Plasma exosomes characterization reveals a perioperative protein signature in older patients undergoing different types of on-pump cardiac surgery

Abstract

Although exosomes are extracellular nanovesicles mainly involved in cardioprotection, it is not known whether plasma exosomes of older patients undergoing different types of on-pump cardiac surgery protect cardiomyocytes from apoptosis. Since different exosomal proteins confer pro-survival effects, we have analyzed the protein cargo of exosomes circulating early after aortic unclamping. Plasma exosomes and serum cardiac troponin I levels were measured in older cardiac surgery patients (NYHA II-III) who underwent first-time on-pump coronary artery bypass graft (CABG; n = 15) or minimally invasive heart valve surgery (mitral valve repair, n = 15; aortic valve replacement, n = 15) at induction of anesthesia (T0, baseline), 3 h (T1) and 72 h (T2) after aortic unclamping. Anti-apoptotic role of exosomes was assessed in HL-1 cardiomyocytes exposed to hypoxia/re-oxygenation (H/R) by TUNEL assay. Protein exosomal cargo was characterized by mass spectrometry approach. Exosome levels increased at T1 (P < 0.01) in accord with troponin values in all groups. In CABG group, plasma exosomes further increased at T2 (P < 0.01) whereas troponin levels decreased. In vitro, all T1-exosomes prevented H/R-induced apoptosis. A total of 340 exosomal proteins were identified in all groups, yet 10% of those proteins were unique for each surgery type. In particular, 22 and 12 pro-survival proteins were detected in T1-exosomes of heart valve surgery and CABG patients, respectively. Our results suggest that endogenous intraoperative cardioprotection in older cardiac surgery patients is early mediated by distinct exosomal proteins regardless of surgery type.

Keywords: Aging; Cardiac surgery; Cardioprotection; Exosomes; Proteomics.

Fig. 1

Perioperative cardiac biomarkers and plasma exosomes. Baseline levels of cTnI (a) and CD63+pEXOs (b) isolated from patients undergoing MVR (n = 15), AoVR (n = 15), or CABG (n = 15). Perioperative variations of serum cTnI measured in patients undergoing MVR (n = 15; c), AoVR (n = 15; d), or CABG (n = 15; e) at baseline (T0) and 3 h (T1) and 72 h (T2) after aortic unclamping. f-i Perioperative variations of CD63 levels (n = 15; f) and corresponding profile of pEXOs concentration (particle/ml) (g-i) at T0, T1, and T2 in MVR group. l-o Perioperative variations of CD63 levels (n = 15; l) and corresponding profile of pEXOs concentration (particle/ml) (m-o) at T0, T1, and T2 in AoVR group. p-s Perioperative variations of CD63 levels (n = 15; p) and corresponding profile of pEXOs concentration (particle/ml) (q-s) at T0, T1, and T2 in CABG group. Data are expressed as mean values ± SD. §P < 0.05 vs. MVR; ¶P < 0.05 vs. AoVR;*P < 0.05 vs. T0; #P < 0.05 vs. T1. MVR mitral valve repair, AoVR aortic valve replacement, CABG coronary artery bypass graft, cTnI cardiac troponin I, pEXOs plasma exosomes, a.u. arbitrary unit

Fig. 2

Relationship between pEXOs levels and surgical parameters. Patients undergoning MVR (n = 15) (1 = 10 a.u.): a relationship between pEXOs values at T1 normalized to values at T0 (T1/T0) and cTnI (x₁), ECCT (x₂), and AoCCT (x₃) normalized values (linear model: y = 1 + x₁ + x₂ + x₃; R² = 1, adjusted R² = 0.995, P = 8.73e−06). b Relationship between pEXOs values at T2 normalized to values at T0 (T2/T0) and cTnI (x₁), ECCT (x₂), and AoCCT (x₃) normalized values (interactive model: y = 1 + x₁x₂ + x₁x₃; y = 1 + x₁x₂ + x₁x₃; R² = 0.779, adjusted R² = 0.641, P = 0.016). Patients undergoing AoVR (n = 15) (1 = 1 a.u.): c relationship between pEXOs values at T1 normalized to values at T0 (T1/T0) and cTnI (x₁), ECCT (x₂), and AoCCT (x₃) normalized values (polynomial model: y = 1 + x₁x₂ + x₂x₃ + x₃₂ + x₂/x₃₂; R² = 0.983, adjusted R² = 0.953, P = 0.00225). d Relationship between pEXOs values at T2 normalized to values at T0 (T2/T0) and cTnI (x₁), ECCT (x₂), and AoCCT (x₃) normalized values (polynomial model: y = 1 + x₁x₂ + x₁x₃ + x₂x₃ + x₃₂ + x₁/x₃₂ + x₂/x₃₂; R² = 0.972, adjusted R² = 0.908, P = 0.0093). Patients undergoing CABG (n = 15) (1 = 0.05 a.u.): e relationship between pEXOs values at T1 normalized to values at T0 (T1/T0) and cTnI (x₁), ECCT (x₂), and AoCCT (x₃) normalized values (polynomial model: y = 1 + x₁ + x₁₂ + x₂₂ + x₂x₃ + x₂₃ + x₂₂/x₃; R² = 0.903, adjusted R² = 0.747, P = 0.0343). f Relationship between pEXOs values at T2 normalized to values at T0 (T2/T0) and cTnI (x₁), ECCT (x₂), and AoCCT (x₃) normalized values (polynomial model: y = 1 + x₁₂ + x₁x₂ + x₂₂ + x₁x₃ + x₂x₃ + x₂₂x₃, R² = 0.954, adjusted R² = 0.852, P = 0.0231). MVR mitral valve repair, pEXOs plasma exosomes, cTnI cardiac troponin I, ECCT extracorporeal circulation time, AoCCT aortic cross-clamp time, AoVR aortic valve replacement, CABG coronary artery bypass graft

Fig. 3

Anti-apoptotic role of pEXOs. In vitro TUNEL assay: a HL-1 cardiomyocytes under normoxia or severe hypoxia and re-oxygenation (H/R) or H/R treated with (+) pEXO HSV or CABG. Apoptotic cells (green) and nuclear counterstaining with DAPI (blue). b Apoptotic cells (%) are expressed as mean ± SD. (n = 6 per group; *P < 0.05 vs. normoxia; #P < 0.05 vs. H/R). TUNEL TdT-mediated dUTP-biotin nick end labeling, HVS heart valve surgery, CABG coronary artery bypass graft, pEXO plasma exosome

Fig. 4

Proteomic profiling of pEXOs cargo. Exosomes were isolated from HVS_T0, HVS_T1, CABG_T0, and CABG_T1 plasma samples. a Repeatability of nLC-MS/MS analyses. b Example of spectral count reproducibility in CABG samples. c Example of spectral count reproducibility in HVS samples. d Virtual 2D map (MW vs pI) of proteins identified in HVS_T0, HVS_T1, CABG_T0, and CABG_T1 samples: in yellow proteins identified with an aSpC ≤ 1, in blue proteins identified with an 1 < aSpC ≤ 10, in red proteins identified with aSpC > 10. e, f, g Venn diagrams of proteins identified in HVS_T0, HVS_T1, CABG_T0, and CABG_T1 samples. pEXOs plasma exosomes, HVS heart valve surgery, CABG coronary artery bypass graft, nLC-MS/MS nano-scale liquid chromatographic tandem mass spectrometry, MW molecular weight, pI isoelectric point, SpC spectral count. (n = 6 per group; 3 biological replicates and 2 technical replicates per sample)

Fig. 5

Exosomal protein-protein interactions (PPI). a PPI network of proteins exclusively identified in HVS_T0 and HVS_T1 exosomes (Table S3) and proteins differentially delivered by comparing HVS_T0 vs HVS_T1 exosomes (Table 2). b PPI network of proteins exclusively identified in CABG_T0 and CABG_T1 exosomes (Table S3) and proteins differentially delivered by comparing CABG_T0 vs CABG_T1 exosomes (Table 3). Bolt and greater text (and node) dimension indicate differentially delivered proteins selected by fold change or/and G-test. Homo sapiens PPIs were retrieved by in-house manually curated PPI network and by STRING db. The networks were visualized by Cytoscape 3.5 software, while biological processes were retrieved by BINGO Cytoscape’s plugin. PPI protein-protein interaction, HVS heart valve surgery, CABG coronary artery bypass graft, db database. (n = 6 per group; 3 biological replicates and 2 technical replicates per sample)