B-255 Identification of the protein expression of Androgen Receptor variants using Targeted proteomics in clinical Castration Resistant Prostate Cancer models

Abstract Background Castration Resistant Prostate Cancer (CRPC) is a treatment resistant form of prostate cancer (PCa). Currently, there is not a way to identify which patients will develop this resistance before full blown CRPC develops. Therefore, all PCa treatment approaches are similar yet this results in tumor regression in some cases and progression in others. Targeting the androgen receptor (AR) is still the main focus of current therapies even in CRPC. Emergence of AR splice variants (AR-Vs) after initial treatment is thought to be one of the primary mechanisms of resistance. AR-Vs lack the ligand binding domain rendering Androgen Deprivation Therapy (ADT) ineffective in tumors expressing these variants. Recent work has identified the DNA and RNA species of ARVs in CRPC but investigation into whether the protein is translated are unknown. One exception is the approval of an antibody test that detects a specific AR-v, AR-v7, from the blood of PC patients and predicts ADT response. However, there are instances where a patient may not express ARv7 and are still resistant to ADT. This may suggest that other AR-Vs, not currently detected at the protein level, are important predictors for ADT response. Methods The primary goal of this work is to expand the protein identification of known and unknown AR-Vs that may predict response to ADT using targeted liquid chromatography tandem mass spectrometry (LC-MS/MS). We developed the AR-V targeted LC-MS/MS assay by first designing AR splice variants from unique splice regions via RNA-guided sequences translated to amino acid sequences. The AR-variants full amino acid sequences were then in-silico trypsin digested using Expasy software tool that generated tryptic peptides. Blast analysis of the AR-V peptides was performed where no overlap with other human proteins was observed. These AR-V peptide sequences were commercially synthesized along with heavy isotopically analogs to be used as internal standards in the LC-MS/MS method. The method examined 9 peptides and included 2 to 3 MRM transitions per peptide and calibration curves were constructed using unlabeled peptides. Specimens examined consisted of a panel of PCa cell lines (n= 6), CRPC patient derived xenografts tumors (PDX) (n=4) and a negative control cell line (no androgen receptor expression), Cos-1, which served with a dual function and was used as matrix for peptide MRM optimization experiments. Linearity and sensitivity for each peptide included in the method was examined using calibration curves. Results The targeted MRM approached was successfully achieved for the quantification of 9 peptides from 8 proteins in 6 PCa cell lines and 4 PDX tumor specimens. Out of the 8 proteins evaluated, we observed 3 known targets (AR-exon1, AR-V7 and AR-V12) and 4 novel AR-Vs in 22Rv1 PCa cell lines. Additionally, CRPC PDX were assessed where 3 novel AR-Vs were detected. Conclusions These preliminary results using a LC-MS/MS platform show promising identification and quantification of novel AR-Vs that have not been measured before at the protein level. Subsequent analysis of larger cohorts will further elucidate AR-Vs role in PCa treatment resistance mechanisms and possibly assist in future treatment approaches.