The Effects of Globin on Microarray-based Gene Expression Analysis of Mouse Blood

Peripheral whole blood-based gene expression profiling has become one of the most common strategies exploited in the development of clinically relevant biomarkers. However, the ability to identify biologically meaningful conclusions from gene expression patterns in whole blood is highly problematic. First, it is difficult to know whether or not expression patterns in whole blood capture those in primary tissues. Second, if explicit steps are not taken to accommodate the extremely elevated expression levels of globin in blood then large-scale multi-probe microarray-based studies can be severely compromised. Many studies consider the use of mouse blood as a model for human blood in addition to considering blood gene expression levels as a general surrogate for gene expression levels in other tissues. We explored the effects of globin reduction on peripheral mouse blood in the detection of genes known to be expressed in human tissues. Globin reduction resulted in 1.) a significant increase in the number of probes detected (5840 ± 944 vs 12411 ± 1904); 2.) increased expression for 4128 probe sets compared to non-globin reduced blood (p < .001, ≥ two-fold); 3.) improved detection of genes associated with many biological pathways and diseases; and 4.) an increased ability to detect genes expressed in 27 human tissues (p < 10-4). This study suggests that although microarray-based mouse blood gene expression studies that do not consider the effects of globin are severely compromised, globin-reduced mouse whole blood gene expression studies can be used to capture the expression profiles of genes known to contribute to various human diseases. Whole and GLOBINclear-treated paired blood samples from 18 mice representing 3 mice from 6 different isogenic mice strains.