Epithelial Cell Origins: Stem Cells in Head and Neck SCC

The Chapter by Gonzalez-Moles in the present Volume provides a comprehensive review of this topic. Briefer coverage here places the concepts into perspective. In health, the human genome continues in its entirety from a fertilised ovum through multiple divisions and steps in differentiation and specialisation to form adult tissues of all types. It is the progressive shutting down of blocks of genes, whereby repressor proteins attach to silencer regions of particular genes, and the switching on or up-regulation of others, which lead to differentiation of tissue type. Once distinct tissues develop—and function, environmental messages can influence gene expression, and these may be heritable. Any healthy adult cell can theoretically recreate a whole tissue type, even a whole animal, given the right growth factors and environment. In practice embryonic stem cells, or those that remain multi-potent in the adult, such as can be derived from bone marrow or dental pulp, are easier to engineer. There is a vast market nowadays in stem cell treatments for tissue regeneration in many diseases, e.g. following stroke, macular degeneration. The field is advancing rapidly and seminal papers appear in many issues of the world’s leading scientific journals, e.g., at the time of writing, a series of outstanding papers in the December 2014 issue of Nature (Tonge et al. 2014). The progress of this work can be tracked through the user-friendly website of the international consortium “Project Grandiose” whose portal is at http://www.stemformatics.org.

The concept of cancer stem cells is, however, somewhat different. It has long been known that bone marrow contained a population of stem cells which, through a series of differentiation steps, gives rise to erythrocytes, lymphocytes or granulocytes. In epithelia, a subpopulation of cells have equivalent behaviour: This is a situation where a minority of cells, when they divide, duplicate their DNA template for a single daughter cell which moves away and the parent cell retains its anatomical position. The daughter cell then divides producing two third generation cells and each of these produce an expansion of the required cell numbers, for a genetically pre-determined number of divisions. There is a genetically predetermined switch to keratinocyte differentiation. From the pioneering work of Mackenzie and of Potten we know that, in skin, a single epithelial stem cell serves an anatomically defined EPU (Epithelial Proliferation Unit) (Mackenzie 1970; Mackenzie et al. 1981; Potten 1974). Similarly, the microvilli of the gut have a distinct organisation into compartments; in mucosal epithelia (Humphries and Wright 2008); these however, have proved harder to visualise.

Thus the rate of division of stem cells and of expansion compartment cells is critical in homeostasis. Equally critical is the accuracy of the process of DNA replication. Random errors can produce immortalisation in a clone of progeny. If cells still in the stem cell compartment or in the viable expansion compartment undergo a mutation or other alteration in one—more likely several—of the genes listed in Table 1.1, above, a malignant neoplasm can result. Theoretically this could be either from a failure of the genetic control of the switch to differentiation pathway, and/or a failure of control of the rate of division of stem cells. There are no reliable methods for making such a distinction but a very recent paper by Tomasetti and Vogelstein (2015) has attracted much interest in this regard. These authors argue that the risk of a malignancy in a particular tissue is proportional to

Table 1.1 Genes which are typically altered in HNSCC, and the functions affected

Tumour suppressors

p53, Rb, MTSI, RARb, p21, DOC-1R







Cancer susceptibility


DNA instability (MS1)

3p, 4q, 5q, 6p, 7q, 7p, 9p, 9q, 11q, 13q, 14q, 18q, 17p

Mismatch repair

hMLH-1, hMSH-2

Angiogenic factors


Heat shock proteins

HSP70, HSP47, HSP27

Proliferative markers

Ki67, PCNA, MYB1

Invasion metastasis


Drug resistance genes


the number of (stem) cell divisions in that tissue required throughout life for homeostasis: a stochastic model in which the chance of a random mutation or mutations being sufficient for neoplastic transformation increases with the number of times DNA has been replicated. The argument relates to stem cells, but could equally apply to expanding cell compartments in epithelial tissues. They test this concept by plotting the rate of (stem) cell division in human tissues against incidence of cancer in that tissue or organ and show a strong positive correlation of *0.81. They argue that about two thirds of cancers arise simply due to “bad luck”—due to such random errors, the remaining third being explained by environmental carcinogens. This is proffered as an explanation for the low incidence of malignancy in tissues such as brain, and for sarcomata in general compared to epithelial cancers, or for small bowel v colon. The epidemiological data come from the SEER database in the USA, so more work will be necessary to see if the theory is dented by the large differences around the globe in many cancers, notably tobacco-associated SCC.

This thinking has profound impact on public health policy making. Although it will vary from population to population, depending on the prevalence of established risk factors therein, primary prevention will have limited impact. For random cancers to be managed on a population basis, increased secondary prevention— screening for early detection of smaller, perhaps asymptomatic, more curable, lesions is needed.

The Cairns hypothesis, promulgated in a classic paper in 1975, and revisited with mathematical modelling in 2002, proposes that stem cells retain their original DNA template throughout life (Cairns 1975, 2002). In this case if a mutation arose during replication of a stem cell, this would be present only in a member of the expansion compartment, and would automatically die out. This was postulated as a mechanism for protection against cancer or other genetic diseases. Conversely, however, if a stem cell carries an oncogenic mutation, a neoplasm is inevitable. Successful treatment would necessitate elimination of affected stem cells and this has become something of a dogma in contemporary cancer research.

The phenomenon of “tumour progression” (see below) is more consistent with “sternness” being a fluid or changeable property of sub-populations of cells within a solid neoplasm, as in the stochastic model proposed by Antoniou et al. (2013) (Fig. 1.6).

< Prev   CONTENTS   Source   Next >