Coronary artery bypass grafting is associated with immunoparalysis of monocytes and dendritic cells.

Abstract Coronary artery bypass grafting (CABG) triggers a systemic inflammatory response that may contribute to adverse outcomes. Dendritic cells (DC) and monocytes are immunoregulatory cells potentially affected by CABG, contributing to an altered immune state. This study investigated changes in DC and monocyte responses in CABG patients at 5 time‐points: admission, peri‐operative, ICU, day 3 and day 5. Whole blood from 49 CABG patients was used in an ex vivo whole blood culture model to prospectively assess DC and monocyte responses. Lipopolysaccharide (LPS) was added in parallel to model responses to an infectious complication. Co‐stimulatory and adhesion molecule expression and intracellular mediator production was measured by flow cytometry. CABG modulated monocyte and DC responses. In addition, DC and monocytes were immunoparalysed, evidenced by failure of co‐stimulatory and adhesion molecules (eg HLA‐DR), and intracellular mediators (eg IL‐6) to respond to LPS stimulation. DC and monocyte modulation was associated with prolonged ICU length of stay and post‐operative atrial fibrillation. DC and monocyte cytokine production did not recover by day 5 post‐surgery. This study provides evidence that CABG modulates DC and monocyte responses. Using an ex vivo model to assess immune competency of CABG patients may help identify biomarkers to predict adverse outcomes.


| INTRODUC TI ON
Coronary artery disease is the leading cause of death globally 1 and can be treated by coronary artery bypass grafting (CABG). Exposure of patient blood to the bypass circuitry and re-perfusion following CABG is associated with activation of leucocytes, complement and cytokine secretion. 2 Activation of the patient's inflammatory response following CABG has been suggested to contribute to, and be a potential indicator of adverse outcomes including infection, increased ICU length of stay (LOS), atrial fibrillation (AF), myocardial infarction, respiratory distress, multiple organ dysfunction (MOD), acute kidney injury and mortality. [3][4][5] CABG can trigger the systemic inflammatory response syndrome (SIRS) characterized by leucocytosis, capillary leakage, MOD and an imbalance of pro-inflammatory cytokine production (eg IL-6, IL-8). 6,7 SIRS is partially counteracted by the compensatory anti-inflammatory response syndrome (CARS) characterized by the release of anti-inflammatory cytokines (eg IL-4, IL-10) and has been reported to occur in up to 40% of CABG patients. 8 An imbalance in immune homeostasis due to SIRS and/or CARS may increase the risk of CABG patients developing post-operative infection, particularly when CARS dominates over SIRS. 9 An imbalance in immune homeostasis can also lead to cell immunoparalysis, which renders the immune system unresponsive to a secondary insult (eg inflammation or infection). 10,11 Immunoparalysis or dysfunction of cells involved in the immune response could have detrimental implications for the patient, therefore restoring immune homeostasis is vital to prevent adverse patient outcomes. 12,13 In addition, preventative measures could be employed to manage the effects of a dysfunctional immune response. Changes to the patient immune profile following CABG can be divided into an early-phase and late-phase. 14 Direct contact of patient blood with non-endothelial surfaces associated with bypass circuitry triggers an early-phase response activating multiple cellular pathways (eg coagulation, fibrinolysis, complement) and cell types (eg lymphocytes, monocytes, neutrophils). 15 A late-phase response is then triggered by neutrophil-endothelial cell-mediated heart/lung ischaemia and re-perfusion injury. 14,15 Ischaemia damages endothelial cells resulting in neutrophil activation and the release of pro-inflammatory mediators. 14,15 Pro-inflammatory mediator release may also modulate immunoregulatory cells such as dendritic cells (DC) and monocytes. DC are antigen-presenting cells that play a central role at the interface of the innate and adaptive immune response. In human peripheral blood, DC comprise < 1% of the leucocyte population and release cytokines and chemokines required for antigen presentation and immune regulation. [16][17][18] Monocytes also link the innate and adaptive immune response and mediate antimicrobial host defence and removal of apoptotic cell debris. 19,20 Monocytes comprise 2%-12% of the leucocyte population and are a reservoir for myeloid precursors macrophages and DC. 19 Understanding the role DC and monocytes play in CABG-associated immunomodulation, and whether dysfunction in these cells contributes to adverse patient outcomes is limited. Therefore, we performed a prospective longitudinal assessment of the DC and monocyte phenotype in CABG patients at five time-points: admission, peri-operative, ICU, day 3 (D3) and day 5 (D5).

| Study approval and patient recruitment
Ethics approval (HREC/14/QPCH/117) was obtained from the Human Research Ethics Committee (HREC) at The Prince Charles Hospital (TPCH) which acts in accordance with the Declaration of Helsinki. Patients who were scheduled for CABG and met the inclusion criteria (Table 1) were enrolled with written informed consent. Whole blood (10 mL) was collected into an ethylenediaminetetraacetic acid (EDTA)-anticoagulated phlebotomy tube (Becton Dickinson (BD)) at five time-points-1) admission (just prior to anaesthesia but before surgery commencement), 2) peri-operative (after chest closure), 3) ICU (24 h post), 4) D3 post-operative and 5) D5 post-operative.
Maintenance was achieved with a propofol infusion and an inhalational agent-sevoflurane. The bypass circuit was primed with colloid and crystalloid, with packed red blood cells (PRBC) being reserved for patients with low pre-operative haemoglobin. Additives to the circuit included 2500 IU heparin, 20 mL of 8.4% sodium bicarbonate and 10 mL calcium chloride. The initial heparin dose was 300 IU/kg as a bolus maintaining the activated clotting time (ACT) > 400 s prior to aortic cannulation. The circuit included a hollow fibre membrane oxygenator, non-occlusive roller pumps and a heat exchanger. Flows were calculated on a cardiac index of 2.4 L min -1 m -2 with perfusion pressures maintained at 50-70 mm Hg. Core temperatures were reduced to maintain moderate hypothermia (32-34°C), monitored by nasopharyngeal or bladder temperature probes. Myocardial preservation was achieved with St Thomas I crystalloid solution 30 mL/kg and maintained with blood cardioplegia-4:1 ratio. Heparin was reversed at the end of the procedure with protamine (1 mg for every 100 IU of heparin used).

| Assessment of haematological parameters
Routine haematological parameters were determined by electronic impedance and absorption spectrophotometry (Abbott Cell-Dyn Emerald Haematology Analyser). Reference ranges used were obtained from QML Pathology 21 (Australia).

| Ex vivo whole blood inflammatory response culture model
DC and monocyte specific responses (surface co-stimulatory and adhesion molecule and intracellular cytokine production) were assessed using an established whole blood assay model. 22 Patient blood and RPMI 1640 media (containing 2 mmol L -1 L-glutamine; Gibco by Life Technologies) were cultured 1:1 (1.5 mL total volume) in 2 wells of a 24-well plate (Costar, Corning Life Sciences) for a total of 6 hrs (37°C 5% CO 2 ). Golgi plug (containing brefeldin-A; 1 μg/mL; BD Biosciences) was added after 1-h incubation to facilitate detection of cytokines produced within the cell via flow cytometry. Duplicate plates were run in parallel with the addition of lipopolysaccharide (LPS; TLR4 specific Escherichia coli 055:B5; 1 μg/mL; Sigma) to model an infectious complication and investigate immunoparalysis.

| Assessment of co-stimulatory and adhesion molecules
For wells that did not receive Golgi plug, cells were harvested, centri-

| Assessment of cytokine production
For wells that received Golgi plug, cells were harvested, centrifuged (1000 x g, 2 min) and stained with fluorescently labelled monoclonal antibodies to identify DC and monocyte populations as above (see 'Assessment of co-stimulatory and adhesion molecules').

| Statistical analysis
A repeated measures one-way ANOVA with Dunnett's post-test (with admission sample as the comparator) was used for analyses of haematological parameters, co-stimulatory and adhesion molecules and cytokine production at each time-point (P < 0.050; GraphPad Prism 7, GraphPad software). Monocytes and DC express HLA-DR; therefore, we used the median fluorescent intensity (MFI) to assess changes in monocyte and DC HLA-DR expression. The percentage of positive (%POS) cells was used to assess changes to the monocyte and DC immune profile. MFI was used to assess changes to the monocyte and DC immune profile when co-stimulatory and adhesion marker expression or cytokine production was ≥ 90% at admission. Using the admission sample as the comparator, a ratio was calculated at each time-point to determine changes from baseline.
The log 2 of this ratio was then calculated to equalize data around zero. In this way, +1 indicates a 2-fold increase and −1 indicates a 2-fold decrease. Differences in inflammatory responses were also investigated for association with clinical outcomes post-operative AF and prolonged ICU LOS (long > 24 h; Spearman's correlation, P < 0.050; GraphPad Prism 7).

| Patient characteristics and basic haematological parameters
The cohort of 49 patients was predominantly male (90%), and the average age at surgery was 68 years ( Table 2). The average ICU LOS for the patient cohort was 40 h, and of the 49 patients, 17 developed post-operative AF ( Table 2). As expected, CABG resulted in significant modulation of patient's basic haematological parameters which include an increased white blood cell count over the post-operative period and a reduced red blood cell count and haemoglobin levels particularly during CABG (Table 3).
DC are key immune mediators required to regulate and activate B cells, T cells and natural killer cells. [16][17][18]

| CABG suppressed monocyte and DC cytokine and chemokine production
Dysfunction of monocyte cytokine and chemokine signalling capacity can impact downstream processes associated with the patient immune response. 19,20 Monocyte production of MIP-1α, IL-10 and IL-8 was significantly suppressed during CABG and throughout the postoperative period (Figure 2A-C). In addition, monocyte production of MCP-1, IL-6, IL-12, TNF-α, IP-10 and MIP-1β was also suppressed from the ICU period through to D5 post-CABG ( Figure 2D-I). Of note, while monocyte IL-1α production was increased during CABG, it was followed by a period of post-operative suppression ( Figure 2J).
Monocyte cytokine and chemokine levels did not return to baseline levels by D5 post-surgery.
DC cytokine and chemokine production is an essential secondary signal to trigger antigen presentation and induce T-cell priming towards a pro-inflammatory state. 25  DC capacity to produce cytokines and chemokines did not recover by D5 post-surgery.

| Monocyte and DC co-stimulatory and adhesion molecule expression was immunoparalysed following CABG
Immunoparalysis is a form of acquired immunodeficiency which renders the patient immune system unresponsive to secondary insult (eg LPS) as a result of an imbalance in immune homeostasis. 26

| Monocyte and DC cytokine production was immunoparalysed following CABG
Utilizing the ex vivo model of bacterial infection, we found patient monocytes and DC were immunoparalysed following CABG, evidenced by a F I G U R E 2 Assessment of the monocyte and DC-specific immune profile following CABG. Open blue squares indicate monocytes and open red circles indicate DC. Y-axis indicates the log 2 fold change data. X-axis indicates sample collection time-point (Admission (Admin)) and peri-operative (OT), ICU, D3, D5). Data from 49 CABG patients. ANOVA indicated by horizontal bar with P value above. An asterisk indicates P value for monocytes and a hash indicates P value for DC. Dunnett's post-test (admission vs. sample time-point) is indicated at specific timepoints by an asterisk for monocytes or hash for DC as follows: * (or #) P < 0.050, ** (or ##) P < 0.010, *** (or ###) P < 0.001, **** (or ####) P < 0.0001 deficiency of monocytes and DC to produce cytokines and chemokine in response to LPS stimulation. Monocyte production of MIP-1α, IL-10, IL-8, TNF-α and MIP-1β was suppressed and had a reduced capacity to respond to LPS during surgery and over the entire post-operative period (Figure 4A-C, G, I). Of note, monocyte IL-8 production increased during CABG, before a period of suppression ( Figure 4C). In addition, monocyte production of MCP-1, IL-6, IL-12, IP-10 and IL-1α was down-regulated and had a reduced capacity to respond to LPS from 24 h post-CABG ( Figure 4D-F

| D ISCUSS I ON
CABG remains as one of the most common cardiac surgery procedures worldwide. 32  immunoparalysis. [10][11][12] Immunoparalysis is hypothesized to be a protective adaptation to an overwhelming pro-inflammatory response rendering key cells of the immune response unresponsive to a secondary insult (eg LPS). 10,11 When cells become paralysed, their capacity to respond appropriately to additional stimulus is reduced 10,11 and is characterized by reduced monocyte HLA-DR surface expression and TNF-α production. 41  AF is reported as the most frequent post-operative complication observed in the cardiac setting, with 30%-50% of patients undergoing cardiac surgery developing AF post-operatively. This prolongs admission, may require anticoagulation if persistent and is associated with poorer patient outcomes. 48 Prolonged ICU LOS can impede long-term patient recovery and increase hospital expenses. 49 This is multifactorial, being influenced by infection, ventilation time and the need for blood transfusion. 49 Preliminary findings from our study suggest a modulated DC and monocyte phenotype is associated with prolonged ICU LOS and post-operative AF. Early prediction of the risk for adverse outcomes may guide early patient intervention and management in order to minimize post-operative complications, improving patient outcomes.
Post-operative AF is difficult to prevent, and while we do not suggest that these changes are solely responsible for it, it may provide avenues for further study in terms of prevention. While this study investigates the CABG patient immune profile, our findings warrant further investigation of predictive markers of adverse outcomes.
We provide evidence that CABG suppresses the monocyte and

ACK N OWLED G EM ENTS
Australian governments fund Australian Red Cross Lifeblood to provide blood, blood products and services to the Australian community. This work was supported by a new investigator grant from The Prince Charles Hospital Foundation (grant to AP).

Dr. Perros reports a new investigator grant from The Prince
Charles Hospital Foundation during the conduct of the study. All other authors confirm no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.