Evaluating the relevance of the platelet transcriptome

A recent article in Blood by Gnatenko et al¹  globally profiled mRNA expression in platelets. Using microarray analysis, the authors reported that approximately 2000 transcripts (13%-17% of probed genes) are present in unstimulated platelets isolated from healthy human subjects and concluded that evaluating the platelet transcriptome will be useful for identifying proteins that regulate normal and pathologic platelet and megakaryocyte functions. We have also found similar transcript profiles in platelets with different microarray strategies.^2,3 

An accompanying commentary voiced similar conclusions, but also suggested that the most frequent platelet mRNAs detected by microarray are well-known leukocyte or red cell messages, heightening suspicion that their abundance may be due to contamination by these classes.⁴  These 3 messages were thymosin β₄, neurogranin, and β-globin.¹  Thymosin β₄ and neurogranin protein are also found in platelets,^1,5  however, which is consistent with observations that platelets express gene products that are also present in other cell lineages.^2,3,6,7  Moreover, Gnatenko et al¹  took rigorous measures to account for contributions by leukocytes and red blood cells, and provided persuasive evidence that the thymosin β₄ and neurogranin transcripts are derived from platelets. The significance of detecting β-globin mRNA in their preparations still needs to be resolved, and if platelet-derived, will likely require confirmation by in situ detection methods. However, Gnatenko and colleagues indirectly addressed this issue by isolating total RNA from whole blood and concluded that the globin transcripts observed in their platelet preparations were not supported by erythrocyte contaminant estimates.¹ 

Based on serial analysis of gene expression (SAGE) results, which preferentially targets abundant mRNAs, the commentary also underscored the author's conclusions that the vast majority of messages in platelets are mitochondrially derived.^1,4  We do not argue this point. Indeed, mitochondrial RNAs are continuously transcribed, in contrast to other platelet transcripts, and generate multiple polyadenylated transcripts from individual genes accounting for their enhanced detection by SAGE.¹  An important point, however, is that the mitochondrial genome encodes only 13 mRNAs and 2 rRNAs.¹  Platelets contain over 2000 individual mRNA species,^1-3  including well-known messages for the α_IIb and β₃ integrin subunits,^8,9  that were not detected by SAGE.¹  Thus, although they are more abundant, mitochondrially derived transcripts represent a minute fraction (< 0.01%) of the mRNA species pool present in human platelets.^1-3  There are numerous examples in which the identification of nonmitochondrial mRNAs in platelets has generated important physiologic insights regarding the characterization and functional significance of corresponding proteins.^1-3,6,7,9,10 

Although the total RNA in human platelets is at least 100-fold less than in leukocytes,¹⁰  platelet mRNAs are diverse, polyadenylated, distributed in a fashion that is influenced by cytoskeletal and RNA-binding proteins, and differentially translated in response to outside-in signals.^1-3,6-8,10  In addition to mitochondrial transcripts, many others are basally present.^1-3,6-10  Characterization of the platelet transcriptome, and the proteins that it encodes, will undoubtedly increase our understanding of platelet and megakaryocyte physiology and behavior in health and disease.