Susan R. Atlas

JAK mutations in high-risk childhood acute lymphoblastic leukemia

Pediatric acute lymphoblastic leukemia (ALL) is a heterogeneous disease consisting of distinct clinical and biological subtypes that are characterized by specific chromosomal abnormalities or gene mutations. Mutation of genes encoding tyrosine kinases is uncommon in ALL, with the exception of Philadelphia chromosome-positive ALL, where the t(9,22)(q34;q11) translocation encodes the constitutively active BCR-ABL1 tyrosine kinase. We recently identified a poor prognostic subgroup of pediatric BCR-ABL1-negative ALL patients characterized by deletion of IKZF1 (encoding the lymphoid transcription factor IKAROS) and a gene expression signature similar to BCR-ABL1-positive ALL, raising the possibility of activated tyrosine kinase signaling within this leukemia subtype. Here, we report activating mutations in the Janus kinases JAK1 (n = 3), JAK2 (n = 16), and JAK3 (n = 1) in 20 (10.7%) of 187 BCR-ABL1-negative, high-risk pediatric ALL cases. The JAK1 and JAK2 mutations involved highly conserved residues in the kinase and pseudokinase domains and resulted in constitutive JAK-STAT activation and growth factor independence of Ba/F3-EpoR cells. The presence of JAK mutations was significantly associated with alteration of IKZF1 (70% of all JAK-mutated cases and 87.5% of cases with JAK2 mutations; P = 0.001) and deletion of CDKN2A/B (70% of all JAK-mutated cases and 68.9% of JAK2-mutated cases). The JAK-mutated cases had a gene expression signature similar to BCR-ABL1 pediatric ALL, and they had a poor outcome. These results suggest that inhibition of JAK signaling is a logical target for therapeutic intervention in JAK mutated ALL.

Identification of novel cluster groups in pediatric high-risk B-precursor acute lymphoblastic leukemia with gene expression profiling: correlation with genome-wide DNA copy number alterations, clinical characteristics, and outcome

To resolve the genetic heterogeneity within pediatric high-risk B-precursor acute lymphoblastic leukemia (ALL), a clinically defined poor-risk group with few known recurring cytogenetic abnormalities, we performed gene expression profiling in a cohort of 207 uniformly treated children with high-risk ALL. Expression profiles were correlated with genome-wide DNA copy number abnormalities and clinical and outcome features. Unsupervised clustering of gene expression profiling data revealed 8 unique cluster groups within these high-risk ALL patients, 2 of which were associated with known chromosomal translocations (t(1;19)(TCF3-PBX1) or MLL), and 6 of which lacked any previously known cytogenetic lesion. One unique cluster was characterized by high expression of distinct outlier genes AGAP1, CCNJ, CHST2/7, CLEC12A/B, and PTPRM; ERG DNA deletions; and 4-year relapse-free survival of 94.7% ± 5.1%, compared with 63.5% ± 3.7% for the cohort (P = .01). A second cluster, characterized by high expression of BMPR1B, CRLF2, GPR110, and MUC4; frequent deletion of EBF1, IKZF1, RAG1-2, and IL3RA-CSF2RA; JAK mutations and CRLF2 rearrangements (P < .0001); and Hispanic ethnicity (P < .001) had a very poor 4-year relapse-free survival (21.0% ± 9.5%; P < .001). These studies reveal striking clinical and genetic heterogeneity in high-risk ALL and point to novel genes that may serve as new targets for diagnosis, risk classification, and therapy.

GENE EXPRESSION CLASSIFIERS FOR RELAPSE FREE SURVIVAL AND MINIMAL RESIDUAL DISEASE IMPROVE RISK CLASSIFICATION AND OUTCOME PREDICTION IN PEDIATRIC B-PRECURSOR ACUTE LYMPHOBLASTIC LEUKEMIA

To determine whether gene expression profiling could improve outcome prediction in children with acute lymphoblastic leukemia (ALL) at high risk for relapse, we profiled pretreatment leukemic cells in 207 uniformly treated children with high-risk B-precursor ALL. A 38-gene expression classifier predictive of relapse-free survival (RFS) could distinguish 2 groups with differing relapse risks: low (4-year RFS, 81%, n = 109) versus high (4-year RFS, 50%, n = 98; P < .001). In multivariate analysis, the gene expression classifier (P = .001) and flow cytometric measures of minimal residual disease (MRD; P = .001) each provided independent prognostic information. Together, they could be used to classify children with high-risk ALL into low- (87% RFS), intermediate- (62% RFS), or high- (29% RFS) risk groups (P < .001). A 21-gene expression classifier predictive of end-induction MRD effectively substituted for flow MRD, yielding a combined classifier that could distinguish these 3 risk groups at diagnosis (P < .001). These classifiers were further validated on an independent high-risk ALL cohort (P = .006) and retainedindependent prognostic significance (P < .001) in the presence of other recently described poor prognostic factors (IKAROS/IKZF1 deletions, JAK mutations, and kinase expression signatures). Thus, gene expression classifiers improve ALL risk classification and allow prospective identification of children who respond or fail current treatment regimens. These trials were registered at http://clinicaltrials.gov under NCT00005603.

Gene expression profiles predictive of outcome and age in infant acute lymphoblastic leukemia: a Children's Oncology Group study

Gene expression profiling was performed on 97 cases of infant ALL from Childrens Oncology Group Trial P9407. Statistical modeling of an outcome predictor revealed 3 genes highly predictive of event-free survival (EFS), beyond age and MLL status: FLT3, IRX2, and TACC2. Low FLT3 expression was found in a group of infants with excellent outcome (n = 11; 5-year EFS of 100%), whereas differential expression of IRX2 and TACC2 partitioned the remaining infants into 2 groups with significantly different survivals (5-year EFS of 16% vs 64%; P < .001). When infants with MLL-AFF1 were analyzed separately, a 7-gene classifier was developed that split them into 2 distinct groups with significantly different outcomes (5-year EFS of 20% vs 65%; P < .001). In this classifier, elevated expression of NEGR1 was associated with better EFS, whereas IRX2, EPS8, and TPD52 expression were correlated with worse outcome. This classifier also predicted EFS in an independent infant ALL cohort from the Interfant-99 trial. When evaluating expression profiles as a continuous variable relative to patient age, we further identified striking differences in profiles in infants less than or equal to 90 days of age and those more than 90 days of age. These age-related patterns suggest different mechanisms of leukemogenesis and may underlie the differential outcomes historically seen in these age groups.

A Bayesian Network Classification Methodology for Gene Expression Data

We present new techniques for the application of a Bayesian network learning framework to the problem of classifying gene expression data. The focus on classification permits us to develop techniques that address in several ways the complexities of learning Bayesian nets. Our classification model reduces the Bayesian network learning problem to the problem of learning multiple subnetworks, each consisting of a class label node and its set of parent genes. We argue that this classification model is more appropriate for the gene expression domain than are other structurally similar Bayesian network classification models, such as Naive Bayes and Tree Augmented Naive Bayes (TAN), because our model is consistent with prior domain experience suggesting that a relatively small number of genes, taken in different combinations, is required to predict most clinical classes of interest. Within this framework, we consider two different approaches to identifying parent sets which are supported by the gene expression observations and any other currently available evidence. One approach employs a simple greedy algorithm to search the universe of all genes; the second approach develops and applies a gene selection algorithm whose results are incorporated as a prior to enable an exhaustive search for parent sets over a restricted universe of genes. Two other significant contributions are the construction of classifiers from multiple, competing Bayesian network hypotheses and algorithmic methods for normalizing and binning gene expression data in the absence of prior expert knowledge. Our classifiers are developed under a cross validation regimen and then validated on corresponding out-of-sample test sets. The classifiers attain a classification rate in excess of 90% on out-of-sample test sets for two publicly available datasets. We present an extensive compilation of results reported in the literature for other classification methods run against these same two datasets. Our results are comparable to, or better than, any we have found reported for these two sets, when a train-test protocol as stringent as ours is followed.

GSI-I (Z-LLNle-CHO) inhibits γ-secretase and the proteosome to trigger cell death in precursor-B acute lymphoblastic leukemia

Gamma secretase inhibitors (GSIs) comprise a growing class of compounds that interfere with the membrane-bound Notch signaling protein and its downstream intra-nuclear transcriptional targets. As GSI-I (Z-LLNle-CHO) is also a derivative of a widely used proteosome inhibitor MG-132, we hypothesized that this compound might be active in precursor-B acute lymphoblastic leukemia (ALL) cell lines and patient samples. We found that GSI-I treatment of precursor-B ALL blasts induced apoptotic cell death within 18–24 h. With confirmation using RNA and protein analyses, GSI-I blocked nuclear accumulation of cleaved Notch1 and Notch2, and inhibited Notch targets Hey2 and Myc. Microarray analyses of 207 children with high-risk precursor-B ALL demonstrate that Notch pathway expression is a common feature of these neoplasms. However, microarray studies also implicated additional transcriptional targets in GSI-I-dependent cell death, including genes in the unfolded protein response, nuclear factor-κB and p53 pathways. Z-LLNle-CHO blocks both γ-secretase and proteosome activity, inducing more robust cell death in precursor-B ALL cells than either proteosome-selective or γ-secretase-selective inhibitors alone. Using Z-LLNle-CHO in a nonobese diabetes/severe combined immunodeficiency (NOD/SCID) precursor-B ALL xenograft model, we found that GSI-I alone delayed or prevented engraftment of B-lymphoblasts in 50% of the animals comprising the experimental group, suggesting that this compound is worthy of additional testing.

Gene expression overlap affects karyotype prediction in pediatric acute lymphoblastic leukemia.

Gene expression overlap affects karyotype prediction in pediatric acute lymphoblastic leukemia

Integrated statistical and pathway approach to next-generation sequencing analysis: a family-based study of hypertension

Genome wide association studies (GWAS) have been used to search for associations between genetic variants and a phenotypic trait of interest. New technologies, such as next-generation sequencing, hold the potential to revolutionize GWAS. However, millions of polymorphisms are identified with next-generation sequencing technology. Consequently, researchers must be careful when performing such a large number of statistical tests, and corrections are typically made to account for multiple testing. Additionally, for typical GWAS, the p value cutoff is set quite low (approximately <10−8). As a result of this p value stringency, it is likely that there are many true associations that do not meet this threshold. To account for this we have incorporated a priori biological knowledge to help identify true associations that may not have reached statistical significance. We propose the application of a pipelined series of statistical and bioinformatic methods, to enable the assessment of the association of genetic polymorphisms with a disease phenotype--here, hypertension--as well as the identification of statistically significant pathways of genes that may play a role in the disease process.

REGULARIZATION STRATEGIES FOR HYPERPLANE CLASSIFIERS: APPLICATION TO CANCER CLASSIFICATION WITH GENE EXPRESSION DATA

Linear discrimination, from the point of view of numerical linear algebra, can be treated as solving an ill-posed system of linear equations. In order to generate a solution that is robust in the presence of noise, these problems require regularization. Here, we examine the ill-posedness involved in the linear discrimination of cancer gene expression data with respect to outcome and tumor subclasses. We show that a filter factor representation, based upon Singular Value Decomposition, yields insight into the numerical ill-posedness of the hyperplane-based separation when applied to gene expression data. We also show that this representation yields useful diagnostic tools for guiding the selection of classifier parameters, thus leading to improved performance.

Defining genetic intra-tumor heterogeneity: a chronological annotation of mutational pathways

Tumor heterogeneity is believed to be important in tumor progression and its response to therapies. However, despite numerous mutations being reported in human tumors, genetic intra-tumor heterogeneity remains poorly defined. We have developed a novel strategy to provide a chronological annotation of mutational events in a tumor. We used an endometrial tumor from a patient and transplanted it into athymic mice to create many tumor xenografts. While the patient tumor xenografts were initially responsive to raloxifene treatment, xenografts created with cancer cell clones isolated from the same patient tumor showed dramatic differences in response to raloxifene, indicating existence of intratumor heterogeneity with some subpopulations inherently resistant to the drug. A 250K