Recent evidence supports that alterations in hepatocyte-derived exosomes (HDE) may play a role in the pathogenesis of drug-induced liver injury (DILI). HDE-based biomarkers also hold promise to improve the sensitivity of existing in vitro assays for predicting DILI liability. Primary human hepatocytes (PHH) provide a physiologically relevant in vitro model to explore the mechanistic and biomarker potential of HDE in DILI. However, optimal methods to study exosomes in this culture system have not been defined. Here we use HepG2 and HepaRG cells along with PHH to optimize methods for in vitro HDE research. We compared the quantity and purity of HDE enriched from HepG2 cell culture medium by 3 widely used methods: ultracentrifugation (UC), OptiPrep density gradient ultracentrifugation (ODG), and ExoQuick (EQ)-a commercially available exosome precipitation reagent. Although EQ resulted in the highest number of particles, UC resulted in more exosomes as indicated by the relative abundance of exosomal CD63 to cellular prohibitin-1 as well as the comparative absence of contaminating extravesicular material. To determine culture conditions that best supported exosome release, we also assessed the effect of Matrigel matrix overlay at concentrations ranging from 0 to 0.25 mg/ml in HepaRG cells and compared exosome release from fresh and cryopreserved PHH from same donor. Sandwich culture did not impair exosome release, and freshly prepared PHH yielded a higher number of HDE overall. Taken together, our data support the use of UC-based enrichment from fresh preparations of sandwich-cultured PHH for future studies of HDE in DILI.