lncRNA Signature for Predicting Cerebral Vasospasm in Patients with SAH: Implications for Precision Neurosurgery

Abstract

Subarachnoid hemorrhage (SAH) patients' surgery is performed to prevent extravasation of blood into the subarachnoid space. Cerebral vasospasm (CVS; narrowing of cerebral arteries) occurs following SAH and represents a major cause of associated mortality and morbidity. To improve postsurgery care of SAH patients and their prognosis, the ability to predict CVS onset is critical. We report a long noncoding RNA (lncRNA) signature to distinguish SAH patients with CVS from SAH patients without CVS. Cerebrospinal fluid (CSF) was obtained from SAH patients without CVS (n = 10) and SAH patients with CVS (n = 10). lncRNAs ZFAS1 and MALAT1 were significantly upregulated (p < 0.05), whereas lncRNAs LINC00261 and LINC01619 were significantly downregulated in SAH patients with CVS (p < 0.05) compared to SAH patients without CVS. We applied this lncRNA signature to retrospectively predict CVS in SAH patients (n = 38 for SAH patients without CVS, and n = 27 for SAH patients with CVS). The 4-lncRNA signature was found to be predictive in >40% of samples and the 2-lncRNA comprising MALAT1 and LINC01619 accurately predicted CVS in ∼90% cases. These results are initial steps toward personalized management of SAH patients in clinics and provide novel CSF biomarkers that can substantially improve the clinical management of SAH patients.

Keywords: LINC01619; MALAT1; cerebral vasospasm; long non-coding RNA; subarachnoid hemorrhage.

Graphical abstract

Figure 1

lncRNAs ZFAS1 and MALAT1 in SAH Patients Expression levels of lncRNAs ZFAS1 and MALAT1 in pooled samples representing patients with SAH (without CVS) and those with SAH and CVS. Values plotted are mean ± 95% confidence interval (CI).

Figure 2

lncRNAs LINC00261 and LINC01619 in SAH Patients Expression levels of lncRNAs LINC00261 and LINC01619 in pooled samples representing patients with SAH (without CVS) and those with SAH and CVS. Values plotted are mean ± 95% CI.

Figure 3

Expression of ZFAS1 and MALAT1 in Individual Patient Samples Expression levels of lncRNAs ZFAS1 and MALAT1 in individual samples representing patients with SAH (without CVS) and those with SAH and CVS. Values plotted are mean ± 95% CI.

Figure 4

Expression of LINC00261 and LINC01619 in Individual Patient Samples Expression levels of lncRNAs LINC00261 and LINC01619 in individual samples representing patients with SAH (without CVS) and those with SAH and CVS. Values plotted are mean ± 95% CI.

Figure 5

Validation of lncRNA Signature in Archived Samples Expression levels of lncRNAs ZFAS1, MALAT1, LINC00261, and LINC01619 in individual archived samples are shown. 38 samples were from patients without CVS, whereas 27 samples represented patients with SAH and CVS. Values plotted are mean ± 95% CI. ns, nonsignificant.

Figure 6

Predictive Power of lncRNA Signatures Predictive power of a 4-lncRNA signature versus 2-lncRNA signature in predicting occurrence of CVS in patients with SAH. 4-lncRNA signature consisted of lncRNAs ZFAS, MALAT1, LINC00261, and LINC01619, whereas the 2-lncRNA signature only considered MALAT1 and LINC01619.