Pharmacokinetic and Pharmacodynamic Analysis of Gemcitabine and Birinapant Combinations in Pancreatic Cancer
Pancreatic cancer is the one of the leading causes of cancer-related deaths in the United States and is characterized with low survival rate and high drug resistance. Because of the redundant and highly mutated signaling pathways in pancreatic cancer, numerous combinational therapies have been sought. Currently the selection of drug combinations is largely empirical and methods of evaluating and optimizing drug combinations have not been standardized. An important reason for this is the lack of comprehensive characterization of drug mechanisms of action and causes for drug resistance. The purposes of this dissertation are: first, to set up a paradigm for evaluating drug combinations mathematically and translating the evaluation methods from in vitro to in vivo preclinical systems; second, to serve as an example for characterizing the biological signaling pathways and drug pharmacology comprehensively with systems modeling approaches, supported with “big data” from advanced techniques such as proteomic analysis; and third, using such systems models, further selecting and optimizing drug combinations to reverse drug resistance and enhance efficacy. The two drugs selected are gemcitabine, a major component in the therapies for pancreatic cancer treatment, and birinapant, an antagonist of inhibitor of apoptosis proteins (IAP). In Chapter 1, the efficacy of this drug combination was evaluated in PANC-1 cells. A basic pharmacodynamic (PD) model was developed to characterize the temporal changes in the numbers of attached and floating cells after treatments, and synergistic effects were observed for both proliferation inhibition and death induction. Measurements of cell cycle distributions and apoptosis were then obtained and a mechanism-based PD model was developed to reveal more details and capture the major features of the beneficial interactions. From the mechanism-based PD model, different exposure schedules were tested and an optimal one to achieve maximal efficacy was proposed. Assumptions were made in developing the mechanism-based PD model in Chapter 1. In Chapter 2, a proteomic approach was utilized for a comprehensive, unbiased study of proteins perturbed by gemcitabine and birinapant to test previous hypotheses. The mechanisms of action for both drugs were characterized more intensively, and additional details were incorporated into the interaction knowledge described previously. Based on the proteomics data, reasons for gemcitabine resistance were discussed, and regulators of DNA damage responses involving DNA repair, anti-apoptosis, and pro-migration and invasion proteins were proposed as promising candidates for therapeutic targeting. With the rich quantitative proteomics data, a network modeling approach was attempted in Chapter 3. Quantitative relationships were developed for selected signaling pathways of cell cycle regulation, DNA damage responses, DNA repair, apoptosis, NF-κB, and MAPK-p38, which were then linked to describe the cell cycle progression and apoptosis, and finally to changes in cell numbers. Based on the developed network model, simulations were made under different conditions and compared with observations, serving as a validation process. The impact of p53 mutation and p53 silencing on the efficacy of gemcitabine was tested with this model. Sobol Sensitivity Analysis was applied to select promising targets to be combined with gemcitabine. In addition, the efficacy of curcumin combined with gemcitabine was evaluated based on the model simulation. With extensive evaluation and comprehensive characterization of the mechanisms of this drug combination in cell culture, efforts were continued to investigate the effects of the combination in a mice xenograft model. In Chapter 4, pharmacokinetic information for gemcitabine and birinapant was gathered from the literature and full physiologically-based pharmacokinetic models (PBPK) were developed to characterize drug distribution in the body and into the pancreatic tumor. The tumor concentrations then were used to drive inhibition in tumor growth and a semi-mechanistic PBPK/PD model was developed to evaluate the efficacy of the drug combination in vivo. Their joint effects were revealed as merely additive. The network model developed in Chapter 3 was introduced to bridge the PBPK and PD models, and reasons for the discrepancies in vitro and in vivo were explored. Model predictions showed that simultaneous dosing was preferable to sequential dosing in vivo with stronger suppression of the DNA repair signaling. In summary, this dissertation proposed a paradigm for evaluating drug combinations quantitatively in preclinical systems of cell lines and xenograft models. Comprehensive characterization of drug mechanisms of action and biological systems through network modeling can facilitate the selection and optimization of candidates for anti-cancer combination therapy. The bridging of knowledge in different scales with mathematical models in different complexity helps to minimize the gap of translating from in vitro to in vivo or even from preclinical to clinical research.