MODELING SERUM-DRUG KINETICS FOR THE PREDICTION OF CANCER RESISTANCE TO SECOND LINE THERAPEUTICS
Open Access
- Author:
- Mc Ilroy, Kyle
- Graduate Program:
- Bioengineering
- Degree:
- Master of Science
- Document Type:
- Master Thesis
- Date of Defense:
- April 06, 2020
- Committee Members:
- Justin R Pritchard, Thesis Advisor/Co-Advisor
William O Hancock, Program Head/Chair
Deb Kelly, Committee Member
William O Hancock, Committee Member - Keywords:
- cancer
- Abstract:
- Targeted first-line cancer therapies can achieve 60-80% objective response rates in advanced patients. A BCR-ABL1 mutation, which causes chronic myeloid leukemia (CML), for instance, is inhibited extremely effectively by imatinib. However, many patients who relapse on treatment show further mutations that hinder imatinib binding. Patients who relapse on targeted therapies require second and even third line therapies that are tailored to the genetic mutations present in the tumor. Oncologists have developed methods to predict how to give the right drug to the right patient. However, most personalized medicine studies do not account for the effects of human serum binding. The current clinical standard to correct for this is a corrective factor which is developed by dialyzing the targeted drug against serum protein. This method uses an in-vitro equilibrium measurement that ignores the in-vivo competitive equilibrium between the drug target (with nm affinity) and the nonspecific serum binding effects. While this effect is rarely accounted for in cancer studies, the anti-microbial field frequently acknowledges relative drug kinetics through the application of a shift-assay. Here we propose a strategy that utilizes an in-vitro measurement of drug sensitivity in the presence of serum to predict the clinical efficacy of tyrosine kinase inhibitors for many target driven point mutations in gastrointestinal stromal tumor (GIST), CML, and toxicity for EGFR inhibitors. The clinical data used were obtained from the literature and were used to create a statistical method to predict drug resistance. For individual resistant mutant single point mutations were genetically engineered into [Ba/F3,A431] human cells. Isogenic measurement of point mutations is preferred because patient samples are rare and have variable genetic backgrounds. Following the generation of mutant cell lines, two sets of IC50 experiments were conducted to measure the difference that the addition of protein serum would have on cell viability. Serum-added IC50 experiments were conducted similarly, but the medium also had both 341mM human serum albumin and 1 mg/ml human a1-acid glycoprotein (to mimic nonspecific human blood serum). Cell viability for all experiments was plotted against log drug concentration and Hill equations were fitted to the data. In spite of the astonishing genetic and microenvironmental complexity of human tumors, qualitative clinical responses for patients with on-target mutations in tyrosine kinases can be predicted with careful in-vitro pharmacology that accounts for competitive binding with human serum. Isolating the effect of tyrosine kinase mutations and dosing at a clinically relevant concentrations in the presence of physiologic levels of human serum competitively inhibits drug efficacy. Quantitating this effect across multiple diseases and targets allows for a dramatic improvement in predictive ROC of clinical efficacy and toxicity in mutation targeted TKI therapy. Despite diverse in-vitro cancer models that incorporate many in-vivo features into models of cancer (including organoids), in oncogene addicted TKI systems. Our predictions are as accurate as those recently published in GI organoids.