Utilizing a high-throughput approach to identify effective systemic agents for aggressive thyroid cancer variants

Presentation: AHNS-003
Topic: Endocrine
Type: Oral
Date: Wednesday, May 1, 2019
Session: 1:10 PM - 1:55 PM Scientific Session 1 - Endocrine
Authors: Abdallah S Mohamed, MD1, Ying Henderson1, Yunyun Chen1, Clifford Stephan2, Gilbert Cote1, Maria Cabanillas1, Mark Zafereo1, Vlad Sandulache3, Stephen Y Lai1
Institution(s): 1MD Anderson Cancer Center, 2Institute of Biosciences and Technology, Texas A&M University, 3Baylor College of Medicine

Background: Despite the use of aggressive, multimodality treatment most anaplastic thyroid carcinoma (ATC) patients die within a year of diagnosis. Although the combination of BRAF and MEK inhibition has been recently approved for use in ATC, it remains effective in a minority of patients who ultimately develop drug resistance. 

Objective: To identify effective systemic agents against aggressive thyroid carcinoma (TC) variants.

Methods: Twelve short tandem repeat (STR) validated human TC-derived cell lines were used (ATC n=7, poorly differentiated (PDTC) n=1, papillary (PTC) n=4). All cell lines underwent comprehensive genomic characterization prior to drug testing. The details of the types and the mutational profiles of these cell lines are summarized in Table 1. High-throughput drug screens were performed using the NCI’s Approved Oncology Set V (n=114) and a custom collection of FDA approved drugs, investigational agents and mechanistically annotated compounds (n=153). The effect of drugs on cell growth and survival was measured after 72 hours of drug exposure. To identify the most effective drugs, we selected individual agents with maximal growth inhibition at each dose level relative to wells examined on the day of treatment (top 25th percentile) and subsequently used non-parametric statistics to compare effect size with other drugs and controls. The concentration-response curves from biological replicates of different passage number were fitted using a non-linear regression analysis to a 4-parameter logistic equation and the AUC was calculated and used for the development of pharmacologic trees (Figure 1). Confirmatory testing was completed for the most effective drug classes which were then stratified by cell line type and genomic background. 

Results: We were able to identify the most effective compounds for each cell line. The most effective classes of agents against ATC cell lines were: antimetabolites, inducers of reactive of oxygen species (ROS), proteasome and microtubule inhibitors. These agent classes in addition to HDAC inhibitors achieved the highest effectiveness for PTC cell lines at 0.1μM dose level but only proteasome and microtubule inhibitors remained effective at 0.01μM dose level. TP53 mutational status impacted drug sensitivity; mutant TP53 cell lines demonstrated enhanced sensitivity to pralatrexate and vinca alkaloids, while wild-type TP53 cell lines demonstrated preferential sensitivity to HDAC inhibitors. Likewise, BRAF mutational status affected drug sensitivity with higher sensitivity to taxanes and protein kinase inhibitors in V600E mutation compared with preferential sensitivity to proteasome and HDAC inhibitors in wild-type. Confirmatory testing validated initial screening results.

Conclusion: High-throughput screening identified classes of systemic agents which demonstrate preferential effectiveness against aggressive TC variants, particularly those with mutant BRAF and TP53. These agents provide a basis for in vivo preclinical validation prior to clinical trial development.  

Table 1. TC cell lines.

Figure 1. Example of ATC cell line (MDA-T178) results. A) Boxplots of effective drugs in the initial screen compared with DMSO and other ineffective drugs; B) Selected drugs at a 0.1μM concentration with equivalent efficacy compared with 1μM dose; C) Drug tree analysis where the size of the colored dots represents relative effectiveness of each individual agent, and D) Confirmatory 8-dose drug response curves.