In this study, we enrolled a total of 203 participants from our prior investigation [12], categorized into three cohorts. The initial group, designated as the sequencing group, encompassed 15 individuals diagnosed with acute myocardial infarction (AMI) and 15 individuals with non-cardiac chest pain (NCCP). The first validation cohort, included 20 AMI patients and 20 NCCP patients. The second validation cohort, comprised 85 AMI patients and 48 NCCP patients. The research adhered to the guidelines outlined in the Declaration of Helsinki and received approval from the Ethics Committee of Beijing Chao-Yang Hospital (2016-ke-101). All participants provided consent after receiving pertinent information.

Participants in this study were selected based on strict inclusion and exclusion criteria to ensure the integrity of our findings. For inclusion, AMI patients were identified through clinical symptoms, specific electrocardiogram (ECG) changes, and elevated cardiac biomarkers, namely troponin-I (TnI) and creatine kinase MB (CKMB), following the Universal Definition of Myocardial Infarction. NCCP controls included individuals with chest pain but normal cardiac biomarkers and no coronary stenosis, as confirmed by angiography. Exclusion criteria were applied rigorously to minimize confounding factors, excluding individuals with malignant diseases, severe organ failure (renal, hepatic, or heart failure), thyroid dysfunction, autoimmune diseases, acute or chronic infectious diseases, and pregnant women.

We employed standard venipuncture techniques to collect fasting venous blood samples from all participants. Subsequently, the obtained blood underwent processing for plasma extraction through centrifugation at 4 °C for 10 minutes at 3000 RPM. Following meticulous separation, the plasma samples were partitioned into smaller aliquots and preserved at a temperature of –80 °C until subsequent analysis.

Plasma-derived exosomes were acquired using the exoEasy Maxi kit (Qiagen, Hilden, Germany, 76064), following the producer’s instructions. In brief, plasma previously frozen at –80 °C was thawed at 37 °C and then subjected to centrifugation at 10,000 g for 15 minutes to isolate the supernatant. This 4 mL supernatant was transferred to a new 15 mL centrifuge tube. Subsequently, an equivalent volume of buffer was gently mixed with the plasma, and the mixture was incubated at room temperature. The blend was loaded into a column, centrifuged at 500 g for 1 minute, and the waste was discarded. The column was repositioned into the centrifuge tube. Next, 10 mL of buffer was added to the column, followed by centrifugation at 3000 g for 5 minutes. The column was then moved to a new centrifuge tube. Finally, 1 mL of elution solution was introduced to the column, incubated for 1 minute, and centrifuged at 500 g for 5 minutes. The filtered liquid was collected, reintroduced to the column for an additional 60 seconds, and subsequently subjected to centrifugation at a force of 500 g for 5 minutes. The resulting filtrate was gathered in a fresh eppendorf tube devoid of RNase and stored at –80 °C for future use.

For exosome size analysis, 1 mL of phosphate-buffered saline (PBS) was added to the exosome suspension and thoroughly mixed. Measurement was performed using the Zeta View-Particle Metrix nanoparticle tracking analysis instrument (ZetaView 8.04.02, Particle Metrix GmbH, Meerbusch, North Rhine-Westphalia, Germany). The sample was introduced into the sample chamber, and the software was initiated. Adjustments were made for brightness and focus, and the field of view was set for observation. The software then analyzed particle size by tracking the time and displacement of the particle’s motion trajectory.

For transmission electron microscopy (TEM) analysis of exosomes, the isolated exosomes underwent treatment with a 2% paraformaldehyde solution (P0099, Beyotime, Beijing, China). A small volume of the exosome solution was then applied to a Formvar-carbon-coated copper grid and allowed to incubate for 10 minutes at room temperature. The grid was rinsed with phosphate-buffered saline (PBS) solution and subsequently treated with a 1% glutaraldehyde solution (G6257, Merk, Darmstadt, Germany), followed by rinsing with distilled water. Afterward, the grid underwent treatment with uranyl acetate solution (U25690, Acmec, Shanghai, China) and methyl cellulose solution (ST1510, Beyotime, Beijing, China) before being air-dried. Finally, the copper grid was positioned in a sample holder and imaged using TEM (JEM-2100, JEOL Ltd., Tokyo, Japan) at 80 kV, with appropriate adjustments made to focus and brightness settings.

For Western blotting, the supernatant was mixed with 1/5 of 5$\times$ loading buffer and heated for 5 minutes. Subsequently, a 20 µL sample was loaded onto a 10% SDS-polyacrylamide gel for electrophoresis. After transferring the proteins onto a polyvinylidene fluoride (PVDF) membrane (3010040001, Merck KGaA, Darmstadt, Germany), they were blocked with 5% bovine serum albumin (BSA) (ST023, Beyotime, Beijing, China) for 1 hour. The membrane was then incubated overnight at 4 °C with antibodies specifically cluster of differentiation 63 (CD63) antibody (ab68418, 1:1000, abcam, Cambridge, MA, USA), heat shock protein 70 (HSP70) antibody (ab2787, 1:3000, abcam, Cambridge, MA, USA), or GAPDH antibody (ab9485, 1:5000, abcam, Cambridge, MA, USA). Following a TBST wash, the membrane was exposed to a 1:15,000 dilution of either anti-rabbit (92632211, LI-COR Biosciences, Lincoln, NE, USA) or anti-mouse secondary antibody (92632210, LI-COR Biosciences, Lincoln, NE, USA) and incubated for 1 hour at room temperature. After additional washes, the membrane was exposed for detection.

Total RNA from exosomes was extracted using the exosome RNA extraction kit (217184, Qiagen, Hilden, Germany). The exosome solution was mixed with lysis buffer, and after centrifugation, the supernatant was collected. Chloroform was added to separate the aqueous phase, which was then combined with ethanol. This mixture was applied to columns provided in the kit, subjected to centrifugation, and washed. The columns were transferred to new tubes, and the filter membranes were dried. Elution buffer was introduced, and after centrifugation, the columns were discarded. The filtrate was then collected and stored at –80 °C. The RNA concentration was determined using NanoDrop 2000 (Thermo Fisher Scientific, Waltham, MA, USA), which measured the quantity of RNA, its concentration, and the absorbance ratio at 260/280.

For sequencing profile generation, plasma samples from every 5 patients with AMI and every 5 patients with NCCP were amalgamated into one ‘sample’, leading to the creation of ‘3 AMI’ and ‘3 control’ samples for profiling. Initially, exosome RNA undergoes preliminary processing steps, including purification. Subsequently, cDNA is synthesized through reverse transcription, followed by end repair and adapter ligation to prepare the library. Once passing quality control, the libraries undergo sequencing.

After acquiring the raw sequencing data, the initial steps involve data filtering, adapter trimming, and processing of low-quality reads. This is followed by an assessment of sequencing quality and removal of ribosomal RNA (rRNA) to ensure the generation of high-quality data, termed as Clean Data. The clean data is initially mapped using Tophat2 software (Version 2.1.1.1, http://ccb.jhu.edu/software/tophat/index.shtml) for a standard alignment, excluding gene fusions (unfusion). This process aims to identify reads that cannot be directly mapped to the genome (unmapped reads). Subsequently, Tophat2’s Tophat-fusion module is employed to align these unmapped reads for fusion gene (fusion) mapping, capturing potential circRNA-forming reads. CircRNAs are identified using Python software (Version 3.9, https://www.python.org/) based on their unique back-splicing junctions, where exon sequences are rearranged. Junction reads are detected and refined with Tophat-fusion, focusing on typical GT/AG splice sites for precise circRNA identification. The expression levels of circRNAs are determined based on the predicted circRNA sequences. This involves counting the circRNA-associated reads, including both TopHat mapping and TopHat-fusion mapping reads, and calculating their RPM (Reads Per Million mapped reads). The count value for each circRNA is calculated using HTseq software (Version 2.0.0, https://htseq.readthedocs.io/en/latest/index.html), which then determines its RPM value. To calculate RPM, the number of reads mapping to circRNA is divided by the total number of mapping reads and multiplied by 10${}^6$.

Based on RPM results, Audic’s method is used to calculate the p-values and q-values for genes in comparison groups. CircRNAs with a fold change (log2(Fold change) $>$1) and a significance level (q-value 0.001) are selected to determine differential expression among samples. The overall distribution of differentially expressed circRNAs is represented using a Volcano Plot. Additionally, we analyzed genes associated with differentially expressed circRNA sequences within Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways to understand their impact on biological processes. By examining these pathways, this analysis provides an indirect approach to investigating the possible functions of circRNAs, establishing a foundation for future studies on their functionality. The Fisher Exact Test was employed to determine the significance (p 0.05) of these genes in KEGG pathways, identifying statistically significant signal transduction and disease pathways.

The expression levels of exo-circRNAs were quantified using the ViiA 7 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The relative fold-change was calculated using the 2${}^{-\mathrm{\Delta }\mathrm{\Delta }\text{CT}}$ method. The primer sequences employed for amplification are listed in Table 1.

Our previous study involved a total of 203 participants, distributed across three cohorts, as detailed earlier [12]. The sequencing group comprised 15 individuals diagnosed with acute myocardial infarction (AMI) and 15 individuals with non-cardiac chest pain (NCCP). This group aimed to investigate the profile of circulating exosomal circRNAs and identify potential candidate circRNAs through high-throughput sequencing. The first validation cohort included 20 AMI patients and 20 NCCP patients. Its primary objective was to validate the candidate exosomal circRNAs identified in the sequencing cohort using real-time PCR (RT-PCR). This cohort served as a small-scale validation. The second validation cohort comprised 85 AMI patients and 48 NCCP patients. The purpose of this cohort was to confirm the presence of exosomal circRNAs identified in the first validation cohort but on a larger scale. Fig. 1 illustrates the study flow and cohort distribution.

Fig. 1.

Study flow diagram. AMI, acute myocardial infarction; NCCP, non-cardiogenic chest pain.

Plasma exosomes were identified through Western blotting, nanoparticle tracking analysis (NTA), and transmission electron microscopy (TEM) following their isolation from blood samples. Western blotting revealed the presence of specific exosomal marker proteins CD63 and HSP70 in both groups (Fig. 2A). NTA indicated that the exosomes had a diameter of around 100 nm (Fig. 2B). TEM analysis illustrated the vesicular structure of exosomes, characterized by a circular, bilayer membrane (Fig. 2C). These findings are consistent with the typical features of exosomes.

Fig. 2.

Identification of plasma exosomes. (A) Western blotting results. (B) NTA results. (C) TEM results. HSP70, heat shock protein 70; AMI, acute myocardial infarction; NCCP, non-cardiogenic chest pain; NTA, nanoparticle tracking analysis; TEM, Transmission electron microscopy.

To investigate the profile of circRNA in plasma exosomes of individuals with AMI, we conducted a comprehensive sequencing analysis on samples from a group of 15 NCCP participants and another group of 15 AMI patients, referred to as the sequencing cohort. The baseline characteristics of the subjects used for sequencing are available in our previously published data [12]. No significant disparities in age and gender were observed between the AMI and NCCP cohorts. Consistent with the clinical features of AMI patients, levels of cTNT, CK-MB, and NT-proBNP were notably elevated in the AMI group compared to the NCCP group.

High-throughput sequencing revealed significant disparities in the expression profile of circRNAs in plasma exosomes of AMI patients compared to the control group. Based on the criteria of $|$log2(Fold change)$|$ $>$1 and q-value 0.001, a total of 893 circRNAs with differential expression were identified in AMI plasma exosomes. Among these, 118 circRNAs were upregulated, while 775 circRNAs were downregulated (Fig. 3A).

Fig. 3.

Differential expression and pathway analysis of circulating exosomal circRNAs in AMI patients. (A) Volcano plot displaying the differential expression of circRNAs in AMI patients and NCCP controls. (B) Distribution of KEGG Pathway Enrichment for genes matched with exo-circRNA sequences showing differential expression. (C) Five up-regulated circRNAs in AMI selected for the first validation cohort. AMI stands for acute myocardial infarction, and NCCP refers to non-cardiogenic chest pain. exo-circRNA, exosomal circular RNA; KEGG, kyoto encyclopedia of genes and genomes; RPM, reads per million mapped reads.

Studies have shown that some circRNAs function by regulating the linear transcripts of their corresponding genes [13, 14], indicating a significant association between circRNAs and the functions of their corresponding genes. Therefore, we analyzed genes associated with differentially expressed circRNA sequences within specific KEGG pathways. KEGG categorizes these pathways into four separate groups: Cellular Processes, Environmental Information Processing, Human Diseases, and Organismal Systems. The analysis revealed that genes associated with these circRNAs were notably concentrated in focal adhesion (Cellular Processes), mammalian target of rapamycin (mTOR) signaling pathway (Environmental Information Processing), pathways in cancer (Human Diseases), and thyroid hormone signaling pathway (Organismal Systems) (Fig. 3B). This suggests a direct correlation between the varying expression of specific circRNAs and their influence on these crucial biological pathways.

Among these differentially expressed circRNAs, 107 were annotated in circBase. To validate the results obtained through high-throughput sequencing and assess their potential as biomarkers, we selected five circRNAs that were upregulated according to the sequencing results. These circRNAs exhibited both high expression levels and substantial fold changes within the AMI group. Specifically, the selected circRNAs were: hsa_circ_0001535, hsa_circ_0003270, hsa_circ_0000972, hsa_circ_0001119, and hsa_circ_0001558 (Fig. 3C). To validate these findings, we conducted RT-PCR experiments on a small set of specimens, comprising our first validation cohort.

The RT-PCR validation results revealed that hsa_circ_0001119 consistently displayed a low and undetectable expression level in the majority of samples. Additionally, there was no significant difference in the expression of hsa_circ_0003270 between patients with acute myocardial infarction (AMI) and those with non-cardiac chest pain (NCCP). However, the detection results for hsa_circ_0001535, hsa_circ_0000972, and hsa_circ_0001558 were consistent with the sequencing results. Specifically, the relative amount of hsa_circ_0001535 in the plasma exosomes of AMI patients (8.49 (5.34, 13.94)) was significantly higher than that in the NCCP group (5.45 (3.50, 7.57)) (p 0.05), indicating a 1.56-fold increase in the AMI group (Fig. 4A). Fig. 4B depicts that the calculated area under the Receiver Operating Characteristic (ROC) curve for diagnosing AMI in this cohort is 0.685.

Fig. 4.

Validation results with the first validation cohort. (A–C) Relative expression levels of hsa_circ_0001535 (A), hsa_circ_0000972 (B) and hsa_circ_0001558 (C) in AMI and NCCP. (D–F) ROC curves of hsa_circ_0001535 (D), hsa_circ_0000972 (E) and hsa_circ_0001558 (F) for diagnosing AMI. ROC, receiver operating characteristic.

Furthermore, the relative expression level of hsa_circ_0000972 in patients with acute myocardial infarction (22.17 (15.73, 32.66)) was significantly higher than that observed in the non-cardiac chest pain group (12.40 (3.43, 25.08)) (p 0.05), indicating a 1.79-fold increase in the acute myocardial infarction cohort. The ROC curve’s area for diagnosing AMI in this group was found to be 0.683 (Fig. 4C,D). In comparison, the plasma exosomes of AMI patients exhibited a considerably higher relative expression level of hsa_circ_0001558 (27.62 (20.47, 44.81)) compared to the NCCP group (7.42 (4.21, 31.08)) (p 0.01), indicating a 3.72-fold increase in the AMI group. The calculated value for the area under the ROC curve in diagnosing AMI within this group was 0.790, as shown in Fig. 4E,F. Given that hsa_circ_0001558 demonstrated the highest area under the ROC curve for diagnosing AMI among these three circRNAs, it was selected for further validation using a larger sample size cohort.

To validate the exosomal circRNAs identified in the initial validation group, we conducted a larger sample validation with 85 AMI patients and 48 NCCP patients, focusing on hsa_circ_0001558. For insights into their clinical characteristics, please refer to Table 1 in our previously published data [12]. The demographic characteristics between the two groups were largely evenly distributed.

The findings from the second validation cohort of AMI and NCCP patients demonstrated similar results for the validation of plasma exosomal hsa_circ_0001558 as the first validation cohort. The exosomal hsa_circ_0001558 in the plasma of AMI patients (8.81 (4.87, 17.76)) exhibited a significantly higher relative expression level compared to the NCCP group (1.98 (0.74, 6.44)), showing a substantial disparity (p 0.01), with the AMI group having a 4.45-fold increase compared to the NCCP group (Fig. 5A). Fig. 5B shows that the ROC curve for diagnosing AMI had an area of 0.793.

Fig. 5.

Validation results with the second validation cohort. (A) Expression Levels of hsa_circ_0001558 in the plasma exosomes of NCCP and AMI Patients. (B) ROC Curve for the diagnosis of AMI by plasma exosomal hsa_circ_0001558. (C) Expression Levels of hsa_circ_0001558 in the plasma exosomes of NCCP, NSTEMI, and STEMI Patients. (D,E) ROC Curve for the diagnosis of NSTEMI (D) and STEMI (E) by plasma exosomal hsa_circ_0001558. (F) ROC curve for discriminating STEMI from NSTEMI by plasma exosomal hsa_circ_0001558. NSTEMI, non-ST-segment elevation myocardial infarction; STEMI, ST-segment elevation myocardial infarction.

Additionally, patients with AMI were categorized into two groups: ST-segment elevation myocardial infarction (STEMI) and non-ST-segment elevation myocardial infarction (NSTEMI). The NSTEMI group exhibited a significantly higher relative expression of plasma exosomal hsa_circ_0001558 (5.54 (4.00, 11.96)) compared to the NCCP group (1.98 (0.74, 6.44)), with a 2.80-fold increase in NSTEMI compared to the NCCP group (Fig. 5C). The ROC curve’s area under the curve (AUC) for the diagnosis of NSTEMI was 0.72 (Fig. 5D). The STEMI group exhibited a 5.27-fold increase in the relative expression of plasma exosomal hsa_circ_0001558 (10.44 (6.95, 22.93)) compared to the NCCP group and 1.88-fold increase compared to the NSTEMI group (Fig. 5C). Additionally, the area under the ROC curve demonstrated a value of 0.831 for diagnosing STEMI (Fig. 5E) and a value of 0.687 for differentiating STEMI from NSTEMI (Fig. 5F).