<p>Bacterial L-asparaginases (E.C.3.5.1.1) are enzymes that catalyze the hydrolysis of L-asparagine to aspartic acid. For the past 30 years these enzymes have been used as therapeutic agents in the treatment of acute childhood lymphoblastic leukemia. The presence of a low rate glutaminase activity however causes serious side-effects to patients in treatment, as glutamine depletion give rise to neurotoxicity, anaphylaxis, and other hypersensitivity reactions. The interest in the enzyme from Erwinia carotovora originates from the fact that it shows a decreased glutaminase activity, and therefore the enzyme is expected to exhibit fewer side effects when used in therapy.</p><p>The main focus of this thesis is the crystal structure determination of L-asparaginase from Erwinia carotovora in the presence of aspartic acid at 2.5 Å resolution. The structure was refined to an R/Rfree factor of 19.9/28.6 with good stereochemistry.</p><p>L-Asparaginases are homotetrameric enzymes with a known 222 symmetry and an identical fold. The Erwinia carotovora asparaginase consists of eight monomers of 330 amino acid residues each. In this case the enzyme is active as a dimer of tetramers. The two tetramers have an inner twofold non-crystallographic symmetry. Each monomer forms two identifiable domains a large N-domain and a small C-domain. The active sites are found at a topological switch-point between those domains.</p>

Bacterial L-asparaginases (E.C.3.5.1.1) are enzymes that catalyze the hydrolysis of L-asparagine to aspartic acid. For the past 30 years these enzymes have been used as therapeutic agents in the treatment of acute childhood lymphoblastic leukemia. The presence of a low rate glutaminase activity however causes serious side-effects to patients in treatment, as glutamine depletion give rise to neurotoxicity, anaphylaxis, and other hypersensitivity reactions. The interest in the enzyme from Erwinia carotovora originates from the fact that it shows a decreased glutaminase activity, and therefore the enzyme is expected to exhibit fewer side effects when used in therapy. The main focus of this thesis is the crystal structure determination of L-asparaginase from Erwinia carotovora in the presence of aspartic acid at 2.5 Å resolution. The structure was refined to an R/Rfree factor of 19.9/28.6 with good stereochemistry. L-Asparaginases are homotetrameric enzymes with a known 222 symmetry and an identical fold. The Erwinia carotovora asparaginase consists of eight monomers of 330 amino acid residues each. In this case the enzyme is active as a dimer of tetramers. The two tetramers have an inner twofold non-crystallographic symmetry. Each monomer forms two identifiable domains a large N-domain and a small C-domain. The active sites are found at a topological switch-point between those domains.

Ethyl pyruvate, a known ROS scavenger and anti-inflammatory drug was found to combat leukemia cells. Tumor cell killing was achieved by concerted action of necrosis/apoptosis induction, ATP depletion, and inhibition of glycolytic and para-glycolytic enzymes. Ethyl lactate was less harmful to leukemia cells but was found to arrest cell cycle in the G0/G1 phase. Both, ethyl pyruvate and ethyl lactate were identified as new inhibitors of GSK-3β. Despite the strong effect of ethyl pyruvate on leukemia cells, human cognate blood cells were only marginally affected. The data were compiled by immune blotting, flow cytometry, enzyme activity assay and gene array analysis. Our results inform new mechanisms of ethyl pyruvate-induced cell death, offering thereby a new treatment regime with a high therapeutic window for leukemic tumors.

Adequate cell production for adoptive cell transfer therapies such as Chimeric Antigen Receptor (CAR)-T cell therapy remains a critical barrier to treatment for indications that fail to achieve clinical success. One such disease is Chronic Lymphocytic Leukemia (CLL), a B-cell lymphoma with their characteristically exhausted T cells, marked by a progressive loss of the ability to secrete cytokines and proliferate, as well as an increase in the expression of checkpoint inhibitor molecules such as PD-1. The goal of this thesis is to characterize the functional differences or specific biomarkers within the CLL patient population that is indicative of the proliferation outcomes. Conventional clinical markers such as Rai stage or PD-1 expression alone were inadequate to describe the complex variability among patients. In order to better characterize exhaustion using microscopy-based cell function assays, we developed a sample sparing microscopy chamber that requires as little as 1000 cells per sample. The microscopy chambers were mass produced via injection molding, and made compatible with the antibody microcontact printing technique developed in the Kam lab. The chambers typically reduced cell usage per experiment by 20-fold. This reduction allowed us to measure IL-2 secretion, T cell arrest response to activating antibody patterns (pattern alignment), and motility of scarce human samples simultaneously from a single experiment. Results from these functional readouts along with other clinical markers were used as inputs for a multifactor exploratory analysis to cluster patients according to their functional similarities from the combination of responses in an unbiased manner. The resulting clusters based on the combination of the top 3 parameters IL-2, pattern alignment, and PD-1 resulted in better separation of patient groups and provided a basis for predicting max doubling outcomes from these inputs. We further used motility measurements as a way to understand initial T cell response to activation before the stop response, which was measured as pattern alignment previously. The time it takes for cells to come to a stop at the signal was most informative for translating T cell activation response to a stop response, and eventually to downstream effector functions of cytokine secretion and proliferation. The results of this work provide a powerful framework to describe different donors, and can be applied to cells from additional donors to guide future cell expansion studies.

Treatment of acute myeloid leukemia, an aggressive hematopoietic malignancy of myeloid progenitors, has remained rather stagnant over the course of several decades. Infusions of cytarabine and anthracycline antibiotics have dominated the landscape of AML therapy, with minor changes to dosing schedule occasionally making slight adjustments to efficacy or tolerability. Improvements in prognosis have been bittersweet, with most progress seen in younger populations less likely to get the disease, and already more likely to achieve remission and to meet survival milestones. Much of this progress is attributed to other factors, such as improved supportive care and availability of hematopoietic stem cell and platelet transfusion. In most patients, occupying the 60-and-above demographic, improvements in survival have not been significant. In turn, the population impact of AML has changed little over time. While accounting for about one-third of total leukemia cases and one percent of total cancer cases, AML accounts for about one half of total leukemia deaths and two percent of total cancer deaths. Most advances straying away from standard treatment have been in important pathways that could be impactful in subsets of the overall AML patient population. Tyrosine kinases are implicated in numerous cancers including AML, with activity-enhancing mutations conferring growth advantages to malignant cells. About one-third of AML patients have mutations in one such kinase, FLT3, and may benefit from inhibitors to tyrosine kinases overall and from FLT3- specific agents. Mutations in isocitrate dehydrogenases highlight another subpopulation, about one-fifth of AML patients, who might benefit from emerging agents that inhibit these pathways from creating a leukemia-favoring environment in the bone marrow. Other pathways similarly implicated in numerous cancers including AML are being targeted with new agents that can benefit some AML patients, such as Hedgehog signaling and apoptotic regulation. Still, breakthroughs are needed that can help most AML patients, particularly in the cases of relapsed leukemia that occurs in most patients within a year or two after remission is achieved. CD33 is among a few molecular targets for AML, though it is just as ubiquitously expressed on healthy myeloid cells. Antibody-drug conjugates like Mylotarg have made progress in this approach, though hematopoietic toxicities have made treatment difficult in older populations. Clever techniques such as ablation of CD33 from healthy myeloid progenitors may be supportive in CD33-based approaches, and immunotherapy involving CD33-targeting is a rapidly growing research focus. This dissertation describes a new type of bispecific antibody that binds CD33 on AML and CD3 on cytotoxic T cells in a proof-of-concept study. Various formats for bifunctional molecules have been created and used clinically, including antibody-drug conjugates and bispecific antibodies that simultaneously engage antigens on two different types of cells. Those like the one described here, bispecific T-cell engagers, have typically taken the form of single-chain fusion proteins containing the variable regions binding to both antigens of interest. Other bispecific antibodies have imitated naturally-occurring immunoglobulin structures, boasting superior pharmacokinetics while facing steep obstacles in large-scale production. The single-chain fusions, easier to produce, can face difficulties in full engagement, with loss of function sometimes seen in fusion partners at the C-terminus. We propose a new format, believed to present two antigen-binding domains in N-terminal positions on a two-chain heterodimeric structure. Capitalizing on an elegantly designed system of hydrophobic cores and hydrogen-bonding networks generating an orthogonal heterodimer, we added an immunoglobulin hinge region to secure a permanently-bound heterodimer, and attached domains binding to CD3 and CD33. We hypothesized that this design, ensured to present its antibody components at N-termini, could bind two antigens at a distance appropriate for facilitating T cell cytotoxicity to AML. After expressing and purifying these proteins in mammalian cells, we demonstrated their ability to persist as a bispecific heterodimer. We showed in vitro that our bispecific heterodimers could bind both CD3+ and CD33+ cells, and that they bolstered T cell cytotoxicity to AML cell lines in a dose-dependent manner. Monomeric components bound only CD3+ or CD33+ cells depending on antibody variable domain present, and had no effect on T cell cytotoxicity. In a mouse model of minimal residual disease, T cells alone did not have a significant effect on the growth of AML, nor did they have an effect on overall survival. T cells with bispecific heterodimer greatly extended survival, and mice of this treatment group were free of leukemia. These findings suggest that this format for bispecific proteins allows for robust simultaneous engagement with both antigens of interest in a manner conducive to T cell cytotoxicity against AML. We believe this presents a compelling modular system for bispecific antibodies, where CD3- and CD33-binding domains can be readily swapped with domains binding to other cancer- or immune cell-specific antigens, and can be further developed into a trispecific system engaging other immune cells or extending half-life with anti-albumin or Fc domains.

Synthetic biology enables the engineering of interactions between living medicines to overcome the specific limitations of monotherapies. A major challenge facing tumor-antigen targeting therapies like chimeric antigen receptor (CAR)-T cells is the identification of suitable targets that are specifically and uniformly expressed on heterogeneous solid tumors. In contrast, certain strains of bacteria are gaining recognition as a new class of antigen-agnostic cell therapy due to their selective growth within the immunosuppressive niche of the solid tumor microenvironment (TME). In response, this dissertation aims to pair the cytotoxicity of CAR-T cells with the antigen-independent specificity of tumor-colonizing bacteria to create a new strategy for solid tumor recognition. Here, we reprogram the probiotic strain of E. coli Nissle 1917 to release synthetic CAR targets and human chemokines directly within the solid tumor core. To enable universal targeting, we design synthetic targets to bind ubiquitous components of the TME and broadly tag tumor tissue for CAR-mediated lysis. We demonstrate that these targets robustly coat the surface of cancer cell lines and lead to effective killing by CAR-T cells across various cancer types. We additionally show that injected probiotics selectively grow within the tumor core and maintain target production ¬ in situ – leading to therapeutic efficacy across multiple genetically distinct tumor models. Within this dissertation, we also reveal that intratumoral bacteria provide natural adjuvant effects that serve to activate and increase the effector functions of CAR-T cells in vivo. However, we discover that this can lead to early T cell exhaustion and terminal effector differentiation. To mitigate the counterproductive effects of overstimulation, we generate a new probiotic strain with reduced inflammatory properties that significantly improves CAR-T cell phenotype – leading to enhanced therapeutic benefit in a human model of leukemia. We conclude by discussing the numerous avenues available to optimize cross-Kingdom signaling and to ultimately leverage the full therapeutic benefit of combined cell therapies for future translation. Altogether, this dissertation highlights the potential of the probiotic-guided CAR-T cell (ProCAR) platform to address the critical roadblock of identifying suitable CAR targets by providing an antigen in situ that is orthogonal to both healthy tissue and tumor genetics – and, in turn, aims to establish the foundation for engineered communities of living medicines.

Immune cell function varies tremendously between individuals, posing a major challenge to the development and success of emerging cellular immunotherapies. In the context of T cell therapy for cancer, long-term diseases such as Chronic Lymphocytic Leukemia (CLL) often induce T cell deficiencies resembling cellular exhaustion, complicating the preparation of therapeutic quantities of cells and maintaining efficacy once reintroduced to patients. The ability to rapidly estimate the responsiveness of an individual’s T cells could provide a powerful tool for tailoring treatment conditions and monitoring T cell functionality over the course of therapy. This dissertation investigates the use of short-term cellular behavior assays as a predictive indicator of long-term T cell function. Specifically, the short-term spreading of T cells on functionalized planar, elastic surfaces was quantified by 11 morphological parameters. These parameters were analyzed to discern the impact of both intrinsic factors, such as disease state, and extrinsic factors, such as substrate stiffness. This study identified morphological features that varied between T cells isolated from healthy donors and those from patients being treated for CLL. Combining multiple features through a machine learning approach such as Decision Tree or Random Forest provided an effective means for identifying whether T cells came from healthy or CLL donors. To further automate this assay and enhance the classification outcome, an image-based deep learning workflow was developed. The image-based deep learning approach notably outperformed morphometric analysis and showed great promise in classifying both intrinsic disease states and extrinsic environmental stiffness. Furthermore, we applied this imaging-based deep learning method to predict T cell proliferative capacity under different stiffness conditions, enabling rapid and efficient optimization of T cell expansion conditions to better guide cellular immunotherapy. Looking ahead, future efforts will focus on optimizing and generalizing the model to enhance its predictive accuracy and applicability across diverse patient populations. Additionally, we aim to incorporate multi-channel imaging that captures detailed T cell subset information, enabling the model to better understand the complex interactions between different cellular features and their influence on long-term proliferation. Our ultimate vision is to translate this technology into an automated device that offers a streamlined and efficient assessment of T cell functions. This device could serve as a critical tool in optimizing T cell production and monitoring T cell functions for both autologous and allogeneic cell therapies, significantly improving the effectiveness and personalization of cancer immunotherapy.

La leucémie lymphoïde chronique (LLC) est la plus fréquente des leucémies qui touchent le monde occidental. Cette pathologie est toujours incurable et se caractérise par une hétérogénéité d’évolution clinique marquée par une survie globale oscillant de quelques mois à des dizaines d’années. Il est donc primordial de savoir à quel type d’évolution le patient sera confronté afin d’adapter au mieux le suivi de la maladie et la précocité ou non du traitement.<p><p>\ / Doctorat en Sciences biomédicales et pharmaceutiques / info:eu-repo/semantics/nonPublished

La thèse est consacrée au développement de nouvelles méthodes de modélisations mathématiques en biologie et en médecine, du type “off-lattice" modèles hybrides discret-continus, et de leurs applications à l’hématopoïèse et aux maladies sanguines telles la leucémie et l’anémie. Dans cette approche, les cellules biologiques sont considérées comme des objets discrets alors que les réseaux intracellulaire et extracellulaire sont décrits avec des modèles continus régis par des équations aux dérivées partielles et des équations différentielles ordinaires. Les cellules interagissent mécaniquement et biochimiquement entre elles et avec le milieu environnant. Elles peuvent se diviser, mourir par apoptose ou se différencier. Le comportement des cellules est déterminé par le réseau de régulation intracellulaire et influencé par le contrôle local des cellules voisines ou par la régulation globale d’autres organes. Dans la première partie de la thèse, les modèles hybrides du type “off-lattice" dynamiques sont introduits. Des exemples de modèles, spécifiques aux processus biologiques, qui décrivent au sein de chaque cellule la concurrence entre la prolifération et l’apoptose, la prolifération et la différenciation et entre le cycle cellulaire et de l’état de repos sont étudiés. L’émergence des structures biologiques est étudiée avec les modèles hybrides. L’application à la modélisation des filamente de bactéries est illustrée. Dans le chapitre suivant, les modèle hybrides sont appliqués afin de modéliser l’érythropoïèse ou production de globules rouges dans la moelle osseuse. Le modèle inclut des cellules sanguines immatures appelées progéniteurs érythroïdes, qui peuvent s’auto-renouveler, se différencier ou mourir par apoptose, des cellules plus matures appelées les réticulocytes, qui influent les progéniteurs érythroïdes par le facteur de croissance Fas-ligand, et des macrophages, qui sont présents dans les îlots érythroblastiques in vivo. Les régulations intracellulaire et extracellulaire par les protéines et les facteurs de croissance sont précisées et les rétrocontrôles par les hormones érythropoïétine et glucocorticoïdes sont pris en compte. Le rôle des macrophages pour stabiliser les îlots érythroblastiques est montré. La comparaison des résultats de modélisation avec les expériences sur l’anémie chez les souris est effectuée. Le quatrième chapitre est consacré à la modélisation et au traitement de la leucémie. L’érythroleucémie, un sous-type de leucémie myéloblastique aigüe (LAM), se développe à cause de la différenciation insuffisante des progéniteurs érythroïdes et de leur auto-renouvellement excessif. Un modèle de type “Physiologically Based Pharmacokinetics-Pharmacodynamic” du traitement de la leucémie par AraC et un modèle de traitement chronothérapeutique de la leucémie sont examinés. La comparaison avec les données cliniques sur le nombre de blast dans le sang est effectuée. Le dernier chapitre traite du passage d’un modèle hybride à un modèle continu dans le cas 1D. Un théorème de convergence est prouvé. Les simulations numériques confirment un bon accord entre ces deux approches. / This dissertation is devoted to the development of new methods of mathematical modeling in biology and medicine, off-lattice discrete-continuous hybrid models, and their applications to modelling of hematopoiesis and blood disorders, such as leukemia and anemia. In this approach, biological cells are considered as discrete objects while intracellular and extracellular networks are described with continuous models, ordinary or partial differential equations. Cells interact mechanically and biochemically between each other and with the surrounding medium. They can divide, die by apoptosis or differentiate. Their fate is determined by intracellular regulation and influenced by local control from the surrounding cells or by global regulation from other organs. In the first part of the thesis, hybrid models with off-lattice cell dynamics are introduced. Model examples specific for biological processes and describing competition between cell proliferation and apoptosis, proliferation and differentiation and between cell cycling and quiescent state are investigated. Biological pattern formation with hybrid models is discussed. Application to bacteria filament is illustrated. In the next chapter, hybrid model are applied in order to model erythropoiesis, red blood cell production in the bone marrow. The model includes immature blood cells, erythroid progenitors, which can self-renew, differentiate or die by apoptosis, more mature cells, reticulocytes, which influence erythroid progenitors by means of growth factor Fas-ligand, and macrophages, which are present in erythroblastic islands in vivo. Intracellular and extracellular regulation by proteins and growth factors are specified and the feedback by the hormones erythropoietin and glucocorticoids is taken into account. The role of macrophages to stabilize erythroblastic islands is shown. Comparison of modelling with experiments on anemia in mice is carried out. The following chapter is devoted to leukemia modelling and treatment. Erythroleukemia, a subtype of Acute Myeloblastic Leukemia (AML), develops due to insufficient differentiation of erythroid progenitors and their excessive slef-renewal. A Physiologically Based Pharmacokinetics-Pharmacodynamics (PBPKPD) model of leukemia treatment with AraC drug and chronotherapeutic treatments of leukemia are examined. Comparison with clinical data on blast count in blood is carried out. The last chapter deals with the passage from a hybrid model to a continuous model in the 1D case. A convergence theorem is proved. Numerical simulations confirm a good agreement between these approaches.