Abstract
Abstract 1634
Poster Board I-660
Background
MicroRNAs (miRNAs) are small non-coding RNAs that regulate gene expression by base- pairing with the target mRNAs and inducing degradation or protein translation inhibition. MiRNAs expression is de-regulated in AML, but the corresponding functional miRNA-controlled pathways remain unknown. The availability of data from gene and miRNA expression profiling may allow generation of functional links from the whole transcriptome and miRNome that are involved in myeloid leukemogenesis. Therefore, here we proposed to attain distinct miRNA-associated gene expression signatures by correlating miRNA expression levels detected by miRNA microarrays with mRNA expression levels assessed by Affymetrix microarrays from newly diagnosed AML patients.
Methods
miRNA and mRNA gene expression levels were assessed by Affymetrix HGU133 microchip and OSUCCC miRNA microchip version 3, respectively, in diagnostic samples from 48 newly diagnosed AML patients. Patient characteristics include; median age 59 years (range 24-81) with cytogenetically normal (n=32), core binding factor (n=3), t11q23 (n=8), t(15;17) (n=4) and del9q (n=1). The miRNA/mRNA correlations were performed using quantitative trait analysis (Spearman correlation test) within the BRB tools. We derived gene expression signatures associated with the following miRNAs; miR-10a,-10b,-155,-181,-191,-23a,-26a,-126,-17,-20,-25,-145,-146,-21,-29a,-29b,-196,-15a and -16-1, which we found to be deregulated in AML (Garzon et al, Blood 2008). To asses functional characteristics, we determined which biologic terms included in the Gene Ontology (GO) project were enriched in the distinct miRNA-associated signatures, at a level of significance of P<0.01.
Results
In table 1, we report the numbers of probes that correlate positively and negatively with the selected miRNAs at the p-value<0.01 (Spearman), and the GO terms that are over-represented in each miRNA-associated signature at the p-value<0.01. Since some of the Affymetrix probes detect miRNA expression (i.e. miR-155 probe) we used them as quality controls. The miR-155 Affymetrix probe expression value was indeed positively correlated with the miR-155 expression value detected by miRNA microarrays (r=0.64, p-value<0.001). Our analysis identified a strong positive correlation of HOX related genes with miR-10 and miR-20a. Furthermore, we observed a negative correlation between miR-181a and –b, -155 and -146 expression with that of genes involved in immunity and inflammatory response (e.g. IRF7 and TLR4), the miR-29b, miR-23a and miR-26a expression with that of antiapoptotic genes (e.g.,MCL-1) and the miR-145 expression with that of pro-apoptotic genes (e.g., Bim and PTEN). These correlations were confirmed by GO analyses, which evidenced the enrichment of members of the homeobox, immunity and inflammation and apoptosis biologic process, respectively. Furthermore, we observed correlation of miR-17,-191,-196a, 29b, -145 and -16 with gene encoding for members of the GO term chromatin modification.
Conclusions
Our results indicate that by correlating data from two different platforms that allow assessment of genome-wide gene and microRNA expression profiles, putative functional miRNA-mRNA interactions were identified in AML samples. These interactions appear to take place within pathways controlling hematopoiesis, innate immunity, apoptosis and chromatin remodeling. By integrating the transcriptome and miRNome in AML cells is possible to derive previously unidentified functional subset of AML that can be treated with specific targeted therapeutic approaches.
Disclosures
No relevant conflicts of interest to declare.
Protein structure–based gene expression signatures