• membrane microvesicles
• serum
• plasma
• coagulation
• lupus

We have read with great interest the article by Kato and colleagues regarding the ‘Apoptosis-derived membrane vesicles drive the cGAS–STING pathway and enhance type I IFN production in systemic lupus erythematosus’.1 However, we are concerned that the authors studied exclusively blood serum, that is, all blood samples from their patients with lupus and healthy controls were coagulated ex vivo before analysis.

At least three major changes can occur to the population of membrane microvesicles (MVs) during coagulation in a test tube. First, a subset of circulating MVs participate in, and become consumed by, clotting. For example, studies from our group and others indicated that over half of the total membrane MVs in blood plasma2 or the in vitro generated MVs3 4 display phosphatidylserine (PS) on their surface. PS-positive MVs are procoagulant relative to those that are PS negative2 3 5 because PS forms a catalytic membrane surface that promotes assembly and catalysis of coagulation factor complexes.3 6 It is especially true for apoptotic MVs,3 5 7 8 since membrane surface exposure of PS is a hallmark of cell apoptosis.9 Thus, PS-positive MVs could be preferentially involved in clot formation in the sampling tube when blood is drawn without anticoagulants. The remaining MVs in the serum tube may not be representative of the original population of MVs that were in the circulation.4

A second change during clotting ex vivo is the generation of new population of MVs that were not originally presented in the circulation. Platelets become activated during clot formation in a test tube and release artificially generated MVs in vitro during the sampling process.10 Platelet-derived MVs generated in the sample tube may account for 50% of the MVs in serum.10 Other cells in blood may also release MVs during the ex vivo clotting. These ‘artificial’ MVs in serum cannot represent the pathophysiological condition of the circulating blood in patients with lupus and healthy controls.

A third change during clotting is the possible cleavage of proteins on MVs by proteases, that is, thrombin, generated during the coagulation cascade. This problem has not been widely studied in MV research field, but it is known in other fields. For example, thrombin digests apolipoprotein-B into fragments,11 and the ‘intact’ lipoproteins are isolated from plasma, not serum.

In our opinion, blood plasma12 prepared in the presence of anticoagulants – should be used in MV research because it avoids consumption of the ‘original circulating’ MVs, and production of the ‘artificial’ MVs, as well as exposure to proteases, during clotting ex vivo in the serum sample tubes. But the study results with serum, as used by Kato et al and other groups, should be confirmed with blood plasma. Artefacts caused by the clotting of blood in a test tube may largely affect results and conclusions of any studies of circulating membrane MVs in clinical translational investigations. Therefore, blood sampling of the circulating membrane MVs in human subjects is an important point that needs to be clarified among the investigators who conduct clinical translational researches.