Acute stroke, with ischemic stroke comprising 80% of all cerebrovascular incidents, has been recognized as one of the core problems in clinical medicine in need of prevention and treatment. It remains one of the most deleterious diseases that produce high social and economic costs worldwide. In 2016, of the 17.8 million deaths due to cardiovascular disease worldwide, 5.7 million were due to strokes1. In 2013, there were almost 25.7 million stroke survivors globally (71% with ischemic stroke) and 113 million disability adjusted life years due to stroke (58% due to ischemic stroke)2.
Introduction of acute reperfusion treatment: intravenous thrombolysis, and in the past 5 years—the mechanical thrombectomy, in combination with specific and general stroke unit procedures, have greatly improved functional outcome after ischemic stroke. Thrombolytic treatment with recombinant tissue plasminogen activator (rtPA) was the first breakthrough in the reperfusion, thus in some sense causative, stroke treatment. Intravenous rtPA is the mainstay and the highest class evidence based method of acute ischemic stroke treatment, and is currently recommended 0–4.5 h after stroke onset3. Moreover, some patients selected with dedicated neuroimaging and special clinical characteristics might be treated even beyond the standard time window4.
In most patients decision on i.v. rtPA administration is straightforward, however, in some cases it might be complex. The biggest concern is the symptomatic intracranial hemorrhage (sICH), which occurs in 3–7% of all treated patients, and is associated with worse 90-day functional outcome and higher disability than in those untreated3,5. There are various measures to quantify the extension of rtPA-associated ICH in clinical trials including the European Cooperative Acute Stroke Study (ECASS) scale (Table 1). Clinical measures used to quantify the neurological deterioration due to rtPA-related sICH include SITS-MOST (The Safe Implementation of Thrombolysis in Stroke-Monitoring Study) and ECASS III3. The former is defined as intracerebral hemorrhage classified as local or remote PH2 within 24–36 h after rtPA bolus administration with the clinically important deterioration of neurological status, whereas the latter is defined as any hemorrhage with clinical deterioration of 4 or more points on the National Institutes of Health Stroke Scale (NIHSS) score.
Meta-analysis of nine trials of intravenous rtPA administration versus control6 showed increased risk of PH2 [6.8% vs 1.3%, OR 5.55 (4.01–7.70)]. SITS-MOST registry also revealed higher frequency of PH2 [3.7% vs 0.6%, OR 6.67 (4.11–10.84)] and fatal intracerebral hemorrhage [2.7% vs 0.4%, OR 7.14 (3.98–12.79)].
Apart from the most common intraparenchymal hemorrhages anatomically related to ischemic lesion, in some patients treatment with i.v. rtPA is complicated by bleeding located remotely from the lesion, which may constitute a substantial portion of all ICHs and affect as many as 27.5% of patients7. However, there is little information on the prevalence of the remote-ICHs from large clinical trials on rtPA therapy for stroke.
Finding a method of the powerful (highly specific and selective) identification of patients at highest risk of sICH, in order to increase the percentage of stroke patients safely treated with rtPA, is one of the most important challenges in stroke research. The SITS-MOST identified nine independent risk factors for sICH: baseline NIHSS score, serum glucose, systolic blood pressure, history of hypertension, age, body weight, stroke onset to treatment time, aspirin monotherapy, and dual antiplatelet therapy with aspirin and clopidogrel. Unfortunately, any patterns or combinations of these and other characteristics into scoring systems (GRASPS, Glucose, Race, Age, Sex, Pressure, Stroke Severity; DRAGON (Dense Artery, Rankin Score, Age, Glucose, Onset to Treatment Time, National Institutes of Health Stroke Scale (NIHSS)), SEDAN (Sugar, Early Infarct Signs, Dense Artery, Age, NIHSS)) still perform poorly and have only modest predictive value for identifying patients at risk8.
To address this problem we designed a major and complex project to identify blood, neuroimaging, and clinical biomarkers combined for prospective assessment of the risk of intracranial hemorrhage (ICH) after thrombolytic treatment of acute ischemic stroke (Investigator Initiated Study funded from Siemens Healthineers, 2018, to B. Karaszewski, Medical University of Gdansk) with analysis of multiple-origin data with deep learning techniques (hypothesis-free approach). The study was designed to recruit 400 ischemic stroke patients treated with i.v. rtPA, but herein we present results of blood proteomic and peptide analysis of pilot 41 patients.
The proteomic approach has already been separately recognized as a valuable and comprehensive method enabling insights into the pathophysiology of stroke with proteomic profile assessed in the brain of stroke patients9,10,11,12, in the endothelial progenitor cells13, in platelets14 or in thrombi-emboli retrieved during the mechanical thrombectomy15. In 2015, a proteomics chip study on large Swedish PIVUS and ULSAM cohorts proved ten proteins to be related to the incident of stroke16. In another study on H-type hypertension related stroke, with use of iTRAQ-based LC–MS/MS proteomics approach, AT-3, CRP, ApoB, and AHSG were proved to be the strongest predictors of this type of stroke17. The same iTRAQ-based LC–MS approach was used in 50 stroke patients and 60 proteins showed a ≈1.5-fold change, with candidate proteins vWF, ADAMTS13, S100A7, and DLG4 confirmed through ELISA to corroborate with the experimental findings18. In the two phase SpecTRA study using liquid chromatography/multiple reaction monitoring-mass spectrometry, insulin-like growth factor-binding protein 3 and serum paraoxonase/lactonase 3 were found to be reliable and reproducible biomarkers for TIA in the Emergency Department settings19. In another study on TIA/minor stroke, ceruloplasmin, complement component C8 gamma (C8γ), and platelet basic protein were significantly different between the ischemic group (TIA and minor stroke) and the controls20 Finally, the SWATH method was used for analysis of serum of 20 ischemic stroke patients and 11 proteins were defined as candidate biomarkers21.
There are only two proteomic studies that concern rtPA treatment in ischemic stroke. In one of them plasma from acute stroke patients was analyzed pre- and post-intravenous tPA using tandem mass spectrometry and protein array profiling. The rtPA treated patients presented with distinct and elongated degradomic patterns in comparison to non-tPA treated patients22. In the second study, high-resolution mass spectrometry and long high-performance liquid chromatography were used to investigate changes in blood proteins after stroke and as a result of thrombolysis treatment. In this study ten patients were treated with rtPA and had up to 5 blood samples collected at different time points after stroke with 26 proteins being proved to be expressed differently and 23 proteins showing significant changes of expression over time23. However, up to date there have been no proteomic studies specifically confronting the possible biomarkers of hemorrhagic complications in stroke patients treated with rtPA, which make the release of the partial patient sample data reasonable. Due to its small size, and assumed biomolecular character of this paper, we do not here combine all individual data (clinical, neuroimaging, proteomic) into the scoring systems. Herein, we reveal our general methodological approach with shortlisting of blood peptide or protein candidates selected with Sequential Window Acquisition of All Theoretical Mass Spectra (SWATH-MS) that in the future might increase sensitivity and selectivity of the rtPA-associated sICH risk calculations.
There are obviously plenty other studies aiming at identifying serum or plasma prognostic biomarkers of rtPA-related hemorrhagic transformation in patients with acute ischemic stroke. However, in general they are based on far different methodological, technical and analytical approach to that described in this paper, and the selected biomarker—candidates have been characterized by relatively low sensitivity or selectivity thus so far not being applicable for clinical practice, and need further investigations24,25,26,27,28,29,30,31,32,33.
