Prostate cancer is a highly heterogeneous disease and its manifestations can vary from indolent localized tumor to widespread metastases. This heterogeneity is also observed at the molecular level both inter- and intra-patient. Intra-patient heterogeneity in the clinical setting of men with castration resistant prostate cancer (CRPC) might be informative in terms of treatment decision. Here I present analytical work on two approaches relevant to the characterization of intra-patient heterogeneity and applied to unpublished CRPC patients sequencing data. The first is based on the genome wide interrogation of multiple metastatic and primary tissue biopsies from single patients. I present genomic analyses to decipher the content of multiple tumor biopsies from CRPC patients and provide comparisons to highlight similarities and differences and to identify alternative patterns of aberrations. The second approach, alternative to tissue biopsies that might under-represent the genomic landscape of the patient’s disease, relies on liquid biopsies, a minimally invasive test that is also amenable to serial sampling. Liquid biopsies contain circulating cell free DNA (cfDNA) released from widespread tumor cells, potentially uncovering the full tumor landscape. By using next generation sequencing on cfDNA obtained from plasma, I developed strategies aimed at systematically tracking the reiterative process of genetic diversification leading to disease evolution and to detect genomic aberrations. I specifically focused on an ad hoc computational procedure (ABEMUS) to detect somatic point mutations that could emerge under treatment pressure and as drug resistance mechanism. The work I present is relevant to the context of precision oncology that exploits detailed patient-specific molecular information to diagnose and follow cancer progression with the ultimate goal of promptly guiding treatment decisions to improve clinical outcome with transdisciplinary strategies. The analytical work I developed can be applied to the study of any tumor type.

Post-transcriptional regulation is emerging as a fundamental step in gene expression that, when altered, can contribute to carcinogenesis. To identify potentially altered translational networks in glioblastoma multiforme (GBM), the worst of brain tumors, we correlated genomic alterations and mRNA levels of RNA-binding proteins (RBPs) in 372 publicly available GBM samples and identified 56 RBPs whose copy number alterations correlate with their altered expression levels in at least 15% of the samples. Among the genes identified using our parameters, HuB (ELAVL2) was deleted in 48% of the samples and down-regulated in more than 90% of these. Given the evidence for HuB activity as a differentiation factor in neuronal cells, we hypothesized that it may act as an oncosuppressor. However, ELAVL2 maps to the same chromosomal band (9p21.3) as CDKN2A, the most frequently inactivated oncosuppressor in gliomas, giving the possibility that ELAVL2 loss is simply a consequence to the CDKN2A deletion and therefore a passenger, albeit very frequent, mutation. To test this possibility, we analyzed the structure of the deletion spanning the two loci by qPCR analysis of 233 GBM samples, testing for the presence of the intervening region, and we obtained evidence that two independent focal deletions occurred in about 20% of the GBM samples bearing homo- and/or heterozygous deletion at CDKN2A and ELAVL2. This result highlights that, in rare but recurrent cases, ELAVL2 deletion occurs independently from CDKN2A deletion, supporting the hypothesis that loss of HuB activity is a condition contributing to tumor progression. This hypothesis was tested using U87MG cells, a commonly used glioma cell line homozygously deleted for CDKN2A and heterozygously deleted for ELAVL2 and primary glioma initiating cells (GICs) homozygously deleted for both genes. Migration, invasion and capacity to form nuerospheres were determined in U87MG cells upon HuB silencing and overexpression, and in GICs upon HuB expression. HuB expression in both cell models resulted in a decreased in migration, invasivity and the capacity to form neurospheres, supporting the hypothesis that HuB act as a tumor suppressor. We finally showed that HuB is able to determine an increase of p21 protein in normal murine neuroblast cells, providing a possible mechanism for HuB-mediated suppression of glioma cell clonality.

Cancer is the second leading cause of death worldwide. Tumor cells contain several mutations that can generate neoepitopes, targets of an effective anti-cancer T cell response. Increasing evidence demonstrated that cancer vaccines targeting neoepitopes are effective and safe both in preclinical models and human patients. Bacterial Outer Membrane Vesicles (OMVs) are naturally produced by all Gram-negative bacteria. They contain several Microbe-Associated-Molecular Patterns (MAMPs), crucial for stimulating innate immunity and promoting adaptive immune responses. The ability to engineer OMVs with cancer epitopes together with their unique adjuvanticity and safety make them a particularly interesting vaccine platform. In this study, we have demonstrated that immunization of mice with OMVs activate both innate and adaptive immunity and induce a Th1 immune response, fundamental for an effective cancer vaccine. OMV immunization also caused upregulation of genes involved in MAMPs detection and signal transduction, a central component of the inflammasome and pro-inflammatory cytokines. OMV vaccination induced an upregulation of Th1 key transcription factor and cytokines, while inducing a downregulation of transcription factor and cytokines associated to Th2 response. Moreover, cytokines released by activated macrophages, DCs, T cells and natural killer (NK) cells were induced by OMV vaccination, together with a key chemokine and a protein for immune cell recruitment and adhesion, respectively. We have successfully engineered OMVs on the surface and in the lumen with OVA(257-264) CD8 T cell model epitope. These OVA-engineered OMVs induced a high percentage of OVA(257-264) specific CD8 T cells and protected mice from OVA-expressing tumors. We have shown that OMVs engineered with a tumor specific antigen (TSA) induced a protective response and promoted a significant recruitment of CD4 and CD8 T cells into tumors, while reducing both CD4 regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs). We have also shown in the same mouse model that vaccination with OMVs engineered with two TSAs have a synergistic protective activity in controlling tumor growth. Finally, we have demonstrated that therapeutic OMV vaccines targeting five different neoepitopes protect mice from tumor growth. Taken together, our results show that OMVs are a promising platform for effective personalized cancer vaccines.

Genetic engineering has been widely used to reprogram cells for a variety of purposes, suggesting a wide range of possible applications in industrial and academic research. Although the techniques available are very well established, the mechanisms of cellular life are not completely understood. Therefore, despite offering a versatile tool, engineered cells are prone to possible unexpected behaviors. Investigations of more controllable systems are partially focused on the creation of cellular mimics assembled from discrete components with defined properties. The controlled assembly of molecules allows the creation of entities able to form compartments in water solutions and to carry out enzymatic reactions or gene expression. These artificial cells are able to establish communication pathways with natural cells and may be further developed to fight pathogens or cancer cells, for example. Despite these promising results, technological applications based on cellular mimics necessitate further technical improvements. A considerable defect of artificial cells is the lack of some mechanisms for selfsustainment that are instead present in engineered living cells. Besides few strategies aimed at energy restoration, artificial cells are not yet able to efficiently use the available resources in their environment. Considering these technical limitations, this thesis proposes to improve communication pathways between artificial and natural cells by exploiting multiple kinds of cellular mimics. Artificial cells can vary in composition and if engineered to coordinate activity, could be capable of overcoming individual weaknesses. To investigate the possibility of creating communities of artificial cells that collaborate with each other, the work described here was focused on establishing molecular communication pathways between two kinds of artificial cells. The designed communication was based on the exchange of chemical messages between two cellular mimics resulting either in genetic regulation or enzymatic reactions. On one side, lipid vesicles carrying gene expression through in vitro transcription and translation reactions and on the other side a novel structure composed of modified proteins, named proteinosome, to carry out enzymatic reactions. Each part of the communication pathway was separately investigated. Some efforts were put into the characterization of genetic switches so as to be able to better tune gene expression. All the other components were then singularly tested before combining together. One way that artificial cells, either alone or in a community, can function as a useful technology is if the artificial cells are able to sense and respond to environmental changes. The sensing functionality can be conferred by natural or synthetic transcriptional regulators. It is possible to modify biological macromolecules to interact with chemical messages released by natural cells. The second part of the thesis summarizes two distinct works aimed at developing two kinds of biosensors with potential applications within artificial cells. Several technical problems arose while testing the communication pathway, and it was necessary to change the initial strategy to include engineered cells. Nonetheless, the work presented here offers a method for the establishment of molecular communication pathways within communities of artificial cells that could serve as the basis for future implementation in more efficient communication systems.

Thesis summary: Traditional approaches investigate the linear effect of high-fat over-consumption on transcriptomic data. However, many biological processes occur in a nonlinear fashion. The objective of this PhD thesis is to study the response types of dietary fat mediated differentially expressed genes (i.e., linear response; nonlinear response: logarithmic, exponential, quadratic and cubic) along the longitudinal axis of the small intestine. The main goal is to assess the presence of: (1) gene-specific response type, which refers to genes that preserve the same response pattern in the whole small intestine; (2) section-specific response type, which explains how particular genes behave in different ways according to the intestinal section where they are expressed. This thesis presents two genome-wide transcriptome analyses that quantified the effects of two dietary lipid interventions in different sections of the small intestine of C57BL/6J mice. The first study includes the analysis of transcriptomic data collected from three sections of the small intestine (i.e., proximal, middle and distal). Data were collected after four weeks of dietary intervention during which mice were fed with various levels of fat (i.e., fat providing 10%, 20%, 30% or 45% kcal out of the total energy intake). Different intestinal sections were expected to exhibit regionalized functionality and display unique digestive and absorptive capacity. The goal was investigating linear and nonlinear (i.e., logarithmic, exponential, quadratic or cubic) transcriptomic responses as a continuous function of dietary fat intake along the longitudinal axis of small intestine. Middle section was the most responsive to dietary fat, and the majority of the genes showed linear response to fat intake. The highest relative importance of logarithmic (saturated) and exponential (unsaturated) response was in the proximal and in the distal section, respectively. Such pattern is coherent with the progressive absorption that occurs along the longitudinal axis of small intestine, with the hypothesis that when the absorptive capacity of the intestinal epithelia is overloaded the remaining fat will overflow to more distal sections. Processes related to inflammation responded in a gene-specific, linear way in the whole small intestine (mostly with up-regulated genes). Cholesterol transport and efflux exhibited section-specific response (i.e., genes were down-regulated with linear and exponential response in the proximal and middle intestine, respectively). The second analysis was performed on transcriptomic data collected from ten sections of the small intestine. The intervention lasted two weeks and mice were fed with three types of diets (i.e., high-fat, low-fat or chow). Three types of transcriptomic responses (i.e., linear, logarithmic or quadratic) to dietary fat were tested. Dietary fat triggered the over-expression of processes related to metabolism and transport of lipids while the processes related to carbohydrate metabolism were down-regulated. Transport and metabolic processes of lipids were significantly associated to logarithmic and linear response, respectively. Middle and distal sections were the most sensitive to fat whereas the proximal section was not responsive to diet (this is due to the prevalence of digestive functions in the proximal small intestine). Considering the complex physiology of small intestine and the influence of environmental effects such as diet and microbiota, nonlinear modelling of transcriptomic response to fat intake can provide a promising new avenue for research in molecular nutrition. The approach of studying nonlinear gene expression as a continuous function of fat intake allows the dynamic characterization of small intestine functioning. It can be used to illustrate the metabolic capacity of the small intestine or the presence of tipping points beyond which the relationship between fat intake and gene expression either weakens (logarithmic response) or strengthens (exponential response). Illustrating characteristic patterns of the gene- and section-specific response to dietary fat along the longitudinal axis of small intestine can contribute to better understand the relationships linking lipid intake to the development of obesity.

As a class of short non-coding RNAs, microRNAs can regulate gene expression by a post-transcriptional pathway through repressing or degrading mRNA. In the evolutionary history, many plant microRNAs are highly conserved from green alge to land plants. In recent years, dramatic studies demonstrate that microRNAs play a crucial in plant growth and development, in response to environmental stresses. Some microRNAs can response to plant hormones, while some others are tissue or cell specific. The understanding of how these microRNAs are regulated at the transcriptional level is just initiated. With the aim to understand the regulatory mechanism of plant microRNA in evolutionary tem, and identify the most relevant cisregulatory elements in some microRNAs for improving the agriculture in the future, this study was carried out. microRNA390 is one of the many conserved microRNAs, it can indirectly regulate the ARFs expression level by targeting TAS3, and consequently regulate lateral organ and later root development in plants. In order to understand the regulatory mechanism of microRNAs in an evolutionary term, microRNA390a and microRNA390b in Arabidopsis were chosen and studied. In 16 phylogenetically related species within Brassicaceae, we analyzed the microRNA promoter sequences and identified overall conserved cREs in microRNA390 promoter regions, and accompanied with functional characterization, we obtained a good view of microRNA390 regulatory network. Based on 454 sequencing technique, took the microRNA sequences of sequenced Arabidopsis as reference, by assembling and aligning the microRNA promoter sequences, calculated the PWM and predicted the putative motifs with both MEME program and PlantCARE database, subsequently compared the motif similarities by TOMTOM program, we eventually obtained the putative ones met the required E-value. In the meantime, we reconstructed the phylogenetic trees of both paralogs by MEGA7 program. We identified 6 and 5 overall conserved cREs. Subsequently, we experimentally validated the putative cREs by Arabidopsis transformation and site-specific mutagenesis. The results we have obtained were as follows: (1) There were totally 29 microRNA loci in 9 families identified to be highly conserved, and totally 104 putative motifs were predicted in their promoter regions. (2) The reconstructed phylogenetic trees based on miRNA390a and miRNA390b promoter sequences respectively were compared with the the phylogenetic relationships (species trees) in known Brassicaceae phylogeny. The data derived from both promoter sequences were inconsistent with Brassicaceae phylogeny. This implied that there might be multiple copies of specific cREs in some specific species, hence the promoter sequences evolution of microRNA is not reflective of species phylogeny. (3) Took Arabidopsis thaliana as model plant, we successfully constructed GUS-fused promoters of miRNA390a and miRNA390b. The GUS histochemical assay indicated that the two paralogs expressed in different tissues in transgenic Arabidopsis. miRNA390a expressed in lateral root primordia, true leaves, cotyledons, as well as in the floral organs, yet it was absent from lateral root tip and shoot apical meristem; whereas miRNA390b specifically expressed on lateral root tips, and a more restricted expression pattern was detected on aerial part of true leaves and floral organ. These differences indicated possible sub-functionalization with respect to their ancestral miR390 during the evolutionary process. (4) Based on the six putative cREs identified in miRNA390a and the reliable WT constructs, we also constructed six GUS-fused promoters that undergone site-specific mutagenesis. The GUS assay demonstrated that the activity of putative cis-elements varied with distance to TSS. Mutations of proximal sites (m2 and m3) enhanced expression thereby M2 and M3 were likely to be silencers; while mutations of distal elements (m5 and m6) tended to decrease the promoter expression, hence M5 and M6 probably work as enhancers. These evidences suggest there was a specific modular cooperativity of miR390a cREs in regulating gene expression and mediating plant development. Furthermore, we treated the 7d-old transgenic seedlings with iron-deficiency, both the GUS assay and qRT-PCR data conferred the iron responsiveness of putative iron-deficiency related E-box M3 and the iron-deficiency responsive cis-element 1 M6.

Brassicaceae is a large family of flowering plants, characterized by cruciform corolla, tetradynamous stamen and capsular fruit. In light of the important economic and scientific values of Brassicaceae, many phylogenetic and systematic studies were carried out. One recent and important phylogenetic analysis revealed three major lineages (I, II and III), however, classification at different taxonomic levels (tribe, genus, and species) remained problematic and evolutionary relationships among and within these lineages were still largely unclear. This is partly due to the fact that the past studies lacked information, as they mainly utilized the morphological data, nuclear DNA, partial chloroplast (cp) genes and so on. Nowadays, next generation sequencing (NGS) technology provides the possibility to make use of big data in phylogeny and evolutionary studies. Thus, we sequenced the chloroplast genomes of 80 representative species, using additional 15 reference chloroplast genomes from the NCBI database, and carried out both the phylogenetic reconstruction and the study of protein coding genes evolution in this novel dataset with different methods. Several novel results were obtained. 1 Successful application of NGS technology in chloroplast genome sequencing. During the final assembly, I could reconstruct full chloroplast genomes and the structure maps for 14 out of 80 sampled species, while the remaining were assembled nearly completely with only few gaps remaining. 2 Characterization of chloroplast genome structure. Gene number and order, single sequence repeat (SSR) as well as variety and distribution of large repeat sequence were characterized. 3 The difference of codon usage frequency was calculated between Cardamine resedifolia and Cardamine impatiens. Twelve genes with signatures of positive selection were identified at a family-wide level. 4 Three major lineages (I – III) were confirmed with high support values. Besides, the positions of various tribes were reclassified. Relationships among and within these lineages were highly resolved and supported in the final tree. Most of the tribes in the analyses were inferred to be monophyletic, only Thlaspideae was paraphyletic. Anastaticeae was for the first time classified into position of expanded lineage II, and position of tribe Lepidieae was delimited with relatively low support values in the final phylogenetic tree. This study was a new and successful application of NGS in large-scale Brassicaceae phylogeny and evolution, which offered the chance to look in details of the structural and functional features of the chloroplast genome. These results provided a paradigm on how to proceed towards the full elucidation of the evolutionary relationships among various biological species in the tree of life.

Retinal dystrophies are a large and extremely variegate group of diseases causing visual impairments and eventually blindness. As of today, no cures are available for such conditions, and the most advanced clinical approaches are based on classical gene transfer techniques. The recent development of genetic tools designed to manipulate splicing offers a unique opportunity to target several of these diseases, potentially reaching a highly specific and controlled effect in contrast to the more investigated gene supplementation therapies. These molecular tools are mainly part of two classes: antisense oligonucleotides (AONs) and chimeric or adapted small nuclear RNAs (snRNAs). They function by altering how the splicing machinery recognize a target sequence, usually an exon, allowing to increase its presence (exon inclusion) or decrease it (exon skipping) in mature mRNA. The first aim of this work has been the identification of genes causing retinal dystrophies having mutations potentially targetable by splicing-correction approaches. Following screening of mutation databases we identified three genes, each of them having mutations with characteristic interesting for a possible application of such approaches: CACNA2D4, RPGR, and USH2A. The CACNA2D4 gene encodes for an accessory subunit of high-voltage-activated (HVA) calcium channels and, when mutated, causes retinal cone dystrophy 4 (RCD4). It was selected because of the presence of a mouse model of retinal dystrophy (Wycisk et al., 2006a), allowing an in vivo test of the designed strategy. Since the approach consisted in inducing skipping of the exon hosting the mutation, we analysed the splicing pattern of the target exon in vivo and in vitro; the functionality of the rescued protein resulting from therapeutic exon skipping. The analysis revealed on one side that the rescued protein was not functional, therefore showing how the approach was not feasible for CACNA2D4. Interestingly, we found evidence of two newly identified splicing isoforms of the gene, one of which mimics the effect of the mutation. A mutation in RPGR intron 9 has been shown to cause retinitis pigmentosa by increasing the inclusion in mature mRNA of an alternatively spliced exon (E9a) (Neidhardt et al., 2007). Using splicing reporter minigene assays, we were able to design efficient chimeric U1snRNAs able to mediate E9a skipping thus correcting the genetic defect. In the USH2A gene we instead identified an interesting region hosting 5% of known mutation which could be approachable with exon skipping. The implementation of assays able to assess functionality of USH2A and RPGR will be fundamental for the development of new splicing-correction approaches for these genes.

Nup358 is the largest nucleoporin of around 358 kDa molecular weight that localizes on the cytoplasmic face of the nuclear pore complex (NPC). It takes part in the overall activity of the NPC mediating the transport of nucleic acids and proteins between the nucleus and the cytoplasm. However, due to its multi-domain configuration, Nup358 has a more pleiotropic function in several cellular mechanisms such as mediating the stability of microtubules and axon specification. Since little is known about the non-conventional role of Nup358 in neuronal polarity, my PhD thesis was focused on characterizing the subcellular distribution and expression pattern of Nup358 protein in cultured neurons. My results show Nup358 present at the nuclear rim of neuronal cells, associated with NPCs, also in the cytoplasm having a spotted pattern along the neuronal processes. Interestingly, Nup358 was remarkably clustered at the axon initial segment (AIS) of mature neurons and dependent on a prior recruitment of the master AIS scaffold, Ankyrin-G (AnkG), to this specific region. Of the distinct domains present in Nup358 protein, the N-terminal region was found to be crucial for its localization at the AIS. Further, changes in Nup358 protein expression were monitored during neuronal development. Indeed, I detected the presence of a shorter isoform of Nup358 that was increasing as neurons develop whereas the full-length protein had an opposite decreasing trend. To gain knowledge about the functional role of Nup358 in neurons, I investigated its protein expression/distribution in response to an increasing or decreasing neuronal activity with specific drug treatments. Surprisingly, Nup358 protein expression was reduced following the stimulation or the depolarization of neurons, mediated by calcium influx and NMDA receptors. Overall, my results show that Nup358 has a unique subcellular distribution in neurons, being enriched at the AIS at advanced stages of development, therefore, suggest an involvement of Nup358 protein in maintaining and modulating neuronal polarity and activity.

Post-transcriptional regulation of gene expression (PTR) is the process responsible for modulating mRNA levels and the related amount of protein. Initially thought to have a limited impact on cell phenotype, it has become increasingly recognized as a strong determinant of the quantitative changes in proteomes, and therefore a driving force for cell phenotypes. Untranslated regions of mRNAs (UTRs) are the core mediator of this process, containing sequence and structural elements bound by various kind of regulators, which influence nuclear export, localization, stability of mRNAs and their translation rates, as well as capping, alternative splicing and polyadenylation of the transcribed pre-mRNA. One of the most important classes of PTR factors are the RNA-binding proteins (RBPs), whose human genome complement is at least 800 genes, characterized by the presence of different functional domains. RBPs bind to the 5’UTR of a transcript often to modulate translation initiation, and to the 3’UTR usually to influence its stability or translatability. Another major group of actors in PTR are noncoding RNAs (ncRNAs). Among them are various classes of long ncRNAs (lncRNAs), the intensively studied microRNAs (miRNAs), siRNAs (small-interfering RNAs) and several other RNA types. miRNAs bind to 3’UTRs by means of short regions of perfect sequence complementation or with some mismatches. Both RBPs and ncRNAs bind mRNAs to the so-called cis-elements, found primarily in 5’ and 3’ UTRs. These elements can be represented as recurring RNA sequences or secondary structures to which the trans factors bind to exert a control over the mRNA. In order to integrate the available experimental data, we have developed AURA, a database offering a comprehensive view of the phenomena through regulatory data including RBP and miRNA binding sites, cis-element annotations, secondary structures, phylogenetic conservation, SNPs, RNA-editing data, gene expression profiles and more. A dynamic graphical interface allows the user to browse through the UTRs in an easy and seamless way. To further enrich this body of data, we also implemented a pipeline for the identification of hyperconserved elements in human UTRs, which we applied to both 5’ and 3’UTRs. We were thus able to recover known and novel PTR mechanisms involving RBPs, including an RBP network controlled by HuR. We are eventually applying the results of these works to infer altered, and thus potentially disease-related, PTR mechanisms in an highthroughput neuroblastoma dataset.