Development of separation methods to produce uniform exosomes subpopulations using field-flow fractionation techniques

Abstract

Exosomes are nanometer-sized vesicles (30-120 nm) with a defined density range (1.08-1.22 g/cm3), secreted by most cells in the body, including tumors. Exosomes contain specific proteins, lipids, and functional microRNA (miRNA) and messenger RNA (mRNA) that can be transferred from cell to cell which suggests a role in cell-to-cell communication. Despite the tremendous strides that have been made in the last years toward understanding the properties and function of exosomes, this goal remains a challenging task. Exosomes have a potential role in early diagnosis of cancer but recent reports suggest controversial activities in cancer therapy. Exosomes have an anti-tumorigenic ability by priming the immune system. However, other studies have demonstrated that exosomes can stimulate tumor growth and metastasis and promote cancer resistance to chemotherapeutic drugs. This paradox may be the result of heterogeneity and contaminants in exosomes preparations and highlights the need for standardized protocols. Ultracentrifugation is the most common isolation technique used for exosomes, but preparations obtained by this method frequently contain other extracellular vesicles such as microvesicles and apoptotic bodies that overlap in densities and sizes with exosomes. This doctoral work seeks to address this challenge by demonstrating the ability of field-flow fractionation (FFF) techniques to be used as an orthogonal approach, along with the primary ultracentrifugation method, to remove non-exosomal byproducts and to produce and characterize homogenous exosomes subpopulations. This thesis involves three main projects with a general aim to design FFF approaches that have the capability to meet the challenges in exosomes separation and analysis. The first project aims to develop optimized sedimentation FFF (SdFFF) and asymmetric flow FFF (AF4) to produce effective mass- and size-uniform exosomes subpopulations from heterogeneous exosomes preparations that were derived from various tumor cell lines (e.g., breast, brain, and ovarian cancers) using ultracentrifugation. Fractions collected at the outlet of the FFF channel are characterized with respect to density and size. The SdFFF, which is an effective mass meff based separation technique, was first used to produce meff uniform fractions. The meff is equal to d3[Delta][rho] where d is diameter and [Delta][rho] is the difference in density between the particle and the carrier liquid. The density distribution of breast cancer ascites exosomes was found to be ~1.01 to 1.18 g/cm3, suggesting that exosomes preparations from ultracentrifugation were contained low density contaminants (in the lower SdFFF fractions) that can be eliminated and separated from exosomes fractions (in the higher SdFFF fractions). The exosomes were also introduced into anAF4 system where they underwent a size separation. AF4 demonstrated the ability to narrow the original size distribution of the exosomes from ~ 30-500 nm to ~ 30-140 nm. Furthermore, each exosomal fraction has a narrow size distribution of approximately ± 10 nm. While SdFFF and AF4 were independently implemented in the first project, the second project focused on a novel two-dimensional approach wherein both techniques were used to obtain uniform mass and size subpopulations. Exosomes from different cancer cell lines are first separated using SdFFF and fractions with uniform meff were collected. Each SdFFF fraction with uniform meff is then injected into an AF4 channel and a size separation was implemented. Hence, each fraction that is collected at theAF4 outlet has a uniform d and density. An online focusing stage commonly used in AF4 solves the sample dilution problem encountered when conducting two-dimensional separations. Assessment of separated exosome fractions confirmed that the fractions are indeed uniform in density and size. The average densities of the exosomes fractions after two-dimensional FFF were determined to be between ~ 1.05 to 1.10 g/cm3. These results suggest that non-exosomes components of low densities can be fractionated from exosomes of higher densities. The sizes of exosomes in fractions collected after two-dimensional FFF are determined using dynamic light scattering and ranged from ~ 40 to 80 nm for brain cancer exosomes, ~ 40 to 140 nm for ovarian cancer exosomes and ~ 50 to 145 nm breast ascites exosomes. These ranges, which span the exosomes sizes, represent narrower size distributions than that of the unfractionated exosomes sample. The third project was designed to improve the sample throughput by scaling up the analytical AF4 channel to a semi-preparative AF4 (sP-AF4) channel. Results from this study showed that the brain cancer-derived exosomes sample volume can be increased up to 250 fold to produce ~ 75 [mu]g of exosomes per fraction using a sP-AF4 channel without affecting the resolution of the separation. This allowed the production of large amounts of size uniform exosomes subpopulations with one sample injection, rather than performing tens of experiments on an analytical channel to collect and concentrate combined fractions. Characterizing the exosomes fractions collected from the sP-AF4 channel shows that the fractions have distinct sizes that range from ~ 40 to 150 nm. Exosomes fractions also showed an abundance of proteins with molecular weights between ~ 60 and 70 kDa. The presence of common exosomal protein biomarkers including CD9, Hsp/c70, and Hsp90 in each sP-AF4 fraction was confirmed by Western blot analysis. The results showed that most of the fractions contained different relative concentrations of exosomal protein biomarkers, indicating the presence of exosomes. The data presented in this thesis advances the separation and characterization of exosomes and establishes a route forward towards obtaining purer exosomes preparations and subpopulations. This capability stands to contribute significantly to furthering the understanding of the role of exosomes in intracellular communication, and to assess their potential role and use in cancer diagnosis and progression.