Hossain, Asif

Abstract Abstract 1814 Multiple myeloma (MM) is an incurable hematologic disorder characterized by the accumulation of plasma cells (PCs) in the bone marrow (BM). Like normal B cells and PCs, MM cells express a distinctive immunoglobulin (Ig) molecular fingerprint resulting from heavy and light chain gene rearrangements and unique mutations introduced by the somatic hypermutation (SHM) process. Detailed examination of the Ig molecule may offer insight into the etiology of B cell malignancies as well as the ontogeny of malignant transformation and ongoing genetic evolution. Because prior studies of the Ig fingerprint of MM cells have been limited in size and scope, our current understanding of MM Ig repertoire is minimal. We used massively parallel sequencing of normal and malignant PCs to interrogate: 1) DNA versus RNA (cDNA) as starting material for studies of Ig repertoire; 2) the Ig repertoire of normal and malignant PCs; and 3) intraclonal heterogeneity in MM. Bone marrow PCs (BMPCs) from an untreated MM patient and from a normal control subject were isolated by magnetic bead separation. DNA and RNA were isolated from purified cells and amplified in a multiplex PCR reaction with primers designed to capture the entire IGHV region including the heavy chain complementarity determining region 3 (HCDR3) necessary for defining clonal cells. Using Roche 454 GS-FLX Titanium chemistry and a GS FLX Titanium Genome Sequencer, over 30,000 sequence reads were obtained from each sample. Following a detailed algorithm to minimize Taq and sequencing related errors and removal of non-Ig or non-productive Ig sequences, the resulting sequences were grouped based on HCDR3 amino acid identity. Each unique HCDR3 represented one member of the repertoire and all the sequences with the same HCDR3 defined the clone size. This measurement showed that there was nearly a five-fold higher level of unique HCDR3s in normal BMPCs when the starting PCR template was DNA vs. cDNA. Thus, DNA is clearly a better starting template for repertoire analysis using deep sequencing methods. Thorough analysis of the MM DNA sample identified PCs expressing 37 different IGHV genes, and predictably, the vast majority of sequence reads (83%) had exact nucleotide identity in the IGHV and HCDR3 region (IGHV3-74 with 18 mutations in the IGHV region) thereby defining the dominant MM clone. There were also ∼5000 additional sequences that either shared an identical HCDR3 region but varied within the IGHV3-74 gene sequence (n=3975) or were identical within the IGHV3-74 sequence yet varied within the HCDR3 (n=862). These sequences were further analyzed, with sequences not present 10 or more times or not found in both the 5' and 3' directions eliminated from further scrutiny. This process revealed a total of 64 putative subclones with single point mutations in the IGHV or HCDR3 region (10–56 copies of each unique sequence representing a total of 1583 clonally related sequences). When we used a more stringent subclone frequency of ≥0.1%, 22 subclones remained. Of note, conventional Ig repertoire analysis would require sequencing ≥1000 colonies to detect MM subclones present at this frequency. Finally, 11 subclones displayed a significant number of IGHV gene nucleotide differences from the dominant MM clone despite exhibiting an identical HCDR3 region. Notably, a number of these distinguishing IGHV gene mutations are silent and thus may preserve the antigenic specificity of these clonally related cells. These subclones failed our frequency criteria but the type and quality of deviations from the clonal sequence suggest that they are not artifact and warrant further investigation. Notably, all of the putative subclones were also found in the MM cDNA cohort providing further evidence as genuine subclones. The very large number of identical MM clonal sequences is consistent with the idea that MM is a post germinal center malignancy that can no longer undergo SHM. However, we present clear evidence of MM-related subclones albeit relatively small in number. The subclone sequence deviations discovered may reflect on-going genetic evolution independent of the formal SHM mechanism. Alternatively, these subclones may indeed reflect “sister” cells generated during the original germinal center reaction that resulted in the malignant clone. In summary, our studies demonstrate the extraordinary potential of this methodology to track clonal evolution in MM and its precursor conditions over time. Disclosures: No relevant conflicts of interest to declare.

Abstract Motivation: The Biological Reference Repository (BioR) is a toolkit for annotating variants. BioR stores public and user-specific annotation sources in indexed JSON-encoded flat files (catalogs). The BioR toolkit provides the functionality to combine and retrieve annotation from these catalogs via the command-line interface. Several catalogs from commonly used annotation sources and instructions for creating user-specific catalogs are provided. Commands from the toolkit can be combined with other UNIX commands for advanced annotation processing. We also provide instructions for the development of custom annotation pipelines. Availability and implementation: The package is implemented in Java and makes use of external tools written in Java and Perl. The toolkit can be executed on Mac OS X 10.5 and above or any Linux distribution. The BioR application, quickstart, and user guide documents and many biological examples are available at http://bioinformaticstools.mayo.edu . Contact: Kocher.JeanPierre@mayo.edu Supplementary information: Supplementary data are available at Bioinformatics online.

Abstract Motivation: KRAS is commonly mutated in a variety of cancers including lung cancer. The KRAS gene is frequently mutated at codons 12 and 13 in lung adenocarcinomas in patients with a history of smoking. Tumors harboring an activating KRAS mutation are aggressive and are often resistant to available therapies. In the present study, we set out to identify pathways that are specifically altered in lung cancer patients whose tumors harbor a KRAS mutation using next generation sequencing technology. Methods: We performed 50nt paired-end RNA-sequencing in 15 lung adenocarcinoma samples (8 and 7 samples with and without KRAS mutation). Reads were aligned to genome and exon junctions using Illumina's alignment tool ELAND_RNA. CASAVA and Genome Studio data analysis software were used to obtain read counts for genes, exons and exon junctions. We applied a variety of computational methods (Casper R package, Bowtie, TopHat) and softwares (Partek, R statistical software, JMP, Ingenuity Pathway Anlaysis, DAVID) to carry out our analyses. Results: Differential gene expression and splicing analysis were performed between the two groups. We identified 115 genes that were differentially expressed and 112 genes that consist of splicing variants with at least 2 fold changes and p-value < 0.01. We randomly selected 6/15 samples and performed real-time qRT-PCR for 6 differentially expressed genes. Significant correlations were observed ranging from 0.44 to 0.83, when qRT-PCR results were compared with RNA-sequencing expression data. Pathway analysis with 115 differentially expressed genes and 112 splicing variants revealed that the most significant pathway is composed entirely of NFκB focus genes, suggesting that there is a direct connection between oncogenic KRAS and activation of the NFκB signaling pathway. Conclusions: NFκB has been implicated in KRas-mediated formation of early stage lung adenomas in the LSL-Kas mouse model. Our data indicate that this connection also occurs in human lung cancer and establish NFκB activation as a key manifestation of KRAS mutation in human lung cancer. Hence drugs targeting NFκB and NFκB related genes may potentially be helpful for the treatment of patients with oncogenic KRAS mutations. (Supported in part by grants from the 26.2 with Donna Foundation and the Florida Department of Health Bankhead/Coley program.) 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 4975. doi:10.1158/1538-7445.AM2011-4975

Abstract Background: The role of fusion genes and associated fusion transcripts has long been recognized in hematopoietic malignancies. Until quite recently it has been difficult to detect such events on a genomic scale in solid tumors. Consequently, little is known about the potential role of fusion genes, transcripts, and proteins as driver mutations, biomarkers, or therapeutic targets in breast cancer. Methods: We have developed a novel analytical pipeline, Snowshoes-FTD, for detection of fusion transcripts in breast cancer cell lines and tumor samples (Asmann, et al. NAR 2011; May 27 ePub ahead of print). Preliminary analyses have been carried out with a panel of 8 each ER+, HER2+, and triple negative (TN) primary breast tumors, 8 primary human mammary epithelial cell (HMEC) lines from biopsy samples, plus 16 normal tissues from the Illumina Body Map dataset. Results: We have identified 120 redundant, tumor-specific fusion transcripts, expressed in two or more tumors and in no non-transformed samples. Sixteen of these represent intrachromosomal fusions and 104 arise from fusion of transcripts that map to two different chromosomes. Every breast tumor expressed one or more fusion transcripts. Twenty-nine fusion transcripts appeared to be tumor subtype specific. Among these, we have identified 2 HER2+, 10 ER+, and 17 triple negative specific redundant transcripts. In general, HER2+ tumors expressed fewer fusion transcripts (range 4 to 28/tumor) compared to TN (range11 to 44/tumor). Chromosomal distribution patterns were also markedly different among the tumor subtypes. For example, ER+ tumors expressed a preponderance of redundant fusion transcripts that involve chr1 and 2, whereas TN tumors had no fusion transcripts that map to either chromosome. Conversely, the predominant locus for TN fusion transcripts was chr19, which contains only one HER2+ fusion and no ER+ fusion transcripts. Conclusions: Primary breast tumors express many chimaeric transcripts, which we presume to arise primarily from genomic rearrangements. The majority of these transcripts are redundant, and a subset are tumor subtype specific. These transcripts may mark regions of chromosomal instability. HER2+ tumors, in general, appear to evidence less chromosomal instability, as inferred from fusion transcripts; although some HER2+ tumors appear to be quite unstable. TN tumors contain many more redundant fusion transcripts, implying increased genomic instability, particularly in chr19. We conclude that these fusion transcripts represent a class of heretofore unrecognized biomarkers that may be used for sub-classification of breast tumors. Some of these transcripts appear to encode proteins that may function as tumor-subtype-specific driver mutations and may have potential as therapeutic targets in breast cancer. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P3-06-02.

Abstract Fusion genes and fusion gene products are widely employed as biomarkers and therapeutic targets in hematopoietic cancers, but their applications have yet to be appreciated in solid tumors. Here, we report the use of SnowShoes-FTD, a powerful new analytic pipeline that can identify fusion transcripts and assess their redundancy and tumor subtype-specific distribution in primary tumors. In a study of primary breast tumors, SnowShoes-FTD was used to analyze paired-end mRNA-Seq data from a panel of estrogen receptor (ER)+, HER2+, and triple-negative primary breast tumors, identifying tumor-specific fusion transcripts by comparison with mRNA-Seq data from nontransformed human mammary epithelial cell cultures plus the Illumina Body Map data from normal tissues. We found that every primary breast tumor that was analyzed expressed one or more fusion transcripts. Of the 131 tumor-specific fusion transcripts identified, 86 were “private” (restricted to a single tumor) and 45 were “redundant” (distributed among multiple tumors). Among the redundant fusion transcripts, 7 were unique to ER+ tumors and 8 were unique to triple-negative tumors. In contrast, none of the redundant fusion transcripts were unique to HER2+ tumors. Both private and redundant fusion transcripts were widely expressed in primary breast tumors, with many mapping to genomic loci implicated in breast carcinogenesis and/or risk. Our finding that some fusion transcripts are tumor subtype-specific suggests that these entities may be critical determinants in the etiology of breast cancer subtypes, useful as biomarkers for tumor stratification, or exploitable as cancer-specific therapeutic targets. Cancer Res; 72(8); 1921–8. ©2012 AACR.