Comparing profiles and processes for the selective isolation and application of extracellular vesicles in regenerative therapeutics: with an initial regenerative dermatology focus

Regenerative medicine has long focused on the application of mesenchymal stem cells (MSCs). However, risk of immunogenicity, evidence of limited engraftment in vivo and pragmatic considerations (e.g. storage) have hindered clinical translation. These limitations have resulted in a move towards extracellular vesicles (EVs) as a potential cell-free alternative. Dermatology is one field looking to apply EVs for regenerative applications targeting tissue form and function. However, there are several challenges to overcome if we are to progress both the clinical and commercial translation of EV therapeutics. This thesis sought to evaluate some of these key challenges.

One major barrier to the translation of EV therapies is an inability to reproducibly isolate high yields of pure EVs due to a lack of optimal isolation methods. The first three experimental chapters of this thesis therefore focused on defining and evaluating parameters relevant to EV isolation. Chapter 2 centres around the optimisation of downstream EV characterisation method western blotting utilised throughout the remainder of the thesis. Chapter 2 also includes the validation of multiple EV isolation methods (ultracentrifugation; UC, polyethylene glycol; PEG precipitation and an aqueous two-phase system; ATPS), which are applied in Chapter 4 to compare EV recovery for therapeutic applications.

Chapter 3 collates the findings of a survey that sought to cross-compare a variety of parameters (e.g. EV source, starting volume, operator experience, application and implementation) governing the selection of EV isolation method(s) across disciplines (e.g. academic, industry, clinical). Outcomes highlighted growth in the clinical application of EVs for therapeutics and diagnostics. In addition to identifying EV isolation method selection to be most impacted by downstream application and implementation (e.g. cost, scalability, time efficiency etc), as well as operator experience.

Chapter 4 continued the comparison of EV isolation methods, evaluating relative EV purity and yield following isolating by UC, PEG precipitation, ATPS, total exosome isolation reagent (TEIR) or size exclusion chromatography (SEC). EV enriched fractions could be obtained via all methods. However, EV recovery varied in terms of both purity and the expression of common protein markers associated with biogenesis (e.g. Alg-2-interacting protein X; Alix, Annexin A2, CD9, CD63 and CD81) between isolation methods. In addition to variations in EV yield and profile, disparity was observed between downstream methods for evaluating sample purity, with notable differences seen between particle to protein (PtP) ratios and tetraspanin (CD9, CD81 and CD63) positive particle recovery. Furthermore, ATPS was shown to compromise downstream analysis. Overall, EV isolation by UC or SEC provided the highest proportion of tetraspanin positive particles. These methods were also selected by EV researchers to have the best overall efficiency, in terms of pragmatic considerations surrounding implementation (e.g. scalability and cost). Based on these outcomes UC was selected for application in the final study in Chapter 5

In Chapter 5 UC was applied to isolate EVs for evaluation of the impact of MSC source on EV profile for prospective applications in regenerative dermatology. EVs were compared from adult (bone marrow; BM) and perinatal (umbilical cord; UB) MSCs, in addition to a commercially available EV product marketed for topical regenerative dermatology applications. Similar profiles were seen when applying a variety of analysis methods (e.g. imaging, marker expression, Raman spectroscopy; RS, proteomics). However, notable differences could be observed in relative tetraspanin expression between the three EV preparations (e.g. upregulation of CD63 in BM-MSC derived EVs). In addition, the commercial EV preparation displayed a comparatively low purity, which was identified both qualitatively (cryo-TEM) and quantitatively (PtP). Proteomic analysis of EVs predicted pathways relevant to regenerative dermatology, such as keratinocyte migration, in all three EV preparations, with pathways indicated to be comparatively upregulated in BM-MSC derived EVs. Whilst all three EV preparations were indicated to have potential for regenerative dermatology applications, variations observed in their profile could impact downstream biological function.

Outcomes of this thesis comment on the selection of cell source, isolation protocol, and pragmatic considerations for regenerative therapeutic applications. It was found that EV researchers selection of EV isolation methods is predominantly impacted by downstream application and analysis, as well as implementation. Upon comparing widely applied EV isolation methods, based on the recovery of common EV markers of biogenesis, as well as overall pragmatic considerations, UC was selected to be the most optimal method for isolating EVs for therapeutic applications in the context of this thesis. When comparing EVs isolated by varying isolation methods or from different MSC sources, EV profiles were shown to differ both in terms of biomolecular content, purity and yield. This variation may hinder our ability to fully define mechanism of action (MOA) and develop a reproducible EV therapeutic, thus limiting both clinical and commercial translation. Further evaluation of the impact of these parameters observed in this thesis on biological function will be a critical next step.