The PhD thesis

Human metastatic melanoma in vitro: ploidy, centrosomal integrity, serum dependency, chromosome 9p21 genetic status, tumour-suppressor gene expression.

The accepted version of the thesis as lodged with the University of Auckland library is available in PDF form suitable for on screen viewing or low resolution printing using the link to the right. Fair use for the purposes of personal study or review is permitted. Reproduction for non-commercial academic use is permitted providing it is correctly attributed. It should be cited as "Charters, Geoffrey A., Human Metastatic Melanoma In Vitro (3ed), Ph.D Thesis, The University of Auckland (2007)", or the equivalent.

Abstract

Thesis excerpts

In his PhD thesis Dr Charters included a number of sections that sought to synthesize, describe, and interpret the state of knowledge as it related to his research, and most were written to be able to stand independently from the thesis itself. A number of these are described below, with links to PDF files of the excerpted versions suitable for on screen viewing or low resolution printing. Similarly, an excerpted and adapted chapter describing one aspect of his experimentatal research is provided.

---

Cancer

"While great advances have been made against many human diseases, cancer still challenges us, and remains a major contributor to human misery. The last 50 years have seen many improvements in treatment and survival, but we are still no better at preventing or curing cancer than in 1950."

This essay is intended for a general readership with a curiosity into the biological basis of cancer and the implications this has for the search for therapies. It begins with descriptions of the defining characteristics of cancer: tissue hyperplasia and dysplasia, angiogenesis, local invasion, and metastasis. From these, the fundamental biological processes involved are identified from first principles. A substantial table with a large supporting bibliography summarises the known hereditary cancer predisposition syndromes, the genes and proteins implicated and their functions, where known, and establishes that the inferred biological processes are in fact very important. The traditional therapeutic modalities of surgery, chemotherapy, and radiotherapy are briefly discussed, as are their shortcomings, notably relatively poor selectivity and non-specific toxicity, and newer approaches that seek to target the key processes identified are introduced.

Melanocyte biology

"Melanin is a dark pigment that occurs in a vast variety of organisms and its production and distribution are the duties of a single class of cell: the melanocyte. While its absorbance at visible wavelengths makes it an important adjunct to communication between animals ranging from camouflage to bold display, absorbance at infrared wavelengths makes it an excellent basis for thermal regulation in poikilothermic animals. The extension of this to the near ultraviolet makes melanin an ideal sink for solar radiation that might otherwise be damaging. This role in photoprotection is underscored by the increased production of melanin in response to ultraviolet irradiation. To achieve this, the melanocyte must survive in an environment that would likely cause the self-destruction of other types of cell. Melanocytes are therefore particularly hardy, but this may come at a price. By enduring genetic damage in order to prevent other cells from sustaining it themselves, melanocytes are uniquely prone to the fruit of such damage, an increased risk of tumorigenic transformation, possibly culminating in melanoma."

This review touches briefly on melanocyte morphology and tissue context before embarking on a detailed description of the molecular biological basis for melanocyte function including the fundamental processes of differentiation, migration, proliferation, and apoptosis, and those peculiar to the melanocyte, most notably the response to genomic damage and melanogenesis. This discussion is accompanied by very detailed illustrations of the key signal transduction channels that link stimulus to response. These are exhaustively cross-referenced to the descriptive text dramatically facilitating the assimilation of the great amount of material presented there. On this basis, a review of the mechanisms underlying the development of melanoma is presented with an emphasis on results of the genetic analysis of tumours and tumour-derived cell lines. In one table spanning three pages, over 50 genes implicated in melanoma tumorigenesis are listed, together with brief notes of the significance of each and references to the primary literature. These are then discussed briefly under the functional headings of maintenance of genomic integrity, altered responsiveness, and regulation of proliferation, apoptosis, angiogenesis, and metastasis.

Immunodeficiency and cancer

"The theory of immune system surveillance as a general defence against cancer has many proponents, but little concrete evidence to support it. One of the telling arguments against this theory comes from the study of hereditary immunodeficiency syndromes. With very few exceptions, cancer is not a feature of these. Where it is, immunodeficiency and cancer predisposition are independent effects of a common molecular flaw, with no causal relationship existing between them. Similarly, arguments based on increased cancer incidence in those with suppressed immune system function do not stand up to close scrutiny. Consequently, suggested cancer therapies based on immune system modulation must be viewed with circumspection."

This brief review seeks to dispel the quite widely held belief that the immune system has a significant role in the prevention of cancer in general, and sounds a caution over the chances of success offered by therapeutic immune system modulation. Clues are sought in the cancer spectrum seen in those with hereditary, acquired, or therapeutically induced immunodeficiencies. While there is good evidence for an immune system role in prevention of cancers caused by pathogenic organisms such as viruses, there is very little to support a role in the general case. Theoretical considerations suggest that the development of cancer therapies based on the modulation of the immune response is inherently limited by the conflicting requirements of increasing sensitivity while avoiding autoimmune responses. A strong argument is made that for hereditary cancer, both prevention by, and successful therapeutic manipulation of the immune system are fundamentally impossible.

The pRB subsystem

"Study of a rare hereditary paediatric cancer has led to the identification of pRB, a tumour-suppressor implicated in human cancer of many types. It plays a crucial role in embryogenesis, differentiation, cellular senescence, and proliferation. The manifold functions of pRB are mediated solely via interactions with over 100 proteins, both individually and in higher-order complexes. Its functions are modulated chiefly post-translationally, with regulated alterations in phosphorylation state being the best understood mechanism. Not surprisingly, many of the elements necessary for regulation of pRB function have themselves been implicated in tumour suppression or tumorigenesis, in particular, the cyclins, the CDKs, and the CKIs.
"This article provides a general review of pRB structure, interaction, and regulation as a basis for a discussion of the mechanism by which pRB exerts control over cell-cycle progression. The relevance that this may have to tumorigenesis in general, and to melanoma in particular, is then addressed."

This is a substantial review of the literature pertaining to the retinoblastoma-associated tumour-suppressor, pRB, and its functional associates, and cites over 400 references.

After a very brief description of the disease, the "two-hit" hypothesis of Knudson and the mapping, identification, genetic structure and transcriptional regulation of the gene responsible, RB1, are described, and the biological significance of pRB is demonstrated through consideration of the results of knock-out mouse models. Evolutionary conservation, inter- and intra-species homology, tissue expression profile, sub-cellular disposition and turnover of pRB are each discussed briefly.

The subject of pRB function is then addressed more fully, commencing with a description of the best-known binding domain present in many of its partners, the LXCXE motif. The status of the related LXSXE motif is then explored, with the conclusion that it may have a larger role to play than has been recognised. A suggestion is then made that the overlapping DLXX(X)E motif may also be significant, and could account for the binding of proteins that lack the LXCXE motif, for example MDM2. This is a novel observation, previously undescribed in the literature. The known regions of pRB implicated in protein binding are then described: N-terminal domains, the 'A' domain and 'B' pocket, the large A/B, and C pockets, and the C-terminal region. A table of 33 pRB-interacting proteins that have been implicated in tumorigenesis is presented that lists the binding domains involved for each partner, where known or suspected, and from these a theoretical competitive binding matrix is derived.

Phosphorylation of pRB as a means of post-translational modulation is introduced and an overview of the principal kinases and phosphatases given. A comprehensive table lists the potential pRB phosphorylation sites, their targeting kinases and phosphatases, the timing of the transitions, and the functional consequences thereof. This part of the discussion culminates in the presentation of an elegant model for the cyclical binding and release of pRB regulated proteins that also neatly resolves the otherwise paradoxical observation that while PP1 is inhibited by pRB binding, in vitro it can dephosphorylate pRB.

A minimal proof of phosphorylation-dependent regulation of cellular proliferation by pRB is given outlining the evidence for cyclin-dependent kinase phosphorylation of pRB causing the release and disinhibition of E2F-1, and consequent production of proteins critical for the G₁-S cell cycle phase transition. This is followed by a detailed step-by-step exposition of this regulation in an idealised cell, from G₁ arrest to release from inhibition and CDK4 activation, to regulation of small A/B pocket interactions, to downstream transcriptional activation, to passage through the "restriction point", to the maintenance of pRB phosphorylation by CDK2 in conjunction with Cyclin A, and the eventual reign of the phosphatases that restores pRB to its inhibitory state. Variations of this basic cycle are then considered including continuous cycling, the action of inhibitory cytokines, cellular senescence, and viral infection. This theoretical model predicts that pRB phosphorylation occurs in five stages, consistent with data from electrophoretic mobility studies.

Attention then turns to the role of the pRB subsystem in cancer, drawing together observations from knock-out models, and genetic analyses of hereditary and sporadic tumours for the members of the pRB family, the D cyclins, the cyclin-dependent kinases, the p16 and p27-related CDK inhibitors and the E2F transcription factors. The involvement of this subsystem in melanoma tumorigenesis in particular is explored further, with consideration of the genes associated with hereditary melanoma predisposition and sporadic melanoma leading to the identification of RB1, CDKN2A, and CDK4 as being highly significant in melanoma. The observation is made that the absence of evidence to implicate CDK2, CDK6, or Cyclin E in melanoma development, despite this having been sought, may mean that deregulation of pRB phosphorylation alone is not sufficient to predispose toward melanoma, and that some facet of pRB function, modulated by CDK4 but no other kinase, may be critical, and surprisingly, this excludes E2F-1 as the functional target. Furthermore, the disparity in melanoma incidence in hereditary syndromes involving CDKN2A and CDK4 raises the suspicion that the alternative protein product of CDKN2A, ARF, may be a contributor to melanoma in its own right, although data are scanty on this.

Finally, there is a discussion of the hypothesis that the pRB subsystem as a whole may function as a melanoma tumour-suppressive mechanism, such that failure of any part predisposes toward tumorigenesis. This may be a good first approximation, but evidence that multiple defects occur in some cases suggests that pRB is not the only significant functional target.

Genome partitioning

"The critical process of spatially aligning replicated genomes during cell division is the province of the centrosome. Where this fails, the maintenance of stable ploidy is compromised, often with adverse consequences for the newly divided cells. Where they are viable, their genetic complement may be imbalanced and in consequence, their inherent activities and their sensitivity and responsiveness to external influences may be aberrant. If this leads to a dysregulation of proliferation, there can be dire consequences for the organism as a whole. The very frequent observation of centrosomal anomalies and ploidy changes in cancer attests to this."

This review explores the structure and function of the centrosome as these pertain to mediating genome partitioning during mitosis. It describes the regulation of the centrosome duplication cycle at the molecular level revealing that the key components are those that also control the nuclear cell division cycle. It suggests that the advent of the centrosome and p53 were essential in the evolution of large, multicellular, long-lived organisms with differentiated tissues, not just to bolster genomic stability, but also to prevent cancer, a disease only possible in such new species.

It begins with a brief description of centriole and centrosome structure and composition and then introduces functional aspects by drawing on the seminal work of Khodjakov et al. who used GFP-tagged centrosomes and laser microdissection to observe the effects of centrosome loss. Consideration of the theoretical consequences of incorrect centrosome number leads to the conclusion that very significant alterations in the genomic content of cells may result, consistent with those reported in cancer. The known molecular regulators of centrosomal function are listed in a summary table with brief comments on each and references to the primary literature.

The molecular biology underlying the centrosome cycle is described in depth from interphase, via centriole replication, centriole severance, and centriole separation, to post-mitotic relocation prior to cytokinesis. The stages are illustrated by highly detailed diagrams of the relevant molecular activities cross-referenced to the descriptive text. In this section, a novel theory is developed to account for the reported observation of just one of two seemingly equivalent centrosomes mediating the last step in cytokinesis, abscission. In this, the widely-held classification of centrosomes into two classes is replaced with a three class model which, in addition to accounting for the observed single centrosome peregrination, also provides for spontaneous correction of any flaw in this that may occur.

Upstream regulation of the centrosome cycle is tackled next, with emphasis on the roles of the cyclin-dependent kinases CDK2 and CDC2, and on their own regulatory systems: the cyclins, the CDK inhibitors, activating and inhibitory phosphorylations, and synthesis and proteolysis. The manner in which p53 influences these events in response to genomic damage is described and the suggestion made that the epithet "guardian of the centrosome" might well be applied to p53, in addition to its other accolades. Lastly in this section, the key roles played by CDK2 and CDC2 in synchronising the nuclear and centrosomal cell cycles is explored.

The review ends by proposing that while evolution has brought life to the point of complex differentiated multicellular organisms, introducing the centrosome to bolster mitotic fidelity and p53 to oversee the process, these systems are not perfect, and where they fail, the result may be cancer.

---

Centrosomal integrity

"The possibility that centrosomal numerical dysregulation may be contributing to unstable ploidy in the NZM cell-lines was investigated by an immunofluorescence microscopic study of mitotic cells in which α-tubulin, the centrosomal marker protein pericentrin, and DNA were differentially fluorescently labelled. Evidence of dysregulation was abundant, and aberrant mitoses were present in all cell-lines studied. Pericentrin accumulated in nucleoli, and its release from these reservoirs at prometaphase entry suggests this may be a means of initiating spindle formation. The results of this study prove that the loss of centrosome numerical regulation, deduced from the consideration of heteroploidy, actually exists, and further, that events of the types necessary to account for this heteroploidy are occurring in the NZM cell-lines. The possibility that pericentrin regulation is defective in melanoma has arisen, and if this is so, it could lead to the development of new therapies."

This document describes the study undertaken to test the hypothesis that centrosomal dysregulation might be responsible for the unstable heteroploidy seen in the NZM cell lines, particularly in NZM10. It begins with a brief statement of the rationale for performing the study and for the selection of multicolour immunofluorescence microscopy of monolayer cell cultures as the technique to be used. The cell and molecular biological and image processing methods are described briefly at this point, followed by a more substantial description of the basis for assessment of the observations made of mitotic cells with respect to mitotic phase, general cellular aspect, centrosome number, centrosome morphology, centrosome positioning, pericentrin distribution, mitotic symmetry, and genome segregation anomalies.

The findings are then presented and are accompanied by over 150 high quality fluorescence microscopy images. The discussion commences with the validation of the method through consideration of the included controls and an assessment of the technical limitations of the equipment used. It is reported that extended pericentrin structures were found, both centrosomally associated and independently, and the possibility that these could interfere with genome partitioning through their influence on mitotic spindle structure is discussed. While supernumerary centrosomes were frequently detected in mitotic cells, these were not always associated with aberrant mitoses, leading to the development of the concepts of cooperative and bystander centrosomes.

The main results of the survey of mitotic cells are then presented together with their statistical analysis. It is found that for most of the cell lines and most of the characteristics studied the NZM cells are significantly more poorly regulated than the tumour-associated fibroblasts that were used as control cultures, and a strong association between centrosomal dysregulation and flawed genome partitioning exists. Particular examples of the types of abnormalities detected are then presented and discussed, including mitotic cells with fewer than two centrosomes, those with three, with four, or with a greater number.

Additional observations made during the mitotic survey, but incidental to it, are then reported, including findings of potentially great significance. Chief among these were two that related to the spatial and temporal distribution of pericentrin during mitosis. Unexpectedly, it was found that pericentrin accumulated in nucleolar reservoirs during interphase and was released from these to flood the cytoplasm with nuclear envelope breakdown at the beginning of prometaphase. It is hypothesised that this may be a means of synchronising mitotic spindle construction with the availability of condensed chromosomes suitable for incorporation into a metaphase plate. Evidence for the formation of intranuclear tubulin nests, tubulin voids, and asynchronous mitoses and apoptoses in individual nuclei of multinuclear cells is presented, together with an initial study into anomalous nucleic acid bodies detected in some cell lines.

These findings are then placed into the context of the known literature and their significance assessed. The overall conclusion is that the hypothesis that centrosomal dysregulation was occurring and was a likely cause of unstable heteroploidy should be considered proven.

Finally, the brief description of the methodology given at the beginning of the chapter is supplemented by highly detailed procedural descriptions together with information concerning reagents and their preparation, after which the reference list is given.