Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2016. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Cytokine therapy can activate potent antitumor responses, yet collateral toxicity often limits dosages. Although immunocytokines have been designed with the intent to localize cytokine activity, systemic dose-limiting side effects are not fully ameliorated by attempted tumor targeting. In the first part of this work, we used the B 1 6F 10 melanoma model to demonstrate that a nontoxic dose of IL-2 immunocytokine synergized with tumor-specific antibody to significantly enhance therapeutic outcomes compared to monotherapy with immunocytokine, concomitant with increased tumor saturation and intratumoral cytokine responses. Examination of cell subset biodistribution showed that the immunocytokine associated mainly with IL-2R-expressing innate immune cells, with more bound immunocytokine present in systemic organs than in the tumor microenvironment. More surprisingly, immunocytokine antigen specificity and Fc[gamma]R interactions did not appear necessary for therapeutic efficacy or biodistribution patterns, as immunocytokines with irrelevant specificity and/or inactive mutant Fc domains behaved similarly to tumor-specific immunocytokine. IL-2-IL-2R interactions, rather than antibody-antigen targeting, dictated immunocytokine localization; however, the lack of tumor targeting did not preclude successful antibody combination therapy. This study presents a safe, straightforward strategy for augmenting immunocytokine efficacy via supplementary antibody dosing and explores underappreciated factors that can subvert efforts to purposefully alter cytokine biodistribution. Numerous studies have identified cancer immunotherapy combinations that exhibit synergistic antitumor activity, but surprisingly, these studies rarely consider the effects of relative dose timing. In the second part of this work, using established syngeneic tumor models, we found that staggering IFN[alpha] administration after, rather than simultaneously with, serum-persistent IL-2 and tumor-specific antibody significantly increased long-term survival and generated immunological memory. Successful combination therapy required IFNa-induced activation of cross-presenting CD8[alpha]+ DCs following release of antigenic tumor debris by the IL-2-and-antibody-mediated immune response. Due to decreased phagocytic ability post-maturation, DCs activated too early captured much less antigen and could not effectively prime CD8+ T cells. Temporally programming DC activation to occur after tumoricidal activity enhanced tumor control by multiple combination immunotherapies that act through distinct mechanistic pathways, presenting a facile strategy for augmenting efficacy in the combinatorial treatment setting and highlighting dose schedule as an overlooked factor that can profoundly affect the success of multi-component immunotherapies. / by Alice Tzeng. / Ph. D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2012. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Genomes experience an often hostile environment that creates a vast array of damages that can give rise to myriad biological outcomes. Fortunately, cells are equipped with networks such as direct reversal, base excision repair, nucleotide excision repair, homologous recombination, and translesion synthesis that help preserve informational integrity. The first part of this dissertation focuses on whether or not bulky alkyl lesions at the N2 atom of guanine are addressed in vivo by the DinB bypass polymerase. In the work described herein, a collection of N2-guanine lesions was inserted in single-stranded M13 genomes and evaluated in strains possessing or lacking DinB via the competitive replication and adduct bypass (CRAB) and restriction endonuclease and postlabeling (REAP) assays. It was found that DinB could in fact bypass the N2-furfuryl-guanine lesion and its saturated homolog in vivo. The second part of this work describes how we systematically investigated the role that the distance from an origin of replication may have in the mutagenesis of an adduct. Our hypothesis was that a lesion farther from the origin of replication would be less mutagenic since it would be afforded more time for detection and removal before the replicative polymerase traversed it, fixing the mutation. We inserted 0-methylguanine in single-stranded M13 genomes at different distances from the origin of replication and analyzed progeny phage by the REAP assay. Our findings were in contrast with the hypothesis; a higher mutation frequency was obtained at the site distal from the origin of replication. Alternative hypotheses and future experiments are discussed as part of this work. The third part of this dissertation seeks to expand the spectrum of known substrates for the enzyme AIkB, which mediates direct reversal of DNA damage. AlkB is an iron- and CCketoglutarate- dependent dioxygenase that is part of the adaptive response in E. coli, and has homologs in many species. On basis of in vitro data we created the hypothesis that the N 2 guanine lesions as well as 6-methyladenine would be substrates for the enzyme AlkB in vivo. We found, however, in this case the in vitro results did not predict the biology observed in cells. / by Nidhi Shrivastav. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2012. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 161-171). / Normalizing angiogenesis is a promising strategy for treatments of cancer and several disorders plagued by misregulated blood supplies. To address the daunting complexity of angiogenesis arising from multiple phenotypic behaviors governed by multiple stimuli, computational approaches have been developed to predict sprouting angiogenic outcomes. In recent years, the agent based model, in which individual cells are modeled as autonomous decision making entities, has become an important tool for simulating complex phenomena including angiogenesis. The reliability of these models depends on model validation by quantitative experimental characterization of the cellular (agent) behaviors which so far has been lacking. To this end, I develop an experimental and computational method to semi-automatically estimate parameters describing the single-cell decision in the agent based model based on flow cytometry aggregate headcount data and single cell microscopy which yields full panel single cell trajectories of individual endothelial cells. Applying thees method to the single cell decision data, I propose two conceptual models to account for the different state transition patterns and how they are modulated in the presence of opposing inflammatory cytokines. The observed unique state transition patterns in the angiogenic endothelial cell population are consistent with one of these descriptions, the diverse population model (DPM). The DPM interpretation offers an alternative view from the traditional paradigm of cell population heterogeneity. This understanding is important in designing appropriate therapeutic agents that take effect at the cellular level to meet a tissue level therapeutic goal. / by Tharathorn Rimchala. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2009. / Vita. Cataloged from PDF version of thesis. / Includes bibliographical references. / Major therapeutic efforts have been devoted to targeting the epidermal growth factor receptor (EGFR), which is aberrantly expressed in many cancers and is correlated with tumor progression and invasiveness. In the current tumor progression paradigm, individual invasive carcinomas arise upon epithelial-mesenchymal transition (EMT) and migrate through a complex tumor microenvironment to successfully metastasize. While the activation of EGFR enhances invasiveness in vivo, it is still unclear which downstream molecular changes caused by EMT contribute to the invasive phenotype and subsequently, how the invasive cell integrates downstream biophysical processes to invade through a three-dimensional (3D) extracellular matrix (ECM). This thesis addresses these questions from a quantitative, engineering perspective, that cell migration in the context of the invasion microenvironment is an inherently multivariate biochemical and biophysical problem. As such, we developed various carefully controlled, but biologically relevant, in vitro experimental systems with an emphasis on the extracellular microenvironment. These systems were combined with quantitative data-driven parameterization of signaling components and subsequent modeling of migration phenotypes via various 2D and 3D single cell tracking assays. By measuring 2D cell migration of immortalized human mammary epithelial cells conferring pre- or post-EMT states, we respectively identified physiologically relevant, EMT-dependent collective and individual migration modes. A comprehensive systems modeling approach identified the novel activation of a downstream kinase, which acts in a switch-like manner to differentially regulate epithelial, EGFR-dependent migration versus mesenchymal migration. Next, the subsequent mesenchymal migration in 3D, as modeled by a human glioblastoma cell line, was assessed via a quantitative biophysical analysis. EGF-enhanced 3D migration arose from a balance between a cell-intrinsic regulation of cell speed and a matrix- and proteolysis-dependent, extrinsic regulation of directional persistence. Lastly, we quantified fibroblast migration in a porous scaffold of varying pore sizes and stiffness to model contact-guided quasi-3D migration. We surprisingly found that the micro-architecture of guidance structures alone influenced cell speed. Therefore, the combination of biologically relevant experimental systems and quantitative models provided novel mechanistic insights pertinent to early stages of tumor metastasis. The experimental approaches and biological mechanisms in this thesis hold potential in guiding therapeutic targeting of the biophysical responses prompted by the extracellular microenvironment. / by Hyung-Do Kim. / Ph.D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2018. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 72-76). / In systems biology, ordinary differential equation models are used frequently to model the dynamics of molecular and cellular systems. These models are parameterized with rate constants and other quantities that are often estimated from empirical data. When the data are insufficient to fully determine the model parameters, the parameter values are unidentifiable, and many parameter sets are consistent with the data. To cope, many studies sample a collection of parameters to represent the uncertainty or simplify the model to remove parameters. Studies rarely verify that their sampling is sufficient or test alternative model simplifications. There is a need for better practices for uncertainty quantification. In this work, I present two case studies demonstrating the use of biochemical models with unidentifiable parameters to make useful predictions. The first study investigates a model of the complement system, a system of circulating proteins involved in immune response, to find promising drug targets for treatment of sepsis. I compared a sampling method to a worst-case search method for quantifying the uncertainty in responses to hypothetical inhibitors and found that the choice of method significantly impacts the results. I identified mechanistic explanations for the observed inhibitor responses that demonstrate limitations of intuition and suggest strategies for further studies. The second study uses a kinetic model of the thiolase and reductase enzymes of the 3-hydroxyacid metabolic pathway to interpret available in vitro data to determine the kinetic changes induced by a mutation in the thiolase. Sampling approaches cannot identify all combinations of rate constants that could have changed according to the data, so I perform a selective enumeration strategy that identifies all feasible combinations by testing only a limited number. The simplest feasible combinations identify three classes of rate constant changes induced by the mutation. I also use a global sensitivity analysis approach to predict which reaction steps are most likely to positively affect the product selectivity ratio of the system. Together, these studies demonstrate that unidentifiable models can be useful if the correct methods are chosen to quantify their uncertainty and serve as examples of how to choose or design these methods. / by David C. Flowers. / Ph. D.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Biological Engineering Division, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 58-63). / Metastasis is a multistep process during which tumor cells separate from a primary tumor, penetrate the bloodstream, evade host defenses, and colonize distant organs. This final and fatal step in tumor development is the cause of more than 90% of cancer related deaths. Therapies and diagnostics can be targeted to metastasis at three points in its progression: the primary tumor, the secondary tumor, and circulating tumor cells (CTCs). While much work has focused on primary tumors, less effort has concentrated on targeted isolation, detection and therapy of deeply penetrated metastases and CTCs. Here, I discuss cell and nanomaterial-based approaches for detecting and ablating these malignant populations. The number of CTCs in the blood directly correlates with disease progression; however, the lack of definitive markers has limited their isolation and characterization. I have demonstrated the potential use of platelets as a cell-based marker for isolation and detection of CTCs. Using phage display technology, it was possible to identify candidate peptides specific to mesenchymal-like tumor cells that may mimic the motile and aggressive CTC population. In order to detect and ablate metastases and CTCs, M13 bacteriophage was engineered into a platform for simultaneous tumor targeting, imaging, and therapy. Single-walled carbon nanotubes (SWNTs) and doxorubicin, a chemotherapeutic agent, were loaded on phage for fluorescent near-infrared imaging and cytotoxicity of metastatic lesions, respectively. The near-infrared optical properties of SWNTs in the "second window" make them promising candidates for imaging nascent and deeply seeded tumors. This approach provides an 'all-in-one' platform for targeted fluorescence imaging and efficient drug delivery and may allow for real-time monitoring of tumor response to drug regimens. / by Aditya Kohli. / M.Eng.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, February 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Monoclonal antibodies have emerged as an important class of cancer therapeutics due to their ability to specifically bind tumor-expressed antigens. Unfortunately, attempts to treat solid tumors with these drugs are often limited by an inability of the antibodies to fully penetrate the tumor tissue, leaving large regions of untargeted and viable cells. The goal of this thesis is to understand the transport phenomena that contribute to poor antibody distribution in tumors, and engineer novel antibody variants with improved targeting properties. Previous studies identified a core set of parameters that impact tumor uptake including antibody size, binding affinity, plasma clearance rate, and cellular catabolism. Here we probe each of these parameters and its effect on tumor penetration using a combination of computational modeling and protein engineering. In the first part of this thesis, we characterize the cellular internalization kinetics of a series of anti-carcinoembryonic antigen (CEA) antibodies and antibody fragments. We demonstrate that internalization is independent of antibody affinity, stability, and valency, and that the measured rates can be used to mathematically predict antibody penetration distance in tumor spheroids. Next, we examine the effect of antibody size and affinity by developing a computational model of in vivo tumor targeting that incorporates size-dependent trends for capillary permeability, interstitial diffusion, available volume fraction, and plasma clearance. The model predicts that intermediate size antibody fragments (MW ~30 kDa) have the lowest tumor uptake with greater accumulation of small and large proteins. To probe size effects experimentally, we engineered a novel 79 kDa ds(Fv)-Fc antibody fragment that is approximately half the size of an IgG but retains its binding and Fc salvage activity. In mice, the ds(Fv)-Fc fragments are cleared from the plasma more rapidly than IgGs but have similar tumor uptake levels at 24 hours, likely due to higher capillary permeability. In the last section, we develop a series of matrix metalloproteinase (MMP) activatable antibody fragments that bind their target antigen up to 300 times faster following cleavage by the tumor expressed protease MMP-2. We believe that MMP dependent binding should prevent targeting of antigen depots in healthy tissues and further improve tumor specificity. / by Michael M. Schmidt. / Ph.D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2017. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 173-197). / Chemotherapy is widely used in the treatment of solid tumors, but its effects are often associated with cancer relapse, metastasis, and drug resistance. The biological mechanisms that drive the structural and functional changes in cancer cells associated with these features of disease progression remain poorly understood. Consequently, quantitative characterization of molecular signaling pathways and changes in cancer cell phenotypes induced by chemotherapy through the use of in vitro model systems would expand our understanding of drug mechanisms and provide for putative strategies to counteract drug-induced cancer progression. Toward this end, I develop bioimage informatics tools to characterize changes in signaling, structure, and function of cancer cells from fluorescence microscopy data. I first present a generally-applicable probabilistic time-series modeling framework to classify cell shape dynamics. Times-series models draw quantitative comparisons in cell shape dynamics that are used to distinguish and interpret cellular responses to diverse drug perturbations. Next, I investigate the effects of doxorubicin, a DNA-damaging chemotherapeutic drug, on breast cancer cell signaling and phenotype. Bioinformatics analyses of phosphoproteomics data are first used to infer biological processes downstream of DNA damage response signaling networks altered by doxorubicin treatment. These analyses reveal changes in phosphoproteins associated with the actomyosin cytoskeleton and focal adhesions. Live-cell imaging of cell morphology, motility, and apoptosis dynamics reveals a link between doxorubicin-induced cytoskeletal signaling and morphological elongation, directional migration, and enhanced chemo-tolerance. These findings imply that sub-maximal tumor killing can exacerbate disease progression through adaptive resistance to primary chemotherapy treatment through DNA damage response-regulated cytoskeletal signaling. Finally, I combine the results of the phosphoproteomic analysis with phenotypic profiling to characterize doxorubicin-induced changes in actomyosin signaling that affect cancer cell shape and survival. I additionally describe a generally-applicable multiplexed fluorescence imaging framework that uses diffusible nucleic acid probes to detect nearly a dozen subcellular protein targets within the same biological sample. Taken together, these methodologies reveal previously-unappreciated effects of chemotherapy on breast cancer signaling and phenotype, and demonstrate the value of combining bioinformatics analyses of -omics data with quantitative fluorescence microscopy as a general strategy in biological mechanism discovery. / by Simon Gordonov. / Ph. D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2017. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Fluorescence microscopy has proven to be immensely powerful for the study of biological systems at both the cellular and systems biological levels. The ability to specifically label a single molecular species fluorescently has enabled the study of complex cellular structures through the visualization of their constituent components both individually as well as in context of the overall structure. Since the advent of engineered fluorescent proteins (such as GFP) and other proteins capable of being genetically encoded as fusion constructs, the utility of fluorescence microscopy has increased exponentially in terms of the ability to efficiently, specifically label desired molecules while limiting perturbations to the biology under study. With this enhanced ability of visualization came a hand-in-hand evolution of computational techniques to extract quantitative information from microscopy images. In this thesis, I focus on the application of fluorescence imaging at the biophysical level in living cells: analyzing the motion/dynamics of single molecules and complexes, which are small relative to the structures of the cell, in order to elucidate their molecular function and mechanism. The motion of these "particles" within living cells is necessarily related to their functions as well as their interacting partners, which can vary dynamically during their lifetimes. Observation and analysis of this motion using a combination of fluorescence microscopy and robust quantitative analysis allows one to infer these characteristics. Here, I study three diverse biological systems in the context of live-cell fluorescence microscopy and biophysical analysis: 1) the transport of 0-actin mRNA particles in primary mouse neurons, 2) kinetochore motion during cell division, specifically focusing on anaphase dynamics, and 3) the motion of cell-growth-implicated membrane proteins in Bacillus subtilis. / Funded by the NSF Physics of Living Systems PHY 1305537. / by Zachary Thomas Barry. / Ph. D.

Thesis: S.M., Massachusetts Institute of Technology, Department of Biological Engineering, 2015. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 65-67). / Plasmodium falciparum is the deadliest malaria parasite. There is no approved vaccine to prevent this disease, and resistance to available antimalarial drugs is becoming widespread. Identification of parasite genes essential to survival and virulence could facilitate the development of novel therapeutics and vaccines. However, these efforts have been impeded by difficulties in manipulating the parasite's genome and functionally perturbing gene expression in a controlled way. Our lab has developed inducible systems to control P. falciparum gene expression, and has achieved successful editing of the P. falciparum genome using CRISPR/Cas9 technology. We have integrated these capabilities into a modular and scalable framework that can be used to efficiently edit, regulate and delete any target parasite gene after a single genome editing operation. This approach will accelerate studies of parasite gene function, and help prioritize potential drug and vaccine targets. A key requirement in this framework is the efficient assembly of donor vectors for modifying target loci and installing the necessary regulatory parts. This necessitates cloning several large, [A+T]-rich P.falciparum genomic regions that can be quite tedious and rate limiting. In this document, we show a new cloning strategy using linear vectors that facilitates rapid and accurate assembly of vectors capable of transforming P. falciparum. We present evidence of successful chromosomal modification of several genes via spontaneous single crossover, as well as zinc finger nuclease- and Cas9- mediated genome editing strategies using our assembled donor vectors. We also show that these modifications enable controllable expression of several previously uncharacterized genes to elicit phenotypes that we are investigating in further mechanistic detail. Importantly, these transgenic parasites can now be rapidly generated to allow identification of novel essential parasite genes in as little as a month. / by Armiyaw Sebastian Nasamu. / S.M.