Whole Blood Transcriptome Profiling Identifies Women with Myocardial Infarction with Non-Obstructive Coronary Artery Disease (MINOCA): Findings from the American Heart Association Go Red for Women Strategically Focused Research Network

Acute myocardial infarction (MI) occurs in >750,000 patients in the United States each year and is primarily caused by obstructive coronary artery disease (MI-CAD). Myocardial infarction with non-obstructive coronary arteries (MINOCA), defined as <50% diameter stenosis in all major epicardial arteries,¹ occurs in up to 15% of acute MIs and is more common in women.² Patients with MINOCA are at heightened risk for morbidity and mortality.³ The etiology of MINOCA is variable; potential mechanisms include atherosclerotic plaque rupture with non-occlusive thrombus, coronary artery spasm, coronary dissection and the alternate diagnoses of takotsubo syndrome and clinically unrecognized myocarditis.

Whole blood gene expression profiling represents an unbiased method to identify and investigate pathways associated with disease states. We hypothesized this technique would provide unique insight into mechanistic pathways that differ between women with MINOCA as compared to MI-CAD or disease controls.

Participants included women who presented to two urban tertiary-care-centers with MINOCA or MI-CAD and were prospectively enrolled in the Women’s Heart Attack Research Program (HARP) study (ClinicalTrials.gov; Unique identifier: NCT03022552). This study was approved by the New York University School of Medicine Institutional Review Board and all subjects gave informed consent prior to participation. Participants electively referred for invasive coronary angiography without acute coronary syndrome and did not have obstructive CAD were identified as disease controls. Peripheral whole blood was collected into PAXgene Blood RNA tubes during the index hospital admission for MI, typically within 24 hours of diagnosis. RNA was isolated and analyzed by RNA-sequencing (Illumina HiSeq4000 Sequencing). Transcript differential expression analysis was performed with DESeq, characterization of enriched cell types and pathways was assessed by modular gene expression analysis,⁴ and Ingenuity Pathway Analysis (IPA) was performed to assess associations between differentially expressed transcripts and canonical pathways. The significance of the association between transcript expression and pathway analysis was measured by Fisher’s exact test. Ethics and Institutional Review Board approvals were provided by New York University School of Medicine.

Gene expression profiles from women with MINOCA (n=11; median 65 years [IQR 62-70]; 100% white, 64% Hispanic), with MI-CAD (n=12; 60.5 years [53-64]; 92% white; 42% Hispanic), and controls (n=9; 66 years [60.5-74.5]; 67% white; 55% Hispanic) were available for analysis. Hierarchical clustering discriminated females with MINOCA from MI-CAD and disease controls (Figure 1A,B). A distinct 2239-transcript signature was defined in women with MINOCA versus controls (p<0.05; Figure 1A,C) and 2145 differentially expressed transcripts were identified between MINOCA versus MI-CAD (p<0.05; Figure 1B,D). Three hundred and fifty common transcripts were differentially expressed between MINOCA versus controls and MINOCA versus MI-CAD subjects.

Figure 1. Whole Blood Transcriptome Profiling Identifies Women with Myocardial infarction with Non-Obstructive Coronary Artery Disease (MINOCA).

Whole blood was collected into PAXgene Blood RNA tubes, RNA isolated and analyzed by RNA-sequencing (Illumina HiSeq4000 Sequencing). Participants were clustered in hierarchical condition trees with each row representing a single gene and each column an individual subject. Color scale indicates normalized expression values and colored bars indicate the individual’s clinical classification. (A) Transcript intensity for the top 1000 differentially expressed transcripts between MINOCA (n = 11) and control participants (n = 9). (B) Transcript intensity for the top 1000 differentially expressed transcripts between MINOCA (n = 11) and MI-CAD participants (n = 12). (C-D) Volcano plots of differential expressed genes in whole blood from MINOCA versus controls, and MINOCA versus MI-CAD patients. The y-axis corresponds to the p value (-log₁₀), and the x-axis displays the log2 fold change. The red dots represent the significantly differential expressed transcripts (p<0.05). (E) Modular analysis approach. Gene expression levels were compared between MINOCA and MI-CAD or control subjects on a module-by-module basis. Modules correspond to previously identified co-regulated gene clusters that were assigned biological functions by unbiased literature profiling⁴ (see legend grid below). Over-abundance of transcripts in a module is depicted in red, under-abundance in blue. The intensity of the dot corresponds to the percentage of genes in the respective module that are significantly (p<0.05) differentially expressed between groups. (F) Canonical pathway analysis (IPA) of significantly differentially regulated transcripts (p<0.05) in MINOCA versus controls and MINOCA versus MI-CAD patients, and subsequent comparison of pathways identified. The significance of the association between transcript expression and a canonical pathway was measured by Fisher’s exact test, -log(p-value). Estrogen receptor signaling was identified as the top common canonical pathway differentially expressed in both comparisons (MINOCA versus controls, -log(P-value) 4.5; MINOCA versus MI-CAD, -log(P-value) 3.4), followed by mammalian target of rapamycin (mTOR) (MINOCA versus controls, -log(P-value) 4.9; MINOCA versus MI-CAD, -log(P-value) 2.5) and eukaryotic initiation factor 2 (eIF2) signaling (MINOCA versus controls, -log(P-value) 3.1; MINOCA versus MI-CAD, -log(P-value) 4.4) (G) Relative expression of whole blood transcripts linked to estrogen signaling as determined by RNA-sequencing in MINOCA (n = 11), MI-CAD (n = 12) and control (n = 9) subjects.

Modular gene expression analysis was performed to identify enriched cell types or pathways, and found that participants with MINOCA had a reduced number of transcripts from genes associated with protein phosphatases and phosphoinositide 3-kinase (PI3K) family members (Figure 1E, M3.4) versus controls. Transcripts associated with cytotoxic T cells were overrepresented in participants with MINOCA versus MI-CAD (Figure 1E, M2.1).

To identify functional canonical pathway(s) unique to MINOCA, we performed IPA of the transcripts differentially regulated between MINOCA versus controls, MINOCA versus MI-CAD, and a subsequent comparison analysis. Estrogen receptor signaling was identified as the top common canonical pathway differentially expressed in both comparisons, followed by mammalian target of rapamycin (mTOR) and eukaryotic initiation factor 2 signaling (Figure 1F).

The identification of estrogen signaling as the most significantly differentially regulated pathway in females with MINOCA versus MI-CAD and controls is a novel finding and might be relevant to understanding mechanisms of MINOCA. Estrogens have multiple beneficial effects on the heart and arterial wall, including vasodilation, inhibition of inflammation, antioxidant effects, and endothelial/cardiac cell survival following injury.⁵ Additionally, enrichment in inflammatory cytotoxic T cell transcripts and eIF2 signaling highlights a distinct mechanistic inflammatory pathway in MINOCA versus MI-CAD and controls.

Binding of estrogen to one of three receptors, estrogen receptor (ER)α, ERβ or G protein-coupled estrogen receptor (GPER) mediates its effect. ERα and ERβ function as ligand-activated transcription factors mediating effects in hormonally regulated tissues, while GPER is as a major mediator of estrogen’s rapid cellular effects throughout the body. Our study found GPER to be the most abundant estrogen receptor transcript in whole blood cells. GPER expression was reduced in participants with MINOCA compared to both MI-CAD and controls (Figure 1G). Additionally, participants with MINOCA had decreased transcript expression of estrogen signaling related genes (GRIP1, MED13, MED13L, SOS2), estrogen regulated genes (cFOS), and reduced expression of MAPK, a pathway that is rapidly activated following estrogen-GPER.

Our data indicate that women with MINOCA have a unique whole blood transcriptional profile compared to patients with MI-CAD or women referred for invasive coronary angiography without obstructive CAD. We hypothesize these differences are mediated, in part, by estrogen signaling. Consistently, alterations to PI3K and mTOR signaling - established estrogen regulated pathways – was demonstrated in women with MINOCA. Together, these results demonstrate the potential of mRNAs to provide further insight into the molecular mechanisms leading to the development of MINOCA with the potential to identify future therapeutic targets.