Identification of new subgroups and prognostic markers in pediatric B cell precursor acute lymphoblastic leukemia by gene expression profiling

Treatment of pediatric acute lymphoblastic leukemia (ALL) is increasingly successful, achieving cure rates of over 80%. Early identification of patients with high risk for relapse has led to improved outcome, however, two third of patients encountering relapse were initially stratified into low-intermediate risk groups. The identification of better upfront prognostic factors remains an important challenge in childhood ALL. In this thesis, gene expression profiling (GEP) was applied to different research approaches aiming to dissect subgroups and to find novel therapeutic targets in B cell precursor ALL (BCP ALL). Among BCP ALL, the patients lacking major genomic aberrations (B-others) represent the subgroup that is most in need of in depth investigations in order to indentify new prognostic factors and improve of risk stratification. To advance biological knowledge of B-others we performed an integrated study of gene and non coding RNAs expression and genetic aberrancies. Chapter 1 reports a study on profiling by gene expression arrays of 145 Italian B-others BCP ALL patients and in a representative subcohort of patients microRNAs (miRNAs) expression profiling and genome-wide DNA copy number variation analysis. In this study we found that 25% of Italian B-others patients fits in a group with unique signature and is associated to a favourable outcome. MicroRNAs expression profiling of this group revealed an unique miRNAs signature characterized by over expression of hsa-miR-125b, -125b-2*, -99a, -100, -125a-3p and has-miR-491-5p. Over expression of cluster miR-125b-2 in region 21q21.1 goes along with over expression of genes in the same chromosomal region. Genome-wide analysis excluded copy number alteration of the 21q21.1 region. The frequent involvement of human chromosome 21 (Hsa21) aberrations in ALL (e.g. hyperdiploidy (HD), t(12;21) or iAmp21) and the involvement of the 21q21.1 region suggest a direct and functional contribution of Hsa21 genes to the malignant transformation of hematopoietic cells. There for there is high interest in studying ALL in children with Down Syndrome (DS), where trisomy 21 is constitutional and where the incidence of ALL is approximately 20-fold higher than in the general population. In Chapter 2 is presented a study of genomic analysis of a large group of DS ALLs that characterizes molecular abnormalities specific of this ALL group. Gene expression analysis revealed that DS ALL is a highly heterogeneous disease not definable as a unique ALL subtype with an enrichment of DNA damage and BCL6 responsive genes suggesting B-cell lymphocytic genomic instability. Surprisingly, only a single Hsa21 gene, SON was included in the DS ALL signature and it was only slightly upregulated. Furthermore gene expression data suggested that DS ALL and HD ALL are very different leukemias, reflecting the fundamental differences between constitutional and acquired trisomy such as the developmental stage in which the trisomy occurs and the fact that a constitutional trisomy is present both in the leukemia cells and in their microenvironment. The study further revealed that 62% of the DS ALL samples were characterized by aberrant expression of the type I cytokine receptor CRLF2. Two kind of aberrations involving CRLF2 were identified: a cryptic translocation involving IGH@ and CRLF2 in the pseudoautosomal region (PAR1) of the sex chromosomes and a deletion within PAR1. This aberration resulted in the P2RY8-CRLF2 fusion and leads to overexpression of CRLF2. Furthermore a novel activating somatic mutation, F232C, in CRLF2 was identified. We demonstrated that CRLF2 and mutated JAK2 cooperate in conferring cytokine independent growth to pro-B cells suggesting that the DS ALL children with CRLF2 aberrant expression may benefit from therapy blocking the CRLF2-JAK2 pathway. Since CRLF2 aberrations were found also among non DS patients, we further analyzed the incidence and prognostic impact of this potential new marker in BCP ALL Italian patients enrolled into the AIEOP-BFM ALL2000 study. Chapter 3, presents the study of a representative cohort of 464 non DS BCP ALL patients that was analyzed for the expression levels of CRLF2 and for the occurrence of CRLF2 rearrangements. In this study we found that the P2RY8-CRLF2 rearrangements in association with 20 times over expression of CRLF2 identifies BCP ALL patients with a very poor prognosis and, among them, an important subset of patients currently stratified in the intermediate risk need to be considered for treatment adaptation. Investigating the pathways highlighted by GEP analysis and testing drug effects require a substantial availability of leukemia samples. Primary ALL samples are difficult to culture in vitro and currently available cell lines poorly reflect the heterogeneous nature of the disease. Mouse xenotransplantation models are therefore widely used for in vivo testing and to amplify the number of leukemia cells to be used for various analyses. In Chapter 4 study we assessed the capability of xenografted samples to recapitulate their respective primary leukemia, and we investigated whether the murine microenvironment selects for leukemia initiating cells leading to a bulk tumor markedly different from the diagnostic patient sample. We analysed the gene expression profiles of 7 primary paediatric ALL samples at diagnosis as well as of their respective xenograft leukaemia samples after serial primary, secondary and tertiary passages in the NOD/SCID/huALL transplant model. In this study we demonstrated that the NOD/SCID/huALL transplant model recapitulates the primary human leukaemia, mimics the features of the primary malignancy and retains these characteristics over serial passages without selection for a subclone of the initial leukaemia. Chapter 5 reports on a study that investigated engraftment properties of 50 pediatric ALL samples transplanted into NOD/SCID mice. Time to leukemia (TTL) was determined for each patient sample engrafted as weeks from transplant to overt leukemia. The study shows that short TTL was strongly associated with high risk for early relapse, identifying a new independent prognostic factor. The high risk phenotype is reflected by a gene signature that identified patients with early relapse in an independent patient cohort. Gene expression profiling revealed a set of genes associated with this aggressive phenotype providing a potential strategy to identify these high-risk patients. Most importantly, pathways involving mTOR regulating cell growth were identified, providing targets for alternative therapeutic strategies for these high risk patients. Concluding, ten years after its introduction in oncohematology, GEP constitutes to be a valuable research tool, efficacious in subtype discovery, biomarkers identification and discoveries of deregulated molecular pathways.