Thesis (Ph. D.)--Massachusetts Institute of Technology, Biological Engineering Division, 2007. / This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Includes bibliographical references (p. 195-208). / Citrobacter rodentium is a naturally occurring murine bacterial pathogen which is used to model human diarrheagenic E. coli (EPEC and EHEC) infections in mice. C. rodentium causes colonic hyperplasia and a variable degree of colitis and mortality in the majority of inbred and outbred lines of mice. Differences in C. rodentium-induced disease are determined by the genetic background of the host. Here, C. rodentium infection in resistant outbred Swiss Webster (SW) mice was compared with infection in the cognate inbred FVB strain for the first time. In contrast to subclinical infection in SW mice, adult FVB mice developed overt disease with significant weight loss, severe colitis, and over 75% mortality. Fluid therapy intervention completely prevented mortality in FVB mice, and expression of pro-inflammatory and immunomodulatory genes in the colon was similar in both lines of mice, suggesting that mortality in C. rodentiuminfected FVB mice is due to hypovolemia resulting from severe dehydration. To identify host factors responsible for the development of mortality, gene expression in the distal colon of FVB and SW mice was investigated using a whole mouse genome Affymetrix array. / (cont.) Transcripts represented by 1,547 probe sets (3.4%) were differentially expressed between FVB and SW mice prior to infection and at 4 and 9 days post-inoculation. Data analysis suggested that intestinal ion disturbances rather than immune-related processes are responsible for susceptibility in C. rodentium-infected FVB mice. Marked impairment in intestinal ion homeostasis predicted by microarray analysis was confirmed by quantitative RT-PCR and serum electrolyte measurements that showed hypochloremia and hyponatremia in susceptible FVB mice. C. rodentium infection was next characterized in additional inbred strains of Swiss origin. SWR and SJL mice developed minimal morbidity and no mortality in response to the pathogen, demonstrating resistance to disease. Furthermore, C3H mice developed severe diarrhea and gene expression changes comparable to those in infected FVB mice, suggesting common pathogenic mechanisms in susceptible strains. In conclusion, C. rodentium infection in FVB mice is a useful model for fatal infectious diarrhea. These studies contribute to our understanding of C. rodentium pathogenesis and identify possible candidates for susceptibility to fatal enteric bacterial infection. / by Diana Borenshtein. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Biological Engineering Division, 2008. / Includes bibliographical references (p. 139-153). / Many cellular processes are governed by large and highly-complex networks of chemical interactions and are therefore difficult to intuit. Computational modeling provides a means of encapsulating information about these interactions and can serve as a platform for gaining understanding of the biology and making predictions about cellular response to perturbation. In particular, there has been considerable interest in ordinary differential equation (ODE) models, which have several attractive features: ODEs can describe molecular interactions with mechanistic detail, it is relatively straightforward to implement perturbations, and, in theory, they can predict the concentration and activity of every species as a function of time. However, both the topology and parameters in such models are subject to considerable uncertainty. We explore the ramifications of these sources of uncertainty for making accurate predictions and develop methods of selecting high confidence predictions from uncertain models. In particular, we promote a shift in emphasis from model selection to prediction selection, and use consensus among model ensembles to identify the predictions most likely to be accurate. By constructing decision trees, this consensus can also be used to partition the space of potential perturbations into regions of high and low confidence. We apply our methods to the Fas signaling pathway in apoptosis to satisfy two goals: first, to design a therapeutic cocktail to reduce cell death in the presence of high levels of stimulus, and second, to design experiments that may lead to a better understanding of the biological network. / by Caitlin Anne Bever. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Biological Engineering Division, 2008. / Includes bibliographical references (p. 309-336). / Molecular signaling systems allow cells to sense and respond to environmental stimuli. Quantitative modeling can be a valuable tool for evaluating and extending our understanding of signaling systems. In particular, studies of the mating pheromone response system in yeast (Saccharomyces cerevisiae) have revealed many protein families and regulatory motifs also found in higher eukaryotes. This thesis develops several computational and experimental approaches that facilitate characterization of cellular signaling systems, and tests these approaches using yeast mating response as a model. Limitations in the current approach to building models of molecular systems were addressed first. For example, published computational models are often difficult to evaluate and extend because researchers rarely make available the information and assumptions generated throughout model building. I developed tools that facilitate model construction, evaluation, and extension. I used these tools to develop the YeastPheromoneModel (YPM) information repository, in which construction of an exhaustive model of the yeast mating system is documented (http://www.YeastPheromoneModel.org). Next, motivated by an ability to rapidly make many derivative models from the YPM repository and by carefully measured abundances of mating system proteins, I analyzed a model of the mating system mitogen activated protein kinase cascade. I found that varying the abundance of the scaffold protein Ste5, but not the abundances of other proteins, is expected to result in a quantitative tradeoff between total system output and dynamic range. Thus, the abundance of scaffold proteins in signaling systems may generally be under selective pressure to support specific quantitative system behavior. / (cont.) Finally, because traditional methods for characterizing signaling systems can be slow and tedious, I postulated that time-dependent stimulation of signaling systems might increase the richness and value of data derived from individual experiments. To do this, I devised a custom microfluidic device to expose yeast cells to pheromone in a time-dependent manner. I also developed computational approaches to investigate the use of time-dependent stimulation to characterize receptor and G protein response dynamics. I found that, at least for the receptor/G protein portion of the mating system, time-dependent stimulation does not appear to offer significant gains for constraining kinetic parameters relative to traditional step-response experiments. / by Ty M. Thomson. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Biological Engineering Division, 2006. / Includes bibliographical references (v. 2, leaves 195-258). / The anticancer drug cisplatin is in widespread use but its mechanism of action is only poorly understood. Moreover, human cancers acquire resistance to the drug, which limits its clinical utility. A paradox in the field is how loss of mismatch DNA repair leads to clinical resistance to this widely used drug. The phenomenon of cisplatin tolerance in mismatch repair deficient cells was initially discovered in E. coli, where methylation deficient dam mutants show high sensitivity to cisplatin and dam mutants with an additional mutation in either of the mismatch repair genes mutS or mutL show near wildtype levels of resistance. A prevalent explanation for this observation is the abortive repair model, which proposes that in dam mutants, where the strand discrimination signal is lost, mismatch repair attempts futile cycles of repair opposite cisplatin-DNA adducts. Previous findings have supported this model to the extent that MutS, the E. coli mismatch recognition protein, specifically recognizes DNA modified with cisplatin. However it has recently been shown that MutS binding to cisplatin adducts may contribute to toxicity by instead preventing the recombinational repair of a cisplatin-modified substrate, and we have previously shown that recombination is an essential mechanism for tolerating cisplatin damage. / (cont.) In the present study, we examined the global transcriptional responses of wildtype, dam, dam mutS, and mutS mutant E. coli after treatment with a toxic dose of cisplatin. We also determined any dose-response at the transcriptional level of several SOS response genes and other genes involved in DNA repair by real time RT-PCR. Furthermore, we performed single-cell electrophoresis in order to determine the effect of mismatch repair on the level of double-strand break formation in cisplatin-treated cells. Our results show that Dam-deficient strains exhibit unique gene regulation that may be due to mismatch-repair induced DNA damage in the absence of adenine methylation. In addition, cisplatin treatment induces double-strand break formation and the SOS response in a dose-dependent manner, and both break formation and the SOS response are greatest in the hypersensitive dam mutant strain. The higher level of cisplatin-induced double-strand breaks in the dam mutant may be dependent on functional mismatch repair. / by Jennifer L. Robbins. / Ph.D.

Thesis (S.M. in Toxicology)--Massachusetts Institute of Technology, Biological Engineering Division, 2004. / Includes bibliographical references (leaves 45-47). / Design and fabrication of a microfluidics system capable of generating reproducible and controlled micro-biochemical environments that can be used as a diagnostic assay and microreactor is important. Here, a simple technique was developed to create a robust microfluidics system capable of generating precise gradients of biochemical properties within its channels. Through this approach, it is possible to create a gradient generator with mammalian cells patterned and seeded under its poly(dimethylsiloxane) (PDMS) channels. Cells that were seeded and patterned under the PDMS channels remained viable and capable of performing intracellular reactions. Using the gradient generator within the PDMS microfluidic device, a gradient of specific and controlled biochemicals can be flowed on seeded cells allowing for high-throughput molecular interaction analysis. The microfluidics system provides a way to study and analyze cell response in the presence of a combination of biochemical signals. / by Guan-Jong Chen. / S.M.in Toxicology

Thesis (Ph. D.)--Massachusetts Institute of Technology, Biological Engineering Division, 2006. / Includes bibliographical references (p. 139-145). / Developments in light microscopy over the past three centuries have opened new windows into cell structure and function, yet many questions remain unanswered by current imaging approaches. Deep ultraviolet microscopy received attention in the 1950s as a way to generate image contrast from the strong absorbance of proteins and nucleic acids at wavelengths shorter than 300 nm. However, the lethal effects of these wavelengths limited their usefulness in studies of cell function, separating the contributions of protein and nucleic acid proved difficult, and scattering artifacts were a significant concern. We have used short exposures of deep-ultraviolet light synchronized with an ultraviolet-sensitive camera to observe mitosis and motility in living cells without causing necrosis, and quantified absorbance at 280 nm and 260 nm together with tryptophan native fluorescence in order to calculate maps of nucleic acid mass, protein mass, and quantum yield in unlabeled cells. We have also developed a method using images acquired at 320nm and 340nm, and an equation for Mie scattering, to determine a scattering correction factor for each pixel at 260nm and 280nm. These developments overcome the three main obstacles to previous deep UV microscopy efforts, creating a new approach to imaging unlabeled living cells that acquires quantitative information about protein and nucleic acid as a function of position and time. / by Benjamin J. Zeskind. / Ph.D.

Thesis (S.M.)--Massachusetts Institute of Technology, Biological Engineering Division, 2005. / Includes bibliographical references (leaves 38-44). / Nitric oxide is postulated to protect cancer cells from the death-inducing effects of tumour necrosis factor alpha by S-nitrosating the active site cysteines, inhibiting cleavage of caspase-9. We aimed to test this hypothesis and to determine its validity across cancer cell types. In addition, we hoped to explain the involvement of certain kinases in nitric oxide-induced apoptosis. The experimental setup involved stimulating human colorectal cancer cells, HT-29 and HCT- 116, and human prostate cancer cells, LNCaP, with cytokines in order to induce cell death. Then, we observed the effects of NO inhibitors, kinase inhibitors, and activation of Akt, a kinase up-stream of the caspase cascade, following transfection of a DNA sequence that was proven to protect cells against apoptosis induction. In our series of experiments, inhibition of the nitric oxide synthases removes nitric oxide protection from apoptosis, but inhibition of only the inducible synthase has opposite effects with prostate and colon cancer cells that are considered insignificant, and its effects on the two types of colon cancer cells are in discord. Transformation and transfection of ARK5 into the colorectal cancer cell line, HT-29 did not prove beneficial. Similarly, glucosamine showed no clear pattern of reducing apoptosis in the cells. Therefore, we propose further exploration of the inhibition of constitutive nitric oxide synthases as a potential therapy. / by S. Julie-Ann Lloyd. / S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Biological Engineering Division, 2008. / Vita. / Includes bibliographical references (p. 82-89). / TLR-4 is essential in host defense against bacterial infection. By recognition of specific pathogen-associated molecular patterns such as lipopolysaccharide (LPS), TLR-4 can in tandem initiate a pair of downstream signaling pathways to regulate cytokine/chemokine release, endotoxin tolerance and apoptosis, which have been suggested to directly or indirectly participate in the regulation of innate and adaptive immune responses. However, little is known about their detailed signal-response relationships. In this thesis, we sought to identify these potential signal-response relationships in RAW264.7 cells through systematic analysis. We first measured LPS stimulated dynamic signaling profiles over a range of an inhibitor of p38 MAPK, SB202190 concentrations for a distribution of kinases centrally involved in TLR-4 signaling network. We then applied quantitative analytical approaches to determine the most important signals or signal combinations contributing to induction of either IL-6 and TNF-ao secretion or apoptosis and construct their corresponding predictive mathematical models. Particularly, we found that the partial least squares regression (PLSR) models built using the ratio of phosphorylated Jun N-terminal kinase (JNK) and extracellular signal regulated kinase (ERK) predicted LPS plus SB202190-induced apoptosis accurately even following perturbation with pharmacological inhibitors of JNK and ERK. Thus, by combining experimental and computational approaches, this thesis has proposed two new potential targets, JNK and ERK, for development of drug therapies against bacterial infection. / by Rongcong Wu. / S.M.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Biological Engineering Division, 2007. / MIT Institute Archives copy: DVD inserted in pocket on p. [3] of cover on v. 1. / "c2007"--p. ii. / Includes bibliographical references (v. 2, leaves 292-302). / A grand challenge of systems biology is to model the cell. The cell is an integrated network of cellular functions. Each cellular function, such as immune response, cell division, metabolism or apoptosis, is defined by an interconnected ensemble of biological pathways. Modeling the cell or even one cellular function requires a computational architecture that integrates multiple biological pathway models in a scalable manner while ensuring minimal effort to maintain the resulting integrated model. Scalable is defined as the ease in which more and more biological pathway models can be integrated. Current architectures for integrating biological pathway models are primarily monolithic and involve combining each biological pathway model's software source code to build one large monolithic model that executes on a single computer. Such architectures are not scalable for modeling complex cellular functions or the whole cell. We present Cytosolve, a new computational architecture that integrates a distributed ensemble of biological pathway models and computes solutions in a parallel manner while offering ease of maintenance of the integrated model. The individual biological pathway models can be represented in SBML, CellML or in any number of formats. The EGFR model of Kholodenko with known solutions is used to compare the Cytosolve solution and computational times with a known monolithic approach. A new integrative model of the interferon (IFN) response to virus infection is developed using Cytosolve. Each model within the integrated model, spans different time scales, is created by different authors from four countries and three continents across different disciplines, is written in different software codes, and is built on different hardware platforms. / (cont.) A new quantitative methodology and formalism is then derived for evaluating different types of monolithic and distributed architectures for integrating biological pathway models. As more biological pathway models develop in a disparate and decentralized manner, the Cytosolve architecture offers a unique platform to build and test complex models of cellular function, and eventually the whole cell. / by V.A. Shiva Ayyadurai. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Biological Engineering Division, 2005. / "June 2005." / Includes bibliographical references. / Micro- and nanofabrication approaches have dramatically changed our society through their use in microelectronics and telecommunication industries. These engineering tools are also useful for many biological applications ranging from drug delivery to DNA sequencing, since they can be used to fabricate small features at a low cost and in a reproducible manner. The goal of this thesis was to develop techniques based on the merger of novel materials and nano and microfabrication approaches to manipulate cell microenvironment in culture. To control cell migration and to restrict cell or colony size, cells and proteins were patterned by using molding or printing methods. Poly(ethylene glycol)-based molecules and polysaccharides were used to control cell-substrate interactions and to prevent cell adhesion on specific regions of a substrate. To control cell-cell contact, layer-by-layer deposition of ionic biopolymers (i.e. negatively charged hyaluronic acid and positively charged poly-L-lysine) was used to generate patterned co-cultures. In addition, to control cell-soluble factor interactions, microfluidic-based approaches were developed. To pattern cells and proteins within microchannels, a soft lithographic method was developed to pattern microchannel substrates using printing and molding approaches. / (cont.) To easily immobilize cells within channels, poly(ethylene glycol) microstructures were used to capture cells within low shear stress regions. These techniques also allowed for the fabrication of multiphenotype cell arrays. In addition, techniques were developed to control the interaction of cells within hydrogels by controlling the spatial properties of hydrogels. / by Ali Khademhosseini. / Ph.D.