The PLATOX project: Combining in vitro and in vivo investigations to generate valid toxicity data for risk assessment of graphene nanoplatelets

Carbon-based nanoplatelets (CNP) represent a new class of 2-D nanostructures in multiple variants and with interesting functional properties (e.g. material enforcement and electrical conductivity). The hazard characterization ofCNP is still incomplete. Commercial CNP candidates (ACS Material, USA) were selected, covering single layer (SLG) or multi layer graphene (MLG), carboxyl graphene, single layer graphene oxide, and graphite oxide. Technical soot (Printex 90) served as non-platelet reference. Sterility and endotoxin content of the CNP were analysed and the morphology (SEM pictures) and the specific surface area (BET method) were reevaluated. As in vitro screening models both, primary rat alveolar macrophages (AM) and MRC-5 human lung fibroblast cells were analysed on membrane damage (LDH release) and metabolic activity (AlamarBlue test). Interestingly, the two SLG induced marked concentration-dependent membrane damage in AM after 24h of incubation, with a BMD30 of 3.2 and 2.5␮g/cm2, whereas no such effect was observed for MRC-5 cells. Some LDH release was also shown for single layer graphite oxide (BMD30: 39.3␮g/cm2). The other materials were nearly inactive. Significant effects on metabolic activity were not observed. In AM, SLG additionally induced direct DNA damage and release of PGE2 and TXB2. In conclusion, SLG showed a (geno)toxic potential in vitro in AM, but not in lung fibroblasts. Based on the in vitro screening data and for validation, an SLG (highest) and an MLG (lowest toxic potential) were selected for in vivo investigations. In a dose range finding (DRF) test, with dosing by intratracheal instillation (0.02 and 0.2 mg/rat, each) SLG was confirmed as the stronger inflammogenic sample (bronchoalveolar lavage fluid - BALF analysis) inducing a significant recruitment of neutrophils and - surprisingly - of eosinophils. In a subsequent 4-week nose-only inhalation study with the same total doses (predicted by MPPD model) the pro-inflammatory effect of SLG was weaker and no eosinophils were detected in BALF. Again, SLG but not MLG exhibited some inflammogenic potential. The histopathological examination mirrored the results of the BALF analysis. Granulocyte infiltration and giant cells were observed only in the SLG high dose group. - Funding: FP7- ERA-NET SIINN - BMBFPtJ.