Copy Number Variation Profiles Research Articles

Abstract P3-01-09: Tracing clonal evolution of circulating tumor cells and cell-free DNA in a metastatic breast cancer patient

Abstract Background: Copy number variation analysis has been shown as a valuable tool to infer clonal evolution of tumor cells. While bulk genomic analysis of FFPE tissue can provide information about the characteristics of a primary tumor or a metastatic nodule, these biopsies are rarely repeated due to their invasive nature. Liquid biopsies on the other hand, provide minimally invasive access to tumor derived cells and DNA and can therefore be used to monitor genomic changes of the tumor longitudinally and in real time. Here we evaluate the kinetics of liquid biopsy biomarkers and monitor genomic changes in single circulating tumor cells (CTCs) and cell-free DNA (cfDNA) from 17 longitudinal blood draws collected from a metastatic breast cancer patient over seven treatment regimes. FFPE tissue from the primary tumor and two metastatic sites in liver and bone are available for comparative analysis. Methods: We used the high definition – single cell assay (HD-SCA) platform to detect and enumerate CTCs of the blood of one metastatic breast cancer patient followed for 4 years. CTCs were characterized morphometrically and single CTCs of each draw were sequenced for copy number variation (CNV). cfDNA was extracted from plasma and tissue samples of the primary breast, bone and liver metastasis were micro-dissected. Extracted DNA from tissue samples and plasma underwent CNV analysis and were compared with CNV data of CTCs. Results: CTCs were identified in 14/17 draws and varied from 0 to >3000 cells/ml, spiking at two different times concordant with clinical progression. Changes in cell free tumor DNA fraction (ctDNA) correlated strongly with CTC count. Comparative CNV analysis of CTCs, cfDNA and FFPE tissue confirmed the tumorous origin of both cells and cfDNA. Single cell CNV analysis found four genomically related sub-clones in the patient's blood, three of which were present at time of enrollment, while the fourth clone emerged 3.5 years later concordant with the decline in the patient's health. Genomic copy number losses included tumor suppressor genes such as ARID1A, FOXP1, ATM, BRCA2, RB1, CDH11 and CDH1. Genomic gains were limited but included the proto-oncogenes BCL9, ABL2 and MDM4. It is striking that the sub-clonal structure persisted nearly unchanged for 3.5 years throughout various hormone and cytotoxic treatments, indicating high genomic stability. Despite their genomic differences, cells from the different clones are indistinguishable by immunofluorescence or morphometric analysis. CNV profiles of cfDNA reflect the most abundant CTC clone at each time-point. Conclusion: Using the HD-SCA assay cellular (CTC), and soluble (cfDNA) fractions of liquid biopsies can be compared at the molecular level over a long time course, and verified by comparison with tissue biopsy. This work confirms that liquid biopsies can be used to access the molecular state of a patient's cancer in near real-time and provide insight into the response to treatment years after the original characterization of the primary tissue. Here we show in a single patient example the power of single CTC analysis and highlight how CNV analysis of liquid biopsies could serve as minimally invasive tool to monitor tumor evolution longitudinally. Citation Format: Welter L, Xu L, McKinley D, Dago AE, Restrepo-Vassalli S, Rodriguez Lee M, Kolatkar A, Nieva J, Hicks J, Kuhn P. Tracing clonal evolution of circulating tumor cells and cell-free DNA in a metastatic breast cancer patient [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P3-01-09.

Protein Disulfide Isomerases Are Promising Targets for Predicting the Survival and Tumor Progression in Glioma Patients

Background: Protein disulfide isomerase (PDI) family mediates various biological functions. However, the gene family exert different effects on tumorigenesis and glioma development. The present study focused on the expression patterns, prognostic values and potential mechanism of the PDI family in gliomas. Methods: Construction of the PDI signature, cluster and risk score models of glioma was done using GSVA, consensus clustering analysis, and LASSO Cox regression analysis respectively. Analysis of the distinctive genomic alterations in gliomas was achieved using published data of somatic mutation and copy number variation profiles. GSVA analysis revealed the underlying mechanisms of PDIs in gliomas. A novel clinical prognosis model was established via nomogram and ROC. qRT-PCR and immunohistochemistry validated the mRNA and protein expression patterns of the PDI genes respectively. Results: Most PDI family members' mRNA expressions were observed significantly different between gliomas classified by clinical features. High values of PDI signature/ risk score and cluster 1 in gliomas were associated with malignant clinicopathological characteristics and poor prognosis. Genomic alterations analysis revealed that many cases having high PDI signature and risk score were associated with genomic aberrations of driver oncogenes. GSVA analysis showed that PDI family was involved in many signaling pathways in ERAD, apoptosis, and MHC class I among many more. Prognostic nomogram revealed that the risk score was a good prognosis indicator for gliomas. The qRT-PCR and immunohistochemistry confirmed that P4HB, PDIA4 and PDIA5 were overexpressed in gliomas. Conclusions: This research highlighted the clinical importance of PDI family in tumorigenesis and progression in gliomas. Funding Statement: This research was supported by the National Natural Science Foundation of China (NO.81703622), China Postdoctoral Science Foundation (NO. 2018M633002), Hunan Provincial Natural Science Foundation of China (NO.2018JJ3838), Foundation of Health Commission of Hunan Province of China (C2019186), and the Key Project of The Second People’s Hospital of Hunan Province (2018A06). Declaration of Interests: The authors declare that there are no potential conflicts of interest. Ethics Approval Statement: Not required.

Integration of genomic copy number variations and chemotherapy-response biomarkers in pediatric sarcoma

BackgroundWhile most pediatric sarcomas respond to front-line therapy, some bone sarcomas do not show radiographic response like soft-tissue sarcomas (rhabdomyosarccomas) but do show 90% necrosis. Though, new therapies are urgently needed to improve survival and quality of life in pediatric patients with sarcomas. Complex chromosomal aberrations such as amplifications and deletions of DNA sequences are frequently observed in pediatric sarcomas. Evaluation of copy number variations (CNVs) associated with pediatric sarcoma patients at the time of diagnosis or following therapy offers an opportunity to assess dysregulated molecular targets and signaling pathways that may drive sarcoma development, progression, or relapse. The objective of this study was to utilize publicly available data sets to identify potential predictive biomarkers of chemotherapeutic response in pediatric Osteosarcoma (OS), Rhabdomyosarcoma (RMS) and Ewing’s Sarcoma Family of Tumors (ESFTs) based on CNVs following chemotherapy (OS n = 117, RMS n = 64, ESFTs n = 25 tumor biopsies).MethodsThere were 206 CNV profiles derived from pediatric sarcoma biopsies collected from the public databases TARGET and NCBI-Gene Expression Omnibus (GEO). Through our comparative genomic analyses of OS, RMS, and ESFTs and 22,255 healthy individuals called from the Database of Genomic Variants (DGV), we identified CNVs (amplifications and deletions) pattern of genomic instability in these pediatric sarcomas. By integrating CNVs of Cancer Cell Line Encyclopedia (CCLE) identified in the pool of genes with drug-response data from sarcoma cell lines (n = 27) from Cancer Therapeutics Response Portal (CTRP) Version 2, potential predictive biomarkers of therapeutic response were identified.ResultsGenes associated with survival and/recurrence of these sarcomas with statistical significance were found on long arm of chromosome 8 and smaller aberrations were also identified at chromosomes 1q, 12q and x in OS, RMS, and ESFTs. A pool of 63 genes that harbored amplifications and/or deletions were frequently associated with recurrence across OS, RMS, and ESFTs. Correlation analysis of CNVs from CCLE with drug-response data of CTRP in 27 sarcoma cell lines, 33 CNVs out of 63 genes correlated with either sensitivity or resistance to 17 chemotherapies from which actionable CNV signatures such as IGF1R, MYC, MAPK1, ATF1, and MDM2 were identified. These CNV signatures could potentially be used to delineate patient populations that will respond versus those that will not respond to a particular chemotherapy.ConclusionsThe large-scale analyses of CNV-drug screening provides a platform to evaluate genetic alterations across aggressive pediatric sarcomas. Additionally, this study provides novel insights into the potential utilization of CNVs as not only prognostic but also as predictive biomarkers of therapeutic response. Information obtained in this study may help guide and prioritize patient-specific therapeutic options in pediatric bone and soft-tissue sarcomas.

Clinical impact of chromothriptic complex chromosomal rearrangements in newly diagnosed multiple myeloma

Complex Chromosomal Rearrangements (CCRs) are increasingly being reported as genetic risk factors of clinical significance in cancer owing to their identification using high resolution whole genome profiling technologies. This study employed high resolution CGH + SNP microarrays for whole genome copy number variations (CNV) profiling and identified CCRs in 11/107(10%) newly diagnosed Multiple Myeloma (MM) patients. Six patients exhibited Chromothripsis (CTH) among seven chromosomes that were confirmed with automated CTLPscanner web tool and; five cases displayed chromoplexy (CPL) which involved multiple chromosomes. Presence of chromothripsis in chromosome 17 in three out of six patients indicate a link between TP53 aberrations and incidence of CTH. Multivariable Cox regression model demonstrated a significant association of CTH with poor PFS (HR = 3.09, p = 0.010) and OS (HR = 3.31, p = 0.024) which suggests that CTH is an additional independent prognostic marker in multiple myeloma. Addition of CTH in risk stratification models in clinical setting in multiple myeloma may help in upfront identification of high risk patients for suitable customized therapy.

Copy number variation profile in noninvasive prenatal testing (NIPT) can identify co-existing maternal malignancies: Case reports and a literature review

ObjectiveThe coexistence of maternal malignancy and pregnancy has received increasing attention in Noninvasive prenatal testing (NIPT) studies. Malignancy in pregnant women potentially affects the copy number variation (CNV) profile in NIPT results. Only one case of hematologic cancer has been reported in a Hong-Kong pregnant women, and solid tumors have never been reported in pregnant Chinese women. Case reportThe patients with dysgerminoma and cervical cancer showed aberrant chromosomal aneuploidies in NIPT and concordant patterns of genome disruption in tumor tissues. The genomic aberrations in the gastric cancer patient had similar copy number variation pattern of gastric cancer. ConclusionThe findings in this study and the literature review further validate the effect of maternal malignancy on the copy number variation profile in NIPT data and strengthen the possibility of detecting malignant tumors with NIPT in the future.

CODEX2: full-spectrum copy number variation detection by high-throughput DNA sequencing

High-throughput DNA sequencing enables detection of copy number variations (CNVs) on the genome-wide scale with finer resolution compared to array-based methods but suffers from biases and artifacts that lead to false discoveries and low sensitivity. We describe CODEX2, as a statistical framework for full-spectrum CNV profiling that is sensitive for variants with both common and rare population frequencies and that is applicable to study designs with and without negative control samples. We demonstrate and evaluate CODEX2 on whole-exome and targeted sequencing data, where biases are the most prominent. CODEX2 outperforms existing methods and, in particular, significantly improves sensitivity for common CNVs.

Cell-Free DNA From Metastatic Pancreatic Neuroendocrine Tumor Patients Contains Tumor-Specific Mutations and Copy Number Variations.

Background: Detection of tumor-specific alterations in cell-free DNA (cfDNA) has proven valuable as a liquid biopsy for several types of cancer. So far, use of cfDNA remains unexplored for pancreatic neuroendocrine tumor (PNET) patients.Methods: From 10 PNET patients, fresh frozen tumor tissue, buffy coat and plasma samples were collected. Whole-exome sequencing of primary tumor and germline DNA was performed to identify tumor-specific variants and copy number variations (CNVs). Subsequently, tumor-specific variants were quantified in plasma cfDNA with droplet digital PCR. In addition, CNV analysis of cfDNA was performed using shallow whole-genome sequencing.Results: Tumor-specific variants were detected in perioperative plasma samples of two PNET patients, at variant allele fractions (VAFs) of respectively 19 and 21%. Both patients had metastatic disease at time of surgery, while the other patients presented with localized disease. In the metastatic patients, CNV profiles of tumor tissue and cfDNA were significantly correlated. A follow-up plasma sample of a metastatic patient demonstrated an increased VAF (57%) and an increased chromosomal instability, in parallel with an increase in tumor burden.Conclusions: We are the first to report the presence of tumor-specific genetic alterations in cfDNA of metastatic PNET patients and their evolution during disease progression. Additionally, CNV analysis in cfDNA shows potential as a liquid biopsy.

Deletion of exon 4 in LAMA2 is the most frequent mutation in Chinese patients with laminin \u03b12-related muscular dystrophy

Although recessive mutations in LAMA2 are already known to cause laminin α2-related muscular dystrophy, a rare neuromuscular disorder, large deletions or duplications within this gene are not well-characterized. In this study, we applied next-generation sequencing-based copy number variation profiling in 114 individuals clinically diagnosed with laminin α2-related muscular dystrophy, including 96 who harboured LAMA2 mutations and 34 who harboured intragenic rearrangements. In total, we detected 18 distinct LAMA2 copy number variations that have been reported only among Chinese, 10 of which are novel. The frequency of CNVs in the cohort was 19.3%. Deletion of exon 4 was detected in 10 alleles of eight patients, accounting for 27% of all copy number variations. These patients are Han Chinese and were found to have the same haplotype and sequence at the breakpoint junction, suggesting that exon 4 deletion is a founder mutation in Chinese Han and a mutation hotspot. Moreover, the data highlight our approach, a modified next-generation sequencing assay, as a robust and sensitive tool to detect LAMA2 variants; the assay identifies 85.7% of breakpoint junctions directly alongside sequence information. The method can be applied to clinical samples to determine causal variants underlying various Mendelian disorders.

Pan-Cancer Analysis Reveals Differential Susceptibility of Bidirectional Gene Promoters to DNA Methylation, Somatic Mutations, and Copy Number Alterations.

Bidirectional gene promoters affect the transcription of two genes, leading to the hypothesis that they should exhibit protection against genetic or epigenetic changes in cancer. Therefore, they provide an excellent opportunity to learn about promoter susceptibility to somatic alteration in tumors. We tested this hypothesis using data from genome-scale DNA methylation (14 cancer types), simple somatic mutation (10 cancer types), and copy number variation profiling (14 cancer types). For DNA methylation, the difference in rank differential methylation between tumor and tumor-adjacent normal matched samples based on promoter type was tested by the Wilcoxon rank sum test. Logistic regression was used to compare differences in simple somatic mutations. For copy number alteration, a mixed effects logistic regression model was used. The change in methylation between non-diseased tissues and their tumor counterparts was significantly greater in single compared to bidirectional promoters across all 14 cancer types examined. Similarly, the extent of copy number alteration was greater in single gene compared to bidirectional promoters for all 14 cancer types. Furthermore, among 10 cancer types with available simple somatic mutation data, bidirectional promoters were slightly more susceptible. These results suggest that selective pressures related with specific functional impacts during carcinogenesis drive the susceptibility of promoter regions to somatic alteration.

A population genomic characterization of copy number variation in the opportunistic fungal pathogen Aspergillus fumigatus.

Aspergillus fumigatus is a potentially deadly opportunistic fungal pathogen. Molecular studies have shaped our understanding of the genes, proteins, and molecules that contribute to A. fumigatus pathogenicity, but few studies have characterized genome-wide patterns of genetic variation at the population level. Of A. fumigatus genomic studies to-date, most focus mainly on single nucleotide polymorphisms and large structural variants, while overlooking the contribution of copy number variation (CNV). CNV is a class of small structural variation defined as loci that vary in their number of copies between individuals due to duplication, gain, or deletion. CNV can influence phenotype, including fungal virulence. In the present study, we characterized the population genomic patterns of CNV in a diverse collection of 71 A. fumigatus isolates using publicly available sequencing data. We used genome-wide single nucleotide polymorphisms to infer the population structure of these isolates and identified three populations consisting of at least 8 isolates. We then computationally predicted genome-wide CNV profiles for each isolate and conducted analyses at the species-, population-, and individual levels. Our results suggest that CNV contributes to genetic variation in A. fumigatus, with ~10% of the genome being CN variable. Our analysis indicates that CNV is non-randomly distributed across the A. fumigatus genome, and is overrepresented in subtelomeric regions. Analysis of gene ontology categories in genes that overlapped CN variants revealed an enrichment of genes related to transposable element and secondary metabolism functions. We further identified 72 loci containing 33 genes that showed divergent copy number profiles between the three A. fumigatus populations. Many of these genes encode proteins that interact with the cell surface or are involved in pathogenicity. Our results suggest that CNV is an important source of genetic variation that could account for some of the phenotypic differences between A. fumigatus populations and isolates.

Machine Learning for Better Prognostic Stratification and Driver Gene Identification Using Somatic Copy Number Variations in Anaplastic Oligodendroglioma.

1p/19q-codeleted anaplastic gliomas have variable clinical behavior. We have recently shown that the common 9p21.3 allelic loss is an independent prognostic factor in this tumor type. The aim of this study is to identify less frequent genomic copy number variations (CNVs) with clinical importance that may shed light on molecular oncogenesis of this tumor type. A cohort of 197 patients with anaplastic oligodendroglioma was collected as part of the French POLA network. Clinical, pathological, and molecular information was recorded. CNV analysis was performed using single-nucleotide polymorphism arrays. Computational biology and feature selection based on the random forests method were used to identify CNV events associated with overall survival and other clinical-pathological variables. Recurrent chromosomal events were identified in chromosomes 4, 9, and 11. Forty-six focal amplification events and 22 focal deletion events were identified. Twenty-four focal CNV areas were associated with survival, and five of them were significantly associated with survival after multivariable analysis. Nine out of 24 CNV events were validated using an external cohort of The Cancer Genome Atlas. Five of the validated events contain a cancer-related gene or microRNA: CDKN2A deletion, SS18L1 amplification, RHOA/MIR191 copy-neutral loss of heterozygosity, FGFR3 amplification, and ARNT amplification. The CNV profile contributes to better survival prediction compared with clinical-based risk assessment. Several recurrent CNV events, detected in anaplastic oligodendroglioma, enable better survival prediction. More importantly, they help in identifying potential genes for understanding oncogenesis and for personalized therapy. Genomic analysis of 197 anaplastic oligodendroglioma tumors reveals recurrent somatic copy number variation areas that may help in understanding oncogenesis and target identification for precision medicine. A machine learning multivariable model built using this genomic information enables better survival prediction.

Abstract 2963: Characterization of disease evolution in sequential sampled metastatic breast cancer using liquid biopsy

Abstract Background: It is established that tumors evolve over time due to treatment pressure or physiological constraints. Tissue biopsies give a one-time window into the characteristics of a primary tumor or metastatic nodule. Their main pitfall is procedure risk, pain, and cost, which is why they are rarely repeated. Liquid biopsies, however, are performed with minimal discomfort and risk, providing an easier way to monitor disease. Circulating tumor cells (CTCs) and cell-free DNA (cfDNA) have proven to be of high prognostic and diagnostic value. Here we evaluate 17 longitudinal blood draws collected from a metastatic breast cancer patient over 4 years and assess the power of single-cell analysis relative to analysis of tissue biopsies and cfDNA. Methods: CTCs were enumerated using the high-definition single-cell assay (HD-SCA) and single CTCs were isolated by micromanipulation. DNA of single cells underwent whole genome amplification and Illumina library construction for copy number variation (CNV) analysis. Genomic DNA from microdissected FFPE tissue and cfDNA extracted from plasma were converted into Illumina sequencing libraries for CNV analysis. CA27.29 levels were reported by the clinic. Results: Levels of CTCs and circulating tumor DNA (ctDNA) fractions were positively correlated with CA27.29. CNV analysis of single tumor cells at time of enrollment revealed 3 morphometrically indistinguishable clones that persisted during the 4-year follow-up. Genomic analysis of primary breast tissue and bone metastasis detected at time of diagnosis matched CNV data of the evolutionary first clone, while a later liver metastasis was comprised solely of clone 2. Remarkably, CTCs in clone 2 were identified years before discovery of the liver metastasis. To date, we have not detected a metastatic site harboring clone 3. We found that CNV profiles from cfDNA represented the aggregate of CTC clones. Conclusion: Longitudinal blood sampling enabled tracing of tumor evolution robustly and less invasively than solid tumor biopsies. Metastasis associated clones were identified years before measureable disease, providing an opportunity for earlier interventions. While plasma-based assays detected the presence of tumor-derived DNA, the single-cell high-content resolution of the HD-SCA workflow was required to deconvolute 3 genomically distinct clones present in this patient. We propose that enumeration of CTCs and analysis of ctDNA could be used as additional clinical markers to monitor disease recurrence, adding valuable morphogenomic information about the primary tumor and metastatic sites. Additionally, we are currently analyzing the results of single nucleotide variation analysis of single CTCs, cfDNA and FFPE tissue from this patient. Citation Format: Lisa Welter, Liya Xu, Dillon McKinley, Sara Restrepo-Vassalli, Angel Dago, Mariam Rodriguez Lee, Anand Kolatkar, James Hicks, Jorge Nieva, Peter Kuhn. Characterization of disease evolution in sequential sampled metastatic breast cancer using liquid biopsy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2963.

Abstract 4584: High-definition single-cell analysis of liquid and solid biopsies reveals heterogeneity of circulating tumor cells in metastatic colorectal cancer

Abstract Treatment of cancer patients is transitioning to personalized strategies using targeted drugs with the required companion biomarkers. Many current biomarkers are tissue-based, requiring a solid biopsy and limiting sampling frequency. Given this, there is significant interest in determining whether noninvasive, blood-based liquid biopsies, utilizing measurements of factors that exist in the blood such as circulating tumor cells (CTCs), reflect solid tumors and can serve as substitutes for solid tumor biopsies. Colorectal cancer with hepatic metastases (mCRC) is one of very few scenarios where metastatic patients routinely undergo surgical resection, providing an opportunity for accessible tissue and simultaneous blood sampling. We collected blood and tumor touch preparations from patients undergoing resection of CRC with the purpose to determine the correlation of solid biopsies and the liquid phase of solid tumors in the mCRC setting. Samples were processed using the previously described HD-SCA (High-Definition Single-Cell Analysis) workflow for high-resolution imaging and characterization of tumor cells at a single-cell level. Briefly, tumor cells were identified by presence of nuclear stain, pan-cytokeratin, and absence of CD45. The colon-specific marker CDX2 was included as a fourth color for further characterization. Subsets of cells were subjected to copy number variation (CNV) profiling. The comparison of pre- and post-surgery blood samples showed that the concentration of CTCs decreased significantly after surgery. The concentration was markedly higher in patients with synchronous disease undergoing simultaneous resection of the primary tumor, indicating that the presence of CTCs is related to tumor burden. Principal component analysis demonstrated that the CTCs were more similar to the metastatic cells than to the primary colon cells. While protein levels varied between the liquid and solid biopsy cell populations, morphologic parameters including, e.g., cell and nuclear area and roundness, were similar in the two. The intrapatient analysis showed heterogeneous CTC populations, in general with a small subset of cells bearing characteristics more similar to the solid tumor cells than the bulk of CTCs. The presence of subpopulations was further indicated by the CNV analysis, which showed that only a minority of CTCs had genomic rearrangements and that, in contrast to the solid biopsy cells, the rearranged cells were largely nonclonal. Based on the high-content HD-SCA data, the CTCs were further grouped into different subtypes. In conclusion, we introduce an approach for high-content comparison of liquid and solid biopsies at a single-cell level. We show that the concentration of CTCs is related to tumor burden, and present a subclassification of rare circulating cells in mCRC. Citation Format: Anna S. Gerdtsson, Jana-Aletta Thiele, Randolph Schaffer, Kelly Bethel, Steven Curley, Heinz-Josef Lenz, Diana L. Hanna, Jorge Nieva, Anand Kolatkar, Mariam Rodriguez Lee, James Hicks, Peter Kuhn. High-definition single-cell analysis of liquid and solid biopsies reveals heterogeneity of circulating tumor cells in metastatic colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4584.

Abstract 1994: Pre-existence of poly-resistant cancer stem cells in high-grade ovarian cancer

Abstract Introduction: High-grade ovarian cancer (HGOC) shows excellent responses to standard-of-care surgery and paclitaxel/carboplatin therapy only to relapse 6-24 months later with typically resistant disease. While the origin of this recurrent, resistant disease is unclear, most believe it is acquired by the action of chemotherapeutics. Using novel stem cell technology that enables the cloning of cancer stem cells (CSCs) from epithelial cancers, we have generated large libraries of CSCs from multiple cases of HGOC. And while the vast majority of these CSC clones are killed by standard-of-care chemotherapeutic drugs, a minor fraction shows profound resistance not only to paclitaxel/carboplatin but to a wide range of structurally unrelated chemotherapeutic drugs to which these cells had no prior exposure. We describe screens for drugs that selectively target this resistant CSC population. Methods: Libraries of 10- to 100,000 CSC clones were generated from individual, therapy naïve, HGOC resections using technology we developed for cloning so-called “adult” stem cells from normal columnar epithelia (Wang et al., 2015). Results: Paclitaxel/carboplatin resistant CSCs were identified in CSC libraries derived from therapy naïve tumors at ratios of 1:50 to 1:300. By copy number variation, these resistant variant clones proved distinct from the bulk of CSCs, and by gene expression analysis varied from sensitive clones by more than 700 differentially expressed genes. Independent resistant clones from the same library clustered with other resistant clones by both copy number variation and gene expression profiles, suggesting the possibility that resistance within a single tumor is dominated by a single type of resistant CSCs. Clones resistant to paclitaxel/carboplatin were screened in a 384-well format against a wide range of experimental drug-like molecules. These pre-existing resistant clones also proved to be profoundly resistant to a large number of structurally unrelated chemotherapeutic drugs. This same screening program identified drugs that act alone or in combination with paclitaxel to eliminate these resistant clones, suggesting a route to personalized medicine for addressing the problem of recurrent disease in HGOC. Conclusions: Tumors from patients with HGOC possess clonogenic CSCs including variants that are resistant to a broad spectrum of chemotherapeutics to which they have not been exposed. It is likely that such CSCs would survive standard-of-care chemotherapy and contribute to the recurrent disease seen in HGOC. We have identified known and experimental drugs that specifically eliminate these resistant variants and the overall platform represents a potential strategy to addressing the problem of recurrent disease in these patients. Citation Format: Wa Xian, Frank McKeon, Christopher Crum, Jingzhong Xie, Bailiang Wang, Yusuke Yamamoto, Suzy Torti, Molly Brewer, Matthew Anderson, Amir Jazaeri, Giulio Draetta, Shan Wang, Rajasekaran Mahalingam, Hong Chen, Marcin Duleba, Wei Rao. Pre-existence of poly-resistant cancer stem cells in high-grade ovarian cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1994.

Abstract 1281: Genomic structure variation in large screening for pediatric sarcoma therapy

Abstract Complex chromosomal aberrations such as amplification and deletion of DNA copy number are frequently seen in sarcoma. Fifty-five DNA structure variation has been listed as standard clinical diagnosis for sarcoma by standard of National Comprehensive Cancer Network (NCCN). However, copy number variation (CNV) as a biomarker of drug treatment for pediatrics sarcoma is still unclear, especially for relapsed and high recurrent patients of pediatric sarcoma. The paper aims to detect the prognosis biomarkers for rhabdomyosarcoma, Ewing's sarcoma (ES), and osteosarcoma based on copy number variation for 128 FDA-approved cancer drugs systematically. The 182 copy number variation (CNV) profiles from clinical sarcoma patients across three types of sarcoma, including osteosarcoma, rhabdomyosarcoma and ES, are observed. Cox survival regression model is used to select significant cancer-related CNVs systematically by correlation with overall survival analysis. 532 significant common CNVs and 782 specific CNVs for different types of sarcoma are observed by p-value <0.01. By comparing with healthy persons, COSMIC database, 947 cancer cells and our INHOUSE tumor datasets, 32 out of 532 CNVs for specific sarcoma types are validated. This systematic CNV comparison makes it possible to map DNA copy number changes and identify chromosomal regions containing “target genes” responsible for tumor development and/or progression. Integrating large-scale drug screening with copy number variation analysis on 38 sarcoma cancer cells, 95 out of 532 are confirmed as drugs resistance and sensitive biomarkers for different types of sarcoma. 15 chemotherapies show strong signals connecting with 42 biomarkers, including genes MYC, RAD21, and RB1 amplification. By bridging molecular biomarkers between tumors and sarcoma cell lines, 5 optimum drugs associated with 14 enzyme biomarkers are recommended for pediatric rhabdomyosarcoma, ES, and osteosarcoma treatment respectively. The research not only detects copy number variations both of sarcoma cell lines and tumors, but also provides novel insights into drug biomarkers based on copy number amplification or deletion for bone sarcoma and soft tissue sarcoma, respectively. Citation Format: Lijun Cheng, Pooja Chandra, Limei Wang, Karen Pollok, Pankita Pandya, Mary Murray, Jacquelyn Carter, Michael Ferguson, Mohammad Reza, Mashall Mark, Lang Li, Jamie Renbarger. Genomic structure variation in large screening for pediatric sarcoma therapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1281.

PO-329 Genomic aberration in pancreatic neuroendocrine tumours (PNET)

IntroductionRare tumours are poorly investigated and therefore heavily limited in possible therapeutic approaches, and subsequently therapeutic success. One such tumour entity, representing only 1%–2% of all pancreatic malignancies is Pancreatic Neuroendocrine tumours (PNET). PNET are cancers of endocrine cells of the islets of Langerhans and only a handful of studies thus far have investigated the genetic makeup of these tumours. Apart from surgery and chemotherapy for localised tumours, the only approved targeted drug for treatment of advanced PNETs is Everolimus, specifically targeting the mTOR protein. Much is still lacking in the area of targeted therapies and possible options for treating metastasized tumours. Our aim is to therefore establish a genetic profile of PNET tumours in order to determine more targets connected to mTOR signalling and consequently, increase options for improved therapy approaches.Material and methodsWe performed deep coverage sequencing, averaging a depth of 1000 reads, using an established in-house PNET panel. Our cohort included Formaldehyde Fixed Paraffin Embedded (FFPE) patient samples comprising of 50 primary tumours of varying grades, and 14 metastases.Results and discussionsUsing variant calling pipelines, we determined single nucleotide variations (SNV) and copy number variations (CNV) in our cohort and elucidated potential heterogeneity within the PNET tumours. Out of the 62 sequenced samples, 26 primary tumours, and 6 metastases showed SNV alterations with the most recurring being MEN1 alterations (50%) irrespective of whether the tumour is primary or metastasis, followed by DAXX (27%) which was exclusively mutated in primary tumours. In addition, mutation in a novel candidate, polycomb protein EED has been determined specifically in G3 PNET and the alteration falls in one of the β-transducing repeat of the protein. CNV was determined in 38 samples with a recurring copy number increase for RICTOR (~37%). Intra-tumour heterogeneity, based on allelic frequency, is evident in some samples while the combined SNV and CNV profile of these patients reveal a clear inter-tumour heterogeneity in PNETs.ConclusionOur study reveals previously identified and novel alterations, which require further investigation at the protein level to determine new possible therapeutic approaches for PNET patients. In addition, our data clearly reveals that due to the high level of inter-tumour heterogeneity present in these tumours, it is vital to consider a more personalised targeted approach to therapy.

Abstract A01: Somatic engineering of the mammary gland for the development of novel mouse models of triple-negative breast cancer

Abstract Large-scale sequencing studies are rapidly identifying putative oncogenic mutations in human tumors. However, discrimination between passenger and driver events in tumorigenesis remains challenging and requires in vivo validation studies in reliable animal models of human cancer. For these reasons, new technologies are needed to expand the genetic toolbox of cancer biologists and allow a more rapid and systematic in vivo interrogation of gene perturbations. In this regard, the advent of CRISPR/Cas9 technologies for somatic genome editing has already paved the way for a new generation of non-germline animal tumor models. For example, liver-specific gene disruption was achieved by transient delivery of components of the CRISPR/Cas9 system in the tail veins of mice, leading to hepatocellular carcinoma (Xue et al., 2014; Weber et al., 2015). Similar approaches have been used to deliver targeted oncogenic mutations to the lung (Platt et al., 2014; Sánchez-Rivera et al., 2014), brain (Zuckermann et al., 2015), and pancreas (Chiou et al., 2015). We have previously shown that the mammary tissue is also amenable to somatic gene editing using intraductal injection of lentiviral vectors encoding Cre recombinase, the CRISPR/Cas9 system, or both in the mammary glands of female mice carrying conditional predisposing alleles (Annunziato et al., 2016). This approach was successfully applied to validate in vivo candidate tumor suppressors implicated in invasive lobular carcinoma (Kas et al., 2017), but it is conceivable that more breast cancer subtypes can be modeled via somatic engineering in mice with distinct predisposing mutations. In this study, we show how somatic engineering via intraductal injection of lentiviruses can be used to develop innovative mouse models of triple-negative breast cancer (TNBC), one of the subtypes with the poorest prognosis in the clinics. Using intraductal injection of Cre-encoding lentiviruses in mice with conditional Brca1 and Trp53 alleles (B1P), we were able to target TNBC-initiating cells from the basal compartment, and induce tumors that retained all the hallmark features of the transgenic Cre-driven tumor counterparts, including latency, histopathology, and DNA copy-number variation (CNV) pattern. Moreover, we could dramatically accelerate tumorigenesis when lentiviral Cre was injected in B1PM mice that harbored, in addition to conditional Brca1 and Trp53 alleles, a conditional knock-in allele overexpressing the Myc oncogene, a candidate driver in BRCA1-associated cancers. Alternatively, somatic Myc expression and loss of BRCA1 and p53 was induced in the mammary glands of B1P mice via intraductal injection of lentiviral vectors encoding Myc and Cre under a viral promoter. In both cases the resulting tumors displayed a reshaped CNV profile that strongly implicated a limited number of recurrent focal amplifications and deletions in the tumorigenic cascade. We reiterated the process by lentiviral overexpression of these candidate driver oncogenes and Cre in the mammary glands of B1PM mice, or by achieving in situ CRISPR/Cas9-mediated somatic gene disruption of one or multiple candidate tumor-suppressor genes in Cas9-transgenic WB1P-Cas9 mice. This versatile somatic platform allows for rapid, flexible, and multiplexable in vivo testing of putative oncogenic hits and deals with the need for increasingly complex mouse models while avoiding the bottlenecks of classical transgenesis. Notably, we show here that we can extend the applicability of non-germline breast cancer modeling in mice to produce genetically and histologically accurate TNBC, which allows in vivo investigation of the underlying biology as well as rapid generation of tailored preclinical models for this aggressive breast cancer subtype. Citation Format: Stefano Annunziato, Julian R. de Ruiter, Chiara S. Brambillasca, Sjors M. Kas, Federica Ferrante, Catrin Lutz, Bjorn Siteur, Bas van Gerwen, Marieke van de Ven, Martine H. van Miltenburg, Linda Henneman, Jos Jonkers. Somatic engineering of the mammary gland for the development of novel mouse models of triple-negative breast cancer [abstract]. In: Proceedings of the AACR Special Conference: Advances in Modeling Cancer in Mice: Technology, Biology, and Beyond; 2017 Sep 24-27; Orlando, Florida. Philadelphia (PA): AACR; Cancer Res 2018;78(10 Suppl):Abstract nr A01.

Novel, improved grading system(s) for IDH-mutant astrocytic gliomas.

According to the 2016 World Health Organization Classification of Tumors of the Central Nervous System (2016 CNS WHO), IDH-mutant astrocytic gliomas comprised WHO grade II diffuse astrocytoma, IDH-mutant (AIIIDHmut), WHO grade III anaplastic astrocytoma, IDH-mutant (AAIIIIDHmut), and WHO grade IV glioblastoma, IDH-mutant (GBMIDHmut). Notably, IDH gene status has been made the major criterion for classification while the manner of grading has remained unchanged: it is based on histological criteria that arose from studies which antedated knowledge of the importance of IDH status in diffuse astrocytic tumor prognostic assessment. Several studies have now demonstrated that the anticipated differences in survival between the newly defined AIIIDHmut and AAIIIIDHmut have lost their significance. In contrast, GBMIDHmut still exhibits a significantly worse outcome than its lower grade IDH-mutant counterparts. To address the problem of establishing prognostically significant grading for IDH-mutant astrocytic gliomas in the IDH era, we undertook a comprehensive study that included assessment of histological and genetic approaches to prognosis in these tumors. A discovery cohort of 211 IDH-mutant astrocytic gliomas with an extended observation was subjected to histological review, image analysis, and DNA methylation studies. Tumor group-specific methylation profiles and copy number variation (CNV) profiles were established for all gliomas. Algorithms for automated CNV analysis were developed. All tumors exhibiting 1p/19q codeletion were excluded from the series. We developed algorithms for grading, based on molecular, morphological and clinical data. Performance of these algorithms was compared with that of WHO grading. Three independent cohorts of 108, 154 and 224 IDH-mutant astrocytic gliomas were used to validate this approach. In the discovery cohort several molecular and clinical parameters were of prognostic relevance. Most relevant for overall survival (OS) was CDKN2A/B homozygous deletion. Other parameters with major influence were necrosis and the total number of CNV. Proliferation as assessed by mitotic count, which is a key parameter in 2016 CNS WHO grading, was of only minor influence. Employing the parameters most relevant for OS in our discovery set, we developed two models for grading these tumors. These models performed significantly better than WHO grading in both the discovery and the validation sets. Our novel algorithms for grading IDH-mutant astrocytic gliomas overcome the challenges caused by introduction of IDH status into the WHO classification of diffuse astrocytic tumors. We propose that these revised approaches be used for grading of these tumors and incorporated into future WHO criteria.

Bamgineer: Introduction of simulated allele-specific copy number variants into exome and targeted sequence data sets.

Somatic copy number variations (CNVs) play a crucial role in development of many human cancers. The broad availability of next-generation sequencing data has enabled the development of algorithms to computationally infer CNV profiles from a variety of data types including exome and targeted sequence data; currently the most prevalent types of cancer genomics data. However, systemic evaluation and comparison of these tools remains challenging due to a lack of ground truth reference sets. To address this need, we have developed Bamgineer, a tool written in Python to introduce user-defined haplotype-phased allele-specific copy number events into an existing Binary Alignment Mapping (BAM) file, with a focus on targeted and exome sequencing experiments. As input, this tool requires a read alignment file (BAM format), lists of non-overlapping genome coordinates for introduction of gains and losses (bed file), and an optional file defining known haplotypes (vcf format). To improve runtime performance, Bamgineer introduces the desired CNVs in parallel using queuing and parallel processing on a local machine or on a high-performance computing cluster. As proof-of-principle, we applied Bamgineer to a single high-coverage (mean: 220X) exome sequence file from a blood sample to simulate copy number profiles of 3 exemplar tumors from each of 10 tumor types at 5 tumor cellularity levels (20–100%, 150 BAM files in total). To demonstrate feasibility beyond exome data, we introduced read alignments to a targeted 5-gene cell-free DNA sequencing library to simulate EGFR amplifications at frequencies consistent with circulating tumor DNA (10, 1, 0.1 and 0.01%) while retaining the multimodal insert size distribution of the original data. We expect Bamgineer to be of use for development and systematic benchmarking of CNV calling algorithms by users using locally-generated data for a variety of applications. The source code is freely available at http://github.com/pughlab/bamgineer.

Association study between copy number variation and beef fatty acid profile of Nellore cattle.

The aim of this study was to analyze the association between the copy number variation regions (CNVRs) and fatty acid profile phenotypes for saturated (SFA), monosaturated (MUFA), polyunsaturated (PUFA), ω6 and ω3 fatty acids, PUFA/SFA and ω6/ω3 ratios, as well as for their sums, in Nellore cattle (Bos primigenius indicus). A total of 963 males were finished in feedlot and slaughtered with approximately 2years of age. Animals were genotyped with the BovineHD BeadChip (Illumina Inc., San Diego, CA, USA). The copy number variation (CNV) detection was performed using the PennCNV algorithm. Log R ratio (LRR) and allele B frequency (BAF) were used to estimate the CNVs. The association analyses were done using the CNVRuler software and applying a logistic regression model. The phenotype was adjusted using a linear model considering the fixed effects of contemporary group and the animal age at slaughter. The fatty acid profile was analyzed on samples of longissimus thoracis muscle using gas chromatography with a 100-m capillary column. For the association analysis, the adjusted phenotypic values were considered for the traits, while the data was adjusted for the effects of the farm and year of birth, management groups at birth, weaning, and superannuation. A total of 186 CNVRs were significant for SFA (43), MUFA (42), PUFA (66), and omega fatty acid (35) groups, totaling 278 known genes. On the basis of the results, several genes were associated with several fatty acids of different saturations. Olfactory receptor genes were associated with C12:0, C14:0, and C18:0 fatty acids. The SAMD8 and BSCL2 genes, both related to lipid metabolic process, were associated with C12:0. The RAPGEF6 gene was found to be associated with C18:2 cis-9 cis-12 n-6, and its function is related to regulation of GTPase activity. Among the results, we highlighted the olfactory receptor activity (GO:0004984), G-protein-coupled receptor activity (GO:0004930), potassium:proton antiporter activity (GO:0015386), sodium:proton antiporter activity (GO:0015385), and odorant-binding (GO:0005549) molecular functions. A large number of genes associated with fatty acid profile within the CNVRs were identified in this study. These findings must contribute to better elucidate the genetic mechanism underlying the fatty acid profile of intramuscular fat in Nellore cattle.