Proteomic profiling of epicardial fat in heart failure with preserved versus reduced and mildly reduced ejection fraction

Abstract In order to explore the proteomic signatures of epicardial adipose tissue (EAT) related to the mechanism of heart failure with reduced and mildly reduced ejection fraction (HFrEF/HFmrEF) and heart failure (HF) with preserved ejection fraction (HFpEF), a comprehensive proteomic analysis of EAT was made in HFrEF/HFmrEF (n = 5) and HFpEF (n = 5) patients with liquid chromatography–tandem mass spectrometry experiments. The selected differential proteins were verified between HFrEF/HFmrEF (n = 20) and HFpEF (n = 40) by ELISA (enzyme‐linked immunosorbent assay). A total of 599 EAT proteins were significantly different in expression between HFrEF/HFmrEF and HFpEF. Among the 599 proteins, 58 proteins increased in HFrEF/HFmrEF compared to HFpEF, whereas 541 proteins decreased in HFrEF/HFmrEF. Of these proteins, TGM2 in EAT was down‐regulated in HFrEF/HFmrEF patients and was confirmed to decrease in circulating plasma of the HFrEF/HFmrEF group (p = 0.019). Multivariate logistic regression analysis confirmed plasma TGM2 could be an independent predictor of HFrEF/HFmrEF (p = 0.033). Receiver operating curve analysis indicated that the combination of TGM2 and Gensini score improved the diagnostic value of HFrEF/HFmrEF (p = 0.002). In summary, for the first time, we described the proteome in EAT in both HFpEF and HFrEF/HFmrEF and identified a comprehensive dimension of potential targets for the mechanism behind the EF spectrum. Exploring the role of EAT may offer potential targets for preventive intervention of HF.


Abstract
In order to explore the proteomic signatures of epicardial adipose tissue (EAT) related to the mechanism of heart failure with reduced and mildly reduced ejection fraction (HFrEF/HFmrEF) and heart failure (HF) with preserved ejection fraction (HFpEF), a comprehensive proteomic analysis of EAT was made in HFrEF/HFmrEF (n = 5) and HFpEF (n = 5) patients with liquid chromatography-tandem mass spectrometry experiments. The selected differential proteins were verified between HFrEF/HFmrEF (n = 20) and HFpEF (n = 40) by ELISA (enzyme-linked immunosorbent assay). A total of 599 EAT proteins were significantly different in expression between HFrEF/HFmrEF and HFpEF. Among the 599 proteins, 58 proteins increased in HFrEF/HFmrEF compared to HFpEF, whereas 541 proteins decreased in HFrEF/HFmrEF. Of these proteins, TGM2 in EAT was down-regulated in HFrEF/HFmrEF patients and was confirmed to decrease in circulating plasma of the HFrEF/HFmrEF group (p = 0.019). Multivariate logistic regression analysis confirmed plasma TGM2 could be an independent predictor of HFrEF/HFmrEF (p = 0.033). Receiver operating curve analysis indicated that the combination of TGM2 and Gensini score improved the diagnostic value of HFrEF/ HFmrEF (p = 0.002). In summary, for the first time, we described the proteome in EAT in both HFpEF and HFrEF/HFmrEF and identified a comprehensive dimension of potential targets for the mechanism behind the EF spectrum. Exploring the role of EAT may offer potential targets for preventive intervention of HF.

K E Y W O R D S
epicardial adipose tissue, heart failure with mildly reduced ejection fraction, heart failure with preserved ejection fraction, heart failure with reduced ejection fraction, proteomics comorbidities contribute to cardiac function is not yet fully understood but is clearly related to metabolic duress and low-grade inflammation. [2][3][4] A growing body of evidence have also demonstrated that patients with HF are suffering lipid-metabolic derangements, 5,6 substantially impacting long-term outcome. Recently, adipose tissue of particular interest in the major risk factor of HF, especially HFpEF, is epicardial adipose tissue (EAT), since these fat deposits lie directly over the surface of the myocardium with no fascia separating the two issues.
EAT is located between the visceral pericardium and the outer myocardium, and they share the same microvasculature. Depending on different physiological conditions, EAT acts as a buffer to absorb fatty acids or take pathologic metabolic activities after a 'phenotypic' transformation. 5,7 The relationship between increased EAT volume and HF has already been established. As previously described, increased EAT correlated with incident HF in the general population, especially HFpEF, 8,9 while a single study suggested EAT was reduced in patients with HFrEF. 10 Recently, a study by Pugliese et al. 11 posed an intriguing observation that indicated the role of EAT on HF might be divergent, in which, for HFpEF, EAT accumulation is related to worse haemodynamic and metabolic profiles, whereas in HFrEF, less EAT suggested more severe cardiac function and adverse prognosis. The results of Pugliese provided compelling clinical evidence on the diverging role of EAT across the EF spectrum, but the molecular signatures behind the pathophysiological of EAT on HFrEF/HFmrEF and HFpEF have not been systematically explored.
Besides, Jin et al. 12 reconfirmed a greater thickness EAT in patients with HFpEF than HFrEF/HFmrEF, and increased EAT thickness is associated with worse left atrial and ventricular function in HFpEF but opposite in HFrEF/HFmrEF. Our previous study 5 has pictured a comprehensive profile for connecting EAT with the pathogenesis of HF. Nevertheless, the proteomic difference of EAT between HFrEF/ HFmrEF and HFpEF remains largely unknown. Accordingly, in the present study, we intend to take a comprehensive proteomic analysis of EAT in patients with HFrEF/HFmrEF and HFpEF and investigate the substrate molecular changes that may be involved in this pathology.

| Study population and tissue sample
EAT were taken from 10 patients who were diagnosed with HF undergoing coronary artery bypass grafting or cardiac surgery for valve replacement between May 2021 and August 2021 and were divided into HFrEF/HFmrEF (n = 5) or HFpEF (n = 5) groups. The clinical manifestation and examinations [brain natriuretic peptide (BNP) >500 ng/L, enlarged left ventricular end-diastolic diameter and reduced left ventricular ejection fraction (<50%)] were the criteria for diagnosing HF. HFrEF/HFmrEF was defined by an LVEF <50%, while HFpEF required an LVEF ≥50%, N-terminal pro-B-type natriuretic peptide (NT-proBNP) >125 pg/mL and the additional presence of relevant structural heart disease or diastolic dysfunction. 13 Gensini Score which was a degree of stenosis score was multiplied by the lesion site score, and the sum of the lesion score was taken as the final score. 14 Baseline demographic characteristics and clinical data were taken at admission. Approximately 2 cm 3 EAT samples were taken from the left-interventricular groove, cut into small pieces, washed three times with ice-cold phosphate buffer saline, frozen for 10 min in liquid nitrogen and then stored at −80°C until analysis. All samples were collected a few minutes before the extracorporeal circulation under the same haemodynamic conditions. All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (the Ethical Committee of Beijing Chaoyang Hospital, China).

| Extraction of protein from epicardial adipose tissue
Proteins of EAT were extracted using RadioImmuno Precipitation

| Label-free quantitative proteomics by LC-MS/MS
Extracted peptides were separated at a 300 nL/min flow rate using a reversed-phase column (100 μm × 150 mm, 3 μm ReproSil-Pur 120 C18-AQ, 1.9 μm, Dr Math), followed by analysing with the EASY- Resolution of 70,000 for MS1 (at 200 m/z) and 17,500 for MS2 in an orbitrap analyser were processed in the following analyses. The automatic gain control target for MS1 and MS2 were set to 3.0 E +6 with max IT 50 ms and 5.0 E +4 with max IT 100 ms respectively. Thirty seconds were set for dynamic exclusion.

| Bioinformatics analysis
MaxQuant software (version 1.5.6.0) was used to perform the raw data. The following parameters were carbamidomethyl [C] as fixed modification, oxidation [M] and acetyl [protein N-term] as variable modifications, three missed cleavages were allowed. The false discovery rates (FDRs) of the peptide-spectra matches were set to less than 0.01 which were determined by a decoy database (Uniprot_ human_2016_09). Only unique and razor peptides were quantified.
All the others were reserved as default. Proteins established more than 99% probability with obtained two correct assigned peptides were considered to be identified. Normalized spectral protein intensity (LFQ intensity) was used to calculate the protein abundance.
Perseus program was used to calculate the significance of Log 2 LFQ intensity of proteins between HFrEF/HFmrEF and HFpEF groups.

The analysis of intracellular pathways was performed by searching Gene Ontology (GO) and Kyoto Encyclopedia of Genes and
Genomes (KEGG) database.

| Measurement of TGM2 levels in plasma
Peripheral venous blood was collected in a tube containing potassium EDTA and was centrifuged at 1500 g for 10 min at ambient temperature immediately. Plasma from the supernatant was collected and frozen at −80°C until analysis. Plasma samples were diluted 1:50000 when tested. Plasma concentrations of TGM2 (Proteinglutamine gamma-glutamyltransferase 2) were measured by ELISA (enzyme-linked immunosorbent assay) from a commercially available microplate kit (CUSABIO Biotech Co. Ltd) in accordance with the manufacturer's instructions. The intra-and inter-assay coefficients of variation were both less than 5%.

| Statistics
Continuous variables were expressed as the mean ± standard deviation for normal distribution or median (interquartile range) for non-normal distribution. Normality was processed using the Kolmogorov-Smirnov test. Categorical variables were expressed as frequencies and percentages. Comparisons between groups were

| Patient characteristics
Initially, a total of 10 HF patients were enrolled and were divided into HFrEF/HFmrEF or HFpEF groups. The demographic and clinical parameters of the patients are illustrated in Table 1. EF (42.7 ± 9.65 vs 61.8 ± 4.92) and LV end-diastolic dimension (60.4 ± 5.59 vs 50.0 ± 5.43) were significantly different between HFrEF/HFmrEF and HFpEF groups among all parameters (p = 0.004, 0.018, respectively).

| Human EAT proteome profile between HFrEF/HFmrEF and HFpEF group
We got a core set of 2407 quantified proteins in HFrEF/HFmrEF and HFpEF groups by importing the raw data to LC-MS/MS datH-FrEF/HFmrEF and HFpEF abase and searching it. Of these, 599 EAT proteins were significantly differentially expressed between the two groups. Of those 599 proteins, 58 proteins increased in HFrEF/HFmrEF relative to HFpEF, whereas 541 proteins decreased in HFrEF/HFmrEF. The volcano plots ( Figure 1A) showed the significant distribution and variation range of differential proteins between the HFrEF/HFmrEF and HFpEF group. The cluster heat map ( Figure 1B), separated groups with the EF in HF patients, using hierarchical analysis with a Pearson correlation. Based on the protein ratios of the 599 significant expressed EAT proteins, we identified the 10 most significantly up and down expressed proteins between the HFrEF/HFmrEF and HFpEF group ( Table 2). Of these, TGM2 was significantly down-regulated in the HFrEF/HFmrEF group (p < 0.001).

| Analyses of the differential EAT proteome between HFrEF/HFmrEF and HFpEF groups
We compared differentially expressed proteins between HFrEF/  (Figure 2A,B).

| Verified the selected protein
Ten candidate EAT proteins which were most significantly up-and down-regulated using label-free LC-MS/MS were obtained as primary targets for further clinical validation in larger cohorts. As presented in    (Figure 4 and Table 5). The value of the combination of TGM2 with Gensini scores for predicting HFrEF/HFmrEF was significantly higher than each parameter alone.

| DISCUSS ION
EAT is full of pro-and anti-inflammatory adipokines such as adiponectin, which maintains proper cardiac contractility in physiological conditions. 15 EAT seems to play an opposite prognostic role in HF patients, as EAT reduction is detrimental in HFrEF/ HFmrEF patients, while increased EAT has a positive role in HFpEF patients. 11 In HFpEF patients, EAT thickness is associated with worse haemodynamic and metabolic outcomes, for example, it is positively correlated with C-reactive protein, interleukin-6, uric acid and troponin T levels in HFpEF, which all together give rise to a pro-inflammatory status in myocardium associated with oxidative stress and fibrosis and are participated in the development of Abbreviations: EAT, epicardial adipose tissue; HFrEF, heart failure with reduced ejection fraction; HFmrEF, heart failure with mildly reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction.
HFrEF expressed significantly lower thermogenic genes in EAT, 18 suggesting a loss of functional EAT brown-like features would be deleterious. In the present study, we describe a comprehensive   Note: Data are mean ± SD or number (%); Gensini score and SYNTAX score, two effective tools used to evaluate the severity of coronary artery disease; TGM2, gene name of Protein-glutamine gamma-glutamyltransferase 2. *P < 0.05, ***P < 0.001.

TA B L E 3
Baseline characteristics of the 60 heart failure patients.
Physiologically, epicardial fat has important functions to maintain cardioprotective effects, including secretion of antiinflammatory cytokines, supplement of fatty acids to myocardium during stress situations, and mechanical protection of myocardium and coronary arteries. 19 Nonetheless, EAT may shift toward a cardiotoxic status under pathological conditions. The oxidative stress and inflammatory factors associated with HFrEF/HFmrEF in our network including unregulated CPM and CRP also took part in macrophage differentiation 20 and innate immune response 21 in the previous study. The recruitment of pro-inflammatory immunocytes and releasing pro-inflammatory mediators lay a foundation between EAT and cardiac impairment. Also increasingly expressed APOA, FABPH and low levels of DBI in response to reduced EF indicated a disorder of lipid metabolism in EAT, with direct lipotoxic effects on the myocardium. 22 In addition, HFpEF up-regulated EAT proteins, participate in collagen fibril organization, such as COL6A6, TNXB, PRTN3 and RAP1A, contributing to increased collagen turnover and severe diastolic dysfunction. 23 We then constructed proteinprotein-interactive network and validated the candidate protein with a larger cohort. However, future research is warranted to investigate the exact mechanism behind these associations of significantly expressed proteins.
In proteome analysis, TGM2 in EAT is the third down-regulated protein in HFrEF/HFmrEF patients. Following the validated stage, the peripheral blood concentration of TGM2 is identically decreased in HFrEF/HFmrEF than in HFpEF groups. Nonetheless, we acknowledge the following limitations in this study. First, a relatively small cohort limited further valuable information on the role of candidate EAT proteins. TGM2 should be further verified in a larger-scale prospective study with a longer duration of follow-up. Second, The BMI of the HFpEF patients is lower than that of usual HFpEF patients. This may be also related to the small sample size due to the precious source of epicardial tissue. Thirdly, a lack of a control group of patients without HF undergoing EAT proteomic profiling represents another limitation. Fourthly, there is no clear evidence that TGM2 is secreted by EAT and then released into the peripheral circulation. The source still needs further study. Finally, our study is on the discovery phase towards the role of EAT across the EF spectrum. Further basic investigation will be followed.
In conclusion, this current study, for the first time, compares the proteome in EAT between HFpEF and HFrEF/HFmrEF and identifies a comprehensive dimension of potential targets for the mechanism behind the EF spectrum. Our findings suggest EAT is not equal-it has heterogeneous properties that modify cardiovascular and HF risk, offering potential targets for preventive intervention.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The mass spectrometry proteomics data have been deposited to the Note: TGM2, gene name of Protein-glutamine gammaglutamyltransferase 2; Gensini score, effective tool used to evaluate the severity of coronary artery disease.