Cancer cell-derived EVs can be used as effective carriers of Paclitaxel to their parental cells, bringing the drug into the cells through an endocytic pathway and increasing its cytotoxicity. However, due to the increased cell viability, the use of cancer cell-derived EVs must be further investigated before any clinical applications can be designed.

Current analysis of exosomes focuses primarily on bulk analysis, where exosome-to-exosome variability cannot be assessed. In this study, we used Raman spectroscopy to study the chemical composition of single exosomes. We measured spectra of individual exosomes from 8 cell lines. Cell-line-averaged spectra varied considerably, reflecting the variation in total exosomal protein, lipid, genetic, and cytosolic content. Unexpectedly, single exosomes isolated from the same cell type also exhibited high spectral variability. Subsequent spectral analysis revealed clustering of single exosomes into 4 distinct groups that were not cell-line specific. Each group contained exosomes from multiple cell lines, and most cell lines had exosomes in multiple groups. The differences between these groups are related to chemical differences primarily due to differing membrane composition. Through a principal components analysis, we identified that the major sources of spectral variation among the exosomes were in cholesterol content, relative expression of phospholipids to cholesterol, and surface protein expression. For example, exosomes derived from cancerous versus non-cancerous cell lines can be largely separated based on their relative expression of cholesterol and phospholipids. We are the first to indicate that exosome subpopulations are shared among cell types, suggesting distributed exosome functionality. The origins of these differences are likely related to the specific role of extracellular vesicle subpopulations in both normal cell function and carcinogenesis, and they may provide diagnostic potential at the single exosome level.

Cholesterol is an important component of all biological membranes as well as drug delivery liposomes. We show here that increasing the level of cholesterol in a phospholipid membrane decreases surface charge in the physiological environment. Through molecular dynamics simulation we have shown that increasing the level of cholesterol decreases Na+ ion binding. Complementary experimental ζ – potential measurements have shown a decreased ζ – potential with increasing cholesterol content, indicative of reduced surface charge. Both experiments and simulations have been carried out on both saturated 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) and monounsaturated 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) membranes. This result is particularly important because membrane surface charge plays an important role in the interactions of biomembranes with peripheral membrane proteins and drug delivery liposomes with the immune system.

Two types of amphiphilic gold nanoparticles (AuNP-1 and -2) grafted with a mixture of poly(N-isopropylacrylamide) (PNIPAM) and polystyrene (PS) chains in two different compositions have been successfully prepared with the "grafting-to" method in a homogeneous THF phase. These AuNPs were thoroughly characterized by FTIR, 1 H NMR, UV-vis, high-resolution transmission electron microscopy, thermogravimetric analysis, and dynamic light scattering to determine the total number of polymer chains bound to the gold nanoparticles, the ratio between PNIPAM and PS chains, and the size of the gold core. Langmuir monolayer experiments at the air-water interface of the two types of AuNPs revealed different compression isotherms of the surface pressure vs particle area (π-A curve) conducted at 20 °C. These amphiphilic gold nanoparticles can be regarded as analogues of amphiphilic diblock copolymers at the air-water interface. The compression isotherm of AuNP-2 with a PNIPAM:PS ratio of 2:1 showed several characteristic regions that can be attributed to the polymer conformational transitions from the pancake, the pancake to brush transition, to the brush. However, the monolayer of AuNP-1, with a ratio of 5:1 of PNIPAM:PS, never reaches a brush stage but showed an extension of the pseudoplateau region upon compression. These differences may be due to the more hydrophilic nature and the more stretched PNIPAM chains. Furthermore, the sessile drop contact angle measurements, conducted at room temperature on both upper and lower surfaces of the AuNP-2 monolayer transferred at 35 mN/m onto either hydrophilic or hydrophobic substrates, are slightly different, 82 ( 2°and 77 ( 2°, respectively. After comparing with the literature data of the contact angles of water on either the pure PS film or the PNIPAM brush, we concluded that the chemically different PNIPAM and PS chains grafted on the surface of the gold core tend to be phase-separated.

Combined experimental and computational studies of lipid membranes and liposomes, with the aim to attain mechanistic understanding, result in a synergy that makes possible the rational design of liposomal drug delivery system (LDS) based therapies. The LDS is the leading form of nanoscale drug delivery platform, an avenue in drug research, known as "nanomedicine", that holds the promise to transcend the current paradigm of drug development that has led to diminishing returns. Unfortunately this field of research has, so far, been far more successful in generating publications than new drug therapies. This partly results from the trial and error based methodologies used. We discuss experimental techniques capable of obtaining mechanistic insight into LDS structure and behavior. Insight obtained purely experimentally is, however, limited; computational modeling using molecular dynamics simulation can provide insight not otherwise available. We review computational research, that makes use of the multiscale modeling paradigm, simulating the phospholipid membrane with all atom resolution and the entire liposome with coarse grained models. We discuss in greater detail the computational modeling of liposome PEGylation. Overall, we wish to convey the power that lies in the combined use of experimental and computational methodologies; we hope to provide a roadmap for the rational design of LDS based therapies. Computational modeling is able to provide mechanistic insight that explains the context of experimental results and can also take the lead and inspire new directions for experimental research into LDS development. This article is part of a Special Issue entitled: Biosimulations edited by Ilpo Vattulainen and Tomasz Róg.