Tesis sobre el tema "Drugs repurposing"

Despite improvements in cancer therapy, overall survival for most cancer types has changed a little in the past decades. Drug repositioning represents a promising approach for discovering new therapeutic strategies for cancer therapy. Since several epidemiological studies reported lower cancer incidence in individuals receiving long term psychotropic drugs treatment, in this project we investigated 27 psychotropic drugs for their cytotoxic activity in several cancer cell lines. Consistent with the cationic amphiphilic structure of the most cytotoxic compounds, we investigated their effect on mitochondrial and lysosomal compartments. Penfluridol, ebastine, pimozide, fluoxetine, fluspirilene and nefazodone showed significant cytotoxicity, in the low micromolar range, in all cell lines tested. In MCF7 cells these drugs triggered mitochondrial membrane depolarization, increased the acidic vesicular compartments and induced phospholipidosis. Neither caspase nor autophagy inhibitors rescued cells from death induced by fluoxetine, fluspirilene and nefazodone. Treatment with 3-methyladenine rescued cell death induced by pimozide and spiperone. Conversely, inhibition of lysosomal cathepsins significantly reduced cell death induced by ebastin, penfluridol, pimozide, spiperone and mildly by fluoxetine. Lastly, spiperone caused apoptosis in colorectal and breast. Our unpublished data on the characterization of spiperone activity on adherent and stem-like colorectal cancer cells demonstrated that its cytotoxicity is linked to perturbations of intracellular calcium (Ca2+) homeostasis, which likely result in mitochondrial Ca2+ overload and membrane depolarization, cell cycle block in G1 phase, and apoptosis. Spiperone induced a PLC dependent Ca2+ release from the endoplasmic reticulum (ER) along with ER stress and unfolded protein response activation, resulting in CHOP upregulation. Altogether these data support the clinical development of psychotropic drugs for cancer therapy.

Even though it is confirmed that ion channels are at the centre of many diseases, approved drugs are only available for small percentage of these proteins, and yet many pathologically important ion channels like transient receptor potential (TRP) cation channels remain without approved drugs. One reason could be the time-consuming and expensive process in drug discovery. Which has high possibility of failure in any step even after approval and marketing. Therefore, repurposing approved drugs might be considered as a solution and may offer an accelerated procedure in finding new treatments for patients. For the present research we selected TRPM8 ion channel as a neglected target despite growing number of studies regarding its association with numerous diseases. In this project we have first identified potent antagonists for TRPM8 ion channel among approved drugs, by using mainly the automated patch clamp device IonFlux 16. Such device allowed us to screen blocking potency of drugs against TRPM8 ion channel in time efficient way. Our approach consisted of using ligand-based virtual screening method, to optimize our screening by identifying candidates for further screening. We also studied possible interactions of identified drugs with antagonist binding site on TRPM8 channel by molecular docking. Furthermore, we have evaluated the effects of identified antagonists against different types of pancreatic ductal adenocarcinoma (PDAC) cells. We were able to identify four drugs with IC50 lower than 50 µM including propranolol, propafenone, carvedilol and nebivolol. Among them nebivolol with IC50 = 0.97± 0.15 µM and carvedilol with IC50 = 9.1 ± 0.6 µM were the most potent blockers. Studying the interactions of identified drugs with known binding site of TRPM8 by molecular docking, revealed high possibility of direct binding of nebivolol to binding site of TRPM8. Nebivolol was the most cytotoxic drug against PDACs, but it was also toxic against non-cancerous HEK-293 cells. While carvedilol had cytotoxic against PDACs, interestingly it wasn’t cytotoxic against HEK-293 cells. Result of these study will provide promising candidates for drug repurposing and will propose promising lead compound in drug discovery for new antagonists of TRPM8 ion channel. Also, our method of approach for identifying candidate drugs as agonist or antagonist could be applied for other ion channels.

Includes bibliographical references
New chemotherapeutics are urgently needed to combat Mycobacterium tuberculosis (Mtb), the causative agent of tuberculosis (TB). The development of compounds that could potentiate the activity of known antimycobacterial drugs is a relatively unexplored approach to new TB drug discovery. This study aimed to generate metabolites of chlorpromazine (CPZ), a phenothiazine with demonstrated in vitro activity against Mtb, and to investigate their potential utility in combination with anti-TB drugs. 7-HydroxyCPZ (M2), CPZ-N-oxide (M3), CPZ sulfoxide (M1), nor-CPZ (M5), nor-CPZ sulfoxide (M6b) and CPZ-N-S-dioxide (M4b) were generated from CPZ using various biotransformation systems and identified by Liquid Chromatography - Mass Spectrometry (LC/MS). The identity of M2 was confirmed with reference to a 7-hydroxyCPZ standard. M3, M1, M5, M6b and M4b were synthesized de novo and used to identify the metabolites generated in the biotransformation samples. Individually, CPZ and its metabolites (M2, M3, M5) were weakly active (MIC99 >50μM) against M. smegmatis (Msm) and Mtb while M1, M6b & M4b did not exhibit a MIC99 even at very high concentrations. Generally, an improvement in activity was observed where CPZ or its metabolites were used in combination with known anti-TB drugs. The combinations that exhibited a fractional inhibition concentration index (FICI) of < 0.5 were defined as synergistic. A combination of M2 and spectinomycin (SPEC) exhibited the highest synergism against Msm (FICI 0.19) and Mtb (FICI 0.13). In vitro assays established that CPZ and M2 are bactericidal against Mtb whereas M3 and M5 are bacteriostatic on their own. In combination assays, the use of RIF with M3 and M5, bedaquiline (BDQ) with M2, and SPEC with M3 were bactericidal. At 140μM, CPZ and M1, M2, M3 treated samples exhibited a 2-fold up-regulation of the cydA (Rv1623c) gene which encodes an essential subunit of the cytochrome bd-type menaquinol oxidase in Mtb. The same observation was made for RIF/M2 and RIF/M5 treated samples. These results suggest that the metabolites retain the mechanism of action (MoA) as the parental CPZ. The Mtb 16S rRNA gene, rrs (MTB000019) was identified as the biological target for SPEC. This brought into perspective the underlying mechanisms at play when SPEC is used in combination with CPZ, its metabolites or other drugs, against mycobacteria. This study establishes the utility of combination assays in confirming the active metabolite(s) of known drugs and provides proof of concept data to support follow-up investigations of CPZ and its metabolites as potential compounds for novel combination therapies for anti-TB drug development.

Rare diseases caused by single-gene mutations affect almost one million Canadians. According to the Online Mendelian Inheritance in Man database, ~4,500 rare monogenic diseases have a known cause; but less than 5% of patients have access to disease-modifying drugs. The dearth of accessible drugs for patients suffering from rare genetic diseases is partly due to the astronomical costs of traditional drug development which, when combined with the small target population, make rare disease therapeutics unattractive ventures for the pharmaceutical establishment. The paucity of cost-effective treatments for rare diseases has resulted in the promotion of clinic-ready drug repurposing as a tenable strategy for rare disease therapeutics. To identify repurposed candidates for rare neurogenetic diseases, I conducted a transcriptome-wide drug screen in mouse primary cerebrocortical cultures. RNA sequencing was used to develop a database of transcriptome-wide differential expression for 218 clinic-tested drugs. The “Neuron Screen” database was queried to identify drugs that upregulate ~60 rare neurogenetic disease genes (type I hits). Gene set enrichment pathway analysis by Ingenuity Pathway Analysis (IPA) was used to identify network associated drug-gene interactions (type II hits). Both types of drug-gene hits were further assessed in vitro and in vivo by qRT-PCR and western blot analysis. This analysis showed that the IPA-based network-associated approach reduces the false positive rate when identifying differentially expressed genes in transcriptome-wide data-sets. The analysis also identified two drug-gene interactions with genes that cause rare neurogenetic disease, thyroid hormone-Pmp22 and dexamethasone-Mfsd2a, that merit further investigation. This work proves the utility of the Neuron Screen database to connect rare disease genes with transcript-modulating drugs and provides a starting point to understand the transcriptional effects of pharmacologic agents on the mammalian brain.

A quimioterapia disponível para o tratamento das leishmanioses conta com um número reduzido de fármacos, com efeitos adversos severos e progressivo aumento de resistência. O reposicionamento de fármacos oferece uma grande oportunidade para introdução de novas terapias. Antidepressivos orais têm demonstrado eficácia tanto in vitro quanto in vivo contra espécies de Leishmania spp. Neste estudo, o antidepressivo sertralina (SRT), e o fármaco ciclobenzaprina (CBP), um relaxante muscular de estrutura tricíclica análoga a antidepressivos, foram avaliados quanto a atividade contra Leishmania (L.) infantum. Estudos metabolômicos não dirigidos utilizando multiplataforma analítica, foram combinados a análises de parâmetros celulares, essenciais para obtenção de uma ampla descrição dos mecanismos de ação. A CBP mostrou uma atividade leishmanicida in vitro, com valor de CE50 de 4,3 ?M contra formas promastigotas e 8,6 ?M contra formas amastigotas intracelulares. O fármaco apresentou uma citotoxicidade (CC50) de 70,6 ?M em células NCTC, e um índice de seletividade similar a miltefosina. Os estudos de mecanismo de ação, sugeriram que a CBP se difunde pela membrana plasmática, causando diminuição do ??p e no interior citoplasmático, parece induzir um estresse do RE com liberação de Ca+2; concomitantemente, induz um desacoplamento brando da cadeia respiratória mitocondrial e depleção dos níveis de ATP. Com o efeito prolongado, a liberação de Ca+2 parece ativar a autofagia, e seu influxo para a mitocôndria potencializar os efeitos deletérios, diminuindo o ??m e aumentando a produção de ROS. A longo prazo, o CBP induz uma extensa alteração metabólica, caracterizada aumento dos níveis da maioria dos metabólitos identificados e atividade desregulada de transportadores de membrana, gerando alto gasto energético associado a condições insuficientes de produção de energia mitocondrial, resultando em morte celular. A sertralina também apresentou atividade leishmanicida in vitro, com valor de CE50 de 2 ?M contra formas promastigotas e 3,9 ?M contra formas amastigotas intracelulares. Sua toxicidade em células NCTC foi de 19,6 ?M, resultando em um índice de seletividade similar a miltefosina. Nossos estudos confirmaram a mitocôndria de Leishmania como alvo primário e, o efeito de desacoplamento da cadeia respiratória associado ao colapso energético, estresse oxidativo seguido da despolarização do ??m como a possível origem desta disfunção mitocondrial. Estudos metabolômicos evidenciaram que a extensão do desarranjo metabólico, abrange diminuição da capacidade de detoxificação do metabolismo tiol-redox, uma severa depleção do pool intracelular de aminoácidos e poliaminas, evidenciando uma completa deterioração do metabolismo energético, por meio de um mecanismo multialvo direcionado a vias metabólicas essências do parasita. Finalmente, este estudo descreve a atividade anti-Leishmania de dois fármacos orais aprovados, com mecanismos de ação letais e irreversíveis no parasita, encorajando o prosseguimento para futuros estudos pré-clínicos na leishmaniose visceral americana
The available chemotherapy for the treatment of leishmaniasis has a reduced number of drugs, with severe adverse effects and progressive increase of resistance. The drug repurposing offers a great opportunity for the introduction of new therapies. Oral antidepressants have been demonstrated efficacy both in vitro and in vivo against Leishmania spp. In this study, the antidepressant sertraline (SRT), and the drug cyclobenzaprine (CBP), a muscle relaxant with tricyclic structure analogous to antidepressants, were evaluated against Leishmania (L.) infantum. Untargeted metabolomic studies using multiplataform analysis were combined to cellular parameters to a broad description of the mechanisms of action. Cyclobenzaprine showed an in vitro leishmanicidal activity with an EC50 value of 4.3 ?M against promastigotes and 8.6 ?M against intracellular amastigote forms. The drug showed a cytotoxicity (CC50) of 70.6 ?M in NCTC cells, and a selectivity index similar to miltefosine. Mechanism of action studies suggested that CBP diffuses through the plasma membrane, causing a decrease of the ??p and inside the cytoplasm, the drug seems to induce an ER stress, with release of Ca+2; concomitantly, it induces a mild decoupling of the mitochondrial respiratory chain and depletion of ATP levels. With the prolonged effect, a release of Ca+ 2 appears to activate an autophagy, and its mitochondrial influx results in a potentiation of deleterious effects as decreasing of ??m and increasing ROS production. In long term, CBP induces an extensive metabolic alteration, characterized increased levels of most of the identified metabolites and unregulated activity of membrane transporters. These generates a high energy expenditure associated to limited conditions of mitochondrial energy production, resulting in the cellular death. Sertraline also showed in vitro leishmanicidal activity, with an EC50 value of 2 ?M against promastigotes and 3.9 ?M against intracellular amastigote forms. Its toxicity in NCTC cells was 19.6 ?M, resulting in a selectivity index similar to miltefosine. Our studies confirmed the mitochondria of Leishmania as the primary target, and the uncoupling of the respiratory chain associated with energy collapse, oxidative stress, and the depolarization of ??m as the possible origin of this mitochondrial dysfunction. Metabolomics evidenced an extended metabolic disarray caused by SRT encompassing a decrease in the scavenging capacity of the thiol-redox metabolism and a severe depletion of the intracellular pool of amino acids and polyamines. The complete deterioration of energetic metabolism was evident through a multi-target mechanism, affecting the main metabolic pathways of the parasite. Finally, this study describes an anti-Leishmania activity of two approved oral drugs with lethal and irreversible mechanisms of action in the parasite, encouraging future preclinical studies in American visceral leishmaniasis.

Le développement de novo de médicaments est un processus long et coûteux. De plus en plus, les développeurs de médicaments cherchent à mettre en oeuvre des stratégies rentables et à moindre risque pour le développement de produits pharmaceutiques. Le processus de trouver de nouveaux usages pour des médicaments existants en dehors de l'indication initiale pour laquelle ils ont été initialement approuvé est couramment désigné comme « repositionnement », « réorientation » ou « reprofilage ». Le développement de formulations différentes pour un même médicament pharmaceutique est communément désigné comme « reformulation » et le processus de trouver une autre utilisation thérapeutique d'un médicament déjà connu est dénommé « repositionnement ». Ces deux stratégies sont devenues un courant dominant dans le développement des médicaments. Les principaux objectifs de la recherche menée dans cette thèse sont de parvenir à proposer une nomenclature et la taxonomie solide et valable pour l'identification et la classification des stratégies de « repurposing » de médicaments ; évaluer les voies de régulation de stratégies de repositionnement et de reformulation, par types de stratégies et dans les 2 régions géographiques étudiées ; et déterminer les paramètres qui ont un impact sur la probabilité d'un résultat positif sur le prix, le remboursement et l'accès au marché vis-à-vis des conditions accordées pour le médicament original dans les deux régions géographiques dans l'étude
De novo drug development is a costly and lengthy process. As a result of such market forces, drug developers are increasingly striving to find cost effective and reduced-risk strategies for developing drug products and to protect existing products from competition, as well as to extend their patent protection time. The process of finding new uses for existing drugs outside the scope of the original indication for which they were initially approved is variously referred as repositioning, redirecting, repurposing, or reprofiling. The development of different formulations for a same pharmaceutical drug is commonly designated as “reformulation” and the process of finding a new therapeutic use for an already known drug is referred to as “repositioning”. Both strategies have become a mainstream in drug development. The main objectives of the research conducted in this thesis are to propose a robust and valid nomenclature and taxonomy for identification and classification of drug repurposing strategies, to evaluate which regulatory pathways and trends are taken by drug repositioning and reformulation, by repurposed types and within the Europe and the US and determine which parameters have the most and least impact on the probability of a successful outcome on pricing, reimbursement and market access in repurposing vis-à-vis the conditions granted for the original drug

Pharmaceutical industry is today struggling with its productivity as products keep failing after long and expensive development programs. The protability is further threatened by erce competition from cheaper product copies. As an attempt to increase the pipeline output, pharmaceutical companies have lately turned to the strategy of drug repositioning. By applying an already developed drug in new disease areas the lifetime of the product is prolonged and return time on already made investments elongated.  Such development is imbued by less risk than a de novo development and has proven to be a faster and cheaper way to meet the medical demand. With limited company budgets and the often many repositioning possibilities, an informed repositioning selection must be made. As such theoretical model is not publicly available this thesis takes on the task to determine which parameters to take into consideration and how these should be weighted in relation to each other in order to evaluate di erent drug repositioning possibilities. Six main topics are identied to a ect the repositioning success, these are: medical need, economic return, scientic support, timing, life cycle extenders and external relations. These ndings are derived from empirics collected during interviews with employees from ve di erent competence areas involved in repositioning initiatives, na mely: research & development, clinical studies, regulatory a airs, pricing, and commercial. By further support from literature within the elds of drug repositioning and R&D project selection a three-step repositioning model was developed. The first step in the three-step repositioning model consists of primary parameters, these are essential parameters that have to be fullled in order to perform a repositioning strategy. If any of the primary parameters are not fullled, the repositioning opportunity should be killed in a go/no-go decision. In a second step, the secondary parameters are evaluated in a scoring model in order to determine the economical outlook of each repositioning opportunity. The opportunities showing greatest economical outlook should further be evaluated in the third and nal step in the three-step repositioning model. In this nal step the di erent repositioning opportunities are evaluated by their coherence with an overall corporate strategy. By applying this repositioning model to a repositioning selection scarce company resources  ay be focused on the repositioning opportunities showing best future prospect. Evaluating the potential of repositioning opportunities in a structured way should also increase chances to succeed. If successful, a repositioning initiative may a ect both company and society as the company improves return on earlier investments, while more patients in need of treatment will receive access to it. However, the three-step repositioning model presented in this thesis should be tested for more cases and perhaps be complemented with additional parameters or di erent gradings in order to optimize the selection.

Le repositionnement de médicaments consiste à trouver de nouvelles utilisations thérapeutiques pour des médicaments existants qui sont déjà approuvés pour traiter d’autres pathologies. Cette approche profite des connaissances déjà existantes sur ces molécules, permettant ainsi un développement plus rapide et moins coûteux par rapport à la création de nouveaux médicaments. Le repositionnement est particulièrement utile pour répondre à des besoins médicaux non satisfaits, comme par exemple pour les maladies rares ou émergentes. Ces dernières années, le développement de graphes de connaissances a permis de concentrer toutes ces informations biomédicales autour du médicament issues de grandes bases de données ou de connaissances. Un graphe de connaissances est une représentation structurée d’informations provenant de différentes sources, qui relie ces informations les unes aux autres par l’utilisation de relations. Cette représentation est particulièrement utile pour mieux comprendre les relations complexes qui structurent nos connaissances sur un médicament. Elle est utilisée de nos jours pour la tâche de repositionnement en particulier. Une façon efficace de repositionner des médicaments à partir de ces graphes est d’utiliser des méthodes d’intelligence artificielle qui prédisent de nouveaux liens entre les objets du graphe. De cette manière, un modèle correctement entraîné sera capable de proposer une nouvelle connexion entre un médicament et une maladie, indiquant une potentielle opportunité de repositionnement. Cette méthodologie présente cependant un gros désavantage : les modèles pour la prédiction de liens fournissent souvent des résultats opaques, qui ne peuvent pas être interprétés par l’utilisateur final des prédictions. Cette thèse propose d’étudier l’utilisation de méthodes d’intelligence artificielle explicables dans le but de repositionner des médicaments à partir de données biomédicales représentées dans des graphes de connaissances. Dans un premier temps, nous analysons l’impact du pré-entraînement sur les modèles de multihop reasoning pour la prédiction de liens. Nous montrons que la construction des représentations des entités du graphe avant l’entraînement du modèle permet une amélioration des performances prédictives, ainsi que de la quantité et la diversité des explications. Dans un second temps, nous étudions comment l’ajout de relations dans un graphe de connaissances affecte les résultats de prédiction de liens. Nous montrons que l’ajout de liens dans trois graphes biomédicaux permet une amélioration des performances prédictives du modèle SQUIRE, et ce sur différents types de relations lien avec le repositionnement de médicaments. Une analyse de l’impact sur l’explicabilité du modèle est aussi menée à la suite de l’ajout de ces relations. Enfin, nous proposons une nouvelle méthodologie pour la tâche de classification de liens dans un graphe de connaissances, basée sur l’utilisation de forêts aléatoires. À partir des informations concernant le voisinage de chaque noeud dans le graphe, nous montrons qu’un modèle de forêts aléatoires est capable de prédire correctement l’existence ou non d’un lien entre deux noeuds. Ces résultats permettent une visualisation des noeuds utilisés pour réaliser la prédiction. Enfin, nous appliquons cette méthode au repositionnement de médicaments pour la sclérose latérale amyotrophique (SLA)
Drug repositioning involves finding new therapeutic uses for existing medications that are already approved to treat other conditions. This approach takes advantage of the existing knowledge about these molecules, enabling faster and less costly development compared to creating new drugs. Repositioning is particularly useful for addressing unmet medical needs, such as rare or emerging diseases. In recent years, the development of knowledge graphs has enabled the consolidation of all this biomedical information around drugs, coming from large data sources or knowledge repositories. A knowledge graph is a structured representation of information integrated from different sources, linking these pieces of information together using relationships. This representation is especially useful for understanding the complex relationships that structure knowledge about drugs. Nowadays, it is widely used for the task of drug repositioning. An effective way to reposition drugs using these graphs is to employ artificial intelligence (AI) methods that predict new links between objects in the graph. In this way, a well-trained model can suggest a new connection between a drug and a disease, indicating a potential opportunity for repositioning. However, this methodology has a significant disadvantage : link prediction models often provide opaque results that cannot be easily interpreted by the end users. This thesis proposes to explore the use of explainable AI methods for the purpose of repositioning drugs based on biomedical data represented in knowledge graphs. First, we analyze the impact of pre-training on multihop reasoning models for link prediction. We demonstrate that building representations of the graph entities before model training improves the predictive performance, as well as the quantity and diversity of explanations. Secondly, we examine how the addition of relationships in a knowledge graph affects link prediction results. We show that adding links in three biomedical knowledge graphs improves the predictive performance of the SQUIRE model across different types of relationships related to drug repositioning. An analysis of the impact on model explainability is also conducted, following the addition of these relationships. Finally, we propose a new methodology for the task of link classification in a knowledge graph, based on the use of random forests. Using information about the neighborhood of each node in the graph, we show that a random forest model can accurately predict the existence or absence of a link between two nodes. These results allow for a visualization of the nodes used to make the predictions. Lastly, we apply this method to drug repositioning for amyotrophic lateral sclerosis (ALS)

The goal of this study was to identify novel drug repurposing opportunities in cancer by utilizing the off-target profiles of clinically relevant kinase inhibitors. This was based on the observation that the global target profiles of compounds are largely ignored and that many compounds have activity that cannot be explained by their cognate target alone. Additionally, by utilizing clinically relevant compounds, any results would hold a high potential for eventual clinical implementation. We utilized a systems pharmacology approach utilizing cell viability-based drug screening to identify compounds with beneficial off-target activity and then using chemical and phosphoproteomics in order to elucidate the mechanisms of action of these compounds. We found that tivantinib has off-target activity in NSCLC cells through inhibition of GSK3. Based on tivantinib’s ability to inhibit GSK3, we hypothesized that tivantinib would therefore have activity in acute myeloid leukemia (AML). We found that tivantinib had potent activity in AML through inhibition of GSK3. We also identified a highly synergistic combination with ABT-199 by drug synergy screening which was effective in HL60 cells and patient derived AML cells. We also found that the anaplastic lymphoma kinase (ALK) inhibitor, ceritinib, had activity across several ALK-negative lung cancer cell lines. We utilized integrated functional proteomics to identify the new targets and network-wide signaling effects. Combining pharmacological inhibitors and RNA interference revealed a polypharmacology mechanism involving the noncanonical targets IGF1R, FAK1, RSK1 and RSK2. Mutating the downstream signaling hub YB1 protected cells from ceritinib. Consistent with YB1 signaling being known to cause taxol resistance, combination of ceritinib with paclitaxel displayed strong synergy, particularly in cells expressing high FAK autophosphorylation, which we show to be prevalent in lung cancer. Together, we present a systems chemical biology platform for elucidating multikinase inhibitor mechanisms, synergistic drug combinations, mechanistic biomarker candidates and identifying novel drug repurposing opportunities.

Background and Rationale: Although the majority of patients with endometrial cancer (EC) are diagnosed early when disease is confined in the uterus and prognosis is excellent, there is a subset of patients with dismal prognosis. Carboplatin and paclitaxel is the standard chemotherapeutic regimen for EC. Given that response to chemotherapy impacts disease prognosis, especially in advanced, recurrent and metastatic disease, novel chemotherapeutic agents with improved safety profile are necessary to improve response rates and outcomes in these patients. Quinacrine (QC) is an inexpensive antimalarial drug with a predictable safety profile which recently surfaced as a promising anticancer agent thought to be associated with decreased risk of developing chemo-resistance through targeting multiple pathways simultaneously.

Objective: To generate preclinical data on the effect of QC in inhibiting tumorigenesis in EC both in vitro and in vivo as well as explore its role as an adjunct to standard chemotherapy in vivo in an EC mouse xenograft model.

Methods: Five different EC cell lines (Ishikawa, Hec-1B, KLE, ARK-2, and SPEC-2) representing different histologies, grades of EC, sensitivity to cisplatin and p53 status were used for the in vitro studies. MTT and colony formation assays were used to examine QC’s ability to inhibit cell viability in vitro. Drug combination studies were performed and the Chou-Talalay methodology was employed in order to examine synergism between QC and cisplatin, carboplatin or paclitaxel. A cisplatin-resistant EC subcutaneous mouse xenograft model was used in order to explore QC’s anticancer activity in vivo and assess its role as maintenance therapy.

Results: QC exhibited strong synergism in vitro when combined with cisplatin, carboplatin or paclitaxel with the highest level of the synergism being observed in the most chemo-resistant EC cell line. Neither QC monotherapy nor standard chemotherapy significantly delayed tumor growth in the mouse xenografts. Co-administration of QC with standard chemotherapy significantly augmented the antiproliferative ability of these chemotherapeutic agents as evidenced by the significant decrease in tumor burden. Combination treatment was associated with a 14-week prolongation of median survival compared to standard chemotherapy alone. Maintenance therapy with QC following standard chemotherapy was proven superior to standard chemotherapy as it resulted in long-term stabilization of disease evidenced by lack of significant tumor progression and further prolongation of overall survival. QC treatment alone, in combination with standard chemotherapy or as maintenance therapy was well-tolerated and was not associated with weight loss compared to control mice. A yellow skin discoloration was noted during active treatment with QC which was entirely reversible within a few days upon discontinuation of treatment.

Conclusions: QC exhibited significant antitumor activity against EC cell lines in vitro and was successful as maintenance therapy in chemo-resistant EC mouse xenografts. This preclinical data suggest that QC may be an important adjunct to standard platinum-based chemotherapeutic regimens for patients with recurrent EC.

De nombreux systèmes complexes du monde réel peuvent être représentés par des graphes, où les nœuds représentent des entités et les liens des relations entre les paires de nœuds. La prédiction de liens (LP) est l'un des problèmes les plus intéressants et les plus anciens dans le domaine de l'exploration de graphes ; elle prédit la probabilité d'un lien entre deux nœuds non connectés. Cette thèse étudie le problème LP dans les graphes simples et les graphes de connaissances (KGs). La première partie de cette thèse se concentre sur le problème LP dans les graphes simples. Dans la première étude, des approches basées sur la similarité et sur l'encastrement sont évaluées et comparées sur des graphes simples de différents domaines. L'étude a également identifié la difficulté de fixer le seuil du score de similarité pour calculer la métrique de précision des approches basées sur la similarité et a proposé une nouvelle méthode pour calculer la métrique. Les résultats ont montré la supériorité attendue des approches basées sur l'intégration. Cependant, chaque approche basée sur la similarité s'est avérée compétitive sur des graphes aux propriétés spécifiques. Nous avons pu vérifier expérimentalement que les approches basées sur la similarité sont explicables mais manquent de généralisation, tandis que les approches basées sur l'encastrement sont générales mais non explicables. La deuxième étude tente de surmonter la limitation de l'inexplicabilité des approches basées sur l'encastrement en découvrant des connexions intéressantes entre elles et les approches basées sur la similarité. La troisième étude démontre comment les approches basées sur la similarité peuvent être assemblées pour concevoir une approche LP supervisée explicable. Il est intéressant de noter que l'étude montre des performances LP élevées pour l'approche supervisée sur différents graphes, ce qui est très satisfaisant. La deuxième partie de la thèse se concentre sur les LP dans les KGs. Un KG est représenté comme une collection de triplets RDF, (head,relation,tail) où les entités head et tail sont reliées par une relation spécifique. Le problème de LP dans un KG est formulé comme la prédiction de la tête ou de la queue manquante dans un triplet. La LP basée sur l'incorporation de KG est devenue très populaire ces dernières années, et la génération de triplets négatifs est une tâche importante dans les méthodes d'incorporation. La quatrième étude traite d'une nouvelle méthode appelée SNS pour générer des triplets négatifs de haute qualité. Nos résultats montrent une meilleure performance LP lorsque SNS est utilisé que lorsque d'autres méthodes d'échantillonnage négatif sont utilisées. La deuxième étude traite d'une nouvelle méthode d'extraction de règles neuro-symboliques et d'une stratégie d'abduction pour expliquer les LP par une approche basée sur l'intégration en utilisant les règles apprises. La troisième étude applique notre LP explicable pour développer une nouvelle approche de repositionnement des médicaments pour COVID-19. L'approche apprend un ensemble d'enchâssements d'entités et de relations dans un KG centré sur COVID-19 pour obtenir un meilleur enchâssement des éléments du graphe. Pour la première fois à notre connaissance, des méthodes de criblage virtuel sont ensuite utilisées pour évaluer les prédictions obtenues à l'aide des embeddings. L'évaluation moléculaire et les chemins explicatifs apportent de la fiabilité aux résultats de prédiction et sont de nouvelles méthodes complémentaires et réutilisables pour mieux évaluer les molécules proposées pour le repositionnement. La dernière étude propose une architecture distribuée pour l'apprentissage des KG embeddings dans des environnements distribués et parallèles. Les résultats révèlent que l'apprentissage dans l'environnement distribué proposé, par rapport à un apprentissage centralisé, réduit considérablement le temps de calcul des méthodes d'incorporation KG sans affecter les performances des LP
Many real-world complex systems can be well-represented with graphs, where nodes represent objects or entities and links/relations represent interactions between pairs of nodes. Link prediction (LP) is one of the most interesting and long-standing problems in the field of graph mining; it predicts the probability of a link between two unconnected nodes based on available information in the current graph. This thesis studies the LP problem in graphs. It consists of two parts: LP in simple graphs and LP knowledge graphs (KGs). In the first part, the LP problem is defined as predicting the probability of a link between a pair of nodes in a simple graph. In the first study, a few similarity-based and embedding-based LP approaches are evaluated and compared on simple graphs from various domains. he study also criticizes the traditional way of computing the precision metric of similarity-based approaches as the computation faces the difficulty of tuning the threshold for deciding the link existence based on the similarity score. We proposed a new way of computing the precision metric. The results showed the expected superiority of embedding-based approaches. Still, each of the similarity-based approaches is competitive on graphs with specific properties. We could check experimentally that similarity-based approaches are fully explainable but lack generalization due to their heuristic nature, whereas embedding-based approaches are general but not explainable. The second study tries to alleviate the unexplainability limitation of embedding-based approaches by uncovering interesting connections between them and similarity-based approaches to get an idea of what is learned in embedding-based approaches. The third study demonstrates how the similarity-based approaches can be ensembled to design an explainable supervised LP approach. Interestingly, the study shows high LP performance for the supervised approach across various graphs, which is competitive with embedding-based approaches.The second part of the thesis focuses on LP in KGs. A KG is represented as a collection of RDF triples, (head,relation,tail) where the head and the tail are two entities which are connected by a specific relation. The LP problem in a KG is formulated as predicting missing head or tail entities in a triple. LP approaches based on the embeddings of entities and relations of a KG have become very popular in recent years, and generating negative triples is an important task in KG embedding methods. The first study in this part discusses a new method called SNS to generate high-quality negative triples during the training of embedding methods for learning embeddings of KGs. The results we produced show better LP performance when SNS is injected into an embedding approach than when injecting state-of-the-art negative triple sampling methods. The second study in the second part discusses a new neuro-symbolic method of mining rules and an abduction strategy to explain LP by an embedding-based approach utilizing the learned rules. The third study applies the explainable LP to a COVID-19 KG to develop a new drug repurposing approach for COVID-19. The approach learns ”ensemble embeddings” of entities and relations in a COVID-19 centric KG, in order to get a better latent representation of the graph elements. For the first time to our knowledge, molecular docking is then used to evaluate the predictions obtained from drug repurposing using KG embedding. Molecular evaluation and explanatory paths bring reliability to prediction results and constitute new complementary and reusable methods for assessing KG-based drug repurposing. The last study proposes a distributed architecture for learning KG embeddings in distributed and parallel settings. The results of the study that the computational time of embedding methods improves remarkably without affecting LP performance when they are trained in the proposed distributed settings than the traditional centralized settings

Phenotypic screening of drug molecules relies on the generation of a specific response; however the means by which this is elicited often remains unknown. Affinity chromatography is a valuable tool in the discovery of drug binding partners and may even allow the elucidation of the wider interactome of the initial drug target. The introduction of easily cleavable linkers and affinity-independent elution protocols to affinity chromatography is of current interest, since they render the technique much more adaptable with respect to the characterisation of biologically active species of interest. This thesis details the application of a novel azobenzene linker developed by the Hulme group for use in affinity-independent chromatography. The first chapter reviews recent developments in affinity chromatography and describes the synthesis of an affinity linker toolbox with both affinity-dependent and affinity-independent linkers. These linkers are functionalised with an azide moiety for use in CuAAC coupling to alkynyl derivatives of bioactive small molecules and have been modified to include photoreactive groups giving a series of linkers for use in the identification of less abundant, or low affinity, proteins. The first drug investigated, anisomycin (ANS), is a small molecule which was initially introduced as an antibiotic drug (Flagecidin). At nanomolar concentrations ANS has been shown to affect the mitogen activated protein kinase (MAPK) pathways; downstream effects of these pathways are thought to play a role in a range of pathological disorders such as Alzheimer’s disease, cancer and spinal muscular atrophy (SMA). ANS is thus a candidate for drug repurposing. Although the downstream effects of MAPK/SAPK pathway activation induced by anisomycin are well-documented, the cellular target has yet to be revealed. Previous work by the Hulme group has shown that the N-propargyl anisomycin derivative (I) retains the biological activity of the lead compound ANS. Thus to evaluate the cellular protein targets, N-propargyl ANS (I) was coupled onto the linker toolbox to create an ANS affinity probe library as described in chapter 2. The second drug investigated, fingolimod, was introduced as an immunomodulating drug (Glienya) for the treatment of multiple sclerosis (MS). This small molecule has also been shown to have anti-cancer properties in a range of cancer cell lines; however the precise mechanism by which this is effected is unknown. Literature precedent shows that terminal modification of fingolimod generates analogues which still retain biological activity. Thus a novel fingolimod alkyne derivative (II) was synthesised and used to create an affinity probe library as described in chapter 3. Chapter 4 describes affinity pull-down experiments conducted with the aim of finding the protein target(s) of ANS and fingolimod, using the affinity probe libraries generated in chapters 2 and 3. This chapter concludes with a discussion of the implications of these findings and directions for future study.

Acute myeloid leukaemia (AML) is a group of aggressive haematopoietic malignancies associated with adverse outcome. Fms-like tyrosine kinase 3 (FLT3) receptor mutations confer a particularly poor prognosis to AML patients. There is no satisfactory treatment against this disease, especially for the cases harbouring FLT3 mutations, and the quest for novel therapeutic options continues. Drug repurposing represents a powerful strategy to single out existing agents active in novel therapeutic contexts. We performed a high-throughput drug screening, designed to search for agents that inhibit the growth of AML cell lines with mutated FLT3 within libraries of FDA-approved compounds or molecules in advanced phases of clinical trials. Auranofin, an antirheumatic drug, and pyrvinium pamoate, an anthelmintic agent, were identified and chosen from the list of 290 hits for in vitro and in vivo validation. We confirmed that in vitro treatment with auranofin and pyrvinium pamoate reduces AML cell growth through a cytotoxic and cytostatic effect, respectively. We identified the synergies/additivities of the two molecules with standard anti-AML drugs (e.g., cytarabine, doxorubicine) and a specific FLT3 inhibitor (quizartinib). Auranofin synergised with cytarabine and its effect was additive when combined with quizartinib; pyrvinium pamoate showed an additive effect when used with doxorubicin and quizartinib. Next, we determined that both auranofin and pyrvinium pamoate act through their described mechanism of action, i.e., inhibit thioredoxin reductase (auranofin) and Wnt signalling (pyrvinium pamoate). In addition, we identified a novel mechanism of action for the two agents: the induction of the endoplasmic reticulum stress and the unfolded protein response that follows. Our results support the potential of auranofin (less so in the case of pyrvinium pamoate) for the treatment of AML patients, including those with FLT3 mutations, provided that the ongoing in vivo validation is successful.

Las leishmaniasis, la tripanosomiasis Americana y Africana, y la malaria son enfermedades parasitarias que constituyen un importante problema de salud global que afecta mayoritariamente a países en desarrollo. El aumento del número de resistencias a sus tratamientos actuales, su toxicidad y la necesidad de asistencia sanitaria para la aplicación de los mismos reflejan la urgente necesidad de desarrollar vacunas eficaces y nuevos tratamientos económicos, fáciles de administrar y resistentes a condiciones de almacenamiento adversas. Basándonos en que estas enfermedades parasitarias comparten requerimientos metabólicos con patologías mejor estudiadas, proponemos el reposicionamiento de fármacos para tratarlas. Bajo esta premisa, seis fármacos de eficacia probada en la investigación contra el cáncer ―dicloroacetato (DCA), 3‐bromopiruvato (3BP), 2‐ deoxi‐D‐glucosa (2DG), lonidamina (LND), metformina (MET) y sirolimus (SIR)― fueron seleccionados por su habilidad para modular rutas metabólicas relacionadas con la producción de energía y proliferación. El objetivo de este estudio fue validar el uso de estos moduladores bioenergéticos para el control de la leishmaniasis visceral, malaria y tripanosomiasis americana y africana como tratamiento o como potenciadores de la protección de una vacuna frente a L. infantum. Para ello, se evaluó la eficacia de estos compuestos en modelos in vitro de cada parasito ―enfermedad de Chagas (Trypanosoma cruzi), tripanosomiasis Africana (Trypanosoma brucei), leishmaniasis visceral (Leishmania infantum) y malaria (Plasmodium falciparum)―. El 3BP y el DCA indujeron una reducción dosis‐dependiente del crecimiento de los amastigotes intracelulares de L. infantum con IC50 de 17.19 μM y 631.5 μM, respectivamente. En el modelo in vitro de T. brucei, todos los compuestos testados, a excepción de 2DG, afectaron a la viabilidad del parásito: DCA (IC50 = 1.24 mM), 3BP (IC50 = 76.57 μM), LND (IC50 = 26.76 μM), SIR (IC50 = 2.14 μM), y MET (IC50 = 17.30 Mm). En el caso de los amastigotes intracelulares de T. cruzi, DCA, 3BP, 2DG, LND, y MET tuvieron efecto parasiticida con valores de IC50 de 27.07 mM, 27.63 μM, 7.27 mM, 78.37 μM, y 18.48 mM, respectivamente. DCA (IC50 = 5.39 mM), 2DG (IC50 = 4.19 mM), LND (IC50 = 209.13 μM), MET (IC50 = 1.32 mM), y SIR (IC50 = 2.50 μM), mostraron efecto antiparasitario sobre trofozoitos de P. falciparum. Estos resultados sugieren que estos fármacos podrían ser útiles para tratar estas enfermedades parasitarias. Sin embargo, cuando los compuestos eficaces en los modelos in vitro fueron administrados en modelos in vivo de roedor para cada una de las enfermedades, ninguno de ellos contribuyó al control de la enfermedad o de la carga parasitaria. Los resultados obtenidos en el modelo de leishmaniasis visceral en hámster revelaron una disminución de la activación del sistema inmune en los animales tratados con DCA y 3BP, lo cual podría haber contribuido al fracaso del tratamiento. Por último, se estudió la capacidad del SIR para potenciar el efecto protector de una vacuna frente a la leishmaniasis visceral en el modelo hámster. Para ello se administró SIR durante la fase de expansión y contracción del sistema inmune producido por una vacuna de DNA portadora de los genes LACK, TRYP, PAPLE22, y KMPII de Leishmania, y se estudió la respuesta frente al posterior desafío con L. infantum. Los resultados muestran que la vacuna de DNA indujo la reducción eficaz de la carga parasitaria en piel (P = 0.0004) y linfonodos (P = 0.0452), lo cual potenció la administración del SIR alcanzándose también protección parasitológica en bazo (P = 0.0004). El estudio de los marcadores inmunológicos en dicho órgano sugiere que la producción controlada de IFN‐γ y el incremento en la expresión de FoxP3 podrían ser los responsables de la protección alcanzada.
Leishmaniases, African and American trypanosomiases and malaria are parasitic diseases that constitute a major global health problem. The increasing number of drug‐resistances to their current treatments, toxicity cases and the health assistance often required for their administration, makes it urgently necessary to develop efficient vaccines for humans and new affordable therapies, easy to apply and resistant to harsh storage conditions. Due to the fact that these diseases share similar metabolic requirements with better studied diseases, we chose drug repurposing as a potentially effective approach against them. With this purpose, six different compounds used in anti‐cancer research —dichloroacetate (DCA), 3‐bromopyruvate (3BP), 2‐deoxy‐D‐glucose (2DG), lonidamine (LND), metformin (MET), and sirolimus (SIR)— were selected according to their ability to modulate energy production and proliferation related metabolic pathways. The aim of this study was to validate the suitability of these bioenergetics modulators for the management of visceral leishmaniasis, malaria and African and American trypanosomiasis as a treatment, or as a preventive tool by enhancing the protective power of a vaccine against L. infantum. The effectiveness of these compounds was first evaluated on in vitro models of each parasite ― Chagas disease (Trypanosoma cruzi), human African trypanosomiasis (Trypanosoma brucei), visceral leishmaniasis (Leishmania infantum) and malaria (Plasmodium falciparum)―. L. infantum promastigotes were not susceptible to these compounds, whereas L. infantum intracellular amastigote growth was dose‐dependently reduced by 3BP (IC50 = 17.19 μM) and DCA (IC50 = 631.5 μM). In the T. brucei in vitro model all the tested compounds, with the exception of 2DG, affected parasite survival with IC50 values of 1.24 mM for DCA, 76.57 μM for 3BP, 26.76 μM for LND, 2.14 μM for SIR, and 17.30 mM for MET. In the case of T. cruzi, DCA, 3BP, 2DG, LND, and MET showed parasite‐killing activity with IC50 values of 27.07 mM, 27.63 μM, 7.27 mM, 78.37 μM, and 18.48 mM, respectively. For P. falciparum DCA (IC50 = 5.39 mM), 2DG (IC50 = 4.19 mM), LND (IC50 = 209.13 μM), MET (IC50 = 1.32 mM), and SIR (IC50 = 2.50 μM), showed antiplasmodial activity. These results reinforce the hypothesis that drugs with proven efficacy in the treatment of cancer by interfering with energy production might be useful in treating threatening parasitic diseases and provide new opportunities for their repurposing. However, when compounds that were effective in the in vitro approach were administered to the in vivo rodent models of these diseases, none of them contributed to disease management or parasite load control. Immunological analysis in the VL hamster model revealed a significant downregulation of immune‐activation in infected animals treated with DCA and 3BP, which may also contribute to treatment failure. In the last chapter of this work, the suitability of sirolimus as an immunomodulatory compound to boost the activity of a preventive vaccine against VL was analyzed. Sirolimus is an already marketed compound that has been described to boost immune protection against different disease models. In our study, Syrian hamsters were treated with sirolimus concomitantly with the administration of a plasmid DNA vaccine carrying the Leishmania genes LACK, TRYP, PAPLE22 and KMPII, and the subsequent response towards a L. infantum challenge was studied. Our results show that the DNA vaccine itself efficiently reduced the burden of parasites in skin (P = 0.0004) and lymph nodes (P = 0.0452), which was potentiated by SIR administration by also inducing parasitological protection in the spleen (P = 0.0004). The study of immune markers in spleen suggests that lower production of IFN‐γ and the concurrent increase of FoxP3+ expression may be responsible for the protection mediated by the DNA vaccine that was potentiated by sirolimus.

Drugs exert their (therapeutic) effects via molecular-level interactions with proteins and other biomolecules. Computational prediction of drug-protein interactions plays a significant role in the effort to improve our current and limited knowledge of these interactions. The use of the putative drug-protein interactions could facilitate the discovery of novel applications of drugs, assist in cataloging their targets, and help to explain the details of medicinal efficacy and side-effects of drugs. We investigate current studies related to the computational prediction of drug-protein interactions and categorize them into protein structure-based and similarity-based methods. We evaluate three representative structure-based predictors and develop a Protein-Drug Interaction Database (PDID) that includes the putative drug targets generated by these three methods for the entire structural human proteome. To address the fact that only a limited set of proteins has known structures, we study the similarity-based methods that do not require this information. We review a comprehensive set of 35 high-impact similarity-based predictors and develop a novel, high-quality benchmark database. We group these predictors based on three types of similarities and their combinations that they use. We discuss and compare key architectural aspects of these methods including their source databases, internal databases and predictive models. Using our novel benchmark database, we perform comparative empirical analysis of predictive performance of seven types of representative predictors that utilize each type of similarity individually or in all possible combinations. We assess predictive quality at the database-wide drug-protein interaction level and we are the first to also include evaluation across individual drugs. Our comprehensive analysis shows that predictors that use more similarity types outperform methods that employ fewer similarities, and that the model combining all three types of similarities secures AUC of 0.93. We offer a first-of-its-kind analysis of sensitivity of predictive performance to intrinsic and extrinsic characteristics of the considered predictors. We find that predictive performance is sensitive to low levels of similarities between sequences of the drug targets and several extrinsic properties of the input drug structures, drug profiles and drug targets.

The number of known monogenic rare diseases (~7000) exceeds the number of effective treatments (~500) by more than an order of magnitude underlining the pressing need for generalizable therapeutic approaches for this class of conditions. In this regard, the majority of recessive and x-linked recessive disorders are caused by missense mutations encoding proteins that frequently have residual function but are rapidly degraded by the 26S proteasome. Bortezomib is a small molecule that inhibits the 26S proteasome and has been approved for use in patients for an unrelated condition; multiple myeloma. Previous work has shown that, for a small number of disorders, bortezomib can inhibit the degradation of the mutant protein, thereby increasing the protein level and activity, holding out the promise of a beneficial therapeutic effect by the repurposing of this agent. We present here a high level western blot based survey of nine recessive disorders to characterize the general effectiveness of such an approach. Thirteen patient fibroblast cell lines comprising 9 different diseases with 19 known mutations were selected on the basis of missense mutations protein expression data when available. The cell lines were incubated with bortezomib (10 nM and 50 nM; 24 hrs) and levels of the mutated protein were quantified by western blot. Unfortunately, no consistent, appreciable increase was observed for any of the conditions tested. The general therapeutic value of re-purposing bortezomib for recessive and x-linked diseases appears limited at best. The few reported cases of bortezomib successfully working in increasing mutated protein levels appear to be the exceptions and not the norm. Moreover successes are more often limited to cell lines carrying a transgene expressing the mutated protein rather than endogenous mutated protein in patient cell lines.

Down syndrome (DS) is the main genetic cause of intellectual disability. Decreased proliferation of Neural Progenitor Cells (NPC), widespread neurogenesis impairment and increased astrogliogenesis are considered among the major determinants of brain atrophy and intellectual disability in DS individuals. The best characterized and studied animal model for DS is the Ts65Dn mouse line which recapitulates several features of the human pathology, including cognitive impairment. In the recent years studies suggested that perinatal pharmacotherapies targeting NPC alterations may represent potential interventions in DS. However, at present, no pharmacotherapies are suitable for clinical application. Thus, the need to identify new and safe pharmacotherapies to improve intellectual disability in DS patients. Based on these data, our overall goal was to unravel novel mechanisms underlying We generated and phenotypically characterized NPC from Ts65Dn and euploid pups (P1, 2). We then targeted NPC alterations using at first a phenotypic-based approach that identified 30 FDA/EMA approved drugs able to correct trisomic NPC defective proliferation. Importantly, among the potential hits we identified the immunosuppressant cyclosporine A (CSA). We showed that a neonatal treatment with CSA (P3-P15) corrected the whole triad of defects of DS brain. In parallel, we used a target based approach, exploiting the effect of an agonist of the tropomyosin receptor kinase B, 7,8-DHF, and corn oil on neurogenesis in vitro and in vivo, also evaluating the cognitive performances. In the last part of this thesis we identified new molecular mechanisms altered in trisomic NPC in response to the key astrocytic signal thrombospondin-1. Taken together these data showed that trisomic NPC dysfunctions are pharmacologically relevant targets in DS.

Drug repurposing through fasting could pinpoint new cancer cell liabilities and define new treatment options. By screening over 800 approved drugs in PK9 pancreatic ductal adenocarcinoma (PDAC) cells, we identified azoles, inhibitors of cholesterol biosynthesis, as agents whose cytotoxic activity against cancer cells was synergistically enhanced by starvation conditions. We hypothesized that starvation and azoles would cooperate by blunting cholesterol production in PDAC cells. Interestingly, we found other cholesterol inhibitors had their antitumor effects strongly enhanced by starvation. In addition, combined starvation and cholesterol inhibitors reduced tumor growth in gastrointestinal mouse models and intracellular cholesterol levels both in vitro and in vivo. Furthermore, methyl-beta-cyclodextrin, which depletes intracellular cholesterol, and starvation showed a synergistic cytotoxic effect against Capan-1 and MIA PaCa-2 cells (another PDAC cell line). Adding back cholesterol or LDL in gastrointestinal tumor cells/xenografts prevented the synergistic interaction between starvation and cholesterol inhibitors. Combined therapy inhibited pAKT (partially restored with LDL add-back) whereas, its antitumor activity was abolished by the simultaneous supplementation with circulating growth-promoting factors (IGF1, insulin, and leptin). In addition, combined therapy reduced mitochondrial oxidative phosphorylation (OXPHOS) and energy status in gastrointestinal tumor xenografts, whereas cholesterol restoration abolishes this effect. Taken together, these findings support the rationale for conducting clinical studies to assess the safety, feasibility, and activity of combining periodic cycles of fasting with inhibitors of cholesterol biosynthesis in cancer patients.

La transtirretina (TTR) és una proteïna tetramèrica amiloidogènica (55 kDa) present al plasma humà i responsable del transport de la hormona T4 i del retinol a través de la proteïna d’unió a retinol (RBP). La proteïna TTR està associada amb diverses amiloïdosis, concretament la polineuropatia amiloide familiar (FAP), la cardiomiopatia amiloide familiar (FAC) i l’amiloïdosi senil sistèmica (SSA). La variabilitat associada a la TTR es deu tant a mutacions puntuals al gen codificant per aquesta com a modificacions post-traduccionals (PTMs) al residu Cys-10. Les PTMs més comuns associades a la Cys-10 de la TTR són la S-Sulfonació (S-Sulfo), la S-Glicinilcisteinilació (S-CysGly), la S-Cisteinilació (S-Cys) i la S-Glutationilació (S-GSH). Es creu que dites PTMs associades a la Cys-10 podrien jugar un paper biològic important en l’inici i procés patològic de les diferents amiloïdosis lligades a TTR. Hem tractat les amiloïdosis lligades a TTR des de dues perspectives diferents i) Intervencions terapèutiques i ii) Diagnòstic i monitorització de FAP. Referent a la primera part del projecte, hem portat a terme el cribratge de 41 possibles inhibidors de fibril·logènesis seleccionats mitjançant estratègies bioinformàtiques de repurposing de fàrmacs. Com a resultat de l’estudi, s’han trobat 4 nous estabilitzadors del tetràmer de TTR i, per tant, nous candidats pel tractament d’amiloïdosis lligades a TTR. Pel que fa a l’aproximació diagnòstica d’aquest treball, hem desenvolupat una metodologia per a la quantificació de PTMs en mostres de sèrum, així com per a la determinació dels nivells de TTR en aquest, tant en individus sans (wt) com en individus portadors de TTR amiloïdogènica (mutació V30M). Dita metodologia consisteix en una primera etapa d’enriquiment en TTR mitjançant immunoprecipitació, seguit de l’anàlisi de la TTR per espectrometria de masses de i) la proteïna intacta i ii) els pèptids de TTR portadors de les PTMs d’interès mitjançant l’anàlisi dirigit per LC-MS. L’anàlisi de les mostres de sèrum per la combinació d’ambdues estratègies aporta informació sobre la quantificació relativa i absoluta de les diferents PTMs presents a la TTR. Ha sigut possible mostrar que els mètodes basats en proteïna intacta es troben esbiaixats per algunes de les PTMs, donat que assumeixen un factor de resposta constant per les diferents isoformes. Contràriament, el nou mètode de LC-MS dirigit permet la quantificació absoluta de les diferents PTMs i els nivells totals de TTR (wt i mutant). La metodologia reportada ha sigut aplicada en l’anàlisi de dos grups de mostres clíniques. Com a resultat de l’estudi de mostres humanes de pacients de FAP en els diferents estadis de la malaltia, suggerim de forma preliminar les isoformes S-GSH i S-CysGly com a biomarcadors de progressió de la malaltia, permetent la diferenciació entre pacients en estadi 0 i 1 i, per tant, indicant l’aparició de la malaltia. Mitjançant l’anàlisi de mostres de pacients de FAP a diferents temps després de sotmetre’s a un transplantament de fetge (LT) i de pacients receptors de transplantament de fetge dominó provinent d’individus portadors de la mutació V30M, hem caracteritzat la progressió de la relació wt:V30M, així com l’evolució dels nivells de PTMs a la Cys-10, des de la intervenció fins a 9 anys després. Addicionalment, hem observat diferències significatives en els nivells de S-GSH i S-CysGly en comparar pacients de LT i DLT, resultats anàlegs als obtinguts en la comparació d’individus wt (sans) i pacients de FAP en estadi 0.
La transtirretina (TTR) es una proteína tetramérica amiloidogénica (55 kDa) presente en el plasma humano y la responsable del transporte de la hormona T4 y el retinol, a través de la proteína de unión al retinol (RBP). La proteína TTR está asociada con varias amiloidosis, concretamente la polineuropatía amiloide familiar (FAP), la cardiomiopatía amiloide familiar (FAC) y la amiloidosis senil sistémica (SSA). La variabilidad encontrada en la TTR se debe tanto a mutaciones puntuales encontradas en el gen que codifica para ésta como a modificaciones post-traduccionales (PTMs) en el residuo Cys-10. Las PTMs más comunes asociadas a la Cys-10 de la TTR son la S-Sulfonación (S-Sulfo), la S-Glicinilcisteinilación (S-CysGly), la S-Cisteinilación (S-Cys) y la S-Glutationilación (S-GSH). Se cree que dichas PTMs asociadas a la Cys-10 podrían jugar un papel biológico importante en el inicio y proceso patológico de las distintas amiloidosis ligadas a TTR. Hemos abordado las amiloidosis ligadas a TTR desde dos perspectivas distintas i) Intervenciones terapéuticas y ii) Diagnóstico y monitorización de FAP. Respecto a la primera parte del proyecto, hemos llevado a cabo el cribado de 41 posibles inhibidores de fibrilogenesis seleccionados mediante estrategias bioinformáticas de repurposing de fármacos. De este modo, se han encontrado 4 nuevos estabilizadores del tetrámero de TTR y por tanto, nuevos candidatos para el tratamiento de amiloidosis ligadas a TTR. En relación a la aproximación diagnóstica de este trabajo, hemos desarrollado una metodología para la cuantificación de PTMs en muestras de suero, así como para la determinación de los niveles de TTR en éste, tanto en individuos sanos (wt) como en individuos portadores de TTR amiloidogénica (mutación V30M). Dicha metodología consiste en una primera etapa de enriquecimiento en TTR mediante immunoprecipitación, seguido por el análisis de ésta mediante espectrometría de masas de i) la proteína TTR intacta y ii) de los péptidos de TTR portadores de las PTMs de interés mediante análisis dirigido por LC-MS. El análisis de muestras de suero mediante la combinación de ambas estrategias aporta información sobre la cuantificación relativa y absoluta de las distintas PTMs en TTR. Ha sido posible mostrar que los métodos basados en proteína intacta se encuentran sesgados para algunas de las PTMs, dado que asumen un factor de respuesta constante para las distintas isoformas. Por el contrario, el nuevo método de LC-MS dirigido permite la cuantificación absoluta de las distintas PTMs y los niveles totales de TTR (wt y mutante). La metodología reportada ha sido aplicada en el análisis de dos grupos de muestras clínicas. Como resultado del estudio de muestras humanas de pacientes de FAP en los distintos estadios de la enfermedad, sugerimos de forma preliminar las isoformas S-GSH y S-CysGly como biomarcadores de progresión de la enfermedad, permitiendo la diferenciación entre pacientes en estadio 0 y 1 y, por lo tanto, indicando la aparición de la enfermedad. Mediante el análisis de muestras de pacientes de FAP a distintos tiempos después de someterse a un trasplante de hígado (LT) y de pacientes receptores de trasplante de hígado dominó proveniente de individuos portadores de la mutación V30M, hemos caracterizado la progresión del ratio wt:V30M así como la evolución de los niveles de PTMs en la Cys-10, des de la intervención hasta 9 años después. Adicionalmente, hemos observado diferencias significativas en los niveles de S-GSH y S-CysGly en comparar pacientes de LT y DLT, resultados análogos a los obtenidos en la comparación de individuos wt (sanos) y pacientes de FAP en estadio 0.
Transthyretin (TTR) is an amyloidogenic tetrameric protein (55kDa) present in human plasma, transporting T4 hormone and retinol, through the retinol binding protein (RBP). TTR is associated with several amyloidosis, namely familial amyloidotic polyneuropathy (FAP), familial amyloidotic cardiomyopathy (FAC) and senile systemic amyloidosis (SSA). Variability of TTR is not only due to point mutations in the encoding gene but also to post-translational modifications (PTMs) at Cys-10, the most common PTMs being the S-Sulfonation (S-Sulfo), S-Glycinylcysteinylation (S-CysGly), S-Cysteinylation (S-Cys) and S-Glutathionylation (S-GSH). It is thought that PTMs at Cys-10 may play an important biological role in the onset and pathological process of amyloidosis related to TTR. We have aimed TTR amyloidosis from two different perspectives i) Therapeutic interventions and ii) FAP diagnosis and monitoring. Regarding the first branch of the project, we have performed the screening of a library of 41 possible fibrillogenesis inhibitors selected by a bioinformatic repurposing workflow, finding 4 new TTR tetramer stabilizers and thus, new potential candidates for TTR amyloidosis treatment. Concerning the clinical approach of this work, we have developed a methodology for quantification of PTMs in serum samples, as well as for the determination of serum TTR levels, from healthy (wt) and TTR-amyloidotic (V30M mutation) individuals. It involves an enrichment step by immunoprecipitation followed by mass spectrometry analysis of (i) the intact TTR protein and (ii) targeted LC-MS analysis of peptides carrying the PTMs of interest. Analysis of serum samples by the combination of the two methods affords complementary information on the relative and absolute amounts of the selected TTR PTM forms. It is shown that methods based on intact protein are biased for specific PTMs since they assume constant response factors, whereas the novel targeted LC-MS method provides absolute quantification of PTMs and total TTR variants. The reported methodology has been applied to two different sets of clinical samples. As a result of the study of human samples of FAP patients at different disease stages, we preliminary pointed out S-GSH and S-CysGly isoforms as biomarkers of disease progression, allowing the differentiation between FAP stage 0 and 1 and therefore indicating disease onset. Through the analysis of a time series from FAP patients having undergone liver transplantation (LT) and from domino liver transplantation (DLT) recipients from V30M carriers, we have characterized the progression of the wt:V30M ratios, as well as the evolution of the Cys-10 PTMs, from transplantation and up to 9 years after. Additionally, we have observed significant differences in the levels of S-GSH and S-CysGly when comparing liver and domino liver transplanted patients, analogous to the results obtained in the comparison of wt individuals and FAP stage 0 patients.

Philosophiae Doctor - PhD
Glioblastoma multiforme (GBM) remains the most malignant of all primary adult brain tumours. It is a highly invasive and vascularized neoplasm with limited treatment options and very low survival rate. GBM tumours are heterogeneous in nature with cellular hierarchy and at the apex of this hierarchy are the glioblastoma stem cells, known to promote tumourigenesis and resistance to chemotherapeutic agents and tumour recurrence. Currently, the standard care for GBM involves surgical resection, radiation, and chemotherapy treatment with temozolomide. Unfortunately, median survival after treatment is still daunting and tumour relapse is very frequent. Indeed, patients with recurrent glioblastoma have less than a year survival. To address this, novel therapies need to be developed with the early introduction of promising agents into clinical trials and subsequent approval for use. Importantly, for these novel therapies to be approved for GBM, they need to be safe, effective as well as being able to penetrate the blood-brain barrier (BBB). Due to the high cost and process time for the development of new drugs, existing approved drugs are currently being repurposed for new indications and this is gaining significance in clinical pharmacology as it allows rapid delivery of useful drugs from bench to bedside. Drugs of the antipsychotic class are well known to cross the BBB due to their neuroleptic action. To this end, the aim of this study was to identify and characterize the anti-cancer activities of novel phenothiazine-derivatives belonging to the antipsychotic class of drugs in glioblastoma. To achieve this, several novel phenothiazine-derivatives were initially screened for possible anti-cancer activity in the U87 and U251 malignant GBM cells. Two lead compounds, DS00326 and DS00329, were identified and their anti-cancer activities were determined in U87 and U251 cells as well as in primary patient-derived xenograft (PDX) glioblastoma cultures. DS00326 and DS00329 significantly inhibited glioblastoma cell viability, with minimal effects observed in the non-cancerous FG0 fibroblasts. The IC50 values of DS00326 and DS00329 for U251, U87 and PDX cells ranged from 1.61 to 12.53μM. Flow cytometry analyses showed that DS00326 and DS00329 treatment led to an increase in the G1 population of cells. Additionally, DS00326 and DS00329 induced double-strand DNA breaks, which lead to activation of the canonical DNA damage response pathway. Furthermore, DS00326 and DS00329 induced apoptosis as shown by morphological markers, flow cytometry with annexin V-FITC/propidium iodide staining, as well as western blotting with an antibody to detect levels of cleaved PARP. Interestingly, treatment with DS00326 and DS00329 also induced autophagy as evident by the increase of acidic vesicular organelles in cells following staining with acridine orange as well as an increase in levels of the autophagy marker LC3-II. Autophagy was seen as a pro-death pathway in the U87 and U251 cells as inhibition of autophagy led to a reversal of cytotoxicity and consequently increased cell survival. Moreover, it was demonstrated that DS00326 and DS00329 inhibited the PI3/Akt pathway while modulating the mitogen-activated protein kinases p38, ERK1/2 and JNK signalling pathways. Importantly DS00326 and DS00329 displayed anti-cancer stem cell activities by the inhibition of neurosphere formation and regulation of stem cell markers SOX2 and GFAP in PDX cells. Together, the findings from this study suggest that DS00326 and DS00329 may be effective in the treatment of glioblastoma and provide a strong rationale for further clinical studies exploiting phenothiazines and their derivatives as treatments for glioblastoma.
2021-09-01

This PhD thesis combines four chapters on different fields of basic research and sets the focus on two circulating viruses of global concern, the orthomyxovirus influenza A virus (IAV) and the alphavirus Ross River virus (RRV). The first three chapters include swine influenza A virus (sIAV) surveillance for the detection and characterisation of IAV subtypes, an in vitro high throughput screening (HTS) on host micro RNAs (miRNAs) for the discovery of novel anti-IAV (H7N9) targets and their underlying mechanisms, and an approach to reduce disease pathogenesis in mice infected with H7N9 by targeting the pro-inflammatory factor CCL2. In a fourth chapter, drug repurposing with the interleukin-1 (IL-1) inhibitor anakinra was investigated to treat RRV-induced bone loss in mice. By combining these four chapters, a broad range of drug discovery is covered in this PhD thesis; Surveillance, HTS target discovery and the application of drug repurposing in animal models of viral diseases.
Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
Institute for Glycomics
Science, Environment, Engineering and Technology
Full Text

Les infections fongiques invasives constituent un sérieux problème de santé publique dans le monde ; cette situation se complique par la disponibilité d’un faible nombre d’antifongiques utilisés en pratique clinique. De ce fait,1280 molécules médicamenteuses, constituant la chimiothèque Prestwick, ont été testées sur des souches de champignons multirésistants d’intérêt clinique, isolées à l’Hôpital la Timone de Marseille. Le criblage primitif à une concentration de 10 µM avait permis l’identification de plusieurs molécules capables d’inhiber la croissance fongique.Par la suite, on s’est focalisé sur deux molécules médicamenteuses : la colistine et la ribavirine. Les concentrations minimales inhibitrices de ces dernières ont été déterminées, de même que leur activité fongicide ou fongistatique sur une large collection de souches. Des combinaisons synergiques avec les antifongiques habituels ont été mises au point notamment celles de la ribavirine en association avec l’amphotéricine B, l’itraconazole et le voriconazole qui sont actives sur les souches de Candida albicans multirésistantes. Le but de notre troisième travail a été de comprendre le mécanisme d’action de la ribavirine, un antiviral, sur les Candida albicans et d’identifier sa potentielle cible. Pour se faire, les analogues des cibles de la ribavirine chez le virus de l’hépatite C, retrouvés chez les Candida albicans notamment les enzymes inosine-5’-monophosphate déshydrogénase (IMPDH) et l’ARN polymérase ont été étudiés. Des systèmes PCR et séquençage ont été développés pour détecter et analyser les gènes IMH3 et RPO21 qui codent pour les enzymes IMPDH et ARN polymérase respectivement chez les Candida
The increasing incidence of invasive infections caused by pathogenic fungi is a major worldwide concern; a serious situation to which the limited number of available effective antifungals to face it, is another problem. Hence, there is a constant need for other compounds that possess antifungal properties. So by applying drug-repurposing approach, Prestwick Chemical Library containing 1,280 compounds previously approved by the FDA was tested against multidrug-resistant fungi recovered from La Timone Hospital in Marseille. Primary FDA approved drugs screening at fixed concentration of 10 µM, allowed us to identify several fungal growth inhibitors.Among these non-standard antifungals, we focused our study on both colistin and ribavirin drugs. Minimum inhibitory concentrations of these compounds were determined against a large collection of strains, and time-kill curves were performed to establish their fungicidal or fungistatic activity. Moreover, synergistic combinations with the current antifungal agents were examined; notably, association of ribavirin with either amphotericin B, itraconazole or voriconazole active against multidrug-resistant Candida albicans. The aim of the third part of our work was to identify the mechanism of action of ribavirin, an antiviral compound, on Candida albicans and its potential target. So, we focused our work on the analogue of ribavirin target in hepatitis C virus, present in Candida albicans namely inosine-5'-monophosphate dehydrogenase (IMPDH) and RNA polymerase enzymes. We designed PCRs and sequencing systems to detect and analyse IMH3 and RPO21 genes that encode IMPDH and RNA polymerase enzymes respectively

La tradizionale scoperta di farmaci è un processo lungo e costoso, spesso ostacolato da alti tassi di fallimento. Una valida alternativa è il riposizionamento dei farmaci (drug repurposing), definito come l'identificazione di nuove indicazioni terapeutiche per farmaci già noti o candidati farmaci, prodotti sintetici e naturali. Il riposizionamento dei farmaci permette di ridurre i tempi, i rischi e i costi associati alle tradizionali procedure di scoperta, poiché la maggior parte dei composti ha in molti casi già superato studi di sicurezza e tossicità. Inoltre, l'aumento dei dati biologici, clinici e chimici ha creato nuove opportunità per il riposizionamento dei farmaci. Perciò, l'uso su larga scala di approcci in silico ha dimostrato di essere una strategia efficiente e conveniente. Tuttavia, ad oggi rimane un forte bisogno di protocolli razionali e nuove metodologie per aiutare i ricercatori in questo campo. Sulla base di queste premesse, l'obiettivo del progetto di dottorato è stato focalizzato su due aree principali del riposizionamento dei farmaci in silico: i) l'applicazione di protocolli su misura per specifiche campagne di riposizionamento; e ii) lo sviluppo di nuovi metodi e approcci generali. Durante il corso di dottorato, sono state messe a punto molteplici applicazioni di protocolli integranti diversi approcci computazionali per il riposizionamento di prodotti di origine sia naturale che sintetica. Il data mining da noti database pubblici ha permesso di eseguire valutazioni della similarità 2D e 3D, e di selezionare target per eseguire studi di docking molecolare. Ogni protocollo è stato personalizzato tenendo conto delle caratteristiche delle molecole sotto studio. Infine, i test in vitro su proteine o cellule isolate hanno permesso di convalidare sperimentalmente le previsioni. Parallelamente, il progetto di dottorato è stato incentrato sullo sviluppo di protocolli innovativi in grado di fornire nuove risorse. Per esempio, è stata sviluppata una piattaforma basata sulle tecniche di machine learning (ML) per prevedere il profilo di selettività tra diverse isoforme enzimatiche. Inoltre, è stato realizzato il sito web LigAdvisor, una piattaforma integrata per il repurposing e la polifarmacologia. I progetti implementati hanno fornito risultati molto soddisfacenti. Infatti, nel corso dei tre anni è stato possibile: riposizionare una libreria di composti di origine sintetica, identificando un potente inibitore della Anidrasi Carbonica (hCA) II, e poi progettare suoi derivati con attività duale su hCA e sui Recettori degli Estrogeni (ER); riposizionare prodotti naturali su ERβ; identificare candidati per l'inibizione della proteasi principale (Mpro) di SARS-CoV-2. La piattaforma di screening tramite machine learning ha fornito eccellenti prestazioni predittive, migliori di quelle ottenute con altri approcci tradizionali. Infine, lo sviluppo del sito web LigAdvisor, liberamente accessibile, permette anche ai non esperti del settore di reperire una grande quantità di dati di alta qualità e di eseguire una varietà di ricerche. In conclusione, i risultati riportati in questa tesi dimostrano come l'uso di approcci computazionali, intelligenza artificiale e tecniche di data mining sia realmente utile nella progettazione di campagne di riposizionamento. Uno degli aspetti innovativi dei progetti realizzati è infatti rappresentato dall'integrazione di diversi metodi consolidati in nuovi protocolli e piattaforme, aumentando così la loro usabilità e migliorando le possibilità di sviluppare campagne di riposizionamento di successo. I dati qui presentati sono stati oggetto di molteplici pubblicazioni su riviste internazionali, e le nuove piattaforme proposte sono state rese disponibili al pubblico.
Traditional de novo drug discovery is a long, expensive process that is often hampered by high failure rates. A viable alternative strategy is drug repurposing (or drug repositioning), defined as the identification of novel therapeutic indications for already known drugs or drug candidates, as well as synthetic and natural products. Drug repurposing allows to reduce times, risks and costs associated with traditional de novo discovery pipelines, as most compounds have in many cases already passed safety and toxicity studies. Moreover, in recent times the increase in biological, clinical and chemical data has enabled the progress of novel attractive drug repurposing opportunities. Accordingly, the large-scale use of integrated in silico approaches has proven to be an efficient and cost-effective strategy. However, to date there is still a large need for rational protocols and new methodologies to help researchers in this field. Based on these premises, the aim of the PhD project was focused on two main areas of in silico drug repurposing: i) the application of tailored protocols for specific repositioning campaigns; and ii) the development of novel methods and general approaches. During the PhD course, several applications of integrated protocols using different computational approaches were developed for the repositioning of products of both natural and synthetic origin. Data mining from well-known public databases allowed to perform 2D and 3D similarity estimations, and to select appropriate targets to perform in-depth molecular docking studies. Each protocol was customized taking into account the specific characteristics of the molecules under study. Finally, in vitro testing on isolated proteins or cells allowed to experimentally validate the predictions. At the same time, the PhD project focused on the development of innovative protocols capable of providing new assets to researchers working with drug repurposing. For instance, a machine learning (ML) based platform was developed to predict selectivity profiles across different enzyme isoforms. Moreover, the development of the LigAdvisor website, an integrated platform for repurposing and polypharmacology, was also carried out. The implemented projects provided highly satisfactory results. Indeed, over the three years it was possible to: reposition a library of compounds of synthetic origin, by identifying a potent human Carbonic Anhydrase (hCA) II inhibitor, and then design derivatives with dual activity on hCA and the Estrogen Receptors (ER); to reposition natural products on ERβ; to identify candidates for the inhibition of the SARS-CoV-2 main protease (Mpro). The machine learning hCA screening platform provided excellent predictive performances, which remarkably proved to be better than those obtained by other traditional approaches. Finally, the development of the freely accessible LigAdvisor website provides also non-experts in the field to retrieve a large amount of high quality data and perform a variety of different queries. In conclusion, the results reported in this dissertation demonstrate how the use of computational approaches, artificial intelligence and data mining techniques is indeed of great help in the rational design of repurposing campaigns and useful resources. One of the innovative aspects of the projects carried out is indeed represented by the integration of different established methods in new protocols and platforms, thus increasing their usability and improving the chances of developing successful repositioning campaigns. The data featured here were the subject of multiple publications in international journals, and the novel proposed platforms were made available to the public.

El funcionament intern de les cèl·lules pot entendre’s com un conjunt d’interaccions entre biomolècules, formant una xarxa que coneixem amb el nom d’interactoma. Els fàrmacs i malalties poden considerar-se pertorbacions d’aquesta xarxa, modulant directament molècules específiques, però indirectament comunitats de molècules les interaccions de les quals es veuen afectades per la pertorbació. La medicina de xarxes busca representar i analitzar amb precisió les xarxes biològiques, per tal d’entendre millor les malalties i aconseguir tractaments més segurs i eficaços. En aquesta tesi, presento diverses eines in silico basades en la medicina de xarxes, pensades per estudiar els mecanismes moleculars de malalties i fàrmacs. Aquestes eines s’utilitzen en un ampli ventall d’aplicacions de la medicina de xarxes, com per exemple l’estudi de comorbiditats, endofenotips, efectes secundaris, reutilització i combinació de fàrmacs.

L’instabilité d’ADN mitochondrial (ADNmt) peut être quantitative avec la déplétion de l’ADNmt ou qualitative avec des délétions de l’ADNmt. Ces anomalies sont une des causes les plus commmunes des maladies mitochondriales. Un des gènes qui contrôle la stabilité et le maintien de l’ADNmt est POLG. Ce gène code pour la polymerase gamma mitochondriale. Chez l’homme, les mutations dans le gène POLG sont liées aux maladies mitochondriales telle que; l’insuffisance hépatique, le syndrome d’Alpers, le PEO ou Progressive External Ophtalmoplegia, la neuropathie sensorielle et l’ataxie. Des mutations dans le gène POLG sont aussi associées au syndrome de Parkinson. Aujourd’hui, il n’existe aucune thérapie pour ces maladies. Compte tenu de la conservation évolutive de la fonction mitochondriale de la levure à l’homme, nous avons utilisé deux organismes modèles, Saccharomyces cerevisiae et Caenorhabditis elegans, pour identifier des molecules chimiques capables de compenser l’instabilité de l’ADNmt liée à des mutations du gène POLG dans des fibroblastes d’un patient. Nous avons trouvé trois molécules candidates potentielles: MRS2, MRS3 et MRS4, à partir d’un criblage primaire chez la levure, en utilisant une chimiothèque d’environ 2000 molécules chimiques. MRS3 est la molécule candidate la plus efficace pour la stabilization d’ADNmt chez des mutants POLG de la levure, du champignon filamenteux, du nématode et sur des fibroblastes de patients. MRS3, ou clofilium tosylate (CLO), est un agent antiarrhytmique, médicament pour soigner les troubles du rythme cardiaque. Dans cette étude, nous avons aussi montré que deux autres antiarrhythmiques appartenant à la même classe que CLO avaient un effet positive chez un mutant POLG de C. elegans. En utilisant une approche de chemogénomique chez la levure, nous avons identifié Fis1, un acteur de la fission mitochondriale qui pourrait être impliqué dans la mode d’action de CLO. Fis1 est requise pour la viabilité cellulaire en concentration légèrement toxique de CLO et nécesaire pour la stabilization de l’ADNmt par CLO. L’ensemble de ces résultats ont montré que CLO pourrait être la première molécule chimique qui stimule la réplication de l’ADNmt et qui pourrait être développée pour le traitement des maladies liées à des mutations dans le gène POLG. Ces résultats ont aussi permis de mettre en évidence une nouvelle connexion entre replication de l’ADNmt et la fission mitochondriale
The instability of mitochondrial DNA (mtDNA) in form of mtDNA depletion (quantitative instability) or large deletion (qualitative instability) is one of the most common cause of mitochondrial diseases.. One of the genes responsible for human mtDNA stability, POLG, is exploited in this study. POLG encodes the human mitochondrial polymerase gamma. In human, POLG mutations are a major cause of mitochondrial disorders including hepatic insufficiency; Alpers syndrome, progressive external ophthalmoplegia, sensory neuropathy and ataxia. They are also associated with Parkinsonism. Currently, there is no effective and disease-specific therapy for these diseases. Based on the conservation of mitochondrial function from yeast to human, we used Saccharomyces cerevisiae and Caenorhabditis elegans as first pass filters to identify chemical compounds that suppresses mtDNA instability in cultured fibroblasts of a POLG-deficient patient. We found three potential candidates, MRS2, MRS3 and MRS4, from a chemical screening of nearly 2000 compounds in yeast. MRS3 is the most efficacious in stabilizing mtDNA in yeast, filamentous fungi, worm and patient fibroblasts. This unsuspected compound, clofilium tosylate (CLO), belongs to a class of antiarrhythmic agents for cardiovascular disease. Two other antiarrhythmic agents (FDA-approved) sharing common pharmacological properties and chemical structure with CLO also show potential benefit for POLG deficiency in C. elegans. Using a chemogenomic approach in yeast, we also discovered that a mitochondrial fission actor Fis1 is implicated in the mechanism of action of CLO. Fis1 is important for cellular viability in a slightly toxic concentration of CLO and is required for the mtDNA stabilizing potency of CLO. Our findings provide evidence of the first mtDNA-stabilizing compound that may be an effective pharmacological alternative for the treatment of POLG-related diseases and uncover a new connection between the mitochondrial fission process and mtDNA replication
Ketidakstabilan DNA mitokondria (mtDNA) dalam bentuk pengurangan kopi mtDNA di dalam sel (ketidakstabilan kuantitatif), atau pun dalam bentuk delesi pada sekuens mtDNA (ketidakstabilan kualitatif) merupakan salah satu penyebab penyakit mitokondria. Salah satu gen yang bertanggung jawab dalam menjamin kestabilan mtDNA adalah POLG. Gen POLG mengkode protein polimerase gamma pada manusia, yang mereplikasi dan mereparasi mtDNA di dalam mitokondria. Mutasi pada gen POLG dapat menyebabkan penyakit kelainan mitokondria pada manusia, seperti gagal ginjal, sindrom Alpers, Progressive External Ophtalmoplegia, neuropati sensorial, ataxia dan bisa dikaitkan dalam beberapa gejala Parkinsonisme. Saat ini, belum ada terapi obat yang dapat mengatasi penyakit – penyakit tersebut. Berdasarkan kesamaan evolutif dari ragi hingga manusia, pada studi ini kami menggunakan Saccharomyces cerevisiae dan Caenorhabditis elegans untuk mengidentifikasi molekul obat yang berpotensi mengatasi ketidakstabilan mtDNA dari fibroblas pasien manusia yang memiliki mutasi gen POLG. Kami mengidentifikasi tiga kandidat potensial, yakni MRS2, MRS3 dan MRS4 dari penapisan kurang lebih 2000 molekul obat dengan menggunakan ragi. MRS3 adalah kandidat yang paling berkhasiat dan mampu mengatasi ketidakstabilan mtDNA pada ragi, Podospora, cacing dan fibroblas manusia. MRS3 adalah alias bagi clofilium tosylate (CLO), sebuah molekul antiaritmia untuk penyakit kardiovaskuler. Pada studi ini, kami juga menguji aktifitas dua molekul antiaritmia lain yang tergabung dalam kelas yang sama dengan CLO, dan menemukan bahwa kedua molekul ini juga berpotensi mengatasi defisit POLG pada cacing C. elegans. Dengan menggunakan metode kemogenomik pada ragi, kami juga mengidentifikasi sebuah aktor prosesus pembelahan mitokondria, Fis1, yang berpotensi terlibat dalam mekanisme seluler CLO. Fis1 dibutuhkan untuk: (1) kelangsungan hidup ragi pada konsentrasi toksik CLO dan (2) efek CLO dalam menstabilkan mtDNA pada ragi. Keseluruhan studi ini membuktikan potensi CLO sebagai molekul penstabil mtDNA yang pertama, yang dapat dikembangkan sebagai salah satu alternatif terapi obat untuk penyakit – penyakit mitokondria terkait mutasi POLG. Melalui studi ini, juga diungkap adanya hubungan antara kestabilan mtDNA dan prosesus pembelahan mitokondria

Résumé : Les méthodes de détection de similarités de sites de liaison servent entre autres à la prédiction de fonction et à la prédiction de cibles croisées. Ces méthodes peuvent aider à prévenir les effets secondaires, suggérer le repositionnement de médicament existants, identifier des cibles polypharmacologiques et des remplacements bio-isostériques. La plupart des méthodes utilisent des représentations basées sur les atomes, même si les champs d’interaction moléculaire (MIFs) représentent plus directement ce qui cherche à être identifié. Nous avons développé une méthode bio-informatique, IsoMif, qui détecte les similarités de MIF entre différents sites de liaisons et qui ne nécessite aucun alignement de séquence ou de structure. Sa performance a été comparée à d’autres méthodes avec des bancs d’essais, ce qui n’a jamais été fait pour une méthode basée sur les MIFs. IsoMif performe mieux en moyenne et est plus robuste. Nous avons noté des limites intrinsèques à la méthodologie et d’autres qui proviennent de la nature. L’impact de choix de conception sur la performance est discuté. Nous avons développé une interface en ligne qui permet la détection de similarités entre une protéine et différents ensembles de MIFs précalculés ou à des MIFs choisis par l’utilisateur. Des sessions PyMOL peuvent être téléchargées afin de visualiser les similarités identifiées pour différentes interactions intermoléculaires. Nous avons appliqué IsoMif pour identifier des cibles croisées potentielles de drogues lors d’une analyse à large échelle (5,6 millions de comparaisons). Des simulations d’arrimage moléculaire ont également été effectuées pour les prédictions significatives. L’objectif est de générer des hypothèses de repositionnement et de mécanismes d’effets secondaires observés. Plusieurs exemples sont présentés à cet égard.
Abstract : Methods that detect binding site similarities between proteins serve for the prediction of function and the identification of potential off-targets. These methods can help prevent side-effects, suggest drug repurposing and polypharmacological strategies and suggest bioisosteric replacements. Most methods use atom-based representations despite the fact that molecular interaction fields (MIFs) represents more closely the nature of what is meant to be identified. We developped a computational algorithm, IsoMif, that detects MIF similarities between binding sites. We benchmark IsoMif to other methods which has not been previously done for a MIF-based method. IsoMif performed best in average and more consistently accross datasets. We highlight limitations intrinsic to the methodology or to nature. The impact of design choices on performance is discussed. We built a freely available web interface that allows the detection of similarities between a protein and pre-calculated MIFs or user defined MIFs. PyMOL sessions can be downloaded to visualize similarities for the different intermolecular interactions. IsoMif was applied for a large-scale analysis (5,6 millions of comparisons) to predict offtargets of drugs. Docking simulations of the drugs in the binding site of their top hits were performed. The primary objective is to generate hypotheses that can be further investigated and validated regarding drug repurposing opportunities and side-effect mechanisms.

Les cancers triple négatif (TN) représente la forme la plus agressive des cancers du sein, avec un pourcentage de décès importants. Il existe une grande hétérogénéité au sein de cette maladie en termes de présentation clinique initiale, de caractéristiques biologiques, de sensibilité au traitement et d’évolution. Aucun progrès en survie n’a été réalisé depuis l’avènement des protocoles de chimiothérapie standards. En effet, malgré une bonne réponse initiale au traitement, 65% des patientes résistent aux thérapies actuelles et récidivent ce qui leur confère un pronostic particulièrement sombre. Il y a donc urgence à identifier de nouveaux protocoles de traitement et de nouvelles molécules efficaces pour ces patientes.Une stratégie de plus en plus intéressante actuellement est celle du repositionnement de composés médicaux qui n’étaient initialement pas destinés au traitement d’une maladie donnée. Cette approche a pour avantage de profiter de l’effort de recherche et développement initial extrêmement couteux et de profiter des études pharmacologiques déjà disponibles. J’ai ainsi effectué au cours de ma thèse un criblage de composés chimiques à haut débit sur 12 lignées de cancers du sein TN afin de prendre en compte leur hétérogénéité. A l’issue de ce crible, plusieurs composés se sont avérés intéressants de par leur potentiel anti-cancéreux. Plus particulièrement, des molécules psychoactives impliquées dans le métabolisme de la sérotonine (ie antidépresseurs, et notamment les inhibiteurs de recapture de la sérotonine-SSRI) sont apparues comme des « hits » forts.Suite à ces résultats, mon travail de thèse s’est orienté plus particulièrement vers la compréhension de l’activité de ces molécules et du métabolisme de la sérotonine dans nos modèles TN afin de comprendre pourquoi ces composés pouvaient présenter un intérêt dans le traitement de ces cancers. Différents aspects biologiques ont ainsi été investigués pour ces antidépresseurs. J’ai ainsi étudié le rôle exercé par la sérotonine sur mes modèles cellulaires. D’autre part, j’ai entrepris une cartographie des acteurs du métabolisme de la sérotonine afin de caractériser mes modèles. J’ai ainsi découvert deux récepteurs à la sérotonine majoritairement présents, HTR1D et HTR1B, qui ont fait l’objet de recherches approfondies. J’ai ainsi pu démontrer l’intérêt de ces deux récepteurs comme cibles thérapeutiques potentielles dans les cancers triple négatifs. Grace à une étude rétrospective j’ai pu mettre en évidence une corrélation statistiquement significative entre le niveau d’expression de chacun de ces deux récepteurs et la survie des patientes TN. Nous observons ainsi une nette discrimination entre les deux groupes de cancers exprimant peu ou fortement ces gènes. J’ai ainsi pu mettre en évidence que ces deux récepteurs représentent des biomarqueurs pronostics forts des patientes TN. L’étude immunohistochimique, a permis de confirmer la présence de ces récepteurs dans les tumeurs TN. Par ailleurs, j’ai pu identifier un micro ARN régulant l’expression de l’un des récepteurs dans les lignées TN. De façon cohérente, j’ai pu observer un effet pronostic significatif du niveau d’expression de ce micro ARN sur la survie des patientes TN. L’efficacité des composés de type SSRI et d’un antagoniste de nos deux récepteurs a pu être vérifiée sur des cultures ex vivo issues de PDX notamment résistantes aux chimiothérapies. L’évaluation préclinique de ces composés a pu être testée sur un premier modèle murin TN de type PDX mais n’a cependant pas permis de démontrer d’efficacité antitumorale in vivo. En effet, la complexité du métabolisme de la sérotonine, tout comme l’hétérogénéité biologique des TN requièrent des études plus approfondies afin de pouvoir faire la preuve de concept du ciblage thérapeutique de ces récepteurs et de la modulation du métabolisme de la sérotonine dans ces cancers. Ce travail fait l’objet d’un manuscrit en préparation pour publication dans le cadre de cette thèse
Triple negative breast cancer (TNBC) is the most aggressive form of breast cancers. It accounts for 15-20% of breast cancers. No progress in survival has been achieved since the advent of standard chemotherapy protocols. TNBC is an important clinical challenge. They have the worst outcome among breast cancer subgroups. Given their poor prognosis, their assumed hetetogeneity, and absence of any alternative specific targeted therapy, chemotherapy remains the only TNBC treatment. Despite an often good initial response to treatment, more than a half of patients do not achieve a pathological complete response, with a frequent and fast tumor relapse. Several therapeutic approaches have been identified preclinically, but none of these molecules have been shown to be effective on all of these patients. There is a urge for the identification of new treatments.An interesting strategy is the repurposing of medical compounds that were initially not intended for the treatment of a given disease. This strategy takes advantage of the extremely expensive initial research and development effort. This process is potentially efficient and cost-effective as previous clinical trials have been performed and pharmacokinetics/pharmacodynamics and toxicity have been already explored. In order to develop new treatment schemes we addressed the following question: Is there available drugs with strong activity in TNBC? To do so, we performed a high-throughput drug screening on 12 TNBC cell lines to reflect the dramatic heterogeneity of the disease. From this drug discovery program, several interesting compounds were identified with significant anti-tumor potential against TNBC. More particularly, psychoactive compounds regulating serotonin metabolism (ie antidepressant drugs and notably serotonin selective reuptake inhibitors-SSRIs) were found to be highly effective “hits”.My thesis work turned to the comprehension of serotonin implication in TNBC physiopathology to understand if modulating its metabolism could be of therapeutic interest for TNBC management. Different biological aspects were investigated concerning serotonin effects on TNBC cellular models (serotonin adjunction in vitro or endogenous synthesis inhibition). In addition, I established a comprehensive map of the serotonergic landscape in TNBC (biosynthetic capacity, transporters, receptors) that led to the identification of therapeutic targets that would be of interest in the treatment of cancer: HTR1D and HTR1B. Indeed, by blocking these promising targets (with chemical inhibitors or siRNA knockdown) we observed a strong reduction in cell viability in our large panel of TNBC cell lines. Remarkably, we found that their expression levels were associated to poor prognosis in breast cancer, and notably in TNBC subtype with huge dichotomy observed in the outcome, allowing future stratification of TNBC patient management and selection for further targeted therapies. These results pinpoint HTR1D and HTR1B as strong prognosis biomarkers in TNBC. Immunohistochemistry staining was also conducted to confirm the presence of these targets at the protein level in tumor samples. Moreover, I could identify a microRNA regulating one of these receptors: has-miR-599. Consistently, expression levels of this microRNA demonstrated a prognostic impact on TNBC survival. While ex vivo data of one SSRI and the dual antagonist of HTR1D/HTR1D receptors shown encouraging efficacy, their preclinical evaluation assessed in a TN PDX model could not allow to demonstrate any significant effect on tumor growth in vivo. As a matter of fact, serotonin metabolism is a complex system and TNBC heterogeneity does not permit to conclude on the therapeutic proof of concept of the serotonergic modulation in TNBC with this first attempt. A scientific manuscript of this work is being prepared for publication

La Medicina di Precisione (Precision Medicine) è un nuovo paradigma che sta rivoluzionando diversi aspetti delle pratiche cliniche: nella prevenzione e diagnosi, essa è caratterizzata da un approccio diverso dal "one size fits all" proprio della medicina classica. Lo scopo delle Medicina di Precisione è di trovare misure di prevenzione, diagnosi e cura che siano specifiche per ciascun individuo, a partire dalla sua storia personale, stile di vita e fattori genetici. Tre fattori hanno contribuito al rapido sviluppo della Medicina di Precisione: la possibilità di generare rapidamente ed economicamente una vasta quantità di dati omici, in particolare grazie alle nuove tecniche di sequenziamento (Next-Generation Sequencing); la possibilità di diffondere questa enorme quantità di dati grazie al paradigma "Big Data"; la possibilità di estrarre da questi dati tutta una serie di informazioni rilevanti grazie a tecniche di elaborazione innovative ed altamente sofisticate. In particolare, le tecniche di Machine Learning introdotte negli ultimi anni hanno rivoluzionato il modo di analizzare i dati: esse forniscono dei potenti strumenti per l'inferenza statistica e l'estrazione di informazioni rilevanti dai dati in maniera semi-automatica. Al contempo, però, molto spesso richiedono elevate risorse computazionali per poter funzionare efficacemente. Per questo motivo, e per l'elevata mole di dati da elaborare, è necessario sviluppare delle tecniche di Machine Learning orientate al Big Data che utilizzano espressamente tecniche di High Performance Computing, questo per poter sfruttare al meglio le risorse di calcolo disponibili e su diverse scale, dalle singole workstation fino ai super-computer. In questa tesi vengono presentate tre tecniche di Machine Learning sviluppate nel contesto del High Performance Computing e create per affrontare tre questioni fondamentali e ancora irrisolte nel campo della Medicina di Precisione, in particolare la Medicina Genomica: i) l'identificazione di varianti deleterie o patogeniche tra quelle neutrali nelle aree non codificanti del DNA; ii) l'individuazione della attività delle regioni regolatorie in diverse linee cellulari e tessuti; iii) la predizione automatica della funzione delle proteine nel contesto di reti biomolecolari. Per il primo problema è stato sviluppato parSMURF, un innovativo metodo basato su hyper-ensemble in grado di gestire l'elevato grado di sbilanciamento che caratterizza l'identificazione di varianti patogeniche e deleterie in mezzo al "mare" di varianti neutrali nelle aree non-coding del DNA. L'algoritmo è stato implementato per sfruttare appositamente le risorse di supercalcolo del CINECA (Marconi - KNL) e HPC Center Stuttgart (HLRS Apollo HAWK), ottenendo risultati allo stato dell'arte, sia per capacità predittiva, sia per scalabilità. Il secondo problema è stato affrontato tramite lo sviluppo di reti neurali "deep", in particolare Deep Feed Forward e Deep Convolutional Neural Networks per analizzare - rispettivamente - dati di natura epigenetica e sequenze di DNA, con lo scopo di individuare promoter ed enhancer attivi in linee cellulari e tessuti specifici. L'analisi è compiuta "genome-wide" e sono state usate tecniche di parallelizzazione su GPU. Infine, per il terzo problema è stato sviluppato un algoritmo di Machine Learning semi-supervisionato su grafo basato su reti di Hopfield per elaborare efficacemente grandi network biologici, utilizzando ancora tecniche di parallelizzazione su GPU; in particolare, una parte rilevante dell'algoritmo è data dall'introduzione di una tecnica parallela di colorazione del grafo che migliora il classico approccio greedy introdotto da Luby. Tra i futuri lavori e le attività in corso, viene presentato il progetto inerente all'estensione di parSMURF che è stato recentemente premiato dal consorzio Partnership for Advance in Computing in Europe (PRACE) allo scopo di sviluppare ulteriormente l'algoritmo e la sua implementazione, applicarlo a dataset di diversi ordini di grandezza più grandi e inserire i risultati in Genomiser, lo strumento attualmente allo stato dell'arte per l'individuazione di varianti genetiche Mendeliane. Questo progetto è inserito nel contesto di una collaborazione internazionale con i Jackson Lab for Genomic Medicine.
Precision Medicine is a new paradigm which is reshaping several aspects of clinical practice, representing a major departure from the "one size fits all" approach in diagnosis and prevention featured in classical medicine. Its main goal is to find personalized prevention measures and treatments, on the basis of the personal history, lifestyle and specific genetic factors of each individual. Three factors contributed to the rapid rise of Precision Medicine approaches: the ability to quickly and cheaply generate a vast amount of biological and omics data, mainly thanks to Next-Generation Sequencing; the ability to efficiently access this vast amount of data, under the Big Data paradigm; the ability to automatically extract relevant information from data, thanks to innovative and highly sophisticated data processing analytical techniques. Machine Learning in recent years revolutionized data analysis and predictive inference, influencing almost every field of research. Moreover, high-throughput bio-technologies posed additional challenges to effectively manage and process Big Data in Medicine, requiring novel specialized Machine Learning methods and High Performance Computing techniques well-tailored to process and extract knowledge from big bio-medical data. In this thesis we present three High Performance Computing Machine Learning techniques that have been designed and developed for tackling three fundamental and still open questions in the context of Precision and Genomic Medicine: i) identification of pathogenic and deleterious genomic variants among the "sea" of neutral variants in the non-coding regions of the DNA; ii) detection of the activity of regulatory regions across different cell lines and tissues; iii) automatic protein function prediction and drug repurposing in the context of biomolecular networks. For the first problem we developed parSMURF, a novel hyper-ensemble method able to deal with the huge data imbalance that characterizes the detection of pathogenic variants in the non-coding regulatory regions of the human genome. We implemented this approach with highly parallel computational techniques using supercomputing resources at CINECA (Marconi – KNL) and HPC Center Stuttgart (HLRS Apollo HAWK), obtaining state-of-the-art results. For the second problem we developed Deep Feed Forward and Deep Convolutional Neural Networks to respectively process epigenetic and DNA sequence data to detect active promoters and enhancers in specific tissues at genome-wide level using GPU devices to parallelize the computation. Finally we developed scalable semi-supervised graph-based Machine Learning algorithms based on parametrized Hopfield Networks to process in parallel using GPU devices large biological graphs, using a parallel coloring method that improves the classical Luby greedy algorithm. We also present ongoing extensions of parSMURF, very recently awarded by the Partnership for Advance in Computing in Europe (PRACE) consortium to further develop the algorithm, apply them to huge genomic data and embed its results into Genomiser, a state-of-the-art computational tool for the detection of pathogenic variants associated with Mendelian genetic diseases, in the context of an international collaboration with the Jackson Lab for Genomic Medicine.

In the last decade, molecular biology has expanded from a reductionist view to a systems-wide view that tries to unravel the complex interactions of cellular components. Owing to the emergence of high-throughput technology it is now possible to interrogate entire genomes at an unprecedented resolution. The dimension and unstructured nature of these data made it evident that new methodologies and tools are needed to turn data into biological knowledge. To contribute to this challenge we exploited the wealth of publicly available high-throughput genomics data and developed bioinformatics methodologies focused on extracting information at the pathway rather than the single gene level. First, we developed Gene Set Variation Analysis (GSVA), a method that facilitates the organization and condensation of gene expression profiles into gene sets. GSVA enables pathway-centric downstream analyses of microarray and RNA-seq gene expression data. The method estimates sample-wise pathway variation over a population and allows for the integration of heterogeneous biological data sources with pathway-level expression measurements. To illustrate the features of GSVA, we applied it to several use-cases employing different data types and addressing biological questions. GSVA is made available as an R package within the Bioconductor project. Secondly, we developed a pathway-centric genome-based strategy to reposition drugs in type 2 diabetes (T2D). This strategy consists of two steps, first a regulatory network is constructed that is used to identify disease driving modules and then these modules are searched for compounds that might target them. Our strategy is motivated by the observation that disease genes tend to group together in the same neighborhood forming disease modules and that multiple genes might have to be targeted simultaneously to attain an effect on the pathophenotype. To find potential compounds, we used compound exposed genomics data deposited in public databases. We collected about 20,000 samples that have been exposed to about 1,800 compounds. Gene expression can be seen as an intermediate phenotype reflecting underlying dysregulatory pathways in a disease. Hence, genes contained in the disease modules that elicit similar transcriptional responses upon compound exposure are assumed to have a potential therapeutic effect. We applied the strategy to gene expression data of human islets from diabetic and healthy individuals and identified four potential compounds, methimazole, pantoprazole, bitter orange extract and torcetrapib that might have a positive effect on insulin secretion. This is the first time a regulatory network of human islets has been used to reposition compounds for T2D. In conclusion, this thesis contributes with two pathway-centric approaches to important bioinformatic problems, such as the assessment of biological function and in silico drug repositioning. These contributions demonstrate the central role of pathway-based analyses in interpreting high-throughput genomics data.
En l'última dècada, la biologia molecular ha evolucionat des d'una perspectiva reduccionista cap a una perspectiva a nivell de sistemes que intenta desxifrar les complexes interaccions entre els components cel•lulars. Amb l'aparició de les tecnologies d'alt rendiment actualment és possible interrogar genomes sencers amb una resolució sense precedents. La dimensió i la naturalesa desestructurada d'aquestes dades ha posat de manifest la necessitat de desenvolupar noves eines i metodologies per a convertir aquestes dades en coneixement biològic. Per contribuir a aquest repte hem explotat l'abundància de dades genòmiques procedents d'instruments d'alt rendiment i disponibles públicament, i hem desenvolupat mètodes bioinformàtics focalitzats en l'extracció d'informació a nivell de via molecular en comptes de fer-ho al nivell individual de cada gen. En primer lloc, hem desenvolupat GSVA (Gene Set Variation Analysis), un mètode que facilita l'organització i la condensació de perfils d'expressió dels gens en conjunts. GSVA possibilita anàlisis posteriors en termes de vies moleculars amb dades d'expressió gènica provinents de microarrays i RNA-seq. Aquest mètode estima la variació de les vies moleculars a través d'una població de mostres i permet la integració de fonts heterogènies de dades biològiques amb mesures d'expressió a nivell de via molecular. Per il•lustrar les característiques de GSVA, l'hem aplicat a diversos casos usant diferents tipus de dades i adreçant qüestions biològiques. GSVA està disponible com a paquet de programari lliure per R dins el projecte Bioconductor. En segon lloc, hem desenvolupat una estratègia centrada en vies moleculars basada en el genoma per reposicionar fàrmacs per la diabetis tipus 2 (T2D). Aquesta estratègia consisteix en dues fases: primer es construeix una xarxa reguladora que s'utilitza per identificar mòduls de regulació gènica que condueixen a la malaltia; després, a partir d'aquests mòduls es busquen compostos que els podrien afectar. La nostra estratègia ve motivada per l'observació que els gens que provoquen una malaltia tendeixen a agrupar-se, formant mòduls patogènics, i pel fet que podria caldre una actuació simultània sobre múltiples gens per assolir un efecte en el fenotipus de la malaltia. Per trobar compostos potencials, hem usat dades genòmiques exposades a compostos dipositades en bases de dades públiques. Hem recollit unes 20.000 mostres que han estat exposades a uns 1.800 compostos. L'expressió gènica es pot interpretar com un fenotip intermedi que reflecteix les vies moleculars desregulades subjacents a una malaltia. Per tant, considerem que els gens d'un mòdul patològic que responen, a nivell transcripcional, d'una manera similar a l'exposició del medicament tenen potencialment un efecte terapèutic. Hem aplicat aquesta estratègia a dades d'expressió gènica en illots pancreàtics humans corresponents a individus sans i diabètics, i hem identificat quatre compostos potencials (methimazole, pantoprazole, extracte de taronja amarga i torcetrapib) que podrien tenir un efecte positiu sobre la secreció de la insulina. Aquest és el primer cop que una xarxa reguladora d'illots pancreàtics humans s'ha utilitzat per reposicionar compostos per a T2D. En conclusió, aquesta tesi aporta dos enfocaments diferents en termes de vies moleculars a problemes bioinformàtics importants, com ho son el contrast de la funció biològica i el reposicionament de fàrmacs "in silico". Aquestes contribucions demostren el paper central de les anàlisis basades en vies moleculars a l'hora d'interpretar dades genòmiques procedents d'instruments d'alt rendiment.

Relatório de Estágio do Mestrado Integrado em Ciências Farmacêuticas apresentado à Faculdade de Farmácia
A Doença de Alzheimer (DA) é uma doença neurodegenerativa que se caracteriza pela deterioração global, progressiva e irreversível das funções cognitivas tais como, memória, linguagem e pensamento. A DA apresenta duas caraterísticas neuropatológicas importantes: acumulação de agregados insolúveis de peptídeo β-amilóide (Amyloid beta, Aβ) e a agregação da proteína tau em tranças neurofibrilares (Neurofibrillary Tangles, NFTs). Estes agregados danificam as conexões existentes entre os neurónios impossibilitando a sua comunicação e conduzindo à sua morte. As hipóteses globalmente aceites para explicar a patologia subjacente à DA, são a hipótese da cascata amilóide e da patologia tau.Apesar da investigação crescente na área da DA não existe, até ao momento, um tratamento farmacológico capaz de alterar o curso da doença. Desta forma, é evidente a necessidade premente de explorar novos alvos terapêuticos de forma a encontrar uma terapia segura e eficaz que permita atrasar ou reverter a progressão da doença. Assim, o drug repurposing (DR), o processo de pesquisa de novas indicações terapêuticas para fármacos já aprovados, é uma hipótese a considerar.A descoberta de um eventual papel da via dos leucotrienos na neuroinflamação crónica associada à DA e da sua contribuição para as marcas patológicas da doença despertou o interesse sobre o reposicionamento de fármacos antiasmáticos que atuam nesta via.Por outro lado, a DA e a epilepsia partilham características fisiopatológicas comuns, tais como défice de aprendizagem e memória, neurodegeneração e morte celular. Existem igualmente evidências de que a atividade epileptiforme acelera o declínio cognitivo através decrises silenciosas e aumenta a formação do peptídeo Aβ e da proteína tau, o que suscitou o interesse sobre o possível reposicionamento de fármacos antiepiléticos.A evidência de que “fármacos antigos” podem ser a chave para o tratamento da DA, faz do DR uma hipótese a considerar no futuro.
Alzheimer's disease (AD) is a neurodegenerative disease that is characterized by global, progressive and irreversible deterioration of cognitive functions such as memory, language and thinking. AD has two important neuropathological characteristics: accumulation of insolubleaggregates of β-amyloid peptide (Aβ) and the aggregation of tau protein in neurofibrillary tangles (NFTs). These aggregates damage the existing connections between neurons, making communication impossible and leading to their death. The globally accepted hypothesis to explain the pathology underlying AD is the amyloid cascade and tau pathology. Despite the growing research in the field of AD, there is, to date, no pharmacological treatment capable of altering the course of the disease. Thus, there is an evident need to explore new therapeutic targets in order to find a safe and effective therapy that allows todelay or reverse the progression of the disease. Thus, drug repurposing (DR), the process of discovering new therapeutic uses for drugs already approved for other clinical indications, is a hypothesis to consider.The discovery of an eventual role of the leukotriene pathway in chronicneuroinflammation associated with AD and its contribution to the pathological marks of the disease aroused interest in the repositioning of antiasthmatic drugs that act in this pathway.On the other hand, AD and epilepsy share common pathophysiological characteristics, such as learning and memory deficit, neurodegeneration and cell death. There is also evidence that epileptiform activity accelerates cognitive decline through silent crises and increases theformation of Aβ peptide and tau protein, which has raised interest about the possible repositioning of antiepileptic drugs.The evidence that “old drugs” may be the key to the treatment of AD, makes DR a hypothesis to consider in the future.

The most common cancer in women is breast cancer with approximately 1 in 8 women developing this disease in their lifetime. Clinically, breast cancer can be divided into distinct subtypes based on the presence or absence of hormone receptors such as estrogen receptor (ER), progesterone receptor (PR) and expression of the HER2 gene. In this project, one of the most difficult to treat subclasses - triple negative (ER-/PR-/HER2-) breast cancer (TNBC) is studied. Until now, very limited drug treatment strategies are available for TNBCs because of a lack of hormone receptors as potential drug targets. Thus, it is a matter of urgency to seek specific treatment tailored to TNBC patients. This raises the prospect that antiretroviral drugs might be able to repurpose as anticancer drugs for TNBCs. In the mid-1990s, the incidence of the AIDS-related cancers was greatly reduced in the HIV patients due to the introduction of antiretroviral therapy to the HIV patients. It has been suggested that a direct inhibitory effect of antiretroviral drugs on the reverse transcriptase activity of long interspersed nuclear element 1 (LINE-1) in tumour cells could be a crucial factor. LINE-1 is the most important transposon with autonomous retro-transposition ability in humans. It can alter gene regulations and cause somatic mutations. Since LINE-1 has the potential to adversely affect individuals, it is silenced in differentiated tissues by diverse endogenous mechanisms. Nonetheless, LINE-1 is highly expressed in many cancers, especially in breast carcinomas. Therefore, it is worth examining whether antiretroviral drugs can be repurposed as anticancer drugs for treating TNBCs and to further understand the relationship between LINE-1 and TNBCs. Antiretroviral drug induced anticancer effects may be relevant to down-regulation of the fatty acid metabolism pathway. In this project, two antiretroviral drugs - Efavirenz and SPV122, had been shown to cause cell death and cell proliferation retardation, thus, effectively eliminating cancer cells in a range of TNBC cell lines. Additionally, LINE-1 suppression had been observed in the antiretroviral drugs-treated breast cancer cell lines implying a potential link between LINE-1 and TNBCs. Whole genome RNA sequencing data further highlighted the possible mechanisms involved in this anticancer process. It seemed the fatty acid metabolism pathway could be a key regulator in this anticancer process. Many key genes involved in fatty acid metabolism were down-regulated after drug treatments. However, the antiretroviral drugs-treated MCF10AT and MCF10CA1a cells were found to present some mesenchymal markers which are often characteristic signs of poor prognostic outcomes thereby highlighting the complexity of TNBCs. The RNA sequencing data also strongly implied that cancer stem cells (CSCs) could play a role in these confusing results. CSCs are a small group of cancer cells with stem cell-like abilities, and they are thought to be responsible for drug resistance, cancer metastasis, and cancer recurrence. Interestingly, in a series of experiments, different groups of CSCs showed various responses to a range of drugs. ALDHhigh epithelial-type CSCs were significantly reduced after antiretroviral drug treatment; whereas, CD44+/CD24- mesenchymal-type CSCs were increased after treatment. These results highlighted the importance of CSC heterogeneity and implied that mesenchymal-type CSCs have greater resistance to the drugs than other cancer cells. Finally, the functional CSC assay demonstrated that CSCs can be eliminated by the antiretroviral drugs indicating that Efavirenz and SPV122 might be able to target both non-CSCs and CSCs. To combine all the results together, Efavirenz and SPV122 could potentially be valid anticancer drugs for treating TNBCs by regulating cancer fatty acid metabolism. Follow-up experiments are necessary to further understand how antiretroviral drugs impact anticancer processes.

Trabalho Final de Mestrado Integrado, Ciências Farmacêuticas, Universidade de Lisboa, Faculdade de Farmácia, 2015
The aim of this discussion was to explore the concept and purpose of repositioning drugs, understand the boundaries that still underlie the technique and, analyze repurposed drugs, so they can serve of inspiration for the future of drug discovery and development. Drug repurposing, often also mentioned as drug repositioning is defined as the rational use of know drugs for new indications in order to increase pharmaceutical industry productivity and deliver therapeutic options to patients who suffer from chronic, orphan, neglected, rare, untreatable diseases or pathologies with poor therapeutic approaches. The limitations of this method are the same as in de novo drug development, the idiosyncrasy of both the disease and patient, the acquired resistance to therapy, the bureaucracy implicated in the submission of a drug's approval request and the lack of scientific knowledge to target certain pathologies, makes it hard and risky to develop and commercialize whether a new molecule or an old drug for a new indication. The methods of drug repositioning may be classified in treatment oriented, disease or drug oriented. The methodologies may not require an elevated level of scientific knowledge as serendipity testing or may, on the other hand, demand the comprehension of the shape and binding properties of the substance, as molecular docking. Drug repositioning presented the community with useful therapeutic approaches and, at the same time, has increased the profit of drugs which had been already abandoned. Examples such as thalidomide, a drug created for motion-sickness that was found to be teratogenic, it was later on repurposed for multiple myeloma. Sildenafil, a drug that started out as a low efficacy anti-anginous, but proved to be useful in erectile dysfunction and in pulmonary hypertension. Duloxetine, an old anti- depressant repurposed to syndrome of urinary incontinence, neuropathic pain and to generally anxiety disorder in children, due to one same mechanism of action. Drug repurposing is a fructuous approach for the development of new therapeutics, nevertheless several points have to be enlightened and simplified. Protocols of methods have to be created in order to achieve maximization of time and costs. The legal framework should be simplified, by reducing the heterogeneity between international agencies.

Cancer remains one of the most pressing issues of contemporary science and medicine. Incidence of malignant diseases is rising worldwide and they represent a major problem for the society due to both economic and ethical issues they cause. Although the progress in cancer biology, therapy and immunology has led to the introduction of many novel therapeutic protocols, approaches and drugs with specificity defined on a molecular level into clinical practice, many malignancies retain their poor prognosis. Therefore, intense research into new ways to increase our therapeutic options is warranted. Unfortunately, bringing a completely novel drug into clinical use takes extremely high amounts of time and money and entails a high risk of failure. Therefore, a promising approach has been recently adopted which lies in repurposing compounds already used in human medicine for cancer treatment. This form of research can advance through clinical trials for a new indication much easier, faster and cheaper than researching completely new drugs. The aim of this study was to examine the anticancer potential of one such drug, mebendazole. An anthelminthic from the family of benzimidazoles, mebendazole has been in common clinical use from the 1970s and is marked by its low toxicity as well as its very low solubility....

碩士
國立陽明大學
生物藥學研究所
101
Therapy of cancer is an important question needed to be solved, but many anti-neoplastic agents are facing to dilemma of discovery, such as too much cytotoxicity and taking a long time to place a drug into market. Drug repurposing is one of the approaches to tackling the challenges. Moreover, drug repurposing may substantially increase the number and quality of innovative, cost-effective new medicines, without incurring unsustainable R&;D costs. In this project, I aim to analyze drugs from the Connectivity Map (CMap) and design strategy of literature review to understand situation of drug repurposing of focused drugs in this thesis. Briefly, there are 1309 drugs in CMap and 443 of them are in National Health Insurance Research Database (NHIRD). There are FDA-labeled indications, Non FDA-labeled indications, literature review, patent survey, clinical trials in the strategy. According to this strategy, we can understand the patent situation of drug repurposing of focused drugs, including anti-psychotic agents and several novel compounds. Many potential anti-neoplastic agents have seen tested in clinical trials and have not been patented. In conclusion, drug discovery of anti-neoplastic agents through drug repurposing is highly promising.

碩士
國立成功大學
資訊工程學系
102
Drug development is a time-consuming, expensive, and high-risk task. The uncertainty of drug development has led to the emergence of drug repurposing, which is to find the new indications of approved drugs. Approved drugs have completed more clinical trial data than newly developed drugs. Therefore, drug repurposing is safer and faster than conventional approaches of drug development. This study aims to infer disease-drug indirect relations via disease-gene and gene-drug relations from large-scale biomedical literature. We propose a pattern-based relation extraction method using dependency grammar to identify disease, gene, and drug relations to construct disease-gene and gene-drug bipartite networks. In these bipartite networks, we can understand that a disease is caused by the involvement of gene product from disease-gene network. We can also understand the interaction between protein and drug from gene-drug bipartite network. However, these networks produce a large number of indirect relations between disease and drug. We propose a novel ranking method to prioritize the indirect relations. The concept of the ranking method is based on drug similarity which is defined by repurposed drugs and approved drugs. If a repurposed drug and an approved drug have highly similar interactions with the common genes, the repurposed drug might have a new indication that it has similar effects as that of the approved drug. Our pattern-based relation extraction method performs a higher precision of 0.86 than baseline methods. Because our drug similarity method obtains an R-square score of 0.80 with ATC code similarity, our drug vector space is suitable to calculate drug similarity. Therefore, the ranking method achieves a MAP score of 0.37 in top 100 popular diseases. Finally, we select the repurposed drugs of ovarian cancer, prostate cancer, lung cancer, colorectal cancer, leukemia, and breast cancer for validation by literature study and clinical trials.

博士
國立交通大學
生物資訊及系統生物研究所
102
Developing a new drug is extremely time consuming and expensive. Recently, repurposing drugs or pro-drugs has been proposed as an important paradigm for accelerating therapeutic strategies into clinical trials. Many drugs have been indicated that they can interact with more than one target protein and been used for new indications. In addition, drugs that simultaneously target multiple proteins often improve efficacy, particularly in the treatment of complex diseases such as cancers. Therefore, how to identify target proteins of a compound will be helpful for drug repurposing and multi-target drug design. However, it is still an unsolved problem because many target proteins are not similar in their sequences or structures. In this thesis, we propose the core concept "Homopharma", which describe the conserved binding environment and preferred properties between proteins and compounds, to explore the molecular binding mechanisms and drug repurposing. A Homopharma consists of a set of proteins with the conserved binding interface and a set of compounds that share similar structures and functional groups. In order to recognize the conserved binding interfaces, we developed Space-Related Pharma-motifs (SRPmotif) composed of pharma-interfaces and discontinuous pharma-motifs to rapidly search similar binding interfaces against the structure database. Our results show that proteins of the identified pharma-interfaces not only are involved in the similar cellular process, but also perform similar biological functions. Furthermore, the proteins and compounds in a Homopharma share conserved interactions and similar physico-chemical properties; therefore, the compounds can often bind to the proteins. Experimental results show that protein-compound complexes of a Homopharma often preform similar interactions in which formed by conserved binding residues (protein sites) and similar important functional groups (compound sites). According to the Homopharma concept, we successfully discovered 56 novel flavonoid-kinase inhibitions (IC50 ≤ 10 μM) by in vitro enzymatic profiling, whereas the IC50 values of 25 bioassays are less than 1 μM. Some novel flavonoid-kinase inhibitions also suggest that these flavonoids can be considered as potential anticancer compounds such as oral and colorectal cancer drugs. The results indicate that Homopharma can be utilized to recognize key binding environments between proteins and compounds and discover new usages for existing drugs. Moreover, to design selective kinase inhibitors, we also developed KIDFamMap to explore kinase-inhibitor families (KIFs) and kinase-inhibitor-disease relationships for kinase inhibitor selectivity and mechanisms. The kinase-inhibitor interactions of a KIF are often conserved on some consensus KIDFamMap anchors, which represent conserved interactions between the kinase subsites and consensus moieties of their inhibitors. Our results reveal that the members of a KIF often possess similar inhibition profiles. Integrating the concept of Homopharma, inhibitory effects of compounds, and diseases information, a structure-based protein-compound interaction network can be constructed to explore protein-compound-disease relationships. This protein-compound interaction network would not only be helpful for identifying additional targets of repositioning drugs, improving efficacy and understanding toxicity of compounds, but also provides opportunities for revealing new therapeutic strategies of specific diseases. We believe that these concepts proposed in this thesis can have the potential for understanding molecular binding mechanisms and giving new clues for drug repurposing and drug development.

Dissertação de Mestrado, Biologia Molecular e Microbiana, Faculdade de Ciências e Tecnologia, Universidade do Algarve, 2015
O melanoma maligno é considerado a forma mais letal de cancro de pele, com uma elevada tendência para metastizar para o cérebro. Atualmente, as metástases cerebrais são tratadas através de cirurgia, quimioterapia, radioterapia e radiocirurgia, mas o sucesso destes tratamentos são mínimos, como tal novas estratégias terapêuticas são importantes e necessárias. A fim de determinar novas estratégias de terapia cancerígena, são precisos modelos animais adequados para investigar os efeitos dos tratamentos. Em trabalhos anteriores, desenvolveu-se um novo modelo animal, através de injecção, na corrente sanguínea de organismos imunodeficientes (nod/scid), de células de melanoma humano . Estes animais imunodeficientes, para além de desenvolver tumores cerebrais, também desenvolveram metástases em outros órgãos. A partir dos tumores desenvolvidos, foi determinada, por sequenciamento de RNA, uma lista de genes candidatos responsáveis pelas metástases cerebrais,. Com uma análise bioinformática, utilizando o connectivity map (cMAP), foi possível encontrar várias drogas, administradas em pacientes para outros fins terapêuticos, que podem ser também eficazes no tratamento de metástases cerebrais. O objetivo deste trabalho foi testar os fármacos candidatos, determinados pelas sequencias de RNA, em quatro linhas celulares diferentes de melanoma humano. Dos ensaios in vitro realizados com 9 drogas candidatas, apenas 5 mostraram ter algum potencial, mas só 3 foram selecionadas (Tricostatina A, Metildopa e Pentamidina) para serem testadas in vivo. Esta seleção baseou-se na análise do peso molecular, devido às limitações da passagem pela barreira hematoencefálica. Nos resultados in vivo, observou-se um efeito positivo, ainda que transitório, sobre a carga tumoral e o volume do tumor após quatro semanas de tratamento, quando eram expostos com metildopa. Assim, estudos adicionais têm de ser realizados utilizando a metildopa, para se confirmar e validar os resultados obtidos, e perceber se é suficientemente eficaz para ser utilizada de forma preventiva. Com este trabalho também foi possível otimizar-se protocolos de ressonância magnética para facilitar a observação dos tumores, especialmente os de menores dimensões. Para além disso, a ressonância magnética também permite verificar se os tumores estão a invadir tecidos circundantes. Este protocolo tornou-se uma boa ferramenta que pode ser usado em futuros estudos, permitindo estudar de forma mais eficaz as alterações dos tumores, e contribuir assim para o desenvolvimento de melhores tratamentos.