Matthew Loya, MBS (Genmab)
Molecular profiling of formalin-fixed, paraffin-embedded (FFPE) tumor biopsies is routinely performed by low-plex immunohistochemistry or high-plex RNA transcriptomics. Unbiased, high-plex proteomics by mass spectrometry (MS) with FFPE is currently limited by high sample input requirements, low throughput, and complex bioinformatics needed to make sense of the data.
Our company, Genmab, develops antibody therapeutics for solid tumors and hematological malignancies. Directly characterizing the protein levels from limited patient samples can be used to enhance our understanding of target distribution and to build patient stratification or drug combination strategies with more confidence than RNA-based measurements.
We sought to test Biognosys’ promising mass-spec proteomics technology (directDIA) with FFPE tumor samples that would be found in a routine histopathology lab or clinical trial. These samples were hematological (diffuse b-cell lymphoma, follicular lymphoma, normal lymph nodes, and chronic lymphocytic leukemia [CLL]) or solid tumors (triple negative breast cancer, colorectal cancer, and pancreatic ductal adenocarcinoma). We tested a variety of sample sizes (resections or core needle biopsies) and types (unstained, hematoxylin and eosin [H&E] stained, and acid decalcified CLL bone marrow biopsy). Tissues were mounted on slides and processed using trypsin digestion, followed by peptide purification and direct-DIA analysis (n=54); a pool of FFPE tissue was used to control for injection order bias (n=6). Initial experiments to optimize tissue requirements showed single slides had high protein coverage, so all results reported here used single slides.
A key problem in traditional MS/MS proteomics is the high rate of missing data due to the stochastic nature of 2-step peptide detection. New methods, such as directDIA use an alternative detection strategy, yielding more competitive data at the cost of computationally intensive spectral deconvolution. We confirmed the advantage of direct-DIA in FFPE tissues with single slides: a total of ~12,000 proteins were detected, of which ~8,000 were detected in 90% of our diverse sample set. We compared H&E stained tumor slides adjacent to unstained slides and found a strong correlation of expressed proteins. Acid-decalcified CLL bone marrow biopsy yielded 6,800 detected proteins; these samples were also tested for transcriptomics, but RNA was not recovered. Key therapeutic targets in immuno-oncology were readily detectable (PDL1, PD1, LAG3, TIM-3, OX40, CD19, CD20, CD38) along with expected tumor antigens.
Efficient use of precious tumor biopsies with highly multiplexed proteomic profiling can supplement IHC and reduce the burden on patients. Applying the direct-DIA method to single slides, including previously H&E stained or decalcified, may allow the profiling of large tumor banks for translational research or routine use in clinical trials.
Short Bio Matthew Loya:
Matthew Loya is a Principal Research Associate at Genmab with 10 years of research experience. He earned a Bachelor of Science (BS) in Biology from John Carroll University and a Master of Business and Science (MBS), with a concentration in Drug Discovery and Development, from Rutgers University. He began his career in 2013 at the Cleveland Clinic where he was involved in immunotherapy and autoimmunity research. After 4 years in academia, he transitioned to the pharmaceutical industry where he joined the Translational Research department at Bristol Myers Squibb. There he provided program support for multiple immuno-oncology and immunology programs. In March 2021, he joined the Clinical Genomics team at Genmab where he performs transcriptomic profiling of FFPE samples from clinical trials. He has an innovative mindset and has worked closely with internal colleagues and collaborators at Biognosys to bring single FFPE slide proteomics online for Genmab.