Abstract 3885: Peptide-mediated ‘miniprep’ isolation of extracellular vesicles for high-throughput proteomics; method evaluation and application in colon cancer

• Published in: Cancer research (Chicago, Ill.) Ill.), 2016-07, Vol.76 (14_Supplement), p.3885-3885
• Main Authors: Jimenez, Connie R., de Wit, Meike, Knol, Jaco C., de Reus, Inge, Schelfhorst, Tim, Bishop-Currey, Logan, Dusseldorp, Valerie, van Grieken, Nicole, Beekhof, Robin, Piersma, Sander R., Pham, Thang V., Smit, Egbert F., Fijneman, Remond JA, Meijer, Gerrit, Verheul, Henk MW
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2016-3885

Background: Extracellular vesicles (EVs) are cell-secreted membrane vesicles enclosed by a lipid bilayer derived from endosomes or from the plasma membrane. Since they are released into body fluids, and their cargo includes tissue-specific and disease-related molecules, EVs represent a rich source for disease biomarkers. However, standard ultracentrifugation methods for EV isolation (UC-EV) are laborious, time-consuming, and require high inputs. Recently a novel isolation method was described, which can be performed at small ‘miniprep’ scale, utilizes specific Heat Shock Protein (HSP)-binding peptides to aggregate HSP-decorated EVs (Ghosh et al. (2014), PLoS ONE 9:e110443). Using cancer secretome and biofluid samples, the authors showed enrichment of exosome markers for their method (abbreviated HSP-EV here) but a detailed description of the captured EV proteomes in comparison to the gold-standard ultracentrifugation (UC) method and application to small tumor proximal fluid samples is lacking. Approach: Here we used label-free proteomics of replicate EV isolations from HT-29 cancer cell-conditioned medium to compare EV fractions captured using the new HSP peptide method and UC. Subsequently we applied this novel method to profile EVs released from fresh human colorectal tumors (CRC) (n = 17) and colon adenoma (n = 4) tissue as well as patient-matched normal colon tissue. Results: Despite a 30-fold different input scale (UC-EV: 60 ml versus HSP-EV: 2 ml), both methods yielded comparable numbers of identified proteins (3115 versus 3085), with reproducible identifications (72.5% versus 75.5%) and spectral count-based quantification (average CV 31% versus 27%). EVs obtained by either method contained established EV markers and proteins linked to vesicle-related gene ontologies. In the EV fraction of the tissue secretomes 6390 proteins were identified, of which 471 proteins were significantly 5-fold more present in CRC samples than in normal tissue EVs. Gene ontology analysis revealed enrichment of nuclear proteins involved in DNA damage response, chromosome organization and RNA processing in the CRC EVs. Conclusion: The HSP-EV method provides an advantageous, simple and rapid approach for EV isolation from small amounts of biological samples, enabling high-throughput analysis in a biomarker discovery setting. Citation Format: Connie R. Jimenez, Meike de Wit, Jaco C. Knol, Inge de Reus, Tim Schelfhorst, Logan Bishop-Currey, Valerie Dusseldorp, Nicole van Grieke