Platelet activation and endothelial dysfunction biomarkers in acute coronary syndrome: the impact of PCSK9 inhibition

Abstract

Aims

Platelet activation and endothelial dysfunction contribute to adverse outcomes in patients with acute coronary syndromes (ACS). The goals of this study were to assess the impact of proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibition on markers of platelet activation and endothelial dysfunction in ACS patients and the interaction among PCSK9, platelets, and endothelial cells (ECs) on left internal mammary artery (LIMA) vascular endothelium using specimens obtained during coronary artery bypass surgery (CABG).

Methods and Results

Acute coronary syndromes patients enrolled in the Evolocumab in ACS trials were randomized to placebo or a single dose of 420 mg evolocumab within 24 h of hospitalization. Serum samples for analysis of platelet factor 4 (PF4) and P-selectin, markers of platelet activation, and von Willebrand factor (vWF), a marker of endothelial dysfunction, were obtained at baseline and 30 days. Additionally, LIMA segments obtained during CABG from patients who were and were not receiving evolocumab were immunostained with PCSK9; CD61, a platelet-specific marker; and CD31, an endothelial cell-specific marker. Forty-six participants were randomized to placebo or to evolocumab. Controlling for baseline levels, PF4 and vWF were significantly lower in the evolocumab, than in the placebo, group at 30 days. Immunostaining of LIMA specimens from twelve participants undergoing CABG revealed colocalization of PCSK9, CD61, and CD31 at the vascular endothelium. Administration of evolocumab was associated with decreased overlap of PCSK9, CD61, and CD31.

Conclusions

Proprotein Convertase Subtilisin/Kexin 9 inhibition decreases markers of platelet activation and endothelial dysfunction in ACS patients. PCSK9 is associated with platelets and vascular ECs in LIMA segments and PCSK9 inhibition decreases that interaction.

Graphical Abstract

Graphical Abstract.

Patients presenting with an acute coronary syndrome were randomized 1:1 to placebo or to 420 mg of evolocumab within 24 h of hospital admission. Biomarkers of platelet activation and endothelial dysfunction were measured at baseline (before randomization) and 30 days. Administration of evolocumab was associated with decreased levels of platelet factor 4 (PF4) and von Willebrand factor (vWF). In arterial specimens of participants undergoing coronary artery bypass surgery, there was a decreased interaction among PCSK9, platelets, and endothelial cells in the evolocumab group.

Abbreviation list

 ACS:

Acute Coronary Syndrome

CABG:

Coronary Artery Bypass Graft Surgery

CV:

Cardiovascular

ECs:

Endothelial Cells

EVACS:

Evolocumab in Acute Coronary Syndrome

IF:

Immunofluorescence

IHC:

Immunohistochemistry

LDL-C:

Low-Density Lipoprotein-Cholesterol

LDLR:

Low-Density Lipoprotein Receptor

LIMA:

Left Internal Mammary Artery

NO:

Nitric Oxide

PCSK9:

Proprotein Convertase Subtilisin/Kexin 9

PF4:

Platelet Factor 4

vWF:

von Willebrand Factor

Introduction

Despite guideline directed therapies that markedly improve outcomes in patients with an acute coronary syndrome (ACS),^1,2 there remains an increased risk of recurrent ischemic events during the early post-infarction period.³ This may be due, in part, to inflammation-associated coronary endothelial injury resulting in plaque instability and rupture and to platelet activation, promoting a pro-thrombotic milieu.

Proprotein Convertase Subtilisin/Kexin 9 (PCSK9) is a glycoprotein synthesized primarily in the liver and levels are increased in ACS patients. In addition to its cholesterol regulating effects mediated by binding to the low density lipoprotein receptor (LDLR),^4,5 preclinical and clinical studies indicate that PCSK9 also promotes platelet activation⁶ and vascular endothelial dysfunction.⁷ Although PCSK9 inhibition has been studied in stable patients in the outpatient setting,^8–10 the impact of its administration in the acute, hospitalized setting on markers of endothelial dysfunction and platelet activation has not.

We therefore investigated the effect of PCSK9 inhibition with evolocumab on platelet factor 4 and P-selectin, markers of platelet activation, and on vWF, a marker of endothelial dysfunction, in ACS patients enrolled in the double-blind, prospective, randomized, and placebo-controlled EVACS trials and hypothesized that administration of the clinical PCSK9 monoclonal antibody evolocumab would decrease these markers. In addition, we hypothesized that PCSK9 colocalizes with platelets and ECs on the endothelial vascular surface, and that in patients receiving evolocumab this association is disrupted, breaking a link among PCSK9, platelets, and ECs at that site.

Methods

Study population

The Evolocumab in ACS (EVACS I, NCT03515304) and Evolocumab in Patients with Acute Myocardial Infarction (EVACS II, NCT04082442) are two investigator-initiated, prospective, randomized, double-blind, and placebo-controlled trials conducted at the Johns Hopkins Hospital and the Johns Hopkins Bayview Medical Center.^11,12 Patients presenting with either a NSTEMI and a troponin-I ≥5 ng/ml or a STEMI were randomized within 24 h of hospital admission to placebo or to a single dose of 420 mg evolocumab subcutaneous (1:1 ratio). Blood samples were collected before study drug administration and at 30 days following randomization. An independent Data Safety and Monitoring Board monitored the progress of the studies, which were approved by the Johns Hopkins Institutional Review Board.

PF4 measurement

PF4 concentration levels were measured from serum aliquots stored at −80°C using a commercial ELISA kit (R&D Systems). The minimum detectable dose of human PF4 ranged from 0.010 to 0.100 ng/ml and intra‐assay and inter-assay coefficients of variation were 8.1%±0.5% and 11.8%±3.2%, respectively. All measurements were performed using a fully automated ELISA system (DS2, Dynex Technologies, Inc., Chantilly, VA, USA).

P-selectin measurement

P-selectin concentration levels were measured on serum aliquots stored at −80°C using a commercial ELISA kit (R&D Systems). The minimum detectable dose of human P-selectin ranged from 0.023 to 0.121 ng/ml and intra‐assay and inter-assay coefficients of variation were 8.9%±2.3% and 11.9%±3.6%, respectively. All measurements were performed using a fully automated ELISA system (DS2, Dynex Technologies, Inc., Chantilly, VA, USA).

vWF measurement

vWF concentration levels were measured on serum aliquots stored at −80°C using a commercial Luminex Discovery assay kit (R&D Systems). The sensitivity of the assay was 11.9 pg/ml and the bead region was 61. Intra-assay coefficients of variation were <5%, and inter-assay variances were <10%. All measurements were performed using a MAGPIX xPONENT 4.2 System (R&D systems, Biotechne, Minneapolis, MN, USA).

Immunofluorescence

Multiple left internal mammary artery (LIMA) segments from each participant were formalin-fixed and paraffin-embedded. The specimens were cut at a 5 μm thickness and placed on plus-microscope slides. Slides were deparaffinized and rehydrated (via xylene and ethanol, respectively), underwent antigen retrieval (1% Tris-HCL and 0.1% Proteinase-K for 30 mins at 37°C), and then blocked with 5% bovine serum albumin containing 0.1% Triton-X for 1 h at room temperature, as previously described.^13,14 Three 1X phosphate buffered saline (1X PBS) washes were performed (5 mins each) between each step. Slides were then stained overnight at 4°C with the following primary antibodies in blocking buffer: goat polyclonal anti-PCSK9; mouse monoclonal anti-CD61-FITC (to assess the presence of platelets); and mouse monoclonal anti-CD31 (to assess the presence of ECs). The following day, the slides were washed three times (1X PBS, 5 min each), then incubated at room temperature for 1 h with host-specific secondary antibodies (1:5000; except for anti-CD61-FITC, which already contained a conjugated fluorescent tag). Following an additional three washes, slides were cover slipped with ProLong Gold mounting media containing DAPI and allowed to cure overnight. The immunoreactive signals were then imaged on a Keyence BZ-X800 microscope (Keyence; Itasca, IL, USA). Percent overlap of immunoreactive signals was determined by measuring the length of the protein of interest divided by the length of either CD61 or CD31, then multiplied by 100 (Figures 2, 3, 4).

Figure 2.

PCSK9 colocalizes with ECs and platelets on LIMA vascular endothelium. Representative LIMA specimens obtained during CABG, immunostained via immunohistochemistry for (A) PCSK9 and (B) platelets (CD61). Black and white dashed boxes are enlarged images (Ai—Di). Red arrowheads indicate representative immunosignals of PCSK9 (Ai) and CD61 (Bi) in ECs. (C, D) Representative LIMA segments immunostained via immunofluorescence for (C) CD31, PCSK9 and DAPI and (D) for CD31, CD61 and DAPI. Yellow arrowheads indicate colocalization of PCSK9 and ECs (red-green merge). White arrows indicate localization of platelets at the LIMA endothelium (purple). Black and white scale bars: small = 250 µm, large = 100 µm. CABG, coronary artery bypass graft surgery; ECs, endothelial cells; LIMA, left internal mammary artery; PCSK9, proprotein convertase subtilisin/kexin type 9.

Figure 3.

Evolocumab effect on PCSK9 and platelet colocalization on the vascular endothelium. Representative LIMA segments from patients undergoing CABG, immunostained for PCSK9 via immunohistochemistry (A, C) or for CD31, PCSK9, CD61, and DAPI via immunofluorescence. (B, D) (Ai-Aii, Ci) Black and (Bi-Bii, Di) white dashed boxes are enlarged images. Black or white scale bars: small 250 µm, large 100 µm. (Ai-Aii) Red arrowheads indicate PCSK9 immunosignal at ECs. (Ci) Black arrowheads indicate PCSK9 immunosignal primarily localized in the adventitia. (B, D) White arrowheads indicate CD61 and PCSK9 immunosignal overlap. (E) Schematic illustration of the method used for the calculation of percentage overlap (%overlap) between PCSK9 and CD61 (F). Scatterplot depicting data points representing PCSK9/CD61 %overlap in LIMA specimens from the Evolocumab- (blue bar) and Evolocumab+ (red bar) groups (*p = 0.030). CABG, coronary artery bypass graft surgery; ECs, endothelial cells; LIMA, left internal mammary artery; PCSK9, proprotein convertase subtilisin/kexin type 9.

Figure 4.

Evolocumab effect on PCSK9 and ECs colocalization on the vascular endothelium. Representative LIMA segments immunostained for CD31 via immunohistochemistry (A, C) and for CD31, PCSK9, and DAPI via immunofluorescence. (B, D) (Ai, Ci) and (Bi, Di) are enlarged images of the black/white dashed boxes. Black or white scale bars: small 250 µm, large 100 µm. (Ai, Ci) Red arrowheads indicate CD31 immunosignal in ECs. (Bi, Di) White arrowheads indicate CD31 and PCSK9 immunosignal overlap. (E) Schematic illustration of the method used for the calculation of %overlap between PCSK9 and CD31 (F) Scatterplot depicting data points representing PCSK9/CD31 %overlap in LIMA specimens from the Evolocumab- (blue bar) and Evolocumab+ (red bar) groups (**p = 0.007). ECs, endothelial cells; LIMA, left internal mammary artery; PCSK9, proprotein convertase subtilisin/kexin type 9.

Immunohistochemistry

LIMA slides, adjacent to those described above, were baked in a 60°C oven for 20 min, and then incubated in fresh xylenes three times for 5 min each. Next, the slides were incubated in a decreasing ethanol gradient for 1 min each. The slides were transferred to a tub of 1X ImmunoDNA Retriever with citrate and subjected to high pressure and high temperature for 15 min with a Cuisinart CPC-600 6-qt pressure cooker. The slides were stained by hand with 300 μL of peroxidase blocker for 5 min; 300 μL of the primary antibody (either PCSK9 or CD31) in diluent for 45 min; 300 μL of PolyDetector Link for 15 min; 300 μL of PolyDetector HRP for 15 min; and 300 μL of Betazoid DAB for 5 min. The slides were counterstained with Mayer's hematoxylin. Finally, the slides were incubated in an increasing ethanol gradient for 1 min each, and xylenes for 15 s. After staining, the slides were mounted with PermaMounter and digitally scanned at 20 × magnification using a MoticEasyScan Pro (Motic).

Statistical analysis

Continuous variables are presented as mean ± SD or median (Q1, Q3) and compared using the Welch 2 sample t-test or the Wilcoxon rank-sum test, as appropriate. Categorical variables are presented as counts (%) and compared using the Fisher exact test or the chi-square test. The p-values reported are two-sided and a P-value < 0.05 indicated statistical significance. Generalized estimating equation analysis using log scaled values was performed to examine whether the study drug regimen (placebo or evolocumab) significantly influenced the changes between the baseline and 30-day PF4, P-selectin, and vWF values after controlling for the baseline values.¹⁵ Statistical analysis was performed using GraphPad Prism software version 9.4.1 (GraphPad Software, San Diego, CA, USA) and with SAS 9.4 (SAS Institute, Cary, NC, USA). The sample size/power estimation was within the framework of generalized estimating equation (GEE) model according to Ahn, C. et al.¹⁶ to test the hypothesis that the changes in PF4 outcomes are different between the evolocumab and the placebo group, we assume that (1) 50% of the recruited subjects are randomized to the evolocumab and 50% to the placebo group, (2) each subject is examined two times, at baseline and at 30 days, (3) the time-average difference of the results in the evolocumab subjects differs from that of the placebo subjects by more than 1.00 at a significance level of 0.05, and (4) the Wald test was 2-sided. With a sample size of 40 (20 in each group), the power of this study is 0.95.¹⁶

Results

Baseline characteristics

Baseline characteristics, medications at hospital admission, and laboratory results at hospital admission and 30 days following randomization are shown for the total cohort (N = 46) and by randomization to placebo (N = 23) or to evolocumab (N = 23) (Tables 1 and 2). Mean ± SD age was 60.1 ± 13.3 years, 22% were African American, and 48% were women. All participants were on guideline-directed medical treatment for ACS^1,2 and 96% received dual antiplatelet therapy. There were no statistically significant differences in demographic characteristics or past medical history between the two groups, except for age, which was older in the evolocumab group (placebo 56.3 ± 13.5 years and evolocumab 64.0 ± 12.1 years, p = 0.039). Infarct size was assessed by troponin I levels at baseline and were not different between the two groups. There were also no differences in the 30-day troponin values or in the change between the baseline and 30-day values in the two groups.

Table 1.

Baseline characteristics and medications of study participants at randomization

	Total Cohort (N = 46)	Placebo (N = 23)	Evolocumab (N = 23)	P value
Age (years), mean ± SD	60.1 ± 13.3	56.3 ± 13.5	64.0 ± 12.1	0.039
STEMI, n (%)	9 (20%)	3 (13%)	6 (26%)	0.46
NSTEMI, n (%)	37 (80%)	20 (87%)	17 (74%)	0.46
Women, n (%)	22 (48%)	13 (56%)	9 (39%)	0.38
African American, n (%)	10 (22%)	6 (26%)	4 (17%)	0.72
BMI (kg/m2), mean ± SD	28.8 ± 7.2	28.9 ± 7.0	28.6 ± 7.6	0.94
Diabetes mellitus, n (%)	14 (30%)	4 (17%)	10 (43%)	0.11
Insulin-treated, n (%)	6 (13%)	1 (2%)	5 (21%)	0.19
Hypertension, n (%)	36 (78%)	19 (83%)	17 (74%)	0.72
Current cigarette smoking, n (%)	13 (28%)	8 (35%)	5 (22%)	0.51
Previous myocardial infarction, n (%)	19 (41%)	6 (26%)	13 (56%)	0.071
PCI, n (%)	15 (33%)	4 (17%)	11 (48%)	0.057
CABG, n (%)	7 (15%)	3 (13%)	4 (17%)	1.00
Peripheral artery disease, n (%)	5 (11%)	4 (17%)	1 (2%)	0.35
Lipid-lowering treatment
Statin, n (%)	45 (98%)	23 (100%)	22 (96%)	1.00
Ezetimibe, n (%)	7 (15%)	5 (22%)	2(9%)	0.41
Antiplatelets
Dual antiplatelet therapy, n (%)	44 (96%)	21 (91%)	23 (100%)	0.49
Aspirin, n (%)	46 (100%)	23 (100%)	23 (100%)	1.00
Clopidogrel, n (%)	16 (35%)	10 (43%)	6 (26%)	0.35
Ticagrelor, n (%)	28 (61%)	11 (48%)	17 (74%)	0.13
Anticoagulants
Heparin, n (%)	42 (91%)	22 (96%)	20 (87%)	0.61
Warfarin, n (%)	1 (2%)	0 (0%)	1 (4%)	1.00

BMI, body mass index; CABG, coronary artery bypass graft surgery during admission; NSTEMI, non–ST-elevation myocardial infarction; PCI, percutaneous coronary intervention during admission; SD, standard deviation; STEMI, ST-elevation myocardial infarction.

The bold values indicate the baseline characteristics and medications with a statistically significant difference between the two groups, with a P value < 0.05.

Table 2.

Laboratory results at hospital admission and at 30 days

Hospital admission laboratory results	Total cohort (N = 46)	Placebo (N = 23)	Evolocumab (N = 23)	P value
Platelet count (platelets x103/μl blood), mean ± SD	219.5 ± 56.8	225.2 ± 57.2	213.8 ± 57.1	0.5
Hemoglobin (g/dl)	12.53 ± 1.84	12.64 ± 1.99	12.41 ± 1.73	0.68
PF4 (ng/1000 platelets), median (IQR)	8.8 [4.1,12.2]	9.2 [4.2,12.2]	8.0 [4.1,12.1]	0.84
vWF (ng/ml), median (IQR)	26.8 [17.8,36.1]	27.1 [17.9,36.0]	26.7 [17.7,36.4]	0.96
P-selectin (ng/1000 platelets)	0.18 [0.14,0.24]	0.18 [0.12,0.24]	0.18 [0.14,0.21]	0.67
Total cholesterol (mg/100 ml), mean ± SD	163.1 ± 47.0	173.1 ± 48.3	153.2 ± 44.5	0.15
Triglycerides (mg/100 ml), median (IQR)	103.5 [71.5160.5]	125.0 [73.0266.0]	100.0 [67.0153.0]	0.31
HDL-C (mg/100 ml), mean ± SD	48.5 ± 14.2	50.4 ± 16.7	46.7 ± 11.3	0.38
LDL-C (mg/100 ml), mean ± SD	92.8 ± 40.9	102.0 ± 43.5	89.0 ± 37.8	0.27
ApoB (mg/100 ml), mean ± SD	82.7 ± 27.6	87.0 ± 29.9	78.3 ± 24.9	0.29
eGFR (mL/min/1.73m2)	74.27 ± 26.11	81.65 ± 22.7	66.55 ± 27.67	0.05
	Total Cohort	Placebo	Evolocumab
30-Day laboratory results	(N = 46)	(N = 23)	(N = 23)	P value
Platelet count (platelets x103/μl blood), mean ± SD	225.8 ± 67.3	217.5 ± 74.1	234.2 ± 60.3	0.41
Hemoglobin (g/dl)	11.7 ± 2.2	12.17 ± 2.0	11.28 ± 2.3	0.2
PF4 (ng/1000 platelets), median (IQR)	11.3 [8.8,13.7]	13.1 [10.7,14.5]	10.7 [5.9,12.7]	0.036
vWF (ng/ml), median (IQR)	21.8 [15.1,37.5]	27.4 [17.8,38.5]	18.6 [13.8,72.0]	0.089
P-selectin (ng/1000 platelets)	0.18 [0.14,0.23]	0.18 [0.13,0.25]	0.18 [0.15,0.22]	0.76
Total cholesterol (mg/100 ml), mean ± SD	111.3 ± 31.5	121.3 ± 30.4	101.3 ± 30.0	0.029
Triglycerides (mg/100 ml), median (IQR)	87.5 [63.0136,3]	92.0 [79.0146.0]	82.0 [57.0131.0]	0.17
HDL-C (mg/100 ml), mean ± SD	43.7 ± 11.1	42.4 ± 11.4	45.1 ± 10.8	0.41
LDL-C (mg/100 ml), mean ± SD	48.4 ± 27.6	59.8 ± 24.9	37.1 ± 25.8	0.004
ApoB (mg/100 ml), mean ± SD	56.5 ± 22.1	67.4 ± 20.0	45.7 ± 16.6	<0.001
eGFR (mL/min/1.73m2)	67.59 ± 27.72	75.44 ± 23	61.43 ± 29.98	0.098

ApoB, apolipoprotein B; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; IQR, interquartile range; LDL-C, low-density lipoprotein cholesterol; PF4, platelet factor 4; SD, standard deviation; vWF, von Willebrand factor.

The bold values indicate the laboratory results with a statistically significant difference between the two groups, with a P value < 0.05.

In addition, specimens for immunostaining were obtained from 12 patients with coronary artery disease when they were undergoing CABG. Four of these patients were receiving evolocumab, three of whom were in the randomized study; and eight of the patients were not receiving evolocumab, four of whom were in the randomized study. Baseline characteristics are presented for these patients in Table 3. There were no statistically significant differences in demographic characteristics or past medical history between the two groups.

Table 3.

Baseline characteristics of participants undergoing coronary artery bypass surgery

	Total cohort (N = 12)	Evolocumab- (N = 8)	Evolocumab+ (N = 4)	P value
Age (years), mean ± SD	62.1 ± 5.6	59.5 ± 3.4	67.5 ± 5.4	0.052
Women, n (%)	0 (0%)	0 (0%)	0 (0%)	1.00
African American, n (%)	2 (17%)	1 (13%)	1 (25%)	1.00
BMI (kg/m2), mean ± SD	27.9 ± 3.7	27.7 ± 2.9	28.3 ± 5.4	0.85
Diabetes mellitus, n (%)	5 (42%)	4 (50%)	1 (25%)	0.58
Insulin-treated, n (%)	3 (25%)	2 (25%)	1 (25%)	1.00
Hypertension, n (%)	8 (67%)	4 (50%)	4 (100%)	0.21
Current cigarette smoking, n (%)	4 (33%)	2 (25%)	2 (50%)	0.55

BMI, body mass index; SD, Standard Deviation.

Effect of evolocumab on PF4 and P-selectin levels

PF4 and P-selectin levels are expressed as ng/10³ platelets since serum samples contain activated platelets and the number of α-granules, which store and release PF4 and P-selectin is related to the platelet count of each participant. Baseline PF4 concentrations did not differ between the two groups (placebo: 9.3 [4.2–12.2] and evolocumab 8.0 [4.1–12.1], p = 0.84). In the placebo group there was a significant increase in PF4 from baseline to 30 days (9.3 [4.2–12.2] to 13.0 [10.7–14.5] ng/10³ platelets, p = 0.002). In contrast, in the evolocumab group there was no significant change from baseline to 30 days (8.0 [4.1–12.1] to 10.7 [5.9, 12.7] ng/10³ platelets, p = 0.25, Figure 1A and B). Generalized estimating equation analysis revealed that the administration of evolocumab significantly influenced changes in PF4 levels from baseline to 30 days when compared to placebo (p = 0.019). P-selectin levels at baseline and at 30 days were not statistically different between the two groups, nor were the changes between the two groups significant.

Figure 1.

(A, B) Scatterplots depicting data points representing the paired baseline and 30 days PF4 values, presented as medians with interquartile ranges, for the placebo and the evolocumab groups, respectively. (C, D) Scatterplots depicting data points representing the paired baseline and 30 days vWF values, presented as medians with interquartile ranges, for the placebo and the evolocumab groups, respectively. PF4, Platelet Factor 4; vWF, von Willebrand Factor.

Effect of evolocumab on vWF levels

Baseline vWF concentrations were not different between the two groups (placebo: 27.1 [17.9, 36.0] and evolocumab 26.7 [17.7, 36.4], expressed as ng/ml, p = 0.96). In the placebo group there was no significant change from baseline to 30 days (27.1 [17.9, 36.0] to 27.4 [17.8, 38.5] ng/ml, p = 0.94). In contrast there was a significant decrease from baseline to 30 days in vWF levels in the evolocumab group (26.7 [17.7, 36.4] to 18.6 [13.8, 33.8] ng/ml, p = 0.042, Figure 1C and D). Generalized estimating equation analysis revealed that the administration of evolocumab significantly influenced changes in vWF levels from baseline to 30 days when compared to placebo (p = 0.021).

Colocalization of PCSK9, platelets, and ECs on the vascular endothelial surface

LIMA samples from patients undergoing CABG were immunostained to assess the presence of PCSK9, the platelet surface marker CD61 (integrin beta-3/platelet glycoprotein IIIa), and the endothelial cell (EC) transmembrane glycoprotein CD31 (platelet endothelial cell adhesion molecule (PECAM-1)). CD61 is one of the most widely used markers for the identification of platelets.¹⁷ Tabula Sapiens data at Cellxgene validate the strong and consistent expression of CD61 on platelets, while ECs only rarely demonstrate CD61 positivity.¹⁷ CD31 is considered the most sensitive and specific marker for EC differentiation.¹⁸ Immunohistochemical (IHC) analysis revealed colocalization of PCSK9 and CD61 on the vascular endothelial surface (Figure 2A and B). Immunofluorescence (IF) analysis confirmed that PCSK9 and CD61 were indeed colocalized with CD31 on the LIMA vascular endothelial surface (Figure 2C and D) showing a strong immunoreactive signal for both PCSK9/CD31 (Figure 2Ci, yellow arrowheads) and CD61/CD31 (Figure 2Di; white arrowheads) on the vascular endothelial surface.

Effect of evolocumab on the colocalization of PCSK9 and platelets on the vascular endothelial surface

As the findings from both IHC and IF indicated the presence of robust PCSK9 and CD61 signals (Figure 2C and D) on the LIMA vascular endothelial surface, we performed additional IHC/IF staining in specimens obtained from both groups (Figure 3A–D) to determine whether the extent of colocalization differed in those who were receiving evolocumab (Evolocumab+) and those who were not (Evolocumab-) at the time of cardiac surgery. There was significantly higher colocalization of PCSK9 and platelets (marked by CD61) in the specimens obtained from the patients not receiving evolocumab (Evolocumab-) than for those who were (Evolocumab+) (Evolocumab-: 47.2% ± 19.3% and Evolocumab+: 24.5% ± 11.4%, p = 0.030) (Figure 3E and F).

Effect of evolocumab on the colocalization of PCSK9 and ECs on the vascular endothelial surface

Since the initial IHC/IF analysis also revealed colocalization of PCSK9 and CD31 (EC marker) on the vascular endothelial surface, we sought to determine the extent to which PCSK9 colocalized with ECs and whether the administration of evolocumab impacted this interaction. We confirmed a strong CD31 signal on the vascular endothelial surface via IHC and IF, regardless of whether the specimens were collected from participants in the Evolocumab- or the Evolocumab + group (Figure 4A–D). Comparison of the % overlap of PCSK9 and CD31 between the two groups showed significantly higher colocalization of PCSK9 and ECs in the patients who were not receiving evolocumab (Evolocumab-: 39.6% ± 19.4%) than in the patients who were, Evolocumab+: 14.3% ± 4.5%, p = 0.007) (Figure 4E and F).

Effect of evolocumab on the colocalization of platelets and ECs on the vascular endothelial surface

Finally, we performed an additional analysis to identify the extent to which platelets colocalized with ECs on the LIMA vascular endothelial surface and whether the interaction differed in those who were and those who were not receiving evolocumab. When visualizing CD61 alone via IHC, LIMA specimens obtained from participants in the Evolocumab- group showed an intense CD61 signal, both at the EC surface as well as within the atherosclerotic plaque/thrombi (Figure 5A and B). Comparison of the % overlap of CD31 and CD61 revealed significantly higher colocalization of platelets and ECs in individuals who were not receiving evolocumab (Evolocumab-: 43.1% ± 18.7%) than in the patients who were (Evolocumab+: 20.1% ± 11.4%, p = 0.026). (Figure 5C–E).

Figure 5.

Evolocumab effect on platelets and ECs colocalization on the vascular endothelium. Representative LIMA segments from patients undergoing CABG immunostained for CD61 via immunohistochemistry (A, C), or for CD31, CD61, and DAPI via immunofluorescence (B, D). Black and white dashed boxes are enlarged images. Red (A, C) and white (B, D) arrowheads indicate positive immunostaining for CD61 with ECs. Red asterisks represent thrombi/blood clot (A). Black or white scale bars: small 250 µm, large 100 µm. (E) Scatterplot depicting data points representing CD31/CD61 %overlap in LIMA specimens from the Evolocumab- (blue bar) and Evolocumab+ (red bar) groups (*p = 0.026). ECs, endothelial cells; LIMA, left internal mammary artery; PCSK9, proprotein convertase subtilisin/kexin type 9. Figures 3, 4, and 5 were created with Biorender.com.

Discussion

In this study, we demonstrated that serum PF4 and vWF levels in ACS patients randomized to the PCSK9 inhibitor evolocumab were lower than in those randomized to placebo 30 days following study drug administration. Additionally, in a small number of patients undergoing CABG, we observed colocalization of PCSK9, the platelet marker CD61, and the EC marker CD31 on the vascular endothelium of LIMA segments obtained during surgery and that the extent of overlap between PCSK9 and CD61, PCSK9 and CD31, and CD61 and CD31 was lower in those who were receiving evolocumab than in those who were not.

Two large, randomized trials demonstrate that PCSK9 monoclonal antibodies decrease the risk of ischemic events in patients with known disease who were receiving statin therapies.^9,10 Two mechanisms responsible for the benefit impact of PCSK9 inhibition were reviewed by Navarese et al.¹⁹ The first, LDL-cholesterol reduction, is achieved by decreasing free PCSK9, which would otherwise bind to, and eventually lead to degradation of, LDL-cholesterol receptors. This would impact lipid-related progression of atherosclerotic disease and also decrease pro-inflammatory oxidized LDL.

The second proposed mechanism for the beneficial impact of PCSK9 inhibition is plaque stabilization, which is particularly relevant in ACS, a setting associated with platelet activation and endothelial inflammation. Proprotein Convertase Subtilisin/Kexin 9 is significantly correlated with platelet reactivity in patients without known CAD,²⁰ in those with stable CAD,²¹ in those with STEMI²² and in those with ACS following coronary intervention and who were on ticagrelor or prasugrel.²³ Furthermore, human recombinant PCSK9 increases platelet aggregation induced by low levels of epinephrine.²⁴ Several observational clinical studies indicate that PCSK9 inhibition decreases platelet activation. In a review by Peczek et al.,²⁵ the effect is related to reducing LDL-receptor mediated platelet activation. Proprotein Convertase Subtilisin/Kexin 9 inhibition reduced platelet activation in patients with familial hypercholesterolemia, lowered plasma levels of platelet activation factor C16, which directly triggers platelet aggregation and mediates neutrophil migration, and the production of reactive oxygen species and interleukin-6 in human macrophages.²⁶ In another study of hypercholesterolemia patients, 12-months of PCSK9 antibody therapy decreased platelet aggregation and increased sensitivity to aspirin and these findings were associated with decreased platelet membrane expression of CD62P and plasma levels of soluble CD40 ligand, PF4, and P-selectin.²⁷ In an in vitro study, activation of washed platelets by plasma from patients with heterozygous familial hypercholesterolemia was decreased following six months of PCSK9 inhibitor therapy.²⁸

Moreover, PF4, the most abundant protein released by the platelet α-granules following activation,²⁹ induces E-selectin expression by EC,³⁰ promoting leukocyte recruitment and adhesion to the endothelium. P-selectin is also released from platelet α-granules and the Weibel–Palade bodies of ECs,³¹ and plays a prominent role in the inflammatory response by recruiting leucocytes to areas of injury.³² Platelets are also a major source of PCSK9, which further enhances the pro-thrombotic and pro-inflammatory milieu.³³ Our findings that PF4 is significantly lower in those who received evolocumab than in those who received placebo supports this mechanism.

The endothelium is a critically important vascular structure. In addition to maintaining a balance between the production of vasodilator and vasoconstrictor substances, it produces numerous mediators that are associated with hemostasis, fibrinolysis, growth factor synthesis, and vascular permeability.^34,35 Inflammation induces endothelial injury and dysfunction, which is associated with reduced nitric oxide (NO) bioavailability, adhesion of monocytes to the luminal surface, loss of the permeability barrier, and vasoconstriction. We previously reported that PCSK9 levels were elevated and associated with MRI-assessed coronary endothelial dysfunction in people with HIV and those with dyslipidemia⁷ and that coronary endothelial function improved following the initiation of the PCSK9 antibody evolocumab.⁸ Endothelial injury is characterized by the release of vWF from endothelial Weibel–Palade bodies and platelet α-granules, and vWF is a standard marker of endothelial dysfunction and injury.^36–38 von Willebrand factor levels are elevated in ACS and are an independent predictor of adverse clinical outcomes in these patients.³⁹ Decreased vWF in those randomized to evolocumab support the concept that PCSK9 inhibition decreases endothelial injury in ACS patients.

The importance of platelet-EC interaction and platelet activation by PCSK9 was reviewed by Puteri et al.⁴⁰ Platelets do not attach to ECs or become activated under normal conditions. However, in ApoE^-/- mice platelet attachment to carotid artery endothelium was associated with inflammatory gene expression and preceded the development of atherosclerosis.⁴¹ Proprotein convertase subtilisin/kexin type 9 binding to platelet CD36 directly enhances platelet activation and increases platelet responsiveness to other agonists.⁶ Cytokines released by activated platelets can cause EC injury and result in plaque instability and thrombus formation. The association among PCSK9, platelets, and ECs on the LIMA vascular endothelial surface in patients with coronary disease and the decreased association of platelets and ECs in patients receiving a PCSK9 inhibitor further support the concept that PCSK9 mediates platelet and ECs interaction at the vascular endothelium.

Our findings are limited by the sample size, which is insufficient to assess clinical outcomes in the two groups. Samples for these participants were only consistently available at hospital admission and 30 days following the initial event. Thus, we cannot compare any early sequential changes in the platelet and endothelial markers in the two groups and do not know whether the differences would have persisted for longer than 30 days as only one dose of the study drug was administered. Furthermore, use of serum samples may introduce laboratory artifacts during serum preparation and may not be indicative of the in vivo effect of PCSK9 inhibition. Additionally, point-of-care in vivo testing was not conducted and although the ex vivo measures of platelet activation are an indication of the functional capacity of platelets,⁴² they do not necessarily reflect in vivo platelet activation.⁴³ There are additional markers of platelet activation e.g. thromboxane A2 that may provide additional information regarding the extent of platelet inhibition but which are not available. However, the direction and the magnitude of the effects were robust, the placebo-controlled randomized study of hospital administration of a PCSK9 antibody to ACS patient design is novel, as are the results in LIMA segments obtained from patients with coronary disease. These findings could potentially be used to inform the time of initiation and outcomes of larger studies evaluating the impact of PCSK9 inhibition on clinical outcomes in the high-risk early peri-infarction setting. Given the bleeding associated with current dual anti-platelet regimens following percutaneous intervention in ACS patients, PCSK9 inhibition might be an additional or alternative patient centered, ‘personalized’ approach in certain patient subsets, e.g. those with elevated PF4 or vWF levels. The translation of these findings to clinical outcomes requires large scale randomized studies of PCSK9 inhibition in the high-risk ACS patient population.

Conclusions

Administration of the PCSK9 inhibitor evolocumab to patients with ACS decreased markers of platelet activation and endothelial dysfunction in the early post-infarction period. IHC and IF analysis of LIMA segments obtained during CABG demonstrated PCSK9 colocalization with platelets and ECs on the arterial endothelial surface that may contribute to thrombus formation and plaque instability. Treatment with a PCSK9 inhibitor disrupted this interaction. Future studies exploring the extent to which the impact of PCSK9 inhibition on platelets and ECs alters the clinical course of ACS patients may further refine the identification of secondary prevention goals and appropriate pharmacologic interventions in the high-risk post-ACS patient population.

Funding

This work was supported by Amgen Inc. Amgen, the funder of the study, had no role in the design of the study, the collection, management, or interpretation of the data, or the statistical analysis. The funder reviewed the first submitted version of the article but was not involved in the writing or approval of the article or the decision to submit the article for publication.

Conflict of interest: T.M.L. has received research grants from the American Heart Association Career Development Award (19CDA34760040), National Institutes of Health (1R43HL166091), Merck, and Amgen. S.P.C. has received research grants from the National Institutes of Health (RO1HL148348), the American Heart Association Career Development Award (19CDA34760185), and is on the Scientific Advisory Board for Rejuvenate Bio. M.K.H. and E.M.W. have received research grants from National Institutes of Health (R01GM130564, R01HL137811). M.L.V. has received a research grant from the American Heart Association (2021AHAPRE216237). S.L. has received a research grant from the National Institutes of Health (U01DA040325).

Data availability

The data and analytical methods that support the findings of this study are available from the corresponding author on reasonable request.