Development and characterization of a differentiated thyroid cancer cell line resistant to VEGFR-targeted kinase inhibitors

Abstract

Background: Vascular endothelial growth factor-targeted kinase inhibitors have emerged as highly promising therapies for radioiodine-refractory metastatic differentiated thyroid cancer. Unfortunately, drug resistance uniformly develops, limiting their therapeutic efficacies and thereby constituting a major clinical problem.

Approach and methods: To study acquired drug resistance and elucidate underlying mechanisms in this setting, BHP2-7 human differentiated thyroid cancer cells were subjected to prolonged continuous in vitro selection with 18 μM pazopanib, a clinically relevant concentration; acquisition of pazopanib resistance was serially assessed, with the resulting resistant cells thereafter subcloned and characterized to assess potential mechanisms of acquired pazopanib resistance.

Results: Stable 2- to 4-fold in vitro pazopanib resistance emerged in response to pazopanib selection associated with similar in vitro growth characteristics but with markedly more aggressive in vivo xenograft growth. Selected cells were cross-resistant to sunitinib and to a lesser extent sorafenib but not to MAPK kinase (MEK1/2) inhibition by GSK1120212. Genotyping demonstrated acquisition of a novel activating KRAS codon 13 GGC to GTT (glycine to valine) mutation, consistent with the observed resistance to upstream vascular endothelial growth factor receptor inhibition yet sensitivity to downstream MAPK kinase (MEK1/2) inhibition.

Conclusions: Selection of thyroid cancer cells with clinically utilized therapeutics can lead to acquired drug resistance and altered in vivo xenograft behavior that can recapitulate analogous drug resistance observed in patients. This approach has the potential to lead to insights into acquired treatment-related drug resistance in thyroid cancers that can be subjected to subsequent validation in serially collected patient samples and that has the potential to yield preemptive and responsive approaches to dealing with this important clinical problem.

Figure 1.

In vitro pazopanib selection of BHP2–7 cells led to in vitro pazopanib resistance and also to more aggressive in vivo xenograft behavior. A, 18 μM continuous in vitro pazopanib exposure led to pazopanib resistance as assessed via colony forming assays (continuous pazopanib exposures; WT, parental BHP2–7 cells; pazopanib selected, mixed culture pazopanib selected BHP2–7 cells). Figure 1B, subcutaneously implanted pazopanib selected/resistant cell nu/nu mouse flank xenografts demonstrated more rapid tumor growth compared to parental/unselected BHP2–7 cells (P < .01).

Figure 2.

Mutational assessment of pazopanib-resistant BHP2–7 cells indicated acquisition of a novel KRAS mutation. A, Depiction of melting curve analysis from the LightCycler screen assay (Roche) for duplicates of a no reverse transcriptase (RT) control (green and pink), cDNA of a wild-type cell line transcript (green and blue), and the cDNA of the pazopanib-resistant cell line transcript (red and black). The abscissa represents the temperature of the capillary and the ordinate represents the fluorescence intensity as read by the LightCycler. B, Graph of the negative first derivative of the melting curve shown in panel A. Capillary temperature is again on the abscissa and the negative first derivative of the fluorescence [-d(F2/F1)/dT] is on the ordinate. The mutated form of KRAS codon 13 is responsible for the population of cDNA in which the melting curve peaks at 48.76°C. Wild-type cDNA is represented at the 62.09° peak in the curve. C, Sanger sequencing trace of the pazopanib-resistant BHP2–7 line cDNA against wild-type BHP2–7 cDNA. Highlighted in red is the KRAS codon 13 allele GGC showing the mutation to GTT in the resistant line.

Figure 3.

Pazopanib resistance was maintained in cloned sublines of pazopanib-selected/resistant BHP2–7 cells. A and B, Seenteen unique cloned sublines were created from the BHP2–7 pazopanib-resistant/selected BHP, demonstrating maintained and stable pazopanib resistance as assessed via in colony-forming assays (continuous pazopanib exposures, representative data from clones 7 and 14 shown). All 17 sublines were additionally sequenced for KRAS mutations, and all contained the identical KRAS G13V mutation (see Figure 2). C and D, Cross-resistance to sunitinib and, to a lesser extent with respect to sorafenib, was demonstrated in pazopanib-selected/resistant BHP2–7 subline 7 (similar results were seen for subline 14, data not shown).

Figure 4.

Pazopanib selected/resistant BHP2–7 cells were equally sensitive to the MEK1/2 inhibitor GSK1120212 compared with parental BHP2–7 wt cells, with activation of VEGFR downstream signaling through pERK demonstrated in pazopanib selected/resistant/KRAS mutant BHP2–7 cells. A, Cross-resistance to GSK1120212 was not observed in subline 7 KRAS mutant cells compared with parental BHP2–7 KRAS wt cells (assessed by colony forming assays, continuous drug exposures). B, ERK1/2 phosphorylation was attenuated in BHP2–7 wt cells in response to 18 μM pazopanib exposure but not in KRAS mutant subclones 7 and 14. In contrast, ERK1/2 phosphorylation was similarly attenuated in response to MEK inhibition (10 nM GSK1120212) in parental and in KRAS mutant cell lines.