A platelet transcriptomic signature of thromboinflammation predicts cardiovascular risk

Figures

• 

  Figure 1. Flow chart of study participants. Percentiles of monocyte platelet aggregate (MPA) levels were calculated based on MPA values of n = 149 participants at 2 unique time points (TP). Participants with consistently low MPA levels (defined as ≤ 40th percentile) at 2 time points were defined as MPAlo and participants with consistently high MPA levels (defined as ≥ 60th percentile) as MPAhi. Asterisk indicates that RNA-Seq data were available for a total of n = 100 participants with MPA values available at TP1 and in n = 102 of all study participants.

• 

  Figure 2. MPA gating strategy. (A) Using flow cytometric analysis of fixed whole blood, monocytes were identified as CD14+ cells within the CD45+ cell population. CD61+ cells within the CD14+ monocyte population were defined as monocyte-platelet aggregates (MPA). (B) shows MPA levels from patients with available MPA levels at both time points (TP). High MPA is defined as ≥ 14.1% and low MPA as ≤ 11.7%.

• 

  Figure 3. Platelet and monocyte activation markers in MPAhi versus MPAlo groups. Flow cytometry of whole blood was performed to analyze platelet and monocyte activation markers. (A–D) Expression of platelet activation markers PAC-1, P-selectin, CD40L, and CD40 were compared between MPAhi (n = 32) and MPAlo (n = 28) groups in the resting state and in the presence of the agonists epinephrine (epi), thrombin (thr), adenosine diphosphate (ADP), and arachidonic acid (AA). (E) Spearman’s correlation of MPA levels, platelet activation markers in the resting state measured by flow cytometry, and aggregation in response to different agonists measured by light transmission aggregometry (LTA) is shown. (F) Monocyte activation markers CD11b, CD86, and CD40 were measured using flow cytometry. Graphs show median and interquartile range, Man-Whitney U test was applied. *P < 0.05, **P < 0.01, ***P < 0.001.

• 

  Figure 4. Platelet RNA transcriptomics in MPAhi versus MPAlo groups. Platelets were collected from healthy individuals; RNA was isolated and RNA-Seq performed. (A) Volcano plot showing differential gene expression (DGE), assessed using the DESeq method, in MPAhi compared with MPAlo groups. P values adjusted for age, sex, race, and ethnicity. (B) Gene set enrichment analysis showing differentially expressed pathways between MPAhi and MPAlo groups. NES, normalized enrichment score. (C) Based on differentially expressed genes between MPAhi and MPAlo groups with an adjusted P < 0.1 a transcriptomic signature, TIPS, was developed and calculated for each individual using the singscore method. (D) Box plot showing TIPS in MPAhi versus MPAlo individuals in the derivation cohort. Boxes show the interquartile range (IQR) with the median line, whiskers extend to 1.5 × IQR, and dots indicate outliers beyond that range. A t test was applied.

• 

  Figure 5. TIPS validation. TIPS was applied to cohorts with diseases associated with increased monocyte platelet aggregate (MPA) levels. (A and B) TIPS was calculated in hospitalized patients with COVID (n = 8) compared with matched healthy controls (n = 10) using a t test, and in 85 women undergoing coronary angiogram with myocardial infarction (MI) (n = 44) compared with women referred for coronary angiogram without MI (n = 41). Linear regression was applied and adjusted for age, race, and ethnicity as indicated. Boxes show the interquartile range (IQR) with the median line; whiskers extend to 1.5 × IQR; and points beyond that range are shown as outliers. (C) Among 129 patients who underwent lower extremity revascularization (LER) followed for 18 months, TIPS was calculated from platelet RNA samples collected prior to LER. Kaplan-Meyer curves of the occurrence of future cardiovascular events (composite endpoint of death, stroke, myocardial infarction, and amputation) are displayed stratified by TIPS tertiles; log-rank test was applied after adjustment for age, sex, race, and ethnicity.

• 

  Figure 6. Correlation of TIPS and PRESS. TIPS and PRESS were calculated for 129 patients who underwent lower extremity revascularization (LER) followed for 18 months, from platelet RNA samples collected prior to LER. (A) TIPS and PRESS were correlated using Pearson’s correlation (r = 0.21, P = 0.015). (B) PRESS was compared between TIPS tertiles (Kruskal-Wallis). Boxes show the interquartile range (IQR) with the median line; whiskers extend to 1.5 × IQR. (C and D) Patients were stratified based on PRESS score prediction of high versus low platelet reactivity and high versus low TIPS score resulting in 4 groups: PRESS low and TIPS low (n = 51), PRESS high and TIPS low (n = 14), PRESS low and TIPS high (n = 36), and PRESS high and TIPS high (n = 28). (C) Kaplan-Meyer curve of risk of cardiovascular events compared between groups using log-rank test after adjustment for age, sex, race, and ethnicity. (D) Adjusted hazard ratios for cardiovascular event risk with low PRESS/low TIPS group as a reference.

• 

  Figure 7. Effect of antiplatelet therapy on TIPS. (A–C) TIPS was calculated in a cohort of healthy patients that had blood collected for platelet RNA isolation at baseline (> 4 weeks without antiplatelet medication) and after 4 weeks of treatment with 81 mg aspirin (n = 61) (A), 325 mg aspirin (n = 62) (B), and 180 mg ticagrelor (n = 50) (C). Boxes show the interquartile range (IQR) with the median line; whiskers extend to 1.5 × IQR. Paired t test was applied.