Geoffrey A. Charters MSc (Hons) PhD (Pathology)
Molecular oncopathologist
Geoffrey A. Charters MSc (Hons) PhD (Pathology)
Molecular oncopathologist
Biography
For two human metastatic melanoma cell lines, NZM4 and NZM7, having different responses to irradiation, to:
Initially, Geoffrey approached the Head of the School of Biological Sciences, Prof. Dick Bellamy, who was very keen to be helpful, but not entirely enthusiastic about the chances of a long-lost alumnus returning to the fold. He explained that Geoffrey had basically missed the revolution in molecular biology that had occurred, and would find it very difficult indeed to resume his studies after such a long absence.
Prof. Bellamy was exceedingly generous with his time and led Geoffrey on a tour of the research facilities of the School. In the various laboratories he enquired of Masters students present if they felt that their recent undergraduate work had been essential to their ability to undertake their current research. He was a little surprised by the answer he consistently received, that it had not!
Ultimately, Prof. Bellamy conceded that since Geoffrey technically had the necessary prerequisites, from Auckland University no less, if he could find a willing supervisor there was nothing at all to prevent him from enrolling, and wished him all the best.
This was not a trivial exercise since, as it was late in the year, most intending postgraduate students had already found supervisors and few vacancies remained. Supervision had to be by a member of the School of Biological Sciences, and the chances appeared slim for finding a project in the field of molecular genetics, particularly as it related to cancer. Geoffrey was even contemplating proposing a project involving the molecular taxonomy of New Zealand orchids.
Almost as a last resort, he approached Prof. Bruce Baguley, Director of the Cancer Research Laboratory of the Auckland Cancer Society, academically affiliated with the University through the School of Medicine. This might have seemed like an obvious starting point, but since it was not part of the School of Biological Sciences, it had not appeared to be a viable option. Prof. Baguley was most welcoming and sympathetic, and agreed to act as supervisor if some way around the regulations could be found. This was achieved when Dr Don Love agreed to be Geoffrey's nominal supervisor within the School of Biological Sciences, with Prof. Baguley being designated co-supervisor.
The three papers Geoffrey selected were Advanced Cellular and Molecular Biology for Biomedical Research, and Molecular Genetics, both taught in the School of Biological Sciences, and Cancer Biology taught within the Department of Pathology at the School of Medicine.
The extent of progress in the field was astounding to Geoffrey, and he was surprised by the sea-change that had occurred in the approach being taken to analytical biology. A decade earlier, the prevalent analytical modality was deductive, a "bottom-up" approach, with research driven by seeking to understand in detail the mechanics of isolated enzymes, and from that, to piece together so called metabolic pathways, often depicted in huge, complex charts with detail at the atomic level: the Krebs cycle, glycolysis, photosynthesis, lipid metabolism, etc. By the mid-1990s this had been turned upside-down, and a modular, reductive, top-down approach was in vogue. Cellular processes were being compartmentalised into functional subsystems each having significant internal structure, the fruits of the earlier deductive investigations, but with relatively few points of interaction among them. This was the "revolution in molecular biology" to which Prof. Bellamy had referred, and it was indeed profound.
Rather than being discomfited by it, Geoffrey welcomed it. Immediately obvious to him was the analogy between this new view of biology and the structure of the computer systems in which he had been immersed. Each was based on the concurrent execution of multiple inter-communicating processes, each responsive to the state of others and to environmental cues, and cooperating to produce desirable results. Failure of certain of the biological subsystems underlay the origins of cancer, and if modern molecular biology was the reverse-engineering of biological function, then the quest to understand and defeat cancer was biological debugging. The same analytical skills that Geoffrey had honed in the realm of computing for a decade would be equally applicable here; it was just a matter of learning the language and procuring the necessary tools and skills.
One cellular mechanism in particular captured Geoffrey's interest, and that was the regulation of the cell division cycle. It demonstrated remarkable structure, with multiple complex sub-processes being initiated under the control of a small set of regulatory molecules, themselves under the control of a supervisory process that ensured their activation only when all preceding steps had been satisfactorily completed. Additional asynchronous processes were in operation that implemented quality control, and the whole was sensitive both to the internal conditions of the cell, and its external context. Furthermore, it was a cycle: what was done in an orderly fashion was then reset to its initial state with very high precision, time and time again. This process was at the very heart of life, occurring at the molecular level, automatically, and with near perfection, and that it could actually be explored and understood was an epiphany to Geoffrey. Studying such an exquisite and fascinating molecular machine would be a reward in itself, but most importantly for him, understanding the ways it which it failed would provide fundamental insights into the origins of cancer, and perhaps from there, lead to the development of preventative measures or therapies.
In tumour cells the ability to constrain proliferation in response to internal or external stimuli is frequently defective, and several distinct mechanisms are involved. In response to the inhibitory growth factor TGF-β epithelial cells typically arrest in the G1 phase of the cell cycle. This is principally through the induction of the CDK inhibitor p15CDKN2B via SMAD signalling, but other mechanisms must exist as insensitivity to TGF-β is seen in p53 mutants. A possible explanation for this was offered by the report that p53 binds CDK4 mRNA and prevents its translation. The reduction in CDK4 protein that may result could prevent effective pRB phosphorylation allowing it to block cell cycle progression in G1.
Other types of cell cycle arrest are mediated via p53, the best known being through the induction of the CDK inhibitor p21, as occurs in response to radiation damage. However, studies of p21-null cells had shown the existence of a p53-dependent, but p21-independent mechanism of arrest after irradiation. The question therefore arose as to whether the reported p53-dependent blockage of CDK4 translation seen in response to TGF-β might also operate in response to irradiation, and so provide a basis for the p21-independent mechanism.
The principal hypothesis of the project was that it did.
To address this, the levels of CDK4 present at various points in the cell cycle were to be measured using two-dimensional flow cytometry of cells from control cultures and those subjected to irradiation. In order that a contrast was more likely to be found, two melanoma cell lines were chosen that differed in their responsiveness to TGF-β and irradiation: NZM4, inhibited by TGF-β and thought to have normal p53 as it arrests in G1 in response to irradiation; and NZM7, insensitive to TGF-β and thought to be a p53 mutant as it exhibits a diminished radiation response. The rationale was that if a CDK4 reduction were a normal part of the response to both irradiation and TGF-β inhibition, it should be detectable in NZM4 where both sensitivities were intact, but absent from NZM7 where both were abnormal.
Before embarking on the test of this main hypothesis, preliminary work would be carried out in order to gain familiarity with the necessary techniques and with the characteristics of the cell lines to be used. This would comprise an investigation of cell cycle phase changes during growth to high density and upon treatment with TGF-β.
During preliminary work NZM7 was found to be heteroploid, that is, to contain multiple subpopulations with differing ploidy. This was recognised as a potential problem for the accurate estimation of cell cycle phasing through mathematical model fitting. As a result, uniploid strains, or subclones, of the NZM7 cell line were isolated by single cell propagation for use in further experiments, and ten were eventually established. During culture it became evident from clearly visible differences in melanin production that there was phenotypic variation among them, and so a partial characterisation was performed on them in terms of ploidy, melanin production, TGF-β sensitivity and growth rate.
This gave rise to two additional issues. The first was the finding that significant differences in growth rates occurred depending upon the position of a culture within a 96 well plate, and the second that the value derived for the widely used measurement of inhibitory potency, the IC50, was highly dependent upon conditions of culture. This led to the investigation and refinement of both of these practices.
One obstacle developed that resulted in the nature of the project changing substantially. With the preliminary work completed, and the unexpected issues of heteroploidy, NZM7 subclone characterisation and assay basis addressed, the focus moved to CDK4. It was soon found that the anticipated reduction in CDK4 in response to TGF-β did not actually occur in the sensitive NZM4 cell line, or for that matter in NZM7, and the mechanism it had been intended to explore was not operational in the experimental system to be used. Although such a negative finding did constitute a significant result, it would not have satisfied the MSc requirements. Unfortunately, the developmental work needed to arrive at this result had consumed a substantial amount of the strictly limited time permitted for research.
In light of this, a decision was made to continue to explore CDK4 protein expression levels under different conditions that might result in changes, including irradiation, as intended, treatment with etoposide that should engender DNA double-strand breakage, and with the radiomimetic drug bleomycin, which generates intracellular free radicals. The work with TGF-β would be extended to explore the possibilities of production by the NZM cell lines, and that the insensitivity of NZM7 was due to genomic methylation. These results, together with the preliminary findings concerning growth kinetics, heteroploidy, and subclone characterisation would then be presented as a general study. This was far from ideal, but the best alternative available in the circumstances. As the work progressed, a number of other worthy observations were made incidentally.
Having identified NZM7 as being heteroploid from flow cytometric studies and isolated uniploid strains by single cell propagation, it was found that there was considerable phenotypic variation among these. This manifested most obviously in the production of melanin, with NZM7.2 producing copious quantities, while NZM7.4 appeared to produce little if any. Morphologically, all were spindles or lenticular, but in some cases a noticeable number of cells possessed dendritic-like processes. While all had similar doubling times, some lacked thrift in culture as evidenced by an accumulation of sub-cellular debris and flow cytometric data consistent with the occurrence of apoptosis. While the parental NZM7 cell line was considered immune to the inhibitory influences of TGF-β, there was evidence of partial sensitivity for the NZM7.1 and NZM7.4 strains.
NZM4 and NZM7 cultures grown to high density each reached plateaux in cell number, but this was approximately three times as high in NZM7. Flow cytometry revealed that NZM4 entered a stable quiescent state with cells arrested in G1 while in NZM7, population appeared to be stabilised chiefly by cell death. Replacement of supernatant medium stimulated renewed growth in both cell lines establishing that the arrest that had occurred was not solely due to cell density per se, and implicated a soluble factor, either through production of an autocrine inhibitor, or depletion of an essential stimulant or nutrient. TGF-β1 was excluded as the operative agent through the use of a neutralising antibody.
In assessing the sensitivity of the cell lines to TGF-β IC50 values were determined. Since the inherent inhibitory effects of increasing cell density might have been a confounding factor, a series of experiments was conducted with cultures having a range of seeding densities and over a range of incubation periods. It was found that the IC50 derived was highly dependent upon these cultural conditions. On consideration, this should not have been a surprise since where high-affinity binding of a ligand to a receptor is involved, it is not the ambient concentration of the ligand that is critical, but the ratio of ligand concentration to total receptor concentration, and this latter factor is dependent upon the number of cells present. This meant that the common statement of IC50 in terms of concentration alone was inappropriate. Recasting the data in terms of ligand:receptor ratio led to a result consistent over four orders of magnitude of ligand concentration, and independent of seeding density or time in culture.
Inconsistencies in proliferation rates in control cultures were found when 96-well plates were employed, and a search was undertaken to identify the source of this variation. Stacking of plates, and the position of a plate in such a stack, the enclosing of plates in plastic sheaths, position in the incubator, the use of an insulator under the plate, and position of each culture within each plate were considered. It was found that the rate of proliferation was significantly lower in wells at the edge of plates, and the use of these wells was discontinued where consistent growth rate was an important aspect of experimentation.
The sensitivity of NZM4 and the resistance of NZM7 to TGF-β were confirmed. By using NZM4 as an indicator, it was found that both cell lines produced a soluble factor that inhibited the growth of the TGF-β sensitive NZM4 cells when it was either acid or heat treated, but not otherwise, consistent with this factor itself being latent TGF-β. Genomic demethylation did not restore TGF-β sensitivity to NZM7, showing that transcriptional silencing of CDKN2B was unlikely to be the cause of its insensitivity. Early suggestions that low levels of TGF-β may stimulate proliferation in NZM4 were not borne out by closer examination.
It was shown that both NZM4 and NZM7 retained some ability to arrest in G1 or G2 in response to treatments including irradiation, etoposide, bleomycin, and TGF-β but in all cases the response of NZM7 was less robust, and for TGF-β, absent.
Flow cytometric determination of nuclear CDK4 expression with respect to cell cycle phase after various treatments revealed no reduction in G1 CDK4 levels in either cell line in response to radiation, etoposide, or bleomycin treatment, with one exception. In NZM7, reduced levels were seen in response to irradiation, an apparently perverse result given that the cells did not arrest in G1 and the cell line was thought to be a p53 mutant in which this effect was expected to be absent.
Two interesting incidental results were obtained. Nuclei from cells in G0 or G1 were found to contain a range of levels of CDK4, but those in S phase all contained near the maximum, with the level dropping again in G2 or prophase (later phases not being present as only intact nuclei were assayed). This was taken as evidence for a threshold level of CDK4 being required for entry to S phase, an unexpected finding given that the accepted model had it that it was the availability of the D-cyclin partner of the complex, not the CDK that was limiting for S-phase entry. A corollary of this was that the reduction of nuclear CDK4 below the threshold required for S-phase entry would guarantee the occurrence of a G1 phase following mitosis, and so prevents rapid cycling, even in the presence of adequate cyclin-D.
Secondly, when selective gating was used to examine the data for just those G2 or M phase cells with reduced CDK4, it was found from consideration of their forward scatter characteristics that they were the smallest nuclei present. This suggested that the reduction in CDK4 seen may actually be a requirement for chromatin condensation, or that these processes were under joint control.
The resulting thesis was not the focussed work testing a single hypothesis and reaching a robust conclusion that had been anticipated at the outset, but the MSc requirements are broadly defined, and greater emphasis is placed on the work undertaken having been soundly based, reasonably well executed, and adequately reported, rather than on its significance, focus, and novelty. In conjunction with the grades he had obtained in his course work, the thesis was sufficiently well-received to qualify Geoffrey to graduate with Honours (second class, first division).
Although this was a commendable result, Geoffrey still felt that a key question for him remained unanswered, that of whether he was truly capable of making a significant contribution in the field. To continue seeking an answer to this, now with an Honours degree, the next reasonable step was to embark on study for a PhD. In this, Prof. Baguley was again extraordinarily helpful, offering his supervision, arranging for Geoffrey to receive a stipend from the Cancer Society, and providing continuing access to the CRL laboratories, which were about to become the Auckland Cancer Society Research Centre.