**7. Effects of white blood cells on whole blood gene expression profiles**

Whole blood contains a variety of cell types as red blood cells, granulocytes, lymphocytes, and platelets. Most of the nucleated cells in blood are white blood cells such as neutrophils, T-cells, B-cells, and monocytes. The number of white blood cells in humans is known to decrease steadily from infancy to adulthood, and its composition (i.e. lymphocytes, granulocytes) also changes with age [30]. In study of the gene expression profiles change related to aging, hematological data of the fetal stage was unavailable because the amount of collected blood was insufficient for the analysis. From 12 to 30 weeks of age, ANOVA analysis indicated no significant differences in the fractions of lymphocytes, neutrophils, eosinophils, basophils, and monocytes. In addition, these compositions were almost equal to those in human adults. The above result suggests that the gene expression profile change of age-related whole blood RNA is not due to the composition of white blood cell subpopulations.

The intraclass correlation between Staphylococcus enterotoxin B-stimulated and unstimulated blood from healthy subjects was significantly higher in leukocyte-derived samples the in whole blood, suggesting that the method of RNA isolation from whole blood can be a critical step in blood RNA assay [31]. Although PBMCs do not contain neutrophils, eosinophils, basophils, nor platelets, Min et al. reported highly correlated results (r2 = 0.85) for 8,273 genes expressed between the whole blood RNA, by using the PAX gene Blood RNA system, and peripheral blood mononuclear cell (PBMC) RNA samples isolated from healthy volunteers by using a Ficoll-Paque gradient and TRI Reagent (SIGMA) [32]. Other workers conducted a large scale genome-wide expression analysis of white blood cells subpopulations. This study indicates that correlation coefficients for T-cells and monocytes among different healthy subjects were 0.98±0.01 and 0.97±0.01, respectively. However, for the same subjects (n=5), correlation coefficients between T-cells and monocytes was 0.88±0.01, indicating varied correlation between white blood cells subpopulations. In addition, gene expression analysis were showed a varying dependence on the isolation method such as PAXgene, Buffy coat, and lysis. The correlation coefficients between isolation methods were 0.89±0.04, 0.91±0.04, 0.96±0.06, for PAXgene vs. lysis, PAXgene vs. Buffy coat, and Buffy coat vs. lysis, respectively [33]. In order to ensure the reliability for to clinical use of whole blood RNA diagnosis, the development of standard method and measurement standards needs to be sought.

The Gene Ontology (GO) Database was used to categorize gene expression profiles functionally to conduct the effects of white blood cells on whole blood gene expression profiles in our study of hyperlipidemia. As a result, the GO term, related to white blood cell function (GO: 0006954, 0007166), had a high correlation coefficient. In contrast, GO terms related to the repair of damaged organs, including translation (GO: 0006412), positive regulation of growth rate (GO: 0040010), and growth (GO: 004007), showed low correlation coefficients. We, therefore, conclude that the difference in the gene expression profiles between the whole blood and white blood cells are not only caused by differences in experimental protocols, but also by differences in RNA origin [34].
