Abstract 2707: Integrated analysis of colorectal carcinoma by co-extraction of RNA, DNA and protein from FFPE tumor samples

Background: Integrated analysis of multi-omic data is important in understanding oncogenesis, pathology, and drug mechanism of action. Formalin-fixed, paraffin-embedded (FFPE) tissues comprise the bulk of archival specimens in hospitals and clinical trials. Commercial kits are commonly used to extract RNA, DNA, or protein for biomarker studies; however, high-quality extractions are challenging due to crosslinking. We explored the feasibility of performing proteomic, transcriptomic, and genetic analysis from limited FFPE samples collected in clinical trials through a pilot study in colorectal cancer (CRC) and normal tissue. Methods: Qiagen kits were employed, with some modifications, to extract nucleic acid and protein from FFPE sample lysates that would be amenable to next-generation sequencing and liquid chromatography/mass-spectrometry (LC/MS) proteomic profiling. Commercially sourced FFPE slides from CRC biopsies and normal colon, heart, and skin samples were processed (n = 10 each; ~600-mm 2 by 5-μm/slice). RNA-seq library preparation was performed using the Illumina ® TruSeq Access kit. Label-free LC/MS proteomic analysis was carried out using trypsin/Lys-C-digested protein fragments on an EASY-nLC™ 1200 coupled to a Q Exactive™ Plus (Thermo Fisher). MS data were processed with MaxQuant to estimate protein abundance (Cox and Mann, Nat Biotech 2008). Results: Of 40 samples tested, 38 yielded quantifiable nucleic acid and protein of sufficient quality for transcriptional and proteomic analyses. Mean RNA yield was ~300 ng (150 ng-1.2 µg). RNA degradation was significant, with a mean DV >200 of 45% (Agilent Bioanalyzer). Mean DNA yield was 530 ng, with a mean fragment size of 2 kb. Mean protein yield was 50 µg (1 µg-300 µg). RNA libraries had a mapping rate range of 85%-90% and a coding rate range of 70%-80%; ~16,000 genes were reliable detected (log 2 TPM > 1). The number of unique proteins detected ranged from 3,000 to 4,000 for CRC, normal colon, and heart samples; skin samples averaged 1,000 proteins and had the lowest overall yield. Correlation of RNA and protein expression for the same genes was weak (Spearman9s rho ~0.3-0.4) at the per-sample level but statistically significant. Using linear models, we identified differentially expressed transcripts and proteins between the CRC and normal colon samples. Pathway enrichment analysis of both modalities indicated changes in cell cycle, which is consistent with rapid growth of tumor cells. Changes in nucleoside metabolism, extracellular matrix remodeling, and innate immune response were more apparent at the protein level. Conclusion: We found that multi-omic analysis was feasible with FFPE samples, and proteomics can be used to validate RNA results. Additionally, proteomics reveal post-translational events, such as extracellular matrix remodeling, that provide unique insights into cancer pathology. Citation Format: Vishal Patel, Ji Gao, Mingyi Liu, Aiqing He, Xi-Tao Wang, Kathryn Vanderlaag, Ariella Sasson, Stefan Kirov, Sunil Kuppasani, Omar J. Jabado, Kandasamy Ravi, Ashok Dongre, Julie Carman, Heidi LeBlanc. Integrated analysis of colorectal carcinoma by co-extraction of RNA, DNA and protein from FFPE tumor samples [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 2707.