Faculty Bibliography

BACKGROUND: Amplification of the N-myc oncogene is associated with adverse outcomes in the common childhood tumor, neuroblastoma. Because the transforming properties of Myc are related to its ability to modulate gene expression, the authors used cDNA microarrays to identify potential Myc target genes. METHODS: Expression levels of 4608 genes were analyzed in a series of neuroblastoma cell lines. Identical analyses were performed in a panel of medulloblastoma cell lines to identify c-Myc targets and to determine the extent to which N-Myc targets and c-Myc targets were shared. Comparisons were made between cell lines with high levels versus low levels of Myc protein expression. RESULTS: Array analyses yielded 121 genes with increased expression levels (>or= 1.65-fold) and 9 genes with decreased expression levels in N-Myc-expressing versus nonexpressing cell lines. Many of these were newly identified targets of biologic interest. Fifty percent of the N-Myc targets (60 of 121) were mutual c-Myc targets. A significant correlation between the level of N-myc and selected target gene expression was demonstrated independently in 27 neuroblastoma tumor samples and in an N-myc-inducible cell line system. CONCLUSIONS: A number of diverse pathways are modulated by N-Myc in neuroblastoma. Although, overall, there was significant correlation between myc and target transcript expression among cohorts of tumors, great variability in levels of target expression was seen among individual tumor samples, and this biologic heterogeneity in the levels of target gene expression may offer insight into differences in the clinical behavior of neuroblastoma and may prove to be of prognostic significance in the future

The outcome for children with acute lymphoblastic leukemia (ALL) has improved dramatically with current therapy resulting in an event free survival exceeding 75% for most patients. However significant challenges remain including developing better methods to predict which patients can be cured with less toxic treatment and which ones will benefit from augmented therapy. In addition, 25% of patients fail therapy and novel treatments that are focused on undermining specifically the leukemic process are needed urgently. In Section I, Dr. Carroll reviews current approaches to risk classification and proposes a system that incorporates well-established clinical parameters, genetic lesions of the blast as well as early response parameters. He then provides an overview of emerging technologies in genomics and proteomics and how they might lead to more rational, biologically based classification systems. In Section II, Drs. Mary Relling and Stella Davies describe emerging findings that relate to host features that influence outcome, the role of inherited germline variation. They highlight technical breakthroughs in assessing germline differences among patients. Polymorphisms of drug metabolizing genes have been shown to influence toxicity and the best example is the gene thiopurine methyltransferase (TPMT) a key enzyme in the metabolism of 6-mercaptopurine. Polymorphisms are associated with decreased activity that is also associated with increased toxicity. The role of polymorphisms in other genes whose products play an important role in drug metabolism as well as cytokine genes are discussed. In Sections III and IV, Drs. James Downing and Cheryl Willman review their findings using gene expression profiling to classify ALL. Both authors outline challenges in applying this methodology to analysis of clinical samples. Dr. Willman describes her laboratory's examination of infant leukemia and precursor B-ALL where unsupervised approaches have led to the identification of inherent biologic groups not predicted by conventional morphologic, immunophenotypic and cytogenetic variables. Dr. Downing describes his results from a pediatric ALL expression database using over 327 diagnostic samples, with 80% of the dataset consisting of samples from patients treated on a single institutional protocol. Seven distinct leukemia subtypes were identified representing known leukemia subtypes including: BCR-ABL, E2A-PBX1, TEL-AML1, rearrangements in the MLL gene, hyperdiploid karyotype (i.e., > 50 chromosomes), and T-ALL as well as a new leukemia subtype. A subset of genes have been identified whose expression appears to be predictive of outcome but independent verification is needed before this type of analysis can be integrated into treatment assignment. Chemotherapeutic agents kill cancer cells by activating apoptosis, or programmed cell death. In Section V, Dr. John Reed describes major apoptotic pathways and the specific role of key proteins in this response. The expression level of some of these proteins, such as BCL2, BAX, and caspase 3, has been shown to be predictive of ultimate outcome in hematopoietic tumors. New therapeutic approaches that modulate the apoptotic pathway are now available and Dr. Reed highlights those that may be applicable to the treatment of childhood ALL

To identify genes whose expression correlated with biological features of childhood leukemia, we prospectively analyzed the expression profiles of 4608 genes using cDNA microarrays in 51 freshly processed bone marrow samples from children with acute leukemia, over a 24-month period, at a single institution. Two supervised methods of analysis were used to identify the 20 best discriminating genes between the following cohorts: acute myelogenous leukemia (AML) versus acute lymphoblastic leukemia (ALL); B-lineage versus T-lineage ALL; newly diagnosed B-lineage standard-risk versus high-risk ALL; and B-lineage leukemia harboring the TEL-AML 1 fusion versus patients without a molecularly characterized translocation. These methods identified overlapping sets of genes that segregated patients within described subgroups. Cross-validation demonstrated that the majority of patients could be correctly classified based on these genes alone, and hierarchical clustering grouped patients with similar clinical and biological disease features. The potential for select genes to discriminate patients was validated using real-time PCR in samples that were analyzed by microarray profiling and in other uniformly processed leukemic marrow samples. As expected, microarray technology can successfully segregate patients defined by traditional measures such as immunophenotype and cytogenetic alterations. However, among specific subgroups, this preliminary analysis also suggests that microarrays can identify unanticipated similarities and diversity in individual patients and thus may be useful in augmenting risk-group stratification in the future

The majority of pediatric anaplastic large cell lymphomas (ALCLs) carry the t(2;5)(p23;q35) chromosomal translocation that juxtaposes the dimerization domain of nucleophosmin with anaplastic lymphoma kinase (ALK). The nucleophosmin-ALK fusion induces constitutive, ligand-independent activation of the ALK tyrosine kinase leading to aberrant activation of cellular signaling pathways. To study the early consequences of ectopic ALK activation, a GyrB-ALK fusion was constructed that allowed regulated dimerization with the addition of coumermycin. Expression of the fusion protein caused a coumermycin-dependent increase in cellular tyrosine phosphorylation and c-Myc immunoreactivity, which was paralleled by a rise in c-myc RNA. To assess the clinical relevance of this observation, c-Myc expression was determined in pediatric ALK-positive and -negative lymphomas. Co-expression of c-Myc and ALK was seen in tumor cells in 15 of 15 (100%) ALK-positive ALCL samples, whereas no expression of either ALK or c-Myc was seen in six of six cases of ALK-negative T-cell lymphoma. C-Myc may be a downstream target of ALK signaling and its expression a defining characteristic of ALK-positive ALCLs

Lack of adequate statistical methods for the analysis of microarray data remains the most critical deterrent to uncovering the true potential of these promising techniques in basic and translational biological studies. The popular practice of drawing important biological conclusions from just one replicate (slide) should be discouraged. In this paper, we discuss some modern trends in statistical analysis of microarray data with a special focus on statistical classification (pattern recognition) and variable selection. In addressing these issues we consider the utility of some distances between random vectors and their nonparametric estimates obtained from gene expression data. Performance of the proposed distances is tested by computer simulations and analysis of gene expression data on two different types of human leukemia. In experimental settings, the error rate is estimated by cross-validation, while a control sample is generated in computer simulation experiments aimed at testing the proposed gene selection procedures and associated classification rules

Apoptosis is the primary mechanism through which most chemotherapeutic agents induce tumor cell death. The purpose of this study was to determine the extent to which blasts from children with leukemia undergo a uniform apoptotic death pathway in vivo. The expression of pro- and anti-apoptotic proteins p53, p21, MDM-2, BCL-2, BCL-X(L), BCL-X(S), and BAX, and caspase-3 activity was determined in circulating blasts collected from the peripheral blood of children with leukemia prior to, and at serial time points following chemotherapy. Culturing blasts ex vivo for 12 h assessed spontaneous apoptosis and the increment induced by chemotherapy. Baseline apoptosis varied between 3% and 29%. Twenty-four hours following chemotherapy the increase in the percentage of cells undergoing apoptosis ranged from <1% to 38%. Eleven of 20 patients who received initial treatment with a p53-dependent drug showed an increase in p53 expression. In these patients, the levels of p53 target genes were also increased. A uniform pattern of BCL-2 family protein expression was not observed and only a minority of samples showed a change that would favor apoptosis. We conclude that that the initial apoptotic response to chemotherapy in children with leukemia is variable involving both p53-dependent and p53-independent pathways