Shotgun proteomic analytical approach for characterizing “protein coronas” of different liposome nanovectors suitable for gene delivery

This PhD thesis is a part of the study concerning non-viral nanocarriers for gene therapy. In modern molecular medicine, gene therapy is a very promising research. One issue with gene therapy is the efficient delivery of the genetic material into the cell. Naked DNA cannot pass freely the cell membrane because it is made up of very large molecules with a hydrophilic nature provided by the negatively charged phosphate groups; moreover it can be easily degraded by nucleases. For these reasons the development of suitable vectors for an efficient transfection is fundamental in order to facilitate and optimize gene transfer to targeted cells without degradation. One of the most important requirements in gene therapy is the development of safe and efficient gene delivery systems. In the past two decades viral vectors, such as adenovirus and retrovirus, were the most used in several clinical trials. Today non-viral vectors have attracted a growing interest in the scientific community, thanks to their preparation reproducibility, lack of immunogenicity and almost no size limit to the piece of DNA carried into cells. The aim of this PhD thesis was to identify and compare plasma proteins bound to different kind of non-viral nanovetcors commonly employed for gene therapy studies. The knowledge about the interaction between plasma proteins and nanocarriers is fundamental to understand the in vivo biodistribution, thus the transfection efficiency (TE). Protein adsorption onto nanoparticle surface (protein corona) is strongly affected by vector surface characteristics. In general, the primary interaction is thought to be electrostatic, thus surface charge of carriers is supposed to play a central role in protein adsorption. Because protein corona composition can be critical in modifying the interactive surface that is recognized by cells, an insight into its formation onto lipid particles may serve as a fundamental predictive model for the in vivo efficiency of a lipid vector. A shotgun analytical proteomics approach was employed in all studies to compare human plasma protein binding capability to cationic liposomes (CLs), lipoplexes and protamine/DNA (P/DNA) complexes. This approach exploited a centrifugation-based protocol for the separation of the nanoparticle-protein complexes, followed by "in solution” proteolytic digestion of the whole protein mixtures and determination of the resulting peptides by nano-high performance liquid chromatography (nanoHPLC) coupled to a high-resolution linear trap quadrupole (LTQ) Orbitrap XL mass spectrometer. Paper I described and compared the adsorption of human plasma proteins bound to cationic liposomes, made of (3-[N-(N,N-dimethylaminoethane)-carbamoyl])-cholesterol (DC-Chol) and the zwitterionic lipid dioleoylphosphatidylethanolamine (DOPE), and their relative DNA cationic lipoplexes. A shotgun proteomics approach based on HPLC coupled to high resolution mass spectrometry (MS) was used for an efficient identification of proteins adsorbed onto liposome and lipoplex surfaces. The distinct pattern of proteins adsorbed helped to better understand the DNA compaction process. The experimental evidence lead us to hypothesize that polyanionic DNA is associated to the lipoplex surface and can interact with basic plasma proteins. Such a finding is in agreement with recent results showing that lipoplexes are multilamellar DNA/lipid domains partially decorated with DNA at their surface. Proteomics experiments showed that the lipoplex corona is rich of biologically relevant proteins such as fibronectin, histones and complement proteins. Our results provided novel insights to understand how lipoplexes activate the immune system and why they are rapidly cleared from the bloodstream. The differences in the protein adsorption data detected in the presented experiments could be the basis for the establishment of a correlation between protein adsorption pattern and the in vivo fate of intravenously administered nanoparticles, and will require some consideration in the future. Paper II and III investigated the liposomes made of the cationic lipid DC-Chol DOPE and the possible effect of membrane charge density on the formation of the protein corona. The membrane charge density is the average charge per unit area of the membrane and it is controlled by the ratio of neutral to cationic lipid in the liposome formulation. To vary it, the neutral/total lipid molar ratio was changed, “diluting” the cationic lipid in the membrane with the neutral helper lipid. In paper II the proteomic study was carried out in order to portrait the protein corona composition from a qualitative point of view. This initial study was performed validating Mascot search results with the Trans-Proteomic Pipeline (TPP) platform. In paper III the experiment was repeated to add a label-free quantitation based on spectral count (SC) by Scaffold. Fibrinogen displayed higher association with CLs having a higher membrane charge density, while apolipoproteins and C4b-binding protein with CLs having a lower membrane charge density. These results are discussed in terms of the different lipid compositions of CLs and may have a deep biological impact for in vivo applications. Surface charge of nanoparticles is emerging as a relevant factor determining the corona composition after interaction with plasma proteins. Remarkably, it was also shown that the charge of the protein corona formed around CLs is strongly related to their membrane charge density. In paper IV we investigated the compositional evolution of the protein corona of 1,2-dioleoyl-3-trimethylammonium propane (DOTAP) CLs and DOTAP/DNA lipoplexes over a wide range of plasma concentrations (2.5-80%). The composition of the hard corona of lipoplexes was quite stable, differently from that of CLs, which did evolve considerably, instead. We showed that the protein corona of CLs was made of both low-affinity and competitive-binding proteins, the relative abundance of which revealed to be dependent on the plasma concentration to which they were exposed. In paper V we examined the advantages of exploiting a P/DNA complex coated with a lipid envelope made of the cationic lipid DOTAP, for transfecting CHO (Chinese hamster ovary cells), HEK293 (human embryonic kidney cells), NIH 3T3 (mouse embryonal cells), and A17 (murine cancer cells) cells. Recent studies, in fact, showed superiority of LPD-mediated gene transfer over conventional liposomes for delivering a gene to the liver. To demonstrate this evidence, we investigated complex formation, DNA protection ability, surface properties, nanostructure, ability to release DNA upon interaction with cellular lipids, and intracellular trafficking. In papers VI and VII, after investigating the physical-chemical properties of LDP complexes, we employed a shotgun proteomics approach to discuss the protein coronas of DOTAP, DOTAP/DNA and LPD complexes. In paper VI we carried out a qualitative study, whereas in paper VII we also added a label-free quantification performed by two distinct methods: the spectral counting (SC), in which the number of spectra matched to peptides from a protein was used as a surrogate measure of protein abundance, and the area under the curve (AUC) or signal intensity measurement, in which protein abundance was derived from the extracted ion chromatograms. Our results could help in designing gene delivery systems, because some proteins could be more selectively bound than others, thus affecting the biodistribution of liposome nanovectors and provide a more efficient in vivo delivery in gene therapy applications.