During medical school clerkships and pediatric residency, trainees are constantly reminded that “children are not simply small adults.” As shown by Chiaretti and colleagues (page ), it seems that the converse of this statement—adults are not simply big children—is true as well, at least with respect to T-cell acute lymphocytic leukemia (T-ALL). Chiaretti et al use the Affymetrix HGU95aV2 chip to identify gene expression profiles that are associated with either response to therapy or long-term remission in 33 adult patients with T-ALL. The authors identified an expression profile that was associated with induction failure; this profile was characterized by overexpression of a single gene (IL8), and lower expression (compared with the successful induction group) of a larger set of genes, including GFI1, BCL-6, MXI1, and CD10. Long-term remission was characterized by higher levels of CD2, BUB1B, TTK, and CENPF, while relapse was associated with overexpression of AHNAK. Although the number of evaluable patients was small, a 3-gene model based on expression of CD2, TTK, and AHNAK was more effective at predicting relapse than were traditional criteria such as white blood cell count or immunophenotype. The predictive utility of this 3-gene model was then verified by real-time polymerase chain reaction (PCR) on an independent set of 18 additional patients.
The authors also studied the expression of a set of 7 genes previously found by Yeoh et al to be relatively overexpressed in T-ALL cells from pediatric patients with a high likelihood of relapse.1 They found that expression levels of these genes did not correlate with remission duration in their cohort of adult T-ALL patients, and that only 1 of the 3 genes that most effectively predicted long-term remission in adult T-ALL patients was predictive in the pediatric T-ALL series. An independent series of pediatric T-ALL patients2 reported that high levels of HOX11 were associated with long-term remission; in contrast, only 1 of 6 adult T-ALL patients in the series by Chiaretti et al maintained a long-term remission. The observation that genes that predict long-term remission in pediatric T-ALL are not predictive for adult T-ALL is perhaps not surprising, given the documented clinical and biologic differences between pediatric and adult T-ALL.3,4
There are several caveats that should be considered when interpreting the differences between the current adult study and the previously published pediatric series.1,2 The number of patients is relatively small, the patients came from different countries, the RNA preparation was subtly different, and the treatment regimens were different among the 3 series. Nonetheless, this type of study can be regarded as a hypothesis generator, and several lines of future investigation have been opened. Larger numbers of patients can be analyzed in parallel, using identical methods for RNA preparation and hybridization. If confirmed in larger studies, the 3-gene model could be used to identify patients at high risk of relapse, who might then benefit from more aggressive therapy. Moreover, the limited number of genes that has been identified can be studied in detail to determine whether their up- or down-regulation is causally related to the outcome of treatment.
