Translational Immunodiagnostics in Stroke (TrImS)

Overview

In adult patients presenting to emergency departments within 24 hours of symptom onset with suspected acute stroke, we aim: 1. to identify early brain- and pathology-specific circulating, whole blood, plasma and serum panorOmic biomarkers that enable early acute stroke detection, diagnosis, dynamics, differentiation, monitoring, prediction and prognosis. 2. to identify early brain- and pathology-specific, panorOmic biomarkers in saliva that enable early acute stroke detection, diagnosis, dynamics, differentiation, monitoring, prediction and prognosis. 3. to derive biomarker platforms of models for early acute stroke detection, diagnosis, dynamics, differentiation, monitoring, prediction and prognosis 4. to validate these models in independent and external datasets

Introduction Stroke is a leading cause of overall mortality, morbidity and disability worldwide. In 2010 it was the second leading cause of death whilst in 2019 it ranked third overall for disability adjusted life years (DALYs) and second for those aged 50 years and over. Acute ischaemic stroke (AIS) is a leading cause of mortality and morbidity in the USA affecting over 800 000 adults each year. In Hong Kong it is the fourth leading cause of death. It is frequently preceded by a transient ischaemic attack (TIA) which is a harbinger for future cerebral ischaemic events with a 20% risk of stroke within 90 days. Stroke is associated with ischaemic, inflammatory, haemorrhagic and atherosclerotic processes linked to disruption of the blood brain barrier and increased blood borne proteins and nucleic acids. It is estimated that 'the average duration of non-lacunar stroke evolution is 10 hours (range 6 to 18 hours), and the average number of neurones in the human forebrain is 22 billion. In patients experiencing a typical large vessel AIS, 120 million neurones, 830 billion synapses and 714 km (447 miles) of myelinated fibres are lost each hour. In each minute, 1.9 million neurones, 14 billion synapses, and 12 km (7.5 miles) of myelinated fibres are destroyed. Compared with the normal rate of neurone loss in brain aging, the ischaemic brain ages 3.6 years each hour without treatment'. There is also evidence of such injury in the circulation within minutes to hours of major trauma including head injury. The timely diagnosis of stroke and its aetiological classification into stroke-types is important as early appropriate intervention results in vastly improved outcomes. Firstly, determining the cause of the stroke affects which treatment is immediately prescribed. For example, early thrombolysis is indicated for AIS but contraindicated for haemorrhagic stroke (HS). A diagnosis of AIS within 4.5 hours of stroke onset followed by timely thrombolytic intervention improves stroke outcome. Anticoagulant and antiplatelet therapy is indicated for cardioembolic stroke or TIA but contraindicated for atherogenic strokes. In the case of the latter, antiplatelets are recommended. Large vessel occlusion is better treated by surgery than conservatively. Delays in diagnosis and early intervention count as avoidable DALYs, morbidity, healthcare expenditure and early mortality. In the developed world, stroke is suspected and detected by clinical history and examination and confirmed diagnostically by cerebrovascular imaging either by CT and/or MRI. Whilst neuroimaging presents the current best standard for stroke identification and classification it is far from optimal. Neuroimaging provides only gross estimates of neurovascular damage, fails to identify the cause (cryptogenic stroke) in a significant 5 - 15% cases, and provides little information on cellular and molecular pathophysiology. This in turn delays early treatment and limits the search for potential agents for stroke prevention and recovery. Further, the absence of an objective method for determining stroke onset, for example in patients who wake with a suspected stroke, results in 80% of patients with AIS not receiving thrombolysis. Therefore, there is a need to understand the dynamic pathophysiology of stroke at an early stage. Molecular biomarkers such as conventional and high-sensitivity troponin have transformed the detection and risk-stratification of patients with acute coronary syndrome. However, no such markers exist for the early detection, diagnosis, classification and risk-stratification of stroke or TIA. Over the last several decades, investigators have searched for stroke markers which might be useful for screening, detection, diagnosis, classification, monitoring, prediction and prognosis of stroke. Many candidate markers have been identified but none have been sufficiently accurate or early enough in stroke development to find a place in clinical care. The great challenge is identifying early, accessible biomarkers with a high level of accuracy for detecting, diagnosing and differentiating stroke aetiology. Recent technological discoveries suggest that many blood-based markers may be detected both qualitatively and quantitatively in breath and saliva using immunoassay and spectral profiling. Such technologies yield accurate results within minutes to seconds. The challenge is not downstream in the translational pipeline but the upstream discoveries of -omics markers. Once discovered, the technologies for developing rapid assays and point-of-care tests are already in place. Brain-specific biomarkers in Stroke The search for meaningful biomarkers in stroke has been long. It includes genomic, transcriptomic, proteomic, metabolic and lipidomic approaches. For example, recent studies suggest that whole blood transcriptomics may enable accurate AIS differentiation. The Ischemia Care Biomarkers of Acute Stroke Etiology (BASE) investigators are currently evaluating RNA gene expression in peripheral blood in stroke patients presenting within 30 hours of stroke onset (NCT02014896). BASE is a multi-centre, prospective study with an estimated enrolment of 1100 adult patients and 100 age, gender and co-morbidity matched controls who present to the Emergency Department (ED) or hospital with suspected AIS or TIA. The results of these studies are awaited but at present there is no early, accurate biomarkers platform to guide stroke detection and diagnostics. There has been extensive study on brain-specific proteomic biomarkers of glial cells \[e.g. S100 calcium-binding protein B (S100B), glial fibrillary acidic protein (GFAP)\] and neuronal cells \[e.g. ubiquitin C-terminal hydrolase-L1 (UCH-L1), neuron-specific enolase (NSE), alpha II-spectrin breakdown products (e.g. SBDP120, SBDP145, and SBDP150), myelin basic protein (MBP), neurofilament light chain (NF-L), tau protein, visinin-like protein-1 (VLP 1), NR2 peptide injury in cerebrospinal fluid (CSF) and peripheral blood in the search for timely diagnostic information for stroke \[28\]. Yet none have found their way into clinical practice. Non-specific biomarkers in Stroke As previously mentioned, stroke is associated with generic, non-specific, ischaemic, inflammatory, haemorrhagic and atherosclerotic processes that are linked to disruption of the blood brain barrier and increased blood borne proteins and nucleic acids. A detailed review of non-specific but pathology-relevant stroke biomarkers is beyond the scope of this introduction. However, an Omics approach should include a mention of lipidomics. Lipidomics research in stroke has included conventional and advanced lipid approaches. For example, sphingolipids, phospholipids (including lyso- and ether- species), cholesteryl esters, and glycerolipids have been shown to predict future cardiovascular events and cardiovascular death but not with any high degree of accuracy. The addition of seven lipid species to a base model of 14 traditional risk factors and medications improved the prediction of cardiovascular events with a C statistic from 0.680 to 0.700 (P\<0.0001). The addition of four lipid species into the base model improved the prediction of cardiovascular death with a C statistic of 0.740 to 0.760 (P\<0.0001). Whilst statistically impressive the improvement in real-world prediction appears marginal and lacks impact. This approach has not led to a diagnostic platform for assessing early stroke. The unmet clinical need Patients commonly present to EDs, general practices, the community, prehospital and hospital wards with suspected acute stroke where delays to diagnosis result in delays in treatment and inefficient healthcare processes. Clinical acumen is variable and subjective. Accurate, objective, early, safe, minimally-invasive, diagnostic and prognostic tests are needed to inform on stroke. EDs are one of the commonest settings for acute presentations of these illnesses. Unanswered Questions Many questions are currently unanswered. Can brain- and pathology-specific small molecular proteins and nucleic acids be detected quantitatively in the circulation (e.g. whole blood, plasma) and saliva within minutes to hours of the onset of acute stroke and TIA? Do such detections fit a temporal, dynamic and mechanistic role in the ischaemic, inflammatory, immunological, haemorrhagic, apoptotic, atherosclerotic, recovery and regenerative processes implicated in stroke and TIA? Do perturbations (elevations or reductions) in such markers have a role in the detection, diagnosis, stroke-classification, prediction and prognosis in patients with stroke? Do combinations of brain-specific biomarkers and non-specific pathological markers (e.g. atherosclerotic, inflammatory) improve stroke detection, diagnosis and differentiation? Could the discovery of such markers guide clinical pathways and lead ultimately to novel vaccines and therapeutic interventions in such patients? Questions 1. In adult patients presenting to EDs with suspected acute stroke, what combinations of early brain-specific and pathology-specific panorOmics liquid biopsy biomarkers optimise stroke detection, diagnosis, dynamics and differentiation? 2. In adult patients presenting to EDs with suspected acute stroke what early panorOmics platforms of liquid biopsy biomarkers are useful for predicting responsiveness to treatment, monitoring and prognosis in differentiate stroke types? * Stroke is suspected if patients are FAST or LAPSS or ROSIER positive. Suspected stroke includes stroke mimics, TIA, AIS and HS. * Liquid biopsy includes whole blood, plasma, serum, white cell pellet and red cell effluent, and saliva * Liquid biopsy biomarkers include genomic, transcriptomic, proteomic, metabolomic, lipidomic and haematological contents. * Stroke types include HS and AIS. Hypothesis We hypothesise that stroke involves a rapid, complex process of ischaemia, necrosis, apoptosis and brain barrier disruption which releases tissue-specific genomic, epigenomic, transcriptomic and proteomic markers into the blood stream early in the course of cerebrovascular pathophysiology. These biomarkers will be detected early in easily accessible fluids such as whole blood, plasma, serum and saliva. The dynamic changes will be useful for personalised disease detection, diagnosis, risk-stratification, disease and therapeutic monitoring, prediction and prognosis. The dynamic changes may also enable a reasonable assessment of stroke onset and permit early intervention in patients with otherwise unknown stroke onset. Objectives In adult patients presenting to EDs within 24 hours of symptom onset with suspected acute stroke, we aim: 1. In discovery research - to identify early brain- and pathology-specific circulating, whole blood, plasma and serum panorOmic biomarkers that enable early acute stroke detection, diagnosis, dynamics, differentiation, monitoring, prediction and prognosis. In discovery research - to identify early brain- and pathology-specific, panorOmic biomarkers in saliva that enable early acute stroke detection, diagnosis, dynamics, differentiation, monitoring, prediction and prognosis. 2. In developmental research using training sets - to derive biomarker platforms of models for early acute stroke detection, diagnosis, dynamics, differentiation, monitoring, prediction and prognosis 3. In validation research using test sets - to validate these models in independent and external datasets