Abstract 1763

Poster Board I-789


The myelodysplastic syndromes (MDS) are a group of aging-associated hematopoietic disorders characterized by ineffective maturation that, in 6-33% of cases, progress to acute leukemia (AML), a disease of blocked differentiation. Since microRNAs (miRNAs) regulate cell differentiation/maturation as well as cell identity, miRNAs may play critical roles in both the development of MDS as well as transformation to AML. With the advent of new therapies and treatment modalities for MDS and AML, the availability of biomarkers capable of detecting early MDS and predicting progression to AML would have tremendous impact on the management of these patients.


The study utilizes twenty samples of bone marrow mononuclear cells, ten from MDS patients and ten from normal controls, isolated and stored at the University of Pennsylvania Stem Cell Core Facility between 2003 and 2007. The MDS specimens included seven (7) from patients in whom it was known that no transformation to acute leukemia occurred within five (5) years of the available specimen and three (3) from patients in whom acute leukemia was diagnosed within two (2) years of the available specimen. The patient specimens were selected to represent early stage MDS and therefore not at high risk for transformation based on morphologic analysis (circulating blasts '5%) and the 2001 WHO classification system. Total RNA was obtained from each sample and arrayed on a custom Agilent miRNA microarray at the University of Pennsylvania Microarray Facility in a dual channel experiment in which each sample was arrayed against a pool of all twenty samples. Statistical analysis was performed using Genespring v 7.31 and Partek Genomics Suite v6.3. Subsequently, thirteen miRNAs of interest and one amplification control were examined by RT-PCR using ABI commercial primers for those targets. Total RNA extracted from formalin-fixed paraffin-embedded tissue was also examined for the relative expression of these fourteen miRNAs.


Class discovery ANCOVA algorithms were applied to the data to identify 13 miRNAs which differentiate the MDS samples from the normal controls. Real time PCR was performed upon the sample RNAs to verify the microarray results. Of the thirteen candidate miRNAs, miR-150, miR-342, and miR-103, miRNAs documented in other studies to be important in hematopoiesis, demonstrate particular promise as classifiers for MDS based upon the high degree of correlation between the microarray results and the RT-PCR studies as well as their ability to independently differentiate MDS patients from normal controls. When the three miRNAs are taken together, the microarray and RT-PCR results correlate in 83% of cases, with only 2 major discrepancies. Furthermore, comparison between the total RNA isolated from the bone marrow mononuclear cells suspensions of two patients with total RNA isolated from the corresponding formalin-fixed paraffin-embedded clot sections, across fourteen different miRNAs demonstrate extremely high levels of correlation (R2 = 0.94 and 0.85), proving the feasibility of future studies utilizing the readily available paraffin-embedded clot sections as a source of patient samples. Microarray probes to predicted miRNAs were also identified which distinguished the three patients who progressed to acute myeloid leukemia within 24 months from the seven that did not.


A miRNA signature was identified which distinguishes early cases of MDS from normal controls. These miRNAs are predicted to regulate many of those mRNAs identified by traditional transcriptional profiling. Ultimately, it is hoped that assessment of miRNAs or their downstream targets will provide practicing pathologists with new markers for the identification of MDS and provide clinicians with additional prognostic information to guide the use of therapeutic interventions.


Carroll:Sanofi Aventis Corp: Research Funding; Cephalon Oncoloy: Consultancy.

Author notes


Asterisk with author names denotes non-ASH members.