Søren Bache
A conversation with Søren Bache, MD, from the Neurointensive Care Unit, Department of Neuroanaesthesiology and Centre for Genomic Medicine, Rigshospitalet, University of Copenhagen, Denmark, about microRNA changes after subarachnoid hemorrhage.
Interviewed by José G. Merino, MD, Associate Professor of Neurology, University of Maryland School of Medicine.
They will be discussing the paper, “MicroRNA Changes in Cerebrospinal Fluid After Subarachnoid Hemorrhage,” published in the September 2017 issue of Stroke.
Dr. Merino: Thank you for agreeing to the interview. First, I would like you to explain some things about delayed cerebral ischemia (DCI) after subarachnoid hemorrhage (SAH) for our readers: How common is it? How soon after SAH does it develop? How does it affect outcome after SAH?
Dr. Bache: The reported prevalence of DCI after SAH varies, but newer randomized clinical trials have found a risk of 21–38% in patients who survive the initial bleeding and aneurism-securing surgery. The variation in calculated risk may be due to discrepancies both in case definition (i.e. the numerator) and in the definition of which patients are entered into the denominator. Today, most researchers base their case definition of DCI on the criteria suggested by Vergouwen et al. (Vergouwen MD, et al. Stroke. 2010). Before this consensus work, the definition varied even more, and many used their own criteria for DCI, delayed ischemic neurological deficits (DIND) or cerebral vasospasm. However, not all patients are conscious enough to be assessed clinically for a deterioration in consciousness, and such patients may be either included or excluded in the total number of patients; hence, the variation in the denominator. Based on Vergouwen’s criteria, in our center, we found a prevalence of 23% in 450 patients admitted from 2009–12 with SAH (unpublished data). These patients all receive prophylactic nimodipine, which lowers the risk of DCI; therefore, one should expect publications from the pre-nimodipine era to report a higher prevalence of DCI (Dorhout Mees SM, et al. Cochrane Database of Systematic Reviews. 2007).
Delayed cerebral ischemia occurs a median of 6–7 days after hemorrhage, but this varies, with a typical reported range from 3 to 14 days. DCI may be reversible, but in some cases it progresses to permanent brain injury, thereby affecting outcome.
Dr. Merino: Vasospasm has been implicated as a common etiology of DCI after SAH, but you argue in your paper that other mechanisms may be at play. What is the evidence for and against the hypothesis that vasospasm is the cause of DCI?
Dr. Bache: There is substantial evidence that cerebral vasospasm is associated with clinical deterioration after SAH, but no solid evidence of any causal relation. The theory of vasospasm as a causal factor of secondary brain injury probably emerged 40–50 years ago parallel to our ability to visualize the caliber of cerebral arteries and measure cerebral blood flow. Understandably, the observations of these vessel abnormalities gave hope for a therapeutic potential. However, since then it has become clear that SAH patients can have angiographic narrowing of vessels with no clinical deterioration, as well as clinical deterioration with no angiographic narrowing. In fact, approximately 70% of SAH patients have some degree of angiographically evident vasospasm, compared to a prevalence of DCI of approximately 30%. Furthermore, the CONSCIOUS studies demonstrated a reduction in cerebral vasospasm after administration of clazosentan with no improvement in neurological outcome. Also, nimodipine, originally administered to reduce vasospasm, seems to exert its positive effects through another mechanism.
This calls for scientific reporting of these complications as separate entities until there is a better understanding of the pathophysiological mechanisms behind both, as well as their mutual interrelation.
Dr. Merino: You postulate that other cellular processes may be responsible for DCI and look at the expression of microRNAs (miRNAs) in the CSF to try to identify some of these processes. Can you please briefly explain what are miRNAs, their function, and production and distribution (intra- and extra-cellular)? How does the study of miRNAs in CSF help us understand the mechanisms underlying DCI? Are miRNAs in CSF representative of miRNAs that are expressed in the cell?
Dr. Bache: MicroRNA (miRNA) are a large group of small (around 22 nucleotides long), non-coding RNA (as opposed to messenger RNA (mRNA)), each with specific sequences. Each specific miRNA targets a specific mRNA through homologue base pairing. The specific pairing of a miRNA and an mRNA inhibits protein synthesis from that mRNA. As cell function is largely controlled by protein synthesis, miRNA play regulatory roles in all cells. Some miRNAs are tissue-specific and stable in biofluids, which adds further to their potential as biomarkers.
Many experimental models have linked several miRNAs to neuronal apoptosis and neuroinflammation in various types of acute brain injury. However, very few articles have reported clinical data on cerebral miRNAs. In our study, we hypothesized that miRNA changes occur in relation to clinically relevant processes after SAH. Ultimately, we hope to find specific miRNAs with active roles in the processes leading to brain injury from SAH, as miRNAs are modifiable targets and could provide novel therapeutic targets.
Dr. Merino: Can you briefly describe what you did for this study? (A brief explanation for non-experts)
Dr. Bache: We screened the 754 most common miRNAs in CSF from SAH and control patients using real-time quantification PCR (RT-qPCR). The results were validated using a different RT-qPCR technique including extensive quality control. The need for extensive quality control became apparent early in the process, as measuring miRNA in biofluids poses several challenges compared to tissue and mRNA.
Dr. Merino: What did you find?
Dr. Bache: We found marked changes in the miRNA profile after SAH, with 66 miRNAs changing by a median of 34-fold (range of fold changes from 3.3 to 970, and median P = 10-8). Furthermore, we identified miRNAs which could be related to the occurrence of delayed cerebral ischemia and poor functional outcome after SAH.
Dr. Merino: What cellular pathways do the up- and downregulated by the miRNAs that are different in the SAH patients?
Dr. Bache: Our study design does not provide this information, and it will be a large challenge to demonstrate this clinically. Current technology requires samples of brain tissue to study the up- and downregulation of these pathways. Most potential designs including the sampling of brain tissue would face substantial ethical challenges.
Contrary to clinical studies, experimental studies have the potential to provide this information. However, in our opinion, we need further clinical studies before experimental studies can elaborate on the role of miRNAs for clinically relevant mechanisms.
Dr. Merino: Is there any added value at looking at intracellular miRNAs beyond that of intrathecal miRNAs?
Dr. Bache: Yes. The field of extracellular miRNA is very young, and one of the most important questions is, “How does intracellular miRNAs relate to extracellular miRNAs?” Appropriate experimental models should attempt to establish this relationship. Also, studies of intrathecal or systemic administration of mimics or inhibitors of miRNAs should measure intracellular miRNA levels in various types of brain tissue.
Dr. Merino: Taking your results into account, what are the next steps to identify the mechanisms responsible for DCI after SAH?
Dr. Bache: The present study was conducted in a single center reporting measurements of miRNA in CSF only, which were obtained on day 5 after SAH. Daily measurements of selected miRNA in CSF from the time of hemorrhage until the occurrence of DCI would provide a dynamic profile, which will aid us in determining the role played by each miRNA for the development of DCI; in addition, it could potentially indicate the time when measuring a given miRNA provides the most information. Similar measurements of selected miRNA in blood could potentially expand the usefulness to patients without external ventricular drainage.
Also, reproduction of these data in an independent cohort would strengthen our ability to select the right miRNAs for further study.
For selected miRNAs, hypotheses could be built from their regulatory roles as described in the literature or based on in silico target prediction. An experimental design could then be set up focusing mainly on one or a few miRNAs, including measurements of their target mRNA and proteins, as well as in situ hybridization for a deeper understanding of these processes.
Dr. Merino: What other approaches may complement the search for miRNAs in CSF to further our understanding of DCI?
Dr. Bache: With continued advances in technology, we are moving further away from macroscopic evaluation of cerebral artery calibers and cerebral blood flow towards a deeper understanding of the complex cellular processes leading to brain injury. Experimental models have recently shown the activation of various pathways leading to neuronal apoptosis after different types of acute brain injury. One of the most promising theories is the influx of cellular calcium resulting in uncoupling of mitochondrial phosphorylation and oxygen demand. In light of the positive effects and uncertain mode of action of the calcium antagonist, nimodipine, on delayed cerebral ischemia, these mechanisms should be studied further to broaden our understanding.
On a more general note, our current studies are my first encounter with “omics” technology, where a large number of variables are tested in one sample/patient. The number of studies reporting data produced with the use of omics-methods is increasing rapidly. These approaches pose new challenges, especially when it comes to preprocessing and statistical analysis of data, but also with regard to validation of the reported findings.
Dr. Merino: Thank you very much.
Dr. Bache: I would welcome any comment, discussion or need for elaboration you or your readers might have.
