Submitted to the faculty of the Chemical Informatics Graduate Program in partial fulfillment of the requirements for the degree Master of Science in the School of Informatics,Indiana University, December 2005 / The completion of Human Genome Project is seen as a gateway to the discovery of novel drug targets (Jacoby, Schuffenhauer, & Floersheim, 2003). How much of this information is actually translated into knowledge, e.g., the discovery of novel drug targets, is yet to be seen. The traditional route of drug discovery has been from target to compound. Conventional research techniques are focused around studying animal and cellular models which is followed by the development of a chemical concept. Modern approaches that have evolved as a result of progress in molecular biology and genomics start out with molecular targets which usually originate from the discovery of a new gene .Subsequent target validation to establish suitability as a drug target is followed by high throughput screening assays in order to identify new active chemical entities (Hofbauer, 1997). In contrast, chemogenomics takes the opposite approach to drug discovery (Jacoby, Schuffenhauer, & Floersheim, 2003). It puts to the forefront chemical entities as probes to study their effects on biological targets and then links these effects to the genetic pathways of these targets (Figure 1a). The goal of chemogenomics is to rapidly identify new drug molecules and drug targets by establishing chemical and biological connections. Just as classical genetic experiments are classified into forward and reverse, experimental chemogenomics methods can be distinguished as forward and reverse depending on the direction of investigative process i.e. from phenotype to target or from target to phenotype respectively (Jacoby, Schuffenhauer, & Floersheim, 2003). The identification and characterization of protein targets are critical bottlenecks in forward chemogenomics experiments. Currently, methods such as affinity matrix purification (Taunton, Hassig, & Schreiber, 1996) and phage display (Sche, McKenzie, White, & Austin, 1999) are used to determine targets for compounds. None of the current techniques used for target identification after the initial screening are efficient. In silico methods can provide complementary and efficient ways to predict targets by using chemogenomics databases to obtain information about chemical structures and target activities of compounds. Annotated chemogenomics databases integrate chemical and biological domains and can provide a powerful tool to predict and validate new targets for compounds with unknown effects (Figure 1b). A chemogenomics database contains both chemical properties and biological activities associated with a compound. The MDL Drug Data Report (MDDR) (Molecular Design Ltd., San Leandro, California) is one of the well known and widely used databases that contains chemical structures and corresponding biological activities of drug like compounds. The relevance and quality of information that can be derived from these databases depends on their annotation schemes as well as the methods that are used for mining this data. In recent years chemists and biologist have used such databases to carry out similarity searches and lookup biological activities for compounds that are similar to the probe molecules for a given assay. With the emergence of new chemogenomics databases that follow a well-structured and consistent annotation scheme, new automated target prediction methods are possible that can give insights to the biological world based on structural similarity between compounds. The usefulness of such databases lies not only in predicting targets, but also in establishing the genetic connections of the targets discovered, as a consequence of the prediction. The ability to perform automated target prediction relies heavily on a synergy of very recent technologies, which includes: i) Highly structured and consistently annotated chemogenomics databases. Many such databases have surfaced very recently; WOMBAT (Sunset Molecular Discovery LLC, Santa Fe, New Mexico), KinaseChemBioBase (Jubilant Biosys Ltd., Bangalore, India) and StARLITe (Inpharmatica Ltd., London, UK), to name a few. ii) Chemical descriptors (Xue & Bajorath, 2000) that capture the structure-activity relationship of the molecules as well as computational techniques (Kitchen, Stahura, & Bajorath, 2004) that are specifically tailored to extract information from these descriptors. iii) Data pipelining environments that are fast, integrate multiple computational steps, and support large datasets. A combination of all these technologies may be employed to bridge the gap between chemical and biological domains which remains a challenge in the pharmaceutical industry.

Chemical genomics is an interdisciplinary field that combines small molecule perturbation with genomics to understand gene function and to study the mode(s) of drug action. Existing methods for correlating chemical genomic profiles are not ideal as they often require one to define the disrupting effects, commonly known as batch effects. These effects are not always known, and they can mask true biological differences. I present a method, Bucket Evaluations (BE), which surmounts these problems. This method is a non-parametric correlation approach, which is suitable for locating correlations in somewhat perturbed datasets such as chemical genomic profiles. BE can be used on other datasets such as those obtained via gene expression profiling and performs well on both array-based and sequence based readouts. Using BE, along with various correlation methods, on a collection of datasets, showed it to be highly accurate for locating similarity between experiments.

Chemical genomics is an interdisciplinary field that combines small molecule perturbation with genomics to understand gene function and to study the mode(s) of drug action. Existing methods for correlating chemical genomic profiles are not ideal as they often require one to define the disrupting effects, commonly known as batch effects. These effects are not always known, and they can mask true biological differences. I present a method, Bucket Evaluations (BE), which surmounts these problems. This method is a non-parametric correlation approach, which is suitable for locating correlations in somewhat perturbed datasets such as chemical genomic profiles. BE can be used on other datasets such as those obtained via gene expression profiling and performs well on both array-based and sequence based readouts. Using BE, along with various correlation methods, on a collection of datasets, showed it to be highly accurate for locating similarity between experiments.

Comprendre les interactions réalisées entre un candidat médicament et sa protéine cible est un enjeu crucial pour orienter la recherche de nouvelles molécules. En effet, ce processus implique de nombreux paramètres qu’il est nécessaire d’analyser séparément pour mieux comprendre leurs effets.Nous proposons ici deux nouvelles approches observant les relations protéine/ligand. La première se concentre sur la comparaison de cavités formées par les sites de liaison pouvant accueillir une molécule. Cette méthode permet d’inférer la fonction d’une protéine mais surtout de prédire « l’accessibilité » d’un site de liaison pour un médicament. La seconde tactique se focalise sur la comparaison des interactions non-covalentes réalisées entre la protéine et le ligand afin d’améliorer la sélection de molécules potentiellement actives lors de criblages virtuels, et de rechercher de nouveaux fragments moléculaires, structuralement différents mais partageant le même mode d’interaction. / Understanding the interactions between a drug and its target protein is crucial in order to guide drug discovery. Indeed, this process involves many parameters that need to be analyzed separately to better understand their effects.We propose two new approaches to observe protein/ligand relationships. The first focuses on the comparison of cavities formed by binding sites that can accommodate a small molecule. This method allows to infer the function of a protein but also to predict the accessibility of a binding site for a drug. The second method focuses on the comparison of non-covalent interactions made between the protein and the ligand to improve the selection of potentially active molecules in virtual screening, and to find new molecular fragments, structurally different but sharing the same mode of interaction.

RUNX1 est un facteur de transcription essentiel pour l’hématopoïèse et joue un rôle important dans la fonction immunitaire. Des mutations surviennent dans ce gène chez 5 à 13% des patients atteints de leucémie myéloïde aiguë (LMA) (RUNX1mut) et définissent un sous-groupe particulier de LMA associé à un pronostic défavorable. En conséquence, il est nécessaire de procéder à une meilleure caractérisation génétique et de concevoir des stratégies thérapeutiques plus efficaces pour ce sousgroupe particulier de LMA. Bien que la plupart des mutations trouvées dans le gène RUNX1 dans la LMA soient supposément acquises, des mutations germinales dans RUNX1 sont observées chez les patients atteints du syndrome plaquettaire familial avec prédisposition aux hémopathies malignes (RUNX1-FPD, FPD/AML). En outre, 44 % des individus atteints évoluent vers le développement d’une LMA. Suite au séquençage du transcriptome (RNA-Seq) d’échantillons de la cohorte Leucégène, nous avons montré que le dosage allélique de RUNX1 influence l’association avec des mutations coopérantes, le profil d’expression génique et la sensibilité aux médicaments dans les échantillons primaires de LMA RUNX1mut. Aussi, la validation des mutations trouvées chez RUNX1 a mené à la découverte que 30% des mutations identifiées dans notre cohorte de LMA étaient d’origine germinale, révélant une proportion plus élevée qu’attendue de cas de mutations RUNX1 familiales. Un crible chimique a, quant à lui, révélé que la plupart des échantillons RUNX1mut sont sensibles aux glucocorticoïdes (GCs) et nous avons confirmé que les GCs inhibent la prolifération des cellules de LMA et ce, via l’interaction avec le récepteur des glucocorticoïdes (Glucocorticoid Receptor, GR). De plus, nous avons observé que les échantillons contenant des mutations RUNX1 censées entraîner une faible activité résiduelle étaient plus sensibles aux GCs. Nous avons aussi observé que la co-association de certaines mutations, SRSF2mut par exemple, et les niveaux de GR contribuaient à la sensibilité aux GCs. Suite à cela, la sensibilité acquise aux GCs a été obtenue en régulant négativement l’expression de RUNX1 dans des cellules LMA humaines, ce qui a été accompagné par une régulation positive de GR. L’analyse de transcriptome induit par GC a révélé que la différenciation des cellules de LMA induite par GCs pourrait être un mécanisme en jeu dans la réponse antiproliférative associée à ces médicaments. Plus important encore, un criblage génomique fonctionnel a identifié le répresseur transcriptionnel PLZF (ZBTB16) comme un modulateur spécifique de la réponse aux GCs dans les cellules LMA sensibles et résistantes. Ces observations fournissent une caractérisation supplémentaire de la LMA RUNX1mut, soulignant l’importance de procéder à des tests germinaux pour les patients porteurs de mutations RUNX1 délétères. Nos résultats ont également identifié un nouveau rôle pour RUNX1 dans le réseau de signalisation de GR et montrent l’importance d’investiguer le repositionnement des GCs pour traiter la LMA RUNX1mut dans des modèles précliniques. Enfin, nous avons fourni des indications sur le mécanisme d’action des GCs, en montrant que PLZF s’avère un facteur important favorisant la résistance aux GCs dans la LMA. / RUNX1 is an essential transcription factor for definite hematopoiesis and plays important roles in immune function. Mutations in RUNX1 occur in 5-13% of Acute Myeloid Leukemia (AML) patients (RUNX1mut ) and are associated with adverse outcome, thus highlighting the need for better genetic characterization and for the design of efficient therapeutic strategies for this particular AML subgroup. Although most RUNX1 mutations in AML are believed to be acquired, germline RUNX1 mutations are observed in the familial platelet disorder with predisposition to hematologic malignancies (RUNX1-FPD, FPD/AML) in which about 44% of affected individuals progress to AML. By performing RNA-sequencing of the Leucegene collection, we revealed that RUNX1 allele dosage influences the association with cooperating mutations, gene expression profile, and drug sensitivity in RUNX1mut primary AML specimens. Validation of RUNX1 mutations led to the discovery that 30% of RUNX1 mutations in our AML cohort are of germline origin, indicating a greater than expected proportion of cases with familial RUNX1 mutations. Chemical screening showed that most RUNX1mut specimens are sensitive to glucocorticoids (GC) and we confirmed that GCs inhibit AML cell proliferation via interaction with the Glucocorticoid Receptor (GR). We observed that specimens harboring RUNX1 mutations expected to result in low residual RUNX1 activity were most sensitive to GCs, and that co-associating mutations, such as SRSF2mut, as well as GR levels contribute to GC-sensitivity. Accordingly, acquired GC-sensitivity was achieved by negatively regulating RUNX1 expression in human AML cells, which was accompanied by upregulation of the GR. GC-induced transcriptome analysis revealed that GC-induced differentiation of AML cells might be a mechanism at play in the antiproliferative response to these drugs. Most critically, functional genomic screening identified the transcriptional repressor PLZF (ZBTB16) as a specific modulator of the GC response in sensitive and resistant AML cells. These findings provide additional characterization of RUNX1mut AML, further stressing the importance of germline testing for patients carrying deleterious RUNX1 mutations. Our results also identified a novel role for RUNX1 in the GR signaling network and support the rationale of investigating GC repurposing for RUNX1mut AML in preclinical models. Finally, we provided insights into the mechanism of action of GCs, which positions PLZF as an important factor promoting resistance to glucocorticoids in AML.