Nanosized extracellular vesicles (EVs, ∼30-2000 nm) have emerged as important mediators of intercellular communication, offering opportunities for both diagnostics and therapeutics. In particular, small EVs generated from the endolysosomal pathway (∼30-150 nm), referred to as exosomes, have attracted interest as a suitable biomarker for cancer diagnostics and treatment monitoring based on minimally invasive liquid biopsies. This is because exosomes carry valuable biological information (proteins, lipids, genetic material, etc.) reflecting their cells of origin. Using EVs as biomarkers or drug delivery agents in clinical applications requires a full understanding of their cellular origin, functions, and biological relevance. However, due to their small size and very high heterogeneity in molecular and physical features, the analysis of these vesicles is challenged by the limited detection ranges and/or accuracy of the currently available techniques. To overcome some of these challenges, this thesis focuses on developing different techniques for characterization, detection and protein profiling of EVs at both bulk and single particle levels. Specifically, the three methods investigated are scanning electron microscopy, electrokinetic sensing, and combined fluorescence - atomic force microscopy. 

First, a protocol for scanning electron microscopy imaging of EVs was optimized to improve the throughput and image quality of the method while preserving the shape of the vesicles. Application of the developed protocol for analysis of EVs from human serum showed the possibility to use scanning electron microscopy for morphological analysis and high-resolution size-based profiling of EVs over their entire size range. Comparison with nanoparticle tracking analysis, a commonly used technique for EV size estimation, showed a superior sensitivity of scanning electron microscopy for particles smaller than 70-80 nm. Moreover, the study showed process steps that can generate artifacts resembling sEVs and ways to minimize them. 

Secondly, a novel label-free electrokinetic sensor based on streaming current was developed, optimized and multiplexed for EV protein analysis at a bulk level. Using multiple microcapillary sensors functionalized with antibodies, the method showed the capacity for multiplexed detection of different surface markers on small EVs from non-small-cell lung cancer cells. The device performance in the multichannel configuration remained similar to the single-channel one in terms of noise, detection sensitivity, and reproducibility. The application of the technique for analysis of EVs isolated from lung cancer patients with different genomic alterations and after different applied treatments demonstrated the prospect of using EVs from liquid biopsies as a source of biomarker for cancer monitoring. Moreover, the results held promise for the application of the developed method in clinical settings. 

Finally, to increase the understanding of EV subpopulations and heterogeneity, a platform combining fluorescence and atomic force microscopy was developed for multiparametric analysis of EVs at a single particle level. The use of a precise spot identification approach and an efficient vesicle capture protocol allowed to study and correlate for the first time the membrane protein composition, size and mechanical properties (Young modulus) on individual small EVs. The application of the technique to vesicles isolated from different cell lines identified both common and cell line-specific EV subpopulations bearing distinct distributions of the analyzed parameters. For example, a sEV population co-expressing all the three analyzed proteins in relatively high abundance, yet having average diameters of <100 nm and relatively low Young moduli was found in all cell lines. The obtained results highlighted the possibility of using the developed platform to help decipher unsolved questions regarding EV biology. 

Extracellular vesicles (EVs, ∼ 30 nm−5 μm ) are lipid bilayer-enclosed particles expressingvaluable biological information such as proteins, lipids, and nucleic acids that reflecttheir shedding cell. The discovery of their importance in cell-to-cell communicationsparked a boom in research. Their abundance, ability to freely surpass natural barriersin the body, and reflection to the original cell make them suitable players in fieldssuch as treatment monitoring and targeted drug delivery. By investigating how EVsubpopulations interact with cells, we may also gain further insights into theoreticalquestions such as how cells communicate and how cells respond to external stimuli.Virtually all cells in the body release EVs; each cell may contain multiple origin spots forbiogenesis, and EVs may have different intended purposes that are also reflected in theircomposition. Therefore, EVs are extremely heterogeneous in size, expression level ofbiomolecules, and nanomechanical properties such as elasticity. This heterogeneity andthe small size of the vesicle pose technical challenges for the characterization platformsexisting today. EVs may be studied in bulk with a single output for an entire particleensemble or individually, yielding a characterization of individual EVs in a sample.Bulk methods are often faster, offer higher throughput, and may be the only option foranalyzing some parts, such as RNA. However, for complete characterization, we need toretrieve information on single EVs. This thesis explores techniques to characterize EVson a single vesicle level with three different platforms: a fluorescence microscope, anatomic force microscope, and a combined fluorescence and atomic force microscope.First, a fluorescence microscope is used to study EVs released by cells in a cancercell line model study. The cells are either left untreated or treated with two drugs: onethat the cells should respond to and one that they should be immune to. Five relevantsurface proteins were stained, imaged, and analyzed. The study revealed the possibilityof monitoring drug responses through immunofluorescence. Next, the platform was usedto study lung cancer patients undergoing treatment with EVs retrieved through liquidbiopsy. Each patient generated two sets of EVs: one sample from before treatmentand one sample after treatment, but before the tumor stopped responding to the drug.While the study revealed changes in individual proteins when comparing the two sampleswithin each patient, it was difficult to distinguish a pattern regarding the length oftreatment before drug resistance. It was not until we studied the correlation of proteinsand combined all protein expressions in a sample into a joint probability distributionthat trends became clearer. Longer treatments, for example, were found to have astronger positive correlation among the proteins. This highlights the importance ofincluding sophisticated statistical methods to analyze clinical EV samples on a singleEV level. Next, a theoretical model taking into account the EV’s liquid properties was con-structed. The model agrees with force spectroscopy measurements performed with force microscopy. Three EV samples with different protein expression levels were comparedin terms of elasticity moduli. With the low throughput of EVs in the technique, astatistical framework to compare the distributions of stiffness values was developed. Theframework revealed a large variation in stiffness values extracted from a single vesicle,which is hypothetically attributed to thermal fluctuations and diffusion of membranemolecules.Finally, we combined the fluorescence microscope and atomic force microscope toinvestigate subpopulations and heterogeneity in single EVs with both protein expressionand precise mechanical measurements of size and Young’s modulus. The platformrevealed distinct subpopulations with unique properties in the analyzed parameters. These combined measurements are the first of their kind, and a combined platformcharacterizing EVs in multiple ways may offer great insights into EV biology.

Extracellulära vesiklar (EVs, ~30 nm - 5 µm) är lipiddubbelskiktsinneslutna partiklar som uttrycker värdefull biologisk information som proteiner, lipider och nukleinsyror som reflekterar deras föräldracell.Upptäckten av deras betydelse i cell-till-cellkommunikation utlöste en explosion inom forskning. Deras mängd, förmåga att fritt passera naturliga barriärer i kroppen och reflektion av dess ursprungliga cell gör dem till lämpliga aktörer inom områden som behandlingsövervakning och riktad läkemedelsleverans. Genom att undersöka hur EV-subpopulationer interagerar med celler kan vi också få ytterligare insikter i teoretiska frågor som hur celler kommunicerar och hur celler reagerar på yttre stimuli. Praktiskt taget alla celler i kroppen släpper ut EVs; varje cell kan innehålla flera ursprungsplatser för biogenes, och EVs kan ha olika avsedda syften som också återspeglas i deras sammansättning. Därför är EVs extremt heterogena i storlek, uttrycksnivå av biomolekyler och nanomekaniska egenskaper som elasticitet. Denna heterogenitet och den lilla storleken på vesikeln utgör tekniska utmaningar för de karaktäriseringsplattformar som finns idag. EVs kan studeras i bulk med ett enda resultat för en hel partikelensemble eller individuellt, vilket ger en karakterisering av individuella vesiklar i ett prov. Bulkmetoder är ofta snabbare, erbjuder högre genomströmning och kan ibland vara det enda alternativet för att analysera vissa delar, såsom RNA. Men för en fullständig karaktärisering måste vi hämta information om enstaka vesiklar. Denna avhandling utforskar tekniker för att karakterisera EVs på en enstaka vesikelnivå med tre olika plattformar: ett fluorescensmikroskop, ett atomkraftmikroskop och ett kombinerat fluorescens- och atomkraftmikroskop.

Först används ett fluorescensmikroskop för att studera EVs som frigörs av celler i en modellstudie av cancercellinjer. Cellerna lämnas antingen obehandlade eller behandlas med två läkemedel: ett som cellerna ska svara på och ett som de ska vara immuna mot. Fem relevanta ytproteiner taggades, fotograferades och analyserades.Studien avslöjade möjligheten att bevaka läkemedelssvar genom immunfluorescens. Därefter användes plattformen för att studera lungcancerpatienter som genomgick behandling med EVs hämtade genom flytande biopsi.Varje patient genererade två uppsättningar EVs: ett prov från före behandling och ett prov efter behandling, men innan tumören slutade svara på läkemedlet. Medan studien avslöjade förändringar i individuella proteiner när man jämförde de två proverna inom varje patient, var det svårt att särskilja ett mönster angående behandlingslängden före läkemedelsresistens. Det var inte förrän vi studerade korrelationen mellan proteiner och kombinerade alla proteinuttryck i ett prov till en gemensam sannolikhetsfördelning som trenderna blev tydligare. Längre behandlingar visade sig exempelvis ha en starkare positiv korrelation bland proteinerna. Detta understryker vikten av att inkludera sofistikerade statistiska metoder för att analysera kliniska EV-prover på en enda EV-nivå.

Därefter konstruerades en teoretisk modell som tar hänsyn till vesikelns flytande egenskaper. Modellen överensstämmer med kraftspektroskopimätningar utförda med kraftmikroskopi. Tre EV-prover med olika proteinnivåer jämfördes i termer av elasticitetsmoduler. Med den låga genomströmningen av EVs i plattformen utvecklades ett statistiskt ramverk för att jämföra fördelningarna av styvhetsvärden. Ramverket avslöjade en stor variation i styvhetsvärden extraherade från en enda vesikel, vilket hypotetiskt tillskrivs termiska fluktuationer och diffusion av membranmolekyler.

Slutligen kombinerade vi fluorescensmikroskopet och atomkraftmikroskopet för att undersöka subpopulationer och heterogenitet hos individuella EVs i både proteinuttryck och exakta mekaniska mätningar av storlek och Youngs modul. Plattformen avslöjade distinkta subpopulationer med unika egenskaper i de analyserade parametrarna. Dessa kombinerade mätningar är de första i sitt slag, och en kombinerad plattform som karakteriserar EVs på flera sätt kan ge fantastiska insikter om EV-biologi.