*This work was supported by a grant from Biomedic e.V. and by IKFE services.

Correspondence and request reprints to:
Professor Dr Michael Derwahl,
Department of Medicine
St. Hedwig Hospital
Grosse Hamburger Str. 5-11
D-10115 Berlin – Germany
Tel: 0049-30-2311-2503
Fax: 0049-30-2311-2324
E-mail: m.derwahl@alexius.de

DISCLOSURE STATEMENT: The authors have nothing to disclose

INTRODUCTION
The occurrence of nodules by age is a hallmark of the human thyroid gland and some other glands (1). Indeed 20 to more than 50% of the normal population develop one or more thyroid nodules during life-time (2). The majority of these nodules are non-functioning (more than 80%) but only 5% of these scintigraphically “cold” lesions are malignant tumours (3, 4). In 20% of these cases, a follicular neoplasia is cytologically diagnosed by fine needle biopsy, which makes a subsequent surgery obligatory (5). But only in a few percent of these patients histological diagnosis confirms malignancy. Unfortunately, despite many efforts, up to now there is no pre-operative approach to clearly differentiate between benign and malignant tumours in these patients. One reason for this is our limited knowledge of the molecular aetiology of benign non-functioning nodules. Although many pathogenetic factors such as iodine deficiency, mutagenesis, overexpression of growth factors and their related receptors, altered signalling, recent data on gene expression profiles and a genetic predisposition are known a comprehensive concept for the pathogenesis of thyroid nodules and nodular goitres is still missing (6, 7).
Histologically, thyroid nodules are classified as adenoma (with a characteristic capsule), as adenomatous or hyperplastic nodule or lesion (8). However, independent of their histology, molecular biology analysis has revealed that a majority of these nodules are clonal, i.e. derived from a single cell or a naturally occurring clonal cell patch, and thus meet the criteria usually applied to characterize true benign neoplasias (reviewed in 6, 1).
Classical theory considers a differentiated follicular thyroid cell, altered by a sequence of molecular aberrations achieved during cell cycling, as the primary source of thyroid tumourigenesis (9). With this respect the high frequency of tumour formation in the thyroid gland seems to be surprising, since, compared to highly proliferating tissues such as the colon, the growth rate of human thyroid cells is rather low. It has been estimated that human thyrocytes divide only about five times during adulthood which corresponds to a turnover time of about 8.5 years for the follicular thyroid cell (10). Furthermore, tissues with high cell turnover such as the colon are more sensitive to mutagenesis and other molecular mechanisms that initiate tumour formation, whereas in resting tissues such as the thyroid these mechanisms should be less operative.

Adult stem and progenitor cell in thyroid tissue
Stem cells are classified as embryonic or as adult stem cells. An embryonic stem cell is derived from inner cell mass, which is part of the early (d 4 to 5) embryo called the blastocyst, whereas an adult stem cell is an undifferentiated, quiescent or slow cycling cell that occurs in a differentiated tissue (11, 12). Adult stem cells are capable of making identical copies of themselves throughout the organism’s lifetime, which is referred to as “self-renewal”. By “asymmetric” cell division they generate one self-copy and one precursor or progenitor cell, a partly differentiated cell that further divides and gives rise to differentiated cells. Adult stem cells have been detected in different tissues such as colon, skin, pancreas, liver and brain (13-17). Recently, we have identified stem cells in the human thyroid gland (18).
Due to their pluripotency and undifferentiated state stem cells are widely believed to be involved in the pathogenesis of tumours since these cells share many properties with cancer cells such as self-renewal and indefinite growth (11, 19, 20). The hypothesis of thyroid cancer as a stem cell disease has also been proposed by several authors (21, 22). Proof for this assumption comes from the very recent finding of cancer stem cells in some thyroid carcinoma cell lines by us and by other (23, 24). In colorectal cancer, the concept of a stem cell disease has also been extended to the benign tumour precursors, the low-grade adenoma (25). In this benign neoplasia, stemness- and proliferation-associated genes are already activated.

Thyroid nodules, stem cells and their niches
Very recently, evidence has been provided for a stem cell hierarchy in the adult human breast (26). In contrast to the essentially quiescent stem cells, progenitor cells were found to be more actively dividing. If such a hierarchy also exists in the thyroid gland, one may speculate that (analogous to colorectal adenoma) such progenitor cells are the source for the generation of thyroid nodules.
Only about 1-2 of 105 cells of nodular goitres are stem or progenitor cells (18, own unpublished data). Surprisingly, a relative resistance to growth stimulation was observed by us and by others (in mouse thyroid glands) when stem and progenitor cells were maintained in mono-culture (27, 28). Such observations that were also made with other stem cells may be explained by the interaction between stem cells and micro-environmental cells (niche cells) (29). Niche cells provide a sheltering environment that protects stem cells from uncontrolled differentiation stimuli, apoptotic stimuli and on the other hand from excessive proliferation that could lead to cancer (29). The lack of these regulatory cells may explain why adult stem cells when isolated as single cells do not proliferate despite intense growth stimulation.
However, which mechanism renders a quiescent stem or progenitor cell to proliferate and grow out to form a tumour? In vitro, induction of apoptosis reduces the strict control of niche cells (30). Under these conditions, stem cells in thyroid cell cultures gradually escaped niche control and grew out to form three-dimensional spheres, designated thyro-spheres, when stimulated with EGF and basic FGF (27). These stem cell-derived spheres were composed of 5% stem cells and 95% progenitor cells which may indirectly prove the much higher cycling activity of the progenitors as mentioned above. When growth stimulation of fast-cycling progenitor cells was terminated and TSH-enriched medium was added, proliferation rate of progenitor cells slowed down and the differentiation process was initiated.

Thyroid nodules and nodular goitres as a stem cell disease
Can the above described experiment serve as a model for the role of some fast-cycling progenitor cells in the pathogenesis of thyroid nodules? Are there any arguments that support the concept of thyroid nodules and nodular goitres as a stem cell disease?

There is some evidence from different studies that

1. stem cells reside in thyroid tissue (as in all other tissues) for life-time of the organism (29).
2. stem cells and their progeny are under the control of niches that limit proliferation of these undifferentiated cells (29).
3. induction of apoptosis and (excessive) growth stimulation can overcome strict niche control (30).
4. under these conditions actively cycling, more or less differentiated progenitor cells (some of them so far without a (fully) developed iodine metabolism) may grow faster than the surrounding differentiated thyrocytes.

Acknowledgement
The author thanks Dr Eva Nathues for support in the preparation of the manuscript and Professor Studer (Berne) for his valuable comments.