The paper in Nature Communications is here: http://go.nature.com/2FqW45A
For chemists and physicists, the concept of the observer effect is often introduced by way of the classical Davisson-Germer experiment. As modern experimental methods grow ever more complex, careful calibration is necessary to understand any effect on what is being measured. Although our work grew out of the quest to understand observer effects, it started as something else altogether. Traditional probes used to detect neuronal activity are built on silicon hardware, and although very effective over increasingly longer timeframes, are limited by the spatial localization of neurons relative to a recording stud. We and our colleagues initially began collaborating to determine the feasibility of using graphene-based transistors for photocurrent measurements to detect neuronal activity, something that would overcome the spatial limitations of recording stud placement.
We initially intended to perform a fairly quick characterization of neurons grown on monolayer graphene, the configuration used in photocurrent measurements. The relative biocompatibility previously observed in eukaryotic cells suggested that this may be straightforward. As the ultimate benchmark for our intended application would be effects on electrical activity, our line of investigation began with electrophysiological recordings of neurons on graphene. We observed an increase in neuronal firing frequency on graphene; this was substantial and our curiosity was piqued. When we began to look at morphology and more targeted measurements of presynaptic activity, what consistently emerged was surprising: synaptic potentiation on graphene.
Synaptic potentiation on CVD graphene.
Although these results were interesting, and enough to provide a broad characterization of what might be a concern when interpreting photocurrent scanning measurements, we took a step back to think about why and how potentiated neurotransmission might be occurring. Here we took clues from previous research, where the biological effects of graphene have remained an open question. Some types of prokaryotic cells are irreparably damaged after graphene exposure, while use of graphene as a substrate has minimal or even sometimes somewhat beneficial effects on eukaryotic cells. When considering the difference between membrane types, we found it interesting to note that eukaryotic membranes contain cholesterol while prokaryotic membranes largely lack cholesterol. Might cholesterol somehow be involved? This led us to the idea of manipulating membrane cholesterol levels to elucidate the role of cholesterol in synaptic potentiation. As the effects of graphene, mediated by cholesterol, should be generalizable, we sought to extend our results. Cholesterol plays important and fundamental roles to cells in a variety of ways, but with particular relevance to drug discovery we demonstrated the effect of graphene on G protein-coupled receptors, whose activity are mediated by cholesterol.
Although our results demonstrate the role of cholesterol in the functional effects we observe, there may yet be roles for other lipids or even proteins in modulating the cellular effects of graphene. The plasma membrane is a complex and dynamic microenvironment – as is a protein layer known as a ‘corona’ that forms around nanomaterials when exposed to biological systems. Although there have been recent developments, future studies would be aided by advances in the ability to visualize cholesterol and other membrane biomolecules natively at high spatiotemporal resolution.