REVIEW Normal vs cancer thyroid stem cells: the road to transformation M Zane 1,2,4 , E Scavo 1,4 , V Catalano 1 , M Bonanno 1 , M Todaro 1 , R De Maria 3 and G Stassi 1 Recent investigations in thyroid carcinogenesis have led to the isolation and characterisation of a subpopulation of stem-like cells, responsible for tumour initiation, progression and metastasis. Nevertheless, the cellular origin of thyroid cancer stem cells (SCs) remains unknown and it is still necessary to define the process and the target population that sustain malignant transformation of tissue-resident SCs or the reprogramming of a more differentiated cell. Here, we will critically discuss new insights into thyroid SCs as a potential source of cancer formation in light of the available information on the oncogenic role of genetic modifications that occur during thyroid cancer development. Understanding the fine mechanisms that regulate tumour transformation may provide new ground for clinical intervention in terms of prevention, diagnosis and therapy. Oncogene advance online publication, 11 May 2015; doi:10.1038/onc.2015.138 INTRODUCTION Thyroid cancer (TC) accounts for 96% of endocrine malignancies with 62 980 new cases expected to be diagnosed in the US in 2014, where it represents the second most common cancer among adolescents ages 15–19 (www.cancer.org). Despite of a global increase in incidence over the past three decades, the mortality rate remains low. This is a consequence of a favourable prognosis for the more frequent well-differentiated forms, subdivided into papillary (PTC) and follicular TC (FTC). 1 By retaining the differentiated features of normal thyrocytes, includ- ing the ability to concentrate iodine, in most cases these tumours can be treated successfully by surgical resection, followed by radioactive-iodine administration. 2 In contrast, the rare undiffer- entiated anaplastic TCs (ATCs), have a very-poor prognosis because of their invasiveness and metastatic behaviour (Figure 1) as well as their insensitivity to radioactive-iodine treatment for lack of an iodine symporter. 3 Alterations in key signalling pathways are proposed for distinct forms of thyroid transformation. Gain-of-function mutations in the thyrotropin receptor (TSH-R) or Gsα encoding genes, result in increased cAMP accumulation and TSH-independent proliferation, which in turn account for hyperfunctional adenomas, benign lesions without propensity towards malignant progression. Constitutive activation of the MAPK pathway seems to be the hallmark of different forms of TC. 2 Genomic alterations of the proto-oncogene tyrosine-protein kinase receptor Ret, the neuro- trophic tyrosine kinase receptor, as well as the intracellular signal transducer Ras and the serine/threonine-protein kinase B-Raf, have clearly been implicated in the pathogenesis of PTCs. 4 Similarly, the chromosomal translocation t(2;3)(q13;p25), which fuses the transcription factor paired box protein Pax-8 (Pax-8) and peroxisome proliferator-activated receptor gamma (PPAR-γ) encoding genes, has been identified in significant proportions in FTCs. 5 In addition to RAS mutations, another common event of these tumours is the PI3K pathway aberrant activation through mutation of the catalytic subunit p110 (PI3KCA) and loss of PTEN (Figure 2). 6 The multistep carcinogenesis model suggests that ATCs arise by way of a dedifferentiation process from pre-existing FTC or PTC (Figure 3). 7 The additional genetic events involved in the progression towards tumour dedifferentiation are (i) the inactivat- ing point mutation in the TP53 gene 8–10 and (ii) the activating mutation in the β-catenin encoding gene CTNNB1. 11,12 Evidence in favour of this multistep carcinogenesis model includes, the presence of well-differentiated TC within ATC specimens and the coexistence of BRAF gene and TP53 gene mutations in both undifferentiated and differentiated carcinomas. 13,14 However, this model is not in accordance with the rare occurrence of RET/PTC and PAX8/PPARG rearrangements in ATC 15 and the low turnover rate of thyroid follicular cells (about five renewals per lifetime) that reduces the possibility of accumulating the mutations needed for transformation. 8,16–18 The existence of several differentiation degrees has led to the assumption that TC cells are derived from remnants of fetal thyroid cells, such as stem cells (SCs) or precursors, rather than mature follicular cells. 9,19,20 According to this fetal cell carcinogen- esis model supported by gene expression profiling data, ATC arises from fetal thyroid SCs, marked by the onco-fetal fibronectin expression and lack of differentiation markers. Thyroblasts are hypothesised to be at the origin of PTC and are characterised by the concomitant expression of onco-fetal fibronectin, and the more differentiated marker thyroglobulin (Tg). Remnants of pro- thyrocytes, which represent a more differentiated cell type not expressing onco-fetal fibronectin, would result in FTC (Figure 3). 7,20 Genomic alterations, including mutations in TP53 and BRAF genes, as well as RET/PTC and PAX8/PPARG rearrange- ments, have an oncogenic role by conferring proliferative advantages and preventing fetal thyroid cells from differentiating. 1 Department of Surgical and Oncological Sciences, University of Palermo, Palermo, Italy; 2 Department of Surgical, Oncological and Gastroenterological Sciences, University of Padua, Padua, Italy and 3 Regina Elena National Cancer Institute, Rome, Italy. Correspondence: R De Maria, Regina Elena National Cancer Institute, Rome, Italy or Professor G Stassi, Department of Surgical and Oncological Sciences, University of Palermo, Via del Vespro, 131, Palermo 90127, Italy. E-mail: demaria@ifo.it or giorgio.stassi@unipa.it 4 These authors contributed equally to this work Received 1 February 2015; revised 24 March 2015; accepted 30 March 2015 Oncogene (2015), 1 – 11 © 2015 Macmillan Publishers Limited All rights reserved 0950-9232/15 www.nature.com/onc