Olena Morozova

Molecular Profiling Reveals Biologically Discrete Subsets and Pathways of Progression in Diffuse Glioma

Therapy development for adult diffuse glioma is hindered by incomplete knowledge of somatic glioma driving alterations and suboptimal disease classification. We defined the complete set of genes associated with 1,122 diffuse grade II-III-IV gliomas from The Cancer Genome Atlas and used molecular profiles to improve disease classification, identify molecular correlations, and provide insights into the progression from low- to high-grade disease. Whole-genome sequencing data analysis determined that ATRX but not TERT promoter mutations are associated with increased telomere length. Recent advances in glioma classification based on IDH mutation and 1p/19q co-deletion status were recapitulated through analysis of DNA methylation profiles, which identified clinically relevant molecular subsets. A subtype of IDH mutant glioma was associated with DNA demethylation and poor outcome; a group of IDH-wild-type diffuse glioma showed molecular similarity to pilocytic astrocytoma and relatively favorable survival. Understanding of cohesive disease groups may aid improved clinical outcomes.

Comprehensive Pan-Genomic Characterization of Adrenocortical Carcinoma

We describe a comprehensive genomic characterization of adrenocortical carcinoma (ACC). Using this dataset, we expand the catalogue of known ACC driver genes to include PRKAR1A, RPL22, TERF2, CCNE1, and NF1. Genome wide DNA copy-number analysis revealed frequent occurrence of massive DNA loss followed by whole-genome doubling (WGD), which was associated with aggressive clinical course, suggesting WGD is a hallmark of disease progression. Corroborating this hypothesis were increased TERT expression, decreased telomere length, and activation of cell-cycle programs. Integrated subtype analysis identified three ACC subtypes with distinct clinical outcome and molecular alterations which could be captured by a 68-CpG probe DNA-methylation signature, proposing a strategy for clinical stratification of patients based on molecular markers.

Penetrance of biallelic SMARCAL1 mutations is associated with environmental and genetic disturbances of gene expression

Biallelic mutations of the DNA annealing helicase SMARCAL1 (SWI/SNF-related, matrix-associated, actin-dependent regulator of chromatin, subfamily a-like 1) cause Schimke immuno-osseous dysplasia (SIOD, MIM 242900), an incompletely penetrant autosomal recessive disorder. Using human, Drosophila and mouse models, we show that the proteins encoded by SMARCAL1 orthologs localize to transcriptionally active chromatin and modulate gene expression. We also show that, as found in SIOD patients, deficiency of the SMARCAL1 orthologs alone is insufficient to cause disease in fruit flies and mice, although such deficiency causes modest diffuse alterations in gene expression. Rather, disease manifests when SMARCAL1 deficiency interacts with genetic and environmental factors that further alter gene expression. We conclude that the SMARCAL1 annealing helicase buffers fluctuations in gene expression and that alterations in gene expression contribute to the penetrance of SIOD.

DNA Hypermethylation and 1p Loss Silence NHE-1 in Oligodendroglioma

Oligodendroglioma is characterized by mutations of IDH and CIC, 1p/19q loss, and slow growth. We found that NHE‐1 on 1p is silenced in oligodendrogliomas secondary to IDH‐associated hypermethylation and 1p allelic loss. Silencing lowers intracellular pH and attenuates acid load recovery in oligodendroglioma cells. Others have shown that rapid tumor growth cannot occur without NHE‐1–mediated neutralization of the acidosis generated by the Warburg glycolytic shift. Our findings show for the first time that the pH regulator NHE‐1 can be silenced in a human cancer and also suggest that pH deregulation may contribute to the distinctive biology of human oligodendroglioma. Ann Neurol 2012;71:845–849

TumorMap: Exploring the Molecular Similarities of Cancer Samples in an Interactive Portal

Vast amounts of molecular data are being collected on tumor samples, which provide unique opportunities for discovering trends within and between cancer subtypes. Such cross-cancer analyses require computational methods that enable intuitive and interactive browsing of thousands of samples based on their molecular similarity. We created a portal called TumorMap to assist in exploration and statistical interrogation of high-dimensional complex omics data in an interactive and easily interpretable way. In the TumorMap, samples are arranged on a hexagonal grid based on their similarity to one another in the original genomic space and are rendered with Googles Map technology. While the important feature of this public portal is the ability for the users to build maps from their own data, we pre-built genomic maps from several previously published projects. We demonstrate the utility of this portal by presenting results obtained from The Cancer Genome Atlas project data. Cancer Res; 77(21); e111-4. ©2017 AACR.

A Distinct DNA Methylation Shift in a Subset of Glioma CpG Island Methylator Phenotypes during Tumor Recurrence

Glioma diagnosis is based on histomorphology and grading; however, such classification does not have predictive clinical outcome after glioblastomas have developed. To date, no bona fide biomarkers that significantly translate into a survival benefit to glioblastoma patients have been identified. We previously reported that the IDH mutant G-CIMP-high subtype would be a predecessor to the G-CIMP-low subtype. Here, we performed a comprehensive DNA methylation longitudinal analysis of diffuse gliomas from 77 patients (200 tumors) to enlighten thexa0epigenome-based malignant transformation of initially lower-grade gliomas. Intra-subtype heterogeneity among G-CIMP-high primary tumors allowed us to identify predictive biomarkers for assessing the risk of malignant recurrence at early stages of disease. G-CIMP-low recurrence appeared in 9.5% of all gliomas, and these resembled IDH-wild-type primary glioblastoma. G-CIMP-low recurrence can be characterized by distinct epigenetic changes at candidate functional tissue enhancers with AP-1/SOX binding elements, mesenchymal stem cell-like epigenomic phenotype, and genomic instability. Molecular abnormalities of longitudinal G-CIMP offer possibilities to defy glioblastoma progression.

Comparative RNA-Sequencing Analysis Benefits a Pediatric Patient With Relapsed Cancer

Clinical detection of sequence and structural variants in known cancer genes points to viable treatment options for a minority of children with cancer.1 To increase the number of children who benefit from genomic profiling, gene expression information must be considered alongside mutations.2,3 Although high expression has been used to nominate drug targets for pediatric cancers,4,5 its utility has not been evaluated in a systematic way.6 We describe a child with a rare sarcoma that was profiled with whole-genome and RNA sequencing (RNA-Seq) techniques. Although the tumor did not harbor DNA mutations targetable by available therapies, incorporation of gene expression information derived from RNA-Seq analysis led to a therapy that produced a significant clinical response. We use this case to describe a framework for inclusion of gene expression into the clinical genomic evaluation of pediatric tumors.

Abstract 2466: Identifying confidently measured genes in single pediatric cancer patient samples using RNA sequencing

In the UC Santa Cruz Treehouse Childhood Cancer Initiative (treehousegenomics.soe.ucsc.edu), we are exploring the utility of using RNA-Seq analysis of tumor samples from children to identify potential novel therapeutic options for each individual. Within a single RNA-Seq data set, the gene expression measurements are not equally accurate. The identification of activated, druggable pathways requires accurate gene-level expression measurements. We receive samples from a variety of clinical and research settings, and the quantity and complexity of the available input material and the depth of sequencing differ. These factors inspired us to develop a tool that will allow us to identify accurate measurements in most RNA-Seq samples we receive. First, we characterized the relationship between depth of sequencing and the accuracy of the gene expression measurement. We analyzed subsets of reads in samples with more than 50 million Uniquely Mapped, Exonic, Non-duplicate (UMEND) reads. UMEND reads typically constitute over 80% of the reads in a high quality experiment with sufficient starting material. We compared gene expression across the subsets of reads to calculate how many UMEND reads are required to produce consistent measurements. We found that, on average, genes expressed at 1-5 TPM in our data require 30 million reads to be accurately measured. For this calculation, we define accuracy as the condition in which 75% of genes are measured to within 25% of the true value. Secondly, we use these known relationships to identify genes that have been accurately measured in our tumor RNA-Seq samples. For a sample with 15 million UMEND reads, we find that genes expressed above 5 TPM can be accurately measured and are retained. In the first twelve samples analyzed, samples with more than 10 million UMEND reads retained at least 46% of the genes expressed above zero. We exclude as references those samples with fewer than 10 million UMEND reads due to the marked gene loss after thresholding for this group. Using accurately measured genes allows us to more confidently assess similarity to other samples, identify enriched pathways, and confirm the expression of drug targets and related molecules under consideration. For example, we reconsidered the CDK4 inhibitor Palbociclib in one patient because the expression of RB1, downstream effector required for Palbociclib-mediated tumor cell death, was under our accuracy threshold. Accuracy thresholds can also be used in experiment planning. Accuracy thresholding allows us to better assess the value of an RNA-Seq data set and, if necessary, identify the subset of genes whose expression can be confidently considered in a clinical setting. Our experience points to the importance of careful quality control in this process. Citation Format: Holly Beale, Du Linh Lam, John Vivian, Yulia Newton, Avanthi Tayi Shah, Isabel Bjork, Ted Goldstein, Angela N. Brooks, Josh Stuart, Sofie Salama, E. Alejandro Sweet-Cordero, David Haussler1, Olena Morozova. Identifying confidently measured genes in single pediatric cancer patient samples using RNA sequencing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 2466. doi:10.1158/1538-7445.AM2017-2466

Abstract LB-338: A critical evaluation of genomic data sharing: Barriers to accessing pediatric cancer genomic datasets: a Treehouse Childhood Cancer Initiative experience

Genomic data sharing is increasingly recognized as critical to genomic research. The need is acute in pediatric cancer research due to the rarity of pediatric tumor types and paucity of pediatric cancer data, and in translational research to assess the impact of genomic research on human health. However, genomic data sharing is hindered by an absence of standards regarding timing, patient privacy, use agreement standards, and data characterization and quality. At UC Santa Cruz Treehouse Childhood Cancer Initiative (treehousegenomics.soe.ucsc.edu), we examine individual pediatric cancer tumor RNA sequencing profiles against a database of over 11,000 tumor RNA sequencing profiles from public genomic datasets such as The Cancer Genome Atlas, Therapeutically Applicable Research To Generate Effective Treatments, International Cancer Genome Consortium, and Medulloblastoma Advanced Genomics International, and pediatric cancer clinical trials with which we partner, such as those at Stanford University, UC San Francisco, Children’s Hospital of Orange County, and British Columbia Children’s Hospital. For over 18 months, we have worked systematically to enhance the Treehouse dataset by adding pediatric cancer data and presently underrepresented tumor types. The NIH and other leading funding agencies now regularly require grantees to make genomic data generated available to the research community, either post-publication or after an embargo period. We have combed websites and public repositories, searched PubMed, and contacted researchers directly. Finding data requires a mining of literature, often with limited information, and initiating the many different processes for requesting permission for these datasets, with different and often cumbersome data use obligations. The combination of cryptically named datasets, multiple data types and the practice of grouping datasets from multiple papers under a single study accession makes zeroing in on the correct dataset challenging. Downloading the genomic data is time-consuming, such that a dataset of under a 100 files can take up to a week to download under optimal conditions. Matching metadata is inconsistently available, often vague, sparse or error ridden. Only after months of identifying, permissioning for use, committing to use- and sharing-restricting terms, and downloading the genomic and metadata, is it possible to assess the quality, often discovering that data quality is low. We evaluate the barriers to data sharing based on the Treehouse experience and offer guidelines for timing, use agreement standards, and data characterization and quality, to enhance data sharing and outcomes for all pediatric cancer patients. Citation Format: Katrina Learned, Ann Durbin, Robert Currie, Holly Beale, Du Linh Lam, Theodore Goldstein, Sofie R. Salama, David Haussler, Olena Morozova, Isabel Bjork. A critical evaluation of genomic data sharing: Barriers to accessing pediatric cancer genomic datasets: a Treehouse Childhood Cancer Initiative experience [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr LB-338. doi:10.1158/1538-7445.AM2017-LB-338

Abstract 4890: A pan-cancer analysis framework for incorporating gene expression information into clinical interpretation of pediatric cancer genomic data

Genomic characterization used in pediatric cancer clinical trials is limited to the detection of somatic mutations and gene fusions in well-characterized cancer genes. However, these approaches do not reveal actionable therapeutic targets for the majority of pediatric cancer patients. Incorporation of gene expression information into clinical genomic analysis is hindered by the lack of appropriate computational methods, designed for single patients rather than patient cohorts. UC Santa Cruz Treehouse Childhood Cancer Initiative (treehousegenomics.soe.ucsc.edu) enables the incorporation of gene expression information into the genomic evaluation of pediatric cancer patients. We leverage large cancer RNA sequencing datasets, including The Cancer Genome Atlas, Therapeutically Applicable Research to Generate Effective Treatments, Medulloblastoma Advanced Genomics International Consortium, International Cancer Genome Consortium, and published research and clinical studies. Through our “pan-cancer analysis”, we compare each prospective tumor’s RNA sequencing and/or mutational profile to over 11,000 uniformly analyzed tumor profiles using our Tumor Map method. Tumor Map visualizes single tumors together with the reference compendium and identifies samples that are most similar to the given tumor based on the gene expression profiles. We also developed a gene expression outlier analysis to identify transcripts that are over expressed in the given tumor. These pan-cancer gene expression analyses are used in conjunction with mutation data to nominate molecular pathways that may be driving the disease in each child, providing useful information to the medical teams. We aim to evaluate this approach in partnership with pediatric cancer clinical genomic trials at Stanford University, UC San Francisco, Children’s Hospital of Orange County, University of Michigan, Children’s Mercy Hospital, and British Columbia Children’s Hospital. The analysis of the first 27 patients at Stanford, most with refractory solid tumors, provided evidence of the potential clinical utility of incorporating gene expression information into the genomic evaluation of pediatric cancer patients. In all cases, we identified candidate driver molecular pathways that could be targeted by existing FDA-approved therapies or therapies available through a clinical trial. The most frequently identified molecular targets were receptor tyrosine kinases and cyclin-dependent kinases. For 3 patients with no treatment options prior to our work, the analysis contributed to treatment decisions. This study provides a framework for incorporating gene expression information into the clinical interpretation of pediatric cancer genomic data. We underscore the importance of releasing the data to the community immediately following generation, so that they may benefit new patients. Citation Format: Olena Morozova, Yulia Newton, Avanthi Tayi Shah, Holly Beale, Du Linh Lam, John Vivian, Isabel Bjork, Theodore Goldstein, Josh Stuart, Sofie Salama, E. Alejandro Sweet-Cordero, David Haussler. A pan-cancer analysis framework for incorporating gene expression information into clinical interpretation of pediatric cancer genomic data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4890. doi:10.1158/1538-7445.AM2017-4890