From 2000 to 2015, 97% of clinical trials in oncology failed, in part, due to an incomplete understanding of the dynamic tumor microenvironment (TME). The interplay between the developing neoplasm, resident immune and stromal cells, the extracellular matrix, and signaling molecules results in a complex balance that can promote or prevent tumor progression. An improved understanding of the relationship between these features will facilitate the development of novel cancer therapeutics.The primary goal of this thesis is to characterize how cancer cell-secreted small extracellular vesicles (sEV) are transported and distributed within the interstitial space of the TME. sEVs influence cancer progression through interactions with a range of cell populations in the TME. However, how sEVs are physically distributed within the interstitial matrix, and how this distribution is altered over the course of malignant cancer progression is poorly defined. To assess sEV interstitial transport, sEVs were isolated from the MCF10 series—a model human cell line of breast cancer progression. sEV characterization demonstrated increasing presence of laminin-binding integrins α3β1 and α6β1 on sEVs as the malignant potential of the MCF10 cells increased. Diffusion experiments using fluorescence recovery after photobleaching (FRAP) provided quantitative characterization of diffusion and kinetic binding parameters between bulk sEVs and a laminin-rich ECM, and demonstrated increased accumulation of bound sEVs in the matrix as the malignancy of the parent cell increased. In silico finite element models illustrated sEV accumulation in the matrix resulting in higher bound interstitial sEV concentrations as well as the formation of a transient spatial gradient. Subsequent in vitro microfluidic device convective flow experiments confirmed enhanced concentration of sEVs in the matrix and the formation of interstitial concentration gradients mediated by integrin interactions with laminin-rich ECM. Taken together, these studies demonstrate that sEV interstitial transport, concentration, and spatial distribution are partially dependent on integrin binding to laminin, and evolves with cancer cell malignancy.

Extracellular matrix (ECM) derived from whole organ decellularization offers a promising biological scaffold for tissue engineering applications. The native 3D structure and biochemical composition of these matrices can potentially support tissue-specific recellularization strategies. However, decellularization protocols use reagents that can disrupt the ECM resulting in a range of mechanical properties and protein composition. By identifying structural and biochemical features of the ECM that impact cell behavior, we can tailor decellularization protocols to retain those features. Recellularization of decellularized matrices is an intriguing strategy to engineer lung and cardiac tissue that will likely include the fibroblast. However, excessive collagen deposition by fibroblasts could interfere with normal structure and function of the surrounding tissue. Furthermore, integrin expression can influence the expression of intracellular structural proteins such as alpha smooth muscle actin (α-SMA), and extracellular structural proteins such as collagen. However, previous work has not determined the effect of decellularized ECM on fibroblast function and integrin signaling.

In this work, we used multiphoton microscopy (MPM), combined with image correlation spectroscopy (ICS), to characterize structural and mechanical properties of the decellularized cardiac matrix in a non-invasive and non-destructive fashion. ICS amplitude of second harmonic generation (SHG) collagen images (collagen content) and ICS ratio of two-photon fluorescence (TPF) elastin images (elastin alignment) strongly correlated with compressive modulus. We then seeded cardiac and lung fibroblasts on cardiac ECM, lung ECM and their components to determine the effect of substrate composition, tissue specificity and integrin expression on fibroblast phenotype. α-SMA expression increased for stiffer substrates, and lung fibroblasts expressed significantly higher levels of α-SMA than cardiac fibroblasts. Higher expression of β3 integrins in cardiac fibroblasts, combined with increased α-SMA expression resulting from functional blocking of β3 integrins, demonstrates that β3 plays an important role in regulating cardiac fibroblast phenotype. Our findings indicate that ECM stiffness strongly correlates with collagen and elastin alignment in the ECM following decellularization, which can potentially impact fibroblast collagen and α-SMA expression during recellularization. Furthermore, differential expression of β3 integrins in organ-specific fibroblasts impacts α-SMA expression suggesting that both stromal cell source and ECM structure can impact the remodeling response during recellularization.

Engineered thick tissues require rapid blood perfusion upon implantation for survival. We have previously described a method to prevascularize (in vitro development of a vascular network) engineered tissues with endothelial cells and mural cells prior to implantation. That strategy has the potential to overcome the limitation of oxygen delivery by diffusion, and thus increase post-implantation survival rate. Pericytes are recruited to facilitate vessel maturation and stabilization during the formation of blood vessels. We hypothesized that the introduction of pericytes into the prevascularized tissue would enhance vessel formation and stimulate host anastomosis. Fibrin tissues were prevascularized by co-culturing human placental pericytes with endothelial colony forming cell-derived endothelial cells (ECFC-EC) from cord blood and normal human lung fibroblasts (NHLF) for 7 or 14 days in vitro. Tissues were then subcutaneously implanted to the dorsal side of SCID mice and retrieved 7 days later. Tissues with a low pericyte-fibroblast ratio developed a vessel network that was well-perfused with host circulation after 7 days in vivo culture. In contrast, pericytes alone or with a high pericyte-fibroblast ratio failed to develop significant in vitro vessel networks, and did not anastomose with the host circulation. Our results suggest that a low pericyte-fibroblast ratio can enhance the in vivo perfusion of engineered tissues.

Additionally, we designed and constructed a model to control oxygen diffusion during both in vitro and in vivo culture using biocompatible Poly (methyl methacrylate) (PMMA) and low-density polyethylene (LDPE). Tissues were prevascularized in the devices without limiting oxygen diffusion in vitro for 7 days. Once tissues were implanted into the host, tissue access to oxygen was limited to a small opening facing the host skin side. Oxygen could diffuse through this opening, or be transported by convection following anastomosis with host vasculature. After 7 days, tissues were explanted. Blood perfusion is observed around the access point, but not the entire tissues, which is consistent with blood clotting following anastomosis.

Current preclinical methods to evaluate drug safety fail to accurately predict cardiotoxicity, the leading cause of drug withdrawal from the market. Human stem cell-derived cardiomyocytes represent an intriguing new source of cells for the development of in vitro drug screening platforms. However, questions of phenotypic immaturity, lack of analytical tools to monitor critical cardiomyocyte functions and simplicity of current cardiac tissue models, have delayed the acceptance of these new stem cell-based testing platforms. In this work, we surveyed the many aspects of the stem cell-derived cardiomyocyte phenotype and contrasted them to adult cardiomyocytes. Phasor fluorescent lifetime imaging microscopy (FLIM) analysis monitors metabolism of cells in a nondestructive and noninvasive manner. Phasor FLIM analysis was used to assess the transient metabolic signature of cardiac spheroids and to characterize the acute effect of cyanide poisoning on cardiomyocyte metabolism. Future cardiac drug testing platforms can be used in conjunction with phasor FLIM analysis to elucidate the metabolic effect of drugs. Finally, the effect of interstitial flow on a model of vascularized cardiac tissue was examined. Increased interstitial flow rates enhanced vascular network formation and significantly increased cardiomyocyte growth in cardiac tissues. Methods to increase the complexity and maturity of cardiac tissue can potentially improve the predictive capability of stem cell-based drug testing platforms, and ultimately prevent unnecessary mortality by cardiac drug side effects.

Assessing B cell affinity to pathogen-specific antigens prior to or following exposure could facilitate the assessment of immune status. Current standard tools to assess antigen-specific B cell responses focus on equilibrium binding of the secreted antibody in serum. These methods are costly, time-consuming, and assess antibody affinity under zero-force. Recent findings indicate that force may influence BCR-antigen binding interactions, cell response, and thus immune status. Here, we designed a simple laminar flow microfluidic chamber in which the antigen (hemagglutinin of influenza A or hen egg lysozyme) is bound to the chamber surface to assess antigen-specific BCR binding affinity of five hemagglutinin-specific hybridomas under 65- to 650-pN force range. Our results demonstrate that both increasing shear force and bound lifetime can be used to enrich antigen-specific high-affinity B cells. The affinity constant (KA) of the membrane-bound BCR in the flow chamber correlates well with the affinity of the matched antibodies measured in solution. These findings demonstrate that a microfluidic strategy can rapidly assess intrinsic BCR-antigen binding properties and identify antigen-specific high affinity B cells. This strategy has the potential to both assess functional immune status of a heterogenous population of B cells and be a cost-effective way of identifying individual B cells as antibody sources for a range of clinical applications.

Treatments for cancer remain elusive due, in large part, to the dynamic and unstable genome of most cancer cells. More recently it has become evident that tumor growth and progression to metastasis depends on the ability to recruit normal cells, such as endothelial cells, fibroblasts, as key accomplices. These observations suggest that selective targeting of normal cells, which have a stable genome, could be an effective alternative or complimentary approach in the overall management of the disease. Understanding of such relationship is key for the design of anti-metastatic therapeutics. However, much of the data reported in this field has been performed in xenograft models and/or 2D cultures; which are limited by the number of controllable variables, extrapolation to human tumor physiology, and not amenable for a high-throughput design. This work aims to address the role of the tumor microenvironment using a novel in vitro platform that combines microfluidic and tissue engineering technology to create a 3D tumor microarray in which the tumors receive their nutrients through perfused human microcirculation. This model is capable of replicating the physiology of the in vivo tumor microenvironment; thus providing relevant physiological results. Most importantly, the impact of creating an in vitro 3D metastasis model with perfused human capillary bed could significantly enhance high-throughput anti-metastatic drug screening.

Cancer drug development remains a costly and inefficient endeavor that often translates to limited clinical success. While most therapies focus on stalling the growth of or eradicating tumor cells directly, the microenvironment in which these cells inhabit plays a hugely influential role in defining drug efficacy and disease progression. The supporting vasculature system as well as the interstitial extracellular matrix are particularly consequential to the transport, distribution, and uptake of therapeutics. While it is known that the host microenvironment may enable the advancement of malignancy and even the development of resistance, much of the mechanistic understanding by which this regulation occurs remains unclear. This is due in part to a lack of physiologically relevant models, though advancements in the emerging field of "tumor engineering" are beginning to challenge this. The transition away from incompatible animal models and simplified two-dimensional cultures has brought about the creation of advanced three-dimensional models in order to better simulate and test the microenvironmental characteristics that define human cancers. Nonetheless, few systems are able to capture the full range of authentic, complex in vivo events such as neovascularization, intravasation, and variable oxygen distribution. This work will explore the details of developing biologically-inspired, highly controlled in vitro tumor microenvironments to replicate and investigate these events. Such systems have the potential to mediate successful translation of preclinical research to clinical significance, while also providing mechanistic insight into the early stages of tumor progression and metastasis.

T cells discriminate peptide-major histocompatibility complexes (pMHCs) expressed on antigen presenting cells (APCs) via their T Cell Receptors (TCRs). The adaptability of the human immune system is demonstrated, in part, by the ability to generate a stunningly large number of possible TCRs (>1020-1061). However, matching TCRs with specific pMHC targets to invoke an appropriate and targeted immune response remains challenging and has created a need for a deeper understanding of how the TCR engages its antigen to induce an immunogenic response. Utilizing experimental observation and molecular dynamics, we propose a force-dependent kinetic proofreading discrimination model whereby the TCR must sustain and form transient bonds under load for sufficient time to initiate biochemical signaling. This computational model is utilized to construct the building blocks of TCR design by (1) machine learning the physiochemical determinants of TCR dissociation kinetics, (2) homology modelling patient-specific TCRs to a target pMHC, and (3) predicting TCR function to a target pMHC from primary amino acid sequence. The creation of highly immunogenic, tumor-specific TCRs will require rapid and efficient screening of TCR information space. The success of these techniques will be measured by the ability to accurately predict in vitro T cell immunogenicity and will depend on the generation of high-fidelity datasets. Moreover, additional biological complexity may need to be integrated into this computational framework to augment predictive power. Hence, we investigate the effects of glycosylation, coreceptors (CD3 and CD4), and the phospholipid bilayer on the TCR interaction with the pMHC. In addition, methodology and software is developed to analyze the non-equilibrium receptor-ligand kinetics in a microfluidic flow chamber. The methods proposed in this work are suggested to provide an architecture that may inform the design of novel TCRs for immunotherapies.

Chimeric Antigen Receptor (CAR) T-lymphocyte immunotherapy, effective in blood cancers, shows limited success in solid tumors (prostate, breast) due, in part, to an immunosuppressive tumor microenvironment (TME). Immunosuppression affects various cell-types, including tumor cells, macrophages, and endothelial cells. Therefore, a systematic analysis of these constituents is necessary to improve CAR T-lymphocytes’ tumoricidal function. Conventional murine-based models offer limited concordance with human immunology/biology. Therefore, we have developed a human “tumor-on-a-chip” (TOC) platform to study immunosuppression at high spatiotemporal resolution. Our polydimethylsiloxane (PDMS) microfluidic TOC features an endothelial cell-lined (EC) channel that mimics features of an in vivo capillary, such as molecular and cell transport across the endothelium and into the TME. We tested this using 70kDa Dextran and fluorescence-recovery-after-photobleaching (FRAP) measurements, confirming physiologic interstitial flow velocities (1.1-2 μm/sec). In vivo, soluble factors from a solid tumor impact the adjacent endothelial phenotype. Our device demonstrates that particles diffuse in the opposite direction of interstitial flow to reach the endothelium, 100 μm away, at concentrations as high as 20% of those at the tumor. Additionally, our results demonstrate that M2-like macrophages and endothelial cells affect tumor cell (prostate cancer cell line DU145) growth, clustering, and migration. M2-like macrophages in the TME also induce PD-L1 and abrogate ICAM-1 gene expression on the adjacent endothelium. These results are consistent with in vivo observations of limited CAR-T extravasation and effector function, pointing to a specific role of the M2-like macrophage.