Search

Abstract As part of the advancement in therapeutic decision-making for brain tumor patients at St. Jude Children’s Research Hospital (SJCRH), we developed three robust classifiers, a deep learning neural network (NN), k-nearest neighbor (kNN), and random forest (RF), trained on a reference series DNA-methylation profiles to classify central nervous system (CNS) tumor types. The models’ performance was rigorously validated against 2054 samples from two independent cohorts. In addition to classic metrics of model performance, we compared the robustness of the three models to reduced tumor purity, a critical consideration in the clinical utility of such classifiers. Our findings revealed that the NN model exhibited the highest accuracy and maintained a balance between precision and recall. The NN model was the most resistant to drops in performance associated with a reduction in tumor purity, showing good performance until the purity fell below 50%. Through rigorous validation, our study emphasizes the potential of DNA-methylation-based deep learning methods to improve precision medicine for brain tumor classification in the clinical setting.

Abstract In this study, we leveraged machine-learning tools by evaluating expression of genes of pharmacological relevance to standard-AML chemotherapy (ara-C/daunorubicin/etoposide) in a discovery-cohort of pediatric AML patients (N = 163; NCT00136084 ) and defined a 5-gene-drug resistance score (ADE-RS5) that was predictive of outcome (high MRD1 positivity p = 0.013; lower EFS p

SAMD9 and SAMD9L germline mutations have recently emerged as a new class of predispositions to pediatric myeloid neoplasms. Patients commonly have impaired hematopoiesis, hypocellular marrows, and a greater risk of developing clonal chromosome 7 deletions leading to MDS and AML. We recently demonstrated that expressing SAMD9 or SAMD9L mutations in hematopoietic cells suppresses their proliferation and induces cell death. Here, we generated a mouse model that conditionally expresses mutant Samd9l to assess the in vivo impact on hematopoiesis. Using a range of in vivo and ex vivo assays, we showed that cells with heterozygous Samd9l mutations have impaired stemness relative to wild-type counterparts, which was exacerbated by inflammatory stimuli, and ultimately led to bone marrow hypocellularity. Genomic and phenotypic analyses recapitulated many of the hematopoietic cellular phenotypes observed in patients with SAMD9 or SAMD9L mutations, including lymphopenia, and pinpointed TGF-β as a potential targetable pathway. Further, we observed nonrandom genetic deletion of the mutant Samd9l locus on mouse chromosome 6, mimicking chromosome 7 deletions observed in patients. Collectively, our study has enhanced our understanding of mutant Samd9l hematopoietic phenotypes, emphasized the synergistic role of inflammation in exaggerating the associated hematopoietic defects, and provided insights into potential therapeutic options for patients.

Paediatric therapy-related myeloid neoplasms (tMN) have a dismal prognosis and have not been comprehensively profiled. Here the authors characterise the molecular landscape of 84 paediatric tMN patients, and find that, unlike adult tMNs, these do not emerge from pre-existing clones and that MECOM dysregulation is frequent.

Abstract Background There is currently no method to precisely measure the errors that occur in the sequencing instrument/sequencer, which is critical for next-generation sequencing applications aimed at discovering the genetic makeup of heterogeneous cellular populations. Results We propose a novel computational method, SequencErr, to address this challenge by measuring the base correspondence between overlapping regions in forward and reverse reads. An analysis of 3777 public datasets from 75 research institutions in 18 countries revealed the sequencer error rate to be ~ 10 per million (pm) and 1.4% of sequencers and 2.7% of flow cells have error rates > 100 pm. At the flow cell level, error rates are elevated in the bottom surfaces and > 90% of HiSeq and NovaSeq flow cells have at least one outlier error-prone tile. By sequencing a common DNA library on different sequencers, we demonstrate that sequencers with high error rates have reduced overall sequencing accuracy, and removal of outlier error-prone tiles improves sequencing accuracy. We demonstrate that SequencErr can reveal novel insights relative to the popular quality control method FastQC and achieve a 10-fold lower error rate than popular error correction methods including Lighter and Musket. Conclusions Our study reveals novel insights into the nature of DNA sequencing errors incurred on DNA sequencers. Our method can be used to assess, calibrate, and monitor sequencer accuracy, and to computationally suppress sequencer errors in existing datasets.

BackgroundAcute myeloid leukemia (AML) is a hematological malignancy with a dismal prognosis. For over four decades, AML has primarily been treated by cytarabine combined with an anthracycline. Although a significant proportion of patients achieve remission with this regimen, roughly 40% of children and 70% of adults relapse. Over 90% of patients with resistant or relapsed AML die within 3 years. Thus, relapsed and resistant disease following treatment with standard therapy are the most common clinical failures that occur in treating this disease. In this study, we evaluated the relationship between AML cell line global metabolomes and variation in chemosensitivity.MethodsWe performed global metabolomics on seven AML cell lines with varying chemosensitivity to cytarabine and the anthracycline doxorubicin (MV4.11, KG-1, HL-60, Kasumi-1, AML-193, ME1, THP-1) using ultra-high performance liquid chromatography – mass spectrometry (UHPLC-MS). Univariate and multivariate analyses were performed on the metabolite peak intensity values from UHPLC-MS using MetaboAnalyst to identify cellular metabolites associated with drug chemosensitivity.ResultsA total of 1,624 metabolic features were detected across the leukemic cell lines. Of these, 187 were annotated to known metabolites. With respect to doxorubicin, we observed significantly greater abundance of a carboxylic acid (1-aminocyclopropane-1-carboxylate) and several amino acids in resistant cell lines. Pathway analysis found enrichment of several amino acid biosynthesis and metabolic pathways. For cytarabine resistance, nine annotated metabolites were significantly different in resistance vs. sensitive cell lines, including D-raffinose, guanosine, inosine, guanine, aldopentose, two xenobiotics (allopurinol and 4-hydroxy-L-phenylglycine) and glucosamine/mannosamine. Pathway analysis associated these metabolites with the purine metabolic pathway.ConclusionOverall, our results demonstrate that metabolomics differences contribute toward drug resistance. In addition, it could potentially identify predictive biomarkers for chemosensitivity to various anti-leukemic drugs. Our results provide opportunity to further explore these metabolites in patient samples for association with clinical response.

Abstract Consolidation therapies for children with intermediate- or high-risk acute myeloid leukemia (AML) are urgently needed to achieve higher cure rates while limiting therapy-related toxicities. We determined if adoptive transfer of natural killer (NK) cells from haploidentical killer immunoglobulin–like receptor (KIR)–human leukocyte antigen (HLA)-mismatched donors may prolong event-free survival in children with intermediate-risk AML who were in first complete remission after chemotherapy. Patients received cyclophosphamide (Day − 7), fludarabine (Days − 6 through − 2), and subcutaneous interleukin-2 (Days − 1, 1, 3, 5, 7, and 9). Purified, unmanipulated NK cells were infused on Day 0, and NK cell chimerism and phenotyping from peripheral blood were performed on Days 7, 14, 21, and 28. As primary endpoint, the event-free survival was compared to a cohort of 55 patients who completed chemotherapy and were in first complete remission but did not receive NK cells. Donor NK cell kinetics were determined as secondary endpoints. Twenty-one patients (median age at diagnosis, 6.0 years [range, 0.1–15.3 years]) received a median of 12.5 × 106 NK cells/kg (range, 3.6–62.2 × 106 cells/kg) without major side effects. All but 3 demonstrated transient engraftment with donor NK cells (median peak donor chimerism, 4% [range, 0–43%]). KIR–HLA-mismatched NK cells expanded in 17 patients (81%) and contracted in 4 (19%). However, adoptive transfer of NK cells did not decrease the cumulative incidence of relapse (0.393 [95% confidence interval: 0.182–0.599] vs. 0.35 [0.209–0.495]; P = .556) and did not improve event-free (60.7 ± 10.9% vs. 69.1 ± 6.8%; P = .553) or overall survival (84.2 ± 8.5% vs. 79.1 ± 6.6%; P = .663) over chemotherapy alone. The lack of benefit may result from insufficient numbers and limited persistence of alloreactive donor NK cells but does not preclude its potential usefulness during other phases of therapy, or in combination with other immunotherapeutic agents. Trial registration www.clinicaltrials.gov, NCT00703820. Registered 24 June 2008.

Retinoblastoma is a childhood cancer of the developing retina that begins in utero and is diagnosed in the first years of life. Biallelic RB1 gene inactivation is the initiating genetic lesion in retinoblastoma. The p53 gene is intact in human retinoblastoma but the pathway is believed to be suppressed by increased expression of MDM4 (MDMX) and MDM2. Here we quantify the expression of MDM4 and MDM2 mRNA and protein in human fetal retinae, primary retinoblastomas, retinoblastoma cell lines and several independent orthotopic retinoblastoma xenografts. We found that MDM4 is the major p53 antagonist expressed in retinoblastoma and in the developing human retina. We also discovered that MDM4 protein steady state levels are much higher in retinoblastoma than in human fetal retinae. This increase would not have been predicted based on the mRNA levels. We explored several possible post-transcriptional mechanisms that may contribute to the elevated levels of MDM4 protein. A proportion of MDM4 transcripts are alternatively spliced to produce protein products that are reported to be more stable and oncogenic. We also discovered that a microRNA predicted to target MDM4 (miR191) was downregulated in retinoblastoma relative to human fetal retinae and a subset of samples had somatic mutations that eliminated the miR-191 binding site in the MDM4 mRNA. Taken together, these data suggest that post-transcriptional mechanisms may contribute to stabilization of the MDM4 protein in retinoblastoma.

Outcome after recurrence of childhood acute myeloid leukemia (AML) is poor. We performed this study to identify prognostic factors for recurrence and for survival after recurrence of AML.The clinical characteristics, biological features, treatment modalities, and outcomes of children with de novo AML who were enrolled on 3 consecutive clinical protocols from 1987 to 2002 at St. Jude Children's Research Hospital were studied. Regression modeling was used to identify prognostic factors for recurrence and for survival after recurrence.The outcome after recurrence was poor, with a 5‐year survival estimate of only 23.3% ± 5.7%. Multivariable analysis indicated that male sex (P = .005), autologous stem cell transplant before recurrence (P = .097), each additional month from diagnosis to recurrence (P = .041), and stem cell transplant after recurrence (P < .001) were associated with a better survival after recurrence, whereas M5 or M7 morphology (P = .001) were significantly predictive of a lower survival estimate after recurrence.Survival after recurrence was poor in children with AML. Novel therapies are urgently needed to prevent or to treat recurring AML. Cancer 2007. © 2006 American Cancer Society. [ABSTRACT FROM AUTHOR]

To compare the transcriptosome of early‐passage nonlesional dermal fibroblasts from systemic sclerosis (SSc) patients with diffuse disease and matched normal controls in order to gain further understanding of the gene activation patterns that occur in early disease. Total RNA was isolated from early‐passage fibroblasts obtained from nonlesional skin biopsy specimens from 21 patients with diffuse SSc (disease duration <5 years in all but 1) and 18 healthy controls who were matched to the cases by age (±5 years), sex, and race. Array experiments were performed on a 16,659‐oligonucleotide microarray utilizing a reference experimental design. Supervised methods were used to select differentially expressed genes. Quantitative polymerase chain reaction (PCR) was used to independently validate the array results.Of the 8,324 genes that passed filtering criteria, classification analysis revealed that <5% were differentially expressed between SSc and normal fibroblasts. Individually, differentially expressed genes included COL7A1, COL18A1 (endostatin), DAF, COMP, and VEGFB. Using the panel of genes discovered through classification analysis, a set of model predictors that achieved reasonably high predictive accuracy was developed. Analysis of 1,297 gene ontology (GO) classes revealed 35 classes that were significantly dysregulated in SSc fibroblasts. These GO classes included anchoring collagen (30934), extracellular matrix structural constituent (5201), and complement activation (6958, 6956). Validation by quantitative PCR demonstrated that 7 of 7 genes selected were concordant with the array results. Fibroblasts cultured from nonlesional skin of patients with SSc already have detectable abnormalities in a variety of genes and cellular processes, including those involved in extracellular matrix formation, fibrillogenesis, complement activation, and angiogenesis. [ABSTRACT FROM AUTHOR]