CD8+CD28null T Lymphocytes are Associated with the Development of Atrial Fibrillation after Elective Cardiac Surgery

Abstract Background Postoperative atrial fibrillation (POAF) is assumed as a complex and multifactorial interaction of different pathogenic factors. Data suggests an inflammatory process as the main trigger of this specific type of atrial fibrillation. CD8+ T lymphocytes that lack the surface protein CD28 were found to be crucially involved in chronic inflammatory processes within the cardiovascular system. Of utmost interest, these so-called CD8+CD28null T cells are known to present with autoaggressive behavior and deleterious cytotoxic effects on human tissue. Methods A total of 129 patients undergoing elective cardiac valve and/or coronary artery bypass graft surgery were enrolled. Fluorescence-activated cell sorting was performed to investigate lymphocyte subsets. Patients were stratified in two subgroups according to patients developing POAF (n = 60) and individuals free of POAF (n = 69). Results Comparing patients developing POAF to individuals free of POAF, the fraction of CD8+ lymphocytes was significantly higher in individuals developing POAF (30.5% [POAF] vs. 25.7% [no-POAF]; p = 0.021). Interestingly, also the fraction of CD8+CD28null T lymphocytes was significantly higher in the POAF subgroup (66.7% [POAF] vs. 61.6% [non-POAF]; p = 0.043). Multivariate logistic regression proved that the fraction of CD8+CD28null cells is a strong and independent prognosticator for the development of POAF with an adjusted odds ratio per 1 standard deviation of 3.21 (95% confidence interval 1.01–10.18; p = 0.048). Conclusion We found that cytotoxic CD8+CD28null T lymphocytes proved to be a strong and independent predictor for the development of POAF after elective cardiac surgery. Our results potentially indicate an autoimmune impact of this preexisting, highly cytotoxic T cell subset in the pathogenesis of POAF.


Introduction
Atrial fibrillation (AF) is the most common supraventricular tachycardia and cardiac arrhythmia. It is estimated that currently a total of 33.5 million people worldwide are affected by AF. 1 In the European Union, a progression up to 17.9 million in the group of 55 years and older is predicted till 2060. 2 Particularly one type of AF has been out of investigational focus in the past years. Postoperative AF (POAF) is a specific type of AF and the most common complication after cardiac surgery, occurring with an incidence ranging between 15.7 and 41.6%. POAF is associated with a higher risk for chronic heart failure, thromboembolic events, and poor patient outcome. 3 However, the development of POAF is assumed as a complex and multifactorial interaction of different pathogenic factors and is still not completely understood. Several series highlighted that elevated levels of inflammatory markers were associated with the development of POAF, consecutively assuming an inflammatory background as an important driver in the occurrence of POAF [4][5][6][7][8] While CD8 þ T lymphocytes play a pivotal cytotoxic role within the healthy immune system, there is growing evidence that persistent inflammatory activation, for example, diseases with chronic immune activation, leads to loss and irreversible downregulation of their costimulatory CD28 receptor. The malfunction in this so-called CD8 þ CD28 null T lymphocytes seems to negatively impact immune responses, leading over to enhanced cytotoxicity and regulatory functions, resulting in cardiac remodeling. 9 With the reference of several previous studies describing the impact of inflammation on the onset of AF, 10 the present study aimed to elucidate the inflammatory impact of CD8 þ CD28 null T cells on the development of POAF after elective cardiac surgery.

Study Design
A total of 129 patients were enrolled in this prospective observational single-center study, undergoing elective cardiac surgery of either coronary artery bypass graft (CABG) or valve surgery (repair or replacement) including minimal invasive procedures or a combination of both. All patients had to be at least 18 years old at the time of study inclusion, give written informed consent, and be free of any type of AF at time of study enrolment and as well 6 months prior surgical procedure. To ensure the absence of any active infections, chronic inflammatory conditions, autoimmune diseases, or malignancies, medical records and inflammatory markers including C-reactive protein (CRP) and total leucocyte count were screened prior to inclusion and if applicable resulted in exclusion from enrolment. Patients were included between August 2015 and September 2016, 1 day prior to surgical intervention at the Division of Cardiac Surgery, Department of Surgery of the General Hospital of Vienna (Medical University of Vienna) based on a standardized questioner. Additionally, peripheral venous blood samples were taken for further fluorescence-activated cell sorting (FACS) analysis at the time of study enrolment.
During postoperative hospitalization, standardized laboratory values were determined within 24 hours after surgical intervention and CRP levels were measured daily to assess baseline level and the maximum increase after surgery in accordance to the laboratory standards of the Medical University of Vienna (Cobas C System CHE2-Roche Diagnostics, Switzerland). Participants were equipped with a permanent 6-lead surface electrocardiographic monitoring until discharge to detect episodes of AF. Any episode of AF was documented via a 12-led surface electrocardiogram and POAF defined in accordance to the guidelines of the European Society of Cardiology as new-onset AF after major cardiac surgery. 11 Quality of data was guaranteed due to the regularly performed range and consistency checks. Additionally, with the prospective character of this study integrity of any tested variable was feasible. The study was approved by the local ethical committee of the Medical University of Vienna (No. 1110/2013) and complies with the 1975 Declaration of Helsinki.

Flow Cytometry
To examine the role of CD8 þ CD28 null T cells, lymphocytes were separated from fresh heparinized ethylenediaminetetraacetic acid blood samples using a standardized protocol. Lymphocytes and peripheral mononuclear cells were separated based on their buoyant density by performing gradient centrifugation with a high-grade polyethylene barrier. After centrifugation, the lymphocyte layer was harvested and washed. Finally, a cryopreservation medium (Synth-a-Freeze Cryopreservation Medium, Thermo Fisher, Waltham, Massachusetts, United States) was added for storage at -80°C.
Samples were unfrozen, underwent several cell washing procedures using warmed phosphate-buffered saline, and were stained and incubated for 30 minutes with fluorescein isothiocyanate conjugated Anti-CD8 (Biolegend, San Diego, California, United States) and APC/Alexa 647 conjugated Anti-CD28 (Biolegend). Flow Cytometry was performed using a Becton Dickinson FACS Canto II flow cytometry and FACS Diva. To asses CD8 þ CD28 null cells, gating for CD8, CD28, and total lymphocyte count was performed (►Supplementary Fig. S1, available in the online version). The total number of lymphocytes was calculated based on counting beads for all FACS samples. Results for CD8 þ CD28 null are quoted as the percentage of CD8 cells and CD8 as the percentage of lymphocytes.

Statistical Analysis
Categorical data are shown as frequency and percentage. Continuous variables were log-transformed prior to analysis and presented as median and interquartile range (IQR). Categorical data were analyzed with chi-square test and continuous data using Mann-Whitney U test.
The prognostic impact of inflammatory biomarker and immune cells on the development of POAF was evaluated using binary logistic regression analysis. Results were shown in odds ratio (OR) with a 95% confidence interval (CI). The multivariate model was adjusted for potential confounders, including age, gender and type of surgery.
Thrombosis and Haemostasis Vol. 120 No. 8/2020 Statistical significance was defined by a two-sided p-value of < 0.05. Within our cohort study with the given incidence of POAF, a sample size of more than 100 patients (according to at least 50 individuals per group) allows identifying a risk factor, which increases the assumed relative risk by 30% (power 80%, α 0.05). STATA 11 software (StataCorp LP, United States) and PASW 18.0 (IBM SPSS, United States) was used for statistical calculations.

Baseline Characteristics
In total, 129 patients were enrolled in this study. Out of them, 60 patients (46.5%) developed POAF and 69 (53.5%) were free of POAF (non-POAF group). The median time until the first onset of POAF was the second postoperative day (IQR: day 2-3). The study population median ages were 69 (IQR: 61-74) in the POAF subgroup and 68 (IQR: 57-74) in the non-POAF subgroup. CABG surgery (38.3%) was the most common surgery followed by valve surgery (35%) and the composition of CABG and valve surgery (26.7%). Baseline characteristics were overall comparable between the two subgroups. However, we observed a significantly higher number of individuals with severe aortic insufficiency in the non-POAF subgroup. Detailed baseline characteristics divided into the subgroups POAF and non-POAF are shown in ►Table 1.

Distribution of T Cell Subsets
In the bivariate analysis, the median total number of lymphocytes was lower in the POAF subgroup than in the non-POAF subgroup (12,

Binary Logistic Regression Analysis
We performed a binary logistic regression analysis to determine the correlation between the rate of CD8 þ CD28 null cells and the incidence of POAF. The percentages of CD8 þ CD28 null cells were log-transformed before binary logistic regression analysis. A high rate of CD8 þ CD28 null cells showed to be significantly associated with the development of POAF

Discussion
Several studies could demonstrate a strong association of inflammatory markers-such as CRP, tumor necrosis factor-α, interleukin (IL)-2, IL-6, and IL-8-and POAF. 4,6,12 CD8 þ CD28 null T cells are presumably associated with autoimmune diseases and chronic inflammation. They affect the immune response and also the immune deficit of elderly people adversely 9,13 leading to an increased perforin and granzyme b release, which is known to have an autoaggressive behavior against myocardial tissue and results in cardiac remodeling.
The current study assessed the association of CD8 þ CD28 null T cells with the development of POAF after elective cardiac valve and/or CABG surgery. These data are the first in the literature describing the influence of these cytotoxic cells on the development of POAF, indicating that an inflammatory predisposition plays a pivotal role in the occurrence of POAF. Intraoperative inflammation triggered due to surgical trauma, together with atrial fibrosis and preexisting risk factors (e.g., heart failure, myocardial infarction, obesity), possibly exceeds a certain threshold and POAF is developed. The highest peaks of CRP levels in the POAF subgroup were measured on the second postoperative day (IQR: day 2-3) which is accompanied by the median time until the first onset of POAF. This accords to recent evidence in the literature and suggests an inflammatory reaction based on intraoperative trauma rather than a systemic inflammation. 3 There was no difference observed in frequencies of Additionally, the adjustment for the respective type of surgery did not modulate the predictive potential of CD8 þ CD28 null T cells on POAF. Since mitral valve disease mirrors a major driver for the development of AF, a sensitivity analysis in individuals free of mitral stenosis also showed no modulation of the prognostic value of this highly cytotoxic T cell line.
A preoperative statin therapy was inversely associated with POAF and showed lower CRP levels in several studies. 14,15 Those findings were in line with our results showing an OR of 0.27 ([95% CI: 0.13-0.75; p < 0.001]; data not shown) for the development of POAF. However, we were not able to show a clear impact of prior statin therapy on frequencies of CD8 þ CD28 null cells (data not shown).

Limitations
At this point, a small study population size has to be mentioned. Considering the complexity of lymphocyte separation and FACS analysis, nevertheless, a respectable number of samples has been analyzed. Moreover, assessment of interand intra-assay coefficients and variability could not be done due to the lack of multiple measurements. However, a potential variability of measurements was limited by the use of Thrombosis and Haemostasis Vol. 120 No. 8/2020 standard operating procedures-including lymphocyte separation and cryopreservation-during the analysis of blood samples.
Considering the performed cryopreservation procedure of T cells and lack of a viability dye staining to exclude dead cells, we are not aware of modulations of the fraction of CD28 null cells at the time of analysis. We can reasonably rely on robust data available in literature, that the performed protocol for cryopreservation and Ficoll is not associated with a measurable loss of total CD4 þ or CD8 þ T cells or modulations in CD28 null frequencies after cell isolation and cryopreservation. 16 Unfortunately, a standardized assessment of left atrial diameter measures for all patients was not available for analysis. Since a comprehensive adjustment including valve disease (e.g., mitral stenosis as a major driver for left atrial enlargement) did not modulate the predictive potential of CD8 þ CD28 null cells on POAF, we might overcome a potential bias.
Furthermore, the study has been limited by a singlecenter study design and low ethnical diversification of our population.

Conclusion
Within the present investigation, increased frequencies of CD8 þ CD28 null T lymphocytes were found to be associated with the development of POAF. The surgical intervention per se potentially acts as an inflammatory trigger for cell activation. From a cellular perspective, the downregulation of the CD28 surface protein and both the associated autoaggressive and cytotoxic potential of this specific cell line might lead to scar formation beyond atrial myocardial tissue. Myocardial microscar formation and atrial fibrosis potentially impact on atrial electric conduction and foster the development of atrial fibrillating impulses after surgery.
From a clinical perspective, CD8 þ CD28 null T cells could be assessed preoperatively to stratify the postoperative risk    after elective valve and/or CABG surgery considering a personalized prognostic and treatment approach. Further investigations seem crucial to implement CD8 þ CD28 null assessment as a reliable predictor in routine use.
What is known about this topic?
• Downregulation of the CD28 surface protein might lead to scar formation. • Surgical intervention potentially acts as inflammatory trigger for cell activation What does this paper add?
• CD8 þ CD28 null T cells were found to be associated with the development of POAF and preoperatively stratify the postoperative risk.

Funding
This study was funded by Austrian Science Fund FWF (SFB-54).