Beyond Media Composition: Cell Plasma Membrane Disruptions by Graphene Oxide

In their letter, Bussy and Kostarelos reproduced aspects of our recent in vitro study of graphene oxide (GO)-induced mammalian cell plasma membrane (PM) ruffling and shedding and performed additional tests on the effect of different culture media. GO-induced acute responses of NIH/3T3 fibroblast cells are less pronounced in their selected media compared to two buffered salt solutions (BSS and PBS) commonly used in immunofluorescence and other cell studies in vitro. This fibroblast cell line was evaluated in both studies, and differences in membrane shedding were attributed by Bussy and Kostarelos to the varied electrolyte composition, which might affect the solution behavior of nanomaterials. Differences between cell types complicate unambiguous interpretation of their results for GO-treated A548 and BEAS-2B cells, but these cell lines also showed less pronounced responses when cultured in their selected media.

We are appreciative and intrigued by their results in light of our continuing efforts to understand GO-induced PM ruffling and shedding. Interpretations of effects from different extracellular solutions hinge upon the mechanism behind these elusive cell responses to GO. PM shedding through biophysical bending of the PM has seen numerous examples with biomacromolecules (McMahon and Gallop, 2005), yet not with nanomaterials. Additionally, GO may disrupt cell signaling at the PM (Chen et al., 2012). Because of GO’s promiscuity at the nano-bio interface, it remains unclear whether PM shedding involves specific cell signaling and active cellular processes. We observed inhibition of intracellular Ca²⁺ mobilization at GO concentrations relevant for inducing PM shedding (Sun et al., 2016) and suppression of degranulation in rat basophilic leukemia cells by GO (unpublished results). These observations suggest that GO affects cell signaling. However, it is unknown the degree to which any of these observations is relevant to PM shedding. Furthermore, PM ruffling and shedding likely involve disruptions of the linkages between the cytoskeleton and the PM. Different extracellular solutions may also alter cell responses by affecting these linkages. Therefore, changes of extracellular solutions may both affect the solution behavior of GO and alter cell physiology in response to GO as an irritant.

Serum-free DMEM and BSS contain similar monovalent (Na⁺, K⁺) and divalent (Mg²⁺, Ca²⁺) cations with comparable ionic strengths. However, DMEM contains diverse multi-valent anions such as sulfates and phosphates. Graphene oxide nanosheets are two-dimensional polyanionic macromolecules with extraordinary conformational flexibility in aqueous suspension (Poulin et al., 2016). Its solvation and interactions with biomolecules are likely affected by the electrolytes, such that electrolyte composition may change GO’s solution structure and the biological corona surrounding GO. Furthermore, we previously found that GO primarily distributes on the PM with insignificant internalization after 4 h (Sun et al., 2016). Changes of solvation environment may also affect the complex and ill-defined interface between GO and the cell surface.

In addition, serum-free DMEM contains nutrients such as essential amino acids and vitamins, which are absent in the minimal balanced salt buffer (BSS) that we originally used (Sun et al., 2016). We have used serum-free DMEM for GO treatment with MDA-MB-231 lung metastasized breast cancer cells and observed similar ruffling and shedding. But it was a different cell line and the extent of shedding was difficult to quantify under those experimental conditions. The minimal BSS media suited our original immunofluorescence study because the duration of the experiments was only several hours. However, nutrient supply has proven relevant in various contexts related to cell metabolism, genetic stability and particularly the highly adaptive cancer physiology (Lukey et al., 2016; Reynolds et al., 1996). For example, lung metastases from primary breast cancers show remarkable metabolic reprogramming in the lung microenvironment in vivo (Christen et al., 2016). Therefore, mammalian cells likely cope with environmental stress (in this case, GO) differently in different nutrient environments.

Overall, the effect of different extracellular solutions is an interesting phenomenon that requires careful investigation to differentiate between effects arising from electrolytes, nutrients, and other components. 2D nanomaterials capable of stimulating cells are now emerging. Graphene oxide is a non-lethal candidate that demonstrates intriguing effects for modulating cell functions. Mechanistic studies will be rewarding but challenging. We join Bussy and Kostarelos in promoting precise and detailed reports of experimental procedures when conducting experiments with nanomaterials and biological systems. We also advise caution in drawing overly broad conclusions about the biological effects of nanomaterials from a single study. Understanding and best practices will emerge from the combined conclusions of careful studies that both test and extend prior work. We view these Letters as a constructive part of this process.