The quest for circulating endothelial cell standardization: The peril of platelets

In the January 2007 issue of Cytometry Part B: Clinical Cytometry, Strijbos et al. (1) present their observations that critically addresses the issue of defining circulating endothelial cells (CECs) by current methods of FACS and its reliability. Some poignant issues merit further discussion.

Strijbos et al. (1) focuses on validating the popular single platform FACS assay for CECs, as per the approach of Mancuso et al. (2), using their previously published definition [forward scatter (FSC)^{low to intermediate}, side scatter (SSC)^low, CD45⁻/CD31⁺⁺/CD146⁺]. Using methods as diverse as electron microscopy, reverse transcriptase polymerase chain reaction, and fluorescence in situ hybridization, they demonstrated that “CECs” defined thus are in fact, just large platelets. Comparing the protocols from Strijbos et al. (1) and Mancuso et al. (2), both methods are virtually identical, warranting the critical conclusions by Strijbos et al. (1). This in part will certainly explain the observed disparity between previously published numbers of CECs in healthy controls (and disease groups), which differ significantly between investigators, sometimes by a factor of several thousand, which is a disconcerting fact if the same cells are being measured.

Our own work (and others) on CEC assay standardization (3, 4) proves that FACS analysis of CECs is comparable to other proven methods such as immunomagnetic cell separation (with comparable CEC numbers in healthy controls and patients), when set up with strict CEC definition criteria in mind. Firstly, FSC^high events should be included (most reports state CEC ≥ 10 μM; the cells reported by Strijbos et al. (1) are typically ≤6 μM), bearing in mind that typical endothelial cells such as human umbilical vein endothelial cells (HUVEC) are ≥20 μM; secondly, blocking steps must be employed to reduce nonspecific staining; and thirdly, CD45^dim cells should not be included as these include endothelial progenitor cells (3). From our work, we also confirm the findings of Strijbos et al. (1), that the titration of antibodies as well as the type used is paramount (referring to anti-CD146 Ab), with nonspecific binding and background fluorescence resulting from inappropriately large staining concentrations.

The use of HUVECs as an archetypal CEC is well-known but needs more consideration. As mentioned, the typical size of HUVECs is ≥20 μM, and they can differ from CECs phenotypically and morphologically (5). Also, Strijbos et al. (1) should have assessed the recovery of HUVECs to gauge the sensitivity/specificity of their assay, as performed by Mancuso et al. (2), who report >90% specificity. The authors also report no CD34 expression on HUVEC, which is not accurate and does not represent the majority of published work. Strijbos et al. (1) also point out a fact largely ignored by many: CD31 expression on lymphocytes is largely bright, and the small subgroup of CD146⁺/CD31⁺ lymphocytes might well complicate the issue, and if not correctly gated out they will be mistaken for CECs. Besides staining the majority of the leukocyte population, CD31 is also expressed significantly on platelets and haematopoetic progenitors (6) and may actually be more of an impediment than a strength when used in CEC analysis. If used, investigators must include a platelet discriminator, e.g. CD42, CD41, (1) to rule out these “large platelets,” CD133 (a progenitor marker) and ideally, a nuclear stain like DAPI. As demonstrated by Strijbos et al. (1), vWf is also present within both platelets and endothelial cells, thus cannot be used to discriminate them apart.

Overall, Strijbos et al. (1) elegantly provide an explanation for the significantly differing CD146⁺ CEC counts reported thus far, making the point that any methodological assessment of these rare cells needs considerable validation and confirmation. Further interest in these “large platelets” (assuming these are what they are) would require greater elucidation of their primary role in health and disease.

Patrick K. Y. Goon*, Timothy Watson*, Paul S. Stonelake*, Gregory Y. H. Lip*, * Haemostasis, Thrombosis and Vascular Biology Unit, University Department of Medicine City Hospital, Birmingham B18 7QH England, United Kingdom.