Background
Immunotherapies, like PD-1 inhibitors, have revolutionised the treatment of people diagnosed with metastatic melanoma. However, over half of patients do not respond to what is an expensive and sometimes toxic therapy. PD-L1-EV quantitation has been touted as a promising biomarker for diagnosis, prognosis and importantly, as predictors of immunotherapy response. We aimed to recapitulate these findings in a small New Zealand based cohort.
Methods
EV isolation using sequential (ultra)centrifugation was compared to Size Exclusion Chromatography (SEC) for particle and protein yield. SEC was selected to isolate EVs from 2mL plasma taken at two timepoints before (P1) and during (P2; 6 weeks) immunotherapy for 19 consented patients (approval #16/NTA/180). EV PD-L1 was detected using an optimised commercial ELISA.
Results
EV PD-L1 was limited in quantity, near the lower limit of the ELISA. PD-L1 normalised to particle count at P1 was associated with disease specific survival (P=0.035), but not at P2. Significance was however highly dependent on the normalisation approach for PD-L1 abundance. In contrast to prior studies, EV-PD-L1 was not associated with RECIST response to immunotherapy. Careful review of 13 publications highlighted significant variations in methods and data analyses which likely underly the discordance seen between studies regarding for example the overall yields of PD-L1 on EVs and whether they have biomarker utility.
Conclusions
Current studies using PD-L1 EVs as biomarkers for immunotherapy response are widely discordant. Standardisation of protocols and cohort expansion are required to answer the question of whether PD-L1 EVs are truly clinically applicable biomarkers.