Thesis: Ph. D. in Medical Engineering and Medical Physics, Harvard-MIT Program in Health Sciences and Technology, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages [165]-191). / Optical microscopy techniques are widely used to study cellular physiology in their native tissue environments. In particular, the use of fluorescent probes to tag different cell populations, subcellular compartments, specific proteins and nucleotide sequences has enabled examination of cellular phenotypes with increasingly sophisticated detail. Recent efforts to combine physiological imaging and single-cell molecular analysis seek complete understanding of cellular identity and function within complex tissues. Specific cells of interest can be selected and isolated from tissues for downstream molecular analyses using techniques such as laser capture micro-dissection, or cell tagging with photo-conversion. However, high-throughput, unbiased molecular profiling of every cell imaged within a tissue remains an elusive challenge. A fundamental obstacle in previous approaches is spectral overlap due to the relatively broad emission of typical fluorescent probes, which limits their capabilities for multiplexed tagging. The first part of this thesis describes methods for studying cellular physiology in mice at single-cell resolution using two-photon fluorescence microscopy. The second part of this thesis describes the development of a novel class of imaging probes, called laser particles, which rely on narrowband laser emission for massively multiplexed cell tagging. This work establishes laser particles as promising tools for comprehensive single-cell analyses. / by Sheldon J.J. Kwok. / Ph. D. in Medical Engineering and Medical Physics / Ph.D.inMedicalEngineeringandMedicalPhysics Harvard-MIT Program in Health Sciences and Technology

Thesis: Ph. D. in Medical Engineering and Medical Physics, Harvard-MIT Program in Health Sciences and Technology, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references. / Departing from the traditional role as a carrier of genetic information, DNA has emerged as an engineering material for construction of nano-devices. The advances in the field of DNA nanotechnology have enabled design and synthesis of DNA nanostructures of arbitrary shapes and manipulation of the nanostructures' conformations in a programmable way. DNA-based systems offer potential applications in medicine by manipulating the biological components and processes that occur at the nanometer scale. To accelerate the translation of DNA-based systems for medical applications, we identified some of the challenges that are hindering our ability to construct biomedical nano-devices and addressed these challenges through advances in both structural and dynamic DNA nanotechnology. First, we tested the stability of DNA nanostructures in biological environments to highlight the necessity of and path towards protection strategies for prolonged integrity of biomedical nano-devices. Then, we constructed a platform for robust 3D molecular integration using DNA origami technique and implemented the platform for a nanofactory capable of production of therapeutic RNA to overcome the challenges in RNA delivery. Moreover, we established a mechanism to drive DNA devices by changing temperature with prolonged dynamic behavior that was previously challenging to accomplish without special modification of DNA and/or equipment not readily available in a typical lab setting. Together, the progress made in this thesis bring us another step closer to realization of medical applications of DNA nanotechnology by focusing on the challenges in both structural and dynamic aspects of the technology. / by Jaeseung Hahn. / Ph. D. in Medical Engineering and Medical Physics / Ph.D.inMedicalEngineeringandMedicalPhysics Harvard-MIT Program in Health Sciences and Technology

Thesis: Ph. D. in Medical Engineering and Medical Physics, Harvard-MIT Program in Health Sciences and Technology, September, September, 2020 / Cataloged from student-submitted PDF version of thesis. / Includes bibliographical references. / Organs for transplantation has continuously been in short supply and, given COVID-19's propensity to adversely impact solid organs, the shortage will likely become exacerbated. For decades, the field of tissue engineering has developed innovative methods to generate model tissues de novo. Top-down approaches, such as microfluidics and 3D bioprinting, provide spatial control by patterning cell types with high resolution, but face challenges in reproducing physiologically accurate cell types and interactions. Bottom-up methods, such as organoids, induce pluripotent cells to differentiate into aggregates that resemble their in vivo counterparts, yet the size and complexity of these structures are limited by nutrient diffusion and the morphology cannot be controlled. An ideal system would allow for high spatial control while retaining native cell-cell interactions formed through developmental progression. / To approach this capability, we aimed to create a sequential gene expression system that programmatically aggregate and differentiate cells, merging both top-down and bottom-up characteristics. First, we curated and characterized 28 recombinases to determine efficiency and pairwise compatibility for use in mammalian recombinase genetic circuits (RGC). From this set, we designed an RGC capable of expressing 12 genes in sequence, providing a framework for simulating the gene expression cascades of development. To elucidate the temporal dynamics of recombinase action in mammalian cells, we formulated a mathematical model for recombinase expression and catalysis and validated it with experimental data. We found that recombinases have variable expression levels, catalytic rates, and binding affinities, which should be accounted for when designing RGCs. / Separately, we designed a platform for engineering novel membrane proteins for inducing specific cell-cell interactions using coiled-coils, called helixCAM. We demonstrated that helixCAMs are capable of inducing patterned cell binding in E. coli, yeast, and human cells, and further utilized a library-on-library approach to engineer new helixCAM-optimized coiled-coils. Taken together, the genetic tools described in this thesis establish groundwork towards hybrid tissue engineering strategies capable of high-resolution patterning while enabling endogenous cell differentiation and cell-cell interactions to form, ultimately serving as a template for engineering large-scale tissue and organs de novo. / by Chung-Yun (George) Chao. / Ph. D. in Medical Engineering and Medical Physics / Ph.D.inMedicalEngineeringandMedicalPhysics Harvard-MIT Program in Health Sciences and Technology

Thesis: Ph. D. in Medical Engineering and Medical Physics, Harvard-MIT Program in Health Sciences and Technology, May, 2020 / Cataloged from student-submitted PDF version of thesis. / Includes bibliographical references (pages 128-139). / Effective disease management requires high quality and accurate information about disease state. As science and technology have evolved, the history and physical exam, once the foundations of the diagnostic workflow, have been supplemented with modalities that allow physicians to peer inside the body and acquire otherwise inaccessible information. To gain maximal information about a disease, a promising approach would be to administer a probe that can detect disease activity inside the body and emit a signal to the outside world. To this end, our group has developed "activity-based nanosensors", which detect dysregulated protease activity at the site of disease and release a reporter that can be measured in the urine. Because proteases are implicated in multiple diseases, including cancer, activity-based nanosensors have the potential to enable quantitative, noninvasive, and real-time monitoring of disease activity. / Respiratory diseases are leading causes of death and disability, owing in large part to the constant exposure of the lungs to the external environment. Though this accessibility makes the lungs vulnerable to carcinogens and pathogens, it also provides a unique diagnostic opportunity. In this thesis, we aimed to optimize activity-based nanosensors for lung disease sensing in two settings: early detection and treatment response monitoring. Finally, we sought to establish a generalizable pipeline to rationally design such tools for human disease. We first delivered a multiplexed panel of sensors via intrapulmonary administration in two genetically engineered mouse models of lung adenocarcinoma. We found that our sensor panel diagnosed lung cancer in both models, detecting tumors as small as 2.8 mm³ without false positives from benign lung inflammation. We then evaluated this approach in monitoring treatment response in mouse models of malignant and benign pulmonary disease. / We observed dramatic treatment-induced shifts in pulmonary protease activity in both models, enabling rapid, noninvasive, and quantitative evaluation of drug response. Finally, we established a suite of ex vivo assays that enabled the bottom-up design of a protease-activated diagnostic probe, opening the door for translation to human disease. Collectively, this thesis provides a framework for the clinical development of activity-based nanosensors for pulmonary disease diagnosis and monitoring. / by Jesse D. Kirkpatrick. / Ph. D. in Medical Engineering and Medical Physics / Ph.D.inMedicalEngineeringandMedicalPhysics Harvard-MIT Program in Health Sciences and Technology

Thesis: Ph. D., Harvard-MIT Program in Health Sciences and Technology, September, February, 2021 / Cataloged from the official PDF version of thesis. / Includes bibliographical references (pages 145-159). / X-ray computed tomography (CT) and planar x-ray imaging are mainstays of modern clinical care. The electron generation mechanism in standard x-ray tubes - specifically, a thermionic cathode - is reliable and capable of high current. However, thermionic cathodes are bulky, and cannot be pulsed quickly. Non-thermionic ('cold-cathode') electron generation can be exploited to make a smaller and rapidly pulsable x-ray source. Such an x-ray source could improve not just the portability of x-ray devices, but would allow for a CT system to operate by pulsing a distributed ring of x-ray sources instead of rotating a single large x-ray source. Furthermore, cold-cathode x-ray sources could allow for new signal acquisition and processing paradigms in the x-ray domain. This includes time-based image acquisition techniques, such as elastography and photon-counting measurements. In this dissertation, I discuss (1) the development of two novel types of cold-cathode x-ray sources: an ultraviolet photocathode-based source, and a silicon field emission chip; (2) novel methods for planar x-ray image acquisition, including a demonstration of dynamic x-ray elastography using a pulsed photocathode x-ray source; and (3) applications of modern signal processing techniques to the tomographic image reconstruction problem. In an epilogue, I discuss our research on N95 respirator sterilization and re-use for crisis situations. / by Avilash Cramer. / Ph. D. / Ph.D. Harvard-MIT Program in Health Sciences and Technology

Thesis: Ph. D., Harvard-MIT Program in Health Sciences and Technology, February, 2021 / Cataloged from the official PDF version of thesis. / Includes bibliographical references (pages 133-148). / Interest in the use of exoskeletons (wearable robotic devices tightly-coupled to the user's body) for human gait augmentation has soared recently, with research flourishing in system design, control, and use efficacy. Use cases span many fields, from military (e.g. load carriage assistance) to medicine (e.g. gait rehabilitation or restoration) and industry (e.g. injury prevention). Evaluating the human factors of human-exoskeleton interaction is an essential step towards operationalization. Unexplained variation in gait strategy and adaptation observed across individual operators must be better understood to enable safe and effective exoskeleton use in real-life environments. Cognitive fit is an individuals' understanding and ability to operate a system. Exoskeletons and similar tightly-coupled lower-extremity (TCLE) systems entail new interaction modalities that may affect cognitive fit. / This thesis explores how cognitive factors and alternative interaction modalities impact individuals' gait and task performance. Two studies were conducted, one evaluating inhibitory control as measured by a modified Simon task using interaction modalities relevant to TCLE system use, i.e. tactile cues and lower-extremity responses. Second, the Human-Exoskeleton Strategy & Adaptation (HESA) study was implemented, in which individuals completed tasks assessing cognitive factors, i.e. inhibitory control and attention, then walked with an ankle exoskeleton. Evaluation of inhibitory control with tactile cues and lower-extremity responses resulted in slower response times and decreased response accuracy. A probe of attention in the HESA study, i.e. completion of a walking task on a self-paced treadmill, showed modified gait characteristics under increased attentional loads, particularly at slower walking speeds and with the addition of a secondary task. / Individualized variation in exoskeleton gait, quantified by spatiotemporal gait characteristics, was explicitly presented for the first time, showing that distinct individuals initially prioritize goals like stability and coordination with an ankle exoskeleton differently. Finally, select measures of cognitive function were found to be correlated to individuals' exoskeleton gait strategy. Individual differences in baseline factors like inhibitory control and ability to perform tasks under divided attention impact individuals' cognitive fit with exoskeleton systems. This individualized variation, as well as broader population patterns, should inform exoskeleton design and training by encouraging gait strategies that support desired exoskeleton use goals. For example, stroke patients using an exoskeleton to restore their gait and mitigate fall risk should prioritize stability during system use, while factory workers should prioritize system coordination to minimize injury risk. / This thesis provides foundational insights into human-exoskeleton interaction and gait strategy from a human factors perspective. / by Aditi Gupta. / Ph. D. / Ph.D. Harvard-MIT Program in Health Sciences and Technology

This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Thesis: Ph. D., Harvard-MIT Program in Health Sciences and Technology, 2020 / Cataloged from student-submitted PDF version of thesis. / Includes bibliographical references (pages 171-185). / Magnetic Particle Imaging (MPI) is an emerging tracer-based imaging modality that uniquely images the nonlinear magnetization of superparamagnetic iron oxide nanoparticles (SPIOs). MPI boasts high sensitivity, zero background signal, positive contrast, fast temporal resolution, and quantitative detection. The field of MPI is currently preclinical, and this work aims to scale MPI to human sizes by developing and validating it for two clinical applications: tumor detection and imaging for intraoperative margin assessment during breast-conserving surgery (BCS), and functional neuroimaging. For margin assessment in BCS, a hand-held Magnetic Particle detector and a small-bore MPI imager are assessed for intraoperative use along with an injected SPIO agent. The goal is to detect positive margins during surgery and thus reduce the need for future reexcision. Both hardware systems are validated using clinically relevant phantoms. For functional Magnetic Particle Imaging (fMPI) of the brain, a continuous time-series MPI imager is developed and validated for imaging of cerebral blood volume (CBV) changes during functional activation. The goal is improved sensitivity beyond the capabilities of current functional imaging modalities. We present initial results of in vivo rodent fMPI in a small-bore imager, and the design of a human head-sized system, with implementation underway. Through the collective development of these MPI hardware systems and validation of their potential for these two clinical applications, this work aims to catalyze the expansion of MPI into the clinical setting. / by Erica Ellis Mason. / Ph. D. / Ph.D. Harvard-MIT Program in Health Sciences and Technology

Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2004. / Includes bibliographical references (p. 37-39). / This paper demonstrates that one can infer with respectable accuracy a physician's view of the therapeutic relationship that he or she has with a given patient, using data available in the patient's electronic medical record. In this study, we differentiate between the active primary relationship, the inactive primary or non-attending relationship, and the coverage relationship. We demonstrate that a single model built using the Averaged One-Dependence Estimator (AODE) classifier and learned with six attributes taken from patient visit history and physician practice characteristics can, for most of the 18 physicians tested, differentiate patients with a coverage relationship to a given physician from those with a primary relationship, achieving accuracies of 0.90 or greater as determined by the area under the receiver operating characteristic curve. Three of the 18 datasets had too few coverage patients to adequately characterize. We also demonstrate that, surprisingly, physicians are generally of like mind when assessing the therapeutic relationship that they have with a given patient. We find that for all physicians in our sample, a model built individually with any one physician's assessments performs statistically identically to the model built from the assessments of all other physicians combined. As a sub-goal of this research, we test the performance of different attribute selection methods on our dataset, comparing greedy vs. randomized search and wrapper vs. filter evaluators and finding no practical difference between them for our data. We also test the performance of several different classifiers, with AODE emerging as the best choice for this dataset. Lastly, we test the performance of linear vs. non-linear meta-learners for Stacked / (cont.) Generalization on our dataset, and find no increase in accuracy for the more complex meta-learners. / by Thomas A. Lasko. / S.M.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, June 2011. / "May 2011." Cataloged from PDF version of thesis. / Includes bibliographical references (p. 109-121). / This thesis describes the biochemical regulatory impact of endothelial cells, the cells that line all blood vessels, in cancer. Our work draws from concepts in vascular repair and tissue engineering and extends the view of tumor vessels from perfusing tubes to delivery platforms lined with potent paracrine regulatory cells. We focus on how the endothelial cells themselves regulate tumor biology in a state-dependent fashion. We found that healthy endothelial cells inhibit cancer cell proliferation, invasiveness, and inflammatory signaling and that a defined perturbation of the healthy endothelial cell state - silencing of the gene encoding perlecan - causes loss of the invasion-inhibitory capabilities of endothelial cells by transcriptional upregulation of IL-6. The use of matrixembedded endothelial implants enabled the effects in cell culture to be expanded and validated in animal models. Moreover, endothelial cells exposed to a pathologically activating and inflammatory culture environment, similar to endothelial cells exposed to the atherosclerotic milieu, were leaky and inflamed, with dysregulated proliferative and leukocyte binding properties. Unlike healthy endothelial cells, which suppress cancer cell proliferation and metastasis, these dysfunctional endothelial cells instead aggressively stimulated cancer cell inflammatory signaling and invasiveness, which correlated with stimulation of spontaneous metastasis when implanted as matrixembedded cell implants adjacent to tumors. Fascinatingly we were able to identify markers of endothelial dysfunction, including reduction of endothelial perlecan expression, in human non-small cell lung carcinoma specimens. The state-dependent impact of endothelial cells on cancer biology adds another element to stromal regulation of cancer and brings together a range of disciplines and disparate findings regarding vascular control of tumors. That healthy endothelial cells suppress and dysfunctional cells promote tumor aggression may help to explain undesired effects of therapies that target tumor blood vessels. The harnessing of tissue engineering to regulate vascular and cancer biology may motivate the development of innovative pharmacologic and cell-based therapies for cancer. / by Joseph W. Franses. / Ph.D.

Thesis (Ph. D. in Chemical and Biomedical Engineering)--Harvard-MIT Program in Health Sciences and Technology, 2013. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Bacterial infections continue to be a significant concern particularly in healthcare settings and in the developing world. Current challenges include the increasing spread of drug resistant (DR) organisms, the side effects of antibiotic therapy, the negative consequences of clearing the commensal bacterial flora, and difficulties in developing prophylactic vaccines. This thesis was an investigation of the potential of a class of polymeric nanoparticles (NP) to contribute to the management of bacterial infections. More specifically, steps were taken towards using these NPs (1) to achieve greater spatiotemporal control over drug therapy by more targeted antibiotic delivery to bacteria, and (2) to develop a prophylactic vaccine formulation against the common bacterial sexually transmitted disease (STD) caused by Chlamydia trachomatis. In the first part, we synthesized polymeric NPs containing poly(lactic-co-glycolic acid)- block-poly(L-histidine)-block-poly(ethylene glycol) (PLGA-PLH-PEG). We show that these NPs are able to bind to bacteria under model acidic infection conditions and are able to encapsulate and deliver vancomycin to inhibit the growth of Staphylococcus aureus bacteria in vitro. Further work showed that the PLGA-PLH-PEG-based NPs demonstrated the potential for competition for binding bacteria at a site of infection from soluble protein and model phagocytic and tissue-resident cells in a NP composition dependent manner. The NPs demonstrated low toxicity in vitro, were well tolerated by mice in vivo, and circulated in the blood on timescales comparable to control PLGA-PEG NPs. In the second part, we used PLGA-PLH-PEG-based NPs to design a prophylactic vaccine against the obligate intracellular bacterium Chlamydia trachomatis, the most common cause of bacterial STD in the world. Currently, no vaccines against this pathogen are approved for use in humans. We first formulated NPs encapsulating the TLR7 agonist R848 conjugated to poly(lactic acid) (R848-PLA) in PLGA-PLH-PEG-based NPs, then incubated these R848-NPs with UV-inactivated C. trachomatis bacteria in acidity, forming a construct. Mice immunized with this vaccine via genital or intranasal routes demonstrated protection from genital infection post immunization in a primarily CD4⁺ T cell-dependent manner. These results may suggest avenues for future work in designing and developing more targeted drug therapies or vaccine formulations for managing bacterial infections using polymeric nanoparticles. / by Aleksandar Filip Radovic-Moreno. / Ph.D.in Chemical and Biomedical Engineering