In AML, the majority of patients <60 years of age will enter remission but at least 50% will subsequently relapse, therefore the monitoring of minimal residual disease (MRD) during treatment has become an important issue. Current molecular markers for MRD are mainly limited to the RT-PCR detection of the fusion genes resulting from recurrent translocations which paradoxically are mostly limited to favourable risk groups who are least likely to relapse leaving the majority of the patients, including those with a normal karyotype, without a molecular marker suitable for monitoring.

Two hundred and twenty patients have been assessed by gene expression profiling using Affymetrix U133A chips and the data analysed with the aim of identifying novel MRD markers for patients who do not currently have a suitable marker. As an initial “proof of principle”, we have identified possible MRD markers for patients with either t(15;17), t(8;21) or inv(16) and correlated with changes in expression of these markers with clinical changes as measured by established molecular MRD markers (PML-RARα or WT1).

Of the expression profile from 22,283 probe sets in 29 cases of t(15;17), 20 genes were identified which had at least a two fold over expression which was unique to the t(15;17) subgroup. Of these several of the probe sets were related to the same gene, but from the reduced gene list 2 (HGF and ILGF binding protein) were selected for quantitation by quantitative PCR. Similarly the expression profile identified 20 genes which were unique to the 15 cases of t(8;21), and 20 genes which were unique to the 19 cases of inv(16). These included ETO and MYH11 representing the respective 3′ end of the respective fusion transcript. Three other genes (PRAME, POU4F1, and IL5RA) were selected for the t(8;21) cases and ST18, CLIP-170 and MNI for the inv(16) cases. When the relative quantitative expression of each of these “unique” genes was correlated with the expression of the established markers of minimal residual disease (PML-RARα or WTI) there was good correlation. These data suggest that gene expression profiling can identify ‘unique’ genes which can be used to develop specific markers for minimal residual disease monitoring for a larger proportion of cases of AML than is currently available.

Author notes

Corresponding author