Early T-Cell Precursor-ALL in Adult T-ALL.

Recently, a small subgroup of pediatric acute T-lymphoblastic leukemia (T-ALL) was described, which is closely associated with the gene expression profile of early T-cell precursors (ETPs). This subtype, termed ETP-ALL, showed a highly unfavorable outcome compared to non-ETP(='typical')-ALL. Based on the results of Coustan-Smith et al. (Lancet Oncology, 2009), the Italian national study Associazione Italiana Ematologia Oncologia Pediatrica (AIEOP) and St-Jude Children's hospital modified their treatment in children with ETP-ALL to a more intensive regime including stem cell transplantation. ETP-ALL is characterized by a specific immunophenotype (CD1a^-, CD8^-, CD5^weak with expression of stem cell or myeloid markers). Here we explored the existence of ETP-ALL in adults and further studied the molecular characteristics of this specific T-ALL subtype.

We examined the gene expression profiles of 86 adult T-ALL patients obtained from the Microarray Innovations in LEukemia (MILE) multicenter study (HG-U133 Plus 2.0, Affymetrix, Haferlach et al., JCO in press). In addition, bone marrow of 296 patients from the German Acute Lymphoblastic Leukemia Multicenter Study Group (GMALL) were analyzed by flow cytometry and expression levels of BAALC, IGFBP7, MN1, and WT1 were determined by real-time-PCR.

Using the published list of differentially expressed genes in ETPs (Coustan-Smith et al. 2009) we performed unsupervised clustering analyses of the 86 T-ALL samples. A cluster of 17 samples (19.8%) displayed an ETP-associated gene expression profile and were defined as ETP-ALL. Comparing the gene expression profiles of ETP-ALL and typical T-ALL, 2065 probe sets were differentially expressed in ETP-ALL (FDR 0.05). In addition to genes used for classification, we also identified genes known to be involved in the pathogenesis of T-ALL (e.g. PROM1, BCL2, LMO2, LYL1). In particular, stem cell associated genes such as, BAALC (2.52-fold, p=0.003), IGFBP7 (2.76-fold, p=0.002) or MN1 (3.41-fold, p<0.001) were upregulated in ETP-ALL, whereas HOX11 (45-fold, p=0.004), a marker for thymic T-ALL, was downregulated. An independent cohort of 297 patient samples from the GMALL study group was examined by flow cytometry and real-time PCR. 19 (6.4%) samples revealed the ETP-ALL immunophenotype. As expected, all patient samples were found in the group of early T-ALL, representing 23.5% of all early T-ALLs. There was a significant correlation between a lower leukocyte count at first diagnosis and the classification of ETP-ALL (p=0.001). Gene expression measured by real-time-PCR was performed for genes associated with poor outcome in T-ALL: BAALC (2.11-fold, p<0.001) and IGFBP7 (3.59-fold, p=0.003) were significantly upregulated in the group of ETP-ALL. Similarly, the genes MN1 (4.52-fold, p<0.001) and WT1 (2.76-fold, p=0.036), described as poor prognostic markers in cytogenetically normal AML, were also upregulated in ETP-ALL.

In adult T-ALL, a subset of patients shares the gene expression profil and immunophenotype of ETP-ALL, which is in line with recent findings in pediatric patients. The gene expression profile of this subset is significantly correlated to stem cell associated markers predictive for inferior outcome in T-ALL. Interestingly, adverse factors in CN-AML are also aberrantly expressed in ETP-ALL suggesting a myeloid origin of ETPs and indicating a closer relationship between ETP-ALL and AML. The prognostic impact and the determination of the most appropiate set of markers needs to be further investigated. These results support the GMALL strategy to regard early T-ALL patients as high risk with assignment to stem cell transplantation.

Haferlach:MLL Munich Leukemia Laboratory: Equity Ownership.