An Early Thymic Precursor Phenotype Predicts Outcome Exclusively in HOXA-Overexpressing Adult T-ALL: A GRAALL Study

Abstract

▪Introduction: Gene expression studies have consistently identified a HOXA positive (HOXAPos) subgroup of T-cell acute lymphoblastic leukemia (T-ALL) (Ferrando et al, Cancer Cell 2002, Soulier et al, Blood 2005, Homminga et al, Cancer Cell 2011). It is however unclear if HOXAPos T-ALL constitutes a distinct and homogeneous clinical entity, and the biological consequences of HOXA over-expression have not been systematically examined.Methods: We identified and characterized the biological characteristics and clinical outcome of 55 HOXAPos cases among a cohort of 209 adult T-ALL patients who were uniformly treated as part of the Group for Research on Adult Acute Lymphoblastic Leukemia (GRAALL)-2003 and -2005 studies.Results: HOXAPos patients had higher rates of an early thymic precursor (ETP)-like immunophenotype (38% v 13.9%, p = 0.0008), early bone marrow chemoresistance (59.3% v 40.8%, p = 0.026) and positive minimal residual disease (MRD, 51.5% v 23.5%, p = 0.01) than the HOXANeg group. These differences were due to a particularly high frequency of chemoresistant ETP-ALL among HOXAPos cases harboring leukemic fusion proteins that trans-activate the HOXA locus (e.g. PICALM-MLLT10, SET-NUP214).Strikingly, the presence of an ETP-like immunophenotype conferred marked differences in outcome within the HOXAPos group (5 year event-free survival (EFS) 25% for HOXAPos ETP v 52.2% for HOXAPos non-ETP, p = 0.02), which were mirrored by corresponding increases in cumulative incidence of relapse (CIR, 57.1% v 25%, p = 0.01, Figure 1). In contrast, these survival differences were not seen in the HOXANeg patients, where ETP and non-ETP cases had similar 5 year EFS (54.9% v 50%, p = 0.73) and CIR (34.5% v 41.2%, p = 0.57).Multivariate analysis revealed that early bone marrow chemosensitivity was the clinico-biological covariate that had the strongest prognostic interaction with HOXA status. HOXA positivity conferred significant decreases in both the EFS and CIR of chemoresistant patients (p = 0.053 and 0.039 respectively), that was independent of white blood cell count (WCC), stem cell transplant (SCT), ETP phenotype, EGIL classification, and our recently reported risk classifier that integrates the prognostic effects of mutations of NOTCH1, FBXW7, RAS and PTEN (Trinquand et al, J Clin Oncol 2013). There were corresponding marked survival differences within the HOXAPos cohort between chemoresistant and chemosensitive cases. These disparities were not seen in the HOXANeg group, indicating that the prognostic value of chemosensitivity in adult T-ALL is specific to HOXAPos patients.Discussion: Our data show that clinico-biological phenotype is intimately linked to the underlying mechanism of HOXA locus deregulation, and we identify HOXA overexpression as a novel prognostic variable in ETP-ALL. Multivariate analysis suggests that this poor outcome is strongly related to intrinsic treatment resistance, and that this effect is exclusive to the HOXAPos cohort. Patients in the GRAALL-2003 and -2005 studies received enhanced induction and/ or salvage therapy in the event of poor early treatment response. Our results suggest that pediatric regimen-based intensification provides significant survival benefits for HOXANeg chemoresistant cases. In contrast, these modifications are inadequate for therapeutic rescue of the majority of HOXAPos chemoresistant ETP-ALL. The dramatically inferior prognosis of this group mandates consideration for alternative treatments in future clinical trials. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.