Previously, the CCR5 receptor was found to be a good target for treating prostate cancer (PCa). Dr. Yan Zhang’s laboratory designed several CCR5 antagonists, which were screened for their inhibitory effect on the growth and invasion of the M12, DU145 and PC-3 PCa cell lines. Primary in vitro screening showed one compound (Drug 17) significantly inhibited the proliferation of PCa cells at 1μM concentration, with a half-maximal inhibitory concentration of 237.68 nM. Further in vitro assays including a proliferation, cytotoxicity and invasion assay confirmed the inhibitory effect of drug 17. The physiological effect of drug 17 was tested by the Ware laboratory in vivo by subcutaneous injection of M12 cells into male, athymic nude mice. Tumor growth was slowed in mice receiving injections of drug 17 compared to sham injected controls. Thus, in vitro and in vivo assays suggest drug 17 might be an effective therapy to block PCa progression.

The human body is colonized by more than 100 trillion microbes which make up an essential part of the body and plays a significant role in health. We now know the over use and misuse of broad-spectrum antibiotics can disrupt this microbiome contributing to the onset of disease and runs the risk of promoting antibiotic resistance. With antibiotic research still on the decline, new strategies are greatly needed to combat emerging pathogens while maintaining a healthy microbiome. We therefore set out to present a novel species-selective antimicrobial drug discovery strategy. Disruption of the homeostasis within the oral cavity can trigger the onset of one of the most common bacterial infections, periodontal disease. Even though the oral cavity is one of the most diverse sites on the human body, the Gram-negative colonizer, Porphyromonas gingivalis has long been considered a key player in the initiation of periodontitis, suggesting the potential for novel narrow-spectrum therapeutics. By targeting key pathogens, it may be possible to treat periodontitis while allowing for the recolonization of the beneficial, healthy flora. Therefore, we set out to use P. gingivalis and periodontal disease as a model for pathogen-specific antimicrobial drug discovery. In this study we present a unique approach to predict essential gene targets selective for the periodontal pathogen within the oral environment. Using our knowledge of metabolic networks and essential genes we identified a “druggable” essential target, meso-diaminopimelate dehydrogenase, which is found in a limited number of species. This enzyme, meso-diaminopimelate dehydrogenase from P. gingivalis, was first expressed and purified, then characterized for enzymatic inhibitor screening studies. We then applied a computer-based drug discovery method, combining pharmacophore models, high-throughput virtual screening and molecular docking. Utilizing the ZINC database we virtually screened over 9 million small-molecules to identify several potential target-specific inhibitors. Finally, we used target-based and whole-cell based biochemical screening to assess in vitro activity. We conclude that the establishment of this target and screening strategy provides a framework for the future development of new antimicrobials and drug discovery.

The 14-3-3 family of proteins are important signalling proteins involved in a number of cellular processes. These include cell cycle regulation, apoptosis, signal transduction and cell signalling. There is also considerable evidence in the literature that 14-3-3 proteins play a vital role in the pathology of neurodegenerative diseases, including Alzheimer’s, Parkinson’s, Huntington’s and Prion disease. The neurodegenerative disease of focus in this research is Spinocerebellar Ataxia Type 1 (SCA1). SCA1 is a polyglutamine-repeat disease and the interaction of the disease protein ataxin-1 with 14-3-3 proteins leads to the toxic accumulation and subsequent protein aggregation which is characteristic of this disease. This study focused on attempting to elucidate the structure of various domains of the disease protein and also in identifying potential inhibitors of this deleterious interaction. Unfortunately, structural studies were not successful due to a number of caveats encountered in the expression and purification of the ataxin-1 protein domains. By utilising computational methods and small molecule inhibitors, a number of potential lead compounds which possess the ability to at least partly disrupt the interaction of 14- 3-3ζ have been identified. As 14-3-3 proteins play roles in other neurodegenerative diseases, successful identification of potential drug lead treatments can have far reaching benefits in a number of neurodegenerative diseases including SCA1. Lipid rafts are also involved in neurodegenerative disease pathology. Lipid rafts are cholesterol and sphingolipid rich domains which organise the plasma membrane into discrete microdomains and act as signalling platforms and processing centres which attach specific proteins and lipids. A number of disease proteins are processed at these membrane regions, including those involved in Alzheimer’s, Parkinson’s and Prion disease. This processing is a step which is critical in the pathology of disease and abnormal processing leads to the formation of toxic protein aggregates. Previous research in the lab identified the association of low levels of the five main brain isoforms of 14-3-3 proteins with rafts. This study expanded on this to positively identify the presence of the two phospho-forms of 14-3-3, α and δ. The mechanism by which 14-3-3 proteins associate with rafts was also investigated, indicating that 14-3-3 associates with rafts via an unidentified raftbound protein(s). In addition, the phosphorylation status and quaternary structure of 14-3-3 in the presence of sphingolipids has been explored.

The misfolding and aggregation of proteins is closely associated with more than fifty human disorders, including Alzheimer's and Parkinson's diseases, all of which are currently incurable and many represent a major threat to human life. The mechanism of protein aggregation is subject to extensive studies. The damaging effects associated with protein aggregation have been attributed to amyloidogenic species that are present during the misfolding process. In particular, oligomeric species are, however, intrinsically difficult to study as a consequence of their low abundance and highly heterogeneous nature. The first chapter of my thesis gives an introduction into the field of protein folding and misfolding with a focus on the study of protein aggregation, and toxic effects relevant to human disorders. The second chapter of my thesis describes the development of a methodology that enables the study of aggregate induced lipid bilayer permeability, possibly the most general mechanism of protein aggregate toxicity. Surface-tethered lipid vesicles functioning as optochemical probes sensitive to membrane integrity are imaged using total internal reflection microscopy. It is shown that oligomeric species of the 42-residue form of the Aβ peptide (Aβ42) are responsible for the membrane disruption. The methodology can be applied to the study of other proteins such as α-synuclein and tau, and the ability of antibodies and chaperones to counteract the aggregate induced lipid bilayer permeability can be assessed. Furthermore, lipid bilayer permeability induced by aggregates formed in human induced pluripotent stem cells can be studied. The third chapter presents a new approach for the measurement of protein aggregation kinetics by following the development of the lipid bilayer permeability over the course of the aggregation process of Aβ42. The aggregation kinetics can be modulated with molecular chaperones and pre-formed seed fibrils, which allows secondary nucleation to be identified as the process that drives the formation of species responsible for the lipid bilayer permeability. The fourth chapter describes the development of a three-pronged strategy to study the mechanism of α-synuclein amyloid formation. The aggregation is studied in the presence of lipid vesicles or pre-formed fibrils at neutral or acidic pH of the solution. The influence of single-point mutations on the aggregation of α-synuclein is described. Furthermore, the strategy is applied to the characterisation of the ability of antibodies and small molecules to inhibit the aggregation, and thus has the potential for the development of therapeutical agents. The work presented in the fifth chapter characterises the amyloid fibril populations formed by α-synuclein and mutational variants associated with familial Parkinson's disease. X-ray crystallography, circular dichroism spectroscopy, Fourier transform infrared spectroscopy, transmission electron microscopy and atomic force microscopy have all been applied to the analysis of these amyloid fibrils. Finally, the sixth chapter summarises the results described in this thesis and points out future opportunities in the context of fundamental and translational studies related to the research area of protein misfolding disorders.

Submitted by Fabíola Silva (fabiola.silva@famerp.br) on 2018-02-19T12:03:16Z No. of bitstreams: 1 thiagohenrique_tese.pdf: 1021380 bytes, checksum: c215716ce3befdfb390cce0309052b80 (MD5) / Made available in DSpace on 2018-02-19T12:03:16Z (GMT). No. of bitstreams: 1 thiagohenrique_tese.pdf: 1021380 bytes, checksum: c215716ce3befdfb390cce0309052b80 (MD5) Previous issue date: 2016-05-16 / Introduction: The total amount of scientific literature on cancer has grown rapidly in recent years. This makes it difficult, if not impossible, to manually retrieve relevant information on the mechanisms that govern the neoplastic process. Furthermore, cancer is a complex disease, and its. Heterogeneity is particularly evident in head and neck squamous cell carcinoma (HNSCC); one of the most common types of cancer worldwide. Objectives: The present study aimed: a) to identify genes/proteins related to HNSCC; b) to identify ligands that specifically target molecular biomarkers of interest; c) to evaluate computationally protein-ligand complexes and d) to evaluate the effect of ligands on gene expression and on carcinoma cell behavior. Methods: The search for potential markers related to HNSCC was performed by literature mining, following a flow chart that included selection of scientific articles in PubMed by MeSH terms, association of articles with genes/proteins through the gene2pubmed file, selection of genes in external data bases and manual curation steps. In order to identify potential ligands, proteins related to HNSCC and involved in inflammatory processes were used to perform molecular docking assays with known anti-inflammatory drugs. Finally, the role of piplartine, a substance extracted from the Piper longum with anti-inflammatory and antineoplastic effects, on proliferation, migration and gene expression was investigated in neoplastic cells. Results: The curated gene-to-publication assignment yielded a total of 1,370 genes related to HNSCC, with specificity of 74% and sensibility of 87%. The diversity of results allowed identifying new and mostly unexplored gene associations, revealing, for example, that processes linked to response to steroid hormone stimulus are significantly enriched with genes related to HNSCC. The results also showed that piplartine decreases viability and cell migration, and alters expression of genes involved in inflammatory responses. Conclusion: This approach allows the identification of genes related to HNSCC and is able to reveal new associations that deserve to be further studied. Piplartine, the compound selected for in vitro studies, interacts with molecular targets similar to known anti-inflammatory drugs, decreases proliferation and cell migration, and alter the expression of genes associated with HNSCC and inflammatory processes. / Introdução: A literatura científica sobre câncer tem crescido rapidamente nos últimos anos, o que torna difícil, se não impossível, a tarefa de recuperar e analisar manualmente as informações relevantes sobre os mecanismos que governam o processo neoplásico. Além disso, o câncer é uma doença complexa e sua heterogeneidade é particularmente evidente no câncer epidermoide de cabeça e pescoço (CECP), um dos tipos mais comuns de câncer em todo o mundo. Objetivos: Os objetivos do estudo foram (a) identificar genes/proteinas relacionados a CECP a partir de dados da literatura, (b) identificar ligantes que interajam eficiente e especificamente com alvos moleculares selecionados, (c) avaliar computacionalmente o complexo proteína/ligante e (d) avaliar a ação de ligantes na expressão gênica e no comportamento de células de carcinoma. Métodos: A busca de marcadores potenciais relacionados a CECP utilizou a mineração da literatura disponível publicamente, seguindo um fluxograma que incluiu seleção de artigos científicos no PubMed por termos MeSH, associação de artigos com genes/proteínas por meio do arquivo gene2pubmed, seleção de genes em bancos de dados externos e etapas de curação manual. As proteínas identificadas como sendo relacionadas a CECP que apresentaram envolvimento em processos inflamatórios foram submetidas a experimento de docking molecular para identificação de seus ligantes entre drogas disponíveis no mercado. Finalmente, o papel da piplartina, uma substância natural extraída da pimenta Piper longum com evidências de ação anti-inflamatória e antineoplásica, foi avaliado na proliferação, na migração e na expressão gênica de células neoplásicas. Resultados: Um total de 1370 genes relacionados a CECP foi identificado pela abordagem proposta, que mostrou especificidade de 74% e sensibilidade de 87%. A diversidade dos dados permitiu obter associações potenciais ainda não exploradas, revelando, por exemplo, que a resposta ao estímulo esteróide hormonal está significativamente enriquecida com genes relacionados a CECP. Os resultados também mostraram que a piplartina reduz a viabilidade e a migração celular e modifica o padrão de expressão de um painel de genes que atuam em processos inflamatórios. Conclusão: A abordagem empregada permite a identificação de genes relacionados à CECP e revela novas associações que merecem ser estudadas. O composto piplartina selecionado para estudos in vitro interage com alvos moleculares de forma semelhante à de medicamentos anti-inflamatórios conhecidos e é capaz de diminuir a proliferação e a migração celular e de alterar a expressão de genes relacionados à CECP e a processos inflamatórios.

Sequences are ubiquitous in the domain of biology. One of the current best machine learning techniques for analysing sequences is the long short-term memory (LSTM) network. Owing to significant barriers to adoption in biology, focussed efforts are required to realize the use of LSTMs in practice. Thus, the aim of this work is to improve the state of LSTMs for biology, and we focus on biological tasks pertaining to physiological signals, peripheral neural signals, and molecules. This goal drives the three subplots in this thesis: biological applications, visualizations, and extensions. We start by demonstrating the utility of LSTMs for biological applications. On two new physiological-signal datasets, LSTMs were found to outperform hidden Markov models. LSTM-based models, implemented by other researchers, also constituted the majority of the best performing approaches on publicly available medical datasets. However, even if these models achieve the best performance on such datasets, their adoption will be limited if they fail to indicate when they are likely mistaken. Thus, we demonstrate on medical data that it is straightforward to use LSTMs in a Bayesian framework via dropout, providing model predictions with corresponding uncertainty estimates. Another dataset used to show the utility of LSTMs is a novel collection of peripheral neural signals. Manual labelling of this dataset is prohibitively expensive, and as a remedy, we propose a sequence-to-sequence model regularized by Wasserstein adversarial networks. The results indicate that the proposed model is able to infer which actions a subject performed based on its peripheral neural signals with reasonable accuracy. As these LSTMs achieve state-of-the-art performance on many biological datasets, one of the main concerns for their practical adoption is their interpretability. We explore various visualization techniques for LSTMs applied to continuous-valued medical time series and find that learning a mask to optimally delete information in the input provides useful interpretations. Furthermore, we find that the input features looked for by the LSTM align well with medical theory. For many applications, extensions of the LSTM can provide enhanced suitability. One such application is drug discovery -- another important aspect of biology. Deep learning can aid drug discovery by means of generative models, but they often produce invalid molecules due to their complex discrete structures. As a solution, we propose a version of active learning that leverages the sequential nature of the LSTM along with its Bayesian capabilities. This approach enables efficient learning of the grammar that governs the generation of discrete-valued sequences such as molecules. Efficiency is achieved by reducing the search space from one over sequences to one over the set of possible elements at each time step -- a much smaller space. Having demonstrated the suitability of LSTMs for biological applications, we seek a hardware efficient implementation. Given the success of the gated recurrent unit (GRU), which has two gates, a natural question is whether any of the LSTM gates are redundant. Research has shown that the forget gate is one of the most important gates in the LSTM. Hence, we propose a forget-gate-only version of the LSTM -- the JANET -- which outperforms both the LSTM and some of the best contemporary models on benchmark datasets, while also reducing computational cost.

Os receptores ativadores da proliferação de peroxissomos (PPARs) são fatores de transcrição dependentes da ligação de ligantes e possuem um papel chave no controle do metabolismo dos lipídios e da glicose. Existem três isotipos desse receptor: PPARα, PPARβ e PPARγ. O PPARγ é alvo molecular para os compostos TZDs, os quais são fármacos usados clinicamente no controle da diabetes do tipo 2, aumentando a sensibilidade à insulina. Enquanto que os fibratos são os fármacos que atuam no PPARα e são utilizados para diminuir os níveis de triglicerídeos. A maioria dos pacientes que sofrem com a diabetes do tipo 2 apresentam desordens no metabolismo de lipídios. Mesmo com a existência de fármacos capazes de controlar estas desordens metabólicas, a busca de um agonista dual para os PPARα e PPARγ é um grande desafio no controle da síndrome metabólica, uma vez que este composto pode combinar os dois efeitos terapêuticos em uma única molécula. O GL479 é um agonista dual que foi sintetizado com dois grupos farmacóforos, ligando-se tanto ao PPARα quanto ao PPARγ. Dentro desse contexto, este estudo apresenta as bases estruturais de interação do agonista dual GL479 aos PPARs por meio da determinação estrutural dos complexos PPARα-LBD:GL479 e PPARγ-LBD:GL479. A análise detalhada desses complexos revelou diferentes modos de interação do ligante em cada receptor, porém em ambos os casos o GL479 interage com a Tyr da H12. Na estrutura do PPARα-LBD, o ligante adquiriu a característica de um agonista total e no caso do PPARγ-LBD, o GL479 adotou características de um agonista parcial dependente da interação com a H12. Além das analises do agonista dual, 16 compostos foram identificados por docking como ligantes do PPARγ. Três desses ligantes (8, 10 e 15) foram caracterizados por ThermoFluor e fluorescência de polarização com valores de IC50 menor que 10 µM. Adicionalmente, um dos compostos identificados no docking (16) foi cocristalizado com PPARγ-LBD. A conformação adotada pelo ligante não permitiu que ele interagisse diretamente com a H12, sugerindo que este composto possa atuar como um agonista parcial independente da H12. Todas estas descobertas podem ser exploradas no desenho de novos moduladores dos PPARs com menores efeitos adversos ou até mesmo na busca de agonistas duais PPARα ⁄γ, que combine os efeitos terapêuticos no tratamento da diabetes do tipo 2 e da dislipidemia. / Peroxisome proliferator-activated receptors (PPARs) are ligand-dependent transcription factors that control various functions in human organism and they play key roles in the control of glucose and lipid metabolism. There are three different PPAR isotypes: PPARα, PPARβ e PPARγ. PPARγ is a molecular target of TZD agonists, which are clinically used drugs in the control of type 2 diabetes by increasing insulin sensitivity. Whereas fibrates are drugs that act on PPARα and are used to lower serum triglyceride levels. The most patients who have type 2 diabetes also display lipid metabolism disorders. Even with the existence of drugs that can control these metabolic disorders, the search of dual agonist for PPARα and PPAR γ is a major challenge in the control of metabolic syndrome, because this compound could combine both therapeutic effects in a single molecule. GL479 is a dual agonist that was synthesized based on a combination of two key pharmacophores, with the ability to bind in the both PPARs, α, and γ. Thus, this study reveals the structural basis for this dual agonist GL479 by structural determination of the complexes PPARα-LBD:GL479 and PPARγ-LBD:GL479. The detailed analysis of these complexes showed different ligand binding modes for each receptor, however, in the both cases the GL479 interacted with the Tyr of H12. In the PPARα-LBD structure the ligand acquired the features of full agonist and in the case of PPARγ-LBD, GL479 adopted features of a partial agonist dependent of H12 interaction. In addition to the dual agonist analysis, sixteen compounds were identified as PPARγ ligand by docking. Three of these ligands were characterized by ThermoFluor and fluorescence polarization, which resulted in IC50 values smaller than 10 µM. Additionally, one of the compounds, identified by docking, was co-crystallized with PPARγ. The ligand conformation adopted would not allow it a direct interaction with the H12. These contacts were mediated by one water molecule, suggesting this compound might also act as a partial agonist, independent of H12 interaction. All these findings may be explored for the design of PPARs novel modulators with lower side effects, as well, in the exploration of dual agonists PPARα ⁄ γ that combines the therapeutic effects in the treatment of type 2 diabetes and dyslipidemia.

Conventional screening of potential drug candidates through wet lab affinity experiments using libraries of thousands of modified molecules is time and resource consuming, along with the fact that it contributes to the widening time gap between the discovery of disease-causing mutations and the implementation of resulting novel treatments. It is necessary to explore whether the preliminary use of high-throughput virtual screening (HTVS) software such as PyRx will curb both the time and money spent in discovering novel treatments for diseases such as congenital heart defects (CHDs). For example, AXIN2, a protein involved in a negative feedback loop inhibiting the Wnt/β-catenin signaling pathway important for cardiogenesis, has recently been associated with CHD. The loss-of-function mutation L10F on the tankyrase-binding domain of AXIN2 has been shown to upregulate the pathway by loss of inhibition ability, leading to the accumulation of intracellular β-catenin. In a different paper, however, AXIN2 has been shown to be stabilized using XAV-939, a small-molecule drug which targets tankyrase. PyRx and VMD will be used to modify the drug in order to increase its binding affinity to AXIN2, stabilizing the protein and reinstating its inhibitory property to treat CHDs. When used in adjunction to wet lab experiments, HTVS software may decrease costs and the time required to bring a potentially life-saving treatment into use.

In order to reduce the time associated with and the costs of drug discovery, machine learning is being used to automate much of the work in this process. However the size and complex nature of molecular data makes the application of machine learning especially challenging. Much work must go into the process of engineering features that are then used to train machine learning models, costing considerable amounts of time and requiring the knowledge of domain experts to be most effective. The purpose of this work is to demonstrate data driven approaches to perform the feature selection and extraction steps in order to decrease the amount of expert knowledge required to model interactions between proteins and drug molecules.

La possibilité d'utiliser une nouvelle méthodologie pour l'échange de liaisons peptidiques dans des conditions aérobiques et biocompatibles a été étudiée. Nous décrivons l'optimisation de la ligation chimique native combinatoire dynamique (dynNCL) au niveau d'un résidu N-méthylcystéine en utilisant des peptides modèles. Nous employons en outre cette méthode optimisée pour le criblage de bibliothèques de peptides combinatoires dynamiques en présence et en l'absence d'un gabarit de type anticorps. L'effet que l'incorporation d'une jonction dynamique au sein de ligands non-anticorps a sur l'affinité a également été étudié. Nous proposons que dynNCL peut être utilisé pour la conception d'une nouvelle classe de séquences pouvant potentiellement conduire au premier exemple d'épissage de protéines artificielles. / The possibility to use a new methodology for peptide bond exchange in aerobic biocompatible conditions has been investigated. We describe the assay optimization of dynamic combinatorial native chemical ligation (dynNCL) at the N-methyl-cysteine residue using model peptides. We further employ this optimized method for the screening of dynamic combinatorial peptide libraries in the presence and the absence of an antibody template. The effect that dynamic junction incorporation into non-antibody ligands has upon affinity was also studied. We propose that dynNCL can be used for the creation of a new class of designer sequences which can potentially provide the first example of artificial protein splicing.