Search

Acute Lymphoblastic Leukemia (ALL) patients that are positive for minimal residual disease (MRD) after therapy or have leukemic infiltration into the central nervous system (CNS) are considered high risk and receive intensive chemotherapy regimens. Current methods to diagnose MRD and CNS infiltration rely on detecting leukemic cells in patient samples using pathology, flow cytometry, or next-generation sequencing. However, leukemic blasts may persist in the patient but not be physically present in bone marrow biopsy or biofluid sample, leading to inaccurate or delayed patient diagnosis. We have developed a nanopore sequencing workflow to detect B-ALL-associated cell-free DNA (cfDNA) in blood and cerebrospinal fluid (CSF) samples. Quantitation of B-cell specific VDJ recombination events in cfDNA samples defined B-ALL clonal heterogeneity. This workflow allowed us to track the response of individual B-ALL clones throughout treatment. Detection of cfDNA also predicted the clinical diagnosis of MRD and CNS disease. Importantly, we identified patients diagnosed as CNS negative who had low B-cell derived cell-free DNA levels in their CSF sample that correlated with B-cell clones present in the bone marrow. These data suggest that cfDNA assays may be useful in detecting the presence of ALL in the patient even when blasts are not in the biofluid sample. Nanopore analysis of cell-free DNA is a simple, rapid, and inexpensive assay that can serve as a valuable complement to traditional clinical diagnostic approaches for ALL.Abstract Figure

Chimeric antigen receptor T cells (CAR-T) are an effective and potentially durable treatment for refractory and multiply relapsed B-cell acute lymphoblastic leukemia. Neurotoxicity is frequent after CAR-T cell therapy. Mechanisms driving neurotoxicity are incompletely understood, and symptoms can range from transient and mild to severe and life-threatening. Providers have exercised caution in providing CAR-T to patients with neurological comorbidities or extramedullary disease. Here, we report three patients with prior significant neurologic morbidity who safely tolerated CAR-T cell infusion after bridging therapy with conventional chemotherapy.

Acute lymphoblastic leukemia (ALL) is the most common pediatric cancer, with a relapse rate of 15%. The presence of lymphoblasts in the central nervous system (CNS) is a negative prognostic indicator and a frequent site of disease relapse. CNS disease is defined as more than five white blood cells in the cerebrospinal fluid (CSF) with positive lymphoblast cytomorphology. Often leukemic blasts are detected by flow cytometry of the CSF prior to reaching the five white blood cell threshold, requiring repeat lumbar punctures at 2-4 week intervals until the arbitrarily defined clinical criteria for diagnosis are met. Additionally, more than 50% of autopsied brains from ALL patients show evidence of CNS disease, despite non-detectable CNS involvement via cytology of CSF. These data suggest that patients may be harboring CNS disease before it expands enough to be clinically diagnosed. We have developed a new assay that may allow for earlier detection of CNS disease and relapse by quantifying cell-free, circulating tumor DNA (ctDNA) in the CSF. ctDNA is a proven biomarker of relapse and metastasis in solid tumors in pre-clinical testing; however, its utility at predicting CNS disease in ALL has not been examined. We examined two possible methods for using ctDNA as a biomarker: leukemia cell clonality and DNA methylation profiling. We collected bone marrow aspirate, blood, and CSF samples from 11 newly diagnosed patients prior to the start of chemotherapy treatment to use as a training data set. ctDNA was isolated from blood and CSF samples at diagnosis and throughout treatment. Genomic DNA was isolated from bone marrow and peripheral blood mononuclear cell (PBMC) samples at diagnosis. For leukemia cell clonality assays, Invivoscribe Lymphotrack PCR assays combined with MinION (Oxford Nanopore Technologies) sequencing were used to identify the VDJ sequence of the immunoglobulin (B-ALL) or T-cell receptor (T-ALL) rearrangements in the clones comprising each leukemia. Throughout the course of treatment, ctDNA samples were run on the MinION sequencer, examining the abundance of major clones present and ctDNA quantification was done on patient plasma and CSF samples over time. While this assay relies on patient specific VDJ sequencing, we are also in the process of developing a more universal assay that utilizes recurrent methylation changes in T-ALL or B-ALL, compared to normal PBMCs. We performed methylation sequencing on 3 control PBMC samples and 7 ALL patient samples to identify differentially methylated regions (DMRs) present specifically in ALL samples. This sequencing identified 9,222 DMRs present in the ALL samples. These sites were then compared with publicly available datasets, yielding 55 overlapping regions and 19 specific overlapping methylation sites commonly present in ALL samples and not in PBMCs. We are now in the process of validating these differentially methylated sites by droplet digital polymerase chain reaction (ddPCR) to come up with a panel of biomarkers to track ALL disease over time. The end goal of our study is to develop an assay to rapidly detect relapse and CNS disease in ALL that is more sensitive and less invasive than the current clinical assays. Earlier detection of relapse and CNS disease will provide patients with more treatment options, which may ultimately improve patient outcome. Results will ultimately be correlated with patient response to therapy, the presence of CNS disease, and overall outcomes as determined by standard clinical diagnostic procedures. Disclosures No relevant conflicts of interest to declare.

10034 Background: Acute lymphoblastic leukemia (ALL) is the most common pediatric cancer, with a relatively high relapse rate, which is associated with poor prognosis. Currently, minimal residual disease (MRD) at the end of induction and consolidation therapy is the best predictor of patient relapse, however obtaining bone marrow aspirate is invasive and not always accurate. Another major concern in ALL is the presence of central nervous system (CNS) disease, which is often present long before clinical diagnosis can be made by flow cytometry. To circumvent these clinical challenges, we developed a new assay quantifying cell-free, circulating tumor (cfDNA) as a biomarker of disease progression, which can be correlated with MRD status as a predictor of relapse. cfDNA is frequently used to monitor progression of solid tumors, but pediatric leukemias lack common mutations that can be used to distinguish leukemic cfDNA from normal cfDNA. Methods: We examined two possible methods for using ctDNA as a biomarker: leukemia cell clonality and DNA methylation profiling. We developed a novel workflow for identifying VDJ rearrangements in leukemia cells and tracking their presence in cfDNA. We collected bone marrow, blood, and CSF samples from newly diagnosed patients, and cfDNA was isolated from blood and CSF samples throughout treatment. Invivoscribe Lymphotrack PCR assays combined with MinION (Oxford Nanopore Technologies) sequencing were used to identify the VDJ sequence of the immunoglobulin (B-ALL) or T-cell receptor (T-ALL) rearrangements of leukemic clones in genomic DNA. The MinION assay relies on patient-specific sequencing. We are also in the process of developing a universal assay that utilizes recurrent methylation changes in ALL to identify leukemic cfDNA in patient samples. Results: The MinION workflow was used to follow leukemic cfDNA throughout the course of treatment, and accurately identified MRD and CNS disease in patients. This workflow performed equivalent or better at detecting leukemic clones compared to MiSeq and droplet digital polymerase chain reaction (ddPCR), and is faster and less expensive than traditional Illumina sequencing. Methylation analysis of 865 ALL and 79 healthy samples yielded 55 regions and 19 specific methylation sites that were uniquely present in ALL samples. We are validating these sites by ddPCR to establish a panel of biomarkers to track ALL over time via cfDNA. Conclusions: The end goal of our study is provide a more sensitive and less invasive method for tracking MRD and CNS disease than current approaches. Results will ultimately be correlated with patient response to therapy, the presence of relapse or CNS disease, and overall outcomes determined by standard clinical diagnostic procedures.

Next-generation sequencing technologies will likely be used with increasing frequency in pediatric research. One consequence will be the increased identification of individual genomic research findings that are incidental to the aims of the research. Although researchers and ethicists have raised theoretical concerns about incidental findings in the context of genetic research, next-generation sequencing will make this once largely hypothetical concern an increasing reality. Most commentators have begun to accept the notion that there is some duty to disclose individual genetic research results to research subjects; however, the scope of that duty remains unclear. These issues are especially complicated in the pediatric setting, where subjects cannot currently but typically will eventually be able to make their own medical decisions at the age of adulthood. This article discusses the management of incidental findings in the context of pediatric genomic research. We provide an overview of the current literature and propose a framework to manage incidental findings in this unique context, based on what we believe is a limited responsibility to disclose. We hope this will be a useful source of guidance for investigators, institutional review boards, and bioethicists that anticipates the complicated ethical issues raised by advances in genomic technology.

The molecular mechanisms underlying the aggressive behavior of MYCN driven neuroblastoma (NBL) is under intense investigation; however, little is known about the impact of this family of transcription factors on the splicing program. Here we used high-throughput RNA sequencing to systematically study the expression of RNA isoforms in stage 4 MYCN-amplified NBL, an aggressive subtype of metastatic NBL. We show that MYCN-amplified NBL tumors display a distinct gene splicing pattern affecting multiple cancer hallmark functions. Six splicing factors displayed unique differential expression patterns in MYCN-amplified tumors and cell lines, and the binding motifs for some of these splicing factors are significantly enriched in differentially-spliced genes. Direct binding of MYCN to promoter regions of the splicing factors PTBP1 and HNRNPA1 detected by ChIP-seq demonstrates that MYCN controls the splicing pattern by direct regulation of the expression of these key splicing factors. Furthermore, high expression of PTBP1 and HNRNPA1 was significantly associated with poor overall survival of stage4 NBL patients (p ≤ 0.05). Knocking down PTBP1, HNRNPA1 and their downstream target PKM2, an isoform of pro-tumor-growth, result in repressed growth of NBL cells. Therefore, our study reveals a novel role of MYCN in controlling global splicing program through regulation of splicing factors in addition to its well-known role in the transcription program. These findings suggest a therapeutically potential to target the key splicing factors or gene isoforms in high-risk NBL with MYCN-amplification.

e18018Background: GDT produces higher response rates and improved survival compared to empiric therapy in some series of patients (pts) with refractory cancers. Limited data exists on the utility o...

Neuroblastoma (NB) is the most common solid extracranial tumor in children. Stage 4 NB is characterized with its clinical heterogynous outcome. The special category, stage 4S subgroup (2-5% of all NB) has a cure rate of above 90%. On the other hand, MYCN amplification (25-30% of all NB) is associated with poor outcome of 26% event free survival rate. Alternative splicing events have crucial roles in tumorigenesis and progression, and are nearly universal in human genes. However, biomarkers and drug designs typically focus on the main transcript of coding genes. High-throughput RNA sequencing (RNA-seq) experiments allow us to quantify all genes and their isoforms across samples. The purposes of this study are to systematically exam and compare the splicing programs in stage 4 NB subgroups, to evaluate the effects of splicing program, to identify key factors controlling the splicing program, and also to report isoform candidates that play important roles in dramatic different outcomes of advanced-stage NB subtypes, which represents potential diagnostic, prognostics and therapeutic targets for advanced-stage NB. In this study, we analyzed RNA-seq data of 29 stage 4 NB tumors and discovered an alternative splicing signature for MYCN amplified stage 4 tumors compared to MYCN non-amplified 4S tumors. In summary, 460 genes were alternatively spliced between MYCN non-amplified 4S and MYCN amplified stage 4 NB tumors. Pathway analysis with Ingenuity IPA demonstrated that alternatively spliced genes held enriched roles in cancer hallmarks biological functions including cell death, cell proliferation, apoptosis, cell movement and immune response. In particular, tissue-specific splicing factors such as RBFOX, CELF, and hnRNP families were differentially expressed between tumor subgroups. Regulatory motif sequences of these splicing factors were enriched in adjacent introns of differentially spliced exons. In addition, we used SHEP neuroblastoma Tet21N system and siRNA on IMR32 NB cell line to either induce or knockdown MYCN expression. RNA-seq analysis of MYCN induction and knockdown NB cell lines showed evidence of MYCN regulation of some of the splicing factors. This study systematically examined and compared the splicing programs in stage 4 NB subgroups. Moreover, it demonstrated that in addition to MYCN's well characterized role in transcriptional regulation, it regulates splicing events by controlling the expression of splicing factors. Citation Format: Shile Zhang, Tom Badgett, Jun Wei, Young K. Song, Dominik Bogen, Peter Johansson, Catherine Tolman, Jianbin He, Xinyu Wen, Zhihui Liu, Carol Thiele, Javed Khan. Massively parallel transcriptome sequencing in neuroblastoma reveals specific mRNA splicing patterns controlled by RNA binding splicing factors. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3161. doi:10.1158/1538-7445.AM2013-3161

Genetic alterations are thought to enable cancers to proliferate and survive more effectively or to resist cytotoxic therapies. However, the molecular basis of tumorigenesis and progression is not fully understood. In order to understand the genetic aberrations underlying tumor progression in refractory neuroblastoma, we performed next-generation sequencing on the whole exomes of three sequential tumor samples taken from a high-risk stage 4 neuroblastoma patient at diagnosis, after cytotoxic therapies, and at death, paired with a germline DNA sample from the same patient. Each genomic DNA sample was first partitioned using SureSelect whole human exome kits targeting ∼38Mb of genomic regions, then subjected to sequencing. Each exome sequencing experiment yielded approximately 240 million mappable reads representing 12 billion base pairs. We used DiBayes to identify single nucleotide variants (SNVs). After filtering away the low-quality SNVs, we identified approximately 18,000 high-quality SNVs in each sample, of which ∼7500 are within the protein coding regions. Using the frequencies of the SNVs derived from the sequencing experiments, we discovered a shared LOH signature at the chromosome level among all three tumors indicating they originate from a common cancer progenitor. When we examined the germline DNA, we found about 97% of SNVs for each tumor sample were also found in the germline DNA, and each tumor had more than 300 unique somatic SNVs. These data showed that despite of the shared LOH pattern, tumor genomes are highly unstable at the single nucleotide level. Furthermore, we also found that the three tumor samples shared 64 SNVs which may be potential driver mutations. Twenty of those common 64 SNVs, located on 7 genes, are predicted to be damaging mutations by SIFT algorithm. These mutations are currently being validated using Sanger sequencing, and further RNA-seq in the same tumor samples to investigate for pathway disruption that may lead to chemo-refractory disease. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 4833. doi:10.1158/1538-7445.AM2011-4833

Introduction: Neuroblastoma (NB) is one of the small round blue cell tumors of childhood arising in the peripheral nervous system. Fifty percent of patients present with metastatic or high risk disease and despite aggressive multimodal therapy approximately 60% percent of these patients eventually succumb to NB. Currently only a handful of molecular alterations are known to influence prognosis, but no clear mechanism of pathogenesis has been demonstrated. Recently, targeted sequencing efforts identified recurrent single nucleotide variants (SNVs) in the ALK gene in 6.1-15% of NB. In an effort to better understand the biology driving this tumor and to identify new targets for therapy, we sequenced the transcriptomes of 20 stage four tumors, including ten MYCN amplified and ten MYCN non-amplified samples, using massively parallel sequencing technology. Methods: In our analysis pipeline, 50 base nucleotide reads are filtered against a database of rRNA, tRNA, repetitive regions, and adapter sequences. The remaining reads were aligned to the reference human genome (hg 18) and a database of spice junctions. Reads that align were analyzed for: 1) base coverage, 2) transcript expression levels, 3) calling SNVs and 4) determination of which of the SNVs were damaging by Sorting Intolerant From Tolerant (SIFT) analysis Results: Initial analysis of the first six samples, yielded an average of 86.6 million uniquely mapped reads per sample. On average we detected the expression of 6,000 genes to a depth of 10x. The RNA seq expression profile correlated well with expression array data from the same sample (r=0.62), while the sequencing data identified an additional 3,000 genes that were not detected by microarray. Our sequencing method maintains “strandedness” allowing the identification of multiple novel antisense non-coding transcripts. Using the SAMtool, an average of 1,255 nonsynonymous SNVs predicted per sample. Of these nonsynonymous SNVs, 69-160 per sample were predicted by the SIFT algorithm to be damaging. Interestingly, 3 different genes had damaging nonsynonymous SNVs in at least 30% of the samples. Conclusion: Next generation sequencing of transcriptome is a powerful and more sensitive method than microarrays for expression profiling and allows for the identification of novel transcripts including non-coding RNAs. Here we report the most extensive profiling of the neuroblastoma transcriptome to date using this technology. We identified several hundred protein disrupting SNVs, and of these 3 were commonly altered. Ongoing analysis is underway to validate our results. The identification of recurrent genetic alterations in NB will assist in developing a better understanding of the mechanisms of pathogenesis of this pediatric cancer and lead to new therapeutic targets. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 1137.

Accumulation of genetic and epigenetic alterations is thought to enable cancers to proliferate and survive more effectively than their neighboring cells, or for aggressive tumors to resist cytotoxic therapies. However, the molecular basis of tumorigenesis and chemoresistance is not fully understood. With the advent of next-generation sequencing technologies, we can determine the whole transcriptome or genome sequence of a tumor and thereby detect all the changes that may result in a particular phenotype such as resistance to cytotoxic therapies. To understand the genetic aberrations underlying the chemoresistance in neuroblastoma, we attempted to catalogue all genetic aberrations in two tumor samples before and after cytotoxic therapies using RNA-seq (including whole transcriptome and miRNA) and genomic DNA sequencing from the same neuroblastoma patient who was refractory to standard chemotherapies. First, we performed whole genome sequencing on patient's constitutional DNA to establish a genetic background. Multiple mate-paired libraries were constructed and subjected to next-generation sequencing. We acquired 53 Gb high-quality mappable data, equivalent to 16X coverage of this patient's genome. Next, whole transcriptome sequencing of the two tumors yielded total 345 million 50bp mappable reads representing 14,128 and 14,484 genes (20875 and 21330 transcripts) respectively. The average coverage for each transcript was >14, while 5297 genes for one sample and 6912 genes for the other had an average coverage of >10. With such deep sequencing, single nucleotide variant (SNV) analysis on the transcriptome sequencing data identified 9103 and 10556 exonic SNVs of total 154871 and 191033 SNVs respectively in two samples. Among the exonic SNVs, 3965 and 4470 SNVs were in the coding regions and about half of them were nonsynonymous. Using SIFT analysis, we predicted 353 and 443 SNVs to be damaging before and after chemotherapies. Of them, 165 SNVs were shared among two samples. Therefore, these shared SNVs and newly acquired SNVs after chemotherapies may contain the genetic alterations that are resistant to the standard neuroblastoma therapies in this patient. In addition, 188 SNVs were present in the original tumor but not in the refractory sample indicating genomic instability of tumors. In parallel, we performed microRNA sequencing on the two samples. While most of 987 mature microRNAs were detected at similar levels after normalized against total reads, 64 microRNAs were significantly differentially expressed before and after chemotherapies using stringent selection criteria (total read counts in two samples>500, >5 fold change, P Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 2212.