April 27, 2016
Hosted by MIT
Sponsored by Pfizer Inc.
1st Place in the Student Category: Nikolaos Tsamandouras
Assessment of population variability in hepatic drug metabolism using a perfused 3D human liver bioreactor along with modeling and simulation techniques
Prediction of hepatic clearance of a new chemical entity is of significant importance during the early stages of drug development. Such a prediction should be ideally performed not at the “average individual” but at the “population” level in order to efficiently guide the design of clinical studies in humans. This work aims to employ a perfused 3D hepatocyte culture platform (LiverchipTM) to study hepatic metabolic clearance in vitro along with the associated population variability.
Metabolic depletion of 6 compounds was investigated in the LiverchipTM across 5 different hepatocyte donors. Mixed effects modeling was employed to allow the analysis of these multi-level data (several donors, wells) and the subsequent identification of in vitro metabolic clearance predictors. A population PBPK model was developed for one of the studied compounds (lidocaine) in order to illustrate the framework for translation of the in vitro output to the observed pharmacokinetic variability in vivo.
Substantial variability was observed across different donors. Specifically, intrinsic metabolic clearance along with inter-donor variability (reported as coefficient of variation in parenthesis) was determined to be 3.88 (66.8%), 0.81 (29.3%), 8.91 (24.1%), 4.38 (28.5%), 5.02 (32.6%) and 17.8 (36.2%) μL/min/10^6 cells for propranolol, prednisolone, phenacetin, lidocaine, ibuprofen and diclofenac respectively. Albumin, urea, LDH and CYP mRNA levels during the pre-dose culturing period were identified as significant predictors of in vitro metabolic clearance. The latter showed good correlation with in vivo human values for the studied set of compounds. Stochastic simulations of the developed population PBPK model for lidocaine successfully predicted the observed clinical concentration-time profiles and the associated population variability.
In conclusion, this work employs a novel in vitro system to investigate inter-individual variability in drug metabolism and proposes a framework that allows prediction of concentration-time profiles at the “population” level, during the early stages of drug development.
1st Place in the Non-student Category: Jangir Selimkhanov
Model of mouse energy balance improves statistical power for studies of multiple metabolic endpoints
Jangir Selimkhanov, Juen Guo, Kevin D. Hall, C.J. Musante and W. Clayton Thompson
Background/Objectives: The ability to design well-powered animal studies is key to improving our knowledge of obesity as well as identifying and effectively testing potential anti-obesity drug targets. This process is complicated by the difficulty of precisely measuring endpoints such as body composition and food intake and the lack of understanding of how these endpoints may vary within and between study animals. Due to the complexity of the obesity phenotype and difficulty in obtaining precise measurements, poorly designed studies can have a significant negative impact on an anti-obesity program.
Comprehensive Single and Paired Drug Target Identification in Healthy and Disease Models of NF-kB Pathway
Defective levels of the transcription factor NF-kBn has been associated with cancer, inflammatory, and autoimmune diseases. Inhibiting species in the NF-kB pathway may normalize the levels of NF-kBn. However, inhibition of the various species in the pathway do not all have the same level of effect on NF-kBn’s output level as some might have negligible effects while others greatly normalize it. Therefore, it is important to identify species/targets in the NF-kB pathway that that if inhibited would normalize the levels of NF-kBn. Also, since there is a limit to how effective single drug targets are in complicated diseases, identification of combinations of targets is also necessary. Existing work are (1) disease specific, (2) only analyze a small subset of individual targets (3) generally do not explore combinations of targets. In this project, we (1) computationally modify a healthy model of the pathway to reflect what is defective in different diseases to not be disease specific; (2) score the effectiveness of individual targets based on how much they normalize NF-kBn when inhibited at various degrees of inhibition; (3) classify pairs of targets and obtain general rules for synergism of target combinations. We find in all the models that very few species are effective at low levels of inhibition but many are effective at extremely high levels. We also find that the scores of the species greatly depend on their steady state concentrations, relative reaction parameters, and relative concentrations of species they form complexes with. Furthermore, we find that all species that have some effect on the output are synergistic with one another. The effective targets and general principles identified in this project would lead to better drug discovery for NF-kB associated diseases. Furthermore, this approach can be readily applied to identify effective targets in other pathways associated with diseases.
Accurate Atom-by-Atom Predictions of Solvation Electrostatics Using a Hydration-Shell Poisson-Boltzmann Model
Jay Bardhan, Spencer Goossens, Amirhossein Molavi Tabrizi
To improve implicit-solvent model accuracy without sacrificing speed, we test whether nonlinear hydration-shell effects can be modeled by modifying the usual dielectric boundary conditions. We find that simple nonlinear corrections lead to electrostatic models that 1) do not need individually fitted atom radii, reducing the number of free parameters by a factor of ten; 2) offer better accuracy than standard Poisson-Boltzmann models; 3) reproduce ion solvation thermodynamics; 4) reproduce atom-by-atom charging energies including unfavorable hydrophobic solvation.
(Patho)physiological Crosstalk in Gut-Liver Axis Revealed by Integration of Microdevices, Multi-cellular Models and Computational Systems Analysis
Kelly WL Chen, Collin Edington, Emily Suter, Jeremy Velazquez, Rachel Dyer, Jason Velazquez, Michael Shockley, Rebecca Carrier, Murat Cirit, Linda G. Griffith, Douglas A. Lauffenburger
Complex diseases often arise from network-level dysregulation as a result of perturbations across multiple tissues. Incomplete understanding of tissue crosstalk can undermine accurate diagnosis and treatment of disease conditions. For instance, increased intestinal permeability is a hallmark of many chronic diseases and adverse drug reactions. While ‘leaky gut’ is often considered a symptom of disease, it is increasingly appreciated that chronic disruption of intestinal barrier and concomitant alterations in immune homeostasis can exacerbate disease progression and have far-reaching effect on downstream organs. However, a quantitative understanding of how these multicellular tissues communicate and contribute to overall (patho)physiology is limited. To this end, we have developed and implemented an in vitro platform, together with the immune-competent human liver and gut models, to interrogate gut-liver interaction under normal and perturbed contexts.
Our results demonstrated long-term maintenance of intestinal and hepatic functions in baseline interaction. Interestingly, gene set enrichment analysis of RNAseq data comparing liver in interaction versus isolated controls revealed modulation of bile acid metabolism. Gut-liver crosstalk potentiated feedback inhibition of hepatic Cyp7A1 expression, consistent with known physiology. Under inflammatory conditions, non-linear modulation of cytokine responses was observed in the interacting system as compared to isolated controls. In particular, CXCR3 ligand production was significantly enhanced in the integrated system. Multivariate regression together with gene set enrichment analyses revealed that IFNα, IFNγ and TNFα signaling were likely involved in the synergistic amplification of CXCR3 ligand production.
Our experimental approach, which combines in vitro culture models, microdevices and quantitative analysis, is generalizable to study higher order organ interactions, and has broad applicability towards advancing understanding of human (patho)physiology and drug development.
Role of Sickle Hemoglobin Polymerization in Erythrocyte Aging and Death
Sickle Cell Disease (SCD) is a genetic disorder caused by a point mutation on the beta-chain of adult hemoglobin. In deoxygenated conditions, the mutated hemoglobin (HbS) aggregates and forms long polymers causing deformation and sickling of red blood cells (RBCs). These deformed RBCs occlude the microvascular system, leading to painful inflammatory responses called vaso-occlusive crises (VOCs) and organ damage. In addition sickling affects the survival of RBCs and reduces their lifespan leading to comorbidities like anemia. We are developing a systems model of SCD that can be used to build quantitative links between multi-scale pathobiological events and identify and validate new intervention points. Here we describe our approach and progress in the development of this systems model, specifically focusing on our model of hemoglobin polymerization, sickling and RBC survival sickling. We use this model to explain the roles of oxygen partial pressure (pO2) and intracellular HbS concentration in polymerization kinetics and RBC sickling. Further we show how polymerization and sickling can affect the aging, senescence and survival of RBCs. Finally we use the model to understand the mechanism and viability of different therapeutic strategies such as Hydroxyurea, RBC ion channel blockers and targeting of HbS with small molecule allosteric modifiers.
Experimental and Computational Method Characterizes Non-genetic Drug Resistance Mechanisms
Targeted cancer therapeutics such as tyrosine kinase inhibitors (TKIs) have seen constraint in clinical efficacy due to both intrinsic and acquired resistance. Indicators for sensitivity to TKIs may include genetic mutations or protein-level overexpression of targeted or bypass receptor tyrosine kinases (RTKs). While the latter is often attributed to gene amplification, genetic characterization of tumor biopsies has failed to explain substantial proportions of resistance. Therefore we hypothesize that post-synthesis mechanisms governing RTK levels, including receptor trafficking and proteolytic cleavage by pericellular proteases, may represent underappreciated contributors to drug resistance. We have developed a combined computational and experimental framework novel in its quantitative assessment of the contributions of various mechanisms involved in resistance to TKIs that is widely applicable. When applied to a model of intrinsic resistance to Mek inhibition in a triple negative breast cancer (TNBC) cell line previously characterized for the contribution of protein transcription to resistance, we identify that post-synthesis mechanisms vary greatly. This finding highlights the importance of studying integrated mechanisms that collectively lead to protein level changes to capture the full complexity of the system.
Designing Microphysiological Systems: From Bedside to Bench
Christian Maass, Nikolaos Tsamandouras, Cindy Stokes, Murat Cirit, Linda Griffith
Development of Physiome-on-a-chip or microphysiological systems (MPS) provides an in vitro tool for relevant studies of pharmacodynamics and pharmacokinetics of drugs and biological surrogates before clinical trials. Identifying feasible organ functions and subsequently scaling of such (e.g. drug clearance, insulin production) for MPSs is a critical step. State-of-the-art design approaches include direct miniaturization and allometric scaling. More recently, functional scaling is employed to design MPSs to recapitulate specific organ/tissue functions, e.g. metabolic rates. However, using these approaches, scaling of individual or multiple organs relative to each other and achieving a physiologically relevant environment remains challenging. Therefore, a generalizable approach on how to design MPSs is needed.
A Biomimetic Model of Angiogenesis for Therapeutic Applications
Duc-Huy T. Nguyen, Sarah C. Stapleton, Michael T. Yang, Susie S. Cha, Colin K. Choi, Peter A. Galie, Christopher S. Chen.
Angiogenesis is an outgrowth of new blood vessels from existing vasculature. It plays critical roles in human physiology (fetal growth, wound healing, and tissue repair) and many diseases (retinopathies, hemangiomas, and cancer). It is highly regulated by many different biochemical and mechanical cues in the environment. Many pro- and anti-angiogenic factors have been identified in both physiological and pathological diseases. However, effort to achieve therapeutic angiogenesis has remained limited with no FDA-approved therapeutic drugs. Anti-angiogenic therapy such as anti-VEGF in cancer has achieved some successes to improve the patient survival in combination with chemotherapy. Unfortunately, not all cancers respond equally to anti-VEGF treatment. This suggests that there is an urgent need to develop in vitro screening assays to comprehend better the mechanisms underlying angiogenesis and lower the cost of animal experiments. There has been different in vitro assays to capture the many processes of angiogenesis such as Matrigel assay and bead sprouting assay. However, these assays often lack some physiological features of blood vessels such as correct apical-basal polarity and fluid shear-stress. To overcome the short-comings of these assays, we developed an AngioChip, which has two hollow cylindrical channels embedded in 3D collagen matrix. In one channel, we seeded endothelial cells (ECs) to form a monolayer of endothelium while the other channel contains angiogenic factors to trigger angiogenic response of ECs from the biomimetic blood vessel. Using this model, we screened for the angiogenic potentials of different biochemical factors and identified different cocktails of factors that trigger robust response from ECs. Surprisingly, we also found that one of the cocktails appeared to be non-responsive to VEGF inhibition. Our model suggests the potential applications of our AngioChip to capture not only the morphogenetic processes of angiogenesis but also the capability to identify therapeutic compounds to treat angiogenesis-related-diseases.
Pathway-based modeling finds hidden genes in shRNA screen for regulators of Acute Lymphoblastic Leukemia
Jennifer L. Wilson, Simona Dalin, Sara Gosline, Michael Hemann, Ernest Fraenkel, Doug Lauffenburger
Data integration stands to improve interpretation of RNAi screens which, as a result of off-target effects, typically yield numerous gene hits, of which, only a few validate. These off-target effects can result from seed matches to unintended gene targets (reagent-based) or cellular pathways which can compensate for gene perturbations (biology-based). We focus on the biology-based effects and use network modeling tools to discover pathways de novo around RNAi hits. By looking at hits in a functional context, we can uncover novel biology not identified from any individual ‘omics measurement. We validate this pathway through a combination of pathways enrichment and targeted gene perturbation experiments.
3D Metastatic Breast Cancer Model Recapitulates Dormancy and Re-emergence
Carissa L. Young, et al.
Approximately 90% of cancer-associated mortality is a consequence of distant metastasis, a multistep process whereby cells from the primary tumor migrate to and colonize secondary organs. These cancerous cells may either proliferate immediately or lay dormant (e.g., as pre-malignant micrometastasis) for years before forming clinically overt macrometastases. Clinically, undetected metastases have serious implications for cancer patients; exemplified by ~ 33% of women whom suffer a metastatic relapse within 5 years following removal of breast cancer.
In general, distant metastases remain incurable as current chemotherapies are ineffective against them underscoring the need to develop improved therapeutic approaches, rationally designed and effective across a broad range of patient populations. The liver represents an ideal system to evaluate metastasis and efficacy of cancer therapeutics as:
(1) a common site of metastasis for many carcinomas, e.g., breast, colon, pancreatic, and lung
(2) the major organ for drug metabolism to assess efficacy and limiting toxicities of cancer therapies
(3) a primary location of systemic regulation of circadian rhythms via modulation of nutrients, hormones, and cytokines
The absence of accessible all-human ex vivo metastatic models has hindered the identification of mechanistic insights towards targeted therapeutic intervention, plausible biomarkers of disease progression, and human-specific cross-talk between the tumor and the metastatic microenvironment. We developed a 3D metastatic model - composed of human hepatocytes, non-parenchymal cells, and breast cancer cell lines - enabling the evaluation of tumor growth, dormancy, and re-emergence amongst multiple donors, as a consequence of chemotherapies and inflammatory stimuli. Multivariate statistical analyses integrated metrics of clinical and high-content signaling assays, thereby providing biological insights of plausible signatures and communication networks in early metastatic disease. Collectively, our results recapitulate human disease pathophysiology using an all-human ex vivo microphysiological system, which we believe will ultimately improve therapeutic strategies that target breast cancer dormancy and re-emergence.