(T1230-06-31) A Technical Evaluation of Targeted Protein Quantification in Formalin Fixed Paraffin Embedded Tumor Sections by Mass Spectrometry Including Antibody Drug Conjugates (ADC)

Purpose: Mass spectrometry is quantitative, selective, and capable of measuring dozens of proteins in multiplex from a single sample using targeted methods. Moreover, mass spectrometry is a compatible technique for the analysis of formalin fixed paraffin embedded (FFPE) tissue samples. FFPE tumor biopsies are an invaluable resource for looking at either baseline levels or changes in tumor and therapy associated proteins in a clinical trial. In this study, we conducted a technical evaluation of a targeted mass spectrometry platform and a 26-protein panel analysis in FFPE tumor samples. The developed protein panel included both antibody-drug conjugate (ADC) targets such as TROP2, ROR1 and HER2, as well as a number of low abundance immune cell targets.

Methods: To prepare samples, whole slides of FFPE tissue were scraped into a tube for protein extraction and digestion with trypsin. Surrogate peptides for each protein were quantified by liquid chromatography mass spectrometry with stable isotope labeled standards. All tumor samples used in this evaluation were procured from commercial sources and were not part of any company sponsored clinical trials. Analytical reproducibility was assessed using BxPC3 and Jeko-1 cell line FFPE blocks. Technical replicates (n = 6), each combining 2 serial 5mm depth sections from each cell line block were analyzed and the percent coefficient of variance (% CV) was calculated for each protein. The variation in protein concentration across adjacent sections of a single FFPE tumor was evaluated. Consecutive sections (n = 6) of 20mm depth each (4 serial slides of 5mm depth per section) of a single non-small cell lung (NSCLC) tumor biopsy were analyzed. The variation in protein concentration across different NSCLC tumors (or interspecimen) was evaluated. A 20mm depth section from 10 different tumor biopsies were analyzed. Likewise, interspecimen variation was evaluated in other tumor types, namely pancreatic, colorectal, ovarian, bladder, gastric, and rectal.

Results: The majority (15 out of 26) of the targeted proteins were quantifiable in the cell line replicates. The mean protein concentrations correlated well (R² = 0.764, p < 0.0001) with publicly available RNA-seq values. Less than 15% and 25% CVs (n = 6) were observed in target protein normalized to either β-actin or total protein, respectively, with the exceptions of PD-L1, BCMA and VISTA, which were the 3 proteins lowest in abundance. In adjacent sections of a NSCLC tumor, a different set of 15 (out of 26) proteins were quantifiable and mean protein concentrations spanned 3 orders of magnitude. The variation across adjacent sections of NSCLC tumor was less than 12% and 20% CV normalized to β-actin and total protein, respectively. LAG3, PD-L1, PD-L2, and CD8A were only quantifiable in some, but not all of the adjacent sections and were among the lowest abundant proteins in the tumor. When looking in multiple NSCLC tumors (or interspecimen), 17 proteins (out of 26) were quantifiable in 9 or more of the 10 tumors. CD226, CDK4, TIGIT, BCMA, and CD19 were only quantifiable in 4 or fewer of the 10 NSCLC tumors. TIM3, CTLA4, and PVRIG were not quantifiable in any tumors. The level of interspecimen variation in 10 tumors was determined in NSCLC and in the other tumor types analyzed.

Conclusion: The platform evaluated to quantify proteins from FFPE tumors by mass spectrometry was found to be selective and sensitive. A majority of the 26 proteins in panel, which included ADC targets TROP2, ROR1 and HER2, as well as a number of low abundance immune targets, were successfully quantified from a clinically relevant amount of material in many tumors of diverse types. Moreover, the platform generated protein measurements with acceptable analytical precision, reasonable correlation to RNAseq and minimal variation across adjacent tumor sections.