Source Of Differentially Expressed Genes In Peripheral Whole Blood Of Colon Cancer

In attempt to find early diagnostic mRNA biomarkers of cancer for noninvasive clinical application, researchers have extensively studied gene expression changes in peripheral whole blood of many cancer types. However, the diagnostic biomarkers were selected based on whole blood, which is heterogeneous population of blood cells, may be attributable to differences in the relative abundance of individual leucocyte populations. Additionally, no studies evaluated whether these biomarkers can differentiate cancer from inflammation, which is always a confounding factor for cancer diagnoses. Using the CellMix package provided by R and statistical hypothesis testing T-test, we evaluated the change of individual leucocyte populations between cancer whole blood(or inflammation whole blood)and normal whole blood. Colon cancer and sarcoidosis were taken as examples of cancer and inflammation respectively. It turned out that under cancer(or inflammation) condition the proportion of immune cells derived from myeloid progenitor increased while the proportion of immune cells derived from lymphocyte progenitor decreased compared with normal controls. The consistency of differentially expressed genes both detected between cancer whole blood and normal whole blood and between immune cells derived from myeloid progenitor and immune cells derived from lymphocyte progenitor was a high value of more than 90%, indicating that the expression changes between cancer whole blood and normal whole blood may be attributable to the expression changes between myeloid derived immune cells and lymphocyte derived cells. Same phenomenon was observed in inflammation whole blood compared to normal whole blood. From the obove, it could be concluded that the diagnostic biomarker of cancer extracted by using the whole blood reflected the population shift of individual leucocyte and cannot differentiate cancer from inflammation disease.By comparing the transcriptomes of cancer(or inflammation) circulating CD4+T cells, CD8+T cells with these two cell types’ of normal control, we found that IFN induced signaling pathway and pathways associated with IFN response were enriched by up-regulated genes which detected by comparing cancer CD4+T cells, CD8+T cells with the normal controls, while these down-regulated genes selected by comparing inflammation CD4+T cells, CD8+T cells with the normal controls were over-represented in these pathways. Moreover, ISGs(interferon simulated genes) showed opposite expression change in cancer and inflammation CD4+T cells, CD8+T cells compared with normal controls. The results above suggested that using pured immune cells, like CD4+T cells, CD8+T cells, can capture the changes occurred on the immune cells of cancer, which also have the potential to differentiate cancer from inflammation disease. To overcome the cutoff-dependency of existed diagnostic biomarker of cancer, we used the information of relative expression rank between genes to find gene pairs with stable relative expression rank in normal CD4+ T cells while reversed in cancer CD4+T cells and kept the stable relative rank in inflammation CD4+T cells as in normal. Eventually, we selected 7 gene pairs which can differentiate cancer from normal and inflammation disease theoretically.