RICERCA ITALIANA     

New frontiers in the stratification of acute lymphoblastic leukemia (ALL): integration between non-conventional cytogenetics, genomics and post-genomics.

Università degli Studi di Roma "La Sapienza"

Abstract

Acute lymphoblastic leukemia (ALL) is one the most frequent neoplasms in pediatric cohorts, while it is more rare in adult patients. Prognosis varies widely between the two cohorts: in children, the overall five-year event-free survival rate is nowadays greater than 75%, while in adults the results are much less favorable. The improved results in children have led researchers to try to identify who, among the small patients, may benefit from less intense chemotherapeutic regimes; in addition, the evaluation of Minimal Residual Disease (MRD) during the course of the disease has become a fundamental parameter. This "surrogate" marker, beyond measuring the response to treatment, offers unique opportunities to raise scientific questions with potential clinical implications. In particular, MRD allows to identify subgroups of patients displaying a poor early response to treatment in the absence of any other known prognostic factor, as well as subgroups with very rapid tumor clearance and good outcome, even in subgroups with poor prognostic genetic lesions.
In both groups of patients - children and adults -, there is still the need of understanding the mechanism/s that lead to malignant transformation. In fact, in both adults and children within B-lineage ALL (B-ALL) it has been possible to identify several recurrent molecular aberrations that are strictly associated to leukemia development. Nonetheless, with the exception of BCR/ABL, which is per se causal of the disease, in the remaining scenarios it is not yet clear if there are - and this being the case, which are - the additional lesions that concur to the development of the disease.
In T-lineage ALL (T-ALL), until recently, only few aberrations have been detected. The scenario is, however, dramatically changing and also for these patients there is a handful of novel information, which will lead, ultimately, to revise our knowledge. In this setting, gene expression profiling is now allowing to identify novel subgroups previously unrecognized. It is therefore evident how the integration of genomics, proteomics and miR analysis will make it possible to identify profiles and alternative mechanisms that will allow a refined classification, as well as the recognition of tumor suppressor genes and/or oncogenes contributing to the pathogenesis of ALL.
The 4 research units presenting this ambitious project have ideal conditions to accomplish the above mentioned goals: 1) all the units coordinate and/or participate to national multicentric protocols for the enrolment and treatment of children (AIEOP, Associazione Italiana di Ematologia e Oncologia Pediatrica) and adults with ALL (GIMEMA, Gruppo Italiano Malattie EMatologiche dell’Adulto); 2) overall, they have, in situ and fully operative, novel technologies, which should allow the identification of the pathogenetic mechanisms. 3) They have collaborated on different projects over many years. More specifically, two groups are involved in gene expression profiling studies and partly in miR analysis, a group is currently evaluating the MRD nation-wide and is also performing SNPs analysis, and another group is involved in the evaluation of non-conventional cytogenetics: FISH, CGH, array-CGH and mutational analysis by DHPLC.
In the current project, within the pediatric cohort we aim at:
A) Improving the current classification, by characterizing the role of small chromosomal aberrations and putative concomitant gene expression changes.
B) Evaluating if a different MRD clearance (early or slow clearance) is sustained by specific gene expression profiles and/or chromosomal aberrations (identified by SNP analysis).
C) Utilizing genomic analysis to recognize the biological processes that are specifically deregulated by the presence of additional genomic aberrations.
Within adult patients, we aim at:
D) Evaluating in detail the gene expression profiles of adult T-ALL and, partly, of B-ALL, at the onset of disease.
E) Correlating the gene expression profiles with non-conventional cytogenetics; mutational status of some specific genes will be evaluated as well.
F) Defining the clinical impact of this novel information.
G) Characterizing, in both B-ALL and T-ALL the profile of miR, small non-coding RNA molecules, that degrade and/or repress the expression of their RNA targets.
Finally, we aim at:
H) Comparing the profiles of children, adolescents and adults in order to understand if the different signatures may be associated to the different outcome. <<<

Principal Investigator

Roberto Foà Università degli Studi di ROMA "La Sapienza"

Research Objectives

The analysis of acute lymphoblastic leukemia (ALL) has provided in the last twenty years extremely useful results. In fact, studies performed in this disease have allowed to identify recurrent molecular lesions strictly associated to this neoplasm, such as BCR/ABL, ALL1 rearrangements, E2A/PBX and TEL/AML1. As a general premise, three main issues must be kept in mind: 1) the occurrence of the aforementioned lesions varies widely between children and adults; 2) the presence of these aberrations is more frequently detected in B-ALL, while in T-ALL up to recently the mechanisms of transformation were scarcely known; as a matter of fact, up to 2004, in about 80% of T-ALL patients it was not possible to recognize any genetic lesion; 3) the impact on the clinical behavior is well-established with some lesions, as for example BCR/ABL, being associated with a particularly poor outcome (lower percentage of responses to induction chemotherapy, as well as a higher incidence of relapse), and others, e.g. TEL/AML1, that is almost exclusively detected in children, which are associated to a favorable clinical outcome.
Overall, this body of information has represented an ideal background for the definition of stratification criteria and, ultimately, for the design of tailored chemotherapeutic regimens that in some cases include the use of transplant procedures, while in other circumstances allow to reduce treatment intensity. Moreover, in the last ten years the role of minimal residual disease (MRD) analysis has proven very useful in patients’ stratification; it is important, however, to underline that this parameter is routinely used in pediatric setting, while in adult cohorts the introduction of MRD evaluation is more recent and its effective role needs to be further established.
Although the state of the art and the biologic knowledge of ALL – and more in general of onco-hematology – is more refined than in the majority of solid tumors, several issues are still open. In fact, 1) in about 50% of B-ALL and in the greatest part of T-ALL the transformation mechanism is still unknown; 2) even in patients in which a molecular lesion is detected, the event - or the events – that lead/s to a frank leukemic picture, is/are not known; 3) the contribution of additional lesions is not recognized; 4) it is sometimes difficult to define which lesions are mutually exclusive or, contrariwise, which aberrations are complementary to each other; 5) similarly, the reasons whereby patients with the same major molecular lesions exert different clinical behaviors are not fully elucidated; 6) finally, it is not known if other genetic lesions, as for example the involvement of miRs, may play a role in initiating leukemia; if this hypothesis is true, the mechanisms of deregulation are not clear.
Therefore, the current collaborative project aims, within the pediatric cohort, at:
A) Identifying, through the use of SNPs arrays, chromosomal microdeletions and/or microamplification, previously not identified; this part of the project will be performed in both samples that harbor known molecular lesions, as well as in samples without major genetic aberrations.
Furthermore, taking advantage of a gene expression profile database that includes about 400 pediatric ALL (MILE study, Microarray Innovations in LEukemia), we will attempt to understand if these chromosomal microaberrations may result in gene expression changes, both in terms of a “gene dosage effect” and in terms of deregulation of cell functional pathways, such as apoptosis, cell cycle, cell differentiation and development, etc.
B) Evaluating, through gene expression profile analysis, if changes in the most important cellular functions (see above) occur, regardless of the presence of the aforementioned chromosomal microlesions.
C) Improving the current classification and stratification, through the correlation with the information acquired with the SNPs and gene expression profiling studies.
D) Evalutating the hypothesis that different levels of MRD clearance (early or late response) are sustained by the presence of chromosomal microaberrations and/or specific gene profiles.

Within adult patients, we aim at:
E) Evaluating in detail the gene expression profiles of B-ALL and T-ALL. More specifically, we will first focus on T-ALL analysis, mostly because in this group of patients the current knowledge is more limited and also because preliminary results from one of the proposing groups highlight the presence of several subgroups, previously not recognized, that may be of particular interest for the comprehension of the underlying leukemogenic mechanisms. Subsequently, the analysis will concentrate on B-ALL patients that do not harbor any known major molecular aberration. Also in this context, we will take advantage of the MILE database, that presently includes roughly 300 adult ALL patients.
F) Assessing in B-ALL and T-ALL, by non-conventional cytogenetics, using a combination of FISH with multiple probes (so called “multiplex FISH”), array-CGH and the mutational status of specific genes, to define the incidence, the pathogenetic role and the mutual exclusivity, or otherwise complementarity of these aberrations.
G) Correlating the results obtained with the analysis of gene expression profiling with those obtained by non-conventional cytogenetics (points E + F).
H) Defining the clinical impact of the aberrations thus identified: this will be possible mostly because all patients that will be studied are enrolled in multicentric clinical trials, operating within a nation-wide framework; furthermore, a central database collects the clinical information on all patients. Different clinical parameters will be taken into account, and in particular clinical characteristics (white blood cell count, SNC involvement, bulky disease, etc) at the onset of the disease, response to induction chemotherapy and long-term clinical follow-up.
I) Characterizing, in both B-ALL and T-ALL, the profile of miRs, small non-coding RNA molecules involved in several functional processes, which exert their role by degrading and/or repressing their target genes. Wherever possible, we will not only define the differentially expressed miR among the various subgroups, but will also try to identify their target genes, validating the results in experimental models.
Finally, we will:
L) Compare the expression profiles, within the various molecular subsets, of children, adolescentes, young adults, adults and elderly ALL patients, with the goal of identifying a subset of differentially expressed genes that may contribute to the different clinical outcome. As already mentioned throughout this project, this is made possible because two of the participating units are actively involved in the international MILE study. <<<

First Results

This project is absolutely feasible because all the units are using, with high proficiency, powerful technologies, such as gene expression profiling arrays, array-CGH, SNP array and a handful set of non-conventional cytogenetic techniques.
In the pediatric setting, the expected results are:
1) Discovery, by applying high resolution SNP array-based and CGH array-based technologies, of secondary genomic imbalances of possible prognostic and pathogenic importance for ALL stratification that have so far gone undetected. Hopefully, these lesions will be detected not only in samples without known molecular aberrations, but also in those cases, as recently reported in the literature, that display some known recurrent lesion (for example TEL/AML1), and may therefore be useful in further stratification of patients also in these cases.
2) Re-analysis and revision, in light of the lesions identified by array-CGH and SNP array, the gene expression profiling results of pediatric ALL, and if possible, to explain the reason/s of the presence of subgroups that are identified by unsupervised and semi-supervised approaches that, presently, do not appear to correlate with any of current biological knowledge.
3) Correlation of the results of MRM with those obtained by gene profiling, array-CGH and SNP-array.
4) Investigation of the role of microRNA to: 1) directly infer the variation of transcript/microRNA levels in deleted or amplified regions, and 2) identify variation in gene expression of genes indirectly affected by recurrent genomic alterations or by microRNA differential expression.
5) Evaluation of the alterations of signalling profile of cells following exposure to a therapeutic drug could that could indicate which signalling pathways are affected by the drug. They may represent a new category of biomarkers in ALL and these information could contribute to the understanding of drug resistance mechanisms and thus potentially open new perspectives in targeted therapies in leukaemia.
6) Eventually, as a final goal, identification of novel genetic markers to better stratify patients into even better defined risk groups, to potentially identify new subgroups with specific risk of clinical relapse, as well as new candidate genes for specific new therapies.
Within adult patients, we expect to:

7) to identify molecular pathways which are specific and concur to leukemogenesis in individual cases of B- and T-ALL and to establish new genetic subgroups of clinical entities.
8) To set up "multiplex FISH" assays for the simultaneous analysis of a panel of 30 probes in each case.
9) To subject a large series of patients to mutational analysis (using DHPLC) of known genes and of new candidates emerging from the project.
10) To detect cryptic genomic imbalances (using micro-array CGH and SNPs) and determine the impact of their association with other genomic aberrations.
11) To elaborate diagnostic algorythms by integrating different technologies to diagnose specific molecular subgroups.
12) To identify specific clinical-prognostic features associated with each genomic subgroup.
13) To identify novel subgroups in T-ALL, by using the gene expression profiling, as already suggested by preliminary results that clearly show the presence of several subgroups, previously not recognized. Similarly, to identify subgroups within B-ALL
14) To comprehend the molecular lesion/s underlying the gene expression profile observed, by integration of non-conventional cytogenetics.
15) To understand if this “gene-expression based” classification may be prognostically relevant.
16) As already mentioned for children, the final goal is the identification of novel genetic markers to better stratify patients, to potentially identify new subgroups of patients with different risk of failure to treatment as well as clinical relapse and different outcome.
17) To identify novel candidate genes for targeted therapies.
18) To define, as for pediatric ALL, the role of miRs, their differential expression as well as their putative gene targets.
Finally, as a combined effort of all the proposing units we expect to:
19) Identify genes, that within specifically defined subgroups (as for example, the subgroups harboring defined molecular aberrations and/or exerting specific immunophenotypic features), are differentially expressed between adult and pediatric ALL. These genes may be useful for understanding why these age-related subgroups display such a different clinical outcome.
20) Identify chromosomal aberrations, and in particular micro-chromosomal abnormalities that were previously unrecognized, that occur differently between children, adolescents, young adults, adults and elder and may be of prognostic impact.
21) We finally intend to compare the miR profile between adults and children, with the aim of establishing if these molecules may differently participate to leukemogenesis in the the two cohorts analyzed.
22)Publication of the results obtained on national and/or international scientific journals as well as reports on scientific congresses. <<<

Timescale

24 months

National and international background

Acute lymphoblastic leukemia (ALL) is an heterogeneous disease with a highly heterogeneous clinical course; its incidence varies widely between children and adults. Clinical heterogeneity is known to be correlated with recurrent patterns of genetic changes in the leukemic cells and current acute leukemia subclassifications utilize genomic hallmarks for treatment stratification. The analyses used to recognize leukemia subclasses have evolved along several lines, including flow cytometry for lineage-specific marker proteins and DNA-indexing of blast cells, karyotyping to recognize gross chromosomal aberrations, FISH utilizing probes for recurrent aberrations and molecular biology, which uses RT-PCR and quantitative PCR (Q-PCR) for the screening and quantification of recurrent translocations.
Altogether, these approaches have so far allowed to distinguish among acute leukemias the following B-lineage (B-ALL) subclasses: mature B-ALL with t(8;14), pro-B-ALL with rearrangements involving the ALL1 (MLL) gene, c-ALL/pre-B-ALL with t(9;22), ALL with t(12;21), ALL with t(1;19), ALL with hyperdiploid karyotype, c-ALL/pre-B-ALL without any of the above aberrations. The incidence of these abnormalities is different in pediatric and adult cohorts: more specifically, the t(9;22) is much more frequent in adults (and more so in the elderly), as well as rearrangements involving the MLL, with the exception of infants; contrariwise, the t(12;21) is exclusively detected in children and hyperploid karyotype cases are mostly found in the pediatric cohorts.
T-ALL represents, in both pediatric and adult cohorts, about 25% of patients: in these cases, the biologic knowledge has so far been limited and the mechanism/s underlying the malignant transformation was/were not known: as a matter of fact, up to 2004 genetic lesions could be identified in not more than 20% of cases.
With the development of microarray technologies, more specifically gene expression microarrays, SNPs analysis and array-CGH, the scenario is rapidly changing and, at present, there is the chance of a genome-scale survey of aberrancies in leukemias, with the opportunity of identifying also the micro-aberrations.
Within B-ALL, apart from genomic hallmarks of diagnostic leukemia subclasses, other secondary aberrancies are understood to have an important clinical impact on acute leukemias. For example, the translocation t(12;21), which results in the expression of ETV6/RUNX1 fusion gene, is not sufficient for leukemic transformation and other genetic lesions secondary to EVT6/RUNX1 are found in more than 80% of t(12;21)-positive ALLs, the most common secondary aberration being a deletion of the normal ETV6 gene present in 70% of cases investigated by FISH. Other frequent aberrations include duplications of the normal (20%) or the derivative chromosome 21 (10%), deletions of 6q (18%) and deletions of 9p (7%). Apart from the additional 12p and 21q abnormalities, often found by FISH with probes specific for EVT6 and RUNX1, most of the secondary abnormalities have been detected by chromosome banding analyses alone, or in some instances by multicolor FISH.
Because of the limited resolution levels of these methods, many small and all cytogenetically cryptic-genomic imbalances of possible prognostic and pathogenic importance have so far gone undetected. A large scale genome-wide analysis of leukemic cells from 242 pediatric ALL patients using high-resolution, SNPs arrays and genomic DNA sequencing revealed deletions, amplifications, point mutations and structural rearrangements in genes encoding the principal regulators of B-lymphocyte development and differentiation in 40% of B-progenitor ALL cases. The PAX5 gene was the most frequent target of somatic mutations, being altered in 31.7% of cases. Preliminary results from one of the proposing groups, that used SNP arrays to evaluate 29 children without known molecular aberrations, showed that the vast majority of cases analyzed (83%) displays genomic abnormalities. The most relevant abnormalities were CDKN2A/9p21 deletion (7/29, 24%), TEL/12p13 deletion (3/29, 10%) and intra-chromosomal amplification of chromosome 21 (iAMP21) (3/29, 10%). With a novel approach, this proposing group integrated gene copy number and gene expression data, and identified variations in expression of genes directly or indirectly affected by the genomic alterations. SMAD1 emerged as a down-regulated gene in p16 homozygous deleted cases compared with non-deleted cases. The JAG1 gene encoding for the Jagged1 ligand of the Notch receptor was among a list of up-regulated genes in TEL deleted cases. Furthermore, other frequent abnormalities recently identified within hyperdiploid (&gt;50) karyotypes include trisomy 4, 10, and 17, which appear to be critically associated with prognosis; involvement of the genes belonging to the CEBP family, including CEBPA, as translocation partners of IgH; translocations involving genes other than IgH may activate the MYC gene in B-ALL through a mechanism similar to what has been described in non-Hodgkin lymphomas; and finally, chromosome 21 amplifications. Because of the limited resolution of the previously applied technologies, these small size aberrations, with potential prognostic impact, were, so far, not identified.
Within T-ALL, the picture is dramatically changing as well, with novel lesions emerging: at present, it is in fact possible to detect a genetic lesion in as much as 80% of patients. Among these newly identified lesions, it is worth mentioning the CALM/AF10 rearrangement, NOTCH1 mutations and ABL1 rearrangements. It is important to underline that, at variance from B-ALL, these aberrations are rarely detected using conventional cytogenetics and for this reason their recognition has been somehow troublesome. In line with this, gene expression profiling has provided complementary results: in fact, the first study that evaluated the expression profile of T-ALL cases highlighted that the overexpression of a set of known oncogenes - namely LYL, HOX11 (or TLX1) and TAL1 - induces specific patterns of expression that also correlate with various stages of the T-lymphocyte maturation. Furthermore, this study also identified another subgroup of patients that displays similar, though not identical, profile of that of HOX11+ patients: these patients have high levels of expression of the HOX11L2 (or TLX3) gene. Subsequently, gene expression profiling analysis, performed on primary leukemic cells, has highlighted that in patients harboring the CALM/AF10 rearrangement, there is an overexpression of genes belonging to the Homeobox family, and in particular of the HOXA gene. This result is of importance, because a similar profile is constantly observed in patients with ALL1 rearrangement and indicates that various abnormalities may lead to the activation of the same pathways, which, in turn, may play a fundamental role in the leukemic transformation.
Genomic analysis has also been used to identify patients with ABL1 rearrangement who express high levels of ABL and are therefore easy to detect. Also in this scenario it is important to underline the importance of integrating gene expression profile with other technologies, which allow to confirm the presence of the rearrangements that are responsible of such overexpression. Finally, by gene expression profiling, in cell lines, the mechanism of transformation sustained by NOTCH1 has been clarified: NOTCH1 deregulates the expression of MYC and HES1 that in turn down-modulate PTEN, with consequent activation of AKT, that ultimately leads to a proliferative advantage and to leukemic growth. Overall, these findings indicate how the field is rapidly evolving and pinpoint the importance of integrating the various technologies.
Finally, preliminary results from one of the proposing groups using gene expression profiling has highlighted that by unsupervised analysis of a relatively large cohort of adult patients with T-ALL, it is possible to identify several subgroups, thus confirming and corroborating the need of a more refined understanding of the underlying mechanisms of transformation.
Summarizing, two major genetic characteristics of T-ALL have emerged: reciprocal translocations with a high prevalence of partner gene abnormal expression and the presence of two or more concomitant rearrangements in every case and mutations affecting critical genes. Genome analysis has also detected micro-deletions and duplications, and each case of T-ALL is estimated to bear 2 or 3 of these genomic imbalances. Identification of concomitant molecular rearrangements in leukemic blasts will illuminate the specific leukemogenic pathway in each individual and provide the basis for targeted therapy and for monitoring minimal residual disease (MRD). One example of the success of this approach is the experimental use of gamma-secretase inhibitors in NOTCH1-positive T-ALL.
Another field of particular importance is the evaluation of MRD during the course of the disease: this approach, which is largely used in pediatric ALL, allows to identify subgroups with limited early response to therapy in the absence of other prognostic parameters, or otherwise patients with a rapid MRD clearance of the leukemic clone, even in subgroups with apparent poor prognostic features. This analysis is currently being evaluated also in adult ALL. Most probably, the identification of novel lesions will lead to the detection, and therefore use, of new markers that will be integrated in the MRD analysis; this effort will also permit to correlate the early and/or late clearance levels with these new acquisitions.
Finally, another field that is rapidly expanding, is the analysis of miR, a class of small non-coding RNAs which are highly conserved throughout different species and whose role is to repress and/or degrade, at the post-trascriptional level, their gene targets. Although the role of miR is nowadays clear in hemopoiesis, in myeloid and lymphoid commitment, as well as in acute myeloid leukemia (AML) and chronic lymphocytic leukemia (CLL), miR function in B- and T-ALL is not yet clear. In fact, a single study has shown different expression levels of a set of miR, but so far neither their role nor their targets have been identified. It is of importance to underline that two of the proposing units are already evaluating the profile of miRs.

Selected references
Bungaro S, et al. Haematologica, 91:998-1000, 2006.
Bungaro S, et al.. Leukemia. 2007, Epub.
Burmeister T, et al. Blood, 108:3556-59, 2006.
Chen CZ, et al. Science, 30:83-6, 2004.
Chiaretti S, et al. Clin Cancer Res, 11:7209-19, 2005.
Chiaretti S, et al. Haematologica, 92:619-26, 2007.
De Keersmaecker K, et al. Blood, 105:4849-52, 2005.
Fulci V, et al. Blood, 109:4944-51, 2007.
Garraway LA, et al. Nature, 436:117-22, 2005.
Jenner MW,. Blood, 110:3291-300, 2007.
Haferlach T, et al. Blood (ASH Annual Meeting Abstracts), 108: 103, 2006.
Irish JM,et al. Nat Rev Cancer, 146-55, 2006.
Irving JA, et alCancer Res, 65:3053-8, 2005.
Kawamata N, et al. Blood. 2007 Sep 21; epub
Landgraf P, et al. Cell, 129:1401-14, 2007.
La Starza R, et al. Haematologica, 91:1248-51, 2006.
Mullighan CG, et al.Nature, 446:758-64, 2007.
Perez OD, et al. Immunol Rev, 210:208-28, 2006.
Pui CH, et al. Lancet.;352:1731-8, 1998.
Raghavan M, et al. Cancer Res, 65:375-8, 2005.
Ross ME, et al. Blood, 102:2951-2959, 2003.
Wang Y, et al. Cancer Cell, 11:308-9, 2007. <<<