James IV Article

Dissecting the molecular mechanisms of human cancer: Translating laboratory advances into clinical practice

A.M. Thompson
Department of Surgery and Molecular Oncology, Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 9SY

Correspondence to: A.M., Thompson, Professor of Surgical Oncology, Department of Surgery and Molecular Oncology, Ninewells Hospital and Medical School, University of Dundee, Dundee.
Email:a.m.thompson@dundee.ac.uk

Introduction Mechanisms Which methods

Acknowledgements References

There are multiple molecular mechanisms involved in human cancers with many genes involved, complex interactions and a variety of ways to examine them. Some events are already emerging as clinically important; others will turn out to be bystanders. By focussing on key genes, such as p53, the clinical implications of genetic changes and the pathways they link into are becoming apparent. Using the complex methodologies now available allied to disciplines such as mathematics we are improving our understanding of malignancy. This is beginning to impact on the management of patients with cancer; meanwhile, a good surgical operation performed timeously is still the best chance a patient has of a cure for most types of solid cancers

INTRODUCTION
While surgery remains the mainstay of anticancer treatment for most solid cancers, ultimately understanding the mechanisms involved in cancer cell behaviour should lead to improvements in the prevention, early detection, treatment and prognosis of human malignancy. Human cancers develop through interactions between genetic abnormalities at the cellular level with internal and external environmental influences. There are several types of genetic events which may occur during carcinogenesis, some of which are more common in one cancer type than another, and usually multiple events are required before a normal cell becomes malignant. These events can be investigated at the cellular and subcellular levels of deoxyribonucleic acid (DNA), messenger ribonucleic acid (mRNA) or protein in vitro, on tissue sections or on extracts from cancer tissue.

This overview aims to provide a window onto some of the laboratory techniques and advances using breast cancer, oesophageal and gastric cancer as examples. These techniques may contribute to improvements in the clinical care of patients with cancer in future years.

MECHANISMS
There are many mechanisms which may contribute to human carcinogenesis. These may be divided into those affecting the complex components making up the genetic composition of a cell and specific gene events. Examples of the former including telomere shortening and DNA methylation.^1-3 Near the tip (telomere) of each chromosome, repeated runs of thymidine and guanosine nucleic acids act as stabilising influences, preventing the chromosome arm undergoing events including loss of genetic material which may contain key cancer preventing genes. These telomeres are maintained by telomerase enzymes but as part of ageing and during the development of a cancer, telomeres shorten allowing genetic events to occur.^1,2,4,5 One of the many other mechanisms influencing which genes function is methylation of regions of deoxyribo nucleic acid (DNA), thus silencing the expression and function of one or several genes.^3,6 Much of the investigation of human cancers has been targeted at specific genetic events. For example, examining oncogene overactivity, tumour suppressor gene (p53) loss which allows a cancer to develop, altered signal transduction where the signalling systems within a cell become deranged, and growth factor/growth factor receptor changes where the interaction between growth factors and cellular receptors function aberrantly.^7-12 In addition, there may be overlap and/or interaction between these classes of genes.

The literature for each of these fields is substantial, and beyond the scope of a brief review; however, by considering methods by which genetic events are investigated, illustrative examples can point to some current advances.

WHICH METHODS?
Within every living cell in the body, the genetic code is stored on DNA, the integrity of which is maintained by mechanisms such as the telomeres.^1,2,4

The information encoded in the DNA sequences has to be transcribed in the cell nucleus into messenger ribonucleic acid (mRNA), pass from the nucleus into the cytoplasm where the code held within the mRNA is translated into protein. Subsequently, the protein may be modified in a variety of ways, return to the nucleus (to interact with DNA), lie in the cytoplasm (as a structural protein or as a secondary messenger), integrate into the cell surface (as a receptor) or be excreted or secreted from the cell (for example as a growth factor or hormone).

Investigating how cancer cells function and the differences from normal cells can thus be pursued at several levels (Table 1)

COMPARATIVE GENOMIC HYBRIDISATION (CGH)
Comparative genomic hybridisation (CGH) is a comparison of test (cancer) DNA to reference (normal) DNA allowing the detection of abnormalities in the test (cancer) DNA. Like most molecular techniques the outline description does not do justice to the complexities of the techniques involved. However, essentially the test and reference DNAs are labelled with different coloured fluorochromes and competitively hybridised to normal metaphase chromosomes. The relative abundance of test versus reference DNA from specific loci on individual chromosomes can be deduced from the ratio of fluorochromes. This gives images (Figure 1) which can be analysed with appropriate software to reveal a pattern of losses (which may represent tumour suppressor gene loss) and gains (which may represent oncogene amplification) in the tumour DNA.

Figure 1: CGH of an oesophageal cell line (OE 33) demonstrating (a) the metaphase chromosome spread with amplified tumour DNA showing as green spots (highlighted by white arrow heads) (upper left), (b) the chromosomes aligned with amplification (green) and loss (red) highlighted at several loci and (upper right) (c) the computer analysis whereby the location of gains (green peaks eg on chromosome 17) and losses (red troughs on chromosome 17) in relation to normal DNA are documented for the cancer DNA (lower panel).

Using this technique, it has become evident that proximal (cardia) cancers have a different CGH profile to distal gastric cancer, in keeping with the differing epidemiological profile of the cancers.^13,14 In breast cancer, tumours with amplification on the short arm of chromosome 16 (16p) and deletions on the long arm of chromosome 16 (16q) have a better prognosis after several years follow-up than cancers with chromosome 17p deletions and 17q amplifications.¹⁵ These findings point to regions of DNA which require further investigation to localise and identify individual genes involved in cancer. One such gene on chromosome 17p, termed p53 since the protein appears as a band on protein gels at the 53 kilo Dalton position, has been extensively investigated.

p53
p53 has pleiotrophic functions but plays a key role in integrating cellular responses to damaging agents including chemotherapy and radiotherapy.^10,16 As a consequence of p53 activation, cells may be forced to undergo one of several options including apoptosis (programmed cell death) or cell cycle arrest.^9,10

The involvement of p53 in breast cancer can be assessed in a variety of ways.^17-19 One method, Southern blotting, is to extract DNA from a tumour sample and the patient’s normal tissues (venous blood lymphocyte DNA is a convenient source). Following digestion of the DNA by an enzyme which cuts the DNA at specific points, running the digested DNA on a gel, blotting the digested DNA onto a nitrocellulose membrane and probing the DNA with a radiolabelled probe one can look for loss of genetic material (loss of heterozygosity (LOH) or allele loss) in the cancer sample (Figure 2a). Thus, using a probe that identifies DNA close to the p53 gene, one finds that 62% of breast cancers have evidence of loss of genetic material at that site and, by implication, potential loss of the protective functions of p53.

Similarly, RNA can be extracted from a cancer and p53 mRNA expression examined by Northern blotting (Figure 2b). Using these techniques, 17p13.3 LOH and p53 mRNA expression have been significantly associated with poor prognostic markers and with reduced survival in breast cancer, implicating p53 as a key gene in this as in most other human malignancies.^{17, 20-23} 

Figure 2a: Southern blot demonstrating two bands or alleles in the control (normal) DNA from a patient (B) but loss of one band (or allele) in the tumour sample (T). Figure 2b: Northern blot demonstrating p53 mRNA expression in RNA extracted from a cancer and also from a breast cancer cell line (lower panel) with the control for loading (actin mRNA) in the upper panel.

However, these older blotting techniques required considerable amounts of tissue and with the advent of the polymerase chain reaction (PCR) which multiplies faithfully the RNA or DNA present in a small sample to result in larger quantities of copies of the original nucleic acid sequence, similar data can be obtained on minute quantities of material. Even so, these methods imply a role for p53 as an important gene in human cancer (confirming the wealth of in vitro data), but what functional effect p53 actually has in a particular cancer is less certain.

More recently, a functional yeast assay has been developed as a means to assess whether the p53 in a given cancer is functionally abnormal and, thus, unable to act as a protective tumour suppressor gene.24 Furthermore, the presence of normal (or wild type) p53 in a cancer may indicate whether the cancer is susceptible to chemotherapy although this remains controversial.18,19 Briefly, the functional yeast assay involves taking RNA from fresh (or specially preserved) tumour tissue, amplifying copies of the tumour p53 mRNA using the PCR technique, then inserting the amplified product into modified yeast. When grown on selective media, these yeast activate a reporter system which means that if the original p53 from the cancer was normal (wild type) the colonies grow white, if the p53 was exclusively mutant, the colonies grow red. If there was a mixture of normal and cancer tissue or more than one lineage of cancer cells, a mixture of red and white colonies appear (Figure 3).

Figure 3: Three yeast plates showing gastric cancer cell line AGS (wild type p53; white colonies; upper left), oesophageal cancer cell line OE33 (mutant p53; red colonies; lower left) and an oesophageal adenocarcinoma (mixture of red and white colonies; right side; approximately 50% of the sample was cancer tissue).

Figure 4: p53 sequencing. Sequencing of 3 red colonies from a yeast plate (colonies 1,2,3). A consistent mutation from G to A is identified in comparison with the (wild type) p53 sequence. This represents a mutation in codon 273 of p53, an important functional part of the p53 gene (see figure 5) responsible for binding p53 to DNA to allow the p53 to switch on a range of target genes.

To understand what effect the particular mutation detected in a cancer may have, the p53 within the colonies can be selectively amplified and sequenced (Figure 4). Thus, for a cancer where the sequence shows a mutation at position 273 of the p53 gene, as in the case illustrated (Figure 4), the resultant mutant p53 protein may be unable to bind to DNA as it normally would and, thus, is unable to unleash the protective functions which should prevent the development of a cancer.

In practice, this functional yeast assay may be the most effective way to detect p53 mutation in breast cancer²⁵ and the utility of this assay is being tested in clinical trials. In oesophageal cancer, the assay has demonstrated a higher incidence of p53 mutation than previously presumed with p53 mutation assessed using this technique associated with tumour behaviour.²⁶

This technique may also prove valuable in aiding the decision as to whether tumours are likely to be resistant to chemotherapy or radiotherapy, thus, preventing patients undergoing ineffective treatment and helping to target resources to those most likely to benefit.

IMMUNOHISTOCHEMISTRY
The problem faced by many of the techniques already described in this review is that the architecture and spatial relationships of the cancer tissue to surrounding structures and whether it is the cancer tissue or the stroma or lymphocytes being examined is usually unclear. Immunohistochemistry has the advantage of demonstrating the presence of proteins and in which cells and cellular compartments the protein is present. It can be used to supplement the techniques already outlined or protein analyses such as western blotting or enzyme-linked immuno-sorbent assay (ELISA) techniques and by using a range of antibodies to a particular protein, functional associations can be implied.

Thus, using antibodies for example to p53 (Figure 5) one can examine the overall presence of p53 using DO-1 in tissue sections and look for evidence of p53 activation using the FP3 antibody which detects phosphorylation of one of the serine residues of p53.

Figure 5: Cartoon of the p53 gene identifying three of the key domains: the transactivating, DNA binding and regulatory domains. The location of epitopes detected by two different p53 antibodies noted: D0-1 which detects an epitope for aminoacids 18-26 at the N terminal end of p53 (and hence is likely to detect any p53 protein present) and FP3 which detects phosphorylation of serine 392 at the C terminal end and is a measure of p53 activation. The mutation identified in an oesophageal cancer at position 273 is indicated; since this mutation lies in the DNA binding domain, it can prevent normal functioning of the p53 gene.

Figure 6: Sections of four different breast cancers stained with DO-1 to detect p53 protein. Staining varies greatly from sample to sample and indicates some of the difficulties in interpreting reported studies of p53 immunohistochemistry particularly with regard to prognosis in breast and colorectal cancers.

Even so, there remain problems in how immunohistochemistry should be interpreted as sections from four cancers demonstrate (Figure 6). Assuming conditions of fixation, processing and staining are consistent, what is a positive score? Does it matter which cell compartment (nucleus, cytoplasm or cell membrane) the staining is identified in? Does the presence of detectable protein imply it is abnormal, or might there be other mechanisms involved?

RECENT DEVELOPMENTS
In the last two decades, techniques have moved on from using relatively large quantities of tumour material to look at one aspect of one gene (for example allele loss by Southern blotting) to look at a multiplicity of genes or samples. The current techniques include technologies using RNA as a starting point where thousands of genes can be examined on one cancer sample (for example using the Affymetrix system). Arrays customised to look at a particular range of genes or all the known mutations for one gene such as p53 are now operational. This approach has already suggested that tumour profiling may well allow us to categorise patients into groups suitable for particular therapies, predict outcome and challenge theories of the metastatic process.^27,28 Tissue microarrays, where multiple tiny cores of many tissues or tumours are placed on a slide and can be stained and examined together, are particularly exciting to look at material from large clinical trials where treatment differences and outcome are well documented. Proteomics, looking for novel proteins in tumour or serum samples also looks set to produce new insights into cancer biology.

Often these techniques require supportive evidence (for example confirmatory PCR after Affymetrix array analysis) and substantial bioinformatics and biostatistics analyses before the full clinical import can be realised. In addition, laboratory developments are being applied in tandem with newer imaging modalities such as positron emission tomography (PET) scanning and links to other scientific disciplines such as mathematics, using computerised modelling techniques or neural network analyses may provide future benefits.²⁹

ACKNOWLEDGEMENTS
The author is particularly grateful to the following individuals and organisations who have made this review possible. The collaborative work described in this paper was performed by, or in conjunction with:

Virginia Appleyard, Kathryn Ball, Mei-Ling Ball, Susan Bray, Sir David Carter, Frank Carey, Udi Chetty, Ashley Craig, Sir Alfred Cuschieri, John Evans, Sir Patrick Forrest, Ted Hupp, David Kellock, Neil Kernohan, Sir David Lane, Anita Liem, Helen McDowell, Gillian Nicoll, Stuart Oglesby, Norman Pratt, Colin Purdie, Michael Steel, Bob Steele, Craig Stocks, George Thomson, Emma Warbrick, Roland Wolf, Adam Yagui-Beltran and Dorin Ziyaie. 

The research was funded by Breast Cancer Research Scotland, Cancer Research UK, Chief Scientists Office, Royal College of Surgeons of Edinburgh, Medical Research Council, Melville Trust, Rotary, Scottish Hospitals Endowments Research Trust, Tenovus, the Universities of Dundee and Edinburgh and many individual donors.

REFERENCES
1. Hastie ND, Dempster M, Dunlop MG, Thompson AM, Green D et al. Telomere reduction in human colon cancer and with ageing. Nature 1990; 346: 866-868.
2. McIlrath J, Bouffler SD, Samper E, Cuthbert A, Wojcik A, Szumiel I, et al. Telomere length abnormalities in mammalian radiosensitive cells. Cancer Research 2001; 61: 912-915.
3. Nicoll G, Crichton DN, McDowell HE, Kernohan N, Hupp TR, Thompson AM. Expression of the hypermethylated in cancer gene (HIC-1) is associated with good outcome in human breast cancer. British Journal of Cancer 2001; 85: 1878-1882.
4. Djojosubroto MW, Choi YS, Lee HW, Rudolph KL Telomeres and telomerase in aging, regeneration and cancer. Mol Cells 2003; 15: 164-175.
5. Masutomi K, Hahn WC. Telomerase and tumorigenesis Cancer Lett 2003; 194: 163-172.
6. Esteller M. Relevance of DNA methylation in the management of cancer. Lancet Oncol 2003; 4: 351-358.
7. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene Science 1987; 235: 117-182.
8. Hubbard AL, Reid CP, Thompson AM, Chetty U, Anderson TJ. Critical determination of the frequency of c-erb B2 amplification in breast cancer British Journal of Cancer 1994; 70: 434-439.
9. Steele RJC, Thompson AM, Hall PA, Lane DP. The p53 tumour suppressor gene. British Journal of Surgery 1998; 85: 1460- 1467.
10. Vogelstein B, Lane D, Levine AJ. Surfing the p53 network. Nature 2000; 408: 307-310.
11. Liem AA, Chamberlain MP, Wolf CR, Thompson AM. The role of signal transduction in cancer treatment and drug resistance. European Journal of Surgical Oncology 2002; 28: 679-684.
12. Lucke CD, Philpott A, Metcalfe JC, Thompson AM, Hughes-Davies L, Kemp PR, et al. Inhibiting mutations in the transforming growth factor-B type 2 receptor in recurrent human breast cancer. Cancer Research 2001; 61: 482-485.
13. Stocks SC, Pratt N, Sales M, Johnston DA, Thompson AM et al. Chromosomal imbalances in gastric and oesophageal adenocarcinoma: specific CGH detected abnormalities segregate with junctional adenocarcinomas. Genes, Chromosomes and Cancer 2001; 32: 50-58.
14. Thompson AM. Oesophageal and Stomach cancers. In Scenario Planning for Cancer in Scotland, Chapter 2, Scottish Executive, Edinburgh 2001.
15. Hislop RG, Stocks SC, Steel CM, Sales M, Pratt, N, Goudie D et al. Karyotypic aberrations of chromosomes 16 and 17 in breast cancer are associated with survival. British Journal of Surgery 2002; 89: 1581- 1586.
16. Hupp TR, Lane DP, Ball KL Strategies for manipulating the p53 pathway in the treatment of human cancer. Biochem J 2000; 352: 1-17.
17. Ziyaie D, Hupp TR, Thompson AM. p53 and breast cancer. The Breast 2000; 9: 239-246.
18. Aas T, Borresen Al, Geisler S, Smith-Sorensen B, Johnsen H, Varhaug JE, et al. Specific p53 mutation are associated with de novo resistance to doxorubicin in breast cancer patients. Nat Med 1996; 2: 811-814.
19. Bertheau P, Plassa F, Espie M, Turpin, E, de Roquancourt A, Marty M, et al. Effect of mutated TP53 on respone of advanced breast cancer to high dose chemotherapy.Lancet 2002; 360: 852-854.
20. Coles C, Thompson AM, Elder PA, Cohen BB, Mackenzie IM, Cranston G, et al. At least two genes on chromosome 17 are implicated in human breast carcinogenesis. Lancet 1990; 336: 761-763.
21. Thompson AM, Anderson TJ, Condie A, Prosser J, Chetty U, Carter DC, p53 allele losses, mutations and expression in breast cancer and their relationship to clinico pathological parameters. International Journal of Cancer 1992; 50: 528-532.
22. Thompson AM, Crichton DN, Elton RA, Clay MF, Chetty U, Steel CM. Allelic imbalance at chromosome 17p13.3 (YNZ22) in breast cancer is independent of p53 mutation or p53 overexpression and is associated with poor prognosis at medium-term follow-up. British Journal of Cancer 1998; 77: 797-800.
23. Phelan C, Borg A, Cuny M, Crichton DN, Baldersson T, Andersen TI, et al. Consortium study on 1280 breast carcinomas: allelic loss on chromosome 17 targets sub-regions associated with family history and clinical parameters. Cancer Research 1998; 58: 1004-1012.
24. Flaman JM, Robert V, Lenglet S, Moreau V, Iggo R, Frebourg T. Identification of human p53 mutations with differential effects on th bax and p21 promotors using functional assays in yeast. Oncogene 1998; 16: 1369-1372.
25. Yagui-Beltran A, Hupp TR, Thompson AM. The yeast functional assay for p53 mutation: comparison with the HOT technique and clinical/pathological parameters. British Journal of Cancer 2001; 85: Suppl 1, 79.
26. Oglesby SD, Warbrick E, Johnston D, Dillon JF, Munro AJ, Hupp TR, et al. P53 mutations in oesophageal adenocarcinoma are common and are associated with disease response. Gut 2002; 50: Suppl 11, A92. 
27. Van t’Veer LJ, Hongyue D, van de Vijver MJ et al. Gene expression profiling predicts clinical outcome of breast cancer. Nature 2002; 415: 530-536.
28. Ramaswamy S, Ross KN, Lander ES, Golub TR A molecular signature of metastasis in primary solid tumors. Nature Genetics 2003; 33: 49- 54.
29. Anderson ARA, Chaplain MAJ, Newman EL, Steele RJC, Thompson AM. Mathematical modelling of tumour invasion and metastasis. J. Theoret Med 2000; 2: 129-154.

Copyright: 11 December 2003