Thyrocyte-specific inactivation of p53 and Pten results in anaplastic thyroid carcinomas faithfully recapitulating human tumors

Abstract

Anaplastic thyroid carcinoma (ATC) is the most aggressive form of thyroid cancer, and often derives from pre-existing well-differentiated tumors. Despite a relatively low prevalence, it accounts for a disproportionate number of thyroid cancer-related deaths, due to its resistance to any therapeutic approach. Here we describe the first mouse model of ATC, obtained by combining in the mouse thyroid follicular cells two molecular hallmarks of human ATC: activation of PI3K (via Pten deletion) and inactivation of p53. By 9 months of age, over 75% of the compound mutant mice develop aggressive, undifferentiated thyroid tumors that evolve from pre-existing follicular hyperplasia and carcinoma. These tumors display all the features of their human counterpart, including pleomorphism, epithelial-mesenchymal transition, aneuploidy, local invasion, and distant metastases. Expression profiling of the murine ATCs reveals a significant overlap with genes found deregulated in human ATC, including genes involved in mitosis control. Furthermore, similar to the human tumors, [Pten, p53]thyr-/- tumors and cells are highly glycolytic and remarkably sensitive to glycolysis inhibitors, which synergize with standard chemotherapy. Taken together, our results show that combined PI3K activation and p53 loss faithfully reproduce the development of thyroid anaplastic carcinomas, and provide a compelling rationale for targeting glycolysis to increase chemotherapy response in ATC patients.

Figure 1. Clinicopathological features of [Pten, p53]^thyr−/− mice

(A) Kaplan-Meyer analysis of the effect of progressive p53 deletion on the survival of Pten^thyr−/− mice. (B, C) TSH and T4 serum levels in control, single mutant and double mutant (DM) mice. (D-K) Histopatological features of tumors developing in [Pten, p53]^thyr−/− mice. (D) Anaplastic carcinoma (left) flanking an area of well-differentiated follicular carcinoma (right). Undifferentiated tumors display areas of spindle cell morphology (E), with frequent giant cells (F), and occasional bone metaplasia (G). Tumors invade locally into the muscle (H), and the trachea (I), and metastasize to the lungs (J) or, sporadically, to the liver (K). Bar: 100μm.

Figure 2

Expression profiling by real-time PCR of a panel of thyroid differentiation markers using thyroid pools from 4-month old control and single mutant mice, ATC-free, progressively older (4, 6, 8 months) double mutants, and histologically confirmed ATCs from 8- to 9-month old double mutants.

Figure 3. Mouse ATCs display chromosomal instability and aneuploidy

(A) Chromosome counts in two representative early passage (p2) primary cultures from histologically confirmed ATCs. Note the wide distribution of chromosome numbers within the same culture. (B) Karyotypic analysis of two representative cells from independent tumors, showing chromatid breaks (thick arrows), chromosome breaks (thin arrows), translocations (#), and complex rearrangements (*).

Figure 4. [Pten, p53]^thyr−/− ATCs undergo EMT

(A) H&E staining of a representative tumor area containing both a follicular carcinoma (left) and an anaplastic carcinoma (right) component. (B-D) Immunohistochemical detection of phospho-Smad 2 (B), E-cadherin (C), and vimentin (D). Bar: 100μm.

Figure 5. ATCs are addicted to driver gene signaling

(A) Western blotting analysis of the activation status of Akt and MAPK pathways in 4-month old control and single mutant mice, ATC-free, progressively older (4, 6, 8 months) double mutants, and histologically confirmed ATCs from 8- to 9-month old double mutants. (B) immunohistochemical detection of activated Akt, S6 ribosomal protein, and ERK1/2 in tumor areas with both well-differentiated and anaplastic components. (C) IC₅₀ values for an Akt inhibitor (MK-2206) and a MEK inhibitor (U0126) in a panel of mouse and human anaplastic (ATC) and follicular (FTC) carcinoma cell lines. On the bottom, a Western blotting showing the effect on the activation of Akt and ERK1/2 of one hour exposure to these inhibitors (MK-2206: 500nM, U0126: 10μM).

Figure 6. Expression profiling validates the [Pten, p53]^thyr−/− mouse as a clinically relevant model of human ATC

(A) Hierarchical clustering showing the unique expression profile of mouse ATCs when compared to controls, single mutants, ATC-free double mutants (DM), and follicular carcinomas developing in aging Pten^thyr−/− mice (FTC). (B) Venn's diagram showing the significant overlap between genes differentially regulated in mouse and human ATCs. (C) GSEA analysis of the 430 genes common to mouse and human ATC, showing significant enrichment in genes previously shown to be altered in human ATC cell lines as well as genes involved in the control of mitosis.

Figure 7. Real-time PCR validation of a panel of genes differentially regulated in both mouse and human ATCs in 4-month old control and single mutant mice, ATC-free, progressively older (4, 6, 8 months) double mutants, and histologically confirmed ATCs from 8- to 9-month old double mutants

Figure 8. Mouse ATCs undergo a glycolytic switch (Warburg effect)

(A) Coronal section (transverse in the inset) of representative ¹⁸FDG PET scans of 8-month old control and compound mutant mice, showing high ¹⁸FDG accumulation in the thyroid tumor. (B) H&E and immunohistochemical detection of Hexokinase II and Pyruvate kinase M2 in the anaplastic component of a representative mouse tumor. The line demarcates the boundary between well-differentiated and anaplastic component. (C) Western blotting showing the increased expression of Hif1α, Hexokinase II and Pyruvate kinase M2 in four mouse ATCs (T1-T4). (D) Real-time PCR analysis of the expression of Hexokinase II, Pyruvate kinase M2, and Lactate dehydrogenase A in five mouse ATCs (T1-T5). (E) Lactate content of four wild type thyroids and five ATCs (T1-T5), showing consistent increase in neoplastic glands.

Figure 9. Both mouse and human ATC cell lines are highly sensitive to glycolytic inhibitors

(A) Top panel: dose-response curves showing the effect of 2-deoxyglucose (2-DG) on the viability of two mouse ATC cell lines. Middle panel: dose-response curves showing the. Bottom panel: isobolograms showing strong synergy between the two compounds. (B) Top panel: dose-response curves showing the cooperative effect of a 2-DG/Doxorubicin combination on the viability of two human ATC cell lines. Middle panel: isobolograms showing strong synergy between the two compounds. (C) Isobolograms showing strong synergy between doxorubicin and another glycolytic inhibitor, 3-bromopyruvate (3-BP).