HMGB1 Expression Level in Circulating Platelets is not Significantly Associated with Outcomes in Symptomatic Coronary Artery Disease

High-mobility group box 1 (HMGB1) is a DNA-binding protein typically expressed in the nucleus of mammalian cells that acts as a damage-associated molecular pattern molecule (DAMP) when released by dying or stressed cells [1-2]. Studies addressing the role of HMGB1 in myocardial infarction (MI) have generated conflicting results thus far, as detrimental [3-5] as well as salutary effects of HMGB1 [6-8] have been reported. Platelets store and, upon activation, export HMGB1 to the cell surface and release HMGB1 in significant amounts [9-11]. In a trauma/hemorrhagic shock model in mice with platelet-specific ablation of HMGB1, we have recently shown that platelet-derived HMGB1 is a critical mediator of thrombosis [9]. We and others have also demonstrated that platelet-derived HMGB1 induces the formation of neutrophil extracellular traps (NETs), which exert strong proinflammatory activity in various disease states [9, 12]. Myocardial infarction is intrinsically linked with both NETosis [13] and thrombosis [14]. The specific role of platelet-derived HMGB1 in coronary artery disease (CAD), and in particular in MI, remains unknown.

In this study, we examined HMGB1 expression on the surface of circulating platelets in patients with stable CAD and acute coronary syndrome (ACS) and evaluated potential effects of platelet HMGB1 on myocardial function and other clinical endpoints in these patients.

For the cohort study, blood samples were collected during percutaneous coronary intervention (PCI) and immediately analyzed for platelet surface expression of HMGB1 by flow cytometry. Blood was harvested through the catheter sheet before PCI. All subjects gave written informed consent. Patients were admitted to the Department of Cardiology of the University of Tübingen, Germany. We included 183 consecutive patients with symptomatic CAD (stable CAD: n = 80, ACS: n = 103). Inclusion criteria consisted of stable CAD, ACS and need for cardiac catheterization. Exclusion criteria were defined as age <18 and pregnancy. ACS was defined as worsening of angina, acute myocardial infarction (AMI), or sudden cardiac death. An AMI was diagnosed by a rise and/or fall of cardiac biomarker values [cardiac troponin (cTn)] with at least one value above the 99th percentile upper reference limit and with at least one of the following: symptoms of ischaemia, new or presumed new significant ST-segment–T-wave (ST–T) changes or new left bundle branch block, development of pathological Q waves in the ECG, imaging evidence of new loss of viable myocardium or new regional wall motion abnormality, or identification of an intracoronary thrombus by angiography [15]. The study was approved by the institutional ethics committee (270/2011BO1) and complies with the declaration of Helsinki and the good clinical practice guidelines [16].

During PCI, we investigated left ventricular ejection fraction (LVEF) by laevo-cardiography. Course of LVEF was evaluated using transthoracic echocardiography after a median of 6 months. Two-dimensional echo LVEF was assessed using Simpson’s biplane method of discs by manual planimetry of the endocardial border in end-diastolic and end-systolic frames [17]. All patients were tracked after initial PCI for clinical events including all-cause death and MI for 360 days after study inclusion. 19 patients were lost to follow up (10.4%). The primary combined endpoint was defined as the composite of all cause death (ACD) and/or MI during a 12 month follow-up. We defined secondary endpoints as the single events of ACD and MI during 12 month follow-up. Follow-up was performed until first occurrence of one of the pre-defined major endpoints. Follow-up was performed by telephone interview and/or review of patients´ charts on readmission by investigators blinded to the results of laboratory testing.

Circulating platelets in patients were analyzed by flow cytometry using a FACS Calibur flow cytometer with CellQuest software (BD Biosciences, Heidelberg, Germany). Blood was collected in citrate phosphate dextrose adenine and diluted 1: 30 with phosphate-buffered saline (PBS; Lonza, Verviers, Belgium). The samples were then incubated with PE-conjugated anti-CD42b monoclonal antibody (mouse IgG1; Beckman Coulter, Krefeld, Germany) and Alexa Fluor 488-conjugated HMGB1/HMG-1 monoclonal antibody (mouse IgG2B; R&D Systems, MN, USA). After staining, platelets were fixed with 1% paraformaldehyde and evaluated with flow cytometry.

All statistical analyses were performed using SPSS version 21.0 (SPSS, Inc., Chicago, IL, USA). Non-normally distributed data, including median fluorescence intensities (MFIs) and LVEF, were compared using the Mann–Whitney U test or the Kruskal-Wallis H test. Correlations were assessed by Spearman’s rank correlation coefficient (ρ). MFIs and LVEF are presented as median values with 25th and 75th percentiles. Cumulative event-free survival from all-cause death and all-cause death and/or MI was presented by Kaplan-Meier curves. The log-rank test was applied to compare cumulative survival between patients with HMGB1 values ≤ versus > median.

Patients' characteristics (age, gender, cardiovascular risk factors, co-medication) of the prospective cohort (n=183 patients), stratified according to subgroups stable CAD and ACS, are available in Table 1. Differences in medication between patients with stable CAD, unstable CAD, NSTEMI and STEMI are shown in Table 2. HMGB1 expression on platelets did not significantly differ between patients with ACS and patients with stable CAD [median MFI 9.26 (25^th;75^th percentile 7.24; 13.46) vs. 8.49 (25^th;75^th percentile 6.74; 12.06), (p = 0.220)]. No significant differences of platelet HMGB1 expression were detected between patients with stable CAD, unstable CAD, non-ST segment elevation myocardial infarction (NSTEMI), and ST segment elevation myocardial infarction (STEMI) (p=0.779) (Fig. 1).

Next, we investigated the effect of HMGB1 expression on the surface of circulating platelets on LVEF. Platelet HMGB1 did not significantly correlate with baseline LVEF, neither in the overall patient cohort (n = 133, ρ = 0.097, p = 0.266) nor in the MI patients subgroup (n=56, ρ=0.180, p=0.184). To investigate a possible association of platelet HMGB1 with myocardial repair mechanisms, we performed follow-up of LVEF in patients with MI (n = 46). We classified the course of LVEF into 3 groups: Worsening of LVEF after 6 months (>5% loss of LVEF), equal LVEF after 6 months (≤5% loss to ≤5% gain of LVEF), improvement of LVEF after 6 months (>5% gain of LVEF). Platelet HMGB1 expression did not significantly differ between these groups [median MFI 7.15 (25^th;75^th percentile 6.15; 11.42) vs. median MFI 9.28 (25^th;75^th percentile 8.01; 14.74) vs. median MFI 8.95 (25^th;75^th percentile 6.95; 10.58), respectively, p = 0.195] (Fig. 2). We also investigated the association of platelet HMGB1 and markers of myocardial necrosis (cTnI and CK). Platelet HMGB1 did neither correlate significantly with cTnI nor with CK in a subgroup of STEMI patients (n = 30) (ρ = -0.272, p = 0.198 and ρ = -0.026, p = 0.890, respectively).

All patients were tracked after initial PCI. Number and categories of events are shown in Table 3. No significant differences of baseline platelet HMBG1 expression levels were detected for the primary combined endpoint ACD and/or MI (no/yes) [median MFI 9.11 (25^th;75^th percentile 7.05; 12.16) vs. 8.63 (25^th;75^th percentile 6.97; 11.00), p = 0.897], nor the secondary endpoint all-cause death (no/yes) [median MFI 8.95 (25^th;75^th percentile 7.02; 12.16) vs. 9.42 (25^th;75^th percentile 6.84; 10.82), p = 0.553], nor the secondary endpoint myocardial infarction (no/yes) [median MFI 9.15 (25^th;75^th percentile 7.05; 12.16) vs. 7.83 (25^th;75^th percentile 6.68; 10.63), p = 0.529]. Moreover, patients with platelet HMGB1 expression above the median did not show a significantly altered cumulative event-free survival compared to patients with platelet HMGB1 expression below or equal to the median (log rank 0.951 for primary combined endpoint, log rank 0.557 for secondary endpoint all-cause death, and log rank 0.561 for secondary endpoint myocardial infarction, respectively) (Fig. 3).

In this cohort study, we have investigated the expression of HMGB1 on circulating platelets in patients with symptomatic CAD at the time of percutaneous coronary intervention and have shown that platelet HMGB1 expression i) does not significantly differ between stable CAD, unstable CAD, NSTEMI, and STEMI, ii) does not significantly correlate with LVEF, neither at baseline nor at 6 months follow-up of the MI subgroup, iii) does not exert any significant effect on outcome (composite of ACD and/or MI as well as single events ACD and MI), and iv) does not significantly affect cumulative event-free survival of these patients.

The role of non-platelet-specific HMGB1 in myocardial infarction has been investigated in previous studies, which have generated conflicting results thus far. In an MI mouse model, the administration of recombinant HMGB1 into the peri-infarcted left ventricle upregulated tissue healing through the activation of c-kit positive cells to form new myocytes and significantly improved the ejection fraction and other hemodynamic parameters after 2 and 4 weeks of HMGB1 treatment [8]. In transgenic mice with cardiac-specific overexpression of HMGB1, cardiac function and survival of mice after experimental MI was improved [6]. In an experimental MI model in rats, the subcutaneous administration of a neutralizing HMGB1 antibody resulted in expansion of the infarct scar and marked hypertrophy of the non-infarcted area after 2 weeks [7]. Thus, these studies have proposed a protective effect of HMGB1 on the myocardium after the induction of MI.

However, detrimental effects of HMGB1 in the context of MI have also been reported. HMGB1 serum levels were elevated 30 minutes after induction of I/R injury of the myocardium in mice, and administration of recombinant HMGB1 worsened I/R damage [3]. In patients with STEMI treated with percutaneous coronary intervention, plasma HMGB1 levels were elevated compared with healthy control subjects and were independently associated with increased mortality after a 10-month follow-up period [5]. Increased HMGB1 serum levels were also reported in patients with acute MI, which were associated with adverse clinical outcomes including pump failure, cardiac rupture, and in-hospital deaths [4].

In this study, HMGB1 expressed on the surface of circulating platelets in patients with symptomatic CAD was not significantly elevated in MI and had no significant effects on LVEF, other clinical endpoints, and survival of these patients within the investigation period. In a previous study, we observed elevated HMGB1 expression levels on circulating platelets in patients with trauma and hemorrhagic shock, and identified platelet-derived HMGB1 as a critical mediator of thrombosis [9]. Another study has confirmed a critical prothrombotic role of platelet-derived HMGB1 in a mouse model of venous thrombosis [18]. Although coronary thrombus formation undoubtedly plays an important role in the pathophysiology of myocardial infarction [14], the potential impact of HMGB1 expressed on the surface of circulating platelets on heart attack is unclear and debatable. HMGB1 serum levels are indeed increased in patients with acute MI and adversely associated with outcome [4-5]. The predominant source of serum HMGB1 after ischemic events, however, are dying (necrotic and apoptotic) tissue cells [1]. Thus, HMGB1 expressed by circulating platelets might not be a significant contributor in the context of CAD. However, we have shown previously that platelets are indeed the major source of HMGB1 within arterial thrombi [9]. The potential role of HMGB1 derived from platelets that aggregate locally in the coronary thrombi during myocardial infarction, however, remains unknown and requires further investigation.

In conclusion, HMGB1 expressed on circulating platelets in patients with symptomatic CAD is not significantly elevated and does not exert any significant effects on the disease outcome. Further studies are needed to clarify the specific role of platelet-derived HMGB1 in local and systemic pathophysiological processes in the context of myocardial infarction.

We are aware that our results are barely hypothesis generating. Our study has certain limitations mainly due to the observational character of the study, the low sample size and the single timepoints of HMGB1 measurements. Thus, we did not account for a potential variability of HMGB1 expression over time which might have an influence on outcome.

This work was supported by the DFG KFO 274 (VO 2126/1-1, GA 381/10-2). We acknowledge support by Deutsche Forschungsgemeinschaft and Open Access Publishing Fund of University of Tübingen.

The authors declare no conflict of interest.