Prediction of Response to Treatment with Lenalidomide in Patients with Non-Del 5q Myelodysplastic Syndrome by Gene Expression Profiling Remains Difficult.

Abstract 2781

Poster Board II-757

Lenalidomide belongs to a proprietary class of immunmodulatory drugs showing therapeutic activity in patients with myelodysplastic syndrome (MDS), in particular in those having the 5q-abnormality, but also in patients not showing this cytogenetical aberration. In 2008, Ebert et al. (PLos Med. 2, e35) could demonstrate that there is a specific gene expression profile in bone marrow cells collected from MDS-patients either with 5q- syndrome as well as MDS-patients having no 5q-abnormality which is strongly correlated with the clinical response to treatment with lenalidomide. Whereas this finding is not of clinical importance in patients with MDS del 5q (overall response 75 %) it may play a pivotal role for prediction of clinical response to lenalidomide in non-del 5q MDS-patients. Therefore, we have studied gene expression profile (HG-U133plus2.0, Affymetrix, Santa Clara, CA) of routinely isolated low-density mononuclear bone marrow cells from 8 patients with IPSS low/int-1 risk MDS having no deletion on chromosome 5 but were subsequently treated with lenalidomide 5 mg/day. All of the patients were transfusion dependent for red blood cells. The median duration of treatment with lenalidomide was 22 weeks. RNA was extracted by Trizol and quality controlled by using a Bioanalyzer 2100 system (Agilent, Waldborn, Germany) to exclude RNA degradation. Microarray hybridization was performed according to the standard Affymetrix protocol. Data were analyzed by Microarray Analysis Suites 5.0 (MAS 5.0; Affymetrix) and GeneSpring (Agilent Technologies, Santa Clara, CA). For clustering analysis we utilized the gene list of 68 discriminating genes as published by Ebert et al. the molecular analysis did clearly separate two groups of patients having specific gene expression profiles according to the responder/non-responder group as published previously. Furthermore, single sample prediction could discriminate three out of 8 patients to be possible responders to lenalidomide but this was not correlated to the clinical course of those patients while on treatment with lenalidomide. However, none of the MDS-patients receiving lenalidomide did show significant clinical response as defined by reduction of transfusion requirement by 50 % or transfusion independence. In conclusion, prediction of response to lenalidomide in non-del 5q patients by gene expression profiling so far remains critical. Prospective analysis of molecular changes including DNA analysis in larger clinical trials using lenalidomide in non-del 5q MDS-patients are required to establish reliable predictive markers in MDS.