Arterial access for sickle stroke predictors

Milbauer and colleagues use blood outgrowth endothelial cells and microarrays to probe lists of genes assigned to biological systems that may correlate with risk of ischemic stroke in pediatric sickle cell anemia patients and identify genes associated with inflammation as the major factor.

The endothelium plays an important role in sickle cell anemia (SCA) and in sickle cell stroke, but is inaccessible in individual patients. In this issue of Blood, Milbauer and colleagues compare gene expression in blood outgrowth endothelial cells (BOECs) isolated from pediatric sickle cell anemia patients with and without Circle of Willis (CoW) disease. Lin et al have developed a method for selectively culturing mature endothelial cells from BOECs¹  that can be distinguished from circulating endothelial progenitor cells by negative staining for AC133. The effect of culture on gene expression is a significant concern; however, the authors provide evidence that their culture technique results in minimal change in gene expression. Lists of candidate genes for 9 potential systems that may be predictive of stroke (adhesion, angiogenesis, apoptosis, coagulation, hypoxia response, inflammation, redox signaling, shear stress response, and vasoregulation) were proposed, and RNA was isolated from the cultured cells and probed versus all genes in each of the 9 lists. Expression of all genes in each system was included in the probability score given to each system; inflammation was found to be the only significant system. Adhesion and thrombosis had scores that were near to statistical significance. In non-SCA patients, inflammation is generally invoked as a determinant of postischemic severity, but the role of inflammation in atherosclerosis and other diseases predisposing to stroke in non-SCA patients is clear.

As the authors acknowledge, inclusion or exclusion of a system, such as inflammation or adhesion, may depend both on the genes chosen for the list and the number of patients studied. A larger study might have reached different conclusions (in this study, there were 9 SCA patients, 11 SCA control subjects, 22 black hemoglobin A controls, and 21 white controls). The risk of confounding effects in small samples of SCA patients is considerable because factors that may unequally affectthe systems under study—either those that are genetically determined (for example, α- and Sβ°-thalassemia,²  about 20% and 5% of all SCA patients, respectively), or other genetic or hematological factors³ —may be unevenly distributed in small groups unless they are explicitly accounted for in the study design. Furthermore, the patients themselves consisted of a heterogeneous group: those with elevated transcranial Doppler, or abnormal CoW magnetic resonance angiography, or clinical stroke in childhood.

The endothelium is increasingly recognized as a highly interactive organ in its own right, and studying BOECs may allow gene expression to be sampled in the endothelium of individual patients. However, vascular beds differ significantly,⁴  and gene expression in different vascular beds may also vary. Which vascular beds do BOECs sample? One answer might be that beds suffering from pathology would be expected to selectively shed the mature endothelial cells considered here; but, since sickle cell disease is noted for its multiorgan pathology, this would include cells from the venous microcirculation of various organs and tissues as well as the larger arterial vessels responsible for CoW disease.

The use of transcranial Doppler screening has greatly facilitated prospective diagnosis of pediatric patients at risk for CoW stroke⁵ ; however, less than 15% of nontransfused patients went on to develop stroke despite very high velocities,³  and additional diagnostic tests are much needed. This innovative study may shed light on the state of the vascular system in vivo, both in SCA and in other diseases characterized by vasculopathy, but can only be validated by a prospective study.