April, Craig

Abstract Purpose: Recent progress in understanding the molecular biology of epithelial ovarian cancer has not yet translated into individualized treatment for these women or improvements in their disease outcome. Gene expression has been utilized to identify distinct molecular subtypes, but there have been no reports investigating whether or not molecular subtyping is predictive of response to bevacizumab in ovarian cancer. Experimental Design: DASL gene expression arrays were performed on FFPE tissue from patients enrolled on the ICON7 trial. Patients were stratified into four TCGA molecular subtypes. Associations between molecular subtype and the efficacy of randomly assigned therapy with bevacizumab were assessed. Results: Molecular subtypes were assigned as follows: 122 immunoreactive (34%), 96 proliferative (27%), 73 differentiated (20%), and 68 mesenchymal (19%). In univariate analysis patients with tumors of proliferative subtype obtained the greatest benefit from bevacizumab with a median PFS improvement of 10.1 months [HR, 0.55 (95% CI, 0.34–0.90), P = 0.016]. For the mesenchymal subtype, bevacizumab conferred a nonsignificant improvement in PFS of 8.2 months [HR 0.78 (95% CI, 0.44–1.40), P = 0.41]. Bevacizumab conferred modest improvements in PFS for patients with immunoreactive subtype (3.8 months; P = 0.08) or differentiated subtype (3.7 months; P = 0.61). Multivariate analysis demonstrated significant PFS improvement in proliferative subtype patients only [HR, 0.45 (95% CI, 0.27–0.74), P = 0.0015]. Conclusions: Ovarian carcinoma molecular subtypes with the poorest survival (proliferative and mesenchymal) derive a comparably greater benefit from treatment that includes bevacizumab. Validation of our findings in an independent cohort could enable the use of bevacizumab for those patients most likely to benefit, thereby reducing side effects and healthcare cost. Clin Cancer Res; 23(14); 3794–801. ©2017 AACR.

Abstract Ovarian cancer is a clinically and molecularly heterogeneous disease. The driving forces behind this variability are unknown. Here, we report wide variation in the expression of the DNA cytosine deaminase APOBEC3B, with elevated expression in the majority of ovarian cancer cell lines (three SDs above the mean of normal ovarian surface epithelial cells) and high-grade primary ovarian cancers. APOBEC3B is active in the nucleus of several ovarian cancer cell lines and elicits a biochemical preference for deamination of cytosines in 5′-TC dinucleotides. Importantly, examination of whole-genome sequence from 16 ovarian cancers reveals that APOBEC3B expression correlates with total mutation load as well as elevated levels of transversion mutations. In particular, high APOBEC3B expression correlates with C-to-A and C-to-G transversion mutations within 5′-TC dinucleotide motifs in early-stage high-grade serous ovarian cancer genomes, suggesting that APOBEC3B-catalyzed genomic uracil lesions are further processed by downstream DNA "repair" enzymes including error-prone translesion polymerases. These data identify a potential role for APOBEC3B in serous ovarian cancer genomic instability. Cancer Res; 73(24); 7222–31. ©2013 AACR.

Abstract High throughput sequencing technologies open up a new dimension in cancer genomics by enabling the characterization of cancer genomes at base-pair resolution. To date, the majority of large-scale genomics projects rely on fresh-frozen biospecimens in their studies, which are difficult to obtain and often lack long-term clinical data. Unlike fresh-frozen samples, archived formalin-fixed paraffin-embedded (FFPE) tissues are more readily accessible, and are often associated with known clinical outcomes and more complete clinical annotations. However, sequencing library preparation methods need to be further optimized with regard to applicability to FFPE samples. To test whether next-generation sequencing technologies could overcome previously reported artifacts associated with formalin fixation and report accurate sequencing results, we compared whole-genome and targeted enrichment DNA sequencing data obtained from five FFPE tumor samples for which matching frozen tissues were available. We successfully generated sequencing libraries using 1 µg of genomic DNA as input material and a modified TruSeq™ sample preparation protocol. We designed a custom enrichment pool targeting exons from 215 cancer-related genes and utilized this pool to pull-down and sequence approximately 1.3 Mb region of the genome with an average 400x coverage depth to test if deep sequencing would improve validation of tumor-specific somatic mutations. We used the Illumina sequence analysis pipeline and the Windows-based second generation DNA sequencing software NextGENe® to analyze the data and identify sequence variations that are different from the human reference genome. CNV detection was overall higher among FFPE samples compared with fresh-frozen samples, demonstrating that tissue processing impacts sequencing data quality. We obtained good concordance in variant calls between matched FFPE and fresh-frozen samples. Discordant variant calls were mainly due to low depth of coverage in the regions where variant calls were made. Improvements to DNA sequencing methods for archived samples will significantly enhance cancer research and will result in more reliable prediction of individual cancer therapies. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 3179. doi:1538-7445.AM2012-3179

Abstract Background: Gene expression profiling is an important tool in prediction, prognosis, and cancer modeling. The DASL® assay (cDNA –mediated Annealing, extension, Selection, and Ligation; Illumina; San Diego, CA) is a gene expression profiling system designed for degraded RNAs such as those derived from formalin-fixed paraffin-embedded (FFPE) tumor samples. FFPE tumor samples are invaluable resources for cancer research, being the most widely available materials for which patient outcomes are known. Purpose: The objective of this study was to evaluate the performance of Illumina's Whole Genome DASL platform (24,526; 24K) at distinguishing HER2-positive (HER2+) from HER2-negative (HER2-) FFPE breast tumors. Methods: We measured the relative gene expression of biological and technical replicates using nine HER2+ and 11 HER2- primary breast tumors from patients that were all node-positive and estrogen receptor-positive. The HER2 status of the tumors was determined using the Dako Hercept Test following the 2007 ASCO/CAP guidelines for HER2 positivity. The non-background corrected probe-expression data were exported from Illumina BeadStudio software and normalized using the software R and the function fastlo. Log2 transformed expression values were used to obtain the relative gene expression per probe. Spearman correlations were calculated between both types of replicates. Unequal-variance t-tests were conducted to test for differences in expression levels between HER2+ and HER2- tumors. An overall ERBB2 gene expression level was calculated by taking the average of the three ERBB2 probes for each tumor across all replicates. To simulate the possible outcome if only one replicate of each tumor was measured for gene expression, differential expression was tested for all possible combinations of single replicate probe expression values. Results: Spearman correlations for technical and biological replicates within tumor samples ranged from 0.9804 – 0.9946, and 0.9822 – 0.9945, respectively. The overall ERBB2 gene expression was significantly lower in HER2- tumors than in HER2+ tumors (p = 0.0000018). The top five most differentially expressed genes (p < 0.0005) included ERBB2, GRB7, C17ORF37, PERLD1, and ASF1B. For each of the three ERBB2 probes, there were 65,536 possible ways to combine the individual probe expression values. The p-values for testing differential expression in each of these 65,536 combinations of HER2+ tumors versus HER2- tumors ranged from 0.00029 – 0.0091 for ILMN_1717902, 0.0000078 – 0.00026 for ILMN_1728761, and 0.000000002 – 0.0000032 for ILMN-2352131. Conclusions: Illumina's Whole Genome Platform provided highly reproducible gene expression values among biological and technical replicates of breast tumors. This platform also demonstrated significant differential ERBB2 gene expression between HER2+ and HER- breast tumors, which correctly classified the breast tumors as HER2+ and HER2-. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 6127.