Search

2013/2014 Nei vertebrati i telomeri sono strutture specializzate localizzate all’estremità dei cromosomi costituito da sequenze ripetute TTAGGG. L’ accorciamento progressivo del telomero ad ogni divisione cellulare porta ad una disfunzione telomerica e all’attivazione della risposta di danno al DNA alle estremità cromosomiche. L’attivazione del segnale di danno al DNA provoca senescenza o apoptosi, fenomeni legati all’invecchiamento. Per mantenere la funzione del telomero il complesso della telomerasi, costituito dalla componente ad RNA (hTERC) e dalla subunità catalitica (hTERT), aggiunge ripetizioni telomeriche in cellule con alto potenziale replicativo. Due complessi principali sono coinvolti nella regolazione dei telomeri: il complesso shelterin e il complesso della telomerasi. Shelterin è costituito da sei principali proteine: TRF1, TRF2, POT1, TPP1, TIN2 e RAP1, e controlla vari aspetti della funzione telomerica come la lunghezza telomerica, la ricombinazione e la protezione da fattori di risposta al danno al DNA. l’evasione dalla senescenza replicativa raggiunta con la riattivazione della telomerasi, è un passaggio chiave nel processo di tumorigenesi, come osservato nel 90% dei tumori umani. Inoltre, aumentano anche le dimostrazioni che indicano il ruolo centrale di shelterin nella formazione e nella progressione del cancro. L’espressione del complesso shelterin o della telomerasi è strettamente regolata a livello trascrizionale e post-traduzionale, anche se il ruolo dei miRNAs nella regolazione dei telomeri non è ancora stato studiato. Lo scopo del mio progetto di tesi era quello di identificare miRNAs che controllano l’espressione dei componenti di shelterin o del complesso della telomerasi e valutare la rilevanza clinica di questi miRNAs nel contesto del tumore al seno. Per raggiungere questo obiettivo abbiamo eseguito un high-throughput luciferase reporter screening in cellule HeLa identificando un panello di miRNAs che hanno come bersaglio il 3’UTR di componenti del complesso shelterin (TRF1, TRF2, POT1) o del complesso della telomerasi (TERT, DKC1). Con questo screening abbiamo identificato che l’onco-miRNA miR-155 è un efficiente regolatore di TRF1. Il miR-155 è up-regolato in tutti i tipi di tumore al seno ed alti livelli del miR-155 correlano con una bassa espressione dei livelli proteici di TRF1. Inoltre, abbiamo validato con ulteriori saggi luciferasi il targeting di TRF1 da parte del miR-155 ed abbiamo anche scoperto che il miR-155 controlla l’espressione di TRF1 a livello traduzionale. Soprattutto, la bassa espressione dei livelli di mRNA di TRF1 e la bassa espressione dei geni bersaglio del miR-155 correlano con una ridotta sopravvivenza libera da metastasi distanti (DMFS) e con una ridotta sopravvivenza libera da recidive (RFS) in pazienti con cancro al seno luminale estrogeno-positivo (ER+). Questo indica che il targeting di TRF1 da parte del miR-155 è parte di una signature del miR-155 che determina una scarsa sopravvivenza nel cancro al seno di tipo luminale ER+. Soprattutto abbiamo scoperto che il targeting di TRF1 da parte del miR-155 porta ad un aumento della fragilità telomerica, a fusioni telomeriche dei cromatidi fratelli e all’allungamento dei telomeri. Il nostro lavoro dimostra per la prima volta che la regolazione post-trascrizionale di TRF1, mediata dal miR-155, è un efficiente meccanismo per controllare la fragilità telomerica e l’instabilità genomica nel cancro al seno. Invece, per quanto riguarda i miRNAs che regolano la telomerasi nel cancro al seno, abbiamo studiato il miR-296-5p e il miR-512-5p, i quali bersagliano efficientemente hTERT come mostrato dal high-throughput luciferase reporter screening . Entrambi i miRNAs bersagliano specifiche regioni nel 3’UTR di hTERT portando alla degradazione dell’mRNA di hTERT e alla riduzione dell’attività telomerasica. Dati clinici rivelano che il miR-296-5p e il miR-512-5p sono down-regolati nei tessuti tumorali del seno. Inoltre, abbiamo scoperto che l’aumento dell’espressione di hTERT e l’aumentata espressione dei geni bersaglio del miR-296-5p e del miR-512-5p correlano con una scarsa sopravvivenza in pazienti con cancro al seno di tipo basale. Questo evidenzia l’importanza clinica del miR-296-5p e del miR-512-5p nel cancro al seno (basale). A livello molecolare abbiamo mostrato che il miR-296-5p e il miR-512-5p rallentano la proliferazione cellulare e provocano l’ accorciamento della lunghezza dei telomeri in cellule di tumore al seno basale. Soprattutto, abbiamo scoperto che la ri-espressione epigenetica dei geni del miR-296-5p e del miR-512-5p in cellule di tumore al seno basale, determina l’aumento dell’espressione del miR-296-5p e del miR-512-5p e la conseguente riduzione dell’espressione di hTERT. I nostri dati suggeriscono che l’utilizzo di composti che inducono la ri-espressione del miR-296-5p e del miR-512-5p potrebbe essere una strategia promettente per compromettere i meccanismi di mantenimento dei telomeri e la funzione di hTERT nel cancro al seno. I risultati di questo lavoro dimostrano che i miRNAs rappresentano dei nuovi regolatori della funzione telomerica e hanno un impatto sull’omeostasi dei telomeri nel cancro umano. Questo lavoro identifica i miRNAs come nuovi bersagli per modulare la funzione dei telomeri nelle malattie ad essi correlate come il cancro e l’invecchiamento. Vertebrate telomeres are specialized structures localized at the end of chromosomes that are composed of TTAGGG tandem repeats. Progressive telomere shortening with every cell division, finally leads to telomere dysfunction and the induction of a DNA damage response at chromosome ends. DNA damage signaling provokes senescence or apoptosis, phenomena linked to organismal aging. In order to maintain telomere function the telomerase reverse transcriptase complex, composed by the RNA component (hTERC) and the catalytic subunit (hTERT) replenishes telomere repeats in cells with high replicative potential. Two main complexes are involved in the regulation of telomeres: shelterin and the telomerase complex. Shelterin is composed of six main proteins: TRF1, TRF2, POT1, TPP1, TIN2, and RAP1 and controls various aspects of telomere function such as telomere length, recombination and protection from the DNA damage response factors. The escape from replicative senescence is a key step in tumorigenesis and it is achieved by the reactivation of telomerase activity, as observed in 90% of human cancers. However, increasing body of evidence also indicates a central role of shelterin in cancer formation or progression. The expression of telomerase and shelterin complex is tightly regulated at transcriptional and post-translational level, however the role of miRNAs in telomere regulation is not understood. The aim of my PhD thesis project was to identify miRNAs that control the expression of components of the shelterin or telomerase complex and to evaluate the clinical relevance of these miRNAs in the context of human breast cancer. To achieve this goal we performed a high-throughput luciferase reporter screening in HeLa cells and identified a panel of miRNAs that target the 3’UTR of shelterin (TRF1, TRF2, POT1) or telomerase complex components (TERT, DKC1). In this screening we identified the onco-miRNA miR-155 as efficient regulator of TRF1 expression. miR-155 is efficiently upregulated in across all types of human breast cancer and high miR-155 levels correlate with low TRF1 expression levels. We validated targeting of TRF1 by miR-155 in additional luciferase reporter assays and found that miR-155 controls TRF1 expression on the translational level. Importantly, low TRF1 mRNA expression but also low expression of miR-155 target genes is linked to reduced distant metastasis-free survival and relapse-free survival of estrogen receptor positive luminal breast cancer patients. This indicates that targeting of TRF1 by miR-155 is part of a miR-155 signature that mediates poor survival in ER+ luminal breast cancer. Importantly, we found that targeting of TRF1 expression by miR-155 leads to increased telomere fragility, telomere sister chromatid fusions and telomere elongation. Our work demonstrates, for the first time, that post-transcriptional regulation of TRF1 by miR-155 is an efficient mechanism to control telomere fragility and genomic stability in the context of human breast cancer. Concentrating on miRNA dependent mechanisms of telomerase regulation in human breast cancer we studied miR-296-5p and miR-512-5p that were found to efficiently target hTERT in the high throughput luciferase reporter assay. Both miRNAs target specific motifs in the hTERT 3’UTR, leading to degradation of the hTERT mRNA and reduction of telomerase activity. Clinical data reveal that miR-296-5p and miR-512-5p are downregulated in breast cancer tissues. Importantly, we found that increased hTERT expression but also high expression of miR-296-5p or miR-512-5p target genes is linked to poor survival in basal breast cancer patients. This highlights the clinical relevance of miR-512-5p and miR-296-5p in basal breast cancer. On the molecular levels we showed that miR-296-5p and miR-512-5p impair cell proliferation and provoke progressive telomere shortening in basal breast cancer cell lines. We further found that epigentic de-repression of the miR-296 and miR-512 genes in a basal breast cancer cell line increase the expression of miR-296-5p and miR-512-5p levels to reduce hTERT expression. Our data suggest that the use of drugs that release miR-296-5p or miR-512-5p expression might be a promising strategy to impair telomere maintenance mechanisms and hTERT function in human breast cancer. Altogether, my work demonstrated that miRNAs represent new regulators of telomere function that impact on telomere homeostasis in human cancer. This identifies miRNAs as novel targets to modulate telomere function in telomere related maladies such as cancer and aging. XXVII Ciclo 1986

2013/2014 The Hippo signalling pathway is tumour suppressor cascade with a central role in the regulation of fundamental cellular biological processes, such as cell proliferation, apoptosis, organ size control and stem cell functions. The Hippo pathway transduces external signals that come to the cell into the nucleus, where it can control the expression of specific target genes, mainly involved in cell proliferation and differentiation. The Hippo pathway is an inhibitory pathway that control by phosphorylation and inhibition Yes-associated protein (YAP) coactivator, one of the two nuclear effectors of this signalling, involved in the regulation of proliferation and organ size. As consequence, deregulation of Hippo tumor suppressor pathway or hyperactivation of its downstream effectors is often associated with formation, development and tumour dissemination. Consistently, YAP is often over-expressed in a broad range of different tumours and it has aberrant activity in breast cancer as well as in several other human carcinomas. Up-regulation of YAP activity increases stem cell self-renewal in normal and cancer stem cells. In this work we describe the identification of a new hormonal-dependent layer for YAP regulation in breast cancer by the glucocorticoids and we analyze the mechanisms through which this regulation occurs. We found that Glucocorticoid Receptor (GR) binds directly the YAP promoter and induces the transcription of YAP mRNA after GC stimulation in cancer cells. Moreover, GC lead to efficient YAP de-phosphorylation and transcriptional activation, in a transcription-independent manner, by inducing actin cytoskeleton reorganization. Importantly, inhibition of the GR by means of RU486 (GR competitive antagonist) strongly blunted the expansion of the cancer stem cell pool in breast cancer cells by blunting the GR/YAP axis. XXVII Ciclo 1987

2013/2014 The MCM helicase plays a key role in DNA replication, as it unwinds the double helix, providing a single stranded template for the DNA polymerases. MCM genes are found both in Eukaryotes and Archaea. Eukaryotic genomes present at least 6 different MCM genes, which share significant sequence similarity. The six different polypeptides encoded by MCM genes assemble into a heteroexamer, characterized by a toroidal conformation. Each protein comprises 3 domains: the N-terminal domain binds DNA and is essential for hexamerization; the C-terminal domain folds as a winged-helix motif and may have a regulatory role; the central AAA+ domain is the catalytic core of the helicase that couples the ATP hydrolysis with DNA unwinding. Since the eukaryotic MCM complex is difficult to produce in large amounts, most of the available structural information derive from studies on the archaeal model. For eukaryotic MCM proteins no crystal structure has been determined yet, while low resolution reconstructions by electron microscopy are available only for S. cerevisiae and D. melanogaster complexes. The MCM proteins are present only in proliferating cells and are highly expressed in malignant cancer cells and pre-cancerous cells undergoing malignant transformation. Their expression has been compared with routinely used proliferation markers, showing that they are suitable candidates as biomarkers for cancer in clinical practice. Therefore, a detailed knowledge of their structure and function is a crucial pre-requisite for their potential role in cancer diagnosis and therapy. Based on bioinformatics analysis and structural information from archaeal MCM proteins, fragments corresponding to the AAA+ domains and N-terminal domains of the human MCM helicase were designed and cloned separately into different expression vectors. Various expression strategies were explored and constructs were tested for protein expression under a variety of conditions. AAA+ domains showed a severe problem of solubility: most of them precipitated during purification, aggregated or, in the best cases, precipitated after tag removal. The attempt to co express or co purify the subunits together didn’t improve the solubility of the proteins, also in presence of ATP analogues. The N-terminal domains were soluble and expressed at high levels. The co purification of the six fragments led to the assembly of two different oligomeric states compatible with a double and a single hexamer. Both oligomers fit the biological role of the protein, expected to be loaded onto the double stranded DNA as double hexamer and to unwind as single hexamer. DNA binding experiments were carried out by fluorescence polarization: both assemblies are able to bind different DNA substrates (single strand, double strand, fork) with micromolar dissociation constants, consistent with the DNA binding affinities reported for the archaeal MCM complex. A preliminary biophysical characterization performed with different techniques (thermal shift analysis, circular dichroism, multi angle light scattering, small angle X-ray scattering) revealed that the putative single hexamer is properly folded but quite unstable, as it showed the tendency to disassemble into smaller oligomers. In contrast, the putative double hexamer showed a higher stability, but still a degree of inhomogeneity which didn’t allow a structural study of its architecture. Nevertheless, the high yield of our purification protocol represents a promising starting point for further optimization, allowing crystallography and other structural studies. XXVII Ciclo 1987

2013/2014 It is now recognized that C1q can exert functions unrelated to C activation. Here we show that C1q, but not C4, is expressed in the stroma and vascular endothelium of several human malignant tumours. We found that C1q-deficient (C1qa-/-) mice bearing a syngeneic B16 melanoma have a slower tumour growth and prolonged survival compared to wild-type (WT) or C3- or C5-deficient mice. This effect was not attributable to differences in the tumour infiltrating immune cells. Tumours developing in WT mice displayed early deposition of C1q, higher vascular density and increased number of lung metastases than those growing in C1qa-/- mice. Bone marrow chimeras between C1qa-/- and WT mice allowed the identification of non-bone marrow-derived cells as the main local source of C1q that can promote cancer cell adhesion, migration and proliferation. Together these findings strongly support a role for locally synthesized C1q in promoting tumour growth. XXVII Ciclo 1987

2011/2012 Mesenchymal Stem/Stromal Cells (MSCs) are the precursors of various cell types that compose both normal and cancer tissue microenvironments. In order to support the widely diversified parenchymal cells and tissue organization, MSCs are characterized by a large degree of heterogeneity, although available analyses of molecular and transcriptional data do not provide clear evidence. Moreover a wealth of studies has demonstrated a significant role of the microenvironment and MSCs in tumor growth. In the course of the years MSCs were isolated by different groups from different tissues, both healthy and cancerous. The laboratory in which I conducted my PhD project was able to purify MSCs from different healthy tissues (N-MSCs) and, using an adapted protocol, from High-Grade Serous Ovarian Carcinomas (HG-SOC-MSCs). It was possible to shown that these cells do not possess gross chromosomal aberrations and are not tumorigenic in vivo. To better characterize these cells, an integrative bioinformatics analysis was conducted using the deep-CAGE-derived expression profiles obtained from HG-SOC-MSCs, N-MSCs and the FANTOM5 large comprehensive primary cells and tissues dataset. When compared to the other cells and tissues, HG-SOC-MSCs showed a correlation with mesothelial cells and cells hypothesized to have a mesothelial origin, such as smooth muscle cells and fibroblasts. It is known that mesothelial cells in culture can alternate between epithelioid and fibroblastoid morphologies and express high levels of either keratin or vimentin or both depending on their state of growth and the presence of EGF. When cultivated in the absence of EGF, which is known to induce a morphological switch in mesothelial cells, HG-SOC-MSCs switch from a fibroblast-like to an epithelial-like shape. N-MSCs in the same culture conditions, instead, do not change morphology. Moreover, in absence of EGF, HG-SOC-MSCs but not N-MSCs raise the levels of Keratin 7 both in protein and in mRNA. Starting from the list of up-regulated genes in HG-SOC-MSCs compared to N-MSCs a list of mesothelial-related genes was generated. This mesothelial-related gene list was compared to high-throughput gene expression datasets of MSCs derived from other tissues. Such analysis revealed that the mesothelial-related signature is specific to HG-SOC-MSCs. Moreover, Kaplan-Meier survival analysis conducted on a comprehensive SOC microarray dataset showed that patients with higher levels of the mesothelial-related gene signature displayed shorter progression-free survival time. Such correlation was rather specific for HG-SOC given that its performance was either statistically non-significant in the case of lung cancer or correlated with good prognosis in the case of breast cancer. Altogether, the study allowed us to assign a specific identity for MSCs derived from high-grade serous ovarian cancer. We demonstrated a cell-type specific transcriptional activity associated with HG-SOC-MSCs, which identifies them compared to N-MSCs from other districts and position them close to primary mesothelial and mesothelail-derived cells within the FANTOM5 dataset. XXV Ciclo 1985

2013/2014 Il tumore è una malattia caratterizzata da un’estrema complessità molecolare. Gli approcci di tipo “omic”, collezionando dati sull’intero genoma, sui trascritti e proteine in dataset pubblici, permettono di superare questa complessità e di trovare moduli funzionali che eseguono le funzioni coinvolte nei processi tumorali. Ad esempio, i profili di espressione genica da tessuti vengono usati per definire firme di geni e testarne la rilevanza clinica. Ho usato questo tipo di informazione per caratterizzare specifici geni di interesse in modelli di tumore al seno. Uno dei più recenti progetti di tipo “omic” è il FANTOM5. Questo progetto ha generato una risorsa unica: il primo atlante di espressione in mammifero basato su sequenziamento a singola molecola. Il sistema CAGE (Cap Analysis of Gene Expression) è stato usato per misurare i siti di inizio trascrizione (TSS) e l’utilizzo dei promotori in una collezione di campioni umani: in questo modo sono stati misurati i livelli di espressione di gran parte dei trascritti codificanti e non-codificanti nel genoma umano. Ho usato questo tipo di informazione per caratterizzare una linea staminale mesenchimale/stromale (MSC) derivante da tumori sierosi ovarici di alto grado (HG-SOC-MSCs) o da tessuti normali (N-MSCs) inclusi nel dataset FANTOM5. Ho messo in luce programmi funzionali condivisi tra le due linee cellulari e osservato che le differenze principali tra le funzioni attivate nelle due linee sono di tipo quantitativo più che qualitativo. I risultati suggeriscono inoltre che le HG-SOC-MSCs sono simili alle cellule mesoteliali e alle cellule del tessuto muscolare liscio. Inoltre, ho analizzato l’intero dataset usando ScanAll, un nuovo software utile a predire ab initio la presenza di elementi arricchiti nelle regioni geniche che circondano i promotori trovati del progetto FANTOM5. Ho individuato moduli di regolazione, ossia gruppi di motif che si trovano a distanze predefinite sul genoma uno rispetto all’altro. Questi moduli sono arricchiti in regioni del genoma co-espresse rispetto a sequenze generate casualmente. Infine ho creato un compendio di fattori di trascrizione espressi e che partecipano ad interazione proteina-proteina. Cancer is a disease characterized by an extreme molecular complexity. Omics approaches, collecting data in public databases for all the genome, transcripts and proteins, attempt to overcome this complexity and find the functional modules that perform the functions involved in tumour related processes. For instance, cancer tissues gene expression profiles are widely used to define genes signatures and test their clinical relevance. I used this kind information in order to characterise interesting genes in breast cancer models. On the other hand, cellular models datasets could provide data that permits to focus on specific molecular mechanisms and probe the effects of molecules in a specific cancer model. One of the most recent omics project is the FANTOM5 project, that has generated a unique resource, the first single molecule sequencing-based expression atlas in mammalian systems. Cap analysis of gene expression (CAGE) was used to measure transcription start sites (TSS) and promoter usage across a wide collection of human samples thereby identifying and measuring levels of the majority of coding and non-coding transcripts in the human genome. I used this information to characterize a mesenchymal/stromal stem cell line (MSC) derived from high-grade serous ovarian cancer (HG-SOC-MSCs) or derived from normal tissue (N-MSCs) included in the entire FANTOM5 human dataset. I highlighted shared functional programs between HG-SOC-MSCs and N-MSCs suggesting that the global differences between the two cell lines are based on quantitative levels of transcriptional output rather than on qualitative differences. The results suggested that HG-SOC-MSCs are close relatives of mesothelial cells and smooth muscle cells. Furthermore, we analysed the entire dataset using ScanAll, a newly developed software, to ab initio predict the presence of enriched elements in the genomic regions surrounding FANTOM5 promoters. I pinpointed regulatory modules, i.e. groups of enriched motifs co-occurring in co-expressed regions within a fixed distance. These modules are enriched in the co-expressed sequences in each sample respect to random generated sequences. Finally, I created a Compendium of putative expressed and directly interacting transcription factors. XXVII Ciclo 1986

2013/2014 Dietary proteins are the source of the amino acids required by the body for tissue growth and maintenance. The Population Reference Intake (PRI) for proteins, as defined by the European Food Safety Authority (EFSA) for healthy adults, including the elderly, is 0.83 g/kg body weight/day. This amount is defined on the net balance of body protein (or “nitrogen balance”, given by the difference between dietary nitrogen intake and losses) equivalent to 0.66 g/kg/day plus a safety factor for interpersonal variability and differences in proteins quality of mixed diets. The PRI, however, is the minimum daily amount of protein needed to maintain the nitrogen balance and avoid a progressive loss of lean body mass in healthy people with moderate physical activity. Therefore nitrogen balance may not be adequate to define protein requirement in adults and especially in ageing characterized by loss of muscle mass and function (sarcopenia). Furthermore until recently the prevalent idea was that a protein intake above PRI had no further benefits and on the contrary could impair health. These believes are now under discussion, diets with higher protein intake have been shown beneficial in the prevention and treatment of conditions such as sarcopenia, COPD and type 2 diabetes mellitus. There is a need of more precise methods to define protein requirement. AIM. The aim of the present thesis is to investigate in human healthy volunteers new biomarkers adequate to define optimal protein intake. Recent studies have determined protein needs by measuring whole-body protein metabolism using stable labeled isotope-amino acids. METHODS. Our research group has applied two different metabolic methods based on the most widely used tracer, i.e. D5-Phe stable isotope, in two experimental bed rest campaigns (FP7 PLANHAB and INTERREG PANGaA) in healthy volunteers. BR is a suitable model to investigate physiologic adaptation to inactivity. MAIN RESUTLTS. FP7 PLANHAB. We applied the stable isotope infusion technique, to assess the effect of physical inactivity and/or hypoxic condition on whole body protein turnover as previously described in Biolo et al 2008. Chronic hypoxia has been associated with an overall reduction in protein synthesis and in total plasma and skeletal muscle protein content. During the PLANHAB study we investigated, through a crossover randomization, the net effects of 10 days normobaric hypoxia (4000 mt.), associated with either ambulatory conditions or BR, in 11 young (age 24±4 yr), healthy and normal weight male subjects maintained on eucaloric diets. Main results. Hypoxia in ambulatory conditions significantly decreased whole body protein turnover by reducing both protein synthesis (-8±2%) and protein degradation (-8±3%). Hypoxia during bed rest did not caused significant changes in protein metabolism. INTERREG PANGaA. The skeletal muscle loss in aging is caused mainly by the “anabolic resistance” i.e. the inadequate increase in the rate of protein synthesis in response to nutritional-metabolic stimuli, including exercise, protein and amino acid intake as well as insulin and insulin-like growth factor stimulation. As a consequence, the net protein balance becomes negative leading to sarcopenia. The effects of ageing on the anabolic resistance induced by inactivity are poorly investigated. During the PANGeA study we had the opportunity to perform the second documented experimental BR in in healthy elderly volunteers and the first comparing aged with young subjects. To evaluate the anabolic resistance associated with ageing and inactivity, we enrolled 7 young (23±1yr) and 8 elderly (59±1yr) normal weight individuals, in a 14-d experimental BR protocol. We replaced our previous infusion method with a new, simpler, safer and quicker technique, by which tracers are given orally instead of parenterally, the all procedure is completed in two hours, instead of 6, and only two blood draws versus 7 are sufficient. Main results. At baseline parameters of anabolic sensitivity were comparable between young and elderly individuals. The anabolic resistance significantly increased after BR in both groups (bed-rest effect p

2012/2013 The human peptide LL-37 is an important innate immune effector that contributes to defending the organism against infection in different scenarios, ranging from direct bacterial killing to the modulation of immune responses and acting as a signal molecule for different cell types involved in defence or healing. All these functions are possible despite its relatively simple structure, due to its capacity to interact with different types of biological membranes. This cationic peptide is able to switch from a random coil in aqueous solutions into an amphipathic helical structure in presence of salt ions or in the presence of membranes. Intra- and inter-molecular hydrophobic interactions in the hydrophobic sector of the amphipathic helix, together with ionic interactions in the polar sector, make LL-37 prone to self-association, which strongly affects its multiple biological roles, and these therefore need to be explored in detail. In this thesis I have investigated how folding occurs when LL-37 passes from bulk solution to the surface of membranes, and which structural features could be relevant for its in vitro antimicrobial and host cell modulating activities. For this reason comparative biophysical studies were performed on LL-37 and its structurally diverse orthologue in macaque, RL-37. Fluorescent analogues of LL-37 and RL-37, obtained from the systematic substitution of Phe residues with the non-invasive fluorescent probe, p-cyanophenylalanine, contributed to elucidate differential conformational variations in these peptides. Steady-state and time-resolved fluorescence studies confirmed that unlike RL-37, which is monomeric in physiological solution or in presence of neutral model membranes, LL-37 is in an oligomeric form in solution, with a hydrophobic core likely forming near the central region. Furthermore LL-37 appears to interact with both anionic and neutral membranes in an oligomeric form. Cytofluorimetric studies, performed on bacterial cells treated at toxic peptide concentrations, confirmed that the different conformations of LL- and RL-37 in solution give rise to quite different modes of bacterial permeabilization. Lesions of a larger size were observed for LL-37, consistent with a toroidal pore-forming mechanism, while smaller lesions were detected for RL-37 and may indicate an incipient detergent-like mechanism of bacterial inactivation. I further investigated how many molecules are involved in the self-associated form of LL-37, either in solution or in presence of membranes, replacing the Phe5 residue with an analogue bearing Benzoyl-phenylalanine on the side-chain, which is able to crosslink to adjacent molecules when irradiated with UV light. Photo-crosslinking studies revealed that LL-37 is in equilibrium between forms ranging from monomeric to hexameric in physiological solution, and dissociates to lower order oligomers in the presence of membranes, but does not become completely monomeric. The effects of oligomerization on the biological activities of LL-37 were further explored through testing its antiparallel and parallel covalently bound dimers, linked at either end, designed to mimic an obligatory dimeric form of the peptide. Both types of dimerization reduced the antibacterial potency of LL-37 and its capacity to permeabilize the bacterial membrane. Cytotoxic effect against the host cells was instead different for the three forms of dimers tested, and was related to the degree of freedom of the C- or N-terminal region. From the information obtained on the role of oligomerization on the antibacterial and cytotoxic activity of LL-37, I then rationally designed four different analogues of LL-37, which bear minimal primary structure alterations from the native peptide, such as the replacement of one, or exchange of two adjacent residues, in order to modify inter- and intra-molecular salt-bridging. I attempted in this way to transfer some of the RL-37 characteristics to LL-37 in order to improve its antimicrobial properties through reducing its propensity to self-assemble. I confirmed that variations in the equilibrium between attractions or repulsions in LL-37 significantly affect its antibacterial activity and mode of action. Il peptide antimicrobico dell’uomo LL-37 è un componente importante del sistema immunitario che contribuisce a difendere l’organismo dalle infezioni, per via di diversi meccanismi, che possono variare dall'azione diretta battericida, alla modulazione della risposta immunitaria, agendo da molecola segnale per diversi tipi di cellule coinvolte nella difesa oppure nella guarigione dell'organismo ospite. Tutte queste attività sono dovute alla sua struttura relativamente semplice, ma capace di interagire con diversi tipi di membrane biologiche. Questo peptide carico positivamente è in grado di assumere una conformazione elicoidale in presenza di soluzioni saline oppure in presenza delle membrane. La presenza di interazioni idrofobiche, inter- e intra-molecolari, nel settore idrofobico dell’elica amfipatica insieme alle forze ioniche presenti nella regione polare causano aggregazione di LL-37. Questa caratteristica a sua volta influenza fortemente le diverse attività biologiche di LL-37 e perciò è un aspetto importante che necessita di essere studiato in modo approfondito. In questa tesi io ho investigato come avviene la strutturazione del peptide quando esso passa dalla soluzione alla superficie della membrana, e quali sono gli aspetti strutturali rilevante per l'attività antimicrobica in vitro e per la modulazione della risposta delle cellule dell’ospite. Per questo motivo sono stati usati vari metodi biofisici per confrontare la struttura di LL-37 con il suo ortologo nella scimmia RL-37. Sostituendo un residuo di fenilalanina con la sonda fluorescente non invasiva p-cyano-fenilalanina è stato possibile ottenere vari analoghi fluorescenti dei due peptide nativi, disegnati per ottenere maggior informazione a riguardo le variazioni conformazionali e per investigare il meccanismo d’azione di LL- e RL-37. Gli studi con gli analoghi fluorescenti hanno confermato che diversamente da RL-37, che è monomerico in soluzioni fisiologiche o in presenza di membrane neutre, LL-37 forma un oligomero in soluzione, con un nucleo idrofobico attorno alla regione centrale dell’elica. Inoltre, LL-37 interagisce con le membrane sia cariche positivamente che neutre nella forma oligomerica. Dagli studi fatti con il cyto-fluorimetro, trattando le cellule batteriche con concentrazioni tossiche di peptide, si è notato che i due peptidi con conformazione diversa in soluzione interagiscono e permeabilizzano le membrane in modo diverso. Le lesioni causate nei batteri dopo il trattamento con LL-37 erano regolari e di dimensioni tali da suggerire un meccanismo battericida per via di formazione dei pori di membrana, invece RL-37 causa danni di piccole dimensioni alla membrana batterica, compatibile con un meccanismo d’azione simile a quello di un detergente. Inoltre io ho indagato sul numero di molecole che partecipano alla formazione della forma oligomerica di LL-37, sia in soluzione che in presenza delle membrane. Questo è stato possible tramite la sostituzione di Phe5 con un residuo analogo benzoyl-fenilalanina, il quale permette di legare covalentemente molecole adiacenti, quando viene irradiato con raggi UV. Questi studi hanno dimostrato che LL-37 è in equilibrio tra la forma monomerica ed oligmerica, con estensione fino ad esamero in soluzioni fisiologiche, per poi dissociarsi a forme inferiori oligomeriche in presenza delle membrane. Gli effetti dell’oligomerizzazione del peptide LL-37 sulle sue attività biologiche sono stati ulteriormente investigati tramite lo studio svolto con i dimeri covalenti di LL-37, legati con un ponte di solfuro nella regione C-terminale oppure N-terminale. Questi dimeri, paralleli o antiparalelli sono stati ideati per mimare la fomra oligomerica di LL-37. Entrambi i tipi di dimerizzazione hanno mostrato che l’aggregazione di LL-37 porta ad una riduzione dell’attività antimicrobica e dell’efficienza del peptide a permeabilizzare le membrane batteriche. Invece la cito-tossicità verso le cellule dell’ospite è diversa per tutte e tre forme dimeriche ed è legata al grado di libertà della regione N-terminale del peptide. Date le informazioni ottenute sul ruolo dell’oligomerizzazione sull'attività antibatterica e citotossica di LL-37, io ho disegnato 4 analoghi di LL-37 ottenuti da una minima alterazione della struttura primaria rispetto al peptide nativo, come per esempio lo scambio della posizione di uno o più residui adiacenti in modo da alterare la rete di ponti saline tra le molecole e interamente la struttura del peptide. In questo modo si è voluto attribuire alcune caratteristiche di RL-37 al peptide dell’uomo LL-37, in modo da modificare la sua propensione ad assemblarsi per formare oligomeri. Con questi studi si è confermato che variazioni nel equilibrio delle forze attrattive o repulsive nella catena di LL-37 possono alterare significativamente la sua attività antibatterica e il modo d’azione. XXVI Ciclo 1985

2012/2013 The hepatocarcinogenesis consists of many steps and long term courses from normal to malignant tissue. Besides the various etiological factors, chronic exposure to a wide variety of substances can induce an inflammatory response and oxidative stress with consequent hepatocytes dead and regenerative phenomena. When the replicative ability of hepatocytes is impaired, the activation of the hepatic stem cell (SCs) compartment permits the recovery of the mass and functionality of the liver. The cancer stem cells (CSCs) theory postulates that cancer is composed in a hierarchy consisted of many heterogeneous cells with different grades of differentiation, in which only the CSCs are capable of initiating cancer growth. In exception of their tumor-initiating capacity, the CSCs share similar phenotypic signatures and functional properties with normal stem cells. Until now, the origin of the CSCs in hepatocellular carcinoma (HCC) is still unclear. The general objective of this project is to investigate the involvement of the SCs in the hepatocarcinogenesis, starting from early injury to development of HCC. In this study, we employed the HBV-transgenic mouse C57BL/6J-TG(ALB1HBV)44BRI/J (TG) that develops a progressive hepatic damage that well mimics the natural history of human hepatocarcinogenesis. Based on the analysis of the different SCs markers (CD34, CD133, Epcam, Krt19) expressed in hepatic tissues at different courses of the disease (from 3 months to 18 months), we observed the activation and the progressive amplification of the SCs compartment. In particular, there was a strong correlation between the expression pattern of CSCs markers and different stages of hepatic pathologies: a progressive increase of CD133, CD34 and Afp expression was noticed from normal, early cellular damage, dysplasia, and HCC, while Krt19 was decreased in HCC compared to other pathologies. We also performed the isolation and characterization of putative hepatic SCs and CSCs from both TG and their wild type counterpart C57BL/6J (WT) animals in different growth mediums. The isolated primary cells showed up-regulation of the SCs markers compared to total tissue with low or no expression of hepatocytes markers (Alb and Krt18). They had the ability to form 3D clones in matrigel after 7-10 days plating. In summary, our data indicates that this mouse model is well-corresponded with the natural history of hepatocarcinogenesis in human and can represent a good tool to study the involvement of SCs and CSCs in the progression of the disease. In addition, the possibility to isolate SCs and CSCs will be a potent tool to study the differences of developmental and oncogenic molecular pathways between normal SCs and CSCs and to develop novel CSCs-targeted therapy. L’epatocarcinogenesi è un processo lungo caratterizzato da varie fasi che portano dal tessuto sano a quello tumorale. Indipendentemente dai vari fattori eziologici, l’esposizione cronica a un ampio spettro di sostanze può indurre una risposta infiammatoria e la comparsa di stress ossidativo che insieme portano a fenomeni di morte cellulare e rigenerazione. Quando le capacità replicative degli epatociti sono compromesse, l’attivazione del compartimento staminale epatico (SCs) permette il recupero della massa e delle funzionalità epatiche. La teoria delle cellule tumorali staminali (CSCs) ipotizza che il tumore sia composto da cellule eterogenee con diversi gradi di differenziamento organizzate in maniera gerarchica, tra cui solo le CSCs sono in grado di sostenere la crescita tumorale. Oltre la capacità di iniziare il tumore, le CSCs condividono un fenotipo e una marcatura molecolare simile a quella delle cellule staminali normali (SCs). Ad oggi, l’origine delle CSCs nell’epatocarcinoma (HCC) è ancora poco chiara. Lo scopo generale di questo progetto è lo studio del coinvolgimento delle SCs nel processo di trasformazione neoplastica, dall’iniziale danno epatico allo sviluppo dell’HCC. In questo studio abbiamo utilizzato il topo transgenico per HBV C57BL/6J-TG(ALB1HBV)44BRI/J (TG) che sviluppa un danno epatico progressivo che ripercorre la storia naturale dell’epatocarcinogenesi nell’uomo. In base all’analisi di vari marcatori per le SCs (CD34, CD133, Epcam, Krt19) nel tessuto epatico a vari stadi della patologia (da 3 a 18 mesi), abbiamo osservato l’attivazione del compartimento epatico staminale e una sua progressiva amplificazione durante la progressione del danno. In particolare, abbiamo osservato una forte correlazione tra il pattern di espressione dei marcatori delle SCs e lo stadio della patologia: l’esperssione di CD133, CD34 e Afp è progressivamente aumentata nella progressione da tessuto normale a danno cellulare epatico lieve, displasia e HCC, mentre l’espressione di Krt19 ha dimostrato una riduzione nell’espressione nel HCC rispetto le altre patologie. Abbiamo anche isolato e caratterizzato delle SCs e CSCs putative sia da animali TG che da animali wild type C57BL/6J (WT) cresciute in diversi terreni. Le cellule primarie isolate hanno mostrato una maggiore espressione dei marcatori staminali rispetto al tessuto totale di origine e nessuna o bassa espressione di marcatori per gli epatociti (Alb e Krt18). Inoltre queste cellule hanno dimostrato la capacità di formare cloni tridimensionali in Matrigel dopo 7-10 giorni di coltura. Per concludere, questi dati indicano che il modello murino utilizzato ben rappresenta la storia naturale dell’epatocarcinogenesi umana e può rappresentare un buon strumento per lo studio del ruolo delle SCs e delle CSCs nella progressione della malattia. Inoltre, la possibilità di isolare SCs e CSCs rappresenta un potente strumento per lo studio delle differenze tra le vie molecolari coinvolte nello sviluppo e nella trasformazione oncogenica tra SCs e CSCs e per lo sviluppo di nuove strategie terapeutiche specifche per le CSCs. XXVI Ciclo 1982

2012/2013 Cancer stem cells (CSCs) are proposed to be responsible for breast cancer heterogeneity, chemotherapeutic treatment failure, metastatic spread and disease recurrence. The precise identification of the molecular bases that govern the induction and maintenance of CSCs and their aggressive phenotypes is of utmost importance, since it may provide the rational to develop effective therapeutic strategies. In particular there is a considerable effort in finding common pathways, mutations or histological features that might be targeted for therapy, overcoming breast cancer heterogeneity. Here we now demonstrate that CSC self-renewal, chemoresistance, tumour growth and metastases formation capabilities’ are under direct control of Pin1’s enzymatic activity on the Notch signalling pathway. In particular Pin1 protects the nuclear activated forms of Notch1 and Notch4 (N1/4-ICD) from their E3-ubiquitin-ligase Fbxw7α, thereby boosting their protein levels and transcriptional activity. Fbxw7α acts as a potent inhibitor of CSCs maintenance by promoting protein degradation of N1- and N4-ICD, and, as a consequence, this ubiquitin-ligase strongly decreased tumour growth and metastases dissemination in vivo. Interestingly, concomitant over-expression of Pin1 almost completely recovered all these aggressive breast cancer traits. In tissues from breast cancer patients, we observed Notch signalling over-activation despite presence of the negative regulator Fbxw7α, which relied on high Pin1 protein levels. Notably, activation of the Notch-Pin1 axis correlated with poor prognosis in these patients. As a consequence of our findings, suppression of Pin1 holds promise in reverting aggressive phenotypes in breast cancer though shrinkage of CSCs number and a concomitant gain in chemosensitivity, carrying important implications for breast cancers therapy. XXVI Ciclo 1985

2012/2013 Il peptide 2A è un peptide virale di 18-23 aminoacidi presente in alcuni picornavirus, dei virus a RNA che codificano per una poliproteina che viene poi processata per produrre tutte le proteine virali. Il peptide 2A, in particolare, permette la divisione tra le proteine capsidiche e quelle replicative. Durante la traduzione il peptide 2A agisce come elemento autonomo, in grado di mediare il suo processamento al C-terminale: per farlo modifica l'attività del ribosoma, promuovendo l'idrolisi del legame estere tra il peptide nascente e il tRNA anziché che la formazione del legame peptidico, con il conseguente rilascio prematuro della proteina a monte e la continuazione della traduzione dalla Pro a valle. Abbiamo chiamato questa proprietà "terminazione non convenzionale della traduzione", per distinguerla dalla terminazione convenzionale che si verifica in presenza di un codone di STOP e che viene mediata dai fattori di rilascio 1 e 3. Per questa sua attività, sia degli aminoacidi Asn-Pro-Gly al C-terminale del peptide 2A che la Pro al N-terminale della proteina a valle sono essenziali. Grazie alla sua capacità di autoprocessamento, il peptide 2A può essere utilizzato per produrre due proteine da un unico open reading frame (ORF) . Nella prima parte del lavoro abbiamo testato due differenti peptidi 2A (uno derivante dal Foot-and-Mouth Disease Virus (F2A), e uno dal Porcine Teschovirus (T2A) per trovare eventuali differenze tra la loro attività in termini di terminazione non convenzionale e di reinizio della traduzione dalla Pro a valle. Entrambi questi peptidi si sono mostrati molto efficienti nella terminazione non convenzionale, come si nota dalla quasi totale assenza di proteina di fusione. Tuttavia, il reinizio della traduzione dalla Pro valle 2A non è completo, come si evince dal fatto che la proteina a valle viene prodotta in una quantità inferiore rispetto a quella a monte. Abbiamo determinato che F2A funziona meglio T2A, in quanto consentee la sintesi della proteina a valle con maggiore efficienza. Questi risultati indicano che il peptide F2A è un ottimo sistema per la co-espressione di proteine in vivo. Abbiamo sfruttato questa tecnologia per la produzione di anticorpi ricombinante, riuscendo a produrre sia in vitro che in vivo la catena leggera e pesante dell’anticorpo, che inoltre si assembla correttamente. Nella seconda parte del lavoro abbiamo studiato una caratteristica particolare del peptide 2A: quando viene posizionato immediatamente a monte di un codone di stop, l'espressione della proteina a monte è fortemente compromessa. Questa caratteristica è dovuta alla sequenza amminoacidica, e non a quella nucleotidica. La causa di questa compromissione dell’espressione risiede in un forte stalling del ribosoma al sito Gly-STOP, totalmente dipendente dagli ultimi tredici residui del peptide 2A e dalla presenza degli amminoacidi C-terminali Asn-Pro-Gly terminali. Come conseguenza del blocco, viene attivato un controllo di qualità a livello del ribosoma, che induce l’ubiquitinazione e la degradazione da parte del proteasoma della piccola quantità di proteina prodotta. Questa attività di controllo da parte del peptide 2A sulla traduzione è efficace sia su ribosomi legati alla membrana dell’ER, che su quelli liberi nel citosol. Per i ribosomi legati alla membrana, abbiamo identificato il coinvolgimento di alcuni membri del pathway di retrotraslocazione come l’AAA ATPasi p97 e la deubiquitinase YOD1, suggerendo un meccanismo di controllo di qualità che rileva la presenza di ribosomi in stallo sul lato citosolico della membrana dell’ER e la segnala ad effettori presenti nel lato luminale per indurre la degradazione associata all’ER (ERAD) dei peptidi nascenti. E’ noto che uno dei principali mediatori dell’ERAD è l'AAA ATPasi p97/VCP; si che p97 sia necessaria per la retrotraslocazione e la degradazione da parte del proteasoma dei substrati ERAD, e che la sua perdita provoca la loro ritenzione nel lume dell’ER. Di conseguenza, l’attività ATPasica di p97 è ritenuta essenziale per l'estrazione delle proteine dalla membrana dell’ER . Per studiare il ruolo di p97 in ERAD, abbiamo usato un metodo recentemente sviluppato nel nostro laboratorio (Petris et al., 2011) che consiste nella mono-biotinilazione in vivo, da parte della biotin ligasi citosolica Bira, di substrati proteici taggati con il Biotin Acceptor Peptide BAP (GLNDIFEAQKIEWH). Questo metodo consente una discriminazione precisa tra le proteine che si trovano nel lume dell’ER (non biotinilate) e nel citoplasma (biotinilate). Abbiamo analizzato l'effetto di: 1) una forma dominante negativa di p97 (p97QQ), 2) un siRNA specifico per p97, e 3) l’inibitore chimico di p97 DBeQ, in tre diversi substrati modello di ERAD: NS1, il mutante Null Hong Kong dell’α1-antitripsina (NHK-α1AT), e Tetherin. I nostri risultati sono in contrasto con la visione predominante di p97 come fornitore di energia per l'estrazione delle proteine dall’ER in maniera ATP-dipendente. Abbiamo trovato che l'attività di p97 non è coinvolta nella retrotraslocazione di queste proteine, ma piuttosto nella separazione delle proteine, già nel lato citosolico, e nel reclutamento della PNGase. Risultati simili sono stati ottenuti analizzando l'effetto di una forma dominante negativa della deubiquitinase p97-associata YOD1 (YOD1 C160S). The 2A peptide is a short viral peptide of 18-23 amino acids, present in some picornaviruses. These RNA viruses encode a polyprotein that is late-processed to produce all viral proteins. 2A allows the primary cleavage between the capsidic and the replicative proteins. It works as an autonomous element during translation, capable of mediating self-processing at its C-terminus. It modifies the ribosome activity, promoting hydrolysis of the ester bond between the nascent peptide and the tRNA rather than the formation of the peptide linkage, resulting in a premature release of the upstream protein and continuation of translation from the in-frame downstream Pro codon. We call this property “non-conventional translation termination”, in contrast to the “conventional translation termination” occurring at a STOP codon site and mediated by the Release Factors 1 and 3. For its activity both the amino acids Asn-Pro-Gly at the C-terminus of the 2A peptide, and the Pro at the N-terminus of the downstream protein are essential. Due to its property of self-processing at the C-terminus, in mammalian cells it can be used to produce two proteins from a single open reading frame (ORF). In the first part of the work we tested two different 2A peptides (one derived from the Foot-and-Mouth Disease Virus (F2A) and one from the Porcine Teschovirus (T2A)) in order to see possible differences between their activity in term of non-conventional translation termination and re-initiation of translation from the 2A downstream Pro. We found that both these peptides were very efficient in terminating translation, in fact we observed only a very little amount of fusion protein. However, re-initiation from the 2A downstream Pro was not complete, in fact we observed that the downstream protein was produced in a lower amount than the upstream one. We also determined that F2A worked better than T2A, allowing synthesis of the downstream protein with higher efficiency. These results indicate that F2A peptide is a powerful system for the co-expression of proteins in vivo. We exploited this technology for the production of recombinant antibodies. We found that both in vitro and in vivo antibody’s light and heavy chains were produced, and the complete antibody was properly folded. In the second part of the work we investigated a particular feature of the 2A peptide: when it was positioned immediately upstream of a STOP-codon, expression of the upstream protein was heavily impaired. We found that the compromised expression was due to the amino acidic and not to the nucleotidic sequence of the peptide. We found also that it was a consequence of a strong ribosome stalling at the Gly-STOP-codon boundary, totally dependent on the last thirteen 2A residues and on the presence of Asn-Pro-Gly as C-terminal amino acids. As a consequence of the stalling, a ribosome-mediated quality control is activated, inducing ubiquitination and proteasome degradation of the small amount of protein produced. This 2A regulatory activity on translation was effective on both membrane-bound and free ribosomes. We found that, for membrane-bound ribosomes, members of the retrotranslocation pathway such as the AAA-ATPase p97 and the deubiquitinase YOD1 were involved, thus suggesting a quality control mechanism that senses the presence of stalled ribosomes on the ER membrane cytosolic side and signals to effectors present in the luminal side to induce ERAD (Endoplasmic Reticulum Associated Degradation). It is widely known that one of the main players in ER Associated Degradation is the AAA-ATPase p97/VCP. It is generally thought that p97 ATPase activity is required for retro-translocation and proteasomal degradation, and that its loss causes retention of ERAD substrates in the ER lumen. As a consequence, p97 ATPase activity is considered essential for extraction of proteins from the ER membranes. To study the role of p97 in ERAD, we used a method recently developed in our laboratory (Petris et al., 2011), which is based in the specific in vivo mono-biotinylation by the cytosolic biotin ligase BirA of protein substrates tagged with the 15 amino acids-long Biotin Acceptor Peptide BAP (GLNDIFEAQKIEWH). This method allows a precise discrimination between the proteins located in the Endoplasmic Reticulum (non biotinylated) or in the cytoplasm (biotinylated). We analysed the effect of: 1) a dominant negative form of p97 (p97QQ), 2) a p97 specific siRNA, and 3) the p97 chemical inhibitor DBeQ, on the three different ERAD model substrates NS1, Null Hong Kong mutant of α1 anti-trypsin (NHK-α1AT), and Tetherin. Our results clearly challenge the predominant view of p97 as energy provider for the extraction of unfolded proteins from a putative retrotranslocon in an ATP-dependent manner. We found that p97 activity is not involved in the retrotranslocation of these proteins from ER to cytosol, but rather in the segregation of proteins already in the cytosolic side and in the recruitment of PNGase. Similar results were obtained analysing the effect of a dominant negative form of the p97-associated deubiquitinase YOD1 (YOD1 C160S). XXVI Ciclo 1985

2012/2013 Triple-negative breast cancer (TNBC) is a subtype of breast cancer (BC) characterized by highly aggressive phenotype, unfavorable prognosis and lack of specific therapy. At the molecular level TNBC resembles basal-like subtype of BC, characterized by high prevalence of missense point mutations in the TP53 gene, one of the most frequent genetic events in human cancers. Mutant-p53 proteins on one hand lose wild-type tumor suppressive functions but, on the other, acquire oncogenic properties (“gain-of-function”, GOF) that actively contribute to tumorigenesis. Studies in aggressive breast cancer-derived cell lines underscored a central role of mut-p53 in cell-polarity disruption and acquisition of highly aggressive phenotype. Increasing knowledge about the mechanism that regulates the establishment and maintenance of cell polarity, have reveled its important contributions in preventing acquisition of tumorigenic features, but it is still challenging to understand how cancer genes and pathway influence epithelial architecture. Despite the increasing knowledge of mut-p53 functions, little is known about microRNAs associated with mut-p53. Here we investigated whether mut-p53 GOFs exerted in TNBC could be in part mediated by miRNAs whose expression is promoted mut-p53. In MDA-MB-231 (TNBC cells harboring endogenous mut-p53) we performed a screening of miRNAs previously described to be overexpressed in BC and other solid tumors. Among miRNAs positively regulated by mut-p53 we identified miR-30d. Results of further experiments suggest that mut-p53/miR-30d axis plays an important role in inducing epithelial cell polarity disruption and causes aggressive cancer phenotypes, such as cellular migration and invasion. We also identified and validated targets of the mutp53/miR-30d axis that could be involved in mediating their biological effects. In particular we identified DLG5 as a key gene whose suppression by mut-p53/miR-30d seems to be correlated with epithelial cell polarity disruption and causes cancer aggressiveness. We propose that understanding pathways affected by miR-30d will provide new insights that can be exploited in discovery of novel therapeutic targets and diagnostic markers in TNBC. Il tumore al seno Triplo Negativo rappresenta un sottotipo di tumore caretterizzato da un fenotipo altamente aggressivo, prognosi infausta, e mancanza di specifiche terapie mirate. Dal punto di vista molecolare i tumori Tripli Negativi appartengono al sottotipo Basale, caratterizzato da un’elevata frequenza di mutazioni missenso del gene TP53. Tale mutazione rappresenta una delle più frequenti alterazioni genetiche nei tumori. P53 mutata oltre ad aver perso le funzioni oncosopressive di p53, paradossalmente aquisisce anche nuove caratteristiche oncogeniche. Recenti dati in letteratura mostrano come mut-p53 svolga un ruolo chiave nel causare l’alterazione della polarità, e nel determinare fenotipi altamente aggressivi in modelli cellulari di tumore al seno. Numerose evidenze dimostrano che i meccanismi molecolari, che stanno alla base sia dei processi di determinazione che di mantenimento della polarità cellulare, esercitano un importante ruolo nel prevenire l’acquisizione di caratteristiche tumorali; ciò nonostante è ancora dibattuto come i geni coinvolti nella tumorigenesi o nelle vie di segnalazione alterate nel cancro possano influenzare e sovvertire l’architettura delle cellule epiteliali. Inoltre, nonostante le funzioni di mut-p53 siano sempre più studiate, la regolazione di microRNA da parte di mut-p53 resta ancora da esplorare. Lo scopo di questo lavoro è stato valutare se parte delle funzioni oncogeniche acquisite da p53 mutata nei tumori Triplo Negativo siano mediate da miRNA, la cui espressione è indotta da mut-p53. Pertanto in cellule MDA-MB-231, un modello di cellule di tumore al seno Triplo negativo altamente aggressivo con mutazioni in p53, abbiamo condotto uno screening di miRNA oncogenici, al fine di valutare se la loro espressione fosse influenzata dal silenziamento di mut-p53. Fra i miRNA che sono risultati positivamente regolati da mut-p53 abbiamo identificato il miRNA-30d. I risultati suggeriscono che l’asse costituito da mut-p53/miR-30d provoca l’alterazione della polarità cellulare e causa l’acquisizione di un fenotipo tumorale aggressivo, ad esempio nel promuovere migrazione ed invasione delle cellule tumorali. Inoltre abbiamo ricercato e validato alcuni dei geni bersaglio dell’asse mut-p53/miR-30d, al fine di identificare proteine che stessero mediando i loro effetti biologici. In particolare, è stata identificata DLG5, la cui soppressione da parte dell’asse mut-p53/miR-30d sembra svolgere un ruolo chiave nell’indurre perdita della polarità e correla con aggressività del tumore al seno. Con questo lavoro, noi evidenziamo l’importanza della comprensione delle vie di segnalazione disregolate da miR-30d, in quanto queste possono essere sfruttate per la scoperta di nuovi target terapeutici o marker diagnostici nel tumore al seno triplo negativo. XXV Ciclo 1985

2012/2013 Elevati livelli plasmatici di bilirubina non coniugata (UCB) sono responsabili dell’ittero neontale che è fisiologico nella maggior parte dei casi. L’iperbilirubinemia severa e prolungata nel tempo può causare encefalopatia da bilirubina e Kernicterus che, se non trattati, possono lasciare pesanti sequele neurologiche e nei casi più gravi condurre a morte. La neurotossicità da bilirubina è ancora una delle principali cause di malattie neurologiche nei paesi via di sviluppo ed è un problema riemergente nei paesi sviluppati a causa delle anticipate dimissioni dall’ospedale dei neonati. I meccanismi molecolari responsabili della neurotossicità da bilirubina non sono ancora completamente chiariti. Questo lavoro riporta i risultati ottenuti durante il mio progetto di dottorato volto a studiare il “molecular signalling” coinvolto nella neurotossicità da bilirubina. L’obiettivo principale è stato valutare gli effetti di concentrazioni pro-ossidanti di bilirubina sullo stato redox cellulare e sullo stress del reticolo endoplasmico (ER stress). Ci siamo focalizzati sulla pathway che coinvolge Nrf2, analizzando i geni indotti dalla bilirubina per effetto di Nrf2 e studiando il signalling a monte coinvolto nella sua attivazione. Parallelamente abbiamo anche studiato la cascata di segnali coinvolti nell’ER stress. Tutti gli esperimenti sono stati condotti nella linea cellulare di neuroblastoma umano SH-SY5Y, alcuni ripetuti anche nella linea di epatocarcinoma HepG2 e in colture primarie di astrociti dalla corteccia cerebrale di ratto. I nostri risultati mostrano che concentrazioni tossiche di bilirubina inducono un 40% di mortalità cellulare tra 1 e 4 ore di trattamento che si mantiene stabile fino alle 24 ore di trattamento. Le cellule trattate con UCB mostrano un incremento del livello dei ROS intracellulare dopo 1 ora seguito dall’accumulo nucleare dell’Nrf2 endogeno dopo 3 ore. La bilirubina aumenta l’induzione della trascrizione dell’ARE-GFP reporter gene associata ad una up-regolazione di diversi geni target di Nrf2. L’induzione dell’espressione genica può essere suddivisa in due categorie principali:la risposta precoce (4h-8h) e la risposta tardiva (16h-24h).La risposta precoce inizia con l’induzione dell’espressione di ATF3 dopo 4 ore di trattamento ed è seguita da i trasportatori di amminoacidi (xCT and Gly1) dopo 8h. Per la risposta tardiva abbiamo visto l’induzione dell’espressione genica degli enzimi coinvolti nella sintesi del glutatione. (γGCL and TNX1),nella risposta antiossidante e di detossificazione (HO-1, NQO1, FTH)e nell’omeostasi del NADPH (ME1, and G6PD). In seguito al silenziamento specifico di Nrf2, il trattamento con bilirubina diminuisce l’induzione dell’mRNA solo dell’HO-1 (75%), del NQO1 (56%) e della FTH (40%) Inoltre l’induzione dell’HO-1 è ridotta se le cellule vengono pretrattate con l’antiossidante NAC (65%) e con specifici inibitori per PKC (80%), P38α (40%) and MEK1/2 (25%). Risulta evidente che l’induzione di ATF3 è la prima risposta generata dal trattamento con UCB. Di seguito abbiamo osservato un’induzione sequenziale dei marker dell’ER stress: da quelli coinvolti nel signaling di PERK a 4h (PERK, ATF3, ATF4, CHOP), dalla diminuzione della proteina della ciclina D1 dopo 1 h e dall’induzione di IRE1 (XBP1), ATF6, e BiP dopo 8h di trattamento. Da notare però che il silenzia mento di PERK non riduce l’induzione dell’espressione dell’mRNA di ATFs/CHOP, ma induce l’espressione dell’mRNA di GCN2. Riassumendo noi abbiamo dimostrato che la bilirubina causa mortalità cellulare, produce la formazione di ROS, provoca l’accumulo di Nrf2 nel nucleo e induce la risposta antiossidante mediata dalle sequenze ARE. La bilirubina induce l’espressione di diversi geni coinvolti nella risposta antiossidante, tra tutti l’HO-1 e il NQO1 sono indotti dalla bilirubina in maniera dipendente da Nrf2. Abbiamo anche dimostrato che lo stress ossidativo (OS) e la PKC sono i principali fattori coinvolti nell’attivazione di Nrf2/HO-1. I risultati ottenuti dimostrano che l’induzione di ATFs/CHOP e di PERK sono uno dei primi eventi associati alla tossicità da bilirubina. Allo stesso tempo il silenziamento di PERK non influisce sull’induzione di ATFs/CHOP mentre induce GCN2, suggerendo un meccanismo di compensazione tra il signalling di PERK e GCN2. Concludendo i nostri dati dimostrano che lo stress ossidativo e lo stress del reticolo endoplasmico sono coinvolti nella neurotossicità indotta da UCB nella linea di neuroblastoma umano SH-SY5Y. Le cellule sviluppano una risposta adattativa alla bilirubina inducendo OS and ER stress e aumentando l’espressione dei geni coinvolti nella risposta antiossidante (in parte via Nrf2 pathway) e nello stress del reticolo endoplasmico (UPR). Elevated levels of unconjugated bilirubin (UCB) are responsible for neonatal jaundice, and in some case, severe hyperbilirubinemia exposes babies to bilirubin encephalopathy and kernicterus with the risk of neurological sequela and death. Bilirubin neurotoxicity is still a major cause of neurological injury in the developing countries and is a re-emerged problem in the developed countries, due to the early hospital discharge of newborns after birth. The molecular mechanisms of UCB induced neurotoxicty are incompletely elucidated. Present thesis are reported the results obtained during my PhD course aimed to investigate the molecular signaling involved in UCB induced neurotoxicity .The main goal of this work was to evaluate the effects of the pro-oxidant concentration of UCB on cellular redox state and ER stress. We focused on Nrf2 pathway, analyzing the genes induced by UCB at Nrf2-dependent manner and the up-stream signaling involved in Nrf2 pathway activation. In parallel, we also studied the ER stress cascade signaling. All experiments were conducted in SH-SY5Y neuroblastoma cell line, with some performed in HepG2 cells and primary culture of cortical astrocytes. Our results showed that SH-SY5Y neuroblastoma cells incubated with toxic concentration of UCB suffer a 40% loss of cell viability between 1h to 4h, reaching a plateau until 24h after UCB treatment. Treated cells showed an increased level of intracellular ROS after 1h followed by the nuclear accumulation of endogenous Nrf2 after 3h. UCB enhanced the transcriptional activation of ARE-GFP reporter gene associated with an up-regulation of several Nrf2 target genes. Expression response could be divided into two main categories: early (4h-8h) and late response (16h-24h). As far as early genes, UCB mediates a sequential transcription starting with the ATF3 up-regulation at 4h and followed by the induction of amino acid transporters at 8h (xCT and Gly1). On the contrary, for late genes, we observed an up-regulation of the enzymes involved in GSH synthesis (γGCL and TNX1), antioxidant/detoxification (HO-1, NQO1, FTH), and NADPH homeostasis (ME1, and G6PD). Specific Nrf2 siRNA against Nrf2 decreased the induction only of HO-1 (75%), NQO1 (56%), and FTH (40%) upon UCB exposure. HO-1 induction was reduced in cells pre-treated with antioxidant NAC (65%) and with specific signaling inhibitors for PKC (80%), P38α (40%) and MEK1/2 (25%). It was evident that ATF3 up-regulation at 4h represents the earliest response to UCB exposure. We observed a sequential activation of UPR sensors starting with PERK signaling at 4h (up-regulation of PERK, ATF3, ATF4, CHOP at 4h, and loss of cyclin D1 protein at 1h), followed by IRE1 (XBP1), ATF6, and BiP at 8h after UCB treatment. Interestingly, PERK siRNA does not changed the induction of ATFs/CHOP while induced GCN2 mRNA upon UCB exposure. In summary, we demonstrated that UCB mediates loss of cell viability, ROS generation, Nrf2 nuclear accumulation and induction of ARE. Nrf2 pathway activation was associated with the induction of multiple antioxidant genes, among all, HO-1 and NQO1 are induced by UCB at Nrf2-dependent manner. We observed that OS and PKC are the major up-stream signaling involved in Nrf2/HO-1 activation. Results demonstrated ATFs/CHOP induction and ER stress (initiated by PERK signaling) as one of the earliest event associated with UCB toxicity. However, PERK siRNA does not affected ATFs/CHOP induction by UCB while induced GCN2, suggesting a compensatory mechanism between PERK and GCN2 signaling. In conclusion, our data demonstrate that OS and ER stress are involved in UCB induced neurotoxicity in SH-SY5Y cells. The cells undergo an adaptive response against UCB induced OS and ER stress, through activation of multiple antioxidant genes (in part via Nrf2 pathway), and activation of sequential UPR sensors XXVI Ciclo 1985

2012/2013 The tumor suppressor p53 plays a central role in the protection against DNA damage and other forms of stress, primarily by inducing cell cycle arrest or apoptosis. Missense mutation of p53, which is one of the most frequent genetic alterations detected in human cancers, inactivates these growth regulatory functions; in addition, mutant p53 often acquires tumor-promoting activities (gain-of-function). A complete and thorough understanding of the signaling circuitry that regulates wild-type and mutant p53 functions is therefore a primary objective for basic cancer research, since it may lead to development of important tools for diagnosis and therapy of tumors. One crucial component of such knowledge is the molecular mechanism leading to protein activation. Deregulated ribosome biogenesis is commonly observed in cancers as a result of increased biosynthetic demand due to uncontrolled cell proliferation. Cells actively monitor the fidelity of ribosome biogenesis, activating cellular checkpoints when this process is altered. Indeed, perturbations to many aspects of ribosome biogenesis generate a “nucleolar stress” that triggers a p53 response. Evidence in animal models indicates that the ribosomal-p53 checkpoint is indeed important for tumor suppression in vivo. In the recent past, we identified the nucleolar protein GTPBP4 as a novel p53 interactor, and established a functional link between the two proteins by demonstrating that GTPBP4 depletion promotes p53 accumulation and activation. In this Thesis, I demonstrate that GTPBP4 depletion affects 28S rRNA processing, generating a “nucleolar stress” responsible for p53 stabilization. The molecular mechanism responsible for such accumulation relies on the binding of ribosomal proteins to Mdm2. In parallel, I report that the same molecular mechanism leading to wild-type p53 stabilization upon ribosomal stress, can also promote mutant p53 accumulation in cancer cells. Thus, deregulated ribosome biogenesis may contribute to the stability of mutant p53 in tumors. These results collectively suggest that upon alterations in ribosome biogenesis, the TP53 status may be critical to drive the nucleolar surveillance pathway towards a tumor suppressive or oncogenic outcome. L’oncosoppressore p53 svolge un ruolo cruciale nella protezione della cellula da danni al DNA e da altre forme di stress, principalmente inducendo arresto del ciclo cellulare e apoptosi. Mutazioni missenso nel gene TP53 sono le alterazioni genetiche più frequenti trovate nei tumori umani ed inattivano queste funzioni regolative; oltre a questo, p53 mutata acquisisce nuove proprietà oncogeniche (gain-of-function). Capire i meccanismi regolativi che sono alla base delle funzioni di p53 wild-type e mutata è quindi un fondamentale obiettivo per la ricerca di base nel cancro, dal momento che può portare allo sviluppo di importanti strumenti sia diagnostici che terapeutici. Un aspetto fondamentale di tale ricerca è la comprensione dei meccanismi molecolari responsabili dell’attivazione della proteina. La sintesi dei ribosomi è frequentemente deregolata nei tumori, in quanto le cellule hanno un aumentato fabbisogno biosintetico a causa dell’incontrollata proliferazione. La cellula monitora attivamente la fedeltà della sintesi dei ribosomi, attivando meccanismi molecolari quando questo processo è compromesso. Infatti, alterazioni ad ogni livello nella sintesi dei ribosomi genera uno “stress nucleolare” che attiva la risposta di p53. Alcune evidenze in modelli animali infatti suggeriscono l’importanza delle funzioni oncosoppressive dell’asse molecolare costituito dalle proteine ribosomali e p53. In un lavoro pubblicato recentemente, abbiamo identificato la proteina nucleolare GTPBP4 come nuovo interattore di p53, inoltre abbiamo dimostrato un collegamento funzionale tra le due proteine, in quanto la deplezione di GTPBP4 promuove l’accumulo e l’attivazione di p53. In questa Tesi ho dimostrato che la deplezione di GTPBP4 altera il processamento del RNA ribosomale 28S, causando uno “stress nucleolare” responsabile per la stabilizzazione di p53. Il meccanismo molecolare necessario per la stabilizzazione di p53 dipende dal legame di proteine ribosomali a Mdm2. In parallelo, ho dimostrato che lo stesso meccanismo molecolare responsabile dell’accumulo di p53 wild-type in seguito allo stress ribosomiale, può indurre la stabilizzazione anche di p53 mutata. Quindi, un’alterata sintesi ribosomale può contribuire alla stabilità di mutant p53. Questi risultati suggeriscono che in seguito ad alterazioni nella sintesi dei ribosomi, lo stato di p53 può essere critico nel determinare se la risposta cellulare sia di natura oncosoppressiva oppure oncogenica. XXV Ciclo 1980

2012/2013 Abstract Hyperbilirubinemia is the most common clinical diagnosis during neona- tal life. The neonatal jaundice may be physiological without any clinical consequence, or can lead to an acute form of bilirubin en- cephalopathy with minimal neurological damage, or to more severe and dangerous condition called kernicterus (permanent neurological sequelae). The Gunn rat is one of the models in which bilirubin encephalopathy was studied. It was shown that jaundiced newborn animals display an hepatic compensatory mechanism for the incapacity to eliminate bilirubin (Ugt1A1 mutant enzyme) by upregulation of cyotochrome P450 enzymes (Cyp) 1A1 and 1A2 compared with their not jaun- diced littermates. In addition, their expression in brain was selec- tively induced (cortex higher and early vs. cerebellum lower and late induction) by sudden increase of bilirubin content. Cyotochrome P450 enzymes are the main enzymes involved in the detoxification of drug and endogenous compounds. In liver their role in bilirubin oxidation has been deeply demonstrated. In brain, there are many works that analyzed Cyps expression and induction, but their role in bilirubin oxidation has never been evaluated till now. This thesis focused in; 1) analyze the inducibility of Cyps enzymes, by a known inducer, βNF in cortex and cerebellum astrocytes; 2) Select the optimal conditions to induce specifically each of them; 3) analyze their activity and 4) evaluate their capacity to oxidize bilirubin; 5) Last but not least, assess the protective effect of Cyps induction by challenging the cells with high concentrations of bilirubin.We demonstrated that astrocytes Cyps were inducible (both mRNA and activity) by βNF and this induction varies depending on the brain region and among the Cyps. To explain the differences we hypoth- esized the presence of: i) a different protein isoforms (discarded for Cyp1A1) ii) different maturation stage of astrocytes in the two brain regions (P2 rats: Cll is less developed than Cx), iii) differential regu- latory mechanisms. Our work showed that in Cx astrocytes Cyp1A1 induction (mRNA and activity) was an early event (6h), requiring low concentration of the drug, making this approach theoretically possible in vivo. On the other hand, to reach the higher levels of Cyp1A1 (mRNA and activ- ity) in Cll we needed 24h. In both cases Cyp1A1 was able to oxidize bilirubin producing an increase in viability only after TCB addition. The Cyp1A2 was the major catalyst of bilirubin degradation (with- out the need of TCB), but its modulation was more difficult to be achieved. In addition, while Cyp1A2 modulation increased bilirubin oxidation in cortex (also reflected in an increment in viability), in cerebellum we noticed a slight reduction in bilirubin clearance (no improvements in viability). All this data could lead to the basis of the cerebellum susceptibility to UCB. The observation that the TCB addition in Cyp1A1 astrocytes increases bilirubin oxidation and via- bility suggest the possibility to induce Cll resistance toward bilirubin neurotoxicity by Cyps modulation and uncoupling. --------------------------------------------- Riassunto Una delle condizioni neonatali pi`u comuni `e l’ittero. Esso rappresenta l’aumento fisiologico della bilirubina nel sangue e tessuti. Tuttavia esso pu`o divenire una condizione patologica (encefalopatia da biliru- bina), caratterizzata da un danno neurologico minimo, o condurre a una condizione severa e pericolosa per il neonato chiamata kernit- tero (danni permanenti), in caso di una esposizione prolungata ad alti livelli del pigmento. Attualmente, la patologia da iperbilirubinemia rappresenta la principale causa di riammissione in ospedale nel primo mese di vita. Per il danno neurologico da bilirubina esiste un modello animale, il ratto Gunn, che presenta la stessa mutazione genica dei pazienti Crigler-Najjar I. Come nella condizione umana, esso manca della uridin di fosfo glucuronisil tranferasi 1a1 (Ugt1A1), enzima epatico deputato alla coniugazione della bilirubina. Tuttavia l’animale ri- esce a compensare la mancanza mantenendo bassi i livelli serici di bilirubina, grazie ad una incrementata attivit`a delle citocromo P450 mono-ossigenasi epatiche (Cyps, Cyp1A1 and Cyp1A2). Tali enzimi riducono la concentrazione del pigmento attraverso la sua ossidazione. ` E stato supposto che questo meccanismo possa essere alla base anche della capacit`a di specifiche aree del cervello (cortex e collicoli superiori, vs. i danneggiati cervelletto, ippocampo e collicoli inferiori) di ridurre la concentrazione del pigmento, limitando/impedendo il danno neu- rologico associato. Sebbene il loro ruolo, i meccanismi ed i prodotti della loro attivit`a di clearance della bilirubina siano stati ampiamente dimostrati in fegato, in cervello non sono ancora disponibili evidenze funzionali della loro capacit`a di ossidare la bilirubina. L’obiettivo di questa tesi `e stato di analizzare 1) la modulazione delle Cyps cerebrali (brain Cyp: bCyp) utilizzando un noto induttore, βNF; 2) selezionare le condizione ottimali per una induzione selettiva di cias- cuna bCyp; 3) allo scopo di valutare la loro attivit`a (EROD/MROD);4) e la capacit`a specifica di ossidare bilirubina; 5) conferendo pro- tezione (aumento vitalit`a) alle culture primarie di astrociti (da cortex e cervelletto) esposte a dosi tossiche di bilirubina. Questo studio ha dimostrato la inducibilit`a delle bCyps (mRNA e attivit`a) tramite βNF. Abbiamo evidenziato una modulazione dipen- dente della regione cerebrale da cui sono state prodotte le cellule, in aggiunta ad un comportamento differenziale delle Cyps tra loro. Pos- sibili spiegazioni possono essere: i) la presenza di diverse isoforme tra le regioni (ipotesi scartata in seguito ad analisi tramite western blot); ii) diversa presenza delle vie di segnalazione alla base dei meccanismi di modulazione, o, iii) una diversa risposta dovuta al diverso stadio sviluppo degli astrociti preparati dalle due regioni (ratti P2: il Cll `e meno sviluppato del Cx). Inoltre, abbiamo documentato una precoce (6h) inducibilit`a della Cyp1A1 nei astrociti dal Cx, che inoltre necessit`a di minori concen- trazioni del principio. Tali caratteristiche suggeriscono che la Cyp1A1 sia pi`u facilmente modulabile in vivo. Tuttavia, per ottenere una buona modulazione nel cervelletto (regione pi`u sensibile al danno), bisogna mantenere la concentrazione dell’ induttore per un tempo lungo (24h). In entrambe le regioni, un incremento significativo della capacit`a di ossidare bilirubina, migliorando la vitalit`a delle cellule, `e stato ottenuto solo dopo il co-trattamento con TCB (disaccoppia- mento). La Cyp1A2 `e stata significativamente modulata solo negli astrociti da Cx (24h), in cui abbiamo evidenziato un chiara capacit`a dell’enzima di ossidare la bilirubina e migliorare la vitalit`a in assenza di TCB. I risultati ottenuti permettono di comprendere la suscettibilit`a differenziale (tra regioni del cervello) tipica di questa patologia. La capacit`a di indurre Cyp1A1 e renderli pronti a ossidare la bilirubina nei astrociti da Cll (regione pi`u danneggiata) e Cyp1A1/Cyp1A2 in Cx e oltre migliorando la viabilit`a di questi suggeriscono la possibilit`a di conferire resistenza a la tossicit`a da bilirubina. XXVI Ciclo 1986

2012/2013 Cancer and neurodegeneration are linked by a relation of inverse comorbidity, cancer patients being at lower risk for neurodegenerative disorders and vice versa. Interestingly, many cellular processes and factors contribute to both pathologies, and a central role is played by the transcription factor p53. Best known for its antiproliferative activities following transformation-related stimuli, p53 acts to maintain genetic stability and prevent tumour onset by transcriptional and non-transcriptional mechanisms. Recently, a contribution of p53 also in neuronal development and death was unveiled. In the case of Huntington’s Disease (HD), p53 mediates cytotoxicity in HD cells and animal models, whereas its inhibition prevents this phenotype. On these premises, we were prompted to investigate the signalling pathways and protein interactions that modulate p53 activation in both cancer and neurodegeneration with the aim of identifying critical hubs as new targets for therapeutic intervention. We discovered that expression of HD causative agent, i.e. mutant Huntingtin (mHtt) protein, behaves like a genotoxic stimulus in inducing phosphorylation of p53 on Ser46, that leads to its modification by phosphorylation-dependent prolyl-isomerase Pin1 and consequent induction of apoptotic target genes. Inhibition of Ser46 phosphorylation by targeting HIPK2, PKCδ, or ATM kinases, as well as inhibition of Pin1, prevented mHtt-dependent apoptosis of neuronal cells. These results provide a rationale for the use of inhibitors of stress-responsive kinases and Pin1 as a potential therapeutic strategy for HD treatment. On the other hand, we investigated the contribution of BRD7, a protein involved in epigenetic regulation, to the p53 pathway. We found that BRD7 is required for the onset of oncogene-induced senescence, a main tumour suppressive p53 activity. In addition, we found that upon oncogene activation BRD7 restrains, independently of p53, the acquisition of malignant phenotypes, such as migration/invasion and stem cell traits. We observed a strong induction of inflammatory genes after depletion of BRD7, whose contribution to this process is still under investigation. BRD7 takes part into SWI/SNF and PRC2 chromatin remodelling complexes, whose pleiotropic roles in tumorigenesis make them appealing targets for cancer therapy. We will discuss how this new generated knowledge could be exploited for the treatment of neurodegenerative diseases, in which chromatin alterations are now recognized as drivers of pathogenesis. XXV Ciclo 1983

2012/2013 The YAP and TAZ mediators of the Hippo pathway (hereafter called YAP/TAZ) promote tissue proliferation and organ growth. However, how their biological properties intersect with cellular metabolism remains unexplained. Here, we show that YAP/TAZ activity is controlled by the SREBP/mevalonate pathway. Inhibition of the rate-limiting enzyme of this pathway (HMG-CoA reductase) by statins opposes YAP/TAZ nuclear localization and transcriptional responses. Mechanistically, the geranylgeranyl pyrophosphate produced by the mevalonate cascade is required for activation of Rho GTPases that, in turn, activate YAP/TAZ by inhibiting their phosphorylation and promoting their nuclear accumulation. The mevalonate–YAP/TAZ axis is required for proliferation and self-renewal of breast cancer cells. In Drosophila melanogaster, inhibition of mevalonate biosynthesis and geranylgeranylation blunts the eye overgrowth induced by Yorkie, the YAP/TAZ orthologue. In tumour cells, YAP/TAZ activation is promoted by increased levels of mevalonic acid produced by SREBP transcriptional activity, which is induced by its oncogenic cofactor mutant p53. These findings reveal an additional layer of YAP/TAZ regulation by metabolic cues. XXVI Ciclo 1983

2012/2013 HMGA1 is an oncogene encoding for an architectural transcription factor that affects fundamental cell processes, leading to neoplastic transformation. The two main mechanisms by which HMGA1 protein is known to be involved in cancer concern the regulation of gene expression by altering DNA structure and interacting with a conspicuous number of transcription factors. Here we provide evidence of an additional level of gene expression regulation exploited by HMGA1 to exert its oncogenic activity. Starting from protein-protein interaction data showing that HMGA1 interacts with histones, we show that HMGA1 regulates gene expression by affecting the epigenetic status of cancer cells. In particular, it modulates the signalling cascade mediated by the RAS/RAF/MEKK/ERK/RSK2 pathway regulating the levels of histone H3 phosphorylation at Serine 10 and Serine 28. We demonstrate that the down-regulation of these two H3 post-translational modifications by HMGA1 silencing and by inhibitors of the RAS/RAF/MEKK/ERK pathway is linked to cell migration decrease and morphological changes resembling the mesenchymal to epithelial transition. HMGA1 è un oncogene codificante per un fattore trascrizionale architetturale che influenza fondamentali processi cellulari, portando alla trasformazione neoplastica. I due principali meccanismi tramite cui la proteina HMGA1 è nota essere coinvolta nel cancro riguardano la regolazione dell’espressione genica tramite l’alterazione della struttura del DNA e l’interazione con un cospicuo numero di fattori di trascrizione. Qui forniamo la prova di un addizionale livello di regolazione dell’espressione genica sfruttato da HMGA1 per esercitare la sua attività oncogenica. Partendo da dati d’interazione proteina-proteina che mostrano che HMGA1 interagisce con gli istoni, mostriamo che HMGA1 regola l’espressione genica influenzando lo stato epigenetico delle cellule cancerose. In particolare, essa modula la cascata di segnalazione mediata dalla via di RAS/RAF/MEKK/ERK/RSK2 regolando i livelli di fosforilazione dell’istone H3 sulla Serina 10 e sulla Serina 28. Noi dimostriamo che la down-regolazione di queste due modificazioni post-traduzionali di H3 tramite il silenziamento di HMGA1 e l’utilizzo di inibitori della via di RAS/RAF/MEKK/ERK/RSK2 è correlata alla diminuzione della migrazione cellulare e a cambiamenti morfologici che ricordano la transizione mesenchimo-epiteliale. XXV Ciclo 1983

2012/2013 High Mobility Group A (HMGA1a, HMGA1b and HMGA2) proteins are architectural nuclear factors, physiological expressed during embryonic development and re-expressed at high levels following neoplastic transformation, playing essential functions in both these processes thanks to their particular plasticity and consequently multifunctionality. HMGA are involved in a wide number of cellular processes, including Epithelial-Mesenchymal transition (EMT), a biologic developmental process characterized by the conversion of epithelial cells to motile mesenchymal ones, with increased capacity of migration and invasion. EMT plays a key role during the progression of different tumours, including breast cancer and also HMGA have been linked to these processes in the acquisition of tumourigenic features. Consequently taking advantage of different breast cancer cell lines to recreate an "EMT model" we have investigated the role of HMGA proteins in EMT and breast carcinoma. We have developed a cellular model, stable for the overexpression of HMGA1 using the human breast cancer cell line MCF7. We have explored different aspects of tumourigenesis, performing transwell migration and invasion assays, demonstrating that cells with high levels of HMGA1 migrate and invade at a higher and significant level in comparison to control cells. Moreover this data was also confirmed with the development of an inducible cell line for HMGA1 overexpression. Therefore we have examined the expression status of different genes, including several specific EMT markers at mRNA level with Real Time PCR, observing a pre-malignant change towards mesenchymal status. We have investigated the response after DNA damage induced by doxorubicin drug, by colony formation assay, demonstrating that HMGA1 overexpressing cells confer a survival advantage to the cells, being able to survive and form a significant higher number of colonies in respect to control cells. Therefore to study deeper the role of HMGA in EMT, we have developed other two cellular systems, a human cellular model of EMT in MDA-MB-468 human breast carcinoma cells treated with Epidermal Growth Factor (EGF) and the well known EMT model, elicited by Transforming Growth Factor-β (TGF-β) in murine mammary epithelial NMuMG cells, in which HMGA2 is functionally determinant. We have demonstrated by Real Time PCR of EMT markers, Western Blot analyses and immunofluorescence the effective reliability of these cellular models, confirmed also by a dramatic change in morphology of treated cells, towards a mesenchymal phenotype. Concluding we have interestingly observed that overexpression of HMGA1 could confer some tumourigenic features (i.e. migration, invasion) and survival advantage to the cells in the MCF7 model after a cellular DNA damage induction; therefore we have different suggestions that HMGA are involved in EMT in other different cellular models. Le proteine HMGA (HMGA1a, HMGA1b e HMGA2), definite come fattori architetturali della cromatina, sono fisiologicamente espresse ad alti livelli nel corso dello sviluppo embrionale diminuendo gradualmente la loro espressione nel corso del differenziamento. Sono coinvolte, oltre all'aspetto fisiologico, anche in diverse condizioni patologiche, essendo ad esempio ri-espresse ad alti livelli nel corso della trasformazione neoplastica, esercitando funzioni essenziali grazie alla loro alta plasticità, alle peculiari caratteristiche biochimiche e conseguente multifunzionalità. Le proteine HMGA utilizzano diversi meccanismi per esercitare la loro funzione nell'acquisire capacità trasformanti, inclusa la transizione epitelio-mesenchimale. Questo processo biologico, primariamente identificato come fattore chiave dello sviluppo embrionale, è risultato di essere di fondamentale importanza anche nella trasformazione tumorale. Mediante questo meccanismo una cellula epiteliale, mediante molteplici cambiamenti genetici e biochimici acquisisce caratteristiche tipiche di uno "stato mesenchimale", caratterizzato ad esempio da un'aumentata capacità invasiva e migratoria. La transizione epitelio-mesenchimale esercita un ruolo chiave nel corso della progressione di diverse tipologie tumorali, incluso il cancro al seno, a cui in particolare anche le proteine HMGA sono state associate. L'obiettivo della Tesi è quindi quello di studiare il ruolo delle proteine HMGA nella transizione epitelio-mesenchimale e in particolare nel cancro al seno. A questo scopo abbiamo sviluppato diversi modelli cellulari di transizione epitelio-mesenchimale. Il primo modello ha previsto la creazione di un sistema stabile di over-espressione della proteina HMGA1 nella linea epiteliale di tumore al seno MCF7. Abbiamo analizzato diversi aspetti della tumorigenesi mediante saggi di migrazione ed invasione in transwell, dimostrando come alti livelli della proteina HMGA1 inducano un aumento di entrambi i processi rispetto ad una condizione di controllo. Inoltre i dati di migrazione sono stati confermati in un sistema inducibile per la over-espressione di HMGA1 nella stessa linea cellulare MCF7 e da saggi condotti in condizione di deplezione di HMGA1 attraverso strategie di silenziamento, dimostrando ulteriormente come la migrazione sia un fenomeno HMGA1 dipendente. Abbiamo inoltre esaminato lo stato di espressione di diversi geni, inclusi specifici marker di transizione epitelio-mesenchimale, mediante analisi di Real Time PCR, osservando un cambiamento verso una condizione di tipo pre-maligno e di parziale transizione ad uno stato mesenchimale. Inoltre è stata verificata la risposta al danno indotto da doxorubicina mediante saggio di colony formation, dimostrando come cellule over-esprimenti HMGA1 possiedano un vantaggio in termini di sopravvivenza e di numero di colonie formate, rispetto alle cellule di controllo. Per approfondire ulteriormente il ruolo esercitato dalle HMGA nella transizione epitelio-mesenchimale, sono stati sviluppati altri due modelli cellulari, uno nella linea epiteliale umana di cancro al seno MDA-MB-468 trattata con EGF (Epidermal Growth Factor), l'altro nella linea cellulare murina mammaria di tipo epiteliale NMuMG, trattata con TGF-β (Transforming Growth Factor-β), in cui l'azione di HMGA2 è stato dimostrato avere un ruolo determinante. Mediante analisi di Real Time PCR di marker di transizione epitelio-mesenchimale, di Western Blot e di immunofluorescenza abbiamo dimostrato l'effettiva solidità di questi modelli cellulari, confermato anche dal fatto che è possibile apprezzare un consistente cambio morfologico verso un fenotipo mesenchimale e una concomitante over-espressione delle proteine HMGA. Da questi modelli è stato quindi possibile evincere come le HMGA siano coinvolte nell'acquisizione di caratteristiche di tipo tumorale anche mediante processi di transizione epitelio-mesenchimale e come questi modelli siano utili al fine di semplificare network molecolari. XXV Ciclo 1984

2011/2012 Dendritic cells (DCs) are professional antigen presenting cells. They have the unique ability of recognizing antigens, internalizing and processing them for presentation on MHC complexes. DCs also express pattern-recognition receptors like Toll-like receptors (TLRs), which recognize pathogen-associated molecular patterns and induce DCs maturation. As a consequence DCs increase their antigens presenting ability, their levels of co-stimulatory molecules on the cell surface, and their production of cytokines, which are critical for inducing specific T cell activation. All these functions make DCs key elements in the immunological response against pathogens, viruses and tumors. One of the most important cytokines in T cell priming, interleukine-12 (IL-12), begins to be largely transcribed immediately after TLR activation. We have previously shown that in DCs engaged in antigen specific immunological synapses (IS), IL-12 is found enriched in the proximity of the contact site between DC and T cells, and several evidences suggest that its secretion is polarized towards the T cell. These data indicate that polarized trafficking of cytokines containing vesicles in DCs is a mechanism to optimize T cell priming. Polarized trafficking of soluble mediators through defined subcellular areas has been described as fundamental for the function of several other immune cells, but at present a clear characterization of the pathways and organelles implicated in the regulation of cytokine secretion in dendritic cells is still missing. The family of SNARE (soluble-N ethylmaleimide sensitive-factor accessory-protein (SNAP) receptor) proteins, which regulate the membrane fusion process during the intracellular trafficking of soluble mediators, has emerged as the most informative for the identification of such pathways. Indeed different SNAREs regulate diverse mechanisms such as the constitutive secretion of immune mediators, phagocytosis and endocytosis, or also the release of inflammatory mediators and products from secretory granules. During my thesis I have investigated the role of SNAREs in IL-12 secretion by DCs. I have analyzed the expression and modulation during maturation of three SNAREs, Syntaxin6, VAMP3 and VAMP7, that were previously shown to play a role in other immune cells. The subcellular localization and co-localization with the IL-12 subunits were analyzed in DCs alone and in the context of IS by immunofluorescence. I performed functional analyses by silencing SNAREs with the siRNA strategy to detect impairment in DC cytokines secretion and in T cell priming. These studies led to the identification of the lysosomal-late endosomal VAMP7 as a specific regulator in the polarized secretion of IL-12 during T cell priming. This suggest a role, in the secretion of soluble newly synthesized mediators, of late endosomal organelles, until now thought to be implicated only in the regulated secretion of pre-constituted mediators such as cytotoxic granules. Furthermore VAMP3 was found to downregulate the production of all the cytokines tested; we propose this SNARE as regulator of the TLR9 signaling pathway. The second part of the work stems from the concept that DCs properties become suppressed by the tumor microenvironment and many evidences suggest that tolerogenic DCs may be a crucial element in the development of tumor-mediated immune anergy. Different steps in the antigen presentation process are impaired, but also the pattern of secreted cytokines is skewed. In particular, how secretion of inflammatory cytokines becomes impaired in tumor-exposed DCs has not been investigated. I set up a model of mice bearing Lewis lung carcinoma (3LL) to analyze DCs in vitro (co-colture) and in vivo (spleen and tumor-bearing lungs). In all the models I observed a strong impairment in IL-12 production, upon restimulation with TLR agonists. In spleen DCs subtypes remained unchanged during tumor progression, whereas in tumor-bearing lungs the DC composition varied, with an expansion of CD11bhi DCs and a reduction in CD103+ DCs. DCs were isolated from all the three models to analyze the expression of trafficking proteins, and I found that tumor-exposed DCs downregulate all the SNAREs and in particular VAMP3. Moreover Rab27a, a GTPase implicated in the regulation of the exocytosis of secretory granules and lysosome-related organelles is also downregulated at protein level. These preliminary observations suggest that regulation of trafficking proteins in DCs may represent a novel pathway targeted during immunosuppression. Le cellule dendritiche (DC) sono una importante classe di cellule presentanti l'antigene. Hanno l'esclusiva abilità di riconoscere gli antigeni, internalizzarli e processarli per la loro presentazione con il complesso maggiore di istocompatibilità. Le DC esprimono anche recettori "riconoscitori di pattern" come i recettori Toll-like (TLR), che riconoscono pattern molecolari associati ai patogeni e inducono la maturazione delle DC. Come conseguenza le DC aumentano la loro capacità di presentare gli antigeni, i loro livelli di molecole costimolatorie sulla superficie cellulare, e la loro produzione di citochine, che sono fondamentali per indurre l'attivazione specifica delle cellule T. Tutte queste funzioni rendono le DC elementi chiave nella risposta immunologica contro i patogeni, i virus e i tumori. Una delle più importanti citochine per l'induzione delle cellule T, l'interleuchina 12 (IL-12), inizia ad essere estensivamente trascritta immediatamente dopo l'attivatione dei TLR. Il nostro gruppo ha precedentemente mostrato che nelle DC impegnate in sinapsi immunologiche-antigene specifiche (IS), l'IL-12 si trova arricchita in prossimità del sito di contatto tra la DC e le cellule T, e varie evidenze suggeriscono che la sua secrezione è polarizzata verso la cellula T. Questi dati indicano che il traffico polarizzato di vescicole contenenti citochine nelle DC è un meccanismo per ottimizzare l'induzione delle cellule T. Il traffico polarizzato di mediatori solubili attraverso definite aree intracellulari è stato descritto come fondamentale per la funzione di varie cellule immunitarie, ma attualmente una chiara caratterizzazione delle vie di secrezione delle citochine nelle cellule dendritiche è ancora mancante. La famiglia delle proteine SNARE (soluble-N ethylmaleimide sensitive-factor accessory-protein (SNAP) receptor), che regolano il processo di fusione tra le membrane durante il traffico intracellulare di mediatori solubili, è emersa come la più informativa nell'identificazione di questo tipo di vie. Infatti diverse SNARE regolano diversi meccanismi, come la secrezione costitutiva di mediatori immunologici, la fagocitosi e l'endocitosi, o anche il rilascio di mediatori infiammatori e di prodotti sui granoli secretori. Durante la mia tesi ho investigato il ruolo delle SNARE nella secrezione dell'IL-12 nelle DC. Ho analizzato l'espressione e la modulazione durante la maturazione delle SNARE syntaxin6, VAMP3 e VAMP7, precedentemente descritte in altre cellule immunitarie. La localizzazione subcellulare e la colocalizzazione con le subunità dell'IL-12 sono state analizzate nelle DC da sole e nel contesto dell'IS attraverso l'immunofluorescenza. Ho svolto studi funzionali silenziando le SNARE con la strategia dei siRNA, per osservare diminuzioni nella secrezione delle citochine nelle DC. Questi studi hanno portato all'identificazione della lisosomale VAMP7 come regolatore specifico nella secrezione polarizzata dell'IL-12 durante l'induzione delle cellule T. Questo suggerisce un ruolo nella secrezione di mediatori solubili neosintetizzati, degli organelli endosomiali tardivi , finora ritenuti implicati solo nella secrezione regolata di mediatori pre-costituiti come i granuli citotossici. Inoltre si è visto che VAMP3 deregola la produzione di tutte le cotochine testate; proponiamo che questa SNARE sia un regolatore della via di segnalazione del TLR9. La seconda parte del lavoro parte dal concetto che le proprietà delle DC sono soppresse dal microambiente tumorale e molte evidenze suggeriscono che le DC tollerogeniche sarebbero un elemento fondamentale nello sviluppo dello spegnimento immunologico mediato dal tumore. Diversi punti nel processo di presentazione dell'antigene sono deregolate, ma anche il pattern di citochine secrete è modificato. In particolare, come la secrezione di citochine infiammatorie diventi deregolata nelle DC esposte al tumore non è stato investigato. Ho messo a punto un modello di topi con carcinoma polmonare di Lewis (3LL) per analizzare le DC in vitro (con co-colture) e in vivo (nella milza e nel polmone). In tutti i modelli ho osservato una forte diminuzione nella produzione dell'IL-12, dopo stimolazione con agonisti del TLR. Nella milza i sottotipi di DC sono rimasti simili durante la progressione tumorale, mentre nei polmoni affetti da tumore la composizione delle DC è variata, con un'espansione delle DC CD11b e una riduzione delle DC CD103. Le DC sono state isolate da tutti e tre i modelli per analizzare l'espressione delle proteine di traffico e ho osservato che le DC esposte al tumore deregolano tutte le SNARE e in particolare VAMP3. Inoltre anche Rab27a, una GTPasi implicata nella regolazione dell'esocitosi dei granuli secretori e degli organelli lisosomiali è trovata de-regolata a livello della proteina. Queste osservazioni preliminari suggeriscono che la regolazione delle proteine di traffico nelle DC potrebbe rappresentare una nuova via bersagliata durante lo sviluppo tumorale. XXV Ciclo 1985

2011/2012 Metastasis formation is the final step in a solid tumour progression and is the most common cause of death in cancer patients. Therefore novel therapeutic strategies to prevent development of metastases have a potential impact on cancer mortality. In the last years, work from our lab has identified that, following oncogenic signalling, the phosphorylation-dependent prolyl-isomerase Pin1 amplifies mutant p53 oncogenic functions. In particular, we have demonstrated the relevance of a Pin1/mutant p53 axis in controlling directly a transcriptional “ten genes signature” program relevant for tumour growth and invasion. This signature includes genes relevant for breast cancer progression, correlated to poor outcome of breast cancer patients. Among these, we have unveiled DEPDC1 as a potent promoter of invasiveness and migration in MDA MB231 triple negative breast cancer (TNBC) cells. This gene is has been found overexpressed in different mouse and human tumours, such as breast carcinoma, colorectal and lung adenocarcinomas. The aim of my PhD work was to investigate the role of DEPDC1 in the cancer progression that is linked to cell-biological traits associated with high-grade malignancy - including motility, invasiveness and loss of polarity. We found that the high DEPDC1 expression level positively correlates with clinical outcome and aggressiveness in breast cancer and we demonstrate that DEPDC1 regulates important phenotypes involved in tumour formation and progression. First, we identified a role of DEPDC1 in promoting aggressive cancer phenotypes in vitro by regulating cell motility, polarity and proliferation. Second, we have evaluated that the establishment of lung metastasis in vivo in mice was reduced upon inhibition of DEPDC1. Third, we demonstrate the tumorigenic potential of DEPDC1-V1 alone or cooperating with oncogenic H-Ras on induction of malignant cell transformation. Finally, a preliminary data shows that DEPDC1 is also able to increase the efficiency of mammosphere formation of breast cancer cells, implicating a role in cancer stemness. Until now, our results support the strong impact of DEPDC1 on tumour progression while molecular pathways perturbed by DEPDC1 and by which drive the cancer progression are still unknown. However, our analysis of RNA-seq data upon silencing of DEPDC1, suggests the mechanism by which DEPDC1 could induce an aggressive phenotype altering the gene expression profile of breast cancer cells. Future studies will address the molecular networks by which DEPDC1 drives the metastatic cancer progression that could be useful for discovering the novel therapeutic targets and diagnostic markers in breast cancer. XXIV Ciclo 1983

2011/2012 Bac7(1-35) is the smallest fragment of the proline-rich cathelicidin Bac7 that shows the same antibacterial activity as the whole natural peptide of 60 residues. In this work, we remarked that the unique gene whose deletion can confer resistance to E. coli against Bac7(1-35) is sbmA, coding for an inner membrane protein involved in the penetration of this peptide into bacterial cells. Moreover, we provided evidence that SbmA is also involved in the transmembrane transport of a fragment of another proline-rich antimicrobial peptide, arasin1(1-23), isolated from the spider crab. These findings suggest a general role of this membrane protein in the uptake of proline-rich antimicrobial peptides (PR-AMPs) into Gram-negative bacteria. We then measured the intrabacterial concentration reached by Bac7(1-35) in E. coli, and observed that this increases from micromolar in the medium to millimolar within the bacterial cell, suggesting that it may bind to cytosolic structures. For this reason, we looked for possible interactions between Bac7(1-35) and macromolecules involved in viable processes of bacteria. These studies showed that Bac7(1-35) completely inhibits in vitro the transcription/translation process starting from a concentration of 50 μM. Then we demonstrated that inhibition is: i) specific for Bac7(1-35), since it is not exerted by other cathelicidin-derived AMPs not belonging to the Prorich group, and ii) not stereo-specific, since it is exerted at the same level by the all-D isomer of Bac7(1-35). We also demonstrated the ability of Bac7(1-35) to bind DNA in vitro, but we excluded that this binding may represent the primary mechanism of bactericidal action. We also showed that the peptide does not significantly affect in vitro the transcription process, deducing that the inhibition of the transcription/translation targets primarily the translation process. We verified these data in vivo on E. coli cells measuring the incorporation of radioactive precursors of bacterial macromolecules. We observed that the exposure of bacteria to Bac7(1-35) inhibited the incorporation of radioactive leucine, but not of radioactive thymidine and uridine, indicating a specific block at the protein synthesis level and not of DNA and RNA synthesis. In the near future, a clearer definition of the intrabacterial target(s) of Bac7(1-35) would hopefully lead to the experimentation of this molecule or of its derivatives as a new generation antibiotic drug. Bac7(1-35) è il più breve frammento del peptide Bac7, una catelicidina ricca in prolina di 60 residui, dotato della stessa attività battericida del peptide intero. In questo lavoro di tesi, abbiamo rimarcato che sbmA è l’unico gene la cui delezione conferisce ad E. coli una resistenza a Bac7(1- 35). Tale gene codifica per una proteina della membrana interna coinvolta nell’ingresso del peptide nel citoplasma della cellula batterica. Inoltre, abbiamo dimostrato che SbmA è coinvolta anche nel trasporto transmembrana di un frammento di un altro peptide antimicrobico, l’arasina1(1-23). Tali risultati suggeriscono che questa proteina giochi un ruolo generale nell’internalizzazione di peptidi antimicrobici ricchi in prolina nei batteri Gram-negativi. Abbiamo quindi misurato la concentrazione intrabatterica raggiunta da Bac7(1-35) in E. coli e abbiamo osservato che questa aumenta da valori micromolari nel terreno di coltura a millimolari nel citosol batterico, suggerendo un suo legame a strutture interne della cellula. Per questo abbiamo cercato possibili interazioni tra il peptide e macromolecole coinvolte in processi vitali del batterio. Con questi studi abbiamo appurato che Bac7(1-35) in vitro inibisce completamente il processo di trascrizione/traduzione a partire da una concentrazione di 50 μM. Successivamente abbiamo dimostrato che questa inibizione è una peculiarità di Bac7(1-35), in quanto altri AMP derivati da catelicidine ma non ricchi in prolina non hanno dimostrato un’attività comparabile. Inoltre, questa inibizione non è stereo-specifica, in quanto anche l’isomero D di Bac7(1-35) blocca tale processo esattamente come il suo isomero L. Abbiamo inoltre dimostrato la capacità di Bac7(1-35) di legare in vitro il DNA, ma abbiamo escluso che tale legame rappresenti il meccanismo primario della sua attività battericida. Abbiamo anche dimostrato che il peptide non interferisce in vitro in maniera significativa con il processo di trascrizione, deducendo che l’effetto osservato sul processo di trascrizione/traduzione fosse da attribuirsi prevalentemente all’inibizione della traduzione. Abbiamo verificato tali dati in vivo su cellule di E. coli misurando l’incorporazione di precursori radioattivi delle macromolecole batteriche. Abbiamo osservato che l’esposizione di batteri a Bac7(1-35) bloccava l’incorporazione di leucina radioattiva ma non di timidina ed uridina, indicando un blocco specifico della sintesi proteica ma non di quelle di DNA e RNA. In futuro, una definizione ancora più chiara del target intrabatterico di Bac7(1-35) potrebbe portare alla sperimentazione di tale molecola o di suoi analoghi come farmaci antibiotici di nuova generazione. XXV Ciclo 1985

2011/2012 Background: Breast carcinoma (BC) is the most common form of malignancy in women and the leading cause of cancer-related mortality among females internationally. BC includes a series of heterogeneous tumours with a great variability at histological level, biological level and clinical evolution. Because its complexity, BC treatment recommendations are continually changing with the new advances in this field. However there are still a significant number of patients with similar features that show distinct outcome. In order to detect the existing molecular differences and address patients to more personalized treatment, there is a need to find out new cancer biomarkers. Aim: The main goal of my PhD project is to investigate on the possibility of combining traditional clinical and pathological features with new candidate biomarkers for the prognostication of breast cancer. The first part concerns the prognostic role of molecular classification of primary tumours, according to luminal A, luminal B, HER-2+ and basal-like subtypes. In the second part, the prognostic value of nine candidate genes was investigated at mRNA level. The genes of interest belong to the RB pathway (CDK2, RB1), the RAS pathway (HER-2, PI3K, AKT1, AKT2, AKT3, RAF1) and cellular differentiation mechanism (CK8). Methods: This retrospective study comprises 305 BC patients, both lymph node negative (LN-) and lymph node positive (LN+). The molecular classification of primary tumours was performed by means of immunohistochemistry (IHC), using seven surrogate markers: ER, PR, HER-2, Ki67, CK8 and CK5/6, plus vimentin. Results of molecular classification were analysed with respect to morphologic and pathological features, and outcome. Moreover, the molecular characterization was also performed in a set of loco-regional metastatic lymph nodes, to compare the phenotype of primary tumour cells with their matched metastatic cells colonizing regional nodes. To the second purpose, gene’s expression was investigated in the entire cohort of primary tumours by means of real-time PCR, using the TaqMan chemistry. The expression of the nine genes was investigated in connection with the clinical-pathological factors, molecular classification and BC specific patient’s survival. Results: Regarding molecular classification of primary tumours, luminal A, luminal B, HER-2+ and basal-like accounted for 46%, 34%, 8% and 12% respectively. Luminal A tumours were mainly LN-, well differentiated and stage I, while luminal B and HER2+ showed higher tumour grade, nodal metastases as well as higher proliferation status and stage. Luminal A exhibited better survival in comparison to the other subtypes (p

2010/2011 SbmA/BacA è una proteina della membrana interna presente nei batteri Gram negativi e caratterizzata da omologia di sequenza con i domini transmembrana (TMDs) dei sistemi di trasporto di tipo ABC. Ceppi di E. coli con una delezione completa del gene sbmA mostrano un’aumentata resistenza al peptide antimicrobico Bac7 e ad altri peptidi ricchi in prolina a seguito di una loro ridotta internalizzazione. Gli omologhi di SbmA in altre specie sono essenziali per instaurare un’infezione cronica sia in ospiti di origine vegetale che animale. Lo scopo del presente lavoro è stato caratterizzare l’organizzazione e la funzione del sistema di trasporto rappresentato da SbmA in E. coli. Mediante saggi di cromatografia di affinità e di anisotropia a fluorescenza, è stato mostrato come SbmA sia in grado di legare Bac7 senza la necessità di una proteina di legame del substrato (SBP). Inoltre, mediante la tecnica del doppio ibrido in batterio e saggi di crosslinking in vitro, è stato mostrato come SbmA formi dimeri sia in vivo sia in vitro. Saggi di citofluorimetria in presenza di diversi inibitori metabolici indicano come la presenza di un gradiente protonico sia essenziale per garantire il trasporto del peptide attraverso la membrana batterica, trasporto che sembra essere indipendente dall’energia derivata dall’idrolisi dell’ATP. Allo scopo di identificare possibili interattori di SbmA, è stato analizzato il locus genomico di sbmA ed è stato dimostrato che il gene yaiW, localizzato a valle di sbmA, forma con quest’ultimo un’unica unità trascrizionale. Sebbene i due geni siano cotrascritti, le corrispondenti proteine non sembrano interagire né a livello citoplasmatico né di membrana interna. Tuttavia, l’interazione tra queste due proteine a livello periplasmico non può essere esclusa ed il coinvolgimento di YaiW nell’internalizzazione di Bac7 mediata da SbmA può essere ipotizzato sulla base dell’aumentata resistenza al peptide mostrata dal ceppo con una delezione completa del gene yaiW. Dal vaglio in vivo dell’intero proteoma di E. coli versus SbmA mediante il sistema BACTH non è stato identificato alcun dominio di legame dell’ATP, rafforzando quindi l’ipotesi che il sistema di trasporto rappresentato da SbmA non richieda l’idrolisi dell’ATP come fonte di energia. Questo studio ha permesso di identificare 9 possibili interattori di SbmA, tra i quali la proteina della membrana interna FieF, coinvolta nell’efflusso di ioni Zn(II) dalla cellula batterica. E’ stato mostrato come FieF sia coinvolta nell’internalizzazione di Bac7 mediata da SbmA e come l’espressione di sbmA sia modulata dagli ioni Zn(II) in maniera complessa, suggerendo un possibile ruolo per questa proteina nella risposta allo stress da parte del batterio. Questi risultati suggeriscono come SbmA sia un insolito trasportatore di peptidi derivato dalla famiglia dei trasportatori ABC, che ha successivamente perso e/o cambiato nel corso dell’evoluzione alcune caratteristiche, come ad esempio la necessità di una proteina di legame del substrato o la fonte di energia richiesta per il trasporto dei substrati. Inoltre, l’interazione osservata tra le due proteine SbmA e FieF potrebbe suggerire un possibile ruolo della prima nel processo di detossificazione dello zinco, attività che potrebbe essere modulata dal legame di un substrato di natura peptidica a SbmA. Infine, la regolazione dell’espressione di sbmA mediata dagli ioni Zn(II) potrebbe avere un ruolo fisiologico nell’ospite in quanto promuove una maggiore suscettibilità dei batteri ai peptidi antimicrobici ed è in grado di attivare effettori cellulari come i neutrofili. SbmA/BacA is a protein of the inner membrane of many Gram negative bacteria, predicted to be the transmembrane domain of an ABC transport system. sbmA-deleted E. coli strain shows an increased resistance to the antimicrobial peptide Bac7 and other proline-rich peptides due to their reduced uptake. In addition, BacA of some Gram negative species is essential to establish chronic infection in animal and plant hosts. The aim of this study was to characterize the architecture and function of the SbmA transport system in E. coli. By affinity chromatography and fluorescence anisotropy assays we showed that SbmA is able to bind Bac7 also without the requirement of a substrate-binding protein. By using a bacterial two-hybrid BACTH system and in vitro crosslinking assay we showed that SbmA forms dimers both in vivo and in vitro. Flow cytometry analysis using different metabolic inhibitors indicate that the proton motive force rather than ATP hydrolysis may represent the driving force for the translocation of peptide substrates. Searching for possible interactors of SbmA, we showed that the yaiW gene, located downstream of sbmA, is part of the same operon. Even if the two genes are cotranscribed, the corresponding proteins do not seem to interact neither at the cytoplasmic nor at the inner membrane level. However, ΔyaiW strain showed an increased resistance to Bac7 suggesting the involvement of this protein in the SbmA-mediated uptake of the peptide. An in vivo screening of the whole proteome of E. coli against SbmA by the BACTH system did not identify any nucleotide-binding domain of the transporter, strengthening the hypothesis that the transport system is energized by a mechanism different from ATP hydrolysis. Two-hybrid analysis allowed us to fish out 9 potential interactors and among those our attention was focused on FieF, an inner membrane protein involved in zinc efflux from the bacterial cell. We demonstrated that FieF is involved in the SbmA-mediated uptake of Bac7 and that the expression of sbmA is modulated by zinc ions in a complex manner, suggesting a stress role for this protein in bacteria. Overall, these findings suggest that SbmA is an unusual peptide transporter likely derived from the widespread ABC transporters, which has evolutionary lost/changed some typical features of this transport systems, such as the request of a substrate-binding protein and the ATP binding domain. The interaction between SbmA and FieF points to a possible involvement of the former in zinc detoxification, a function which may be modulated by the binding of an unidentified peptide substrate. The regulation of the expression of SbmA by zinc ions might have a physiological relevance in the host because it makes the target bacteria more susceptible to the AMPs and activates cellular effectors such as the neutrophils. XXIV Ciclo 1984

2010/2011 SUMMARY Introduction. Hepatocellular Carcinoma (HCC) ranks fifth in frequency of cancers in the world. Orthotopic Liver Transplantation (OLT) or liver resection represents the best treatments for HCC. However, most patients cannot be subjected to potential curative OLT or resection because of extensive tumor involvement of the liver, metastasis, invasion of the portal vein or advanced underlying hepatocellular disease at the time of diagnosis. Systemic chemotherapy or chemoembolization represent a good alternative for the treatment, however drug therapy of cancer in general is hampered by multidrug resistance (MDR) that is a phenomenon caused by the up-regulation of the ABC-transporters (ABC) leading to chemotherapy failure. To overcome these problems new therapeutic approaches, such gene therapy, are needed. Selective down-regulation of an essential and specific cancer gene such as telomerase (hTERT) could represent an emerging strategy that could prevent cancer progression and diminish numerous side effects derived from drug usage. The present study include two tasks whose aims are: Task 1: a) Assess if the extent of tumoral differentiation results in a different ABCB1, ABCC1 and ABCG2 expression. b) Assess whether the treatment with a chemotherapeutic drug(s) may affect the expression of the three ABC transporters under study. Task 2: To overcome the obstacle of MDR-induced chemoresistance using new therapeutic approaches such as gene therapy, silencing a cancer essential and specific gene. Results and discussion. Task 1: We assessed the ABCB1, ABCC1 and ABCG2 expression in three hepatic cell lines: IHH (non tumoral control), HuH7 (differentiated tumoral cells) and JHH6 (undifferentiated tumoral cells). Only ABCG2 expression correlates with the degree of tumoral differentiation. Through confocal microscopy analysis we observed that the Doxorubicin (Dox) is able to reach the cell’s nucleus within 10 min. After 24h and 48h Dox is completely concentrated into the nucleus where some nuclear damage occurs. The presence of damaged nuclei could explain the decreased mRNA in most of the ABCs under study. The treatment with Dox doses lower than the LC50 for 24h and 48h has different consequences for all the ABC considered in the three cell lines, with an mRNA expression pattern not in line with the protein one in most of the cases, suggesting that the possible mechanism that determines the ABCs protein upregulation in the tumoral cell lines (HuH7 and JHH6) is not the de-novo transcription but probably something related to the protein turnover. After the treatment ABCC1 protein expression increases in the tumoral cell lines but not in the non tumoral one (IHH). Regarding ABCB1 and ABCG2, these transporters seem to play a role only in HuH7 and JHH6 cells respectively. We were not able to correlate the tumorigenic potential of the two tumoral cell lines with the ABC expression since the different behaviour of ABCs and the different contribution to MDR. Thus in order to better clarify the contribution of each single ABC to MDR our future steps will consider the use specific inhibitors. Task 2: From our in vivo data, among four cancer related genes we selected hTERT as the best candidate for silencing experiments due to its exclusive expression in tumoral samples. A functional non-inflammatory siRNA targeting hTERT was designed: SirTel 1. Silencing experiments were conducted in JHH6 cell line. The hTERT silencing effect was dose dependent, at least at the three considered doses (25-50-100nM). For all the subsequent determinations the experimental concentration was 25nM. After 72h of silencing we observed a significant reduction in both hTERT mRNA expression and enzymatic activity (p

2010/2011 Wiskott-Aldrich syndrome (WAS) is an X-linked primary immunodeficiency characterized by recurrent infections, and a marked predisposition to develop autoimmune phenomena. The disease is caused by mutations in WASp, a key regulator of actin polymerization expressed only in hematopoietic cells. A general impairment of hematopoietic cell functions contributes to the pathogenesis of the disease. Neutrophils, B cells, T cells and DCs deficient for WASp were all shown to have impaired homing ability, a cellular function that strictly depends on spatio-temporal regulation of actin polymerization. WASp null T cells present severe impairment in coupling TCR stimulation to proliferation and fail to organize signaling molecules within the immunological synapse (Badour et al., 2004; Dupre et al., 2002; Sims et al., 2007). B cells intrinsic defects include altered B cell receptor clustering, defective homeostasis of mature B cells and a specific reduction in the expression of the complement receptor (Park et al., 2005; Simon et al., 1992; Westerberg et al., 2001). WASp null DCs fail to assemble podosomes and display late migration from the periphery to lymph nodes (de Noronha et al., 2005). Moreover, WASp expression in DCs is critical to organize the dynamic cytoskeletal changes that facilitate DC-T cell interaction during antigen presentation (Pulecio et al., 2008; Bouma et al., 2011). Together, these cellular alterations provided clues to understand the reduced response to pathogens and the immunodeficiency of WAS patients. However, the mechanisms by which perturbation of actin dynamics promote autoimmune phenomena are less clear. Autoimmune complications occur in 40-72% of children with severe WAS phenotype. The most common autoimmune features that develop in WAS patients include hemolytic anemia, vasculitis, renal disease, and arthritis (Dupuis-Girod et al., 2003; Sullivan et al., 1994; Humblet-baron et al., 2007). Impairment of T and B cell tolerance have been reported in WAS patients and in Was-/- mice, but the exact cellular mechanisms that link loss of WASp function to autoimmunity have not been fully elucidated yet (Becker-Herman et al., 2011; Recher et al., 2012; Marangoni et al., 2007; Maillard et al., 2007; Humblet-Baron et al., 2007). It is increasingly recognized that excessive activation of pDCs and elevated type-I interferon (IFN) levels are pathogenic in several human autoimmune diseases such as SLE, psoriasis, Sjogren’s syndrome. Since WAS autoimmune manifestations partially overlap with those of type-I IFN diseases, we hypothesized that pDCs/IFN-α axis may have a role in WAS-associated autoimmunity. In the first part of results we present the analysis of the pDC compartment in a mouse model of the disease, a mouse knock-out for WASp (WKO mice). We show that pDCs from WKO animals are chronically activated, secrete type-I IFN constitutively and become refractory to further stimulation in vivo. By depleting WASp expression or by interfering with actin dynamics in pDCs we prove that WASp-mediated actin dynamics control the activation of the TLR9/IFN-α pathway in a cell autonomous fashion. Based on the results of previous section, we speculated that WASp may regulate IFN-α production by controlling the correct organization of the endocytic pathway. It is well known, in fact, that type-I IFN production relies on spatio-temporal regulation of TLR9 signaling in the endocytic pathway, at the same time, it is largely established that many steps of the endocytosis are regulated by actin regulatory proteins of the WASp family. Therefore, we proceed presenting a series of experiments exploring the role of WASp in mediating signaling downstream TLR9 activation in pDCs. Tracking of TLR9 agonist in pDCs show that WASp controls cellular architecture and early endosomes size leading to accumulation of large aggregates of TLR9 agonist in WASp-deficient cells. To demonstrate a link between the alteration in the pDCs/IFN-α axis found in WKO mice and the pathogenesis of WAS autoimmunity, we first analyzed the presence in patients of clinical aspects known to be associated with type-I IFN diseases. We report that WAS patients display a moderate type-I interferon signature, indicating an exposure to elevated levels of this cytokine. Finally, we moved back to the mouse model of WAS to start investigating the presence of alterations of innate/adaptive immunity classically attributed to unrestrained type-I IFN production. In line with the well known inducing effects of systemic type-I IFNs, our preliminary data illustrate a generally more activated phenotype of immunocytes in WKO mice. In sum, our work provide 1) the first demonstration of an altered pDCs/IFN-α axis in WAS, 2) the existence of a cell intrinsic mechanism of increased pDCs activation in WASp-null pDC, and 3) a new role for actin in restraining excessive activation of TLR9 in pDCs. These observations add a new layer of complexity to our understanding of the pathophysiology of WAS, and raise important considerations about the treatment of patients with autoimmune phenomena. XXIV Ciclo 1983

2010/2011 The tumor suppressor p53 plays a central role in the protection against DNA damage and other forms of stress, primarily by inducing cell cycle arrest or apoptosis. Mutation of p53, which is one of the most frequent genetic alterations detected in human cancers, inactivates these growth regulatory functions; in addition, very often mutant p53 acquires tumor-promoting activities (gain-of-function). A complete and thorough understanding of the signaling circuitry that regulates wild-type and mutant p53 functions is therefore a primary objective for basic cancer research, since it may lead to development of important tools for diagnosis and therapy of tumors. One crucial component of such knowledge is the protein interaction profile of p53. To gain novel insights on p53 interactions, we used a phylogenetic approach. We reasoned that a comprehensive map of the protein interaction profile of Drosophila p53 might reveal conserved interactions of the entire p53 family in man. Using a genome-scale in vitro expression cloning approach, we identified 91 previously unreported Dmp53 interactors. Next, we studied and characterized the interaction of human orthologs of newly identified Dmp53 interactors with all p53-family proteins, and it resulted in the identification of several novel interactants of this family of tumor suppressors (Part 1 of this Thesis). In parallel, we verified that many of the mammalian orthologs of Dmp53 interactors could also bind to an oncogenic p53 mutant (R175H), and therefore are potential novel targets or effectors of mut-53 gain of function activities. Among those proteins we focused our attention on DAB2IP, a tumor-suppressor gene that functions by counteracting the activation of multiple oncogenic pathways. There are evidences that mutant p53 has a stimulatory role in all the signaling pathways that are normally inhibited by DAB2IP; therefore, we propose that mutant p53 may bind and functionally inactivate DAB2IP as one of the mechanisms of its gain-of-function. Given the crucial role of DAB2IP in modulating cellular responses to TNF, we focus on the potential relevance of this interaction in the connection between inflammation and cancer (Part 2 of this Thesis). XXIII Ciclo 1983

2010/2011 The calpains are a family of intracellular cysteine proteases, among which the best studied isoforms, micro- (CAPN1) and milli-calpain (CAPN2), are heterodimers consisting of a catalytic subunit and a common regulatory subunit, CAPNS1, required for function. Calpain is involved in many processes important for cancer biology, such as autophagy, indeed in calpain-depleted cells autophagy is impaired, with a subsequent increase in apoptosis sensitivity. Calpain is also important in all the stages of the stress response. A proteomic approach was employed for the identification of novel CAPNS1 interacting proteins. Proteins immunoprecipitating with endogenous CAPNS1 in HT1080 cell lysates were analyzed by Mass Spectrometry. We identified novel partners among which the deubiquitinating enzyme USP1, a key regulator of the DNA damage response and genome integrity maintenance via its specific action on FANCD2, involved in DNA repair and protection from chromosome instability, and PCNA, involved in the regulation of translesion DNA synthesis (TLS), that bypasses DNA lesions with low stringency basepairing requirements. We performed co-IP assays in lysates of 293T cells and confirmed that the interaction was specific. Furhermore, we observed that calpain is able to bind a USP1 C-terminal deleted mutant, suggesting that USP1 first 523 aminoacids were sufficient for the binding. To understand what is the effect exerted by calpain upon USP1, we depleted calpain activity in a series of cell lines, and followed the fate of endogenous USP1. We transfected CAPNS1 specific siRNAs, or treated cells with a specific inhibitor of calpain, and we observed a strong decrease in USP1 protein levels. This effect should be at a post-transcriptional level, since any significant change in USP1 mRNA levels is detected. We also obtained the same result by transfecting a siRNA specific for CAPN1, the gene encoding for the catalytic subunit micro-calpain. Moreover, we studied the role of calpain in the PCNA-mediated switch between high fidelity replication and TLS upon UV irradiation. In mouse embryonic fibroblasts knockout for CAPNS1, USP1 downregulation is coupled to an increase in PCNA monoubiquitination. Moreover, CAPNS1-depleted U2OS cells showed an increase in the percentage of nuclei containing PCNA-induced foci upon UV irradiation. Since we demonstrated that calpain can modulate an important regulator of DNA damage response such as USP1, we investigated if calpain could have a role in genome integrity maintenance. CAPNS1 depleted cells showed a reduced rescue in DNA repair compared to control cells, suggesting that increased levels in PCNA monoubiquitination could lead to an increased amount of errore-prone TLS. Calpain plays an important role in autophagy, so we asked if USP1 degradation in absence of calpain activity could involve autophagic pathways. We first blocked macroautophagy by silencing ATG5, and we observed that USP1 was downregulated, suggesting that the depletion of ATG5 could lead to an increased activity of other degradation pathways. To impaire chaperone-mediated autophagy (CMA), we silenced a protein important for autophagosome formation, LAMP-2A. Also in this case we observed a decrease in USP1 protein levels, thus suggesting that USP1 is alternatively degraded by different pathways. However, we observed that USP1 is stabilized upon inhibition of lysosomal enzymes, suggesting that USP1 may be degraded in the lysosome. To better understand the mechanism by which calpain affect USP1 stability we search for an effect of calpain upon USP1 co-factor and activator UAF1/WDR48. CAPNS1-depleted cells showed WDR48 downregulation, but WDR48 overexpression only partially rescue USP1 protein levels in this cells. Furthermore, we provided evidences that calpain regulation of p35/p25 activator of Cdk5 can affect Cdh1 phosphorylation and thus APC/Cdh1 activity, leading to a regulation of USP1 stabilization. In conclusion, we identified USP1 as a novel interactor of calpain, and we found that calpain is important for USP1 stability, since in its absence USP1 is downregulated. The importance of this novel regulation is strengthened by the recent findings that unveiled a role of USP1 in maintenance of a mesenchymal stem cell program in osteosarcoma, and thus placing calpain in a crucial regulatory position for cancer development. XXIV Ciclo 1983

2010/2011 Unconjugated Bilirubin (UCB) is the final product of the heme catabolism. The high serum UCB concentrations in the first days of life of the newborns, due to immature mechanisms for hepatic uptake, conjugation and biliary secretion, is called physiological neonatal jaundice. This common condition is generally a benign and transient phenomenon, but in some cases the hyperbilirubinemia can progress to bilirubin encephalopaties ranging from minimally neurological injury to severe and permanent neurodevelopmental dysfunction. In the present thesis the SH-SY5Y neuroblastoma cell line was used to approach the molecular events associated to bilirubin neurotoxicity and highlight the biochemical and molecular events that are induces in the neurons when get contact with the UCB. Depending on the bilirubin concentration and the time of exposure to UCB, we were able to define experimental setups for the study of bilirubin resistance and bilirubin toxicity. Using the model to study bilirubin resistance, it was demonstrated that the resistance is not entirely achieved by limiting the entrance or increasing extrusion of the pigment from the cell, but rather by enhancing the cellular defensive mechanisms, in particular against the oxidative stress. This was achieved by increasing the intracellular glutathione content via the specific induction of the genes and activity of the System Xc-. Furthermore, the cells exposed to bilirubin over-expressed several additional genes that encode for important antioxidant and detoxifying proteins like Heme Oxygenase-1 and NAD(P)H:quinone oxidoreductase 1. As far as the mechanisms of bilirubin neurotoxicity, we showed that UCB exposure lead to the induction of the intracellular ROS accumulation. Moreover, the data presented report evidences that the bilirubin toxicity could be displayed by a mechanism of excitotoxicity carried out by the cellular release of glutamate. Further studies will be necessary to elucidate the molecular mechanisms by which bilirubin produces neurotoxicity and to understand how the cells avoid the damage. The information presented here could contribute to the identification of targets to avoid the bilirubin damage. La bilirubina non coniugata (UCB) è il prodotto finale del catabolismo dell’eme. L’ alta concentrazione di bilirubina nei primi giorni di vita del neonato è conosciuto come ittero neonatale. Questa condizione è generalmente benigna e transitoria, ma in alcuni casi l’iperbilirubinemia può portare a encefalopatie che vanno da minimi danni neurologici a severe e permanenti disfunzioni neuronali. Nella presente tesi la linea cellulare di neuroblastoma umano SH-SY5Y è stata utilizzata per identificare gli eventi molecolari associati alla neurotossicità della bilirubina ed evidenziare gli eventi biochimici e molecolari che sono indotti nelle cellule neuronali quando queste sono esposte alla UCB. A seconda della concentrazione e del tempo di esposizione all’ UCB che viene utilizzato, si sono definiti due modelli sperimentali per studiare sia la resistenza delle cellule alla bilirubina che la sua tossicità. Utilizzando il modello per studiare la resistenza alla bilirubina, è stato dimostrato che la resistenza non è completamente dovuta alla limitazione dell’ entrata o alla rapida eliminazione del pigmento dalla cellula, ma da un aumento dei meccanismi di difesa cellulare, in particolare quelli contro lo stress ossidativo. Si è dimostrato che questo’ultimo è dovuto all’aumento intracellulare di glutatione attraverso l’induzione specifica di alcuni geni e ad un aumento dell’attività del Sistema Xc-. Inoltre, le cellule esposte a bilirubina overesprimono vari geni addizionali che codificano per importanti proteine antiossidanti e detossificanti come l’emossigenasi 1 e la NAD(P)H quinone ossidoreduttasi. Per quanto riguarda ai meccanismi di neurotossicità della bilirubina, è stato dimostrato che l’ esposizione a UCB induce l’accumulo intracellulare di ROS. Inoltre, i dati presentati evidenziano che la tossicità della bilirubina può essere dovuta a meccanismi di esototossicità prodotti dalla liberazione cellulare di glutammato. Ulteriori studi servirebbero per completare l’ identificazione dei meccanismi molecolari cui attraverso i quali la bilirubina produce il danno neuronale e capire come le cellule riescono a evitare il danno. L’informazione presentata in questo lavoro potrebbe contribuire all’identificazione di possibili targets che potrebbero prevenire il danno indotto da bilirubina. XXIV Ciclo 1982

2009/2010 Il gusto dell'amaro è considerato la qualità gustativa più complessa nell'uomo. La percezione del gusto amaro nell'uomo è mediata dalla sottofamiglia hTAS2R dei recettori di membrana accoppiati alle proteine G (GPCR). Nonostante la conoscenza della struttura tridimensionale di questi recettori sia un passo cruciale verso la loro comprensione anche da un punto di vista funzionale, le informazioni strutturali di questi recettori sono ancora limitate. Quindi, le conoscenze delle basi molecolari del gusto amaro sono anch'esse scarse. Una miglior conoscenza delle relazioni tra struttura e funzione dei recettori hTAS2R porterebbe importanti conseguenze per l'industria alimentare, per la concezione di nuove molecole farmacologiche e per la biologia strutturale in generale. Il recettore hTAS2R38 è il più studiato recettore dell'amaro e lega il feniltiocarbamide (PTC) e composti correlati. Sono conosciute differenze, su base individuale, nella percezione del PTC e sono stati effettuati diversi studi con lo scopo di indagare eventuali correlazioni tra queste differenze e le preferenze alimentari. I risultati sono però tutt'ora contrastanti. Finora, sono state identificate tre posizioni polimorfiche che spiegano tali differenze nella percezione del PTC ma le caratteristiche funzionali e strutturali che stanno dietro allo spettro ricettivo e al meccanismo di attivazione del recettore sono ancora sconosciute. Nel presente progetto, utilizzando metodiche di bioinformatica strutturale, docking molecolare e validazione sperimentale, sono stati identificati residui potenzialmente coinvolti nel legame del PTC e nel meccanismo di attivazione del recettore. Le predizioni computazionali sono validate da esperimenti di mutagenesi sito-diretta mirata su residui specifici localizzati nel sito di legame putativo e nelle eliche transmembrana 6 e 7, probabilmente coinvolte nell'attivazione del recettore. Considerati i risultati ottenuti, viene suggerito che (i) i residui N103, F197 e S260 partecipano attivamente al legame con il PTC; (ii) i residui W99, M100 e S259 contribuiscono a definire l'ampiezza e la forma ottimali della cavità di legame; (iii) W99 e M100 in misura minore, e F255 e V296 più solidamente, giocano un ruolo chiave nell'attivazione del recettore, fornendo degli indizi mai suggeriti finora sull'attivazione di questi recettori. Questi risultati, assieme ai dati precedentemente forniti da altri studi, prevalentemente computazionali, suggeriscono che la struttura globale dei GPCR e diverse posizioni aminoacidiche fondamentali per il legame siano più conservate di quello che ci si aspetterebbe sulla base della bassa identità di sequenza. Inoltre, identificando dei residui probabilmente coinvolti nel legame del ligando, i risultati ottenuti forniscono importanti suggerimenti per la progettazione di antagonisti recettoriali in grado di diminuire o eliminare il gusto amaro da cibi, bevande e farmaci. XXIII Ciclo 1982

2009/2010 Intrinsically disordered proteins (IDPs) are flexible molecules, able to adapt to the surfaces of different molecular partners by means of specific, but easily reversible interactions. IDPs carry out pivotal biological functions participating in almost all cell signaling and regulatory pathways. Importantly, IDPs activities are finely modulated by the addition/removal of numerous post-translational modifications (PTMs) which are important conformational modulators. Prototypes of IDPs are High Mobility Group A (HMGA) proteins, which are expressed at high levels and play essential functions both in embryonic and cancer cells. HMGA protein family (HMGA1a, HMGA1b, and HMGA2) belong to the non-histone HMG chromatin protein super-family and are multifunctional architectural transcription factors. HMGA conformational adaptability and intricate pattern of dynamic and constitutive PTMs are thought to be responsible for this multifunctionality. We performed a liquid chromatography-mass spectrometry (LC-MS) screening in twenty different cell lines in order to evaluate HMGA proteins PTM pattern and we evidenced relevant intra-family differences. Moreover, we focused on the poorly characterized HMGA2 and we mapped HMGA2 phosphorylation sites by mass sequencing demonstrating that, similarly to HMGA1, it is phosphorylated on the acidic C-terminal tail by CK2. Importantly, this modification turned out to affect HMGA2 DNA binding. Since truncated HMGA proteins are more oncogenic than full-length ones and since HMGA are in vivo heavily modified on their C-terminal domain, we dissected the role of this domain and its phosphorylations from a structural point of view. We probed HMGA IDPs compactness and accessibility taking advantage of an innovative approach combining limited proteolysis and MS-based techniques. By limited proteolysis, ESI (electrospray ionization)-MS, and IMS (ion mobility separation)-MS we demonstrated that HMGA can assume a compact conformation and that their compactness degree is dependent upon the presence of the acidic C-terminal domain and its constitutive phosphorylations. Moreover, LC-MS analyses after enzymatic assays showed that HMGA forms with a deletion of acidic C-terminal tail are more susceptible to PTMs, thus supporting the idea that the acidic tail is involved in masking the accessibility of modifying enzymes to their own consensus sites. We evidenced macroscopic differences regarding PTMs affecting the three HMGA family members and provided the first data about in vivo HMGA2 PTMs and their effect on DNA binding. Our structural investigations revealed a structure/PTMs relationship dictated by the presence of the C-terminal domain. This evidence, together with the already known in vivo functional outcome of HMGA C-terminal truncation, suggests a structure/function link between HMGA tails, their PTMs, and their oncogenic properties, paving the way for the development of interfering therapeutic strategies based on targeting HMGA proteins. Le proteine intrinsecamente disordinate (PID) sono molecole flessibili in grado di adattarsi alle superfici di differenti partner molecolari tramite interazioni specifiche, ma facilmente reversibili. Queste proteine compiono funzioni biologiche essenziali partecipando in quasi tutte le vie di segnalazione e regolazione cellulari. La modulazione delle loro attività avviene tramite l’aggiunta/rimozione di numerose modificazioni post-traduzionali (MPT), le quali possono comportare anche cambiamenti conformazionali. Le proteine High Mobility Group A (HMGA) sono considerate prototipo delle PID. Esse sono espresse ad alti livelli e giocano funzioni chiave sia durante lo sviluppo embrionale che durante il processo di trasformazione neoplastica. Le proteine della famiglia HMGA (HMGA1a, HMGA1b e HMGA2) appartengono alla super-famiglia delle proteine cromatiniche non istoniche HMG e sono fattori trascrizionali architetturali multifunzionali. L’adattabilità conformazionale e l’intricato pattern di MPT dinamiche e costitutive delle HMGA sono ritenute essere alla base della loro multifunzionalità. In questo lavoro di tesi è stato effettuato uno screening delle MPT a cui sono soggette le HMGA in venti linee cellulari diverse mediante analisi di cromatografia liquida accoppiata alla spettrometria di massa (LC-MS) e si sono evidenziate rilevanti differenze tra i tre diversi membri della famiglia HMGA. Inoltre, concentrandosi sulla poco caratterizzata proteina HMGA2, abbiamo dimostrato mediante mass sequencing che, al pari della proteina HMGA1a, essa è fosforilata sul C-terminale e che tale modificazione comporta un’alterazione delle sue proprietà di legame al DNA. Diversi tipi di tumore esprimono forme tronche delle proteine HMGA mancanti del dominio acidico C-terminale. Dal momento che queste forme sono state dimostrate essere maggiormente oncogeniche rispetto alle controparti a lunghezza completa abbiamo deciso di investigare il ruolo della coda acidica e delle sue fosforilazioni da un punto di vista strutturale. La compattezza e l’accessibilità delle HMGA è stata esplorata sfruttando un approccio innovativo che combina esperimenti di proteolisi limitata e tecniche di spettrometria di massa (MS). Tramite proteolisi limitata, ESI (electrospray ionization)-MS, and IMS (ion mobility separation)-MS abbiamo dimostrato che le HMGA possono assumere una conformazione compatta e che il loro grado di compattezza è dipendente dalla presenza del dominio acidico C-terminale e delle sue fosforilazioni costitutive. Inoltre, analisi LC-MS effettuate dopo saggi enzimatici hanno mostrato come le forme delle HMGA presentanti una delezione della coda acidica C-terminale sono più suscettibili alle MPT, supportando quindi l’idea che la coda acidica sia coinvolta nel mascherare l’accessibilità degli enzimi di modificazione sui siti consensus lungo la sequenza di HMGA. In questo lavoro abbiamo evidenziato differenze macroscopiche riguardanti le MPT che interessano i tre membri della famiglia HMGA e abbiamo fornito i primi dati circa le MPT di HMGA2 in vivo e il loro effetto sul legame al DNA. Le nostre investigazioni strutturali hanno rivelato una relazione struttura/MPT dettata dalla presenza del dominio C-terminale. Questa evidenza, insieme al già noto esito funzionale in vivo della delezione del C-terminale di HMGA, suggerisce un collegamento struttura/funzione tra la coda acidica delle HMGA, le loro MPT, e le loro proprietà oncogeniche. XXIII Ciclo 1983

2009/2010 In infants and in the Crigler-Najjar syndrome type I patients, severe hyperbilirubinemia due to high levels of unconjugated bilirubin (UCB) may cause Kernicterus, leading to an irreversible and selective brain damage. The Gunn rat is the animal model for the study of these pathologies. It has been suggested that different enzymes of the phase I (cytochrome P-450-dependent mixed function oxygenases 1A1, 1A2, 2A3), phase II (glutathione-S-transferases α2, α3, µ3, µ4, π) and phase III transporters (particularly, Mrp1) seems to be involved in UCB detoxification pathways. However, to date, their in vivo brain expression has been evidenced only at the blood-brain interfaces, while remains largely unexplored in brain parenchyma. Particularly for Mrp1, in vitro evidence reported a role of this transporter in protection of neural primary cultures from dissected cortex, by extruding bilirubin out of the cell. The aim of this study is establish the developmental profile of these genes in brain parenchyma, and assess their alteration in hyperbilirubinemic jj animals. Due to the high regional selectivity of UCB-induced neurotoxicity, cerebellum (Cll), striatum (St), hippocampus (Hip) and cerebral cortex (Cx) were chosen for this study. Our results regard the Mrp1 protein in cerebral cortex of normobilirubinemic (JJ) rats showed that its expression varied during the post-natal age, reaching the highest levels at 9 days after birth. No changes were found between JJ and Jj (having a temporary hyperbilirubinemia in the first week of life) rat for all ages analyzed. Similarly, no differences were detected among JJ/Jj and jj (hyperbilirubinemic) rats at P2, P17 and P60, while a significant increase (p < 0.005) was evidenced in P9 jj rats as compared to age-matched JJ animals. Our Mrp1 mRNA analysis in four regions of P9 animals by Real Time-qPCR revealed the absence of differences among Cx, Cll, St and Hip of P9 normobilirubinemic JJ rats. Moreover, no variations between jj and JJ control animals were detected. Regarding the Mrp1 protein expression in the same four regions by Western blot analysis, our results showed that the levels of this transporter in normobilirubinemic JJ rats were lower in Cx, similar in Cll, St and Hip (p < 0.05 vs Cx). Comparing genotypes, a reduction on Mrp1 in jj animals (compared to Mrp1 amount in the same region of JJ pups) was detected in Cll, St, but reached the statistical significance only in Hip (p < 0.05 vs Hip JJ). The analysis of CYPs gene expression in P9 Gunn rats indicate that CYP1A1, 1A2 and 2A3 mRNA were differently expressed among Cx, Cll, St and Hip of JJ rat. Similarly a region-specific modulation of CYPs expression in jj Gunn rats (compared to JJ) was pointed-out. Surprisingly, UCB seems to generate a plateau effect on CYPs mRNA levels among brain regions of jj rats. In P60 JJ Gunn rats the CYPs expression is higher than in P9 animals, with the following pattern among regions: Cx  CllSt  Hip. A down-regulation (p < 0.05) in St of P60 jj compared to normal animals was observed. Analyzing the GSTs expression in P9 animals, higher variability in the GSTs expression among the four brain areas was evidenced. In hyperbilirubinemic (jj) rats (compared to JJ), statistically relevant down-regulations were detected for GSTα2 (in St;p < 0.05), GSTα3 (in Hip;p < 0.05), µ3 (in Cx;p < 0.01), µ4 (in Cx; p < 0.05) and π (in Cll: p < 0.05); while GSTµ4 was up-regulated in St (p < 0.05). From P9 to P60, in JJ animals: GSTα3 expression increased (13-75-fold depending on the region); while GSTα2 (5-fold), µ3 (p < 0.05), µ4 (2-fold) and π (2-fold) mRNA amounts decreased. In P60 jj Gunn rats, compared to controls (JJ):a relevant up-regulation of GSTα3 was observed in Cll (p < 0.005) and Hip (p < 0.05), while GSTµ3 in jj was down-regulated (p < 0.05). The Mrp1 results obtained in the present in vivo study seems not to be in agreement with the in vitro data reported, to date. Thus, the Mrp1 expression is low in brain parenchyma and bilirubin affect (up-regulation) only marginally the protein amounts in cortex of P9 animals, while in other regions Mrp1 is not modulated, indicating a marginal role in vivo in bilirubin clearance. Similarly, while in liver GSTα2 and α3 act together as ligandin, this seems not happens in brain where the two subunit are expressed at very low levels (P9: α2 77000- α3 1500 fold difference; P60: α2 112000-α3 2200 fold difference with respect to age matched livers). For all genes under analysis, a very complex and variable pattern of expression among brain areas was evidenced. Consequently, no general rules concerning bilirubin-induced modulation could be drawn, as both up and down-regulation were observed. Additionally, in Cx of P9 jj animals, a translational control of Mrp1 might be hypothesized due to a significant increase in Mrp1 protein, without changes in mRNA level. Therefore, the genomic screening made in this work provides the first general overview on the mRNA developmental profiles of several CYPs and GSTs genes in brain parenchyma (specifically Cx, Cll, St and Hip) of normal rats, and of animals suffering from hyperbilirubinemia, underlying the necessity to find functional evidence to finally understand the role of these enzymes associated with the kernicterus and Crigler-Najjar type I syndrome pathologies. XXIII Ciclo 1978

2009/2010 To reach new and relevant insights in biomolecular sciences, new tools for fine and precise measurement are needed. Nowadays advances in the field of micro-electro-mechanical systems (MEMS) offer unique opportunities in the design of ultrasensitive analytical devices to support the molecular sensing investigations. Among them Microcantilever (MC) biosensors are label-free platforms that combine a biologically sensitive with a physical transducer in order to selectively and quantitatively detect the presence of specific compounds in a given external environment. Since they can be operated either as nanomechanical resonator or as surface stress sensor, MCs - activated with DNA probes or antibodies for molecular recognition - enable the measurement of mass with extraordinary sensitivity. In particular, the development of mass detector biosensor based on MC systems would permit to shift from qualitative data to quantitative measurements of key molecules involved in physiological processes. This can lead crucial informations to characterize complex mechanisms such as angiogenesis and tumor progression and to the quantification of small amounts of cancer markers, such as Angiopoietin-1 (ANG-1), and their modulation during the early stages of tumor development. XXIII Ciclo 1979

2009/2010 Introduction. The metabolic syndrome is a cluster of alterations, including insulin resistance, dyslipidaemia, hypertension, hyperglycaemia, abdominal obesity, hyperhomocysteinemia, inflammation and oxidative stress, leading to type II diabetes and cardiovascular disease. The metabolic syndrome is usually associated with sedentary lifestyle and overweight, while regular physical activity and weight loss can counteract these alterations and prevent type II diabetes and cardiovascular disease. Aim of the Thesis. In order to define the net role of inactivity as key factor inducing insulin resistance and metabolic syndrome independently from changes in body fat we have investigated the net impact of experimental bed rest on human metabolism. Experimental bed rest in healthy, young, lean subjects represents a suitable model to determine the effects of inactivity on physiology, avoiding potential interferences and confounding effects of diseases, ageing, energy unbalance and excess body fat. We have focused on inactivity-related development of insulin resistance, dyslipidaemia, hypertension as well as inflammation and oxidative stress. These aspects have been investigated during four different experimental bed rest protocols, lasting 2 months (WISE-Toulouse, France) and 5 weeks (Valdoltra, Slovenia 2006–2007–2008). Energy requirements and intakes were strictly controlled to avoid changes in fat mass. Results and discussion. Muscle atrophy. Muscle atrophy was evidenced after three weeks of bed rest and was worsened by prolonged exposure to inactivity (WISE, Valdoltra studies). However, muscle loss rate was higher in the first 5 weeks of bed rest while it decreased in the second month of inactivity (WISE). Time-course analysis of insulin resistance development. Insulin resistance, measured by an oral glucose tolerance test, rapidly developed in the first week of inactivity and was maintained after 5 weeks of bed rest, as assessed by the ISI-Belfiore index of insulin sensitivity (Valdoltra 2008). Cardiovascular regulation. In the first week of bed rest, baroreflex sensitivity decreased indicating that, in an early phase, alterations in the sympatovagal balance paralleled changes in insulin resistance development. At the end of 5 week-bed rest, heart rate and heart rate variability as well as systolic blood pressure variability, indexes of cardiovascular regulation, were also impaired (Valdoltra 2008). Plasma lipids and lipid metabolism. Five weeks of bed rest induced a decrease in high-density lipoprotein (HDL) cholesterol. During inactivity, cholesteryl ester transfer protein (CETP), a key enzyme involved in HDL metabolism, was up-regulated and changes in CETP inversely correlated with changes in HDL-to-non-HDL cholesterol ratio. Conversely, changes in CETP and HDL were not directly correlated to insulin resistance (Valdoltra studies). Cell membrane lipids. Bed rest reduced monounsaturated FAs, enhanced n-6 polyunsaturated FA total contents and affected activities of both Δ-5 and Δ-9 desaturases, enzymes involved in FA metabolism. These data further support that membrane FA composition and activities of Δ-5 and Δ-9 desaturases are predictive indicators of metabolic syndrome development. Moreover, arachidonic-to-eicosapentaenoic acid ratio, reflecting the competitive role of these FAs in the modulation of inflammatory processes, was shifted towards pro-inflammatory state (Valdoltra studies). Oxidative stress and glutathione kinetics. Bed rest induced oxidative stress as showed by enhanced muscle protein carbonylation, a marker of tissue exposure to oxidative damage, and increased muscle glutathione absolute synthesis, as assessed by a new one-sample, double-isotope tracers infusion method (Valdoltra 2007). Homocysteinemia and homocysteine kinetics. Plasma homocysteine level was increased by bed rest, due to a decrease in homocysteine clearance related to remethylation (WISE). Hyperhomocysteinemia is a further evidence of inactivity-mediated oxidative stress and increased cardiovascular risk. Conclusions. Physical inactivity in healthy young subjects is a suitable model to define the net impact of physical inactivity on the development of metabolic alterations observed in patients with the metabolic syndrome. Our results indicate that inactivity is directly involved in insulin resistance development, low-grade systemic inflammation, dyslipidaemia, hyperhomocysteinemia, oxidative stress and autonomic-cardiovascular abnormalities. XXIII Ciclo 1983

2009/2010 With the completion of the sequencing and annotation of hundreds of genomes, and the accumulation of data on the mammalian transcriptome, greater emphasis has been placed on elucidating the function of non-coding DNA and RNA sequences. It is well known that the non-coding portion of the genome can transcribe functional RNAs. Several categories of non-coding RNAs (ncRNAs) have been defined, such as transport RNAs (tRNAs) ribosomal RNAs (rRNAs), small nuclear RNAs (snRNAs) and small nucleolar RNAs (snoRNAs). A larger group of ncRNAs comprises the so-called microRNAs (miRNAs) and long non-coding RNAs serving key regulatory roles. It has been shown that miRNAs directly target a large number of genes, thus affecting significantly major pathways. In my project, I focused on miR-204, a microRNA that is highly conserved from zebrafish to human and located in the sixth intron of the human TRPM3 gene. I sought to identify mir-204 targets by using the Medaka fish (Oryzias latipes), where mir-204 is expressed at very low levels in the nervous system, as a model for perturbation of the mir-204 network. Transient transgenic Medaka fish were produced to knock down and over-express mir-204. Next-generation sequencing was used to sequence the Medaka transcriptome, dissect the putative targets of miR-204, and thus gain further insight about its function. Potential target genes of mir-204 were selected by choosing genes, which presented lower expression in the wild-type (wt) fish than in the knock down, a lower expression in the over-expression than in the wt and, finally, a higher expression in the knock down than in the over-expression. At the same time, I collected a list of putative miR-204 mouse and human targets using the prediction softwares miRanda, PicTar and TargetScan, obtained the Medaka orthologues and verified that the selected genes in Medaka had a statistically significant enrichment in miR-204 targets as compared to the complete set of genes obtained from the RNA-Sequencing approach. The combined RNA-Sequencing and bioinformatics analysis revealed 147 predicted targets of mir-204, which showed a significant enrichment for the axon guidance pathway. In order to confirm this data, real time quantitative PCR has been performed on total RNA from wt and morphant fish. Results showed a higher expression in the knock down fish for 15 out of 25 putative targets (Neo1, Trim71, Ddx3y, Prkar1a, MyoX, Sema3B, Sema3F, Ptprg, Slit2, Epha4, Epha7, Amot, Lpp, Odz4, Jarid2). I further validated these genes by both Q-PCR and luciferase assays. To this aim, I cloned five putative target sequences into the 3’UTR of a luciferase reporter vector (pGL3-TK-luc Promega) to use them in luciferase assays: co-transfection with miR-204 reduced the luciferase activity of Sema3F, belonging to the class of receptors involved upstream of the axon guidance pathway. These results indicate that mir-204 directly targets key genes involved in the axon guidance pathway such as Sema3F in the nervous system. Further validation of the disruption of axon guidance in the transgenic fish has been undertaken in vivo by our collaborators: the experiment demonstrated a clear role of this microRNA in axon path finding during retinal development. XXII Ciclo 1981

2009/2010 Mutations in the TP53 gene are among the most frequent genetic alterations in human cancers. As a consequence of these mutations p53 loses its tumour suppressor functions and may acquire novel oncogenic activities (gain of function) sustaining tumour formation and progression. Many in vivo studies highlighted that mutant p53 gain of function is associated with elevated protein levels, supporting the notion that in tumour cells altered signalling could stabilize and activate mutant p53, with mechanisms similar to those required to stimulate wild-type p53. The aim of my PhD work was to investigate the mechanisms underlying mutant p53 gain of function, focusing on factors that might link cancer-related signalling with mutant p53 activity. An intriguing candidate for this role is the phosphorylationdependent prolyl isomerase Pin1, that transduces phosphorylation signalling into conformational changes affecting the functions of its substrates, as ours and other laboratories have reported for wild-type p53. Despite Pin1 supports wild-type p53 functions, Pin1 is frequently overexpressed in human tumours and has been shown to promote both Her2/Neu/Ras and Notch1 dependent transformation. So we reasoned that the physiological role of Pin1 as a component of checkpoint mechanisms might be subverted during tumourigenesis, thereby turning it into an essential partner of mutant p53 and a critical amplifier of its oncogenic functions. Indeed, we now demonstrate that Pin1 enhances tumourigenesis in a Li-Fraumeni mouse model and cooperates with mutant p53 in Ras-dependent cell transformation. In human breast cancer cells, Pin1 promotes both mutant p53 dependent inhibition of the anti-metastatic factor p63 and the induction of a mutant p53 transcriptional program to increase tumor aggressiveness. Accordingly, we have identified a transcriptional signature (the Pin1/mutant p53 signature) that is associated with poor prognosis in breast cancer and, in a cohort of patients, Pin1 over-expression influences the prognostic value of p53 mutation. Considering that TP53 mutation is more frequent in tumors with higher risk of recurrence such as triple-negative cases and that some of the Pin1/mutant p53 signature genes are over-expressed in triple negative breast cancers, our findings carry therapeutic implications for this kind of cancers and possibly also for other tumours bearing mutant p53 and high levels of Pin1. XXII Ciclo 1983

2008/2009 Argomento di questa tesi di dottorato sono i polisaccaridi (sia capsulari, CPSs, sia esopolisaccaridi, EPSs), prodotti da batteri e rilasciati nell’ambiente circostante, ed il loro ruolo nelle infezioni batteriche. Sono stati presi in considerazione batteri opportunisti responsabili di infezioni polmonari in pazienti affetti da fibrosi cistica (CF). I risultati riportati in questa tesi di dottorato comprendono diverse linee di ricerca e contribuiscono a chiarire le diverse caratteristiche chimiche e biologiche dei polisaccaridi batterici presi in considerazione. La prima linea di ricerca riguarda due EPSs strutturalmente simili prodotti da Inquilinus limosus, un batterio recentemente isolato dalle secrezioni polmonari di pazienti con CF. Si è voluto valutare se i due EPSs fossero in grado di assumere conformazioni differenti e quindi ipoteticamente producessero diverse attività biologiche. Le proprietà conformazionali sono state valutate mediante misure di dicroismo circolare ed immagini ottenute mediante microscopia a forza atomica. Studi di modellistica molecolare sono stati effettuati dalla Dr.ssa M. Kuttel (Dip. di Chimica, Univ. di Cape Town). I due polimeri esibiscono effettivamente differenti conformazioni elicoidali (strutture secondarie), che vengono stabilizzate da legami idrogeno e caratterizzate dalla presenza di gruppi piruvato carichi esposti sulla superficie esterna dell’elica. Questo risultato conferma l’ipotesi che la biosintesi di più specie polisaccharidiche può risultare un mezzo utile per il batterio per diversificare i propri strumenti di difesa da possibili minacce esterne. La seconda linea di ricerca ha preso in considerazione l’esopolisaccaride cepaciano, prodotto dalla maggioranza dei ceppi appartenenti al complesso della Burkholderia cepacia (Bcc). L’unità ripetitiva (RU) di questo EPS è stata isolata dalla membrana interna batterica, dove si trova legata a un trasportatore lipidico prima della polimerizzazione. Durante il lavoro di questa tesi, è stato sviluppato un protocollo per isolare l’unità ripetitiva e la sua RU biologica è stata determinata mediante analisi ESI-MS. Questo studio, che nasce dalla collaborazione con la Dr.ssa C. Lagatolla (Dip. di Scienze della Vita, Univ. di Trieste), impegnata nella ricerca dei geni coinvolti nella biosintesi del cepaciano, ha inoltre permesso di determinare lo schema di acetilazione della stessa, utile per la possibile comprensione del ruolo biologico dei gruppi acetili, e di investigare alcuni passi del processo biosintetico dell’unità ripetitiva del polisaccaride. L’utilizzo di enzimi in grado di degradare il polisaccaride cepaciano è utile per demolire la matrice esopolisaccaridica presente attorno ai batteri e consentirne l’utilizzo in sinergia con terapie antibiotiche. In precedenza era stato trovato un enzima, prodotto da un ceppo ambientale di Bacillus, con attività liasica verso il cepaciano. Il suo isolamento per un futuro sequenziamento è stato oggetto di questa tesi. Questo enzima è in grado di idrolizzare una catena laterale dello scheletro polisaccaridico del cepaciano, così lasciando intatta la natura polimerica dell’EPS. Benché l’eliminazione di questa catena laterale provochi una decisa diminuzione della viscosità, probabilmente dovuta alla perdita di aggregazione, la ricerca di altri enzimi in grado di idrolizzare lo scheletro polimerico a frammenti a basso peso molecolare sarà oggetto di ricerche future. La quarta linea di ricerca ha coinvolto lo studio dell’interazione tra diversi EPSs con tre peptidi antimicrobici del sistema dell’immunità innata, appartenenti alla famiglia delle catelicidine. Negli esperimenti sono stati utilizzati esopolisaccaridi prodotti da Pseudomonas aeruginosa, Inquilinus limosus e ceppi clinici del Bcc. L’inibizione dell’attività dei peptidi è stata valutata tramite saggi di minima concentrazione inibente su un ceppo di Escherichia coli di riferimento in presenza o meno degli EPSs. La formazione di complessi tra peptidi ed EPSs è stata studiata per mezzo di misure di dicroismo circolare, di spettroscopia di fluorescenza e di microscopia a forza atomica. E’ stato infine proposto un modello in cui peptidi strutturati ad α-elica interagiscono con gli esopolisaccaridi mediante interazioni elettrostatiche e non covalenti. Infine, la ricerca ha coinvolto il polisaccaride capsulare prodotto da Neisseria meningitidis gruppo A, usato per sviluppare un vaccino coniugato dal Prof. N. Ravenscroft (Dip. di Chimica, Univ. di Cape Town). In base alle sue osservazioni sulla scarsa reattività del polisaccaride durante il processo di derivatizzazione necessario per la coniugazione con le proteine, è stata formulata l’ipotesi di una possibile aggregazione del polisaccaride capsulare in presenza di ioni Ca2+, derivanti dal processo di purificazione dello stesso. Per capire e risolvere questo problema, le proprietà conformazionali e morfologiche del CPS in presenza di ioni NH4+, Na+ e Ca2+ sono state studiate nel laboratorio di Trieste, mediante tecniche quali dicroismo circolare, spettroscopia di fluorescenza, viscosimetria e microscopia a forza atomica. La presenza di ione ammonio porta le catene polimeriche ad una conformazione più allungata, mentre lo ione sodio favorisce il ripiegamento delle catene in una conformazione globulare. L’addizione di calcio produce aggregazione di un limitato numero di polisaccaridi globulari per formare una struttura ‘toroidal-like’. Questi risultati giustificano la bassa solubilità del polimero in forma calcio e offrono una spiegazione della sua bassa reattività nella preparazione del coniugato proteico. This PhD thesis focuses on polysaccharides (both capsular, CPSs, and exopolysaccharides, EPSs) produced by bacteria and released in the external environment as well as on their role in bacterial infections. For this, opportunistic bacteria involved in lung infections in cystic fibrosis (CF) patients were investigated. The results reported in the PhD thesis involved several topics and contributed to clarify different biological and chemical characteristics of bacterial polysaccharides. The first topic of the research has taken in account two structurally similar EPSs produced by a bacterium recently isolated from respiratory secretions of CF patients, Inquilinus limosus. The working hypothesis was that the two EPSs assume different conformations and therefore exert different biological activity. The conformational properties of these two EPSs were investigated by circular dichroism (CD) and atomic force microscopy (AFM) techniques, whereas the molecular modelling studies were carried out by Dr. M. Kuttel (Dept. of Chemistry, Univ. of Cape Town). The two polymers were shown to exhibit different helical conformations, stabilised by hydrogen bonding and characterised by charged pyruvate groups on the external helical surface. These findings confirming that the biosynthesis of two polymeric species might offer multiple tools to protect bacteria from external threats. The second part of the research involved cepacian, the EPS produced by the majority of bacteria belonging to the Burkholderia cepacia complex (Bcc). In this case, the isolation of its repeating unit (RU) from the internal bacterial membrane where it is bound to a lipid carrier prior to the EPS polymerisation was performed. A procedure to isolate cepacian repeating units was developed and the structure of the biological RU was determined by ESI-MS analysis. At the same time, the RU was useful to determine the acetylation scheme of cepacian, which could be connected with the biological role of acetyl groups, and to obtain information on its biosynthetic sequence. All the data obtained were considered in connection with the genetic investigation on the cepacian biosynthesis, carried out by Dr. C. Lagatolla (Dept. of Life Sciences, Univ. of Trieste). The third topic involved the search and identification of enzymes that can degrade cepacian in order to find biochemical tools able to degrade the EPS matrix present around bacteria and possibly used in synergy with antibiotic therapies. One lyase was found produced by Bacillus spp. and its isolation and sequencing is in progress. This lyase cleaves one of the lateral chains of cepacian backbone, thus living intact the polymeric nature of the EPS. Although the elimination of this lateral chain produces a marked lowering of the cepacian viscosity probably due to loss in aggregation ability, the search of different enzymes able to cleave the EPS backbone into oligosaccharides is still active. The fourth topic of the research focused on the interactions between different bacterial EPSs and three antimicrobial peptides, belonging to the cathelicidin family of primate’s innate immune system, to investigate the possible EPSs protective role towards bacterial cells. The experiments involved EPSs produced by Pseudomonas aeruginosa, Inquilinus limosus and clinical isolates of the Burkholderia cepacia complex. The inhibition of the peptides activity was assessed by minimum inhibitory concentration assays on a reference Escherichia coli strain in the presence and in the absence of EPSs. The complex formation between peptides and EPSs was investigated by means of CD, fluorescence spectroscopy and AFM. As a result, a model was proposed where peptides with a α-helical conformation interact with the EPSs backbone through electrostatic and non-covalent interactions. The last issue of the research involved the capsular polysaccharide produced by Neisseria meningitidis group A; this CPS is used to develop a protein conjugate vaccine by Prof. N. Ravenscroft (Dept. of Chemistry, Univ. of Cape Town). He observed that the derivatisation process, necessary prior to protein conjugation, yielded less product, in terms of protein-conjugate, than expected, and he thought that this was due to CPS aggregation in the presence of Ca2+, used in the purification steps. In order to solve this problem, the conformational and morphological properties of the CPS, in the presence of NH4+, Na+ and Ca2+ cations, were investigated in the Trieste laboratory. The study was carried out using different techniques such as circular dichroism, fluorescence spectroscopy, viscosity measurements and atomic force microscopy. It was shown that ammonium ions were associated with the presence of single polymer chains in an elongated conformation, whereas sodium ions favoured the folding of chains into a globular conformation. The addition of calcium ions produced the aggregation of a limited number of globular polysaccharide chains to form a ‘toroidal-like’ structure. This effect justified the polymer low solubility in the presence of calcium ions and offered an explanation of the polymer low reactivity in the preparation of the protein conjugate. XXII Ciclo 1981

2008/2009 The design of osseous implants, either load bearing or not, with desired mechanical and surface features that promote integration with bone and avoid risks of bone resorption and implant failure due to shear stresses, is still a challenging endeavour. The mechanical stresses which the skeleton undergoes affect bone formation and resorption processes. Bone remodelling is often promoted by adequate stress/strain conditions which are able to prevent bone mass loss. The largely used metallic implants offer several advantages like easy shape casting and modelling but include also several drawbacks like high stiffness if compared with the mechanical properties of native bone. A new generation of bone prosthesis is therefore indispensable to overcome the limitations of the obsolete metallic devices. In the orthopaedic framework, promising results have been achieved in the recent decades by three-dimensional structures named scaffolds. It is mandatory for any optimal scaffold to act as a temporary three-dimensional support for cell adhesion, growth and mineral matrix deposition. Moreover, ideal scaffolds should be able to integrate into surrounding tissue and mimic the structure and morphology of the natural bone tissue. Strict requirements for scaffolds are biocompatibility, a design closely resembling the natural extracellular matrix, an appropriate surface chemistry to promote cellular attachment, differentiation and proliferation and a sufficient mechanical strength to withstand in vivo stresses and physiological loading. Finally, the degradation of the ideal scaffold should proceed in a controlled way, keeping a sufficient structural integrity until the newly grown tissue has replaced the scaffold's supporting functions. Coupling a three-dimensional porous scaffold and a load bearing structure with suitable mechanical properties it is possible to obtain a device where the osteoconductive and osteoinductive properties of the former are synergistically linked with the mechanical ones of the latter. In this work both the aspects – osteointegration and load bearing – of an ideal prosthesis have been investigated. Alginate/Hydroxyapatite composite scaffolds were developed to be used either as scaffolds for sub-critical defects or as coatings for load bearing non-metallic bone prostheses. In both cases the investigation aimed to select suitable components and casting procedures to obtain the best results. The features of the single components and of the final three-dimensional structure were extensively investigated in order to obtain the most clarifying characterization both in terms of physical-chemical properties and in terms of biological responsiveness. The experimental section of this work involved physical-chemical analysis that helped to characterize both the organic and the inorganic components of the scaffold, respectively alginate and hydroxyapatite, before and after composite assembling. This investigation, based on several techniques (NMR, Rheology, XRD, Raman and TEM) allowed to characterize in detail the scaffold’s components and revealed the possibility of using the hydroxyapatite as a source of calcium ions for the gelification of the alginate without loosing the paramount osteoinductive properties of the mineral. Micro Computed Tomography (µ-CT) was employed to understand quantitatively the architectural features of the three-dimensional matrix obtained after alginate gel casting process. Moreover, this tool allowed to assess the influence of different manufacturing protocols (e.g. concentration of the components, casting temperatures) on the scaffold’s final structure. The results obtained by means of µ-CT coupled with the ones of Scanning electron microscopy (SEM) and confocal laser scanning microscopy (CLSM) analysis of the scaffolds showed an optimal interconnected porous structure with pore sizes ranging between 100 m and 300 m and over 88% porosity. Proliferation assays and SEM observations demonstrated that human osteosarcoma cell lines were able to proliferate, maintain osteoblast-like phenotype and massively colonize the scaffold structure. Once the in vitro behaviour of the structure was clear, in vivo tests were performed. Cone-like Alg/HAp scaffolds were tested on skeletally mature female New Zealand White rabbits and compared with positive (bioactive glass scaffold) and negative (without any implant) controls. Ex vivo investigations of the dissected samples were based on µ-CT and histological analysis and revealed high level of osteointegration and osteoconduction of the scaffolds. Moreover, efforts have been made to link the porous structure to the non-metallic fibre reinforced composite used as load bearing unit. Overall, these combined results indicate that the structure here developed is promising for being employed in orthopaedic applications. XXII Ciclo 1978

2008/2009 I peptidi in difesa dell’ospite (HDPs) esercitano molteplici ruoli nell’immunità, agendo sia come molecole antimicrobiche ad azione diretta sia come agenti immuno-moduatori. Il ruolo all’interfaccia tra immunità innata ed adattativa li rende molecole ideali per la futura applicazione nel trattamento di malattie infettive. Lo scopo di questo lavoro è stato quello di valutare le caratteristiche funzionali e strutturali di Catelicidine e Defensine selezionate al fine di correlare queste proprietà con l’ attività selettiva su cellule eucariotiche e procariotiche. Metodi biofisici e biochimici sono stati applicati alla catelicidina umana LL37 ed alcuni analoghi (ortologhi di primate e peptidi artificiali) con lo scopo di studiare la loro struttura e aggregazione in contatto con membrane biologiche e modello (i). Lo stesso approccio è stato anche applicato alle defensine umane hBD2 e 3 ed a loro analoghi. Inoltre, tecniche di microscopia, quali la microscopia a trasmissione elettronica (TEM), la microspettroscopia infrarossa in trasformata di fourier accoppiata ad una sorgente di sicrotrone (µSR-FTIR) e la citofluorimetria, sono state utilizzate in modo complementare al fine di studiare l’interazione a breve termine di hBD2 con cellule presentanti l’antigene, in particolare le cellule dendritiche immature (ii). (i) Lo studio strutturale e l’interazione di membrana di catelicidine ortologhe ci ha permesso di scoprire come l’evoluzione abbia lavorato sulla sequenza dei peptidi inducendo una diversa capacità di strutturare nei diversi ortologhi. Questo ha portato ad un’interazione differenziata e specifica con le membrane e a diversi meccanismi di lisi di membrana cellulare e probabilmente diversi modi di interagire con le cellule dell’ospite. (ii) Inoltre, abbiamo individuato una rapida interazione di hBD2 con le cellule presentanti l’antigene. hBD2 sembra indurre un riarrangiamento generale dei lipidi cellulari che sembra comportare un aumento nella fluidità di membrana e una ri- organizzazione del sistema endomembranoso. Queste variazioni potrebbero essere responsabili di un cambiamento morfologico delle cellule che promuoverebbe la mobilità cellulare in risposta a stimuli esterni. Questo studio dimostra l’esistenza di un posibile meccanismo alternativo di motilità cellulare rispetto alla chemotassi recettore-mediata, indicando un meccanimo di azione di hBD2 sulle iDC più complesso rispetto quanto riportato fino ad oggi. XXII Ciclo 1980

2008/2009 La medicina rigenerativa applicata al campo ortopedico è considerata una possibile opzione terapeutica per la riparazione del tessuto osseo danneggiato. Si stanno studiano e sviluppando una gran varietà di sostituti ossei sintetici come valida alternativa agli innesti di tipo tissutale. Lo scopo di questo lavoro di tesi è la progettazione e lo sviluppo di un materiale iniettabile per il riempimento dei difetti ossei. In particolare abbiamo sviluppato un riempitivo composito iniettabile usando materiali biodegradabili di tipo polisaccaridico funzionalizzati con elementi bioattivi come mediatori dell’adesione cellulare, fattori di crescita osteoinduttivi e peptidi di tipo antimicrobico. Il costrutto finale è un composito a due fasi, le cui parti, inizialmente, sono state sviluppate separatamente. La struttura è stata progettata come composta da una parte bioattiva, costituita da microsfere disidratate di alginato e idrossiapatite recanti peptidi, immersa in una matrice veicolante rappresentata da una soluzione concentrata di acido ialuronico. Con lo scopo di ottenere le microsfere bioattive, sono state esplorate diverse tecniche per la coniugazione peptide-polisaccaride e sono stati presi in considerazione svariati sistemi di rilascio di sequenze peptidiche. In particolare sono stati considerati tre peptidi noti per la loro bioattività: peptidi di tipo RGD, sequenza favorente l’adesione cellulare, un frammento della proteina BMP-2 (bone morphogenetic protein-2) in grado di promuovere il differenziamento di cellule mesenchimali ad osteoblasti e LL-37 peptide antimicrobico umano. Per ottenere un costrutto con proprietà bioadesive, gli sforzi sono stati volti ad un miglioramento dell’interfaccia fra le sfere di alginato/idrossiapatite e il tessuto osseo. Questo è stato possibile immobilizzando sulla superficie delle sfere dei peptidi contenti la sequenza RGD e assicurandone il gusto orientamento ed un’alta resa di immobilizzazione. Dopo aver testato diverse strategie chimiche, l’immobilizzazione effettuata sfruttando la formazione di un ponte disolfuro fra ChitLac preventivamente modificato con gruppi tiolici e un peptide contenente un residuo di cisteina, è stata valutata come la miglior strategia in termini di resa di reazione; allo stesso tempo le microsfere funzionalizzate in questo modo hanno dimostrato un’alta capacità di promuovere l’adesione e la crescita di osteoblasti in esperimenti effettuati in vitro. Un altro importante tema ha riguardato l’incorporazione di un frammento della proteina BMP-2 nel costrutto al fine di promuovere il differenziamento e quindi la proliferazione cellulare. La sequenza temporale degli eventi fisiologici, ha suggerito lo sviluppo di un sistema che risulta essere la somma di due diverse strategie di rilascio: la prima caratterizzata dal semplice intrappolamento del peptide in un sistema di sfere di alginato capace di un rilascio veloce, ed il secondo caratterizzato da un rilascio lento e costante ottenuto grazie all’azione idrolitica di esterasi. Questo è stato possibile inserendo un legame enzimaticamente idrolizzabile nella struttura contenente il frammento di BMP. Questa seconda strategia ha sfruttato una chimica altamente selettiva quale la “click chemistry”. La sintesi della molecola spaziatrice recante un legame enzimaticamente idrolizzabile è stata effettuata partendo da un γ-valerolattone. Questo spaziatore è stato inoltre progettato per recare un gruppo terminale funzionale adeguato per le reazioni di click chemistry ed è stato legato al ChitLac. La reazione di cicloaddizone è stata poi effettuata fra il ChitLac funzionalizzato con lo spaziatore e il frammento di BMP anch’esso opportunamente funzionalizzato. L’idrolisi enzimatica è stata verificata, inoltre è stata eseguita un’esaustiva caratterizzazione tramite NMR ed elettroforesi capillare. E’ stata poi presa in considerazione l’incorporazione di peptidi antimicrobici nel costrutto. Partendo da dati di dicroismo circolare riguardanti l’influenza che alcuni polisaccaridi hanno sulla conformazione del peptide LL-37 in soluzione, è stata riconosciuta alla miscela LL- 37/alginato la capacità di modulare la citotossicità del peptide. La miscela ha inoltre dimostrato la capacità di mantenere l’attività antimicrobica su batteri Gram negativi e questo ci ha spinto a considerarne l’applicazione su costrutti solidi con finalità ortopediche. In quest’ambito è stato escluso, da esperimenti in vitro, un effetto causato dal semplice contatto fra cellule o batteri e una superficie di alginato caricato col peptide. L’unico possibile meccanismo di rilascio riscontrato in un costrutto di alginato/LL-37 è stato tramite la degradazione di questo. Infine, il costrutto è stato assemblato incorporando le microsfere disidratate in una soluzione concentrata di acido ialuronico, ottenendo un costrutto dall’aspetto simile ad una pasta che è in grado di facilitare il processo di iniezione del riempitivo. Misure di tipo reologico hanno indicato in incremento della componente viscoelastica aggiungendo il particolato nella soluzione di acido ialuronico, questo generalmente è associato ad una buona capacità di recuperare l’elasticità dopo l’iniezione il che è un requisito richiesto ai biomateriali usati come riempitivi ossei. I risultati ottenuti hanno dimostrato l’applicabilità del composito come sostituto osseo dotato di proprietà bioattivite (osteoconduzione, osteoinduzione, bioadesività ed effetto antimicrobico su ceppi di tipo Gram negativo). Bone regenerative medicine is considered a potential therapeutic option for the healing of damaged bone tissue. A variety of synthetic bone graft substitutes have been investigated as alternative to current tissue based bone graft materials. In this study efforts have been made to achieve an injectable material to fill bone defects. We have designed composite injectable filler by using polysaccharides as biodegradable materials, and by functionalizing them with bioactive elements such as adhesion mediators, osteoinductive growth factors and anti-microbial peptides. The final construct is represented by a two-phase composite, whose parts have been initially built separately. The structure was thought as composed by bioactive alginate/hydroxyapatite dried microbeads functionalized with peptides, suspended in a concentrate hyaluronic acid solution as the matrix vehicle. With the aim to obtain bioactive alginate/HAp beads, different techniques of peptides-polysaccharides conjugation were considered and different peptide delivery systems were explored. In detail the peptides considered were three, the RGD active sequence with bioadhesive properties, a fragment of BMP-2 (bone morphogenetic protein-2) that promotes the differentiation of MSCs into osteoblasts, and LL-37 a human antimicrobial peptide. In order to obtain a construct with bioadhesive properties, efforts have been made to improve the tissue-alginate/HAp beads interface by immobilizing RGD peptide sequence in a manner that assure the right motif orientation and high yields. After having tested various chemical strategies, the immobilization by disulfide bridge formation between a ChitLac previously modified with a thiol group and a peptide containing a cysteine residue was evaluated to be the best in terms of yield; at the same time μbeads functionalized in that way exhibit a very high osteoblast adhesion and growth promotion in in vitro experiments. Another important topic regarded the incorporation of BMP-2 epitope fragment into the construct, to enhance differentiation and proliferation. Time-tuned physiological functionality suggested the development of a system being the sum of two delivery strategies: the first characterized by simple entrapment in alginate beads for a fast burst release and the second one giving a slow and constant release, obtained by the action of bone esterases on an enzymatically cleavable linker. This second strategy exploited a high selective chemistry such as that of click reactions. Starting by a γ-valerolactone the synthesis of the enzymatically cleavable spacer was performed. The linker was designed to contain a final functional group for click chemistry. It was bound to ChitLac and then to a modified BMP fragment by a click reaction. Enzymatic cleavage was verified and an exhaustively characterization by means of NMR and capillary electrophoresis was performed. The incorporation of antimicrobial peptides was also taken into account. Starting from the information gained by circular dichroism data on the influence on LL-37 conformation given by several polysaccharides, an optimal modulation activity of the cytotoxic effect of LL-37 was found using the peptide/alginate mixture. The demonstrated maintenance of antimicrobial activity on Gram negative strains drove to an application on solid constructs for orthopaedic applications. An effect caused by the simple contact between cells or bacteria and alginate surface loaded with the peptide, was excluded by in vitro experiments. The only possible release mechanism that an LL-37/alginate construct was able to show was related to the degradation of this one. Finally, the whole construct was assembled by incorporating the dried microbeads in a concentrate hyaluronic acid solution, obtaining a paste-like construct that can facilitate the injectability of the particulate. Rheological measurements indicated an increment of the viscoelastic component by adding particulate into hyaluronate solution; this, generally, is associated to a good capacity to regain the elasticity after the injection, as expected for a biomaterial for bone filler applications. The results collected demonstrate the applicability of the obtained composite as bone graft substitute with bioactive properties (osteoinduction, osteoconduction, bioadhesivity and antibiotic effect on Gram negative strains). XXII Ciclo 1978

2008/2009 Antimicrobial peptides (AMPs) participate in the immunity of both animals and plants. In mammals, the most important AMP families are defensins and cathelicidins. The only human cathelicidin, LL-37, is a 37-residue, α-helical, cationic peptide with a direct membranolytic antibacterial activity and various immunomodulatory activities. The aim of our research was to establish a procedure based on an E. coli knock-out mutant library in the search for mutations conferring altered susceptibility to AMPs. The library was created by random insertions of Tn5 transposon into the bacterial genome. The selection of low-susceptibility mutants and their identification was first tested with a Bac7-derived peptide, for which at least one interactor had been found. The procedure was then applied for the identification of mutations that modulate the susceptibility of E. coli cells to LL-37 with the aim of dissecting the steps of the peptide's mode of action. The differences between the wild-type and the mutant phenotypes were characterised using assays determining the bacterial culture growth inhibition and killing, as well as cell binding and membrane permeabilisation. In the majority of the isolated Bac7-resistant mutants, Tn5 was inserted in the sbmA gene, confirming previous findings that showed an essential role for the SbmA protein in the internalisation of Bac7, allowing it to exert its antimicrobial activity. Tn5 insertion in the gene waaY, identified in 15 out of 20 resistant mutants selected with LL-37, lowered the peptide’s ability to inhibit bacterial growth and to kill cells, decreased peptide binding to the bacterial surface and membrane permeabilisation. However, it did not alter the susceptibility to other AMPs. WaaY encodes a specific kinase that phosphorylates the Hep II residue in the core region of bacterial lipopolysaccharide. The inactivation of this enzyme determines a decreased negative charge of the outer membrane, which in turn causes a decrease in the peptide’s binding to the cell surface and in its antibacterial activity. The results reveal a putative LPS-binding site for LL-37 and stress the importance of its initial binding to the cell surface for antimicrobial efficacy. The established procedure utilising Tn5 mutants in the search for bacterial mutations conferring resistance to AMPs, proved efficient and opened a previously unreported branch of research on the mode of action of the human cathelicidin. I peptidi antimicrobici (AMP) fanno parte dell'immunità innata sia degli animali che delle piante. Le famiglie di AMP più importanti nei mammiferi sono le defensine e le catelicidine. L'unica catelicidina presente nell'uomo, LL-37, è un peptide lineare e cationico di 37 residui con una conformazione ad α-elica, che esercita un’azione microbicida diretta tramite permeabilizzazione di membrana. In aggiunta, il peptide è anche dotato di una serie di funzioni immunomodulatorie. Lo scopo del lavoro è stato quello di mettere a punto una procedura che utilizza una libreria di mutanti knock-out di E. coli, ottenuta mediante inserzioni casuali del trasposone Tn5 nel genoma batterico, per cercare mutazioni che rendano i batteri resistenti agli AMP. La bontà del metodo è stata prima verificata selezionando e identificando mutanti resistenti al frammento 1-16 del peptide Bac7. Una volta ottenuti i risultati attesi, la procedura è stata applicata all'identificazione di mutazioni che modulino la suscettibilità di E. coli a LL-37, con lo scopo di meglio comprenderne il meccanismo d'azione. I fenotipi wild-type e mutante sono stati caratterizzati usando saggi d'inibizione della crescita batterica, di killing e di capacità del peptide di legarsi ai batteri e di permeabilizzarli. Nella maggioranza dei mutanti resistenti al peptide Bac7 isolati, Tn5 ha inattivato il gene sbmA, confermando risultati precedenti che avevano portato a dimostrare l'importanza della proteina SbmA per l'internalizzazione di Bac7. Nel caso di LL-37, in 15 dei 20 mutanti resistenti selezionati il trasposone ha inattivato il gene waaY, che codifica per una chinasi specifica per la fosforilazione del residuo Hep II della regione core del lipopolisaccaride batterico. La mancata fosforilazione nei mutanti selezionati riduce la carica negativa della superficie batterica, che, a sua volta, diminuisce l'inibizione della crescita, il killing, il legame e la permeabilizzazione delle cellule da LL-37, ma non da altri peptidi, suggerendo che la modificazione abbia alterato specificamente l'attività di LL-37, rivelando un suo putativo sito di legame all’LPS e sottolineando l'importanza del legame iniziale alla superficie cellulare per l'efficacia battericida del peptide. La procedura messa a punto per la ricerca di mutazioni che inducono resistenza agli AMP è risultata efficiente ed ha indicato una nuova via per far luce sulla sequenza di eventi responsabile del meccanismo d'azione della catelicidina umana. XXII Ciclo 1982

2008/2009 Le proteine HMG (High Mobility Group) sono la famiglia più ampiamente e meglio caratterizata fra le proteine cromosomiche non istoniche. Esse sono raggruppate in tre sottofamiglie: HMGA, HMGB ed HMGN. Ciascuna sottofamiglia è caraterizzata da una sequenza o motivo funzionale attraverso il quale sono capaci di legarsi a specifiche strutture sul DNA o sulla cromatina. La famiglia HMGA di mammifero consiste di due geni funzionali: HMGA1 and HMGA2. Lo splicing alternativo del trascritto del gene HMGA1 dà origine alle proteine HMGA1a ed HMGA1b. La proteina HMGA2 è codificata dall’altro gene. Le proteine HMGA partecipano a specifiche interazioni proteina-DNA e proteina-proteina inducendo sia cambiamenti strutturali della cromatina che la formazione di complessi macromolecolari su regioni promotrici//enhancer di diversi geni. Pertanto esse sono descritte come fattori archietturali della trascrizione. Grazie alla loro multifunzionalità, ese partecipano a diversi processi biologici quali l’integrazione virale, l’embriogenesi e la differenziazione, l’apoptosi, la senescenza, la trasformazione neoplastica ed il riparo del DNA. La caratteristica di proteine oncofetali attribuita alle HMGA prende origine dal fatto che esse sono (i) altamente espresse durante l’embriogenesi, (ii) assenti o espresse a bassi livelli nei tessuti adulti e (iii) altamente riespresse nelle cellule trasformate. Queste proteine sono, tra le proteine nucleari, quelle più soggette a modificazioni post-traduzionali e le loro attività sono modulate da una ampia varietà di modificazioni post-traduzionali. In questo lavoro di tesi, grazie all’uso di linee cellulari di cancro mammario in cui è stata indotta la reversione del fenotipo neoplastico tramite trattamento farmacologico e successive analisi di spettrometria di massa, si è ricercata una connessione tra le modificazioni post-traduzionali delle HMGA ed il processo di trasformazione neoplastica. Inoltre, grazie ad una nuova strategia in vitro sviluppata nel nostro laboratorio per l’identificazione di partner molecolari delle proteine HMGA, si è dimostrato che la proteina HMGA1a si associa con la proteina Ku70, che è un fattore chiave coinvolto nel riparo delle rotture al doppio filamento di DNA attraverso il “non-homologous end joining”. Numerose evidenze sperimentali suggeriscono che l’inibizione del riparo del DNA da parte delle HMGA posa contribuire alle instabilità genetiche e cromosomiche comunemente ritrovate nelle celle cancerose. Perciò, si è ricercata una relazione funzionale tra le proteine HMGA e le proteine chiave nel meccanismo di riparo del DNA attraverso “non-homologous end joining”, focalizzando l’attenzione in particolare sulla proteina DNA-PK (DNA-dependent protein kinase), che ha una funzione regolatrice chiave in questo processo. XXII Ciclo 1973

2008/2009 Non-syndromic cleft lip with or without cleft palate (NSCLP) is one of the most common birth defect. Genetic studies on human populations have identified numerous predisposing factors, as MYH9 and JARID2, whose role during palatogenesis remains obscure. In order to improve our knowledge on pathogenetic mechanisms, we carried out expression studies during mouse palate development using RNA in situ hybridization. As reference genes (Tgfb3, Irf6, Pvrl1, Foxe1 and Tp63) known to be required for correct palate formation, Jarid2 and Myh9 are expressed in the epithelial cells of the palatine processes before and at the time of contact. Then, this signal decreases and eventually disappears concomitantly with the degradation of the medial epithelial cells. Consistent with these observations, RT-PCR carried out on dissected palatal shelves detected products of Myh9 and Jarid2 from embryonic day E14.0 to E15.0 with a pick at E14.5, the stage when shelves appose in the midline. Taken together, these expression studies strongly support the association studies that designate these genes as predisposing factors for NSCLP. In multifactorial diseases as NSCLP once genetic studies demonstrate association, a major challenge is to identify mutations. In this regard, we started analyzing two in linkage disequilibrium SNPs (rs3752462 of MYH9 and rs2076056 of JARID2) that could be involved in defective splicing mechanisms, as hypothesized on the basis of their localization within splice sites. Using a hybrid minigene assay, however, we demonstrated that none of the allelic variants lead to any aberrant products at least in HeLa cells, the model used for this study. Moreover, we investigated whether alternative splicing events of the Myh9 gene detected in cochlea and brain (Li et al., 2008) could also be found in other tissues and palate. A small insertion of 12 bp between exon 4 and 5 (loop1) due to an alternative splicing was detected in all adult tissues analyzed. The same insertion was not detected in embryonic tissues, such as palatal samples at different stages of development, and brain. These results suggest that the genomic region of loop1 should be investigated in all risk haplotypes to search for pathogenetic variants. In conclusion, further investigations should be planned to explore the role of MYH9 and JARID2 in orofacial development in order to dissect signaling pathways during palate formation and identify the causative variants contributing to NSCLP. XXII Ciclo 1973

2008/2009 Distant metastases of solid tumours are the major cause of cancer death. To improve the effectiveness of chemotherapy, only marginally active on secondary tumours, anti-metastatic agents are needed, i.e. compounds that display the capacity to selectively interfere with metastatic formation and growth. Among metal compounds, NAMI-A, HIm[Ru(III)Cl4Imdmso], has repeatedly shown a peculiar and selective anti-metastatic activity being able to prevent the formation, and to inhibit the growth of established secondary tumours. On these bases, the first phase of the project was aimed to investigate how variations on the NAMI-A chemical structure can influence the anti-metastatic activity. For this purpose, some representative complexes have been chosen: two heterocyclic compounds KP418, HIm[Ru(III)Cl4(Im)2], and KP1019, HInd[Ru(III)Cl4(Ind)2], presenting a different N-donor ligand, and three organometallic compounds, RM175, [(η6-biphenyl)Ru(II)Cl-(ethylendiamine)]PF6, its osmium congener AFAP51, [(η6-biphenyl)Os(II)Cl(ethylene-diamine)]BF4, and RAPTA-T, [RuCl2(η6-toluene)PTA], carrying a PTA ligand instead of ethylendiamine. The effects of the compounds on the interference with some steps of the metastatic progression are evaluated with appropriate in vitro tests, comparing the behaviour of the human MDA-MB-231 highly invasive breast cancer cells to that of human HBL-100 non tumorigenic mammary epithelial cells. To validate the model, the in vitro effects are compared with the in vivo anti-metastatic activity studied in the MCa mammary carcinoma of the CBA mouse. The results obtained highlight the selective activity of the organometallic compound RAPTA-T towards the highly invasive cell line in vitro, accompanied by a selective inhibition of metastasis development in vivo. RAPTA-T seems to act through the modulation of tumour cells-extracellular matrix interactions, and of cell motility. In particular RAPTA-T induces a cytoskeleton remodelling, mainly through the formation of stress fibres, that causes a stiffening of the cell body, particularly evident on MDA-MB-231 cells grown on ECM components. These effects, selectively identified in the highly invasive MDA-MB-231 cells, can be related to the higher ruthenium uptake, detected in this cell line. Cell adhesion, migration and invasion are directly related to actin assembly and disassembly, phenomena regulated by the RhoGTPases. RAPTA-T completely counteracts the increase of trypsin mediated cell detachment induced by the RhoGTPases inhibitor C3 transferase, in MDA-MB-231 cells grown on fibronectin and collagen IV, and on HBL-100 cells grown on collagen IV. These molecular events might stem from the cell surface and involve integrin adhesion molecules, as suggested by the role of ECM components in the functional tests and by the preference of RAPTA-T to bind them, above all, collagen IV, with which RAPTA-T interacts chemically as confirmed by an NMR analysis. The effect of RAPTA-T on cells immediately after the adhesion on the substrate, i.e. mainly adherent with integrinic receptors, suggests that RAPTA-T interacts mainly with integrins in the form already bound with the substrate. Le metastasi dei tumori solidi sono una delle principali cause di morte. Per migliorare l’effetto della chemioterapia, che è attiva solo marginalmente sui tumori secondari, sono necessari agenti anti-metastatici, cioè composti che si mostrino capaci di interferire selettivamente con la formazione e la crescita delle metastasi. Tra i vari composti di sintesi il NAMI-A, HIm[Ru(III)Cl4Imdmso], ha ripetutamente mostrato una selettiva azione anti-metastatica dimostrandosi capace di prevenire la formazione e di inibire la crescita di metastasi già formate. Su queste basi, la prima fase di questo progetto ha avuto lo scopo di indagare come variazioni nella struttura chimica del NAMI-A potessono influenzare l’azione anti-metastatica. A questo scopo sono stati scelti alcuni composti rappresentativi: due composti eterociclici il KP418, HIm[Ru(III)Cl4(Im)2], e il KP1019, HInd[Ru(III)Cl4(Ind)2], e tre composti organometallici, RM175, [(η6-biphenyl)Ru(II)Cl-(ethylendiamine)]PF6, il suo congenere a base di osmio AFAP51, [(η6-biphenyl)Os(II)Cl(ethylene-diamine)]BF4, e RAPTA-T, [RuCl2(η6-toluene)PTA], che al posto dell’etilendiamina presenta il ligando PTA. Gli effetti dei composti sull’interferenza con alcuni step della progressione metastatica sono stati valutati con alcuni test in vitro, paragonando il comportamento della linea cellulare altamente invasiva MDA-MB-231 proveniente da carcinoma mammario, con quello delle HBL-100 linea cellulare dell’epitelio mammario non tumorigenica. Per validare il modello proposto in vitro gli effetti dei composti sono stati paragonati con la loro azione in vivo studiata sul carcinoma mammario (MCa) dei topi CBA. I risultati ottenuti mettono in luce la selettiva azione del composto organometallico RAPTA-T verso la linea cellulare altamente invasiva in vitro, accompagnata dalla inibizione selettiva dello sviluppo delle metastasi in vivo. Il RAPTA-T sembra agire attraverso la modulazione delle interazioni cellula-matrice extracellulare e della motilità cellulare. In particolare il RAPTA-T induce il rimodellamento del citoscheletro, principalmente attraverso la formazione di fibre da stress, questo genera irrigidimento del corpo cellulare, particolarmente evidente sulle MDA-MB-231 cresciute sui componenti della matrice extracellulare. Questi effetti, selettivamente identificati sulla linea altamente invasiva, MDA-MB-231, possono essere correlati ad un più alto assorbimento di rutenio che è stato rilevato in questa linea cellulare. L’adesione cellulare, la migrazione e l’invasione sono direttamente correlate al rimodellamento del citoscheletro actinico, e questi fenomeni sono regolati dalle RhoGTPasi. Il RAPTA-T reverte completamente l’aumentato distacco conseguente al trattamento con la C3 transferasi, agente che inibisce le RhoGTPasi, nelle MDA-MB-231 cresciute su fibronectina e collagene IV e sulle HBL-100 cresciute sul collagene IV. Questi eventi potrebbero coninvolgere l’adesione integrinica ai substrati, come suggerito dal ruolo dei componenti della matrice extracellulare nei test funzionali presi in considerazione, dalla preferenza del RAPTA-T per il legame con questi, e soprattutto il collagene IV con cui il RAPTA-T interagisce chimicamente, come confermato dall’analisi NMR. L’effetto del RAPTA-T su cellule appena dopo l’adesione al substrato, quindi prevalentemente legate attraverso i recettori integrinici, suggerisce che il RAPTA-T interagisca principalmente con le integrine nella forma già legata i substrati. XXII Ciclo 1978

2008/2009 The Protein Arginine Methyltransferase 6 (PRMT6) is an enzyme characterized by a predominant nuclear localization and automethylation activity. Until today, very few information are known about its function and its substrates. To better characterize PRMT6, we looked for new partners and substrates, using the yeast two-hybrid system. This technique allowed us to discover 36 new partners for this enzyme, 19 of these interactions were confirmed with the GST Pull-down assay. Among these partners, 9 resulted to be interactors also for HMGA protein, a chromatin architectural factor that has been previously demonstrated to be an in vivo substrate of PRMT6. The binding between partners and PRMT6 was further assessed in vivo in mammalian cells, and out of nine proteins tested, 7 were confirmed. To test whether among the identified partners there were substrates, we performed an in vitro methylation assay, and discovered 4 new substrates of this enzyme (Macrophage migration inhibitory factor, Human DnaJ homologue, Small nuclear ribonucleoprotein-associated proteins B and B’, and Tubulin beta-2A chain). Moreover, we evaluated if the presence of HMGA1a, could modulate the methylation activity of PRMT6. We demonstrated that MIF’s methylation was enhanced in the presence of HMGA1a protein. Furthermore, to evaluate a possible role played by HMGA, MIF, and PRMT6 together, we studied the effect of MIF on Cyclin A (CycA) promoter. We choose this reporter because it is known the activation effect of HMGA2 protein on this promoter. Our data demonstrated that both MIF and PRMT6 could activate this promoter, but the simultaneous presence of these two proteins result in a lower activation, suggesting that PRMT6 had a negative effect on MIF function. PRMT6 è una Protein Arginine Methyltransferase caratterizzata da una localizzazione prevalentemente nucleare e capacità di autometilazione. Ad oggi, poche informazioni sono note circa la sua funzione ed i suoi substrati. Per meglio caratterizzare questo enzima, abbiamo cercato nuovi partners e substrati utilizzando il sistema del doppio ibrido in lievito. Questa metodica ci ha permesso di identificare 36 nuovi interattori. Per 19 di questi, l’interazione è stata confermata anche in GST Pull-down, e 9 sono risultati essere partners in comune anche con la proteina HMGA, un fattore architetturale della cromatina che è stato precedentemente dimostrato essere metilato in vivo da PRMT6. Successivamente, per 7 dei 9 interattori testati, il legame con PRMT6 è stato confermato anche in vivo in cellule di mammifero. Per testare se tra i vari partners identificati ci fossero anche dei substrati, abbiamo effettuato un esperimento di metilazione in vitro che ha portato all’identificazione di 4 nuovi substrati di PRMT6 (Macrophage migration inhibitory factor, Human DnaJ homologue, Small nuclear ribonucleoprotein-associated proteins B and B’, and Tubulin beta-2A chain). Abbiamo inoltre valutato se la presenza di HMGA1a potesse modulare l’attività di metilazione di PRMT6 e abbiamo dimostrato che la metilazione di MIF aumentava in presenza di HMGA1a. Inoltre, per valutare il possibile ruolo di HMGA, MIF e PRMT6 insieme, abbiamo studiato l’effetto di MIF sul promotore della Ciclina A. E’stato scelto proprio questo promotore in quanto è noto essere attivato della proteina HMGA2. Con i nostri esperimenti abbiamo dimostrato che sia MIF che PRMT6 possono attivarlo, ma la simultanea presenza di queste due proteine risulta in un’attivazione meno forte, suggerendo un effetto negativo per PRMT6 su MIF. XXII Ciclo 1982

2010/2011 Background. A major clinical problem in oncology is that patients with tumors of the same stage and histological type often do not have the same prognosis and they do not show the same response to the therapeutic treatments. This is a major concern in colorectal cancer (CRC), because of its high incidence and impact on public health. Aim. The main goal of the present work is to identify a set of markers of prognosis and therapy efficacy in CRC patients using a reverse translational research approach. In the first part of the work, we focused on the selection criteria to submit stage II colon cancer patients to 5-FU based adjuvant therapy, while in the second part we searched for new molecular targets to predict the response to cetuximab or panitumumab in the metastatic setting. Methods. To the first purpose, we evaluated the mRNA expression levels of thymidylate synthase (TS) by qRT-PCR and IHC, the MMR (mismatch repair system) status by immunohistochemistry (IHC), the CIMP (CpG island methylator phenotype) status by methylation specific PCR (MSP) and the expression of p53, β-catenin, GSK3β and CD8 by IHC on a retrospective case study comprising 60 stage II colon cancer patients submitted to 5-FU adjuvant therapy and 60 colon cancer patients who did not receive therapy. Results of molecular analyses were compared with clinical-pathological factors and histological features of the cancers. To the second purpose, we evaluated KRAS mutational status by direct sequencing and an alteration at the DNA level in one of the genes belonging to the EGFR pathway in a retrospective case study comprising 98 recurrent CRC patients treated with cetuximab or panitumumab and 65 recurrent patients who did not receive treatment. mRNA expression levels of the EGFR-pathway gene were evaluated by qRT-PCR. Information about the gene and the method for alteration assessment cannot however be disclosed because of patent pending. Results. In the first part of the study, we found a better survival in patients with a high mRNA expression level of TS and treated with 5-FU (87% survival for TS high versus 60% for TS low at 5 years of follow up) while an opposite behaviour was displayed by patients who did not received any therapy (83% survival for low TS versus 60% for high TS expressors at five years of follow up). These results were independent from clinical-pathological and the other examined molecular factors. Molecular parameters and clinical-pathological variables did not show any correlation with patients’ survival, with the exception of the presence of tumoral CD8+ infiltrating lymphocytes, which was a good prognostic factor (84% survival for patients with tumor infiltrated versus 63% for those without infiltration at 5 years of follow up), independently of 5-FU treatment. In the second part of the study we found, as expected, an advantage in progression free survival in patients having a wild type K-RAS, but a greater advantage in patients showing one of the types related to the alteration we evaluated (81% survival for patients with the type 1 alteration versus 51% for KRAS wild type patients at 6 months of follow up). The effect of such alteration was specifically related to the treatment with monoclonal antibodies because it was lost in the group of 65 recurrent patients without cetuximab/panitumumab treatment. We also found that higher mRNA levels of the EGFR pathway gene were predictive of better benefit from the therapy with monoclonal antibodies and this effect was shown to be cumulative with a wild type KRAS status. Conclusions. As far it concerns the first part of the study, we therefore identified mRNA expression level of thymidylate synthase (TS), the 5-FU main cellular target, as the most important discriminator in the decision of which stage II colon cancer patients should be submitted to adjuvant 5-FU therapy. This evidence confirms the results obtained previously with a different training set of patients. With respect to the second part of the study, we identified an alteration at the DNA level in one of the genes belonging to the EGFR pathway as a new biomarker of cetuximab/panitumumab treatment efficacy in recurrent CRC patients. XXIV Ciclo

2010/2011 Bacterial exopolysaccharides (EPS) are known to protect microbial cells from a hostile environment, often by forming biofilms. Moreover, pathogens which produce EPS as a defence against the human immune system were demonstrated to be more resistant to host defence mechanisms both in vivo and in vitro. Therefore, the knowledge of EPS structures is a prerequisite to understand their role in bacterial infections. In the Bacterial Polysaccharide Laboratory the research is focussed on bacteria involved in lung infections in cystic fibrosis (CF) patients, in particular on species belonging to the Burkholderia cepacia complex (Bcc) and on Stenotrophomonas maltophilia. This project focuses on the structure-function relationship of bacterial EPS involved in infections, by studying their primary structure and correlating it with functional properties. The research described in this thesis regards three main lines: 1) The definition of the primary structure of the EPS produced by one strain of Burkholderia vietnamiensis and two clinical strains of Stenotrophomonas maltophilia; 2) The scavenging effect of cepacian, the main EPS produced by the Bcc, towards reactive oxygen species (ROS); 3) The preliminary investigation of the role of EPS from the Bcc in the formation of biofilms. I polisaccaridi batterici (EPS) possono proteggere le cellule batteriche da un ambiente inospitale, spesso attraverso la formazione di biofilm. Inoltre batteri patogeni capaci di produrre EPS come difesa nei confronti del sistema immunitario umano, si sono dimostrati più resistenti ai meccanismi di difesa dell'ospite sia in esperimenti in vivo che in vitro. Di conseguenza la conoscenza della struttura primaria degli EPS è un prerequisito fondamentale per la comprensione del loro ruolo nell'infezione batterica. Nel laboratorio di ricerca su polisaccaridi batterici la ricerca è rivolta verso batteri coinvolti nelle infezioni polmonari di pazienti malati di fibrosi cistica (CF), in particolare specie appartenenti al Burkholderia cepacia complex (Bcc) e Stenotrophomonas maltophilia. Questo progetto si concentra sulla relazione tra struttura e funzione in EPS batterici coinvolti in infezioni studiando la loro struttura primaria e correlandola con le proprietà funzionali. La ricerca descritta in questa tesi riguarda tre linee principali: 1) La definizione della struttura primaria di EPS prodotti da un ceppo di Burkholderia vietnamiensis e da due isolati clinici di Stenotrophomonas maltophilia; 2) L'effetto di scavenging del cepaciano, il principale EPS prodotto da batteri del Bcc, nei confronti di specie reattive dell'ossigeno (ROS); 3) Studi preliminari sul ruolo degli EPS del Bcc nella formazione di biofilm. XXIV Ciclo

2011/2012 The retinoblastoma tumor suppressor protein (pRb) belongs to a cellular pathway that is deregulated in several human tumors. A hallmark of the pRb pathway is its ability to control G1-S transition of the cell cycle and to prevent uncontrolled cell proliferation. Nevertheless, research in the last years identified multiple alternative cellular functions regulated by pRb including several aspects of stem cell biology, that in contrast to pRb role in cell cycle control, are poorly understood. The work presented in this thesis aimed to investigate the post-transcriptional regulation of pRb expression and pRb function in the context of human cancer and stem cell biology. In the first part of this thesis work we evaluated the role of miR-335, a microRNA predicted to target RB1, to regulate pRb at post-transcriptional level and the impact of this regulation on pRb tumor suppressor function in cancer cells, focusing on the control of cell proliferation; in the second part we investigate the role of pRb in stem cells context studying the relationship between the control of cell-cycle progression and the control of self-renewal features. In particular we focused our research both on the regulation of its expression by evaluating the impact of miR-335 on pRb in the context of embryonic stem cells, and to link pluripotency transcription factor Oct4 with the regulation of the pRb pathway. MicroRNAs are small non-coding RNAs critically involved in the post-transcriptional regulation of gene expression and influencing all biological processes, including tumorigenesis, stem cell self-renewal and differentiation. Although Rb proteins have a critical role in many cellular processes, information on microRNAs that control Rb family proteins at post-transcriptional levels are still very limited. The initial aim of this project was to address a putative role for miR-335, a microRNA predicted to target mammalian RB1, in the control of pRb expression, cell proliferation and tumorigenesis. We were able to demonstrate that miR-335 is differentially expressed in human cancer cells and that it regulates the expression of human pRb by directly target a conserved sequence motif in its 3’UTR. High conservation of miR-335 and its RB1 target sequence in placental mammals, as well as differential expression in human cancer cell lines indicates an important role in cell cycle control. We found that the impairment of the pRb pathway by miR-335 is paralleled by a significant upregulation of p53 levels and that the activation of the p53 pathway in the context of miR-335 overexpression impairs cell proliferation and neoplastic transformation. In line with this, reducing p53 levels is sufficient to drive hyperproliferation and increased transformation in the context of ectopically increased miR-335 levels. In our study, we also showed that the activation of the p53 pathway during DNA damage significantly increases miR-335 levels and demonstrated that miR-335 mediates efficient cell cycle arrest on DNA damage in a positive feedback loop with p53. In conclusion, these results identify miR-335 as a potent regulator of pRb at posttranscriptional level, indicate that miR-335 helps control proliferation by balancing the activities of the pRb and p53 tumor suppressor pathway and that miR- 335 activation plays an important role in the induction of p53-dependent cell cycle arrest after DNA damage. Recently it has becoming evident that cancer cells exhibit several traits that are also characteristic of embryonic stem cells. In particular, the constitutive hyperphosphorylation of pRb that ensures rapid and indefinite cell proliferation potential of ESCs recapitulates the impaired function of the pRb pathway present in virtually all tumors. In addition, self-renewal expression signature including the pluripotency transcription factors Oct4, Nanog and Sox2 were also found in several types of human cancer. Although these evidences support the idea that the acquisition of stem cell-like gene expression signature represents a key step in tumorigenesis, evidence for a functional link between pluripotency transcription modules and the control of aberrant function of tumor suppressor pathways is limited. The second aim of this thesis work was to better characterize the role of pRb in stem cells context focusing on the regulation of its expression and function, connecting pluripotency transcription factor Oct4 with the Retinoblastoma tumor suppressor pathway. We were able to demonstrate that Oct4 cooperates with pRb in the establishment of self-renewal program of mouse embryonic stem cells (mESCs) and that these two factors are under control of miR-335 that targets conserved sequence motifs in the 3’UTRs of Oct4 and Rb1. We demonstrated that Oct4 is required to ensure the hyperphosphorylation of pRb, thereby permitting rapid cell proliferation of self-renewing mESCs. miR-335 plays a central role in interrupting the self-renewal promoting Oct4-pRb axis at the onset of mESC differentiation by targeting the expression of both proteins at the post- transcriptional level. In particular, we found that in self-renewing mESCs, high Oct4 expression levels drive the expression of Nipp1 and Ccnf to inhibit the activity of the protein phosphatase type 1 (Pp1) thereby establishing pRb hyperphosphorylation as a key feature of rapidly cycling, pluripotent mESCs. Upon induction of differentiation, transcriptional repression of Oct4 in conjunction with targeting of Rb1 and Oct4 by miR-335 causes the collapse of the Oct4-Nipp1/Ccnf1-Pp1-pRb axis, leading to a rapid pRb dephosphorylation, the exit from self-renewal and the establishment of pRb regulated cell cycle profile of differentiated cells. In conclusion, our results introduce a novel regulatory circuit, modulated by miR- 335, which connects the pluripotency transcription factor Oct4 with the pRb pathway to control mESC self-renewal and differentiation, and anticipate an important role for Oct4 in the inactivation of the Retinoblastoma tumor suppressor pathway in human cancers with embryonic stem cell gene expression signatures. XXV Ciclo 1982

2009/2010 Background: Clinical proteomics aims to characterize the protein networks altered by pathologic processes or therapeutic treatment, and to develop biomarker profiling technologies to promptly detect diseases and treat them more effectively. The challenge of translating proteomic profiling to the bedside lies in applying technologies for the analysis of human tissues, which are routinely obtained by biopsy or surgery, without substantially modifying the clinical workflow. Formalin-fixed and paraffin-embedded (FFPE) tissues are the most widely available archive material suitable to discover new biomarkers or therapy targets, for their validation and for their implementation in clinical practice. However, the lack of standardised guidelines for protein analysis of archive tissues has hampered their regular use in the daily practice. Furthermore, methods to extract proteins and to identify and analyse them also quantitatively have only recently been developed. This research was carried out to develop, standardise and apply qualitative and quantitative proteomic methodologies to archive tissues. Objectives: 1. Development, standardisation and validation of protein extraction methods and protocols from fixed and paraffin-embedded tissues with different pre-analytical conditions. 2. Application of protein analysis in two different research settings: a. at diagnostic level: screening of monoclonal antibodies against Human papillomaviruses (HPVs) in cervix uteri lesions suitable for early detection and progression risk assessment; b. at translational research level: identification of a new therapy target for glioblastoma, deepening the functions of human RECQ1 helicase, an enzyme involved in the maintenance of chromosome stability. Methods: 1. In order to develop, standardise and validate protein extraction methods and protocols from fixed and paraffin-embedded tissues, lung, colon and breast cancer specimens were collected by three groups from the European consortium IMPACTS (Integration of Molecular Pathology and Cell and Tissue Structure). Each specimen was divided into two equivalent sets: one was fixed in formalin and paraffin-embedded; the other one was fixed in a new alcohol-based fixative (FineFix) and paraffin-embedded. In order to compare the fixation procedures and the protein extraction efficiency, a total of 81 protein lysates were prepared in five different laboratories of the consortium. Selected marker proteins were compared, at quantitative level, between the two fixation procedures using the reverse phase protein array technology. Tests were applied to determine whether any of the fixation methods caused differences in protein lysate microarray measurements. Pairwise differences between fixation methods were tested. 2. (a) In diagnostics, protein analysis was applied to 138 FFPE cervix uteri squamous lesions provided by a multicentric retrospective study. They were characterized at molecular level for the presence and type of HPV by using PCR-based systems or the Hybrid Capture assay. Thirty-nine monoclonal antibodies directed against the E7 viral oncoprotein were supplied by a private company. They were validated by generating different tissue microarrays in which multiple punches for each bioptic sample were done, taking different cervical lesions (from low grade to invasive squamous cancer). The antibody providing better staining and specificity results was selected and validated with other tissue microarrays, which included a total of 138 tissue cores from different cervical lesions, and with different HPV-types infections. HPV capsid protein L1, the surrogate biomarkers p16, hTERT, p53 and Ubiquitin, and the proliferative marker Ki67 were also tested in the same way. For the selected E7 antibody and the six biomarkers, one-way analysis of variance was performed to detect significant differences in the mean number of positive cells and in the mean staining intensity (evaluated by a score from 0 to 3+). The analysis was carried out among different types of cervical lesions, both at cytoplasmatic and nuclear level, by epithelial layer. We made use of repeated-measure analysis of variance to ascertain significant differences in the mean number of positive cells between E7 monoclonal antibody and each biomarker. Spearman’s rank correlation test and linear regression analysis were employed in order to detect correlations between E7 and the other biomarkers. 2. (b) The application of proteomics to FFPE tissues at translational research level concerned the expression of the RECQ1 helicase in glioblastoma and in perilesional brain tissue, and its expression in lung, colon and thyroid carcinomas and in their perilesional tissues. All specimens were submitted to immunohistochemistry. The expression pattern of RECQ1 was further analysed on a tissue microarray containing 63 glioblastomas and 19 perilesional brain tissues. Furthermore, to test the function of RECQ1 in glioblastoma growth, RNA interference experiments were performed in different glioblastoma cell lines and the effect of gene silencing on cell growth and proliferation was monitored. The Wilcoxon test for matched-samples was performed to evaluate differences in immunohistochemistry outcome measures, i.e. the proportion of positive cells for RECQ1, the intensity of the signal and the number of positive cells for Ki67, between tumour and perilesional tissues. Spearman’s rank correlation test was applied to assess the relationship between intensity score of RECQ1 and the number of positive cells in tumour specimens. Kruskal-Wallis test was carried out to investigate differences in the proportion of positive cells for RECQ1 and Ki67 among groups defined by patient gender and by RECQ1 intensity score. Results: 1. Quantitative comparison between fixatives for the tested antibody panel (β-Actin, E-cadherin, EGFR, HER2 and ER) yielded a higher immunostaining intensity for the FineFix lysates compared to the FFPE ones (p = 0.04). Signal intensities for EGFR (p = 0.007) and HER2 (p = 0.02) were significantly higher in the FineFix samples compared to the FFPE samples, whereas no differences were observed for β-Actin, E-cadherin and ER. 2. (a) The E7 immunostaining was similar to that of p16, which is the most widely-used surrogate marker in clinical practice for all types of cervical squamous lesions. Furthermore, in high grade squamous intraepithelial lesions (HSILs) E7 was similar to hTERT. In the detection of squamous cell carcinoma, E7 was similar to hTERT and Ubiquitin. A comparison between the mean number of E7 and L1 positive cells showed a significant difference in low grade squamous intraepithelial lesions (p = 0.002), among HSILs (p = 0.01) and in squamous cell carcinoma (p = 0.01). Correlation analysis between the two viral proteins, E7 and L1, allows to identify two possible groups of lesions, both in the low grade squamous intraepithelial lesions (LSILs) and in the HSILs. One group was charcharterized by high level of E7 and low level of L1, the other group was characterized by high level of both marker. This groups were also ri-evalueted at morphological level and lead us to hypothesize two possible models in the progression of LSILs: the viral replicative model and the cell proliferative one. 2. (b) RECQ1 was highly expressed both in the perilesional sections and in lesional sections of all tumours analyzed, with the exception of glioblastoma where its expression was significantly higher in tumour (p = 0.001). RECQ1 expression was confined in the nuclei of the tumour cells, thus suggesting that this enzyme might play an important role in glioblastoma growth. As expected, immunohistochemical analyses against Ki67 showed that protein expression was higher in glioblastoma than in the surrounding normal tissue (p = 0.0008). No correlation between the proportion of positive cells for RECQ1 and Ki67 was found (Spearman’s rho = 0.2, p = 0.2), suggesting that the high expression of RECQ1 in brain tumours is not simply related to the higher degree of proliferation of these cells. An essential role of RECQ1 in tumour growth and proliferation was confirmed by RNA interference experiments and clonogenic assay on cell lines. Conclusions: During my doctorate, all the ojectives set were obtained. The goals achieved trough this research are the following:  The standardization of a suitable protein extraction protocol from FFPE tissues. This study demonstrated that it is possible to harmonize protein analysis in archive tissues in different European laboratories by using the same protocols for tissue processing and protein extraction.  The optimization of an appropriate protocol for molecular analysis of tissues fixed with new formalin-free fixative FineFix. The possibility of using proteomic approaches also on tissues fixed with reagents alternative to formalin has great potential for future pathology, when the use of formalin will be banned due to its carcinogenetic effects. Furthermore, FineFix has proven a suitable formalin substitute in clinical practice, because it preserves both specimens’ morphology and immunoreactivity.  The validation of a new antibody against the E7 oncoviral protein for the detection f HPV in cervical squamous lesions. This leads us to conclude that E7 might be a suitable specific marker for the diagnosis of cervical lesions and could be used also for others HPV-linked diseases. The analysis of a battery of E7 monoclonal antibodies has highlighted their great usefulness in differential diagnosis.  An in-depth knowledge of the function of RECQ1 in glioblastoma. RECQ1 expression has been studied in glioblastoma cell lines for the first time ever. This research has demonstrated its important role in glioblastoma growth and proliferation, and in maintaining genome stability.  The recommendation of RECQ1 helicase as therapy target in glioblastoma, based on the evidence that the enzyme expression in this tumour type is higher than in perilesional normal tissues or in cancers of different origins. These findings lead us to consider RECQ1 a reliable target for the development of new anti-cancer therapies to eliminate proliferating tumour cells. Finally, this research demonstrated that archive tissues could be a valuable source of material for proteomics research. XXIII Ciclo

2009/2010 This thesis describes several aspects of stem cell theory in human primary liver cancers, both in vivo and in vitro models, with focus on identification, expression and drug resistance related to the CSCs theory.. XXIII Ciclo