dose corticosteroids to reduce postoperative complications following cardiopulmonary bypass and cardiac surgery in adults (39, 40) and neonates (41), preoperative prednisone to reduce facial swelling after mandibular surgery (42), and preoperative cyclosporine A to reduce inflammation after cataract surgery (43). Pre- and/or intraoperative corticosteroids and glucocorticoids are often embedded in enhanced recovery after surgery protocols for total knee and hip replacements or head and neck oncology procedures requiring microvascular free tissue transfer reconstruction (44–47). We hypothesized that early administration of an appropriate therapeutic agent may help counter “The Ugly” (48) acute proinflammatory insult of PCI with subsequent reperfusion.

This paper reports the postreperfusion cardioprotective effects of a preemptively administered small-molecule dibenzoxazepinone-based MAPK kinase/extracellular signal–regulated kinase (MEK/ERK) inhibitor. Previous studies by our group showed that BT2 inhibits circulating levels of proinflammatory cytokines, such as IL-1β and IL-6 (49, 50), which have high relevance to M/IR injury (51–53) and reduce inflammation and bone erosion in arthritic mice (50). Here, we show that IF size in the area at risk (AAR) in the rodent heart fell ~50 or ~70% when BT2 was delivered once or twice before reperfusion, respectively. BT2 prevented postreperfusion LV remodeling, loss of heart function, and fibrosis. BT2 reduced neutrophil, monocyte, and macrophage content in the AAR. BT2 also reduced transcription factor, cytokine, metalloproteinase, inflammasome, DAMP, DAMP-sensing receptor, and HF biomarker expression in rodent hearts including multiple genes predictive of HF in patients with ACS undergoing PCI. Preemptive administration of a cardioprotective agent like BT2 before revascularization could help counter the expected proinflammatory insult of PCI and reperfusion, a strategy that merits study in unstable angina/non–ST-segment elevation myocardial infarction (UA/NSTEMI) patients prior to intervention.

RESULTS

BT2 reduces IF size in the AAR 24 hours after M/IR injury

Since inflammation aggravates IF size and contributes to tissue damage postreperfusion (7), we evaluated the cardioprotective effects of BT2, a recently discovered anti-inflammatory agent (49). In Sprague-Dawley rats with M/IR injury, after 30 min of left anterior descending (LAD) coronary artery ligation, followed by reperfusion (Fig. 1A), BT2 (100 mg/kg, administered intraperitoneally 24 hours prior to and during ischemia) reduced IF size in the AAR by ~70%, relative to the vehicle (Veh)–treated group, despite similar AAR/LV ratios between groups (Fig. 1B). In contrast, an inactive BT2 analog, BT3 (49), had no effect on IF size. Circulating cardiac troponin I levels, a marker of myocardial injury, decreased by ~50% in the BT2-treated group compared to controls, affirming the protective effect on myocardial damage (Fig. 1C). Immunohistochemical staining revealed that BT2 treatment inhibited the activation of ERK (phosphorylated ERK, p-ERK) in the AAR, consistent with its postulated mechanism of BT2 action (Fig. 1D) (49). BT2 treatment did not affect PR or QTc interval, suggesting no adverse effect on cardiac rhythm (Fig. 1E). Mass spectrometry (MS) confirmed the serum bioavailability of BT2, with levels measuring 45 μg/ml 24 hours after M/IR injury (Fig. 1F).

Fig. 1. BT2 reduces IF size in the AAR 24 hours after M/IR injury.

(A) Sprague-Dawley rats underwent 30 min of LAD ligation then reperfusion for 24 hours or 2 weeks. BT2 (100 mg/kg) was delivered intraperitoneally 24 hours prior to ischemia and during ischemia. Representative ECG lead III traces are shown. IF denotes infarct. AAR indicated by absence of Evans blue perfusion. (B) Assessment of IF size/AAR and AAR/LV 24 hours after M/IR in rats treated with BT2 or BT3. BT3 is an inactive structural analog of BT2. One-way analysis of variance (ANOVA), n = 8 to 9 per group. Veh denotes vehicle (saline with 0.5%, v/v, Tween 80 and 0.01%, v/v, DMSO). Representative cross sections of TTC stained heart. Green outline indicates AAR. White sections within AAR indicate IF. (C) Serum cardiac troponin I levels in BT2 or Veh treated rats 24 hours after M/IR. Mann-Whitney test, n = 9 per group. Data remain statistically significant (**, t test) even when outlier is not included. (D) Representative immunohistochemical staining in AAR 24 hours after M/IR injury. BT2 was delivered during ischemia. Dystrophin, p-ERK, and DAPI staining quantified using Image-Pro Plus. Arrows indicate positive staining. Multiplexed staining shown in Figs. 1D and 4 (C and D) was performed on the same tissue sections from vehicle- or BT2-treated rats. Multiple antigens (pERK, CD68, CD4, Dystrophin) were stained and the indicated combinations are shown. DAPI was used as a nuclear stain. Mean ± SEM of n = 3 to 4 per group. Statistical significance assessed by t test. (E) Heart rate and PR and QTc interval in rats treated with BT2 or vehicle 24 hours after M/IR. PR interval was calculated from ECG lead II data. QT interval was corrected (QTc) using Bazett’s formula (QTc = QT/√RR). t test or MW test, n = 11 to 12 per group. (F) Circulating BT2 levels in rats treated with BT2 or vehicle 24 hours after M/IR determined by MS. Analysis of BT2 by high-resolution ultra-performance liquid chromatography–MS. Extracted ion chromatograms of protonated BT2 (m/z of 327.1340 ± 4 parts per million) for A. Blank, B. Sample BT2, and C. Standard (36.8 fmol). BT2 eluted at 13.9 min. Representative traces shown, n = 11 to 12 per group. ns, not significant; objective.

BT2 preserves heart function and prevents adverse remodeling

Within 24 hours, both ejection fraction (EF) and fractional shortening (FS) levels decreased in the vehicle group. In contrast, BT2 prevented loss of cardiac function (Fig. 2A). EF and FS remained low 2 weeks after reperfusion, yet BT2 retained its ability to protect against loss of heart function. There was no difference in EF or FS levels in BT2-treated rats at 2 weeks compared with those in the sham group (i.e., surgery but no M/IR injury) a fortnight prior (Fig. 2A). This result indicates sustained postreperfusion cardioprotective effects of BT2.

Fig. 2. BT2 preserves heart function and prevents adverse remodeling 24 hours and 2 weeks after M/IR injury.

(A) EF and FS 24 hours and 2 weeks after M/IR injury as determined by ultrasound. One-way ANOVA, n = 6 to 9 per group. (B) LVIDs and LVIDd 24 hours and 2 weeks after M/IR injury as determined by ultrasound. One-way ANOVA, n = 6 to 9 per group.

Ventricular remodeling and impaired contractility, like EF and FS, are hallmarks of ischemic cardiomyopathy. In line with EF and FS levels, ventricular enlargement (i.e., increased LV internal dimension at end-diastole, LVIDd) was evident in the vehicle group after 2 weeks, a change prevented by BT2. There was no difference in LVIDd at 2 weeks in the BT2 group compared with sham (Fig. 2B), indicating, again, sustained cardioprotection by BT2. Impaired contractility evident at 24 hours in the vehicle group was more profound at 2 weeks, yet this was completely inhibited by BT2 at 24 hours and 2 weeks. There was no difference in LV internal dimension at end-systole (LVIDs) in the BT2 group compared with sham at any time examined (Fig. 2B).

Single nucleus RNA sequencing analysis

To gain insights into the mechanism of action of BT2, we isolated nuclei from the AAR of rats treated with vehicle or BT2, 24 hours after M/IR injury, and performed single nucleus RNA sequencing (snRNA-seq). This technique, unlike single cell RNA-sequencing (scRNA-seq), can also facilitate the capture and assessment of gene expression in larger-sized cardiomyocytes among a range of cell types (54). To compare with bulk RNA-seq data from adjacent tissue slices, we collapsed the snRNA-seq into a pseudobulk dataset. Of 4442 genes in the dataset, there were 174 significant differentially-expressed genes (DEGs) [log₂ fold change (log₂FC) < −1 or >1, false discovery rate (FDR) < 0.05] (table S1). The number of up-regulated (log₂FC > 1, FDR < 0.05) and down-regulated (log₂FC < −1, FDR < 0.05) DEGs were 124 and 50, respectively (table S1).

Inspection of individual DEG in the collapsed snRNA dataset (Fig. 3A, top left) revealed that extracellular matrix components (Lama2, Gpc6, and Lamc1), transcription factors (Zeb1, Zfpm2, and Ebf1), a cell adhesion molecule (Sdk1), glycoprotein (B3galt1), protein kinase (Prkg1), and an effector of epithelial-to-mesenchymal transition (Rbms3) (55) were among the 10 most down-regulated DEGs. Glypican-6 (Gpc6), in particular, is a known regulator of HF progression, which controls cardiomyocyte growth through ERK signaling (56), while zinc finger E-box binding homeobox 1 (Zeb1) regulates collagen synthesis by cardiac fibroblasts and myocardial fibrosis (57). Among the 10 most up-regulated DEGs were immune and inflammatory modulators (Slfn4, Ptprc, Dock2, Spp1, Lilrb4, and Dock8), a transcription factor (Zfp710), nucleoside transporter (Slc28a2), phospholipid transporter (Atp8b4), and a tumor suppressor (Rbm47). Gene set enrichment analysis (GSEA) served as an exploratory tool (FDR < 0.25) (58) to provide an initial indication of gene pathways regulated by BT2. This analysis revealed that 1801 of 4803 gene sets were up-regulated (678 gene sets enriched at P < 0.25), whereas 3002 of 4803 gene sets were down-regulated (1310 gene sets enriched at P < 0.25). Among the top 10 up-regulated and down-regulated gene sets ranked by normalized enrichment score (NES) in GSEA analysis, the up-regulated sets were associated with cell differentiation and immune cell receptor-mediated signaling, whereas the down-regulated sets were related to the basement membrane, collagen, and extracellular matrix (Fig. 3A, top right and bottom).

Fig. 3. snRNA-seq reveals that BT2 reduces multiple inflammatory genes and biochemical pathways in a range of cell types.

(A) DEG and GSEA of the collapsed cell snRNA-seq dataset. Top left: Volcano plot showing up- and down-regulated DEGs in the collapsed DEG dataset. Vertical dashed lines indicate log₂FC < −1 or >1, FDR < 0.05. Top right: GSEA showing top 10 up- and down-regulated gene sets in the collapsed DEG dataset. A total of 1801 of 4803 gene sets are up-regulated, 578 gene sets are significant at FDR < 0.25, and 560 gene sets are enriched at P < 0.05. In contrast, 3002 of 4803 gene sets are down-regulated, 1310 gene sets are significant at FDR < 0.25, and 1018 gene sets are enriched at P < 0.05. Bottom: Representative GSEA profiles from the collapsed DEG dataset. Rat genes were remapped against human orthologs (Rat_Gene_Symbol_Remapping_Human_Orthologs_MSigDB.v7.3.chip with database c5.all.v2024.1.Hs.symbols.gmt). (B) Dot plot of all cell types and selected key gene markers that identify those cell types. The full list of all identified markers can be found in table S3. (C) Proportions of cell types and nuclei number in the AAR modulated by BT2 24 hours after M/IR injury. Left: Proportion of cells in the vehicle and BT2 groups expressed as a percentage of total cell type. Right: Absolute numbers of nuclei grouped as cell type showing the effect of BT2 versus vehicle. U prolif denotes unidentified proliferating cells. (D) snRNA-seq DimPlot of identified cell populations. All cell types were clustered from unsupervised analysis, and identities were named from gene expression profiles and gene marker analysis. Each colored dot represents one nucleus. UMAP display is limited to the 2000 most variable expressed genes. (E) Enrichment analysis and bubble plot showing enriched down-regulated pathways in five indicated cell types.

Cell cluster analysis

snRNA-seq data analysis identified 18 individual cell types based on key gene markers (Fig. 3B). Where possible, cell types were labeled with names as determined by Arduini et al. (59). A full list of all identified markers is provided in table S2. snRNA-seq revealed large changes in the proportion of certain cell types between treatments in the AAR 24 hours after M/IR injury. Figure 3C (left) shows that BT2 increased the proportion of Hmcn1⁺Pgm5⁺Npr3⁺ endothelial cells (EC2), but not Flt1⁺Dach1⁺Cyyr1⁺ endothelial cells (EC1) or Unc5c⁺Dnm3⁺Vwf⁺ endothelial cells (EC3a). BT2 also increased the proportion of Gsn⁺Dcn⁺Col3a1⁺ fibroblasts (FB1), but not Col1a1⁺Fbn1⁺Postn⁺ fibroblasts (FB3) or Diaph3⁺Adam12⁺Mki67⁺ myofibroblasts. BT2 did not change the proportion of Ryr2⁺Tnnt2⁺Rbm20⁺ cardiomyocytes or Prox1⁺Sema3a⁺Flt4⁺ lymphatic ECs. BT2 reduced the proportion of inflammatory cell types including Csf3r⁺Mctp2⁺Ipcef1⁺ neutrophils, Ptprc⁺Sirpd⁺Arhgap15⁺ monocytes, Rbm47⁺Arhgap22⁺Tbxas1⁺ macrophages, and Flt3⁺Grap2⁺Ciita⁺ T cells. Figure 3C (right) illustrates that the most abundant cell types in the AAR were macrophages and EC1, followed by Abcc9⁺Lrrc4c⁺Rgs5⁺ pericytes, FB1, EC3a, and cardiomyocytes.

Uniform manifold approximation and projection (UMAP) display revealed excellent separation between multiple cell types, subtypes, and lineages, and reduced density of neutrophils, monocytes, macrophages, and T cells (Fig. 3D). DEGs for each cell cluster were identified (table S3). Enrichment analysis of down-regulated genes (FDR < 0.05) from each cell cluster revealed that BT2 reduced cell migration, motility, and cell adhesion pathways in neutrophils, monocytes, macrophages, T cells, and myofibroblasts (table S4). On the other hand, BT2 enhanced pathways associated with autophagy in neutrophils, sodium channel regulation in monocytes, immune system processes in macrophages, receptor signaling and response to endogenous stimuli in FB1, and response to stress pathways in myofibroblasts (table S5). Pathway enrichment in the collapsed snRNA-seq dataset revealed down-regulation of a range of cellular metabolic processes, and up-regulation of immune processes and cellular signaling (table S6).

Comparison of suppressed pathways enriched across five key cell types highlights that BT2 has vastly different effects on gene expression in various cell types. BT2 had the most profound suppressive effect on genes implicated in a range of biological processes in myofibroblasts, macrophages, and monocytes, yet had minimal effect on FB1 and even less effect on neutrophils (Fig. 3E).

Macrophages, monocytes, and neutrophils

Inspection of individual DEG in macrophage cells (Fig. 4A, top left) identified LOC310926, Cblb, Lama2, and Prkg1 among the most down-regulated DEGs, and Milr1 and Slfn4 among the most up-regulated DEGs. Casitas B lymphoma-b (Cblb) is a key regulator of macrophage activation (60, 61), while mast cell immunoglobulin-like receptor 1 (Milr1), also known as allergin-1, inhibits autoantibody production via up-regulation of macrophage phagocytosis (62). GSEA revealed that 2020 of 4297 gene sets were up-regulated (one gene set enriched at FDR < 0.25), whereas 2277 of 4297 gene sets were down-regulated (24 gene sets enriched at FDR < 0.25). GSEA further showed that among the top 10 gene sets ranked by NES in macrophages were those involved in cholesterol efflux and sterol transport, whereas down-regulated gene sets included those associated with cell structure and cytoskeleton (Fig. 4A, top right and bottom, and table S7). Pathway enrichment analysis of BT2 down-regulated genes (FDR < 0.05) in macrophages identified suppression of pathways related to cell differentiation, cell signaling, locomotion, and migration (Fig. 4B and table S4). This observation agrees with lower CD68⁺ macrophage staining (Fig. 4C) in the AAR 24 hours after M/IR injury in BT2-treated rats. In contrast, BT2 had no significant effect on CD4⁺ T lymphocyte or regulatory T cell (T_reg cell) (CD4⁺FoxP3⁺) populations (Fig. 4D).

Fig. 4. DEG, GSEA, and enrichment analysis in the macrophage cluster.

(A) DEG and GSEA of the macrophage cluster. Top left: Hyperbolic volcano plot of up- and down-regulated DEGs in macrophage dataset. Two hyperbolic curves (y = 1/x) were overlaid as reference boundaries on the conventional volcano plot, with original cutoffs (log₂FC < −1 or >1, FDR < 0.05) used as baseline. Top right: GSEA showing top 10 up- and down-regulated gene sets in macrophage cluster. A total of 2020 of 4297 gene sets are up-regulated, one gene set is significant at FDR < 0.25, 89 gene sets are enriched at P < 0.01, and 250 gene sets are enriched at P < 0.05. In contrast, 2277 of 4297 gene sets are down-regulated, 24 gene sets are significant at FDR < 0.25, 150 gene sets are enriched at P < 0.01, and 283 gene sets are enriched at P < 0.05. Bottom: Representative GSEA profiles from macrophage dataset. Rat genes were remapped against human orthologs (Rat_Gene_Symbol_Remapping_Human_Orthologs_MSigDB.v7.3.chip with database c5.all.v2024.1.Hs.symbols.gmt). (B) g:profiler enrichment analysis and bubble plot showing pathways in macrophages down-regulated by BT2. Number of pathways shown was limited to 20. For the complete list, refer to tables. (C) Immunohistochemical staining in AAR 24 hours after M/IR injury. BT2 was delivered during ischemia. Dystrophin, CD68, and DAPI staining was quantified using Image-Pro Plus. Representative immunohistochemical staining shown with arrows providing examples of positive staining. Mean ± SEM of n = 3 to 4 per group. Statistical significance assessed by t test. (D) Immunohistochemical staining in AAR 24 hours after M/IR injury. BT2 was delivered during ischemia. Dystrophin, CD4, FoxP3, and DAPI staining quantified using Image-Pro Plus. Representative immunohistochemical staining shown with arrows providing examples of positive staining. Multiplexed staining shown in Figs. 1D and 4 (C and D) was performed on the same tissue sections from vehicle- or BT2-treated rats. Multiple antigens (pERK, CD68, CD4, Dystrophin) were stained and the indicated combinations are shown. DAPI was used as a nuclear stain. Mean ± SEM of n = 3 to 4 per group. Statistical significance assessed by t test.

LOC310926, Prkg1, Lama2, and Ebf1 were among the most down-regulated DEGs in the monocyte dataset (fig. S1, top left), whereas Slfn4, Zfp710, and Vps54 were among the most up-regulated DEGs. GSEA showed that 983 of 2210 gene sets were up-regulated (27 gene sets enriched at FDR < 0.25), whereas 1227 of 2210 gene sets were down-regulated (52 gene sets enriched at FDR < 0.25). GSEA analysis further revealed that among the top 10 up-regulated gene sets ranked by NES in monocytes were those associated with STAT (signal transducer and activator of transcription) signaling, natural killer cell–mediated immunity, and cell killing, while down-regulated sets were related to cell cycle transition and collagen-containing extracellular matrix (fig. S1, top right and bottom). BT2 down-regulated genes (FDR < 0.05) identified pathways that were associated with cell motility and migration (fig. S2 and table S4).

In the neutrophil dataset, Serpinb1a, Il1r1, and Nek10 were among the most down-regulated DEGs, whereas Usp32, Zfp710, and Jam1 were among the most up-regulated (fig. S3). GSEA identified 1204 of 3012 gene sets up-regulated (however, no gene sets were enriched at FDR < 0.25), whereas 1808 of 3012 gene sets were down-regulated (no gene sets were enriched at FDR < 0.25). BT2 enriched pathways were associated with autophagy and histone deacetylase recruitment (Mad-Max), whereas gene sets that were down-regulated included those associated with cell migration and motility (fig. S3).

Myofibroblasts and fibroblasts

A distinct proliferation gene signature (Mki67, Ccna2, Tk1, and Top2a) separated myofibroblasts from the FB1 and FB3 clusters (Fig. 5A, left) (63). FB3, like myofibroblasts, were Postn⁺ and Pdgfra⁺, but like FB1, did not express the proliferation gene signature that characterized the myofibroblasts (Fig. 5A, left and right). While BT2, relative to vehicle, had no effect on proliferation marker gene expression in FB1 or FB3 or myofibroblasts (Fig. 5B), there were approximately threefold more FB1 in the BT2 group than the vehicle group (Fig. 3C). In contrast, BT2 did not affect the proportion of cells in the FB3 and myofibroblast groups (Fig. 3C).

Fig. 5. DEG, GSEA, and enrichment analysis in fibroblast and myofibroblast clusters, and BT2’s effect on cardiac fibrosis.

(A) Seurat Do heatmap output of cell and proliferation markers in fibroblast and myofibroblast populations. Left: Proliferation marker expression. Right: Activation marker expression. Colors represent individual nuclei. (B) Seurat Do heatmap function output highlighting BT2’s effects on proliferation markers in fibroblast and myofibroblast populations. Vertical bars within a cell type represent individual nuclei. Veh denotes vehicle. (C) DEG and GSEA of myofibroblast cluster. Top left: Hyperbolic volcano plot of up- and down-regulated DEGs in the myofibroblast dataset. Two hyperbolic curves (y = 1/x) were overlaid as reference boundaries on the volcano plot, with baseline original cutoffs (log₂FC < −1 or >1, FDR < 0.05). Top right: GSEA showing top 10 up- and down-regulated gene sets in myofibroblast cluster. A total of 2457 of 5747 gene sets are up-regulated, one gene set is significant at FDR < 0.25, 64 gene sets are enriched at P < 0.01, and 213 gene sets are enriched at P < 0.05. In contrast, 3290 of 5747 gene sets are down-regulated, 10 gene sets are enriched at FDR < 0.25, 179 gene sets are enriched at P < 0.01, and 483 gene sets are enriched at P < 0.05. Bottom: Representative GSEA profiles from myofibroblast dataset. Rat genes were remapped against human orthologs (Rat_Gene_Symbol_Remapping_Human_Orthologs_MSigDB.v7.3.chip with database c5.all.v2024.1.Hs.symbols.gmt). (D) Enrichment analysis and bubble plot showing pathways in myofibroblasts down-regulated by BT2. Number of pathways shown was limited to 20. For the complete list, refer to tables. (E) BT2 reduces cardiac fibrosis 2 weeks after M/IR injury. Left: ST elevation between groups. Data sourced from lead III traces from six-lead ECG recorded 5 to 10 min after LAD ligation show unbiased ST-elevation between groups. t test, n = 6 per group. Middle and right: Representative cross sections of hearts from rats treated with BT2 or vehicle twice prior to reperfusion left for 2 weeks stained with Sirius Red quantified using Image Pro Plus. t test, n = 6 per group.

In myofibroblasts, Fgd5, Prkg1, Plpp3, and Il4r were among the most down-regulated DEGs, whereas Fam111a, Actn1, and Mfap5 were up-regulated (Fig. 5C, top left). Faciogenital dysplasia 5 (Fgd5)–antisense 1 expression can reduce IF size, enhance cardiac function, and inhibit cardiac fibrosis, whereas protein kinase G1 (Prkg1) activity is associated with cardiac fibrosis (64). GSEA revealed that 2457 of 5747 gene sets were up-regulated (one gene set enriched at FDR < 0.25), whereas many more (3290 of 5747) gene sets were down-regulated (10 gene sets enriched at FDR < 0.25). GSEA revealed that among the top 10 gene sets ranked by NES as up-regulated in myofibroblasts were those involved in cell structure and replication, whereas gene sets that were down-regulated included those associated with cell signaling and extracellular matrix production (Fig. 5C, top right and bottom). BT2 down-regulated genes (FDR < 0.05) from myofibroblasts were enriched in pathways associated with cellular development, signal transduction, cell motility, and migration (Fig. 5D and table S4). GSEA further showed that BT2 significantly inhibited DEGs associated with the MAPK cascade (relative to vehicle) in myofibroblasts (FDR < 0.25), but not in fibroblasts, monocytes, macrophages, neutrophils, cardiomyocytes, ECs, T cells, or vascular smooth muscle cells (fig. S4).

In FB1 (fig. S5, top left), Zeb1, Pcdha4, and Tmlhe were among the most down-regulated DEGs, whereas Aox3, Fhl2, and Zfp710 were among the most up-regulated. GSEA revealed that 1789 of 4171 gene sets were up-regulated (48 gene sets enriched at FDR < 0.25), whereas 2382 of 4171 gene sets were down-regulated (0 gene sets enriched at FDR < 0.25). GSEA revealed that among the top 10 gene sets ranked by NES as up-regulated in FB1 were those involved in cell integrity and structure, whereas gene sets that were down-regulated included those associated with chromosomal structure and aging (fig. S5, top right and bottom). BT2 up-regulated genes from FB1 (FDR < 0.05) were enriched in pathways associated with cell surface receptor signaling, cell adhesion, and anatomical structure (fig. S6).

BT2 reduces cardiac fibrosis 2 weeks after M/IR injury

Interstitial accumulation of extracellular matrix proteins such as collagen typifies cardiac fibrosis and contributes to causing cardiac dysfunction (65). Sirius Red staining visualized collagen in cross sections of hearts from rats treated with BT2 or vehicle 2 weeks after M/IR. Despite no differences in ST segment elevation between Veh and BT2 groups following coronary artery ligation (Fig. 5E, left), BT2 reduced scarring in the LV by ~50% after 2 weeks (Fig. 5E, middle and right). Inhibition of cardiac fibrosis by BT2 after 2 weeks agrees with its ability to preserve heart function and prevent adverse remodeling at this later time (Fig. 2B).

Bulk RNA-seq analysis

To gain further insights into the mechanism of action of BT2, we isolated total RNA from the AAR of rats treated with vehicle or BT2, 4 or 24 hours after M/IR injury, and performed bulk next-generation RNA-seq, an approach that includes cytoplasmic transcripts. Multidimensional scaling (MDS) (Fig. 6A) showed clear separation of gene expression profiles between treatment groups. Of 14,927 genes in the dataset, there were 975 significant DEGs (log₂FC < −1 or >1, FDR < 0.05) at 4 hours. The number of up-regulated (log₂FC > 1, FDR < 0.05) and down-regulated (log₂FC < −1, FDR < 0.05) genes were 339 and 636, respectively (table S8). At 24 hours, there were substantially more (2297) DEGs (log₂FC < −1 or >1, FDR < 0.05); the number of up-regulated (log₂FC > 1, FDR < 0.05) and down-regulated (log₂FC < −1, FDR < 0.05) genes were 1200 and 1097, respectively (table S9).

Fig. 6. Bulk RNA-seq reveals BT2 reduces multiple inflammatory genes and biochemical pathways.

(A) MDS plot. RNA was isolated from the AAR of rats treated with vehicle or BT2, 4 or 24 hours after M/IR injury, and then processed for bulk RNA-seq and bioinformatics analysis. (B) Four hours DEG data plotted as volcano plots. Log₂FC < −1 or >1, FDR < 0.05. (C) Twenty-four hours DEG data plotted as volcano plots. Log₂FC < −1 or >1, FDR < 0.05. (D) Immunohistochemical staining of Trpm2 in the AAR 24 hours after M/IR injury. BT2 was delivered once during ischemia. Integrated optical density (IOD) and tissue area were quantified using Image-Pro Plus. Representative immunohistochemical staining is shown with arrows providing examples of positive staining. Data represent means ± SEM of n = 3 to 4 rats per group. Statistical significance was assessed by t test. (E) KEGG enrichment analysis (performed on down-regulated genes, FDR < 0.05) performed with 4 and 24 hours bulk RNA-seq DEG data. (F and G) Western blotting was performed with extracts of human THP-1 cells incubated with or without 20 ng/ml of IL-1β for various times and pretreated with 10 μM BT2 for 24 hours. Alternatively, Western blotting was performed with extracts of THP-1 cells incubated with or without 20 ng/ml of IL-1β for 30 min and pretreated with various concentrations of BT2 for 24 hours. Membranes were incubated with antibodies to phospho-ERK, ERK, FOS, or β-actin followed by secondary antibodies. Veh denotes medium containing 0.01% DMSO. Approximate positions of molecular weight markers are shown. Data represent three biologically independent experiments. (H) TransAM enzyme-linked immunosorbent assay was performed using nuclear extracts of THP-1 cells incubated with or without 20 ng/ml of IL-1β for 30 min and pretreated with 10 μM BT2 for 24 hours.

BT2 reduces the expression of neutrophil and monocyte/macrophage biomarkers in the AAR

Neutrophils provide the primary cellular response to M/IR injury and mediate inflammation and severity associated with M/IR injury (66, 67). BT2 reduced the expression (log₂FC < −1, FDR < 0.05) of several neutrophil markers in the AAR within 4 hours of M/IR injury such as Cd11b/Itgam, Cd18/Itgb2, Cd33, Cd44, and Cd45/Ptprc (table S8). In addition, compared with vehicle 24 hours following M/IR, BT2 reduced the expression (log₂FC < −1, FDR < 0.05) of CD11b/Itgam, Cd18/Itgb2, Cd33, Cd15/Fut4, Cd16/Fcgr3a, Cd32/Fcgr2b, Cd44, Cd45/Ptprc, and Cd62L/Selplg (table S9). At 24 hours after M/IR injury, BT2 inhibited monocyte (e.g., Cd14 and Cd11b/Itgam) and proinflammatory macrophage (e.g., Cd86, Cd83, and Cd80) biomarker expression (table S9). In line with reduced neutrophil and monocyte/macrophage biomarker expression in the AAR, BT2 inhibited levels of cell surface receptors related to leukocyte recruitment expressed by monocytes and/or neutrophils (e.g., Cxcr2 and CD44) and their cognate endothelial ligands (e.g., Cxcl2 and Sele) (tables S8 and S9).

BT2 reduced the expression of transcription factors, cytokines, metalloproteinases, inflammasome, DAMPs, DAMP-sensing receptors, and HF biomarkers in AAR

Compared with vehicle 4 hours after M/IR injury, BT2 reduced the expression (log₂FC < −1, FDR < 0.05) in the AAR of proinflammatory transcription factors (e.g., Egr1, Fosl1, and Klf5), matrix metalloproteinases (e.g., Mmp3, Mmp8, and Mmp25), proinflammatory cytokines and their receptors (e.g., Osm, Il1β, Il1r2, Tnfsf18, Tnfrsf12a, and Tnfrsf1b), DAMP biomarkers (e.g., S100a8, S100a9, and Hspa1a), DAMP receptors (e.g., Tlr1, Tlr2, Tlr13, Ripk2, Nlrp3, Trem1, Trem3, and Trpm2), CXC chemokines (e.g., Cxcl1 and Cxcl2), chemokines (e.g., Ccl2, Ccl3, Ccl6, Ccl17, Ccl22, and Ccl24), HF biomarkers (e.g., Nppb, B-type natriuretic peptide, Bnp), and growth factors (e.g., Hbegf) (table S10). Among genes with the largest fold inhibition of expression at 4 hours were Mmp3 (log₂FC −4.67), metallopeptidases Adam12 (log₂FC −3.05) and Adam8 (log₂FC −2.74), and chemokines Ccl12 (log₂FC −4.88) and Ccl22 (log₂FC −2.84). Conversely, at 4 hours, BT2 increased the expression of Sned1 (log₂FC 2.49), an anti-inflammatory factor, and Tcam1 (log₂FC 2.91), which mediates T_reg cell suppressive function (Fig. 6B and table S8).

Similarly, compared with vehicle 24 hours after M/IR injury, BT2 in the AAR reduced (log₂FC < −1, FDR < 0.05) the expression of proinflammatory transcription factors (e.g., Atf3, Fos, and Fosl1), matrix metalloproteinases (e.g., Mmp10, Mmp12, and Mmp16), proinflammatory cytokines and their receptors (e.g., Il6, Il17ra, Il18, Tnfsf18, and Tnfrsf12A), proinflammatory macrophage biomarkers (e.g., Cd83), DAMP biomarkers (e.g., Hspb1), DAMP receptors (e.g., Tlr1, Tlr2, Tlr6, Tlr7, Nlrp3, Trem1, Trem3, and Trpm2), CXC chemokines (e.g., Cxcl1 and Cxcl6), chemokines (e.g., Ccl2, Ccl17, and Ccl22), and HF biomarkers (e.g., Nppa, atrial natriuretic peptide, Anp; Nppb, Bnp). Anp and Bnp are well established noninvasive indicators of impaired LV function and HF progression and severity (68, 69). Among genes with the largest fold inhibition of expression with BT2 compared to vehicle at 24 hours were Mmp10 (log₂FC −5.79), Mmp12 (log₂FC −3.60), Il6 (log₂FC −3.70), Ccl12 (log₂FC −3.14), and Ccl7 (log₂FC −3.03). At 24 hours, BT2 increased the expression of the T_reg cell biomarker Foxp3 (log₂FC 2.20), antioxidant enzyme Sod3 (log₂FC 1.92), and the autophagy regulator Trim50 (log₂FC 2.89) among a range of other genes (Fig. 6C and tables S9 and S11). Immunohistochemical analysis confirmed BT2 inhibition of the DAMP receptor Trpm2 (31) in the AAR 24 hours after M/IR injury (Fig. 6D).

Early phase infiltrating monocytes serve as mediators of M/IR injury. For example, small interfering RNA (targeting CCR2) knockdown of monocytes reduces IF size in rodents (70, 71). BT2 strongly reduced mRNA levels of Ccl2 [whose main receptor is CCR2; (72)] at 4 hours (log₂FC −2.21) and 24 hours (log₂FC −2.85). BT2 reduced (log₂FC < −1, FDR < 0.05) mRNA levels of monocyte markers in the AAR within 24 hours of M/IR injury, such as CD62L/Sell at 4 hours, Cd14 at 24 hours, and Cx3cr1 at 24 hours. Monocytes are a key source of macrophages; BT2 reduced mRNA levels of the proinflammatory macrophage marker Cd83 at 4 and 24 hours in the AAR. BT2 also reduced (log₂FC < −1, FDR < 0.05) levels of Nlrp3, which, in monocytes, mediates and exacerbates the inflammatory response following reperfusion onset (73) at 4 and 24 hours (tables S8 and S9).

We next performed Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis to determine whether DEG following BT2 treatment were overrepresented in specific biochemical pathways (FDR < 0.05) (tables S12 to S15). Among the top gene sets with reduced expression at 4 hours were cytokine-cytokine receptor interactions, lipid and atherosclerosis, TNF signaling pathway, chemokine signaling, MAPK signaling, nuclear factor κB (NF-κB) signaling, IL-17 signaling, Toll-like receptor signaling, and leukocyte transendothelial migration (Fig. 6E and table S12). The top genes with reduced expression at 24 hours included those involved in cell cycle, chemokine signaling, cytokine-cytokine receptor interaction, lipid and atherosclerosis, and NF-κB signaling (table S14). BT2 also up-regulated a range of metabolic pathways at 24 hours (fig. S7).

Both snRNA-seq and bulk RNA-seq were performed on AAR tissue 24 hours after M/IR injury to illustrate the impact of BT2 on gene expression using two different but complementary techniques. Partial concordance is expected, given the distinct strengths of each method: Bulk RNA-seq captures average changes across all cell types from full-length mature RNA, whereas snRNA-seq resolves cell type–specific effects and heterogeneity and is generated from the 3′ end of nascent and mature transcripts. Cross-platform comparison of log₂FC < −1 or >1 at FDR < 0.05 (BT2 versus Veh) data identified 37 common DEGs (r² = 0.72) (fig. S8), including Atp8b4, Dock2, Dock8, Cd11b/Itgam, Lilrb4, Ptprc, Rbm47, Slfn4, and Spp1 (table S16). While these top DEGs are noted elsewhere in this study, biologically meaningful processes may also be reflected by genes showing more modest fold change.

A key transcription factor inhibited by BT2 in the AAR at 24 hours was the ERK-dependent transcription factor Fos, a prototypic member of the AP-1 transcription factor family. Western blotting affirmed BT2 inhibition of ERK phosphorylation. In human THP-1 monocytic cells exposed to IL-1β, BT2 inhibited ERK phosphorylation within 30 min, an effect sustained for at least 120 min (Fig. 6F). BT2 also inhibited the inducible expression of the ERK-dependent transcription factor FOS. BT2 inhibited FOS expression in a dose-dependent manner (Fig. 6G) and its interaction with its cognate recognition element (5′-TGAGTCA-3′) (Fig. 6H).

BT2 reduces IF size and preserves heart function after M/IR injury when delivered during myocardial ischemia

The preceding animal studies delivered BT2 twice before reperfusion. To determine whether BT2 also exerts cardioprotection when it is administered once during ischemia, we administered BT2 to rats intraperitoneally prior to reperfusion and assessed IF size and heart function after 24 hours. BT2 inhibited both IF size/AAR (Fig. 7A) and neutrophil infiltration in the border zone (Fig. 7B) by ~50%. Conversely, BT2 increased both EF and FS (Fig. 7C) and reduced LVIDs (Fig. 7D), as it did when administered twice (Fig. 2B).

Fig. 7. BT2 reduces IF size in the AAR and preserves heart function 24 hours after M/IR injury when delivered during myocardial ischemia.

(A) Sprague-Dawley rats were subjected to 30-min LAD ligation followed by reperfusion for 24 hours. BT2 (100 mg/kg) was delivered 5 min after the start of ischemia. IF size/AAR assessment 24 hours after M/IR in rats treated with BT2 or vehicle. Veh denotes vehicle (saline with 0.5%, v/v, Tween 80 and 0.01%, v/v, DMSO). t test, n = 5 per group. (B) Representative H&E-stained 5-μm cross sections of reperfused myocardium in the border region showing hemorrhage and infiltrating neutrophils (arrows). BT2 was delivered once during ischemia. t test, n = 3 to 4 per group. (C) EF and FS 24 hours after M/IR injury as determined by ultrasound. t test, n = 5 per group. (D) LVIDs and LVIDd 24 hours after M/IR injury as determined by ultrasound. Mann-Whitney or t test, n = 5 per group.

BT2 inhibits multiple genes associated with disease severity in patients with coronary artery disease

Interrogation of the STARNET database (74) comparing gene expression in aortic tissue from patients with coronary artery disease (CAD) with those without CAD revealed that genes suppressed by BT2 in rodents are associated with disease severity. This includes EGR1 (FDR = 3.9 × 10⁻¹⁰), FOS (FDR = 3.9 × 10⁻⁵), FOSL1 (FDR = 3.2 × 10⁻¹⁹), JUN (FDR = 3.8 × 10⁻⁴), and ATF3 (FDR = 9.5 × 10⁻¹⁸) (Fig. 8A and tables S8 and S9). In STARNET, EGR1 and ATF3 are highly significant for DUKE CAD severity score (which predicts all-cause mortality) and JUN correlates with SYNTAX (Synergy between Percutaneous Coronary Intervention with Taxus and Cardiac Surgery) risk score (predictive of MACE) (Fig. 8A).

Fig. 8. BT2 inhibits multiple genes associated with CAD severity.

(A) DEGs in aortae of patients with CAD (n = 600) compared with those without CAD (n = 250) (STARNET database). (B) DEGs in monocytes of patients with AMI (n = 12) compared with those with stable disease (n = 11) (GSE129935). (C) DEGs in LV apex from patients with AMI (n = 6) compared with those with nonischemic disease (n = 8) (GSE974). (D) DEGs in circulating ECs from patients with AMI (n = 49) were compared with those from healthy individuals (n = 50) (GSE66360). (E) DEGs in peripheral blood collected from patients with ACS comparing STEMI (n = 7), NSTEMI (n = 10), and UA (n = 9) patients with normal individuals (n = 7) (GSE61145, platform GPL6884).

We next analyzed gene expression in the Holvoet et al. database (75) constructed from monocytes sourced from patients with AMI compared with stable CAD (GSE129935). This comparison showed that mRNA levels of ATF3, FOSL1, FOSL2, KLF5, IL1β, IL6R, TNFRSF21, HBEGF, OSM, CXCL1, CXCL2, TLR2, RIPK2, NLRP3, TREM1, TSC22D1, and NPPA are included in a significantly increased DEG in the setting of AMI (Fig. 8B). Experimentally, BT2 reduced levels of all these DEGs (tables S8 and S9). In contrast, mRNA levels of SNED1 (nidogen) decreased in the setting of AMI (Fig. 8B), and increased in rats treated with BT2 (tables S8 and S9). Nidogen increases heart contractile function and reduces fibrosis in a mouse M/IR model (76).

We also analyzed gene expression in the Hall et al. database (77) constructed from LV apex samples harvested from patients with AMI at the time of implant of an LV assist device compared with apex material sourced from nonischemic patients (GSE974). FOSL2, IL1β, TLR2, and TREM1 mRNA levels were elevated in AMI tissue, whereas FOXP3 levels decreased (Fig. 8C). BT2 reduced Fosl2, Il1β, Tlr2, and Trem1 mRNA levels and increased Foxp3 levels in rat M/IR injury (tables S8 and S9).

We then examined gene expression in the Muse et al. database (78) constructed from circulating ECs sourced from patients with AMI and healthy individuals (GSE66360). IL1β, IL6, OSM, CXCL1, CXCL2, TLR2, NLRP3, TREM, FOSL1, FOSL2, ATF3, EGR1, and HB-EGF mRNA levels were elevated in AMI tissue (Fig. 8D). BT2 reduced mRNA levels of I11β, Il6, Osm, Cxcl1, Cxcl2, Tlr2, Nlrp3, Trem, Fosl1, Fosl2, Atf3, Egr1, and Hbegf in the heart (tables S8 and S9).

We analyzed gene expression in the Park et al. database (79) constructed from peripheral blood collected from patients with ACS comparing patients with STEMI, NSTEMI, and UA with normal individuals (GSE61145, platform GPL6884). OSM, IL6R, IL1R2, CXCL1, TLR2, TREM1, and FOSL2 mRNA levels were elevated in STEMI serum. OSM and TLR2 mRNA levels were also elevated in NSTEMI serum (Fig. 8E). BT2 reduced mRNA levels of Osm, Il6r, Il1r2, Cxcl1, Tlr2, Trem1, and Fosl2 (table S8). These findings indicate that BT2 inhibits multiple genes in rats associated with disease severity in patients with ACS.

BT2 inhibits multiple genes predictive of HF in STEMI patients undergoing PCI

Since our RNA-seq data revealed that BT2 reduced Nppa and Nppb levels in the heart (table S11), we explored whether other genes that were suppressed by BT2 may link to the progression of HF in patients with STEMI. We analyzed gene expression in the Maciejak et al. database (80) constructed from peripheral blood mononuclear cells (PBMCs) isolated from individuals with stable CAD or two groups of patients with STEMI that underwent PCI, one group that developed HF within 6 months and the other that did not develop HF (GSE59867). mRNAs predictive of elevated HF progression were NLRP3, TRPM2, FOSL2, IL1β, CXCL1, CXCR2, CCL24, TLR6, IL1R2, HBEGF, FOS, and FOSL1 (Fig. 9, A and B). BT2 reduced mRNA levels of Nlrp3, Trpm2, Il1β, Cxcl1, Cxcl2, Ccl24, Tlr6, Il1r2, Hbegf, Fos, and Fosl2 (table S8).

Fig. 9. BT2 inhibits multiple genes predictive of HF in patients with STEMI.

DEGs in PBMCs isolated from patients with stable CAD (n = 46), or patients with STEMI who developed HF [blood was collected within 1 day (n = 9), 4 to 6 days (n = 9), 1 month (n = 8), and 6 months (n = 8) of AMI] or did not develop HF [blood collected within 1 day (n = 8), 4 to 6 days (n = 6), 1 month (n = 8), and 6 months (n = 8) of AMI] within 6 months (GSE59867). Levels of (A) NLRP3, TRPM2, FOSL2, and IL1β, and (B) CXCL1, CXCR2, CCL24, TLR6, IL1R2, HBEGF, FOS, and FOSL1 in those who developed HF or did not. Data in (A) and (B) were analyzed using Kruskal-Wallis or one-way ANOVA, as appropriate. (C) Overlap of up-regulated DEGs in patients with STEMI who later developed HF (PBMCs collected within 1 day of AMI, GSE59867) with DEGs down-regulated by BT2 in the AAR in rats 4 or 24 hours after M/IR injury. The upper histogram denotes the number of DEGs up-regulated in patients with STEMI who developed (green, HF up, n = 9) or did not (blue, non-HF up, n = 8) develop HF within 6 months relative to patients with stable CAD (n = 46). The lower histogram denotes the number of DEGs down-regulated in patients with STEMI who developed (orange, HF down, n = 9) or did not (pink, non-HF down, n = 8) develop HF within 6 months. The percentages between the circles in the Venn diagrams indicate the proportion of “HF up” or “HF down” DEGs that overlapped with down- or up-regulated DEGs in the 4 and 24 hours BT2 groups. Expression data were fitted to a negative binomial generalized linear model before undergoing a genewise quasi-likelihood F test to identify DEGs.

To place these data into a broader context, we compared all DEGs in the HF group within 1 day of AMI (all underwent direct PCI) (versus stable CAD, FDR < 0.05) with DEGs in the BT2 treated group (versus vehicle, FDR < 0.05). There were 4175 up-regulated DEGs and 3221 down-regulated DEGs in the HF group at 24 hours. In contrast, there were just 77 up-regulated and 77 down-regulated DEGs in the non-HF group (versus stable CAD) (Fig. 9C). Of the up-regulated DEGs in the HF group, 10.1% (or 423) were the same DEGs that BT2 down-regulated in the rat AAR at 4 hours (Fig. 9C and table S17). Conversely, 6.3% (or 202) of down-regulated DEGs in the HF group were up-regulated by BT2 in the AAR at 4 hours (Fig. 9C and table S18). The overlap of up-regulated and down-regulated DEGs in patients with HF with DEGs down-regulated and up-regulated by BT2, respectively, effectively doubled when DEGs in the HF group were compared with DEGs in the BT2 group at 24 hours. BT2 down-regulated 18.5% (1 in 6) of DEGs up-regulated in the HF group (Fig. 9C and table S19). Conversely, BT2 up-regulated 12.5% (1 in 8) of DEGs down-regulated in the HF group (Fig. 9C and table S20). These observations demonstrate the ability of BT2 to regulate the expression of genes predictive of HF. Together, these data highlight the cardioprotective effects of BT2 in rats with M/IR injury and its potential to reprogram gene expression signatures that epitomize ischemic disease severity and HF in patients with ACS undergoing PCI.

DISCUSSION

Myocardial injury secondary to PCI can follow restoration of coronary blood flow following ischemia. Several large clinical trials have targeted inflammation secondary to M/IR injury with limited promising results arising from studies intervening early in the IL-1β–IL-6 pathway (48). A range of pharmacologic agents have been tested clinically for the limitation of IF size, including mitochondrial permeability transition pore and complement inhibitors, NO donors, antioxidants, cardiolipin inhibitors, Na⁺/H⁺ and Na⁺/Ca²⁺ exchanger inhibitors, PKC inhibitors, p38 MAPK inhibitors, atrial natriuretic peptide, GLP-1 agonists, thrombin inhibitors, complement inhibitors, statins, β-blockers, and ACE inhibitors (81–83). However, most of these interventions have not demonstrated significant benefit in reducing IF size or improving clinical outcomes post-ACS. For example, meta-analysis of multiple randomized trials comparing early intravenously administered β-blockers with placebo or routine care in patients with STEMI undergoing PCI showed no difference in 1-year death or biomarker-based MI size (84). Similarly, the AMISTAD-II trial demonstrated reduced MI size with adenosine infusion, but failed to improve long-term clinical outcomes (85). These findings highlight the urgent need for new therapies that can effectively mitigate M/IR injury and improve patient prognosis.

The present study tested the preemptive postreperfusion cardioprotective effects of the dibenzoxazepinone, BT2 (50). In rats in which the LAD was ligated for 30 min followed by 24 hours or 2 weeks of reperfusion, BT2 proved effective as a small-molecule inhibitor of IF size, cardiac dysfunction, and fibrosis. In contrast, BT3, a structural analog, did not affect IF size, indicating that BT2’s effect on IF size was not due to mass effect. BT2 also reduced serum levels of the cardiac biomarker troponin I, indicating that BT2 attenuates myocyte injury following M/IR injury. BT2 preserved heart function (no decrease in EF or FS) and prevented adverse remodeling (no increase in LVIDd or LVIDs), 24 hours and 2 weeks after M/IR injury. EF, FS, LVIDd, and LVIDs in BT2-treated animals at 2 weeks were equivalent to that of the sham group (rats that underwent the surgical procedure without M/IR injury). LVIDd represents the maximum size of the LV ventricle, reflecting the extent of dilation, whereas LVIDs represents the remaining size after contraction. Of these, LVIDd is considered a better indicator of ventricular remodeling and enlargement (86) and independently predicts all-cause mortality or hospitalization for cardiac outcomes including HF (87–90). LVIDd did not change 24 hours post–M/IR injury (Fig. 2B). At 2 weeks post–M/IR injury, however, when LVIDd increased significantly, BT2 prevented this increase. LVIDd at 2 weeks in M/IR rats treated with BT2 did not differ from that of sham operated rats (Fig. 2B). Our findings, therefore, indicate that BT2 prevents cardiac remodeling (as measured by LVIDd at 2 weeks) secondary to M/IR injury and IF formation (Figs. 1B and 7A, at 24 hours). In both RNA-seq and snRNA-seq, BT2 had no impact on established biomarkers of cardioprotection (Akt1, Akt2, Akt3, Pkc, Ip3k, and Stat3) (log₂FC < −1 or >1, FDR < 0.05).

BT2 prevented myocardial fibrosis 2 weeks after M/IR injury. BT2’s inhibition of cardiac scarring is supported by its suppression of expression of genes linked with LV remodeling, fibrosis, and oxidative stress, namely, growth factors (91, 92) and lysyl oxidase (93) including (but not limited to) the fibrotic mediators Tgfb2 and Lox (tables S9 and S11). Myofibroblasts are key mediators of LV dysfunction, cardiac remodeling, and scar formation (94). BT2 suppressed pathways associated with cellular development, signal transduction, cell motility, and migration in these cells. Moreover, it down-regulated 10 gene sets as compared to one up-regulated gene set in these cells (at FDR < 0.25). As well as reducing IF size (Figs. 1B and 7A) and cardiac fibrosis (Fig. 5E), snRNA-seq analysis shows that, in myofibroblasts, BT2 inhibits the expression of multiple extracellular matrix-related genes involved in collagen production (Fig. 5C, right), extracellular matrix organization, and matrix structure (table S4). As indicated above, GSEA performed with snRNA-seq data revealed that BT2 inhibited the MAPK cascade preferentially in myofibroblasts (FDR < 0.25) (fig. S5). While BT2 treatment resulted in a greater number of FB1, BT2 did not increase proliferation marker expression in these cells, indicating that this was unlikely the result of enhanced cell replication. This finding may result from greater FB1 cell survival. Cardiac fibroblasts can resist stress-induced death (63). Despite the higher proportion of FB1 in BT2-treated AAR, the number of myofibroblasts did not differ between BT2 and vehicle groups. This finding suggests that BT2 maintains fibroblasts in the quiescent FB1 state and restrains transition to myofibroblasts.

This study used both bulk RNA-seq and snRNA-seq analysis. Direct comparison of DEGs from bulk RNA-seq versus collapsed snRNA-seq is challenging, as these differ in RNA type (processed cytoplasmic versus pre-mRNA), number of genes and what these numbers represent (cell-to-cell heterogeneity versus average gene expression), and snRNA-seq having inherent 3′ bias. Comparing fold change in specific genes between bulk RNA-seq and snRNA-seq datasets will not necessarily align since the two outputs represent different stages of RNA processing: Nuclear RNA includes nascent mRNA (i.e., transcripts that have not yet undergone splicing, or polyadenylation), whereas total mRNA includes both pre-mRNA and mature mRNA, the latter being the substrate for protein synthesis. In addition, RNA-seq captures sequences across the entire length of an RNA transcript, whereas snRNA-seq captures a portion of the transcripts near the A-rich region, typically the last 350 nucleotides of the 3′ untranslated region. In addition, bulk RNA-seq depth captures more transcripts than snRNA-seq, whereas snRNA-seq captures information identifying cell types at the expense of gene depth. For example, our bulk RNA-seq dataset spanned 14,927 genes, whereas the snRNA-seq dataset scoped 4442 genes.

Two preemptive strategies were used in this study in light of the anti-inflammatory properties of BT2, the lack of any such drug in the clinic and growing literature, which strengthens the view that CAD is a treatable inflammatory disease (95). BT2 reduced IF size in the AAR after M/IR injury, regardless of whether BT2 was delivered 24 hours before and/or during ischemia prior to reperfusion. Bulk RNA-seq and snRNA-seq were performed with AAR specimens in which BT2 was administered twice before reperfusion, in line with the concept that preemptive administration of a cardioprotective agent, such as BT2, might counter the proinflammatory insult of PCI with reperfusion. This sequence would be analogous to a patient having PCI being given aspirin or heparin prior to revascularization (96) and would apply particularly to patients undergoing planned staged or elective PCI (97, 98).

While infiltrating neutrophils provide the primary innate cellular response following M/IR injury, these cells also mediate myocardial damage and affect adverse LV remodeling and HF development in patients with AMI (99). For example, increased neutrophil infiltration within the IF zone is associated with larger IF size and impaired LV function (100). BT2 strongly reduced the expression of neutrophil markers in the AAR within a few hours of M/IR injury. It also reduced proinflammatory macrophage phenotype biomarker expression. This observation suggests that BT2, which inhibits endothelial vascular cell adhesion molecule–1 and intercellular adhesion molecule–1 expression induced by IL-1β and blocks monocytic cell–endothelial adhesion and transendothelial migration (50), may help to ameliorate this, the most damaging (“The Bad”) phase of inflammation post-AMI (48). BT2 reduced the proportion of neutrophils and other inflammatory cell types including monocytes, macrophages, and T cells.

CANTOS (23) revealed that inhibition of IL-1β reduces the incidence of recurrent MACE (75, 77). In the present study, bulk RNA-seq analysis revealed that in cardiac tissue, BT2, a MEK/ERK inhibitor (49) suppressed the expression of IL-1β, IL-6, and multiple other genes associated with disease severity in patients with STEMI. This includes transcription factors [e.g., Egr1 and Klf5, which are ERK-dependent; (101–106)], cytokines and their receptors (e.g., Il1β, Il6, and Il1r2), metalloproteinases (e.g., Mmp3), the inflammasome (e.g., Nlrp3), DAMPs (e.g., S100a8), and DAMP-sensing receptors (e.g., Tlr2). Circulating levels of IL-1β and IL-6 increased significantly in patients with STEMI, NSTEMI, and UA compared with individuals with normal coronary arteries (107). Many drugs used in clinical trials for CAD have had single targets [namely, canakinumab, anakinra (which inhibits both IL-1 isoforms), tocilizumab, and pexelizumab]. BT2 suppressed levels of Trpm2 protein, a transient receptor potential cation channel that senses reactive oxygen species and increases intracellular Ca²⁺ and production of the potent neutrophil chemoattractant CXCL2 (108). Trpm2, which is inhibited by BT2, exacerbates M/IR injury (109).

BT2 modulated the expression of thousands of genes in the heart 4 and 24 hours postreperfusion. These genes represent signaling molecules, transcription factors, proinflammatory and fibrotic mediators, and extracellular matrix genes, among many others. Accordingly, the breadth of genes affected by BT2 secondary to p-ERK inhibition suggests that changes in the expression of some genes are secondary to others. We note that prior studies indicate that IF size as a proportion of AAR or LV is unchanged by certain MEK/ERK inhibitors such as PD98059 or U0126 in rats and sheep (110–114) and that ERK may even be cardioprotective during ischemic conditioning (18) or mild hypothermia (115). The scope of genes affected by BT2 suggests that this compound is a more potent anti-inflammatory agent than other MEK/ERK inhibitors, as we previously observed with PD98059 (49).

Given the complexity and “deleterious flare of excessive inflammation in the early phase after AMI” (48), single-targeted therapy during revascularization may be akin to “deploying an umbrella in a blizzard.” This may explain why, of all large clinical trials for AMI targeting inflammation thus far (48), anti–IL-1R is the only intervention that reduces HF at 1 year follow-up in a small trial (another clinical trial showed that anti-C5 improved EF after a 90-day follow-up). BT2 appears to have multiple targets, which is why it may succeed clinically in the face of decades of experimental M/IR studies with single targets that have failed to translate to clinical practice.

Inhibition of cardiac fibrosis secondary to M/IR injury is a double-edged sword. While excessive fibrosis contributes to LV systolic-diastolic dysfunction and HF, insufficient fibrotic scar repair can lead to ventricular aneurysm and cardiac rupture. For example, mice deficient in Mmp28 have decreased collagen deposition, fewer myofibroblasts, and increased cardiac rupture post-AMI. These mice also have increased LV volumes, LV dysfunction, and lung edema (116). Medications, such as glucocorticoids and nonsteroidal anti-inflammatory drugs (NSAIDs), can predispose to adverse remodeling and rupture. Therefore, balancing pro- and antifibrotic actions is crucial for healthy healing postinfarction. BT2’s inhibition of Mmp expression may avoid excessive matrix degradation, while its capacity to suppress key fibrotic growth factors such as Tgf and Ctgf may prevent excessive scar formation. Changes in transcription effected by BT2 were accompanied by prevention of loss in adverse remodeling and heart function. Short-(24 hours) and long-term (2 weeks) imaging in this study suggests that BT2 promotes healthy healing rather than adverse remodeling.

Beyond reducing well-established clinical biomarkers of HF (Nppa/BNP and Nppa/ANP), BT2 reduced multiple genes predictive of HF. Our interrogation of the Maciejak et al. database (GSE59867) (80) revealed that NLRP3, TRPM2, FOSL2, IL1β, CXCL1, CXCR2, CCL24, TLR6, IL1R2, HBEGF, FOS, and FOSL1 were elevated in patients with STEMI who later developed HF as compared with patients with STEMI who did not. This observation may provide the first link between early-phase elevation of DAMP receptors, cytokines and their receptors, transcription factors, chemokines, and growth factors and the prediction of HF. BT2 inhibits all of these genes at 4 and/or 24 hours, which makes this a core finding in this paper. Twenty-four hours after M/IR injury, BT2 down-regulated 1 in 6 (774 out of 4175) of the DEGs up-regulated in patients with STEMI 1 day after AMI, who later developed HF. This finding suggests that early inflammation induced by reperfusion injury represents a critical therapeutic window for improving the cardiac prognosis of patients with AMI. The VCU-ART (1 year follow-up) (117) and APEX-AMI substudy (48) demonstrate that anti-inflammatory therapies targeting the early phase of AMI resulted in a decreased incidence of HF or improved EF.

This study had several limitations. First, since this study used male rats, we cannot determine whether the expression patterns reported here apply to females. Second, while the present study provides proof-of-principle evidence of BT2’s capacity to inhibit IF size, prevent adverse remodeling, and preserve heart function, future studies using doses lower than 100 mg/kg could inform the ideal therapeutic window for clinical testing. In addition, as BT2 was given as monotherapy in this study, subsequent studies may determine whether BT2’s beneficial effects are affected in animals given standard medical therapy and whether combinatorial benefit results. Third, while it is feasible in this model (using healthy rats) that BT2’s reduction of IF was secondary to its attenuation of the inflammatory response triggered by M/IR injury, the situation is less clear in patients undergoing PCI, where there is a substantial baseline proinflammatory and plaque burden that precedes the added inflammation triggered by reperfusion injury (48). Moreover, the M/IR model may not separate BT2’s beneficial effects on cardiac remodeling from its inhibition of M/IR injury since the degree of adverse cardiac remodeling is influenced by the size of the IF (11). Fourth, while it is unclear how many of the expression patterns obtained delivering BT2 twice would be similar if BT2 was given once during ischemia prior to reperfusion, it may be clinically feasible for a drug like BT2 to be given to patients more than once within a 24-hour window before a planned staged or elective PCI (97, 98), especially since elective PCI in the United States and Japan accounts for ~34 and 73% of cases, respectively (118). This work advances the concept of preemptive cardioprotection, whereby preemptive or prophylactic administration of a drug like BT2, before planned localized tissue trauma, may be applied to high-risk patients with UA/NSTEMI or those having a scheduled elective PCI. Nonetheless, the schedule used here served to probe the mechanism of action of BT2 in the context of M/IR injury. As early as 4 hours post–M/IR, BT2 reduced in the AAR (log₂FC < −1, FDR < 0.05) the expression of proinflammatory transcription factors (e.g., Atf3, Fos, and Fosl1), matrix metalloproteinases (e.g., Mmp10, Mmp12, and Mmp16), proinflammatory cytokines and their receptors (e.g., Il6, Il17ra, Il18, Tnfsf18, and Tnfrsf12A), proinflammatory macrophage biomarkers (e.g., Cd83), DAMP biomarkers (e.g., Hspb1), DAMP receptors (e.g., Tlr1, Tlr2, Tlr6, Tlr7, Nlrp3, Trem1, Trem3, and Trpm2), CXC chemokines (e.g., Cxcl1 and Cxcl6), chemokines (e.g., Ccl2, Ccl17, and Ccl22), and HF biomarkers (e.g., Nppa, Anp; Nppb, Bnp) (table S11). Last, it is unclear whether the anti-inflammatory properties of BT2 may have unintended effects within the clinical environment. For example, certain NSAIDs, such as the COX-2 inhibitor rofecoxib, are associated with adverse cardiovascular outcomes (119), whereas celecoxib, another COX-2 inhibitor, has not been associated with CV harm at moderate doses (120, 121). Moreover, canakinumab (CANTOS) was associated with a slight but significant increase in fatal infection or sepsis (23), while colchicine [COLCOT (26) and LoDoCo2 (122)] was associated with increased mortality from adverse noncardiovascular events (123, 124). Meta-analysis also suggested that colchicine post-MI reduces MACE but increases adverse gastrointestinal events (125), although in the recent CLEAR SYNERGY (OASIS 9) trial, colchicine’s reported effects on repeat MI, urgent revascularization, and stroke were identical to placebo (27). If the identified reductions in IF size, cardiac remodeling, and fibrosis seen in rodents translate to the clinical setting, then the therapeutic benefit of this ERAS-type protocol would be expected to outweigh the potential risk of medically managing a CV event. While the present study provides important proof-of-principle evidence that BT2 can serve as a small-molecule inhibitor of M/IR injury and cardiac fibrosis, future studies in large animal models and a preclinical safety campaign should further demonstrate BT2’s potential to serve as a preemptive, postreperfusion cardioprotective drug for patients with ACS.

MATERIALS AND METHODS

Cell culture

Human THP-1 monocytic cells were obtained from the American Type Culture Collection and maintained in RPMI-1640 (pH 7.4), supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin, and 1% l-glutamine in a humidified atmosphere of 5% CO₂ at 37°C.

Western blotting

THP-1 cells were seeded into six-well plates (0.6 × 10⁶ cells per well) and treated with various concentrations of BT2 or vehicle [0.01% dimethyl sulfoxide (DMSO) in medium] for 24 hours, and 20 ng/ml of human recombinant IL-1β and total cell lysates were prepared in radioimmunoprecipitation assay buffer. Lysates (10 μg) were resolved by SDS–polyacrylamide gel electrophoresis then transferred to Immobilon-P polyvinylidene difluoride membranes (Millipore, USA). Membranes were blocked with 5% skim milk and incubated with rabbit monoclonal anti-phospho-ERK (Cell Signaling, catalog no. 4370), rabbit monoclonal anti-ERK (Cell Signaling, catalog no. 4695), FOS (Cell Signaling, catalog no. 2250), or mouse monoclonal β-actin (Sigma-Aldrich, catalog no. A5316) antibodies followed by horseradish peroxidase–conjugated secondary goat anti-rabbit antibodies (DAKO, catalog no. P0448) or goat anti-mouse antibodies (DAKO, catalog no. P0447). Chemiluminescence was detected using the Western Lightning Chemiluminescence system (Thermo Fisher Scientific, USA), an ImageQuant LAS 4000 biomolecular imager (GE Healthcare Life Sciences, USA), and ChemiDoc MP Imaging Systems with Image Lab Touch Software (Bio-Rad Laboratories Inc.). All antibodies were validated for immunoblot use by the respective manufacturers.

TransAM assay

Nuclear extracts of THP-1 cells were prepared for the TransAM c-FOS Kit (catalog no. 44096) using a Nuclear Extract Kit (catalog no. 40010) and 6 μg of protein (determined by Bradford Assay) according to the manufacturer’s instructions (Active Motif, Carlsbad, CA, USA). The THP-1 cells were incubated with or without IL-1β (20 ng/ml) for 30 min and pretreated with 10 μM BT2 or BT3 for 24 hours.

Experimental M/IR injury

Adult male Sprague-Dawley rats (~300 to 350 g, 8 to 10 weeks old) were obtained from the Australian Resources Centre and housed for 7 days before experiments to allow acclimatization and consumed a standard chow diet. Males were used because estrogens may protect against I/R injury in rats (126). Rats were anesthetized with isoflurane gas. Twenty-four hours before inducing myocardial ischemia, and/or 5 min after coronary artery ligation, animals received BT2 (49) (100 mg/kg), BT3 (49) (100 mg/kg), or vehicle (saline with 0.5%, v/v, Tween 80 and 0.01%, v/v, DMSO) (all 400 μl) via intraperitoneal injection. Six-lead electrocardiogram (ECG) (eKuore Veterinary) was performed preoperation and during ligation. M/IR injury was induced in rats essentially as previously described (67) with minor modification. Briefly, under mechanical ventilation, a left thoracotomy was performed, and the LAD artery was ligated 3 mm from its point of origin (distal to the junction of the pulmonary artery and left atrial appendage) with 6-0 prolene. Ischemia was confirmed by myocardial blanching and ST segment elevation on ECG. After 30 min, the ligature was untied to allow reperfusion of the ischemic myocardium, and the chest was closed. Using a rodent ultrasound imaging system (Vevo3100, VisualSonics), EF, FS, LVIDd, or LVIDs were measured in anesthetized rats at 24 hours or 2 weeks after LAD ligation.

At 24 hours after LAD ligation, in anesthetized rats, the femoral artery was exposed. A 0.96-mm-diameter catheter was inserted into the femoral artery, and 2 ml of blood was collected into SST tubes for later use. At 4 or 24 hours after LAD ligation, the chest cavity was reopened, and the LAD was religated at the same location to define the AAR. In tetraphenyl tetrazolium chloride (TTC) staining, AAR is defined as the region of myocardium that does not take up Evans blue dye; within that AAR, TTC determines infarcted tissue from viable tissue. For rats with endpoint at 4 or 24 hours after ligation, 2 ml of 2% Evans blue dye was injected directly into the right ventricle. Subsequently, 1 ml of 3 M KCl [which arrests the heart in diastole; (127)] was administered by intracardiac injection to stop the heart beating prior to heart removal. Hearts were collected at 4 and 24 hours after ligation, frozen at −30°C for 1 hour, and cut transversely through the LV into five slices (each ~2 mm) from the apex to the base. Slices of AAR from the 24-hour time point were incubated in 4% TTC for 20 min at 37°C to determine IF size in the AAR (IF/AAR) and AAR/LV area, which was measured using Image-Pro Plus (Cybernetics, Bethesda, MD, USA). Slices from 4 and 24 hours were stored at −80°C for further isolation of total RNA. At 2 weeks after LAD ligation, the chest cavity of rodents in this group was reopened, and 1 ml of 3 M KCl was administered by intracardiac injection prior to heart removal. Hearts were excised and fixed in 10% neutral buffered formalin for histology. Experiments were conducted with University of New South Wales (UNSW) Animal Care and Ethics Committee approval (21-15A, 22-118A, 23-116A).

Serum troponin I detection

Cardiac troponin I levels were determined using a Siemens Atellica High-Sensitivity Troponin I Assay on an Atellica Solution IM Analyzer (Siemens Healthineers, Tarrytown, NY, USA), a chemiluminescence-based system that has been used to measure troponin I levels in blood from multiple species including pigs (128) and dogs (129). Human cardiac troponin I antibodies (Siemens) cross-react with rat troponin I (130, 131). The capture antibody and detection antibody epitopes in the High-Sensitivity Troponin I Assay are >97% conserved (35 of 36 residues) between human and rat cardiac troponin I. Human and rat cardiac troponin I share 93% sequence homology (fig. S9).

Analysis of heart rate, PR and QTc interval, and ST-T segment elevation

ECG was performed using a veterinary ECG monitor with a six-lead configuration (eKuore). Parameters were set to 20 mm/mV amplitude and 25 mm/s paper speed. For each rat, three measurements of heart rate and PR and QT interval were taken, and the mean of QT interval is corrected using Bazett’s formula (QTc = QT/√RR). Elevation in ST-T segment from baseline was measured from lead III.

Ultrasound imaging

Anesthetized rats were positioned in supine or slightly left lateral decubitus position to gain optimal access to the heart. Prior to imaging, ultrasound gel was applied to the rat’s epilated chest area to facilitate a good acoustic interface for optimal imaging quality. MX250s transducer was placed on the rat’s chest, taking care to adjust its position and angle for a clear view of the LV short axis, notably showing the two LV papillary muscles. Ultrasound imaging was performed using M-mode imaging and Vevo3100 system (VisualSonics). The M-mode cursor was aligned to the LV, to obtain maximum diameter in the LV short axis. M-mode recording was used to capture a continuous line of motion across the heart once the transducer and imaging parameters were properly set. The system was used to obtain EF, FS, LVIDd, and LVIDs measurements.

BT2 bioavailability

Serum samples were diluted with acetone and the solution left for 14 hours at 4°C. Precipitated protein was pelleted by centrifugation, and the supernatant was diluted 1:4 for MS analysis and compared against a standard curve generated with increasing amounts of BT2. Samples were run on a Thermo Gold C₁₈ column (50 × 2.1 mm) with solvent A (H₂O:0.1% formic acid) and solvent B (H₂O:CH₃CN 20:80, 0.1% formic acid), with gradient (T = 0, 1% B, T = 26 min, 100% B, T = 27 min, 100% B, T = 27.1, 1% B, T = 30 min, 1% B at 0.2 ml/min) and column temperature of 45°C. MS was performed using a QExactive HF operating in data-dependent mode with MS1 scan, mass/charge ratio (m/z) of 140 to 800, 3 × 10⁶ ions, maximum injection time of 25 ms, resolution of 120,000, and top five MS2.

Immunohistochemical staining and analysis

Twenty-four hours after LAD ligation, hearts were excised and fixed in 10% neutral buffered formalin. The hearts were sectioned into four 3-mm slices from apex to base, processed, and embedded in paraffin. Sections from slice 2 were deparaffinized, rehydrated, and subjected to heat-induced epitope retrieval using citrate buffer (pH 6) for 5 min at 110°C. Sections were blocked with Dual Endogenous Enzyme Block (DAKO, S2003) for 10 min, followed by 2% skim milk for 20 min. Slides were incubated with the primary antibody (rabbit polyclonal anti-Trpm2, catalog no. PA5-119946, Thermo Fisher Scientific) at a 1:400 dilution for 60 min at room temperature. After rinsing with buffer, slides were incubated with the secondary antibody (goat anti-rabbit, catalog no. P0448, DAKO) for 30 min. Subsequently, diaminobenzidine chromogen (catalog no. K3468, DAKO) was applied for 5 min, and the slides were counterstained with hematoxylin and Scott’s blue. After counterstaining, slides were dehydrated in 100% ethanol and xylene and then coverslipped. Immunostained slides were scanned using an Olympus VS200 slide scanner (Olympus, Tokyo, Japan), and images were captured using QuPath software (132). The percentage area of positive staining in 20× objective fields was quantified using Image-Pro Plus software (Cybernetics, Bethesda, MD, USA). Quantification was performed with 10 fields of view per section, with three to four rats per group.

Multiplex immunofluorescence staining and image acquisition

Multiplex immunofluorescence staining and image acquisition were performed by the Katharina Gaus Light Microscopy Facility, Mark Wainwright Analytical Centre, UNSW. Primary antibodies used were rabbit monoclonal anti-phospho-ERK (CST, catalog no. 4370s), rabbit monoclonal anti-CD68 (CST, catalog no. 97778s), rabbit monoclonal anti-CD4 (CST, catalog no. 25229S), rabbit monoclonal anti-Foxp3 (Abcam, catalog no. ab215206), and rabbit monoclonal anti-dystrophin (Abcam, catalog no. ab218198). Multiplex images were captured using HALO software (Indica Labs) and analyzed with Image-Pro Plus (Cybernetics, USA). The necrotic zone was defined as the area with negative dystrophin staining or regions showing incomplete myocardial cells based on dystrophin staining. Regions of the border zone were defined as a 500-μm margin from the edge of the necrotic zone. Quantification of positive cells was performed using four to five random 20× objective fields of view in border zone per tissue section, with three to four rats per group.

Sirius Red and methyl green staining

Cross sections of Sirius Red and methyl green stained slides were scanned using an Aperio ScanScope XT slide scanner (Leica Biosystems, Mt. Waverley, VIC, Australia), and images were captured using ImageScope software (Leica Biosystems). Red staining area and LV area were determined using Image-Pro Plus (Cybernetics, Bethesda, MD, USA).

Hematoxylin and eosin staining and neutrophil counts

Sections were deparaffinized, rehydrated, and then immersed in Harris hematoxylin solution for staining. Following differentiation and bluing steps, sections were counterstained with eosin-phloxine solution. Sections then underwent dehydration and were mounted using a mounting medium. Hematoxylin and eosin (H&E)–stained sections were scanned using an Olympus VS200 scanner (Olympus Life Science), and digital images were captured and analyzed using QuPath software (https://qupath.github.io/). Neutrophils, identified as polymorphonuclear leukocytes, were counted in the border zone displaying hemorrhage. Counting was performed under a 20× objective, using a standardized counting grid field within the software. Ten randomly selected fields per rat (n = 3 to 4 rats) were analyzed to obtain an average neutrophil count per 20× objective field.

snRNA-seq and bioinformatics analysis

Twenty-four hours after LAD ligation, the chest was reopened, and the LAD religated at the same location to define the AAR. A 2% Evans blue dye (2 ml) was injected into the right ventricle, followed by 1 ml of 3 M KCl to arrest the heart in diastole. Hearts were harvested, frozen at −30°C for 1 hour, and sectioned into four 3-mm slices from apex to base on dry ice. The AAR tissue from the LV was collected and stored at −80°C for subsequent analysis. snRNA-seq data were sourced from three biologically independent samples per group pooled for each of the two conditions (Veh versus BT2 at 24 hours postperfusion).

snRNA-seq data were generated from frozen AAR using the chromium single cell gene expression solution (10x Genomics) by the Garvan Genomics Platform. Briefly, frozen AAR (third slice from apex) was thawed, cut into small pieces, and homogenized in a Dounce homogenizer with 5 ml of hypotonic lysis buffer. The lysate was combined with a rinse of 1 ml of buffer, gently pipetted, and centrifuged at 500g for 5 min at 4°C. The pellet was resuspended in 5 ml of washing buffer, filtered through a 40-μm cell strainer, and centrifuged. The nuclear pellet was resuspended in 0.5 ml of buffer, and nuclei integrity was assessed using Trypan blue staining under a bright-field microscope. Single nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) and sorted on a Becton Dickinson FACS AriaI II, using a 70-μm nozzle, and based on the gating strategy in representative fig. S10. Nuclei were sorted to a 1.5-ml Eppendorf tube precoated and containing 100 μl of collection buffer. Sorted nuclei were passed over to the Cellular and Spatial Hub for quality control, capture, and sequencing. Nuclei were reconcentrated to achieve an ~1000 nuclei/μl concentration with 5-μl aliquot of nuclei enumerated using AOPI (Logos Biosystems) staining, with counting on the Luna-FX7 automated cell counter (ATA Scientific). Nuclei were then pooled at an equal ratio from three samples per group. Of this pool, ~28,500 nuclei were aliquoted for capture. This allows a capture for a targeted 20,000 nuclei. Capture for single nuclei was processed through the 10x Genomics 3′ GEM-X assay, using the 10x Genomics Chromium-X instrument. Briefly, the processes include (i) cell coencapsulation with barcode-bearing gel beads in emulsions (GEM generation), (ii) reverse transcription and cDNA amplification, and (iii) 3′ gene expression library construction. The 3′ gene expression libraries were sequenced on a 10 B flow cell on a NovaSeq X plus sequencer at the Ramaciotti Centre for Genomics (UNSW Sydney).

Raw sequencing data were aligned to the rat reference genome (Rnor6-2024-A), and gene expression matrices were generated using the Cell Ranger software (10x Genomics). For data preprocessing and quality control, cell matrices were imported and analyzed using the Seurat R package (v5) (133). To ensure high-quality data, cells were filtered based on standard quality control metrics: Cells with fewer than 200 or more than 5000 detected features (nFeature_RNA) and cells with >5% mitochondrial RNA content were excluded from downstream analyses. Filtered data were normalized using Seurat’s NormalizeData function to scale and transform raw expression values for each gene. Highly variable features were identified using the FindVariableFeatures function. Data were then scaled to remove unwanted variation using ScaleData. Principal components analysis (PCA) was performed using the top 15 principal components as input for downstream analyses. The PCA embedding was used to construct a shared nearest neighbor graph and to identify clusters. Data layers were integrated using the canonical correlation analysis method implemented in Seurat. The integration was performed with the IntegrateLayers function, followed by layer joining using JoinLayers to create a unified dataset. Cell clusters were identified using gene lists generated from FindAllMarkers. DEGs were identified using the MAST (Model-Based Analysis of Single-Cell Transcriptomics) framework, which is specifically designed for single-cell RNA-seq data. The zlm function from the MAST package was used to fit a hurdle model to the single-cell expression data (134).

GSEA (preranked)

GSEA was performed using GSEA software (v4.3.3). For the analysis, unfiltered preranked data, including gene names and log₂FCs, were loaded into the software. The gene set databases used were “c5.all.v2024.1.Hs.symbols.gmt” or “c5.go.v2024.1.Hs.symbols.gmt” (for monocyte analysis). The “Collapse/Remap to gene symbols” option was set to “Collapse,” and the chip platform used was “Rat_Gene_Symbol_Remapping_Human_Orthologs_MSigDB.v2024.1.Hs.chip.” All other settings were left as default. The top 10 enriched gene sets, ranked by NES, were visualized using bar plots generated by the SRplot platform (an online tool for data visualization and graphing available at www.bioinformatics.com.cn) (135).

Enhanced volcano diagram and hyperbolic volcano diagram

DEGs were visualized using enhanced volcano and hyperbolic volcano diagrams. Unfiltered data, including gene names, log₂FCs, and FDRs, were uploaded to the SRplot platform (available at www.bioinformatics.com.cn) for visualization. For the hyperbolic volcano diagram, two hyperbolic curves with the function y = 1/x were added as reference lines to the traditional volcano plot, using the original threshold as the baseline. These curves facilitated the classification of genes, making it easier to identify genes with more significant differential expression.

GO (BP) pathway enrichment

Up- and down-regulated genes were those that had an FDR < 0.05 and a fold change greater or less than 0. Enrichment analysis was performed using the gost R function from g:profiler. Organism was set to rnorvegicus, and default settings were used, which incorporated multiple testing. g:profiler output includes a column titled P value, which as documented is actually an adjusted P value. Column title was renamed to p.adj accordingly. Pathway enrichment results were visualized using bubble charts. The top 20 gene sets, including “Term Name,” “Enrichment” (ratio of intersection size to query size), “p.adj,” and “Count” (intersection size), were uploaded to the SRplot platform (available at www.bioinformatics.com.cn) for visualization.

GO (BP) pathway enrichment bar plots for neutrophils

Pathway enrichment results for neutrophils were visualized using bar plots generated on the SRplot platform (available at www.bioinformatics.com.cn). The minimum gene set size (query size) was set to five genes for the analysis.

Dot plot for comparison of down-regulated pathway enrichment across five cell types

The comparison of down-regulated pathway enrichment across five cell types was visualized using a dot plot. Selected gene sets for each cell type, including “Cell Type,” “Term Name,” “Gene Count,” and “FDR,” were uploaded to the SRplot platform (available at www.bioinformatics.com.cn) for visualization.

Bulk RNA-seq and bioinformatics analysis

Ice-cold heart tissue (AAR) was transferred to labeled Precellys Lysing CK14 tubes, and TRIzol reagent (800 μl) was added. Tubes were placed into the TissueLyser adapter sets, and samples were homogenized thrice for 20 s at 6500 rpm, with 3-min intervals between each round of homogenization. Samples were centrifuged for 2 min, and supernatants were transferred to new 1.5-ml screw cap tubes then inverted to mix the contents, vortexed, and incubated for 15 min at 22°C. Total RNA was isolated from homogenized samples using the RNeasy Mini Kit (QIAGEN, catalog no. 74004). Chloroform was added to the mixture prior to microfuge centrifugation at 13,000 rpm for 15 min at 4°C. The upper aqueous layer (containing total RNA) was transferred to microtubes, and isopropanol was added and loaded into RNeasy columns. Columns were washed with buffers RPE and RW1. Total RNA was eluted with ribonuclease-free water. Samples were submitted to the UNSW Ramaciotti Centre for Genomics for TruSeq Stranded mRNA-seq preparation and sequencing by NextSeq 6000 to produce 75–base pair single-end reads. Bulk RNA-seq was sourced from two biologically independent samples for each of four separate conditions (Veh versus BT2 at 4 hours postperfusion; Veh versus BT2 at 24 hours postperfusion).

Bioinformatics analysis performed on this sequence data was conducted by the Australian Genome Research Facility and included adapter and quality trimming with Trim Galore, alignment to the reference genome with STAR, and gene count quantification with featureCounts. DEGs were identified using EdgeR. Raw expression counts were filtered for lowly expressed genes using a minimum counts per million threshold of 0.5, followed by normalization according to library size factors using the trimmed mean of M-values method. Last, expression data were fit to a negative binomial generalized linear model before undergoing a genewise quasi-likelihood F test to identify DEGs.

Volcano plots from RNA-seq DEG analysis

DEGs were visualized using enhanced volcano diagrams. Unfiltered data, including gene names, log₂FCs, and FDRs, were uploaded to the SRplot platform (www.bioinformatics.com.cn) for visualization.

KEGG enrichment from RNA-seq DEG analysis

Enrichment factors were calculated using the formula: enrichment factor = (DEG in category/total genes in category)/(DEG in background/total genes in background). Enrichment bubble plots and KEGG enrichment were generated using the SRplot web platform (www.bioinformatics.com.cn/en?keywords=bubble). The analysis results (including enrichment factor) were uploaded onto the web platform.

GEO dataset analysis

GEO dataset analysis was performed using the GEO2R interface (www.ncbi.nlm.nih.gov/geo/geo2r/). Data from the GEO2R were exported in xls format and analyzed for statistical significance using GraphPad Prism v9. Deidentified, open access human datasets were analyzed after obtaining UNSW Human Research Ethics Committee approval (HC230045).

The Australian Genome Research Facility conducted an analysis where the default GEO2R parameters were used to generate two datasets of DEG in GSE59867. First, samples collected from patients with STEMI within 1 day of AMI (all underwent direct PCI) who subsequently developed HF (n = 9) were compared to samples from a control group with patients with stable CAD (n = 46), and second, another comparison was made between samples from patients who did not develop HF, collected within 1 day of AMI (n = 8), versus a control group with stable CAD (n = 46). Significant DEGs were isolated based on the adjusted P value threshold of 0.05 and further subdivided into a list of up-regulated and down-regulated DEGs based on fold change for each dataset. The R package VennDiagram was used to detect and visualize the intersection of DEGs between these GEO2R analyzed datasets and the DEGs identified in the EdgeR analysis on BT2-regulated samples.

Statistics and reproducibility

Statistical analysis was performed using GraphPad Prism v9, which does not draw error bars when these are shorter than the height of the symbol. If distribution was not normal, Mann-Whitney or Kruskal-Wallis was performed. Normally distributed data were analyzed by t test or one-way analysis of variance (ANOVA). Plotted data represent mean ± SEM. Differences were considered significant when P ≤ 0.05. Where indicated, *P ≤ 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. ns denotes not significant. Investigators were not blinded during experiments given visual differences in BT2, BT3, and vehicle preparations. Quantitation for animal and cell experiments was performed with automated image analysis tools that ensured that same criteria and settings were applied across experimental samples to negate potential bias. Animals and cultures were randomly allocated to each group and handled and processed identically. Replicates (controls and treatments) were biologically independent and collected and analyzed uniformly both within and between experiments.

Supplementary Materials

The PDF file includes:

Figs. S1 to S10

Legends for tables S1 to S20

Other Supplementary Material for this manuscript includes the following:

Tables S1 to S20