Specific biomarkers for acute stroke: comparison between human and animal models and development of a new way to improve acute stroke conditions

Abstract

Stroke is the third most common cause of death in most Western countries and the major cause of disability. The distinction between stroke subtypes and knowledge of the time of stroke onset is critical in clinical practice. The most specific and biologically powerful treatment for acute ischaemic stroke is thrombolysis with recombinant tissue plasminogen activator (rt-PA) given within the first 4.5 hours of ischaemic stroke onset but this therapy is disappointingly underused. This mainly because of unknown symptoms onset time and uncertainty about stroke diagnostic. Neuroimaging can help decide who and how to treat. Nevertheless, neuroimaging is expensive, has contra-indications and is not always readily available. Cheap and easily measured blood biomarkers serve similar roles for other diseases.

The traditional approach to stroke biomarker discovery has been to select candidate markers based on their known involvement in the stroke pathophysiology. This “pick your best candidate” approach inevitably means selection from only a small pool of what is theoretically available. High throughput technologies such as whole genome microarrays and proteomics permit unbiased selection of molecular markers by examining all of genes and proteins expressed in a tissue. In this thesis, these agnostic approaches were used to identify characteristic blood RNA and protein expression profiles occurring after stroke.

In the first chapters, we show that most acute stroke patients do not have the indicated imaging done in an appropriate time window, therefore blood biomarkers would have a useful niche. We also show that, to date, most candidate blood biomarkers have not found use in the clinic because they have been selected from experiments that do not address the rapidly changing nature of stroke and have been identified by an ad hoc process that only scratches the surface of the potential candidates available.

A pilot gene array experiment performed in rats to specifically look for hyper-acute changes that might be clinically informative in a way not easily realizable in humans, identified gene expression changes of great amplitude which changed markedly with time. This indicated that it should be possible to generate a stroke clock for use in the clinic. A larger follow up confirmation experiment designed also to examine the influence of experimental comorbidities demonstrated that many of the changes could be attributed to the surgical intervention needed to induce stroke in rats. This has important implications for our understanding of stroke inflammatory biology and also suggests the need for a different approach to stroke biomarker discovery.

Examination of protein expression in bloods from a clinical trial of hypothermia identified new candidate biomarkers. A clinical trial to examine blood RNA expression, ultimately performed as a pilot because of recruitment difficulties, was designed to examine temporal change. It confirmed that collection of sequential blood samples in a clinically relevant time frame is possible and identified time dependent gene expression with overlap with the rat data. This indicates that generation of a clinically useful stroke clock should indeed be possible.