Graphene microsheets enter cells through spontaneous membrane penetration at edge asperities and corner sites

Abstract

Understanding and controlling the interaction of graphene-based materials with cell membranes is key to the development of graphene-enabled biomedical technologies and to the management of graphene health and safety issues. Very little is known about the fundamental behavior of cell membranes exposed to ultrathin 2D synthetic materials. Here we investigate the interactions of graphene and few-layer graphene (FLG) microsheets with three cell types and with model lipid bilayers by combining coarse-grained molecular dynamics (MD), all-atom MD, analytical modeling, confocal fluorescence imaging, and electron microscopic imaging. The imaging experiments show edge-first uptake and complete internalization for a range of FLG samples of 0.5- to 10-μm lateral dimension. In contrast, the simulations show large energy barriers relative to kBT for membrane penetration by model graphene or FLG microsheets of similar size. More detailed simulations resolve this paradox by showing that entry is initiated at corners or asperities that are abundant along the irregular edges of fabricated graphene materials. Local piercing by these sharp protrusions initiates membrane propagation along the extended graphene edge and thus avoids the high energy barrier calculated in simple idealized MD simulations. We propose that this mechanism allows cellular uptake of even large multilayer sheets of micrometer-scale lateral dimension, which is consistent with our multimodal bioimaging results for primary human keratinocytes, human lung epithelial cells, and murine macrophages.

Keywords: corner penetration; edge cutting; graphene-cell interaction; lipid membrane; molecular dynamics simulation.

Fig. 1.

(A–H) Coarse-grained molecular dynamics simulations of interactions between a lipid bilayer and (A–D) a small graphene flake or (E–H) a large five-layer graphene sheet with staggered stacking and roughened edge topography. (I) The normalized free energy of the system as a function of the graphene orientation when one of the sharpest corners is fixed at a distance of 0.5 nm above the bilayer. Note that A–D and F–H are time sequences; E is an experimental graphene edge structure (–36).

Fig. 2.

All-atom molecular dynamics simulations of corner piercing of a monolayer graphene across a lipid bilayer. (A) Simulations directly showing that the corner piercing proceeds spontaneously. (B) Graphene–bilayer interaction energy as a function of the penetration distance, showing the existence of an energy barrier of about 5k_BT associated with corner piercing. The mean value of interaction energy is obtained from 11 independent simulation runs and the error bars show SD. The relatively large fluctuations of interaction energy at large penetration distance are mainly due to random translational and rotational movements of graphene relative to the bilayer membrane and random configurational changes of individual lipids adjacent to the graphene. (C) Analytical model of corner piercing.

Fig. 3.

Cellular uptake and internalization of few-layer graphene microsheets. (A–C) Confocal images of human lung epithelial cells (A and B) and mouse macrophages (C) exposed to graphene microsheets (0.5- to 25-μm lateral dimension) after 24 h and 5 h, respectively. The nuclei in A and B are visualized (blue fluorescence) with 4′,6-diamidino-2-phenylindole (DAPI). The microtubules of the lung epithelial cells (A and B) are visualized using antitubulin beta antibodies conjugated with FITC (green fluorescence), whereas the actin cytoskeleton of macrophages shown in C is visualized using rhodamine–phalloidin (red fluorescence). In unexposed lung epithelial cells (A and B, Inset), cytoplasmic microtubules (MT) form a linear network spanning across the cytoplasm. Internalized graphene flakes (yellow arrows, A and B) physically displace the linear microtubular network. In unexposed macrophages (C, Inset), filamentous actin (F) is organized into aggregates beneath the plasma membrane. Internalized graphene flakes with large lateral dimension (yellow arrow, C) induce dense aggregates of actin filaments whereas submicron graphene sheets (yellow arrowhead, C) do not disrupt the actin cytoskeleton. Transmission electron micrographs of macrophages (D) and lung epithelial cells (E) exposed to 10 ppm FLG sheets (∼800 nm in lateral dimension) for 5 h and 24 h show localization in the cytoplasm within membrane-bound vacuoles (blue Insets). Graphene microsheets inside vacuoles appear as electron-dense linear sections (D, Inset) or irregular flakes (E, Inset).

Fig. 4.

Cell membrane interactions with graphene microsheets, showing edge or corner penetration for each of three cell types. (A) Corner penetration observed for a graphene sheet of micrometer-scale lateral dimension on the surface of a human lung epithelial cell at low and high magnification. (B) Edge penetration of multiple microsheets (G) into a macrophage (M). (C) Edge penetration for a 5-μm sheet interacting with primary human keratinocytes, in which the edge entry appears to have been nucleated at an asperity or protrusion (thick yellow arrow). (D) Corner penetration mode at the surface of a primary human keratinocyte. The tilted, upright orientation of the graphene sheet produces subtle e-beam shadows immediately adjacent to the sheet in some images. Highly irregular edge topography is seen on essentially all graphene sheets. All images are field-emission scanning electron micrographs of fixed cells with osmium tetroxide postfixation. Exposure times are 24 h except for that of macrophages, which is a 5-h exposure. Cells in A and B were not subjected to critical point drying during sample preparation for scanning electron microscopy. The graphene microsheets here and in Fig. 3 have layer numbers that range from 4 to 25. (Scale bars, 2 μm.)