Validated In-Vitro Toxicity Testing of Graphene Nanoplatelets, a Methodological Challenge

Graphene-family nanomaterials (GFNs) belong to the group of carbon-based synthetic nanomaterial, which are already on the market in multiple variants. Graphene nanoplatelets (GNPs) thereby represent special nanostructures with nanothickness and interesting multifunctional properties, e.g. improvement of mechanical strength and conductivity of composite materials. Currently, existing data do not point to excessive toxicity of GNPs, but meaningful data are scarce and need expansion for proper risk assessment. In the PLATOX project, funded by the European FP7 SIINN ERA-NET program on Nanosafety, typical, commercially available GNPs (ACS Material) were selected, comprising single layer graphene and graphene oxide, multilayer and carboxyl graphene, and graphite oxide nanoplatelets. The GNPs were first analyzed regarding sterility/endotoxin content, morphology (SEM pictures) and specific surface (BET method). Primary rat alveolar macrophages (AM) and human MRC-5 lung fibroblast cells were used as in vitro screening models to detect cell-type specific toxicity. Membrane damage (LDH release) and metabolic activity (AlamarBlue test) served as first screening endpoints. For cell exposure, GNPs were suspended in culture medium by ultrasonication and doses on a µg/cm 2 basis. But, OD600 measurements subsequently showed that the GNPs exhibited unequal sedimentation kinetics, thus limiting comparison of toxic effects on a mass basis. Furthermore, the assays had to be strongly adapted to GNPs, due to e.g. LDH binding to GNPs and disturbance of photometric measurements. In AM, the single layer GNPs GN1P0005 and GN1PF010 markedly enhanced LDH release after 24h of incubation (BMD30 values: 3.22 and 2.47 µg/cm 2 ). Enhancement in LDH release was also observed for single layer graphite oxide (GN0S0010; BMD30: 39.25 µg/cm 2 ) and experimental pristine GNPs (BMD30: 45 µg/cm 2 ). All other materials were nearly inactive. In contrast, no such effect was evident in MRC-5 cells. Significant effects on metabolic function were absent. But in AM, single layer GNPs furthermore induced DNA damage and PGE 2 release. In conclusion, in vitro, single layer GNPs seem to possess (geno)toxic potential in AM, but not in fibroblasts. Notably, in-vitro assays for hazard estimation of GNPs have to be carefully validated to ensure meaningful data.