Mechanisms of Tumour Development

WCR-S3.Q (composition) 27/01/03 9:26 Page 1

3 Mechanisms of tumour development

The phenotypic changes which a cell undergoes in the process of malignant transformation is a reflection of the sequential acquisition of genetic alterations. This multi-step process is not an abrupt transition from normal to malignant growth, but may take place over 20 years or more. The mutation of critical genes, including suppressor genes, oncogenes and genes involved in DNA repair, leads to genetic instability and progressive loss of differentiation. Tumours enlarge because cancer cells lack the ability to balance cell division by cell death (apoptosis) and by forming their own vascular system (angiogenesis). The transformed cells lose their ability to interact with each other and exhibit uncontrolled growth, invade neighbouring tissues and eventually spread through the blood stream or the lymphatic system to distant organs. WCR-S3.Q (composition) 27/01/03 9:26 Page 84

MULTISTAGE CARCINOGENESIS

SUMMARY During this process the cell develops : - Defects in terminal differentiation > Tumours consist of cells whose growth - Defects in growth control - Resistance to cytotoxicity and morphological characteristics are - Defects in programmed cell death markedly different from those of normal cells. Criteria for malignancy include increased cell proliferation, loss of differ- CHEMICALS Selective entiation, infiltrative growth and metasta- Genetic clonal Genetic Genetic Genetic sis to other organs. change expansion change change change

> Malignant transformation is a multistage VIRUS process, typically a progression from PRE- benign lesions (e.g. adenoma) to malig- NORMAL INITIATED NEOPLASTIC MALIGNANT CLINICAL CANCER nant tumours (e.g. carcinoma). This evo- RADIATION CELL CELL LESION TUMOUR CANCER METASTASIS lution of malignant cells is caused by the sequential accumulation of alterations in genes responsible for the control of cellular proliferation, cell death and the These steps are caused by : maintenance of genetic integrity. - Activation of proto-oncogenes - Inactivation of tumour suppressor genes - Inactivation of genomic stability genes > The development of cancer may be initiated by environmental agents (chemical carcinogens, radiation, viruses) and inherited genetic factors (germline Fig. 3.1 Carcinogenesis is a multistage process involving multiple genetic and epigenetic events in proto- mutations). oncogenes, tumour suppressor genes and anti-metastasis genes.

Cancer arises from a single cell Malignant tumours (or “cancers”) are specific cell populations may be identified cells or tissues exposed to them. DNA- described as monoclonal, meaning that as marking a commitment towards malig- damaging activity may be identified on the each tumour arises from a single cell. The nancy, and these may be exploited as an basis of defined protocols (sometimes development of a malignant tumour from early indicator in the context of carcino- called “short-term tests”, to emphasize a normal cell usually occurs over a con- gen testing [3]. Thus, wholly on morpho- their difference from chronic lifetime siderable fraction of our lifetime. Such a logical grounds, cancer may be perceived bioassay in rodents). Chemicals which long period is reflected, for example, by as the outcome of a complex biological exhibit mutagenic activity in short-term the difference between the age at which process. tests, which typically involve sensitive people start smoking and the age at bacterial strains and cell-free extracts to which diagnosis of lung cancer most Multiple steps are required for a can- catalyse metabolism of the test com- often occurs. The long “latent period” in cer to arise pound, are characterized as “genotoxic” lung cancer and almost all other malig- Animal “models” of cancer development, [5]. Genotoxic agents may be complete nancies is not explicable on the basis of a most commonly involving treatment of carcinogens, but can also act as “initiating single-step transition from a normal cell rodents with carcinogenic chemicals or agents”. After a single treatment with an to malignant one. Rather, the tumour is other cancer-inducing agents, have pro- initiating agent, tumour growth may be the outcome of an evolutionary process vided clear evidence that specific stages facilitated by chemicals (or treatments) involving successive generations of cells, in malignant transformation can occur dis- which stimulate cell proliferation, some- which are progressively further advanced cretely [4]. Chemicals which cause cancer times by inducing mild toxic damage in towards cancerous growth [1]. in animals without the need for other exposed tissue. These agents are termed Human histopathological observations treatment are sometimes called “com- “promoters” (Table 3.1). As well as these support this scenario, and a range of pre- plete carcinogens” (although “carcino- genotoxic chemicals, a range of non-geno- malignant lesions have been identified gens” would be appropriate). Most such toxic agents can cause cancer in humans [2]. Likewise, in experimental animals, carcinogens cause damage to DNA of and/or experimental animals [6].

84 Mechanisms of tumour development WCR-S3.Q (composition) 29/01/03 13:58 Page 85

The stages in tumorigenesis have been designated “initiation”, which encom- passes damage to, and then division of Factor Cancer site/cancer exposed cells such that their growth potential is changed irreversibly, and Hormones Estrogens, progesterone Uterus, mammary gland “progression”, denoting multiple rounds Gonadotrophins Ovary, testis, pituitary of cell replication mediating the gradual Testosterone Prostate gland transition of an initiated cell towards Pharmaceutical Oral contraceptives Liver autonomous, cancerous, growth. Ultimate products Anabolic steroids Liver spread of malignant cells resulting in mul- Analgesics Renal pelvis tiple tumour sites has been termed “metastasis”. The unequivocal identifica- Miscellaneous Bile acids Small intestine tion by the mid-1970s of these various substances Saturated fatty acids Colon phases was one indication that carcino- Salt Stomach genesis is a multistage process. Arguably, Tobacco Oral cavity, lung, bladder etc. the greatest achievement of cancer Saccharin, uracil, melamine, Urinary bladder research during the last decades of the tetraphthalic acid and other 20th century has been the elucidation of xenobiotics causing urinary stones multistage carcinogenesis at the molecu- Dichlorobenzene, trimethylpentane Kidney (lead-free gasoline), perchloroethyl- lar genetic level. ene Butylated hydroxyanisole, propionic Stomach The molecular basis of tumour acid pathology Nitrilotriacetate Kidney In a seminal publication, Vogelstein and colleagues [7] provided evidence that Table 3.1 Promoting agents: non-genotoxic agents that facilitate carcinogenesis by stimulating cell division. the different stages in the cellular evo- Tobacco smoke also contains genotoxic carcinogens. lution of colon cancer in humans, histo- logically identified as hyperplasia, CHROMOSOME: 5q 12p 18q 17p early-stage adenoma, late-stage adeno- ALTERATION: Mutation Mutation Loss Loss ma etc., could be identified with specif- GENE: FAP KRAS DCC? p53 ic successive genetic changes (Fig. 3.2). The genetic changes included oncogene activation by mutation at DNA Other specific sites and loss of chromosomal hypomethylation alterations regions (necessarily involving multiple genes) which were subsequently shown to be the location of tumour suppressor Normal Hyperproliferative Early Intermediate Late Carcinoma Metastasis genes. Since that initial description, epithelium epithelium adenoma adenoma adenoma knowledge of the molecular genetic basis for human colon cancer has been Fig. 3.2 The original Vogelstein model for the genetic and histological evolution of colon cancer. massively extended (Colorectal cancer, (Colorectal cancer, p198). p198). For most tumours, the genetic changes are not inherited from our par- ents but arise in a previously normal Commonality and heterogeneity amplification) are common to a number cell. The progeny of this cell after cell The molecular biological basis of multi- of tumour types. However, each tumour division carry the same genetic change stage carcinogenesis initially described type is associated with a distinctive set but the surrounding cells remain nor- for colon cancer appears to have appli- of gene alterations. The genes in ques- mal. Because these genetic changes cation to all tumour types, although tion are discussed under the subhead- affect only the cancer cells, they are there is marked variation in the extent ing Pathology and genetics for each of not passed on to the children of cancer to which genes relevant to particular the tumour types included in Chapter 5. patients. However, in a minority of tumours have been identified [8]. Some Such enumeration of relevant genes cases some critical changes are inherit- genes, and the corresponding change necessitates a degree of simplification. ed, giving a familial predisposition to associated with tumorigenesis (muta- There is clear heterogeneity between colon or other cancers. tion, overexpression, deletion and/or individual tumours of the same type. In

Multistage carcinogenesis 85 WCR-S3.Q (composition) 27/01/03 9:27 Page 86

Peutz-Jeghers polyp

RER+ cancer (Replication Dysplasia in hamartoma Error Positive)

Loss of mismatch repair Juvenile polyp Normal Early adenoma Intermediate adenoma Late adenoma Cancer

Flat dysplasia Ulcerative colitis-associated colorectal carcinoma

MHAP/Serrated adenoma Cancer in mixed hyperplastic adenomatous polyps (MHAP)

Flat adenoma Flat cancer

Fig. 3.3 Histological representation of the pathogenesis of colorectal cancer. Phenotypic changes in the morphology of the colonic mucosa reflect the sequential acquisition of genetic alterations.

other words, not every tumour will nec- and tumour suppressor genes, p96) have esis are being progressively identified essarily exhibit all the genetic changes been identified in terms of their biological [11]. As noted earlier, members of some established for the tumour type in ques- function [9]. Such genes are among those cancer-susceptible families inherit mutation. Moreover, there is often marked that facilitate transmission of growth con- tions in particular genes that contribute to heterogeneity within an individual trol signals from the cell membrane to the cancer development, and hence to their tumour: adjacent cells differ. Mapping nucleus (that is, signal transduction), that individual risk of disease. However, with and identification of genes involved in mediate cell division, differentiation or cell most cancers, the genetic change critical malignant transformation has been a death and, perhaps most critical of all, to carcinogenesis results from damage to major component of the study of the that maintain the integrity of genetic infor- DNA by chemicals, radiation and viruses molecular mechanisms of carcinogene- mation by DNA repair and similar process- (Fig. 3.1). This damage is not entirely and sis. es (Carcinogen activation and DNA repair, perhaps not predominantly produced by p89). Since mutations are normally infre- exogenous agents but by natural process- Multiple genetic changes required quent events, it seems unlikely that in the es, such as the production of reactive oxy- The emergence of a malignant cell popu- course of a human lifetime a cell would gen species or the spontaneous deamina- lation is understood to be the cumulative acquire all the mutations necessary for tion of the 5-methylcytosine naturally effect of multiple (perhaps five, ten or cancer to develop, unless at some point present in DNA [13]. Furthermore, as more) genetic changes, such changes the developing cell lost its ability to pro- shown as the second step in Fig. 3.2, bio- being accumulated in the course of the tect itself against mutation and gained logical change that is heritable may result evolution of the cell from normal to malig- what is called a “mutator” phenotype [10]. from non-genetic processes including the nant. The genes designated as oncogenes Thus, alterations in gene structure and modulation of gene expression by hyper- and tumour suppressor genes (Oncogenes expression which bring about carcinogen- methylation [12].

86 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 87

PRECURSOR LESIONS IN Engl J Med, 328: 901-906, 1993). Several CHEMOPREVENTION TRIALS epidemiological studies have shown that regular use of aspirin or related drugs is Trials of agents for chemopreventive associated with a reduced adenoma inci- activity which are based on assessment dence (IARC Handbooks of Cancer of malignant disease are almost unman- Prevention. Vol. 1, Lyon, 1997). This pro- ageable because of the long period of vides further confirmation that adenomas time (perhaps decades) potentially are precursor lesions for colon cancer, involved. Attention has therefore been since aspirin is known to reduce the inci- focused on lesions, either cellular or dence of malignant colon cancer. molecular, demonstrated to be valid indi- Fig. 3.6 Tubular adenoma of the colon is a precur- cators of the subsequent development of Potential precursor lesions of carcinogene- sor lesion for colorectal cancer. malignancy. A trial may then evaluate the sis include both phenotypic and genotypic effect of the putative chemopreventive markers (Miller AB et al. Biomarkers in agent on such precursor lesions. Cancer Chemoprevention, IARC Scientific cific gene and general chromosomal dam- Publications 154, Lyon, 2001). Thus oral age, cell growth regulatory molecules, The best-validated precursor lesions are leukoplakia is a recognized precursor for and biochemical activities (e.g. enzyme benign tumours, such as colorectal ade- cancer of the oral cavity. Histological mod- inhibition). Serum proteins are of special nomas. It is established that adenoma ulation of a precancer (often called intraep- interest because of their availability. Thus number, size, and severity of dysplasia are ithelial neoplasia) has been used as a pre- prostate-specific antigen (PSA) is being predictive factors for colorectal cancer cursor lesion in prevention trials (Kelloff GJ used as a “surrogate” marker for prostate incidence. It has been estimated that 2- et al., Cancer Epidemiol Biomarkers Prev, 9: cancer. It is expected that the number and 5% of all colorectal adenomas progress to 127-137, 2000). Additionally, genetic variety of biomarkers for precursor adenocarcinomas if not removed or treat- lesions such as progressive genomic insta- lesions will continue to expand In parallel ed. The risk is greater for large and bility as measured by loss of heterozygosi- with the advances in understanding of the severely dysplastic polyps. Cancer risk is ty or amplification at specific microsatellite genetic and cellular basis of carcinogene- decreased by polyp removal, and a strong loci, have been considered (Califano J et al. sis. correlation exists between the relative Cancer Res, 56: 2488-2492, 1996). Other prevalence of adenomas and cancers potential precursor endpoints include pro- across populations (Winawer SJ et al., N liferation and differentiation markers, spe-

Ageing also at the cellular level. In humans, as well when further maintained in culture, once- Apart from multistage development, cer- as in other mammals, the incidence of normal cells acquire the same characteris- tain other processes are fundamental to cancer rises dramatically with age. An tics as cells cultured from malignant malignant disease. Principal amongst exponential increase occurs from mid-life tumours. These and various other alter- these is ageing, which can be considered [14]. Passage of time is also critical to cell ations in growth characteristics are recog- both in relation to the whole individual, and biology. Normal cells do not divide indefi- nized as the experimental counterpart of nitely due to senescence (Box: Telomeres multistage carcinogenesis through which and Telomerase, p108). Senescent cells tumours develop in intact animals or cannot be stimulated to divide further, humans. The genetic basis for senescence, become resistant to apoptotic cell death and its relationship to malignancy, is a sub- and acquire differentiated functions. ject of intense investigation [15]. Senescence may be an anti-cancer mechanism that limits accumulation of muta- Preventing cancer tions. However, when maintained in cul- The significance of multistage carcino- ture, cells treated with carcinogenic chem- genesis extends beyond facilitating icals or infected with oncogenic viruses understanding of how a transition from Fig. 3.4 Severe intraepithelial neoplasia (dyspla- may avoid senescence and proliferate normal to malignant cell growth occurs. sia) in the epithelium of an intrahepatic large bile indefinitely. Such cell populations are The fundamental cellular studies outlined duct, a condition caused by hepatolithiasis. described as being “transformed” and earlier provide a basis for preventing can-

Multistage carcinogenesis 87 WCR-S3.Q (composition) 27/01/03 9:27 Page 88

cer (see chapter 4). The fact that partic- mechanisms now known to operate in ular patterns of cell morphology and the proliferation of cancer cells provide a growth precede emergence of an basis for the development of new, more unequivocally malignant cell population efficient therapies without the side- is the basis of secondary prevention of effects that currently often afflict cancer cancer. patients [17]. Examples include detection of polyps in the large bowel (Fig. 3.5) and of morphological change which is the basis of the Papanicolaou smear test for early detection of cervical cancer. Moreover, dietary or pharmaceutical interventions calculat- ed to prevent or reverse such lesions are the basis of chemoprevention [16]. Most importantly, knowledge of the genetic basis underlying tumour growth should provide new criteria for individual deter- Fig. 3.5 Pedunculated hyperplastic polyp of the mination of diagnosis and prognosis. The colon.

REFERENCES

1. Foulds L, ed. (1969) Neoplastic Development, Vol. 1, mechanisms: a European project. Mutat Res, 353: 47-63. 13. Marnett LJ, Plastaras JP (2001) Endogenous DNA London, Academic Press. damage and mutation. Trends Genet, 17: 214-221. 7. Vogelstein B, Fearon ER, Hamilton SR, Kern SE, 2. Correa P (1996) Morphology and natural history of can- Preisinger AC, Leppert M, Nakamura Y, White R, Smits AM, 14. Armitage P, Doll R (1954) The age distribution of cancer precursors. In: Schottenfeld D, Fraumeni JF, eds, Bos JL (1988) Genetic alterations during colorectal-tumor cer and a multistage theory of carcinogenesis. Br J Cancer, Cancer Epidemiology and Prevention, New York, Oxford development. N Engl J Med, 319: 525-532. 8: 1-12. University Press, 45-64. 8. Balmain A, Harris CC (2000) Carcinogenesis in mouse 15. Wynford-Thomas D (1999) Cellular senescence and 3. Ito N, Imaida K, Asamoto M, Shirai T (2000) Early and human cells: parallels and paradoxes. Carcinogenesis, cancer. J Pathol, 187: 100-111. detection of carcinogenic substances and modifiers in rats. 21: 371-377. 16. Bartsch H (2000) Studies on biomarkers in cancer Mutat Res, 462 : 209-217. 9. Evan GI, Vousden KH (2001) Proliferation, cell cycle etiology and prevention: a summary and challenge of 20 4. Weinstein IB (1982) Carcinogenesis as a multistage and apoptosis in cancer. Nature, 411: 342-348. years of interdisciplinary research. Mutat Res, 462: 255- process—experimental evidence. In: Bartsch H, Armstong 279. 10. Loeb LA (2001) A mutator phenotype in cancer. B, eds, Host Factors in Human Carcinogenesis (IARC 17. Kallioniemi OP, Wagner U, Kononen J, Sauter G Cancer Res, 61: 3230-3239. Scientific Publications No. 39) Lyon, IARCPress, 9-25. (2001) Tissue microarray technology for high-throughput 5. Vainio H, Magee PN, McGregor DB, McMichael AJ, eds 11. Hahn WC, Counter CM, Lundberg AS, Beijersbergen molecular profiling of cancer. Hum Mol Genet, 10: 657- (1992) Mechanisms of Carcinogenesis in Risk Identification RL, Brooks MW, Weinberg RA (1999) Creation of human 662. (IARC Scientific Publications No. 116), Lyon, IARCPress. tumour cells with defined genetic elements. Nature, 400: 464-468. 6. Yamasaki H, Ashby J, Bignami M, Jongen W, Linnainmaa K, Newbold RF, Nguyen-Ba G, Parodi S, Rivedal E, 12. Esteller M, Corn PG, Baylin SB, Herman JG (2001) A Schiffmann D, Simons JW, Vasseur P (1996) Nongenotoxic gene hypermethylation profile of human cancer. Cancer carcinogens: development of detection methods based on Res, 61: 3225-3229.

88 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 89

CARCINOGEN ACTIVATION AND DNA REPAIR

Carcinogen activation carbons, aromatic amines, N-nitro- SUMMARY The first indication that certain cancers samines, aflatoxins and vinyl halides, were associated with exposure to chemi- which yield electrophilic species through >Many chemical carcinogens require cals arose from observations by clini- phase I activation [3]. Other metabolic spontaneous or enzymatic activation to cians in the 18th and 19th centuries. The pathways are known. For example, produce reactive intermediates which field of experimental chemical carcino- dihaloalkanes are activated to carcino- bind to DNA. The resulting carcinogen- DNA adducts may be eliminated from genesis started in 1915 with the experi- genic metabolites by glutathione trans- DNA by various enzyme-mediated repair ments of Yamagiwa and Ichikawa, who ferases. processes. showed that application of tar to the ears Understanding of carcinogen-DNA inter- of rabbits induced skin tumours. In the actions (Fig. 3.9) has resulted largely >In cells and tissues with deficient DNA 1940s, experiments on mouse skin from the development of sensitive and repair, replication of carcinogen-dam- demonstrated the stepwise evolution of specific methods for determining DNA aged DNA may result in the mutation of cancer and allowed the characterization adducts [4]. The most frequently used genes that regulate cell growth and dif- of two classes of agents, initiators and methods include immunoassays using ferentiation in target cell populations. promoters [1]. Most chemical carcino- adduct-specific anti-sera or antibodies, Such genetic alterations typically lead gens are subject to metabolism that 32P-postlabelling, fluorescence spec- to progressive genetic instability resulting in uncontrolled growth, loss of dif- results in their elimination, but in the troscopy, electrochemical detection and ferentiation, invasion and metastasis. course of which reactive intermediates mass spectrometry. Measurement of car- are generated. Such metabolic activation cinogen-DNA adducts in rodents has results in the modification of cellular revealed correlations between the con- macromolecules (nucleic acids and pro- centration of the carcinogen in the envi- teins) [2]. Accordingly, mutagenicity ronment, DNA adduct levels in tissues tests using bacteria and mammalian cells where tumours may arise and cancer Experimental studies in rodents and in in culture were developed and are exten- incidence. It is therefore accepted that cultured cells have led to the classifica- sively used to identify potential carcino- DNA adducts may be used as indicators tion of chemical carcinogens into two gens. Not all chemicals known to cause broad classes: genotoxic and non-geno- cancer, however, can be demonstrated to toxic. Genotoxic carcinogens alter the bind to DNA and hence be classified as structure of DNA, mostly by covalent “genotoxic”. PROCARCINOGEN binding to nucleophilic sites. These Activation of chemical carcinogens in mam- lesions, that is, the chemical entity of malian tissue mostly occurs through oxida- Detoxification Metabolic activation carcinogen bound to DNA, are called tion by microsomal mono-oxygenases DNA “adducts”. The replication of DNA (cytochromes P450, phase I enzymes). Ultimate carcinogen containing unrepaired adducts may Cytochromes P450 are located in the result either in the generation of endoplasmic reticulum (internal mem- Covalent binding to sequence changes (mutations) in the branes of the cell) and constitute a super- DNA, RNA, proteins newly synthesized daughter strands of family of proteins; about 50 are now Promutagenic DNA DNA or in DNA rearrangements evident known in humans. The oxidation products adducts as chromosome aberrations. are substrates for other families of This critical, irreversible genetic event enzymes (transferases, phase II enzymes) No or error-prone can thus result in fixation of the original which link the carcinogen residues to a DNA repair DNA repair structural change in DNA as permanent, glutathione, acetyl, glucuronide or sulfate transmissible, genetic damage, or in the group; the resulting conjugates are Cell replication Cell replication with loss of genetic information through hydrophilic and thus can be easily excret- with no DNA sequence changes alterations in chromosomes. Such heri- ed. Carcinogenic electrophilic metabo- DNA sequence changes (gene mutations) table change has the potential to per- lites arise as by-products of these meta- turb growth control in the affected cell, bolic reactions. The metabolic pathways NORMAL CELLS INITIATED CELLS and is sometimes referred to as the “ini- are well characterized for the major tiation” step of the tumorigenic process classes of chemical carcinogens (Fig. Fig. 3.7 Critical stages in the process of initiation (Fig. 3.7). 3.8), including polycyclic aromatic hydro- by genotoxic chemicals.

Carcinogen activation and DNA repair 89 WCR-S3.Q (composition) 27/01/03 9:27 Page 90

of the effective biological exposure, and adducts, together with other genetic polymorphisms involved, large-scale stud- hence of carcinogenic risk in humans [5]. determinants that regulate DNA repair or ies and high throughput assays (based on However, analysis of DNA adducts in cell cycle control, for example, and thus DNA microchips, for example) will be human cells and tissues remains diffi- affect the outcome of exposure to DNA- required to fully elucidate the complex cult, due to the very low levels of adducts damaging agents and influence cancer nature of such gene-environment interac- present in DNA (typically, one adduct per risk in different individuals [6]. Many stud- tions. 107-108 parent nucleotides). ies have sought to correlate genetic poly- Activities of the enzymes involved in car- morphisms, adduct levels and cancer risk Mutational spectra cinogen metabolism vary greatly between in human populations (Genetic suscepti- Adducts of DNA and proteins can be used individuals due to induction and inhibition bility, p71). These studies have hitherto as early markers of exposure to carcino- processes or to gene polymorphisms that provided some correlations for risk pre- gens as indicated. However, because can affect activity. These variations can diction at the population level. However, adducts only persist for a short time (typ- affect the formation of carcinogen-DNA due to the great number of enzymes and ically, for a few hours or days for DNA

Fig. 3.8 Carcinogen activation by mammalian enzymes: reactions catalysed during metabolism of benzo[a]pyrene and NNK (4-(methylnitrosamino)-1-(3- pyridyl)-1-butanone), both contained in tobacco, and of aflatoxin B1, produce reactive intermediates (ultimate carcinogens, in box), which bind to DNA. Other reaction pathways leading to the formation of glucuronides and other esters, which are excreted, are not shown. 1. Benzo{a}pyrene-7, 8-diol-9, 10-epoxide; 2. 4-(methyl- 7 nitrosamino)-1-(3-pyridyl)-1-butanol; 3. Diazohydroxide; 4. Diazohydroxide; 5. Aflatoxin B1-8,9-oxide; 6. 2,3-Dihydro-2-(N - guanyl)-3-hydroxyaflatoxin B1.

90 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 91

adducts, a few weeks or months for albu- ribosyl-transferase gene HPRT, when inac- of almost all tumour types. A large data- min or haemoglobin adducts), their use- tivated by mutation, renders cells resist- base of p53 mutations has been generat- fulness as exposure markers is limited. ant to growth inhibition by 6-thioguanine; ed. Mutational spectra have been identi- Mutations in specific genes can be used such mutant cells can therefore be isolat- fied that provide evidence for the direct as longer-term “biomarkers” of early bio- ed by culture in the presence of this action of environmental carcinogens in logical effects or of disease [7]. Indeed, agent. Studies in humans have associated the development of certain cancers (i.e. in gene mutation patterns are probably the increases in the frequency of HPRT muta- these cases, cancer can be linked causal- only biological marker that can be char- tions (measured in circulating lympho- ly to past exposure to a defined carcino- acteristic of a past exposure to a carcino- cytes) with exposure to environmental genic agent). These mutations, which genic agent or mixture. Study of such genotoxic agents. However, in contrast to could in principle be used to identify expo- mutations will increasingly assist in the observations made in rodents, in which sure to particular agents, have been identification of etiologic agents, in risk mutation profiles often reflect the rela- termed “signature” mutations. They result prediction and in cancer prevention stud- tively extreme DNA damage that induced from the formation of specific DNA ies. Mutation spectra can be analysed them, characteristic HPRT mutation spec- adducts. For example, p53 mutations either in normal tissues (including blood tra (i.e. the types and positions of the characteristic of the known or suspected cells) or in tumour tissues. Analysis of base changes within the DNA sequence of etiological agent occur in lung cancer mutations in normal tissues remains diffi- the HPRT gene) are more difficult to (attributable to benzo[a]pyrene in tobacco cult, because the mutant cell or DNA observe in humans. smoke) and hepatocellular carcinomas

must be identified against a background The identification of oncogenes and (due to aflatoxin B1 in contaminated food) of a very large excess of non-mutant cells tumour suppressor genes (Oncogenes and (Box: Geographic variation in mutation or DNA, and a selection or an enrichment tumour suppressor genes, p96) has led to patterns, p102). In general, however, it is step is required. In contrast, mutations in the characterization of gene mutations often not practical to obtain DNA from tumour cells often favour growth and are which are more directly associated with healthy tissue to analyse for potentially amplified due to clonal expansion of the carcinogenesis. The RAS family of onco- tumorigenic mutations, as invasive meth- tumour cell population. genes was among the first that was rec- ods of sampling are required. Fortunately, A few genes are suitable markers ognized as being mutated in a wide variety the protein products of the mutated genes (“reporters”) of mutation induction in of human cancers. p53 is the most com- and, even the mutated DNA itself, can be experimental animals and in humans. monly altered tumour suppressor gene in detected and measured in body fluids or Thus the hypoxanthine-guanine phospho- human cancer, being mutated in over 50% secretions, such as blood plasma, that have been in contact with the malignant tissue. DAMAGING AGENT Presumed signature mutations have also been identified in “normal” tissues (non- X-rays UV light X-rays Replication pathological but probably containing initi- Oxygen radicals Polycyclic aromatic Anti-tumour agents errors Alkylating agents hydrocarbons (cisplatin, mitomycin C) ated cells) from exposed individuals. For Spontaneous reactions example, the p53 mutation associated

with exposure to aflatoxin B1 has been found in liver tissue and in plasma DNA G U T G A from healthy subjects (without cancer) G T G G C C T who have consumed food contaminated T with aflatoxins. Therefore, mutations in cancer genes could be used, in certain cases, as early indicators of risk before disease diagnosis.

Uracil 6-4 Photoproduct Interstrand cross-link A-G Mismatch Abasic site Bulky adduct Double-strand break T-C Mismatch DNA repair 8-Oxoguanine Cyclobutane pyrimidine- Insertion 9 Single-strand break dimer Deletion The 3 x 10 nucleotides of the DNA within each human cell are constantly exposed

Base-excision Nucleotide-excision Recombinational Mismatch repair to an array of damaging agents of both repair repair repair (homologous environmental origin, exemplified by sun- or end-joining) light and tobacco smoke, and of endoge- REPAIR PROCESS nous origin, including water and oxygen Fig. 3.9 Common DNA damaging agents, examples of DNA lesions induced by these agents and the most [8] (Table 3.2). This scenario necessitates important DNA repair mechanism responsible for the removal of these lesions. constant surveillance so that damaged

Carcinogen activation and DNA repair 91 WCR-S3.Q (composition) 27/01/03 9:27 Page 92

nucleotides may be removed and replaced before their presence in a DNA strand at the time of replication leads to the generation of mutations [9]. Restoration of normal DNA structure is achieved in human cells by one of several DNA repair enzymes that cut out the damaged or inappropriate bases and replace them with the normal nucleotide sequence. This type of cellular response is referred to as “excision repair” and there are two major repair pathways which function in this manner: “base excision repair” which works mainly on modifications caused by endogenous agents and “nucleotide excision repair” which removes lesions caused by environmental mutagens. UV light is probably the most common exogenous mutagen to which human cells are exposed and the importance of the nucleotide excision repair pathway in protecting against UV-induced carcinogenesis is clearly demonstrated in the inherited disorder xeroderma pigmentosum. Individuals who have this disease lack one of the enzymes involved in nucleotide excision repair and have a 1,000 times greater risk of developing skin cancer following exposure to sunlight than normal individuals. The genes in question have been named XPA, XPB, etc. [10]. One of the great achievements of the last two decades has been the isolation and characterization of the genes, and their protein products, involved in base excision repair and nucleotide excision repair. It has become apparent that certain proteins so identified are not exclusively involved in DNA repair but play an integral part in other cellular processes such as Fig. 3.10 Nucleotide excision repair (NER). Two NER pathways are predominant for removal of UV light- DNA replication and recombination. and carcinogen-damaged DNA. In global genome NER, the lesion is recognized by the proteins XPC and hHR23B while in transcription-coupled NER of protein-coding genes, the lesion is recognized when it stalls RNA polymerase II. Following recognition, both pathways are similar. The XPB and XPD helicases Excision repair of the multi-subunit transcription factor TFIIH unwind DNA around the lesion (II). Single-stranded binding The first step in both base excision repair protein RPA stabilizes the intermediate structure (III). XPG and ERCC1-XPF cleave the borders of the dam- and nucleotide excision repair is the aged strand, generating a 24-32 base oligonucleotide containing the lesion (IV). The DNA replication machinery then fills in the gap (V). recognition of a modification in DNA by enzymes that detect either specific forms of damage or a distortion in the DNA helix. Recognition of damage is followed Nucleotide excision repair may occur in the lesion in a reaction that uses the ATP- by an excision step in which DNA con- the non-transcribed (non-protein-coding) dependent helicase activities of XPB and taining the modified nucleotide is regions of DNA (Fig. 3.10, steps I to V). A XPD (two of the subunits of TFIIH) and removed. Gap-filling DNA synthesis and distortion in DNA is recognized, probably also involves XPA and RPA (II-III). The ligation of the free ends complete the by the XPC-hHR23B protein (I). An open XPG and ERCC1-XPF nucleases excise repair process. bubble structure is then formed around and release a 24- to 32-residue oligonu-

92 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 93

Reactive oxygen species X-rays Methylation, deamination (single-stranded break)

P OH

Spontaneous hydrolysis (abasic site) PARP DNA glycosylase XRCC1 I

PNK APE1

P II III DNA polβ PCNA δ/ε XRCC1 DNA pol +dNTPs +dGTP

IV VII Fig. 3.12 In the human genome there are numerous places where short sequences of DNA FEN1 are repeated many times. These are called microsatellites. In DNA from a patient with hereditary nonpolyposis colorectal cancer, there are changes in the number of repeats in the VIII microsatellites. Note the difference in the V microsatellite pattern between normal (N) and DNA DNA tumour tissue (T) from the same patient. This ligase 3 ligase 1 microsatellite instability is caused by errors in post-replicative DNA mismatch repair.

VI IX AP1) [12]. Gap-filling may proceed by SHORT-PATCH BASE EXCISION REPAIR LONG-PATCH BASE EXCISION REPAIR replacement of a single base or by resyn- (Main pathway) (Minor pathway) thesis of several bases in the damaged strand (depending on the pathway Fig. 3.11 Stages of base excision repair. Many glycosylases, each of which deals with a relatively narrow employed). spectrum of lesions, are involved. The glycosylase compresses the DNA backbone to flip the suspect More complex and unusual forms of dam- base out of the DNA helix. Inside the glycosylase, the damaged base is cleaved, producing an “abasic” age to DNA, such as double strand breaks, site (I). APE1 endonuclease cleaves the DNA strand at the abasic site (II). In the repair of single-strand- clustered sites of base damage and non- ed breaks, poly(ADP-ribose)polymerase (PARP) and polynucleotide kinase (PNK) may be involved. In the “short-patch” pathway, DNA polymerase β fills the single nucleotide gap and the remaining nick is sealed coding lesions that block the normal repli- by DNA ligase 3. The “long-patch” pathway requires the proliferating cell nuclear antigen (PCNA) and cation machinery are dealt with by alter- polymerases β, ε and δ fill the gap of 2-10 nucleotides. Flap endonuclease (FEN-1) is required to remove native mechanisms. Inherited human dis- the flap of DNA containing the damage and the strand is sealed by DNA ligase 3. eases in which the patient shows extreme sensitivity to ionizing radiation and altered processing of strand breaks, such as atax- cleotide (IV) and the gap is filled in by DNA base excision repair (Fig. 3.11, steps ia telangiectasia and Nijmegen breakage PCNA-dependent polymerases (POL) I to VI or steps III to IX) involves the syndrome, constitute useful models to epsilon and delta and sealed by a DNA lig- removal of a single base by cleavage of study the repair enzymes involved in these ase, presumed to be LIG1 (V). Nucleotide the sugar-base bond by a damage-specific processes. Indeed, if elucidation of base excision repair in regions which are tran- DNA glycosylase (e.g. hNth1 or uracil DNA excision repair and nucleotide excision scribed (and hence code for proteins) glycosylase) and incision by an repair was the great achievement of the requires the action of TFIIH [11]. apurinic/apyrimidinic nuclease (human late 1990s, then understanding strand

Carcinogen activation and DNA repair 93 WCR-S3.Q (composition) 27/01/03 9:27 Page 94

Other repair pathways SINGLE BASE MISPAIRS INSERTION OR DELETION LOOPS Human cells, in common with other α eukaryotic and prokaryotic cells, can also hMutSα hMutL hMutSα hMutLα or hMutSβ perform one very specific form of damage hMSH6 hPMS2 hMSH6 hPMS2 reversal, the conversion of the methylated adduct, O6-methylguanine, in DNA back to hMSH2 hMLH1 hMSH2 hMLH1 the normal base (Fig. 3.14). O6-Methylgua- nine is a miscoding lesion: both RNA and DNA polymerases “read” it incorrectly CA when they transcribe or replicate a DNA CTAGGTTA CACACACA template containing it. As this modified GATCCGAT GTGTGTGT base can pair with both the base cytosine (its correct partner) and the base thymine (an incorrect partner), its presence in DNA hMSH6 hPMS2 hPMS2 can give rise to transition mutations by mispairing of relevant bases. A specific hMSH2 hMSH2 hMLH1 hMLH1 protein, O6-alkylguanine-DNA-alkyltransferase, catalyses transfer of the methyl group from the guanine base to a cysteine CTAGGCTA CACACACA GATCCGAT GTGTGTGT amino acid residue located at the active site of the protein [13]. This error-free Fig. 3.13 Mismatch repair pathways: after DNA synthesis, base pairing mistakes that have escaped the process restores the DNA to its original editing function of DNA polymerase are recognized by mismatch repair proteins. state but results in the inactivation of the repair protein. Consequently, repair can be saturated when cells are exposed to high break repair will probably be the great reduced to try to avoid adverse reactions. doses of alkylating agents and synthesis of achievement of the next decade. This will A better understanding of the possible the transferase protein is required before have important consequences. Certain causes of this radiosensitivity, including repair can continue. cancers are often treated with radiother- characterization of the enzymes involved Mismatched bases in DNA arising from apy (Radiotherapy, p277) and a small per- in the repair of DNA damage produced by errors in DNA replication, for instance gua- centage of patients show considerable ionizing radiation, may lead to better tai- nine paired with thymine rather than cyto- sensitivity to their treatment, with the loring of radiotherapy doses to individual sine, are repaired by several pathways result that treatment schedules are patients. involving either specific glycosylases,

Agent Mutation hotspot Type of mutation Tumours associated (> = changes to)

Benzo[a]pyrene Codons 157, 158, 248, 273 G>T transversions Lung, larynx (tobacco smoke) 4-Aminobiphenyl Codons 280, 285 G>C transversions Bladder (aromatic dyes, tobacco smoke) G>A transitions

Aflatoxin B1 Codon 249 AGG>AGT Hepatocellular carcinoma (arginine > serine) Ultraviolet (UV) Codons 177-179, 278 C>T transitions Skin cancer CC>TT transitions (not melanoma) Vinyl chloride Several codons A>T transversions Angiosarcoma of the liver Endogenous mechanism Codons 175, 248, 273, 282 C>T transitions Colon, stomach (enhanced by nitric oxide) at CpG dinucleotides Brain cancers

Table 3.2 Spectra of p53 mutations caused by environmental carcinogens or endogenous mechanisms.

94 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 95

which remove the mismatched bases, or scans the DNA for a lesion or structure long-patch mismatch repair involving that is not a normal constituent of the homologues of the bacterial genes MUTS DNA. Defects in at least four of the genes and MUTL (Fig. 3.13). Insertion or deletion whose products are involved in mismatch loops at microsatellite sequences can be repair, namely hMSH2, hMLH1, hPMS1 α Me recognized by hMutS (a heterodimer of and hPMS2, have been associated with MGMT hMSH2 and hMSH6) or hMutSβ (a het- hereditary nonpolyposis colorectal cancer. Cys erodimer of hMSH2 and hMSH3). This is one of the most common genetic α MGMT Subsequent recruitment of hMutL (a het- diseases and affects as many as 1 in 200 Cys erodimer of hMLH1 and hPMS2) to the individuals and may account for 4-13% of Me altered DNA targets the area for repair, all colorectal cancers (Colorectal cancer, which requires excision, resynthesis, and p198). Affected individuals also develop Fig. 3.14 The repair of O6-methylguanine by ligation. Single nucleotide mispairing tumours of the endometrium, ovary and O6-alkylguanine-DNA-alkyltransferase. events require hMutSα function for recog- other organs. The DNA of hereditary non- nition. One important requirement of such polyposis colorectal cancer tumours is repair processes is that they are able to characterized by instabilities in simple directly from alterations in the proteins distinguish the correct base from the mono-, di- and trinucleotide repeats which involved in mismatch repair [14]. Generally incorrect one in the mispair. Since both are common in the human genome (Fig. speaking, genomic instability is considered bases are normal constituents of DNA, this 3.12). This instability is also seen in certain an indicator of, and fundamental to the cannot be achieved by an enzyme that sporadic colorectal tumour cells and arises nature of, malignant cell growth.

REFERENCES WEBSITES 1. Miller EC, Miller JA (1979) Milestones in chemical car- 8. Friedberg EC, Walker GC, Siede W, eds (1995) DNA A comprehensive listing of human DNA repair genes: cinogenesis. Semin Oncol, 6: 445-460. Repair and Mutagenesis, Washington DC, ASM Press. http://www.sciencemag.org/cgi/content/abstract/291/ 2. Miller JA, Miller EC (1977) Ultimate chemical carcino- 9. Lindahl T (2000) Suppression of spontaneous mutage- 5507/1284 gens as reactive mutagenic electrophiles. In: Hiatt HH, nesis in human cells by DNA base excision-repair. Mutat DNA Repair Interest Group (NCI): Watson, JD, Winsten, JA eds, Origins of Human Cancer Res, 462: 129-135. http://www.nih.gov:80/sigs/dna-rep/ (Book B), Cold Spring Harbor, Cold Spring Harbor 10. de Boer J, Hoeijmakers JH (2000) Nucleotide excision Laboratory, 605-627. repair and human syndromes. Carcinogenesis, 21: 453- 3. Guengerich FP (2000) Metabolism of chemical car- 460. cinogens. Carcinogenesis, 21: 345-351. 11. Benhamou S, Sarasin A (2000) Variability in 4. Hemminki K, Dipple A, Shuker DEG, Kadlubar FF, nucleotide excision repair and cancer risk: a review. Mutat Segerbäck D, Bartsch H, eds (1994) DNA Adducts. Res, 462: 149-158. Identification and Biological Significance (IARC Scientific 12. Cadet J, Bourdat AG, D'Ham C, Duarte V, Gasparutto Publications No. 125), Lyon, IARCPress. D, Romieu A, Ravanat JL (2000) Oxidative base damage to 5. Toniolo P, Boffetta P, Shuker DEG, Rothman N, Hulka B, DNA: specificity of base excision repair enzymes. Mutat Pearce N, eds (1997) Application of Biomarkers in Cancer Res, 462: 121-128. Epidemiology (IARC Scientific Publications No. 142), Lyon, 13. Pegg AE (2000) Repair of O6-alkylguanine by alkyl- IARCPress. transferases. Mutat Res, 462: 83-100. 6. Vineis P, Malats N, Lang M, d'Errico A, Caporaso N, 14. Pedroni M, Sala E, Scarselli A, Borghi F , Menigatti M, Cuzick J, Boffetta P, eds (1999) Metabolic Polymorphisms Benatti P, Percesepe A, Rossi G, Foroni M, Losi L, Di and Susceptibility to Cancer (IARC Scientific Publications Gregorio C, De Pol A, Nascimbeni R, Di Betta E, Salerni B, No. 148), Lyon, IARCPress. de Leon MP, Roncucci L (2001) Microsatellite instability 7. McGregor DB, Rice JM, Venitt S, eds (1999) The Use of and mismatch-repair protein expression in hereditary and Short- and Medium-Term Tests for Carcinogens and Data sporadic colorectal carcinogenesis. Cancer Res, 61: 896- on Genetic Effects in Carcinogenic Hazard Evaluation 899. (IARC Scientific Publications No. 146), Lyon, IARCPress.

Carcinogen activation and DNA repair 95 WCR-S3.Q (composition) 27/01/03 9:27 Page 96

ONCOGENES AND TUMOUR SUPPRESSOR GENES

mechanisms, including point mutations control the speed or timing of cell division SUMMARY that constitutively activate an enzyme, but rather its accuracy. Caretaker genes deletions that remove negative regulatory are usually involved in DNA repair and in > Human cells become malignant through regions from proteins, or increased the control of genomic stability. Their the activation of oncogenes and expression resulting from promoter dereg- inactivation does not enhance cell prolif- inactivation of tumour suppressor ulation or from multiplication of the num- eration per se but primes the cell for rapid genes. The pattern of genes involved varies markedly at different organ sites. ber of copies of the gene (a phenomenon acquisition of further genetic changes [4]. called “amplification” [3]). Activation of an The combined activation of oncogenes > Oncogenes stimulate cell proliferation oncogene is a dominant mechanism, since and inactivation of tumour suppressor and may be overexpressed by gene alteration of a single allele is sufficient to genes drive the progression of cancer. The amplification (e.g. MYC). In addition, confer a gain of function for cancer onset most evident biological consequences of oncogenes may be activated by muta- or progression. The non-activated coun- these alterations are autonomous cell pro- tions (e.g. the RAS gene family). terpart of an oncogene is sometimes liferation, increased ability to acquire called a “proto-oncogene”. A proto-onco- genetic alterations due to deregulated > Tumour suppressor genes are typically gene is in fact a “normal” gene in all DNA repair, ability to grow in adverse con- inactivated by gene mutations in one respects, often with important functions ditions due to decreased apoptosis, allele (gene copy), followed by loss of the intact allele during cell replication in the control of the signalling of cell pro- (Apoptosis, p113) capacity to invade tis- (two-hit mechanism). This leads to loss liferation, differentiation, motility or sur- sues locally and to form distant metas- of expression and abolition of the sup- vival. tases, and ability to activate the formation pressor function, which is particularly A tumour suppressor gene is a gene of new blood vessels (a process called important in cell cycle control. whose alteration during carcinogenesis angiogenesis). Together, these five biolog- results in the loss of a functional property ical phenomena may be caricatured as >Mutational inactivation of suppressor essential for the maintenance of normal pieces of the “cancer jigsaw” [5] (Fig. genes in germ cells is the underlying cell proliferation. Loss of function of a 3.15). None alone is sufficient in itself, but cause of most inherited tumour tumour suppressor gene is typically a cancer arises when they interact together syndromes. The same type of mutation recessive mechanism. Indeed, in many into a chain of coordinated events that may arise through mutations occurring during an individual’s lifetime. instances both copies of the gene need to profoundly modifies the normal cellular be inactivated in order to switch off the pattern of growth and development. corresponding function. Inactivation of tumour suppressor genes proceeds by loss of alleles (most often through the loss Definitions of entire chromosomal sections encom- The multi-step nature of carcinogenesis passing several dozen genes), small dele- has long been recognized (Multistage car- tions or insertions that scramble the read- Genetic Invasiveness cinogenesis, p84). Over the past 20 years, ing frame of the gene, transcriptional instability experimental studies in animals and silencing by alteration of the promoter molecular pathological studies have con- region, or point mutations that change the verged to establish the notion that each nature of residues that are crucial for the step in malignant transformation is deter- activity of the corresponding protein. Autonomous mined by a limited number of alterations Recently, it has emerged that tumour sup- growth in a small subset of the several thousands pressor genes can be conveniently sub- of cellular genes [1]. The terms “onco- classified into two major groups. The gene” and “tumour suppressor gene” are genes of the first group are nicknamed commonly used to identify the sets of “gatekeepers”. Their products control the genes involved in such sequences of gates on the pathways of cell proliferation. Angiogenesis Unlimited events [2]. Both groups of genes are Typically, gatekeeper genes are negative replicative potential extremely diverse in terms of nature and regulators of the cell cycle, acting as function. An oncogene is a gene whose “brakes” to control cell division. The genes function is activated in cancer. This can be of the second group are called “caretak- Fig. 3.15 The cancer jigsaw: multiple functions achieved by a number of simple molecular ers”, as their primary function is not to must be altered for tumorigenesis to occur.

96 Mechanisms of tumour development WCR-S3.Q (composition) 6/02/03 9:00 Page 97

Common human oncogenes Many common proto-oncogenes encode components of the molecular cascades that regulate the cellular response to mitogenic signals [6]. They include growth factors (e.g. TGFA), growth factor receptors (e.g. the receptors for epidermal growth factor, EGF and its close homologue, ERBB2), receptor-coupled signal transduction molecules (in particular, sev- A B eral small guanosine triphosphate (GTP)- Fig. 3.16 Analysis of the status of the ERBB2 oncogene by fluorescent in situ hybridization (FISH) with a binding proteins located on the inner face rhodamine-labelled ERBB2 probe (pink). In breast tumour cells without amplification of the gene, each of the cell membrane, such as the various nucleus possesses two copies of ERBB2 (A). In tumour cells with high-level amplification of the gene, members of the RAS family), kinases numerous signals are evident in each nucleus (B). (SRC, ABL, RAF1), regulatory subunits of cell cycle kinases (CCND1 and CCNA), gene is located within a region of the with activated tyrosine kinases and act as phosphatases (CDC25B), anti-apoptotic genome which is amplified in about 27% of “amplifiers” to increase the strength of the molecules (BCL2) and transcription fac- advanced breast cancers, leading to a signal generated by the activation of cell- tors (MYC, MYB, FOS, JUN). The cumber- spectacular increase in the density of the surface receptors [8]. In their active form, some nomenclature of these genes (Box: molecule at the cell surface. ERBB2 ras proteins bind guanosine triphosphate Naming genes and proteins, p101) owes encodes a transmembrane protein with (GTP) and catalyse its hydrolysis into much to the way they were discovered and the structure of a cell-surface receptor, the guanosine diphosphate (GDP) returning to identified. The SRC gene, for example, was intracellular portion of which carries a their inactive form. Oncogenic forms of the first oncogene identified, in 1976, as a tyrosine kinase activity. Overexpression of activated RAS genes often carry missense modified version of a cellular gene incor- ERBB2 leads to constitutive activation of mutations at a limited number of codons porated in the genome of a highly trans- the growth-promoting tyrosine phosphory- within the GTP-binding site of the enzyme, formant chicken retrovirus, the Rous sar- lation signal. The elucidation of this mech- making it unable to hydrolyse GTP and thus coma virus. The MYC gene was also origi- anism has led to the development of neu- trapping it in the active form. Activation of nally identified in the genome of an avian tralizing antibodies and specific chemical RAS genes thus induces the cell to behave retrovirus inducing promyelocytic inhibitors of tyrosine kinase activity as as if the upstream, Ras-coupled receptors leukaemia. The RAS genes were first iden- therapeutic approaches to the blocking of were being constantly stimulated. tified as activated genes capable of induc- ERBB2 action. ing the formation of rat sarcomas, and MYC various members of the family were found RAS The MYC oncogene may be seen as a pro- in different murine retroviruses, such as The RAS genes are located one step down- totype of the family of molecules which lies the Harvey sarcoma virus (HRAS) and the stream of ERBB2 in growth signalling cas- at the receiving end of the signal transduc- Kirsten sarcoma virus (KRAS). cades. The protein products of the RAS tion cascades. MYC encodes a transcrip- The most commonly activated oncogenes genes are small proteins anchored at the tion factor which is rapidly activated after in human cancers are ERBB2 (in breast cytoplasmic side of the plasma membrane growth stimulation and which is required and ovarian cancers), members of the RAS by a lipidic moiety. They indirectly interact for the cell to enter into cycle [9]. family (in particular KRAS in lung, colorectal and pancreatic cancers, and MYC (in a large variety of tumours such as cancers of the breast and oesophagus and in some forms of acute and chronic leukaemia). These three examples give an excellent illustration of the diversity of the mechanisms of oncogene activation and of their consequences for cell growth and division.

ERBB2 A B In the case of ERBB2, oncogenic activation Fig. 3.17 In cell cultures, activation of a single oncogene may result in a changed morphology from is almost always the result of amplification “normal” (A) to “transformed” (B) and this often corresponds to a change in growth properties. Malignant of the normal gene [7] (Fig. 3.16). This transformation appears to require the co-operation of at least three genes.

Oncogenes and tumour suppressor genes 97 WCR-S3.Q (composition) 6/02/03 9:00 Page 98

Normal Myc transactivates a number of other cel- Tumour suppressor genes: history of a chromosome 13 lular genes and has a wide spectrum of concept Parent Normal heterozygous Parent molecular effects (a phenomenon that Whereas the study of retroviruses and for RB1 RB1 may explain why Myc is activated in many gene transfection experiments were the different types of cancer cells). keys to the discovery of oncogenes, Chromosome 13 with deletion Activation of Myc often proceeds through tumour suppressor genes were identified of RB1 amplification of the region containing the through the study of large DNA viruses Child heterozygous gene on chromosome 8, but Myc is also and the analysis of familial tumour syn- for RB1 commonly activated by chromosomal dromes. translocation in some forms of B-cell Somatic mutation with high frequency in retinal cell with loss of leukaemia (Leukaemia, p242). Retinoblastoma normal chromosome In 1971, Knudsen proposed the now popu- BCL2 lar “two hits” hypothesis to explain the The BCL2 gene (activated in B cell lym- inheritance of retinoblastoma, a rare phomas) exemplifies another kind of childhood tumour type [11,12] (Genetic oncogene. Initially identified as a gene susceptibility, p71). He postulated that, in located within a chromosomal breakpoint a familial setting, individuals may inherit Proliferation in some forms of leukaemia, BCL2 was only one normal copy of the gene (local- found to encode a protein capable of ized by linkage studies to chromosome extending the life span of a cell by pre- 13q14), the other being either lost, par- venting the onset of programmed cell tially deleted or otherwise inactivated. death, or apoptosis [10] (Apoptosis, Consequently, these individuals would just p113). Biochemical studies have revealed need one additional mutagenic step to Nonmalignant Proliferating retinoblastoma cells cells that BCL2 encodes a regulator of the per- switch off the remaining copy of the gene, meability of the mitochondrial mem- thus totally losing the corresponding func- brane. Mitochondrial damage and cyto- tion (Fig. 3.18). The very same type of can- Fig. 3.18 The retinoblastoma gene is a paradigm plasmic leakage of mitochondrial compo- cer may also occur in a sporadic manner, for tumour suppressor genes: if a child inherits a mutation or deletion of one copy (“allele”) of the nents is one of the important signals that but in this case it would require two con- retinoblastoma gene, the remaining normal copy lead a cell to apoptosis. By helping to secutive “hits” (mutagenic events) to inac- tends to be lost at a high frequency in cells of the keep the mitochondrial permeability tivate the two copies of the gene in the retina, resulting in loss of function and in the for- pores closed, Bcl-2 protein prevents this same cell. This theory paved the way for mation of tumours. The diagram shows loss of the whole normal chromosome but the normal allele leakage and thus allows the survival of the modern concept of recessive tumour can also be lost by mutation, deletion, gene con- cells that would otherwise have been suppressor genes. In 1988, the gene version or mitotic recombination. eliminated by a physiological process. responsible for familial retinoblastoma was identified [13]. The RB1 gene encodes a protein that binds and inactivates tran- Cytotoxic scription factors that are essential for the drugs progression of the cell cycle, thus fulfilling UV Cytokines the functions of a molecular “brake” on X-rays Ribonucleotide cell division. γ-rays depletion Hypoxia Microtubule depletion Large DNA viruses In parallel with events previously outlined, Redox changes Growth factor depletion it became evident that many DNA viruses Temperature? Senescence associated with cancer encode complex viral proteins that are capable of seques- Activation or accumulation of p53 protein tering and inactivating cellular proteins [14]. This is the case of a tumorigenic Induction of p53 target genes Binding to p53-interacting proteins simian virus, SV40, of several adenoma and polyoma viruses and of oncogenic forms of human papillomaviruses. In the case of SV40, the virus encodes a large protein (called LT for Large Tumour anti- Fig. 3.19 Many types of biological stress lead to a p53-mediated response. gen) which binds two cellular proteins, the

98 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 99

product of the RB1 gene (pRb) and an molecules and transcription factors that Others ubiquitous protein that was conservatively regulate cell proliferation. Loss of APC In the case of hereditary Wilms tumours, a called p53. In the case of oncogenic function sets these transcription factors rare type of kidney cancer, the gene identi- human papillomaviruses, the viruses free, an event that favours not only the fied encodes a protein essential for the cor- encode two distinct proteins, E7 (which formation of polyps but also their trans- rect differentiation of the nephron. This very neutralizes pRb) and E6 (which neutralizes formation into adenomas and carcino- specific role may explain why the hereditary p53). Thus it was suggested that pRb and mas. loss of this gene does not seem to be asso- p53 might have similar, complementary ciated with cancers at any other site. functions, operating jointly in the control Breast cancer of cell division. Two genes have been identified as This short overview gives only a few exam- The “missing link” in this conceptual edi- involved in familial breast cancer risk, ples of the diversity of tumour suppressor fice was the discovery of alterations in the BRCA1 and BRCA2 [18]. These genes genes, and there is little doubt that many gene encoding p53. This was achieved in encode large proteins with complex still remain to be identified. Given the 1989, when it emerged that the p53 gene functions in many aspects of cell regu- breadth of the concept of “tumour suppres- was often mutated and/or deleted in lation, such as cell cycle control and sors”, many genes encoding components of many forms of cancers [15]. In 1991, DNA repair. However, how their inacti- stress response pathways have the poten- inherited loss of p53 was found to be vation contributes to the onset or tial to behave in this fashion (as their alter- associated with a rare familial syndrome development of breast cancer is still ation may prevent cells from mounting an of multiple cancers, the Li-Fraumeni syn- largely unknown. adequate response to genotoxic, potentially drome, in which afflicted family members suffer vastly increased incidence of many tumour types [16]. Today, about 215 families worldwide affected by this syndrome have been described and the p53 mutations they exhibit are compiled in a database maintained at IARC.

Tumour suppressor genes and familial cancer syndromes Most familial cancer syndromes are inherited as a recessive trait, and cor- A B respond to the constitutive inactivation Fig. 3.20 Accumulation of p53 in human epidermis after exposure to sunlight. Unexposed skin shows no immunostaining against p53 protein (A). Exposed skin (B) shows a dense dark nuclear coloration of epi- of an important tumour suppressor dermal cells due to positive immunostaining for p53 protein. gene, as described above in the case of familial retinoblastoma. Over the past 15 years, many loci containing tumour

suppressor genes have been identified Transcriptional activation by linkage studies in cancer-prone fam- Transcriptional repression ilies. Protein interactions

Colorectal cancer In colorectal cancers, two different famil- Bcl-2 σ p21 ial cancer syndromes have been found to 14 - 3 -3 p53R2 Bax Unknown be associated with the constitutive alter- RPA IGF/BP3 proteins? TFIIH Killer/DR5 ation of two distinct sets of tumour sup- PAG608 pressor genes (Colorectal cancer, p198). PIG3 Patients with familial adenomatous poly- cdc25 Cdk PCNA Replication/ posis, a disease that predisposes to the G2 G1 G1/S Transcription/ early occurrence of colon cancer, often Cell cycle arrest Apoptosis carry alterations in one copy of the ade- Repair nomatous polyposis coli (APC) gene [17]. This gene plays a central role in a signalling cascade that couples cell-surface receptors, calcium-dependent adhesion Fig. 3.21 Multiple response pathways are triggered by the accumulation of p53 in the cell nucleus.

Oncogenes and tumour suppressor genes 99 WCR-S3.Q (composition) 27/01/03 9:27 Page 100

expression of several dozen genes involved in cell cycle control, in the induction of apoptosis, in DNA repair and in differentiation control. Together these genes exert complex, anti-proliferative effects (Fig. 3.21). Essentially, when cells are subjected to tolerable levels of DNA-damaging agents, activation of p53 will result in cell cycle arrest, temporarily removing the cells from the proliferative pool or mediating differentiation. However, when faced with highly damaging levels of genotoxic stress, p53 will induce apoptosis, a programmed form of suicide that eliminates cells with potentially oncogenic alterations. This complex role in the protection of the cell from DNA damage has resulted in p53 being described as the “guardian of the genome” [20]. Loss of this function by mutation, as often occurs during carcino- Fig. 3.22 Molecular modelling of part of the p53 protein (DNA-binding domain), showing its interaction genesis, will allow cells with damaged with DNA. The amino acids labelled (arginine 175, 248, 273) are important for maintaining biological activity and are among the “hotspots” for mutations in cancer. The zinc atom is required for stabilizing DNA to remain in the proliferative popula- the complex three-dimensional structure of the p53 oligomer. tion, a situation that is essential for the expansion of a clone of cancer cells. The p53 gene differs from most other CDKN2A/ tumour suppressors in its mode of inacti- INK4A vation in human cancers. Whereas most Exon 1 β Exon 1 α Exon 2 Exon 3 gene tumour suppressors are altered by loss of alleles or inactivating deletions or insertions, p53 is commonly the target of p16INK4A point mutations within the portion of the Inhibitor of gene that encodes the DNA-binding cyclin D/CDK4 complexes domain of the protein (Fig. 3.22). These mutations prevent the correct folding of p14ARF this protein domain, and therefore dis- Inhibitor of p53- Mdm2 complex rupt the interactions of p53 with its spe- formation cific DNA targets. However, the mutant Fig. 3.23 Generally, a single segment of DNA codes for a single protein. However, the p16 and p14ARF proteins are often extremely stable and proteins are both encoded by a single region of DNA. P = promoter. therefore accumulate to high levels within the nucleus of cancer cells. This accu- oncogenic forms of stress). The genes are also mutated at variable rates in many mulated protein can often be detected by responsible for complex inherited diseases forms of sporadic cancer. However, two of immunohistochemistry in primary tumours such as ataxia telangiectasia or xeroderma them, p53 and CDKN2A, are very com- as well as in distant metastases. Although pigmentosum (Carcinogen activation and monly altered in almost every kind of not all mutations induce accumulation of DNA repair, p89) belong to this category human cancer. the protein, p53 accumulation provides a [19]. Alteration of such genes results in convenient tool for pathologists to assess many defects, including hypersensitivity to p53, the guardian of the genome the possibility of a p53 dysfunction in radiation and therefore to the development The p53 gene encodes a phosphoprotein cancer specimens [21]. of cancers such as skin tumours. of molecular weight 53,000 daltons, Mutation is not the only way to alter p53 which accumulates in the nucleus in protein in cancer. In cervical cancers, Tumour suppressor genes and spo- response to various forms of stress, in p53 gene mutations are infrequent, but radic cancers particular, DNA damage (Fig. 3.20). In this the protein is inactivated by binding of Many of the tumour suppressor genes context, p53 acts as a transcriptional reg- the viral protein E6 which is produced by associated with familial cancer syndromes ulator, increasing or decreasing the human papillomavirus. This protein cre-

100 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 101

ates a molecular bridge between p53 and restore the function of mutant p53 have reading frame” (often called by the same the protein degradation machinery, shown promising results in experimental name as its mouse homologue, p19ARF), is resulting in the rapid degradation and systems. As knowledge of the p53 path- synthesized from a different portion of the effective elimination of p53 protein. This way improves, it is anticipated that this CDKN2A locus but shares one exon (exon interaction plays an important role in cer- central molecular event in human cancer 2) with p16. However, although the DNA vical cancer (Cancers of the female will provide a basis for developing new sequence encoding the two products is reproductive tract, p215). In normal cells, forms of therapy. identical, p16 and p14ARF use different the degradation of p53 is regulated by reading frames of exon 2, such that their the Mdm2 protein. Mdm2 (“murine dou- CDKN2A: one locus, two genes amino acid sequences are completely dif- ble minute gene 2”) was originally identi- CDKN2, or “cyclin dependent kinase ferent. p14ARF is a protein that controls fied in the mouse as the product of a inhibitor 2”, is known under several names, Mdm2, which in turn regulates p53 protein gene amplified in aberrant chromosome including INK4A (inhibitor of kinase 4A) stability. Activation of p14ARF blocks Mdm2 fragments called “double minute chromo- and MTS1 (multiple tumour suppressor 1). and therefore results in p53 protein accu- somes”. Amplification of MDM2 is com- The CDKN2A locus is located at the mulation and activation. Thus the CDKN2A mon in osteosarcomas and is sometimes extremity of the short arm of chromosome locus behaves as two unrelated but inter- detected in other cancers, such as carci- 9, the letter “A” serving to distinguish it locked genes. The first gene, encoding nomas or brain tumours. MDM2 thus from the CDKN2B gene, which is located p16, directly controls cell cycle progres- behaves as an oncogene, since its activa- just 20 kilobases away. sion and senescence. The second, encod- tion by amplification causes the inactiva- This gene is unique in that it contains two ing p14ARF, controls p53 and all its down- tion of a tumour suppressor gene [22]. distinct reading frames, with two different stream anti-proliferative functions. The p53 gene (and its product) is one of promoters, the same DNA being used to The CDKN2A locus is often altered by loss the most studied genes in human cancer. synthesize two proteins that do not have a of alleles (which removes both p16 and In the 20 years since its discovery in single amino acid sequence in common p14ARF), by mutation (most frequently in 1979, more than 15,000 publications [23] (Fig. 3.23). The first reading frame to exon 2, common to both gene products), have addressed its structure, function be discovered encodes p16, an inhibitor of and by hypermethylation. Increased and alteration in cancers. There have cyclin-dependent kinases 4 and 6, which methylation of specific regions of the DNA been many attempts to exploit this associates with cyclin D1 in G1 phase of within the promoters and some of the cod- knowledge in the development of new the cell cycle (The cell cycle, p104). The ing regions prevents adequate transcrip- therapies based on the control of p53 p16 protein is thus an archetypal cell cycle tion and decreases the levels of protein activity in cancer cells. Experimental “brake”, its loss leading to increased cell synthesized. Loss of expression due to gene therapy has shown that it may be proliferation and, more specifically, to hypermethylation may be the most fre- possible to restore p53 function in cells escape from replicative senescence and quent way of altering the CDKN2A locus in that have lost the gene. More recently, extended cellular life span. The other read- many forms of cancers, particularly carci- drugs designed to specifically target and ing frame, named p14ARF for “alternative nomas.

NAMING GENES AND ative of the successful investigator. inhibitor WAF1 is also known as PROTEINS Identification of a novel gene is often fol- CDKN1A, CAP20, MDA-6, PIC-1 and SDI- lowed by discovery of structurally related 1. In scientific writing, all such names Conventionally, a gene (that is, a specif- or “homologous” genes (and correspon- are given in the first instance, after ic segment of DNA) is identified by a sin- ding proteins) and these may be given which a single name is used consistent- gle name, in upper case and italicized names closely related to the first member ly in any one document. The latest esti- (e.g. the oncogene RAS) that is indica- of the “family” identified. Such an mates from the Human Genome Project tive of the character or function of the approach to nomenclature may be inade- suggest that there may be about 30,000 protein encoded, which is designated by quate, for example, in those instances in human genes. the same name, in lower case (ras in the which a single DNA segment encodes Conventions for the naming of genes case of the present example). Proteins multiple proteins through alternative and proteins are subject to international are also named by reference to their splicing of messenger RNA. Multiple agreement and are continuously subject molecular weight, with the correspon- names for the same gene or protein may to review (HUGO Gene Nomenclature ding gene in superscript (e.g. p21WAF1 ). arise because of independent discovery Committee, http://www.gene.ucl.ac.uk/ The names of genes are often based on (and hence, naming) by different investi- nomenclature/). acronyms, and are generally the prerog- gators. Thus, the cyclin-dependent kinase

Oncogenes and tumour suppressor genes 101 WCR-S3.Q (composition) 27/01/03 9:27 Page 102

GEOGRAPHIC VARIATION IN MUTATION PATTERNS

Mutations in cancer genes are the direct consequence of attack on DNA by exogenous or endogenous agents or of errors in DNA repair systems. By analysing the type and the sequence context of such mutations, it is possible to form hypothe- ses regarding the nature of the mutagenic mechanism involved. The most interest- ing genes in this respect are those altered by missense point mutations, such as members of the RAS family, CDKN2A/ INK4A, and, in particular, the p53 gene.

The p53 gene is the most frequently mutated gene in human cancer, with over 16,000 mutations reported and compiled Fig. 3.24 Geographic variations in the prevalence of p53 gene mutations in breast cancers. in a database maintained at IARC (http://www.iarc.fr/p53). The diversity of these mutations allows the identification class of mycotoxins which contaminates ences have been observed between geo- of patterns which vary depending on the traditional diets (groundnuts) (Food con- graphical areas, which may provide infor- tumour type, the geographic origin and taminants, p43). Experiments in animals mation on the nature of risk factors the risk factors involved. These are often and in cell culture have shown that aflatox- involved. specific for particular agents that have ins can directly induce the mutation at caused these mutations. Thus p53 gene codon 249. This particular mutation is not In many other cancers, mutation patterns mutations in cancers may be seen as “fin- found in liver cancers in areas of the world, also vary from one region of the world to gerprints” left by carcinogens in the such as the USA, where exposure to afla- another. This variability may give clues human genome, which may help to identi- toxins is low. about the genetic heterogeneity of popu- fy the particular carcinogen involved. lations, as well as about the diversity of Specific mutations have also been agents involved in causing cancers. For A typical example of such a “fingerprint” observed in lung cancers from smokers example, in oesophageal cancers, muta- is the mutation at codon 249 observed in (due to tobacco carcinogens). In skin can- tion types widely differ between high-inci- liver cancers of patients from sub- cers, the mutations bear typical chemical dence and low-incidence regions, sug- Saharan Africa and Eastern Asia. In these signatures of the damage inflicted to DNA gesting that specific mutagens are at regions, liver cancer is a consequence of by exposure to solar ultraviolet radiation. In work in causing the excess incidence chronic infection by hepatitis viruses and other instances, exemplified by patterns of seen in some parts of the world, such as of dietary poisoning with aflatoxins, a mutation in breast cancer, marked differ- Northern Iran and Central China.

Lesions in the p16INK4A-cyclin D, CDK4-pRb cancer types have been characterized. cancer process. Moreover, many biolog- and p14ARF-Mdm2-p53 pathways occur so Most of these have a powerful impact ical alterations leading to cancer may frequently in cancer, regardless of patient on tumour growth. However, it is very not be detectable at the DNA level. age or tumour type, that they appear to be likely that many critical genes with less Cancer-causing changes may result fundamental to malignancy [24]. penetrant phenotypes remain to be from modification of RNA levels or pro- identified. In particular, the genes cessing, and of protein structure and Prospects for the molecular analysis of involved in stress reponses, in the con- function through a variety of epigenetic cancer trol of oxygen metabolism and in the phenomena. The systematic profiling of More than 200 genes that are altered at detoxification of xenobiotics are all can- gene expression in cancer cells will variable proportions in different human didates for a role as cofactors in the probably reveal a whole new set of

102 Mechanisms of tumour development 083 A 125 WORLD CANCER 27/06/03 12:05 Page 103

potential cancer genes altered through tory of protein structure, dynamics and provide a basis for the development of less conspicuous mechanisms. Apart interactions is still largely unexplored. new means for treatment and diagnosis. from expression studies, the whole terri- An understanding of such processes will

REFERENCES WEBSITES 1. Fearon ER, Vogelstein B (1990) A genetic model for colo- share the ability to relieve the cell's requirement for cyclin American Tissue Type Collection: rectal tumorigenesis. Cell, 61: 759-767. D1 function in G1. J Cell Biol, 125: 625-638. http://www.atcc.org/ 2. Weinberg RA (1995) The molecular basis of oncogenes 15. Baker SJ, Fearon ER, Nigro JM, Hamilton SR, Centers for Disease Control and Prevention, Atlanta: and tumor suppressor genes. Ann N Y Acad Sci, 758: 331- Preisinger AC, Jessup JM, vanTuinen P, Ledbetter DH, http://www.cdc.gov/ 338. Barker DF, Nakamura Y (1989) Chromosome 17 deletions Cancer Genome Anatomy Project: 3. Savelyeva L, Schwab M (2001) Amplification of onco- and p53 gene mutations in colorectal carcinomas. http://www.ncbi.nlm.nih.gov/ncicgap/ Science, 244: 217-221. genes revisited: from expression profiling to clinical appli- European Bioinformatics Institute: cation. Cancer Lett, 167: 115-123. 16. Birch JM (1994) Li-Fraumeni syndrome. Eur J Cancer, http://www.ebi.ac.uk/ 30A: 1935-1941. 4. Kinzler KW, Vogelstein B (1997) Cancer-susceptibility HotMolecBase (Weizmann Institute, Israel): genes. Gatekeepers and caretakers. Nature, 386: 761, 763. 17. Bresalier RS (1997) The gatekeeper has many keys: http://bioinformatics.weizmann.ac.il/hotmolecbase/ 5. Hanahan D, Weinberg RA (2000) The hallmarks of can- dissecting the function of the APC gene. Gastroenterology, 113: 2009-2010. IARC p53 database: cer. Cell, 100: 57-70. http://www.iarc.fr/p53/ 18. Zheng L, Li S, Boyer TG, Lee WH (2000) Lessons 6. Hunter T (1991) Cooperation between oncogenes. Cell, Kyoto encyclopedia of genes and genomes: 64: 249-270. learned from BRCA1 and BRCA2. Oncogene, 19: 6159- 6175. http://www.genome.ad.jp/kegg/kegg.html 7. Hung MC, Lau YK (1999) Basic science of HER-2/neu: OMIM (Online Mendelian Inheritance in Man): a review. Semin Oncol, 26: 51-59. 19. Shiloh Y, Rotman G (1996) Ataxia-telangiectasia and the ATM gene: linking neurodegeneration, immunodeficien- http://www3.ncbi.nlm.nih.gov/omim/ 8. Wiesmuller L, Wittinghofer F (1994) Signal transduction cy, and cancer to cell cycle checkpoints. J Clin Immunol, Protein Data Bank (a protein structure database): pathways involving Ras. Mini review. Cell Signal, 6: 247- 16: 254-260. http://www.rcsb.org/pdb/ 267. 20. Lane DP (1992) Cancer. p53, guardian of the 9. Henriksson M, Luscher B (1996) Proteins of the Myc genome. Nature, 358: 15-16. network: essential regulators of cell growth and differentiation. Adv Cancer Res, 68: 109-182. 21. Hainaut P, Hollstein M (2000) p53 and human cancer: the first ten thousand mutations. Adv Cancer Res, 77: 10. Antonsson B, Martinou JC (2000) The Bcl-2 protein 81-137. family. Exp Cell Res, 256: 50-57. 22. Momand J, Jung D, Wilczynski S, Niland J (1998) The 11. Knudson AG (1996) Hereditary cancer: two hits revis- MDM2 gene amplification database. Nucleic Acids Res, ited. J Cancer Res Clin Oncol, 122: 135-140. 26: 3453-3459. 12. Knudson AG, Jr. (1971) Mutation and cancer: statisti- 23. Chin L, Pomerantz J, DePinho RA (1998) The cal study of retinoblastoma. Proc Natl Acad Sci U S A, 68: INK4a/ARF tumor suppressor: one gene--two products-- 820-823. two pathways. Trends Biochem Sci, 23: 291-296. 13. Bartek J, Bartkova J, Lukas J (1997) The retinoblas- 24. Sherr CJ (2000) The Pezcoller lecture: cancer cell toma protein pathway in cell cycle control and cancer. Exp cycles revisited. Cancer Res, 60: 3689-3695. Cell Res, 237: 1-6. 14. Lukas J, Muller H, Bartkova J, Spitkovsky D, Kjerulff AA, Jansen-Durr P, Strauss M, Bartek J (1994) DNA tumor virus oncoproteins and retinoblastoma gene mutations

Oncogenes and tumour suppressor genes 103 WCR-S3.Q (composition) 27/01/03 9:27 Page 104

THE CELL CYCLE

tion of cellular materials between the malian cells, these mutations would have SUMMARY daughter cells. Moreover, immediately been lethal and it would therefore have > The control of cell division is critical to before or after the cell cycle, various fac- been impossible to characterize them. normal tissue structure and function. It tors interact to determine whether the cell These mutants were called “cdc”, for cell is regulated by a complex interplay of divides again or whether the cell becomes division cycle mutants, and many of them many genes that control the cell cycle, committed to a programme of differentia- have been accorded wider recognition with DNA replication (S phase) and tion or of cell death. Therefore, the term through the application of their names to mitosis as major checkpoints. “cell cycle” is often used in a broad sense the mammalian homologues correspon- > The cell cycle is tightly regulated to to refer to, as well as the basic, self-repli- ding to the yeast genes. minimize transmission of genetic damage cating cellular process, a number of con- to subsequent cell generations. nected processes which determine pre- > Progression through the cell cycle is and post-mitotic commitments. These primarily controlled by cyclins, asso- may include the commitment to stop ciated kinases and their inhibitors. dividing in order to enter a quiescent Retinoblastoma (RB) and p53 are major state, to undergo senescence or differen- suppressor genes involved in the G1/S tiation, or to leave the quiescent state to checkpoint control. re-enter mitosis. > Cancer may be perceived as the consequence of loss of cell cycle control and Molecular architecture of the cell cycle progressive genetic instability. The molecular ordering of the cell cycle is a complex biological process dependent upon the sequential activation and inactivation of molecular effectors at specific points of the cycle. Most current knowl- Classically, the “cell cycle” refers to the edge of these processes stems from set of ordered molecular and cellular experiments carried out in the oocyte of processes during which genetic material the frog, Xenopus laevis, or in yeast, either Fig. 3.25 Proliferating cells in the basal parts of is replicated and segregates between two Saccharomyces cerevisiae (budding yeast) the colonic crypts, visualized by immunohistochemistry (stained brown). newly generated daughter cells via the or Schizosaccharomyces pombe (fission process of mitosis. The cell cycle can be yeast). The Xenopus oocyte is, by many cri- divided into two phases of major morpho- teria, one of the easiest cells to manipulate logical and biochemical change: M phase in the laboratory. Its large size (over a mil- (“mitosis”), during which division is evi- limetre in diameter) means that cell cycle dent morphologically and S phase (“syn- progression can be monitored visually in thesis”), during which DNA is replicated. single cells. Microinjections can be per- These two phases are separated by so- formed for the purpose of interfering with called G (“gap”) phases. G1 precedes S specific functions of the biochemical phase and G2 precedes M phase. machinery of the cell cycle. The Xenopus During progression through this division oocyte has proven to be an invaluable tool cycle, the cell has to resolve a number of in the study of the biochemistry of the cell critical challenges. These include ensuring cycle, allowing, among other findings, the that sufficient ribonucleotides are avail- elucidation of the composition and regula- able to complete DNA synthesis, proof- tion of maturation promoting factor (MPF), reading, editing and correcting the newly- a complex enzyme comprising a kinase synthesized DNA; that genetic material is (p34cdc2) and a regulatory subunit (cyclin Fig. 3.26 A human osteosarcoma cell nucleus not replicated more than once; that the B), which drives progression from G2 to M during mitosis. Cell division proceeds clockwise from upper right through interphase, prophase spatial organization of the mitotic spindle phase [1]. In contrast, the exceptional (centre), prometaphase, metaphase, anaphase apparatus is operational; that the packing genetic plasticity of yeast has allowed the and telophase. During the cycle, the chromo- and the condensation of chromosomes is identification of scores of mutants with somes are replicated, segregated and distributed optimal; and that there is equal distribu- defects in cell cycle progression; in mam- equally between the two daughter cells.

104 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 105

One of the earliest genes to be identified the activity of cyclin/CDK complexes. functional properties: the WAF1/CIP1 in this way was cdc2. Isolated in S. Downstream of cyclin/CDKs are effector family (p21), the KIP family (p27, p57) and pombe, cdc2 was determined to be able molecules, essentially transcription fac- the INK4 family (p16, p15, p18) (Fig. to correct a G2 cell cycle arrest defect. tors, which control the synthesis of pro- 3.27). The product of this gene, a serine-threo- teins that mediate the molecular and cel- Downstream effectors of cyclin/CDKs nine kinase of molecular weight 32- lular changes occurring during each include proteins mediating three main 34,000 daltons, was subsequently shown phase. functional categories: (1) those involved to be the yeast homologue of the kinase CDKIs are small proteins that form com- in the control of the enzymes responsible contained in the Xenopus MPF. This plexes with both CDKs and cyclins [3]. for DNA replication, proof-reading and enzyme became the paradigm of a class Their role is primarily to inhibit the activi- repair, (2) those involved in chromosome of enzymes now called cyclin-dependent ties of cyclin/CDK complexes and to neg- and chromatin remodelling and in the kinases, or CDKs. In their active form, atively regulate cell cycle progression. control of genomic integrity and (3) those CDKs form heterodimers with cyclins, a They constitute the receiving end of many involved in the mechanics of cell division class of molecules synthesized in a time- of the molecular cascades signalling (including the formation of the centro- dependent manner during the cell cycle. growth promotion or suppression of some and the mitotic spindle, and in the The progression of the cell cycle depends growth. Thus CDKIs may be considered as resorption of the nuclear membrane). upon the sequential activation and inacti- the interface between the cell cycle These processes require the coordinated vation of cyclin/CDK complexes [2], a machinery and the network of molecular synthesis of hundreds of cellular pro- process which requires the synthesis of pathways which signal proliferation, teins. Transcription factors of the E2F cyclins, the formation of a complex death or stress responses. However, by family play a critical role in the control of between a specific cyclin and a CDK and virtue of their complexing properties, gene transcription during cell cycle pro- post-translational modification of the some CDKIs also play a positive role in gression (Fig. 3.28). In G1, factors of the CDK to convert the enzyme to an active cell cycle progression by facilitating the E2F family are bound to their DNA tar- form (Fig. 3.27). assembly of cyclin/CDK complexes. For gets but are maintained in a transcrip- Progression through the cell cycle as example, p21, the product of the CDKN1A tionally inactive state by the binding of mediated by cyclins is, in turn, deter- gene (also known as WAF1/CIP1), pro- proteins of the retinoblastoma (pRb) pro- mined by factors categorized as having motes the assembly of cyclin D/cdk2 tein family. At the G1/S transition, the either regulatory (upstream) or effector complexes in G1 at a stoichiometric 1:1 sequential phosphorylation of pRb by (downstream) roles. Upstream of ratio, but inhibits the activities of these several cyclin/CDKs dissociates pRb cyclin/CDKs are regulatory factors complexes when expressed at higher lev- from the complexes, allowing E2Fs to called cyclin-dependent kinase inhibitors els. There are three main families of interact with transcription co-activators (CDKIs), that regulate the assembly and CDKIs, each with distinct structural and and to initiate mRNA synthesis [4].

Gene (chromosome) Product Type of alteration Role in cell cycle Involvement in cancer

p53 (17p13) p53 Mutations, deletions Control of p21, 14-3-3σ, etc. Altered in over 50% of all cancers

CDKN2A (9p22) p16 and p19ARF Mutations, deletions, Inhibition of CDK4 and 6 Altered in 30-60% of all hypermethylation cancers

RB1 (13q14) pRb Deletions Inhibition of E2Fs Lost in retinoblastomas, altered in 5-10% of other cancers.

CCND1 Cyclin D1 Amplification Progression into G1 10-40% of many carcinomas

CDC25A, CDC25B cdc25 Overexpression Progression in G1, G2 10-50% of many carcinomas

KIP1 p27 Down-regulation Progression in G1/S Breast, colon and prostate cancers

Table 3.3 Cell cycle regulatory genes commonly altered in human cancers.

The cell cycle 105 WCR-S3.Q (composition) 27/01/03 9:27 Page 106

In early G2, cdk1 is in an inactive form. Its activation requires first association with cyclin B, followed by post-translational modification of the kinase itself. WAF1/CIP1 This modification includes phosphorylation of a conserved threonine residue CDK2A (Thr161) by a kinase complex called CAK (CDK-activating kinase), as well as CDK2B dephosphorylation of two residues local- WAF1/CIP1 ized within the active site of the enzyme, INK4D a threonine (Thr14) and a tyrosine (Tyr15). The removal of these phosphate WAF1/CIP1 groups is carried out by the dual-specificity phosphatases of the cdc25 group, KIP1 WAF1/CIP1 comprising three isoforms in humans (A, B and C). Activation of these phos- KIP2 WAF1/CIP1 phatases is therefore crucial for the activation of cyclin B/cdk1 complexes. The KIP1 phosphatase is directly controlled by a number of regulators, including plk1 KIP2 (polo-like kinase), an activating kinase, Fig. 3.27 The progression of the cell cycle depends upon the sequential activation and inactivation of pp2A, (protein phosphatase 2A), an cyclin/CDK complexes. This process requires the synthesis of cyclins, the formation of a complex inhibitory phosphatase and 14-3-3s, a between a specific cyclin and a CDK, and modification of the CDK to convert this enzyme to an active form. The enzyme’s activity may be disrupted by a specific inhibitor, a CDKI. signal transduction molecule which complexes with cdc25, sequesters it in the cytoplasm and thus prevents it from Through this mechanism, E2Fs exert a germ cells during the second division of dephosphorylating its nuclear targets. Of dual function both as transcriptional meiosis: cells that have undergone the course, the action of cdc25 phos- repressors in G1, when bound to pRb, and first, asymmetric division of the meiotic phatases is counteracted by kinases that as transcriptional activators in G1/S and cycle arrest in G2 until completing the sec- restore the phosphorylation of Thr14 and in S phase, after dissociation of pRb from ond division, which is triggered by fertiliza- Tyr15, named wee1 and mik1 [6]. the complex. Recent observations suggest tion. This concept of “cell cycle check- Following the activation process out- that transcriptional repression by E2Fs is points” was later extended to all mam- lined above, the cyclin B/cdk1 complex essential to prevent the premature activa- malian cells [5]. It is now common to envis- is potentially able to catalyse transfer of tion of cell cycle effectors, which would age the mammalian cell cycle as a succes- phosphates to substrate proteins. scramble the temporal sequence of sion of checkpoints that have to be negoti- However, in order to achieve this, it has molecular events and preclude cell cycle ated in order for division to be achieved. to escape the control exerted by CDKIs, progression. There is no clear agreement on how many such as p21. The function of this CDKI is such checkpoints exist in the mammalian itself controlled by several activators, Cell cycle checkpoints cell cycle, or on their exact position. including BRCA1, the product of a breast The notion of “cell cycle checkpoints” is cancer susceptibility gene (Oncogenes also derived from early studies in Xenopus Control of cdk1 at G2/M transition and tumour suppressor genes, p96). The oocytes and in yeast mutants. In S. cere- The regulation of the complex between p21 protein is removed from the com- visiae, commitment to the mitotic cycle cdk1 (also called p34cdc2) and cyclin B plex by a still poorly understood phos- requires the crossing of a “restriction exemplifies how different factors co-oper- phorylation process, which also drives point” called the start transition. Failure to ate to control the activation of rapid degradation of the protein by the cross this transition results in cells being cyclin/CDK complexes at a cell cycle proteasome. This leaves the cyclin blocked in the G1 phase of the cycle. checkpoint. This activation process B/cdk1 complex ready to function, after Another control point has been clearly requires co-operation between three lev- a final step of autophosphorylation, in identified after S phase, at the transition els of regulation: association between the which cdk1 phosphorylates cyclin B. The between G2 and M phases. Cells unable to two partners of the complex, post-trans- complex is now fully active and ready to cross this checkpoint may remain blocked lational modifications of the kinase and of phosphorylate many different sub- in a pre-mitotic, tetraploid state. the cyclin, and escape from the negative strates, such as nuclear lamins, during Physiologically, this checkpoint is active in regulation exerted by the CDKIs. entry into mitosis.

106 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 107

Regulation of the cell cycle and control of genetic stability During the cell cycle, a number of potential problems may result in damage to the genome. These problems may arise at three distinct stages: (1) during DNA replication, especially if the cell is under conditions of stress that favour the formation of DNA damage (irradiation, exposure to carcinogens etc.), (2) following the termination of DNA replication, when the cell effective- ly “switches off” its DNA synthesis machinery and (3) during M phase, when the cell has to negotiate the delicate task of segre- gating chromatids equally. A tight coupling between these processes and cell cycle regulation is therefore crucial to allow the cell to pause during the cell cycle in order to afford the time necessary for the successful completion of all the operations of DNA and chromosome maintenance. Fig. 3.28 Progression from G1 to S phase is regulated by phosphorylation of the retinoblastoma protein Failure to do this may result in both genet- (pRb), in the absence of which DNA replication cannot proceed. ic and genomic instabilities, which are hallmarks of cancer. Genetic instability is char- The cell cycle and cancer pressor activity remains to be deter- acterized by an increased rate of gene Genes involved in cell cycle control are mined). Unlike the CDKN2A/INK4A gene, mutation, deletion or recombination important among those subject to the the CDKN1A gene (encoding p21) is rarely (essentially due to defects in DNA repair). genetic alterations that give rise to can- disrupted in cancer. As p21 plays many Genomic instability results in chromosome cer [8]. However, the proliferation of can- roles in the negative regulation of almost translocations, loss or duplication of large cer cells requires that the cells retain all phases of the cell cycle, loss of this chromosome fragments and aberrant chro- functional cell cycle processes. The cell function might be expected to result in mosome numbers (aneuploidy). cycle alterations seen in cancer are main- uncontrolled cell division. This is appar- Tens of molecules have been identified as ly confined to two major sets of regula- ently not the case, as mice lacking the components of the signalling cascades tors: those involved in the negative con- CDKN1A gene do not show an increased which couple detection of DNA damage trol of cell cycle progression (inactivation frequency of cancer. This observation and regulation of the cell cycle. One of of which leads to accelerated and illustrates one of the most important char- these is the product of the tumour sup- unchecked cell proliferation) and those acteristics of cell cycle regulatory mecha- pressor gene p53 (Oncogenes and tumour involved in coupling the maintenance of nisms: there is a large degree of redun- suppressor genes, p96). p53 is specifically genome integrity to the cell cycle (inacti- dancy and overlap in the function of any activated after various forms of direct DNA vation of which results in cells having particular effector. Therefore, cancer- damage (such as single or double strand gene alterations that progressively accu- causing deregulation of the cell cycle breaks in DNA) and regulates the transcrip- mulate during carcinogenesis) (Table 3.3) requires a combination of many alter- tion of several inhibitors of cell cycle pro- [9]. Most of the genes corresponding to ations in genes encoding proteins that, gression, particularly at the G1/S and these two categories fall within the group either alone or in concert, are critical for G2/M transitions [7]. Other important mol- of tumour suppressors, and many of them the control of cell division. ecules in this coupling process include the are also direct participants in DNA repair Apart from inactivation of negative regula- checkpoint kinases chk1 and chk2. Chk1 is processes. tors, a few cell cycle genes may be acti- activated after replication blockage during The gene which encodes p16 (CDKN2A/ vated as oncogenes, in that their alter- S-phase. In turn, chk1 activates wee1 and INK4A) has been established as a tumour ation results in enhanced activity leading mik1, two kinases that counteract the suppressor gene [10], and mutations and to accelerated cell proliferation. The best action of cdc25 and keep cdk1 in an inac- deletions at this site are commonly found example of such a cell cycle oncogene is tive form. Thus, through activation of chk1, in primary human tumours, especially CCND1, the gene encoding cyclin D1, a the cell triggers an emergency mechanism melanoma (although the contribution of G1-specific cyclin [11]. This gene is locat- that ensures that cells with incompletely another protein encoded by the same ed on chromosome 11p13, within a large replicated DNA cannot enter mitosis. locus on chromosome 9p, p14ARF, to sup- region that is amplified in up to 20% of

The cell cycle 107 WCR-S3.Q (composition) 27/01/03 9:27 Page 108

several carcinomas (e.g. breast, head and phase cyclin) and for activating mutations extremely diverse range of possibilities for neck, oesophageal and lung cancers). of CDK4 (one of the partners of cyclin D1) cancer-associated alterations. In this There is also limited evidence for tran- in some cancers. Indeed, the high com- respect, cancer can be seen as, funda- scriptional activation of cyclin A (an S- plexity of cell cycle effectors provides an mentally, a disease of the cell cycle.

TELOMERES Telomerase assays have not yet entered AND TELOMERASE routine clinical practice, but there is considerable interest in their possible use for cancer diagnosis and prognosis. For exam- The ends of eukaryotic chromosomes are ple, telomerase assays of urine sediments referred to as telomeres. These contain may be useful for diagnosis of urinary tract many copies of a repetitive DNA cancer (Kinoshita H et al., J Natl Cancer sequence, which in vertebrates is the Inst, 89: 724-730, 1997), and telomerase hexanucleotide TTAGGG. The telomeres of activity levels may be a predictor of out- normal human somatic cells shorten by come in neuroblastoma (Hiyama E et al., 50 to 150 base pairs every time cell divi- Nature Medicine, 1: 249-255, 1995). sion occurs. This appears to act as a cell Fig. 3.29 Telomeres contain repetitive DNA division counting mechanism: when a The catalytic subunit of human telomerase, sequences that cap the ends of chromosomes. Quantitative fluorescence in situ hybridization cell's telomeres have shortened below a hTERT, was cloned in 1997 (Lingner J, Cech analysis of human metaphase chromosome critical length, the cell exits permanently TR, Curr Opin Genet Dev 8: 226-232, 1998). It spreads is shown, using oligonucleotide probes from the cell cycle. Normal cells thus has subsequently been shown that genetic specific for telomere (white) and centromere (red) DNA sequences, and the DNA dye DAPI (blue). have a limited proliferative capacity, and manipulations of hTERT which result in inhi- From the laboratory of Drs J.W. Shay and W.E. this acts as a major barrier against car- bition of telomerase activity in tumour cells Wright. cinogenesis. Cells that have accumulated limit their proliferation and often result in cell some carcinogenic changes are unable to death. This raises the possibility that telom- form clinically significant cancers unless erase inhibitors may be a very useful form of A potential challenge facing telomerase this proliferation barrier is breached. therapy for many or most types of cancer. research is the finding that some cancers More than 85% of all cancers achieve this However, in tumours with long telomeres, it maintain their telomeres by a mechanism by expressing an enzyme, telomerase, may take many cell divisions before telom- that does not involve telomerase, that synthesizes new telomeric DNA to erase inhibitors exert an anti-tumour effect. referred to as alternative lengthening of replace the sequences lost during cell When such drugs are developed they will telomeres, ALT (Bryan TM et al., Nature division (Shay JW, Bacchetti S, Eur J therefore need to be carefully integrated with Medicine, 3: 1271-1274, 1997; Reddel RR, Cancer, 33A: 787-791, 1997). other anticancer treatments. J Clin Invest, 108: 665-667, 2001).

REFERENCES WEBSITES 1. Hunt T (1989) Maturation promoting factor, cyclin and 7. Hainaut P, Hollstein M (2000) p53 and human cancer: Animation of the phases of the cell cycle and of mitosis: the control of M-phase. Curr Opin Cell Biol, 1: 268-274. the first ten thousand mutations. Adv Cancer Res, 77: 81- http://www.cellsalive.com/ 137. 2. Pines J (1995) Cyclins and cyclin-dependent kinases: a Nature Reviews, “Focus on cell division”: biochemical view. Biochem J, 308 (Pt 3): 697-711. 8. Hartwell LH, Kastan MB (1994) Cell cycle control and http://www.nature.com/ncb/celldivision/ cancer. Science, 266: 1821-1828. 3. Sherr CJ, Roberts JM (1999) CDK inhibitors: positive The Forsburg laboratory home pages, a guide to the cell and negative regulators of G1-phase progression. Genes 9. Kinzler KW, Vogelstein B (1997) Cancer-susceptibility cycle and DNA replication in S. pombe: Dev, 13: 1501-1512. genes. Gatekeepers and caretakers. Nature, 386: 761, 763. http://pingu.salk.edu/~forsburg/lab.html 4. Weinberg RA (1995) The retinoblastoma protein and 10. Strohmaier H, Spruck CH, Kaiser P, Won KA, Sangfelt cell cycle control. Cell, 81: 323-330. O, Reed SI (2001) Human F-box protein hCdc4 targets cyclin E for proteolysis and is mutated in a breast cancer 5. Hartwell LH, Weinert TA (1989) Checkpoints: controls cell line. Nature, 413: 316-322. that ensure the order of cell cycle events. Science, 246: 629-634. 11. Schuuring E (1995) The involvement of the chromosome 11q13 region in human malignancies: cyclin D1 and 6. Zeng Y, Forbes KC, Wu Z, Moreno S, Piwnica-Worms H, EMS1 are two new candidate oncogenes--a review. Gene, Enoch T (1998) Replication checkpoint requires phospho- 159: 83-96. rylation of the phosphatase Cdc25 by Cds1 or Chk1. Nature, 395: 507-510.

108 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 109

CELL-CELL COMMUNICATION

mony with normal neighbouring cells, it is tal to the nature of malignancy, it has long SUMMARY not surprising that the function of genes been postulated that gap junctional inter- involved in intercellular communication cellular communication is disturbed in > Cells communicate by means of secret- mechanisms is disrupted in many cancer. The first confirmatory evidence ed molecules which affect neighbouring tumours. Thus, several oncogenes (Onco- was the observation that of a reduced cells carrying appropriate receptors, and also by direct cell contact, including genes and tumour suppressor genes, p96) level of gap junctional intercellular com- specifically includes gap junctions. encode products involved in humoral inter- munication in one tumour type [8]. This cellular communication: c-erb, c-erbB2 phenomenon has now been observed in > Cell contact-mediated communication and c-SIS [3]. It has also become clear almost all tumours [4]. Cell lines estab- through gap junctions is controlled by that cell contact-mediated intercellular lished from tumours, as well as cells connexin genes and is often disrupted in communication plays a crucial role in cell transformed in vitro, usually exhibit cancer. This may contribute to uncon- growth control [4] and genes involved are impaired function in respect of gap junc- trolled and autonomous growth. often classified as tumour suppressor tional intercellular communication. Gap genes [5]. Cell adhesion molecules are junctional intercellular communication > Interventions restoring gap junction communication may provide a basis for also involved in cell-cell recognition. There between transformed cells and neigh- therapy. are several lines of evidence which sug- bouring normal counterparts is selective- gest that aberrant functions of cell adhe- ly defective in murine embryonic sion may be involved in tumour invasion BALB/c3T3 cells (Fig. 3.31). A lack of het- and metastasis [6]. erologous gap junctional intercellular communication between transformed and Gap junctional intercellular communi- normal cells has been observed using rat In complex organisms, neighbouring cells cation and cancer liver epithelial cell lines and rat liver behave and function in harmony for the Gap junctional intercellular communica- tumour in vivo. It appears that reduced benefit of the whole organism through the tion is the only means by which cells gap junctional intercellular communica- operation of cell-cell communication. exchange signals directly from the interi- tion is common to many tumour cells. During evolution, various types of intercel- or of one cell to the interior of surround- Further studies with multistage models of lular communication have developed, ing cells [7]. Since the extent to which rat liver and mouse skin carcinogenesis which in mammals take two forms: (1) tumour cells deviate from cells which have revealed that there is, in general, a humoral communication and (2) cell con- exhibit tissue homeostasis is fundamen- progressive decrease in the level of gap tact-mediated communication (Fig. 3.30). Humoral communication is typically mediated by molecules, such as growth factors A. Humoral communication and hormones, excreted from certain cells and received by receptors of other cells. Intercellular communication based on direct cell-cell contact is mediated by various junctions, including adherence junctions, desmosomes and gap junctions. During multistage carcinogenesis, genes critically involved in cell growth are B. Cell contact-mediated communication altered [1]. Most such genes are known to 1.Cell adhesion 2. Gap junction be directly or indirectly involved in the control of cell replication (The cell cycle, p104) or in the death of individual cells [2] (Apoptosis, p113). Genes involved in intercellular communication control cellular growth at another level. These genes function to maintain cell growth in harmony with that of the surrounding tissue. Since Fig. 3.30 Relationships between cells are maintained by different types of intercellular communication, most cancer cells do not proliferate in har- which may (B) or may not (A) require cell contact.

Cell-cell communication 109 WCR-S3.Q (composition) 27/01/03 9:27 Page 110

A Injection into transformed cell B junctional intercellular communication during carcinogenesis and tumour progression. Another line of evidence, that implies a causal role for blockage of intercellular communication in carcinogenesis, is that agents or genes involved in carcinogenesis have been shown to modulate gap junctional intercellular communication. The mouse skin tumour-promoting agent 12-O- Phase-contrast micrographs Fluorescence micrographs tetradecanoylphorbol 13-acetate (TPA) inhibits gap junctional intercellular communication. Many other tumour-promoting agents inhibit gap junctional intercellular communication [9]. In addition to such chemicals, other tumour-promoting stimuli, such as partial hepatectomy and skin wounding, have been demonstrated to inhibit gap junctional intercellular communication. Activation of various oncogenes, C Injection into non-transformed cell D including those which encode src, SV-40 T Fig. 3.31 Selective gap junctional communication: cells transformed by a chemical carcinogen (spindle- antigen, c-erbB2/neu, raf, fps and ras, also shaped and criss-crossed) communicate among themselves, but not with their surrounding non-trans- results in inhibition of gap junctional inter- formed counterparts. When a gap junction-diffusible fluorescent dye is microinjected into a single cell (marked with star) of a transformed focus there is communication between transformed cells but not with cellular communication. Conversely, some the surrounding non-transformed cells (A, B). Injection of the dye into a non-transformed cell which is chemopreventive agents enhance gap junc- located near a transformed focus results in communication between non-transformed cells, but not with tional intercellular communication [10]. transformed cells (C, D).

Cells carrying the tk gene are GCV-P cannot pass through killed by the toxic phosphorylated the cell membrane GCV (GCV-P) they produce No bystander GCV-P effect Transfected cells carrying the HSV-tk gene

Ganciclovir (GCV)

Tumour cell population

Bystander GCV-P effect

Movement of GVC-P from cytoplasm to cytoplasm in gap junctions

Fig. 3.32 Role of cell-cell interaction in gene therapy. In a tumour cell population, only a few cells can be reached by vectors carrying the HSV-tk gene. Expression of the tk gene (orange) makes these cells sensitive to ganciclovir: they produce phosphorylated ganciclovir, which is toxic. As phosphorylated ganciclovir cannot pass through the cell membrane, theoretically only cells expressing the tk gene will die as a result of ganciclovir treatment. Transmembrane dif- fusion of phosphorylated ganciclovir from cytoplasm to cytoplasm can induce a bystander effect sufficient to eradicate a tumour cell population, even if only a few cells express the tk gene [13].

110 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 111

Connexin genes and tumour suppres- Enhancement of cancer therapy by Signal transduction from intercellular sion cell-cell communication network The first indication that normal cells may A decade ago it was demonstrated that The main physiological function of gap suppress the growth of malignant cells gap junctional intercellular communica- junctional intercellular communication is with which they are in contact came from tion could be exploited to distribute ther- probably to maintain homeostasis by the work of Stoker and colleagues [11]. apeutic agents among cancer cells and keeping the level of signals mediated by More recent evidence for such a direct thereby enhance cancer therapy [12]. agents of low molecular weight at equilib- role of gap junctional intercellular com- One principle of gene therapy is the rium among cells linked by gap junctions. munication in tumour suppression has mediation of selective cytotoxicity by the This implies that intercellular communica- come from experiments in which connex- introduction, into malignant cells, of a tion may control cell growth indirectly. As in genes were transfected into gap junc- gene that activates an otherwise innocu- already noted, such an activity is distinct tional intercellular communication-defi- ous drug. In practice, only a fraction of from that mediated by genes which are cient malignant cell lines. In many cases, the total number of tumour cells sought directly involved in cell growth and death. connexin gene expression reduced or to be eliminated, are successfully trans- One particularly important pathway link- eliminated tumorigenicity of recipient fected with the gene in question. ing intercellular interaction to signal trans- cells [10]. However, at least in the case of brain duction involves the cell adhesion mole- Although tumour suppressor genes, most tumour therapy based upon the thymi- cule β-catenin. If the level of β-catenin in notably p53, are mutated in a high pro- dine kinase gene from herpes simplex the cytoplasm and nucleus rises, it acti- portion of tumours, few mutations of con- virus (HSV-tk), not only are the cells vates transcription factors of the TCF(T- nexin genes have yet been found in transfected with the gene affected by cell factor)/LEF family and increases rodent tumours and none has been treatment with the drug ganciclovir, but activity of genes including C-MYC, cyclin reported for any human cancer. Although neighbouring cells are also killed in the D1 and connexin-43. Normally the levels this suggests that connexin gene muta- presence of ganciclovir. Several studies of β-catenin in the cytoplasm and nucleus tions are rare in carcinogenesis, only a have provided strong evidence that this are kept very low because a complex of few studies (all from one laboratory) on a phenomenon, termed “the bystander proteins including the APC gene product limited number of connexin genes (Cx32, effect” (Fig. 3.32), is due to connexin- (adenomatous polyposis coli, Colorectal Cx37[α4] and Cx43) have been conduct- mediated gap junctional intercellular cancer, p198), axin and glycogen synthe- ed. Several polymorphisms in connexin communication; that is, ganciclovir phos- sis kinase 3β bind the free β-catenin and genes in humans and rats have been phorylated by HSV-tk can diffuse through put a phosphate group onto it which described, although there was no appar- gap junctions and even those cells with- marks it for destruction [14]. ent correlation between such polymor- out HSV-tk gene can be killed. The role of In normal cells the level of free β-catenin phisms and the cancer sites examined connexin genes in this effect has been is regulated by the Wnt ("wingless homo- [10]. confirmed [13]. logue") signal from outside the cell, which

Gene Cancer site/cancer Changes observed

Integrin Skin, liver, lung, osteosarcoma Reduced expression

E-cadherin Stomach, colon, breast, prostate Mutations: reduced expression

α-catenin Stomach, colon, breast, prostate, Reduced expression oesophagus, kidney, bladder, etc β-catenin Melanoma, colon Mutations: reduced expression

γ-catenin Breast, colon Loss of expression, translocation into nuclei

Connexins Liver, skin etc Reduced expression, aberrant localization

Table 3.4 Examples of cell-cell interaction genes involved in carcinogenesis [10].

Cell-cell communication 111 WCR-S3.Q (composition) 27/01/03 9:27 Page 112

increases the level by transiently reducing accepts the phosphate allow the levels of occur in many cancers, including those of the activity of the kinase. However, muta- β-catenin to rise; when this happens the the colon, breast and endometrium, tions in either the APC gene or in the part TCF/LEF-controlled genes are permanent- shows the importance of this pathway and of the β-catenin (CTNNB1) gene which ly activated. The fact that mutations of the emphasizes the connection between cell- codes for the part of the molecule which CTNNB1, AXIN1, AXIN2 and APC genes cell contact and signal transduction.

REFERENCES 1. Fearon ER, Vogelstein B (1990) A genetic model for 7. Bruzzone R, White TW, Paul DL (1996) Connections autoradiographic studies using marked cells. J Cell Sci, 2: colorectal tumorigenesis. Cell, 61: 759-767. with connexins: the molecular basis of direct intercellular 293-304. 2. Kinzler KW, Vogelstein B (1997) Cancer-susceptibility signaling. Eur J Biochem, 238: 1-27. 12. Yamasaki H, Katoh F (1988) Novel method for selective genes. Gatekeepers and caretakers. Nature, 386: 761, 763. 8. Loewenstein WR, Kanno Y (1966) Intercellular commu- killing of transformed rodent cells through intercellular 3. Heldin CH (1996) Protein tyrosine kinase receptors. nication and the control of tissue growth: lack of commu- communication, with possible therapeutic applications. Cancer Surv , 27: 7-24. nication between cancer cells. Nature, 209: 1248-1249. Cancer Res, 48: 3203-3207. 4. Krutovskikh V, Yamasaki H (1997) The role of gap junc- 9. Trosko JE, Chang CC, Madhukar BV, Klaunig JE (1990) 13. Mesnil M, Yamasaki H (2000) Bystander effect in her- tional intercellular communication (GJIC) disorders in Chemical, oncogene and growth factor inhibition gap junc- pes simplex virus-thymidine kinase/ganciclovir cancer experimental and human carcinogenesis. Histol tional intercellular communication: an integrative hypothe- gene therapy: role of gap-junctional intercellular communi- Histopathol, 12: 761-768. sis of carcinogenesis. Pathobiology, 58: 265-278. cation. Cancer Res, 60: 3989-3999. 5. Hirohashi S (1998) Inactivation of the E-cadherin-medi- 10. Yamasaki H, Omori Y, Zaidan-Dagli ML, Mironov N, 14. Behrens J, Jerchow BA, Wurtele M, Grimm J, Asbrand ated cell adhesion system in human cancers. Am J Pathol, Mesnil M, Krutovskikh V (1999) Genetic and epigenetic C, Wirtz R, Kuhl M, Wedlich D, Birchmeier W (1998) 153: 333-339. changes of intercellular communication genes during mul- Functional interaction of an axin homolog, conductin, with tistage carcinogenesis. Cancer Detect Prev, 23: 273-279. beta-catenin, APC, and GSK3beta. Science, 280: 596-599. 6. Birchmeier EJ, Behrens J (1994) Cadherin expression in carcinoma: Role in the formation of cell junctions and pre- 11. Stoker MG (1967) Transfer of growth inhibition vention of invasiveness. Biochim Biophys Acta, 1198: 11-26. between normal and virus-transformed cells:

112 Mechanisms of tumour development WCR-S3.Q (composition) 29/01/03 14:06 Page 113

APOPTOSIS

“engulfing” respectively [1]. The regulato- dependent on definition of the ced genes SUMMARY ry phase includes all the signalling path- in the nematode Caenorhabditis elegans, ways that culminate in commitment to members of this gene family being vari- > The term apoptosis refers to a type of cell death. Some of these pathways regu- ously homologous to human BCL2 (which cell death that occurs both physiologi- late only cell death, but many of them suppresses apoptosis), APAF-1 (which cally and in response to external stimuli, including X-rays and anticancer drugs. have overlapping roles in the control of mediates caspase activation) and the cell proliferation, differentiation, respons- caspases themselves (proteases which > Apoptotic cell death is characterized by es to stress and homeostasis. Critical to mediate cell death). The centrality of distinctive morphological changes dif- apoptosis signalling are the “initiator” apoptosis to cancer biology is indicated ferent from those occurring during caspases (including caspase-8, caspase-9 by excess tumorigenesis in BCL2-trans- necrosis, which follows ischaemic injury and caspase-10) whose role is to activate genic and p53-deficient mice. An appreci- or toxic damage. the more abundant “effector” caspases ation of apoptosis provides a basis for the (including caspase-3 and caspase-7) further development of novel and conven- > Apoptosis is regulated by several dis- which, in turn, brings about the morpho- tional cancer therapy tinct signalling pathways. Dysregulation logical change indicative of apoptosis. of apoptosis may result in disordered cell growth and thereby contribute to Finally, the engulfing process involves the The role of cell death in tumour carcinogenesis. recognition of cellular “remains” and their growth elimination by the engulfing activity of Apoptosis, or lack of it, may be critical to > Selective induction of apoptosis in tumour surrounding cells. tumorigenesis [2]. BCL2, a gene mediat- cells is among current strategies for the Identification of genes mediating apopto- ing resistance to apoptotic stimuli, was development of novel cancer therapies. sis in human cells has been critically discovered at the t(14:18) chromosomal

APOPTOSIS Apoptosis is a mode of cell death that facilitates such fundamental processes as Condensation and fragmentation of development (for example, by removal of chromatin unwanted tissue during embryogenesis) and the immune response (for example, by elimination of self-reactive T cells). NECROSIS This type of cell death is distinguished from necrosis both morphologically (Fig. 3.33) and functionally. Specifically, apop- Shrinking/ tosis involves single cells rather than Organelle rounding up, areas of tissue and does not provoke disruption fragmentation inflammation. Tissue homeostasis is and break- of cell and down, nucleus dependent on controlled elimination of cell swelling unwanted cells, often in the context of a continuum in which specialization and maturation is ultimately succeeded by cell death in what may be regarded as the final phase of differentiation. Apart from elimination in a physiological context, Membrane cells that have been lethally exposed to blebbing, cytotoxic drugs or radiation may be sub- residual Engulfing by ‘ghost’ cell neighbouring cell ject to apoptosis. as ‘apoptotic- The process of apoptosis can be body’ described by reference to distinct phases, termed “regulation”, “effector” and Fig. 3.33 Apoptosis and necrosis are distinguished by characteristic morphological changes.

Apoptosis 113 WCR-S3.Q (composition) 27/01/03 9:27 Page 114

translocation in low grade B cell non- Hodgkin lymphoma. It thus became DNA DAMAGE REPLICATION STRESS apparent that neoplastic cell expansion could be attributable to decreased cell death rather than rapid proliferation. Defects in apoptosis allow neoplastic cells to survive beyond senescence, Rad17 thereby providing protection from hypoxia ATM + ATR Rad9, Rad1, Hus1 and oxidative stress as the tumour mass expands. Growth of tumours, specifically

in response to chemical carcinogens, has Chk1 Chk2 Mdm2 p53 BRCA1 Nbs1 cAbl been correlated with altered rates of apoptosis in affected tissues as cell populations with altered proliferative activity Cdc25 p21 emerge. Paradoxically, growth of some 14-3-3θ cancers, specifically including breast, has been positively correlated with increasing Cdk CELL CYCLE APOPTOSIS TRANSCRIPTION DNA REPAIR apoptosis [3]. ARREST

Interrelationships between mitogenic Fig. 3.34 Response to DNA damage is mediated by several signalling pathways and may include apoptosis. and apoptotic pathways A dynamic relationship between regula- damage caused by ionizing radiation, con- tion of growth/mitosis and apoptosis may trolling the initial phosphorylation of probe demonstrated using a variety of rele- teins such as p53, Mdm2, BRCA1, Chk2 vant signalling pathways. Many differing and Nbs1. Other sensors of DNA damage promoters of cell proliferation have been include mammalian homologues of the found to possess pro-apoptotic activity PCNA-like yeast proteins Rad1, Rad9 and [4]. Thus, ectopic expression of the C-MYC Hus1, as well as the yeast homologue of oncogene (normally associated with prolif- replication factor C, Rad17. Specific mole- erative activity) causes apoptosis in cul- cules detect nucleotide mismatch or inap- tured cells subjected to serum deprivation propriate methylation. Following exposure (which otherwise prevents proliferation). of mammalian cells to DNA-damaging Fig. 3.35 Apoptotic cell death requires gap junc- Oncogenes that stimulate mitogenesis agents, p53 is activated and among many tional intercellular communication. Expression can also activate apoptosis. These include “targets” consequently upregulated are the and subcellular location of connexin 43 in healthy oncogenic RAS, MYC and E2F. Mutations cyclin-dependent kinase inhibitor p21 (A) and in apoptotic (B) rat bladder carcinoma in E2F that prevent its interaction with the (which causes G1 arrest) and Bax (which cells. Arrows indicate location of connexin 43 in areas of intercellular contact between apoptotic retinoblastoma protein (pRb) accelerate S induces apoptosis). Thus, the tumour sup- (B) and non-apoptotic (A) cells. Counterstaining of phase entry and apoptosis. A function of pressor gene p53 mediates two responses DNA with propidium iodide reveals fragmentation pRb is to suppress apoptosis: pRb-defi- to DNA damage by radiation or cytotoxic of the nucleus typical of apoptosis (B). cient cells seem to be more susceptible to drugs: cell cycle arrest at the G1 phase of p53-induced apoptosis. the cell cycle and apoptosis (Oncogenes Agents such as radiation or cytotoxic drugs and tumour suppressor genes, p96). The cause cell cycle arrest and/or cell death serine/threonine kinase Chk2 is also able [5]. The DNA damage caused by radiation to positively interact with p53 and BRCA1. or drugs is detected by various means (Fig. Chk2 and the functionally related Chk1 3.34). DNA-dependent protein kinase and kinase appear to have a role in the inhibi- the ataxia-telangiectasia mutated gene tion of entry into mitosis via inhibition of (ATM) (as well as the related ATR protein) the phosphatase Cdc25 (The cell cycle, bind to damaged DNA and initiate phos- p104). phorylation cascades to transmit damage signals. DNA-dependent protein kinase is The regulatory phase Fig. 3.36 Apoptotic cells in an adenoma, visualized by immunohistochemistry (red). Apoptosis is believed to play a key role in the response Two major apoptotic signalling pathways restricted to single cells, unlike necrosis, which to double-stranded DNA breaks. ATM plays have been identified in mammalian cells typically involves groups of cells. Apoptosis does an important part in the response to DNA (Fig. 3.37). The “extrinsic” pathway not produce an inflammatory response.

114 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 115

activates the Fas receptor in the absence of ligand. TRAIL (TNF-related apoptosis- CD95L inducing ligand, Apo-2L) has 28% amino acid identity to FasL. TRAIL induces cell CD95 death only in tumorigenic or transformed cells and not in normal cells [8]. FADD Procaspase-8 The regulation of apoptosis by BCL2 DNA damage c-FLIP family genes

p53 While the members of the “death recep- Bid BclxL tor” family and their ligands have struc- Bax tural elements in common, agents and stimuli initiating the mitochondrial path-

Caspase-8 way to apoptosis are diverse. Common to Bcl2 these stimuli, however, is a change in Truncated bid mitochondrial function, often mediated by members of the BCL2 family [9]. In Procaspase-3 humans, at least 16 homologues of BCL2 Cytochrome c have been identified. Several family mem-

bers (including Bcl-2, Bcl-xL, Bcl-W) sup- Apoptosis-inducing press apoptosis, while others induce factor (AIF) apoptosis and may be subdivided on the Apoptosome basis of their ability to dimerize with Bcl- Caspase-3 Smac/DIABLO 2 protein (Bad, Bik, Bid) or not (Bax, Bak). Apaf-1 Procaspase-9 Phosphorylation of Bad protein by a specific (Akt/PKB) and other kinases prevents dimerization with Bcl-2 and pro- Inhibitors of Apoptotic substrates apoptosis (IAPs) motes cell survival. At least two distinct mechanisms of action are recognized: the Fig. 3.37 Apoptosis occurs when specific proteases (caspases) digest critical proteins in the cell. The binding of Bcl-2 (or other members of the caspases are normally present as inactive procaspases. Two pathways lead to their activation. The death receptor pathway (at the top and left side of the figure) is triggered when ligands bind to death receptors family) with either pro- or anti-apoptotic such as CD95/Fas. The mitochondrial pathway is triggered by internal insults such as DNA damage as members of the Bcl-2 family or the for- well as by extracellular signals. In both pathways, procaspases are brought together. They then cleave mation of pores in mitochondrial mem- each other to release active caspase. The binding of ligand (FasL or CD95L) to CD95 brings procaspase branes. Bcl-x is a potent death suppres- 8 molecules together; release of mitochondrial components bring procaspases 9 together. The active L caspase 8 and 9 then activate other procaspases such as procaspase 3. sor that is upregulated in some tumour types. Bax is a death promoter that is inactivated in certain types of colon can- depends upon the conformational change ramifications. The best-characterized cer, stomach cancer and in haematopoi- in certain cell surface receptors following receptors belong to the tumour necrosis etic malignancies. By dint of relevant the binding of respective ligands. The factor (TNF) receptor gene superfamily binding sites, Bax is under the direct tran- “intrinsic” pathway involves mitochondrial [7]. In addition to a ligand-binding scriptional control of p53. function and is initiated by growth factor domain, death receptors contain homolo- deprivation, corticosteroids or DNA dam- gous cytoplasmic sequence termed the Involvement of mitochondria age induced by radiation or cytotoxic “death domain”. Members of the family Apoptosis induced by cytotoxic drugs is drugs. include Fas/APO-1/CD95 and TNF-1 accompanied by critical changes in mito- receptor (which binds TNFα). Activation chondria. Such apoptotic stimuli induce Cell surface receptors of the Fas (or CD95) receptor by its spe- translocation of Bax from cytosol to mito- Apoptosis may be induced by signalling cific ligand (FasL or CD95L) results in a chondria, which induces release of molecules, usually polypeptides such as conformational change such that the cytochrome c. Loss of transmembrane growth factors or related molecules, “death domain” interacts with the adap- potential follows cytochrome c release which bind to “death” receptors on the tor molecule FADD which then binds pro- and is dependent on caspase activation cell surface [6]. Such cell death was ini- caspase-8. In some cell types, drug- (see below), whereas cytochrome c

tially investigated in relation to the induced apoptosis is associated with Fas release is not. Bcl-2 and Bcl-xL reside immune response, but has much wider activation. Ultraviolet irradiation directly chiefly in the outer mitochondrial mem-

Apoptosis 115 WCR-S3.Q (composition) 27/01/03 9:27 Page 116

brane. Bcl-2, Bcl-xL and Bax can form ion tion of caspases. However, nitric oxide is sors conserved throughout evolution. The channels when they are added to synthet- involved in several aspects of apoptosis IAP protein “survivin” is overexpressed in ic membranes, and this may be related to and may act both as a promoter and a large proportion of human cancers. their impact on mitochondrial biology [10]. inhibitor depending on conditions [11]. Little is known about the involvement of In the cytosol after release from mito- caspase mutations in cancer. chondria, cytochrome c activates the cas- The effector and engulfing phases pases through formation of a complex (the In mammals at least 13 proteases which Caspase substrates and late stages of “apoptosome”) with Apaf-1 (apoptotic- mediate the breakdown of cell structure apoptosis protease activating factor-1), procaspase- during apoptosis have been identified and Apoptosis was initially defined by refer-

9 and ATP. It appears that Bcl-2/Bcl-xL are designated caspases-1 through -13 ence to specific morphological change. In may suppress apoptosis by either prevent- [12]. All possess an active site cysteine fact, both mitosis and apoptosis are char- ing release of cytochrome c or interfering and cleave substrates after aspartic acid acterized by a loss of substrate attach- with caspase activation by cytochrome c residues. They exist as inactive zymogens, ment, condensation of chromatin and and Apaf-1. Sustained production of nitric but are activated by different processes phosphorylation and disassembly of oxide (NO) may cause the release of mito- which most often involve cleavage of their nuclear lamins. These changes are now chondrial cytochrome c into the cyto- pro-forms (designated procaspase-8, etc.) attributable to caspase activation and its plasm and thus contribute to the activa- at particular sites, thereby generating sub- consequences. units which form active proteases consist- Most of the more than 60 known caspase ing of two large and two small subunits. substrates are specifically cleaved by cas- Proteolytic cascades may occur with pase-3 and caspase-3 can process pro- some caspases operating as upstream ini- caspases-2, -6, -7 and -9 [13]. Despite the tiators (which have large N-terminal multiplicity of substrates, protease activi- prodomains and are activated by protein- ty mediated by caspases is specific and protein interaction) and others being seems likely to account for much of the downstream effectors (activated by pro- morphological change associated with tease cleavage). As noted earlier, at least apoptosis. Caspases cleave key compo- two pathways of caspase activation can nents of the cytoskeleton, including actin A be discerned: one involving FADD or simi- as well as nuclear lamins and other struc- lar protein-protein complexes and the tural proteins. Classes of enzymes cleaved other mediated by release of cytochrome by caspases cover proteins involved in c. In the former, affinity labelling suggests DNA metabolism and repair exemplified that caspase-8 activates caspases-3 and by poly(ADP-ribose) polymerase and DNA- -7 and that caspase-3 in turn may activate dependent protein kinase. Other classes caspase-6. On the other hand, release of of substrates include various kinases, pro- cytochrome c into the cytoplasm results in teins in signal transduction pathways and the activation of caspase-9 which in turn proteins involved in cell cycle control, activates caspase-3. exemplified by pRb. Cleavage of some B Though the intrinsic pathway to caspase-3 substrates is cell-type specific. Caspase activation may be distinguished from the activity accounts for internucleosomal extrinsic pathway (i.e. that activated by cleavage of DNA, one of the first charac- Fas, etc.), some interaction is demonstra- terized biochemical indicators of apopto- ble. Thus, caspase-9 is able to activate sis. ICAD/DFF-45 is a binding partner and caspase-8. Nonetheless, the pathways are inhibitor of the CAD (caspase-activated separate to the extent that caspase-8 null DNAase) endonuclease, and cleavage of animals are resistant to Fas- or TNF- ICAD by caspase-3 relieves the inhibition induced apoptosis while still susceptible and promotes the endonuclease activity to chemotherapeutic drugs; cells deficient of CAD. C in caspase-9 are sensitive to killing by Fas/TNF but show resistance to drugs Therapeutic implications Fig. 3.38 Neuroblastoma cells treated with ioni- and dexamethasone. Finally, death of In theory, knowledge of critical signalling zing radiation undergo apoptosis. The TUNEL assay was used to visualize apoptotic cells (green), some cells may occur independently of or effector pathways which bring about before (A) and 24 hours after (B) treatment with X- caspase-3. Caspases-3, -7 and –9 are apoptosis provides a basis for therapeutic rays (5 Gray). Close-up shows that the nuclei of the inactivated by proteins of the inhibitor of intervention, including the development of apoptotic cells are fragmented (C). apoptosis family (IAPs) which are suppres- novel drugs to activate particular path-

116 Mechanisms of tumour development WCR-S3.Q (composition) 6/02/03 9:03 Page 117

ways. Several options are under investiga- tumour types. In experimental systems, suppression of Bcl-2 by an antisense tion [14]. More immediately, attempts are cells acquiring apoptosis defects (e.g. p53 oligonucleotide has been shown to retard being made to exploit knowledge of apop- mutations) can more readily survive hypox- tumour growth and the approach is cur- totic processes to increase the efficacy or ic stress and the effects of cytotoxic drugs rently subject to clinical trial. Likewise, specificity of currently available therapy. [15]. However, clinical studies have not antisense oligonucleotides directed at Simple answers have not emerged. Thus, consistently established that mutation of “survivin” are being evaluated. The possi- for example, relatively increased expres- p53 is associated with poor response to bility of using recombinant TRAIL to sion of Bcl-2 (which, under many experi- chemotherapy [16]. induce apoptosis in malignant cells is mental conditions, inhibits apoptosis) is not The function of Bcl-2 family members may under investigation. TRAIL is implicated as necessarily indicative of poor prognosis be subject to interference by small mole- the basis of all-trans-retinoic treatment of and the reverse appears true for some cules [17]. In preclinical animal models, promyelocytic leukaemia [18]. Also note-

DRUGS TARGETING SIGNAL Ligand TRANSDUCTION PATHWAYS Cell membrane Receptor GRB SOS Ras In complex multicellular organisms, cell proliferation, differentiation and survival PLC PI3K Raf are regulated by a number of extracellular hormones, growth factors and cytokines. DAG Akt MEK These molecules are ligands for cellular receptors and communicate with the nucleus of the cell through a network of intracel- BAD P7056K lular signalling pathways. In cancer cells, key components of these signal transduc- Intracellular messengers PKC tion pathways may be subverted by proto- BclxL ERK oncogenes through over-expression or NUCLEAR TRANSCRIPTION mutation, leading to unregulated cell sig- APOPTOSIS nalling and cellular proliferation. Because a DNA number of these components may be pref- erentially over-expressed or mutated in Signaling pathways targeted by anticancer agents. PI3K = phosphoinositide-3-kinase; PLC = phos- human cancers, the cell signalling cascade pholipase C; PKC = protein kinase C; MEK = mitogen-activated protein kinase kinase; ERK = extracellular sig- provides a variety of targets for anticancer nal-regulated kinase; Akt = protein kinase B (PKB); BAD = Bcl-XL/Bcl-2-associated death protein; VEGF = vascular endothelial growth factor; HER = human epidermal growth factor receptor family; PDGF = platelet therapy (Adjei AA, Current Pharmaceutical derived growth factor; FGF = fibroblast growth factor; SOS = son of sevenless guanine nucleotide exchange Design, 6: 471-488, 2000). protein; GRB = growth factor receptor-bound protein.

Different approaches have been used to attack these targets and include classical ing. The drug Gleevec is already in clinical > Receptors, anti-receptor antibodies and cytotoxic agents as well as small molecule use (Leukaemia, p242). It is hoped that in tyrosine kinase receptor inhibitors drug inhibitors. In addition, antisense future, a combination of agents targeting > RAS farnesyltransferase inhibitors oligonucleotides, vaccines, antibodies, parallel pathways, as well as combinations > RAF inhibitors ribozymes and gene therapy approaches with classical cytotoxic agents will improve > MEK inhibitors have been utilized. the outcome of cancer patients. > Rapamycin analogues Classes of agents and their potential tar- > Protein kinase C (PKC) inhibitors The diagram illustrates cell signalling gets include: > Inhibitors of protein degradation pathways that are targeted by anticancer > Inhibitors of ligands, such as recombinant > Inhibitors of protein trafficking agents currently undergoing clinical test- human antibody to VEGF (rHu mAbVEGF)

Apoptosis 117 WCR-S3.Q (composition) 27/01/03 9:27 Page 118

worthy is the development of caspase 4-hydroxyphenylretinamide. Butyrate, a nase enzyme (COX-2) expression may inhibitors for the treatment of certain short-chain fatty acid produced by bacter- modulate intestinal apoptosis via changes degenerative (non-cancerous) diseases ial fermentation of dietary fibre, inhibits in Bcl-2 expression. Aspirin and similar characterized by excess apoptosis. cell growth in vitro and promotes differen- drugs which inhibit COX-2 may promote Drugs shown to induce apoptosis specifi- tiation; it also induces apoptosis. Both apoptosis and prevent tumour formation. cally include chemopreventive agents roles may contribute to its prevention of (Chemoprevention, p151), exemplified by colorectal cancer. Moreover, cyclo-oxyge-

REFERENCES WEBSITE 1. Strasser A, O'Connor L, Dixit VM (2000) Apoptosis sig- early events that modulate caspase activation during apop- The European Cell Death Organization: naling. Annu Rev Biochem, 69: 217-245. tosis. Nat Cell Biol, 2: 318-325. http://www.ecdo.dote.hu/ 2. Kaufmann SH, Gores GJ (2000) Apoptosis in cancer: 11. Chung HT, Pae HO, Choi BM, Billiar TR, Kim YM (2001) cause and cure. Bioessays, 22: 1007-1017. Nitric oxide as a bioregulator of apoptosis. Biochem Biophys Res Commun, 282: 1075-1079. 3. Parton M, Dowsett M, Smith I (2001) Studies of apoptosis in breast cancer. BMJ, 322: 1528-1532. 12. Kumar S (1999) Regulation of caspase activation in apoptosis: implications in pathogenesis and treatment of 4. Choisy-Rossi C, Yonish-Rouach E (1998) Apoptosis and disease. Clin Exp Pharmacol Physiol, 26: 295-303. the cell cycle: the p53 connection. Cell Death Differ, 5: 129-131. 13. Porter AG, Janicke RU (1999) Emerging roles of caspase-3 in apoptosis. Cell Death Differ, 6: 99-104. 5. Rich T, Allen RL, Wyllie AH (2000) Defying death after DNA damage. Nature, 407: 777-783. 14. Nicholson DW (2000) From bench to clinic with apoptosis-based therapeutic agents. Nature, 407: 810-816. 6. Peter ME, Krammer PH (1998) Mechanisms of CD95 (APO-1/Fas)-mediated apoptosis. Curr Opin Immunol, 10: 15. Zhou BB, Elledge SJ (2000) The DNA damage 545-551. response: putting checkpoints in perspective. Nature, 408: 433-439. 7. Yeh WC, Hakem R, Woo M, Mak TW (1999) Gene targeting in the analysis of mammalian apoptosis and TNF 16. Brown JM, Wouters BG (1999) Apoptosis, p53, and receptor superfamily signaling. Immunol Rev, 169: 283- tumor cell sensitivity to anticancer agents. Cancer Res, 59: 302. 1391-1399. 8. Griffith TS, Lynch DH (1998) TRAIL: a molecule with 17. Zheng TS (2001) Death by design: the big debut of multiple receptors and control mechanisms. Curr Opin small molecules. Nat Cell Biol, 3: E43-E46. Immunol, 10: 559-563. 18. Altucci L, Rossin A, Raffelsberger W, Reitmair A, 9. Gross A, McDonnell JM, Korsmeyer SJ (1999) BCL-2 Chomienne C, Gronemeyer H (2001) Retinoic acid-induced family members and the mitochondria in apoptosis. Genes apoptosis in leukemia cells is mediated by paracrine action Dev, 13: 1899-1911. of tumor-selective death ligand TRAIL. Nat Med, 7: 680- 686. 10. Matsuyama S, Llopis J, Deveraux QL, Tsien RY, Reed JC (2000) Changes in intramitochondrial and cytosolic pH:

118 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:27 Page 119

INVASION AND METASTASIS

synonymous with poor prognosis. Current SUMMARY methods of detecting new tumours, Malignant tumour > The ability of tumour cells to invade and including computed tomography (CT) Growth colonize distant sites is a major feature scans or magnetic resonance imaging Angiogenic Viable cells distinguishing benign growths from (MRI), ultrasound, or measurement of cir- factors malignant cancer. Cytokines Zone of sublethal culating markers such as carcinoembry- hypoxia onic antigen (CEA), prostate-specific anti- >Most human tumours lead to death through Growth Necrosis widespread metastasis rather than the gen (PSA) or cancer antigen 125 (CA125) factors adverse local effects of the primary neo- are not sufficiently sensitive to detect

plasm. micrometastases. A greater understand- Rapid accumulation ing of the molecular mechanisms of Inflammatory of genetic damage > Often, metastatic spread first involves metastasis is required. It is clear that cells in sublethal zone regional lymph nodes, followed by mediated by oxygen metastatic growth may reflect both gain free radicals haematogenous spread throughout the (NO= nitric oxide) body. Metastases may become clinically and loss of function, and indeed the manifest several years after surgical resec- search for “metastasis suppressor” genes Reperfusion of sub- tion of the primary tumour. has been more fruitful than identification lethal zone containing of genes which specifically and reliably cells with varying > Current methods are inadequate for the degrees of genetic potentiate metastasis [1]. damage. Some have routine detection of micrometastases and acquired a metastatic the search for effective, selective phenotype The genetics of metastasis therapies directed toward metastatic growth remains a major challenge. With the publication of the human genome sequence, and various major ini- Metastasis tiatives such as the Cancer Genome Project in the UK and the Cancer Genome Fig. 3.40 The hypoxia hypothesis suggests that the progression of malignant tumours to a Anatomy Project in the USA, the search metastatic phenotype is mediated by deficiency of for genes selectively upregulated, mutat- oxygen and resulting tumour necrosis. Metastasis (from the Greek meaning “change in location”) refers to growth of secondary tumours at sites distant from a • Growth • Angiogenesis • Decreased cell-cell adhesion primary neoplasm. Metastasis thus dis- • Protease activation • Increased protease production tinguishes benign from malignant lesions • Selection • Increased cell-matrix adhesion and is the ultimate step in the multistage 1.Localized tumour 2.Breakthrough 3.Invasion process of tumour progression. Metastatic growth is the major cause of treatment failure and the death of cancer patients. Although secondary tumours may arise by shedding of cells within body cavities, the term metastasis is generally reserved for the dissemination of tumour 4.Transport cells via the blood or lymphatics. Spread in the cerebrospinal fluid and transcoelomic passage may also occur. Most (60-70%) cancer patients have overt or occult metastases at diagnosis, and the prognosis of the majority of these patients is poor (Box: TNM Classification 5.Lodgement 6.Extravasation 7.Metastasis of Malignant Tumours, p124). • Adhesion • Proliferation • Angiogenesis • Protease production • Protease activation/production There is a critical need to identify reliable • Metastasis of metastases indicators of metastatic potential, since Fig. 3.41 The stages in the metastatic process, illustrated in relation to the spread of a primary tumour clinical detection of metastatic spread is from a surface epithelium to the liver.

Invasion and metastasis 119 WCR-S3.Q (composition) 27/01/03 9:27 Page 120

metastasis, and a great deal is known about the cellular and molecular events Gene Cancer type(s) Mechanism that underlie the process. However the ability to predict which patient has occult micrometastases, and the discovery of nm23 Family (H1-6) of Breast Cell migration? nucleoside diphosphate (liver, ovary, melanoma) Signalling via G proteins, effective, selective therapies for metasta- kinases microtubule assembly tic disease, remain major challenges in oncology. PTEN/MMAC1 Prostate, glioma, breast Migration, focal adhesions KAI1/CD82/C33 Prostate, stomach, colon, Cell-cell adhesion, motility The biology of metastasis breast, pancreas, lung Growth of tumours beyond a few millime- tres in diameter cannot progress without CAD1/E-cadherin Many adenocarcinomas Cell-cell adhesion, epithelial organization neovascularization, and there is a growing appreciation of how this phenomenon is MKK4/SEK1 Prostate Cellular response to linked to metastasis [3]. Many genetic stress? changes associated with malignant pro- KiSS-1 Melanoma, breast cancer Signal transduction? gression (mutation of HRAS, over-expres- Regulation of MMP-9? sion of ERBB2 oncogenes, loss of p53) induce an angiogenic phenotype (devel- BRMS1 Breast Cell communication, motility oping blood vessels) via induction of cytokines, such as vascular endothelial DPC4 Colon, pancreas ? growth factor (VEGF-A). VEGF-A is also upregulated by hypoxia in tumours, partly Table 3.5 Putative metastasis suppressor genes. by host cells such as macrophages. The presence of hypoxic areas is a character- ed or lost in metastatic cancers (Table metastasis, although many others so iden- istic of solid tumours and has been relat- 3.5) has gained momentum. It is now pos- tified are also associated with tumour ed to poor response to conventional ther- sible, using laser capture microdissection growth or developmental processes. apies (Fig. 3.40). In addition, activation of and serial analysis of gene expression The events which lead to cancer metasta- epithelial growth factor receptor (EGFR) (SAGE), to isolate invasive cancer cells sis include changes in cell-cell and cell- and other oncogenic signalling pathways and compare their gene or protein expres- matrix adhesion, alterations in cell shape, can also upregulate VEGF-C, a known sion with non-invasive or normal cells deformability and motility, invasion of sur- lymphangiogenic cytokine [4]. The recep- from the same patient [2]. Prior to this, rounding normal tissues, gaining access tors for these cytokines (Flk-1 and Flk-4) transfection of chromosomes or DNA to lymphatic or vascular channels, dis- are expressed on tumour vasculature, from metastatic to non-metastatic cells semination via blood or lymph, survival of and both (in addition to acting as potent (or vice versa), subtractive hybridiza- host defence mechanisms, extravasation mitogens for endothelial cells) also tion/differential display PCR, cDNA array and colonization of secondary sites (Fig. enhance vessel permeability. Thus activa- and other strategies resulted in identifica- 3.41). There are now many features of tion of these signalling pathways may tion of some genes specifically linked to cancer cells recognized to potentiate potentiate both vascular and lymphatic invasion and tumour spread. Basic fibroblast growth factor (bFGF) is often upregulated in cancers, particularly at the invasive edge where tumour cells interact with host cells [5]. Epithelial cells are normally bounded by basement membranes which separate them from the underlying stroma and mesenchymal compartments. Breaching this barrier is the first step in the transition from carcinoma in situ to invasive and potentially metastatic carcinoma. Basement membrane is composed of a Fig. 3.42 Multiple metastatic growths of an intes- Fig. 3.43 Multiple metastases to the brain from a variety of structural proteins including tinal carcinoma in the liver. lung carcinoma. collagen IV (the major component),

120 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:28 Page 121

laminin, entactin and also heparan sulfate affinity can be profoundly influenced by lized during lymphocyte “homing”, and a proteoglycans. Interactions of tumour the local microenvironment and soluble change from the standard “epithelial” pat- cells with the basement membrane have factors, enabling the tumour cell to tern to expression of splice variants asso- been considered to comprise three steps, respond to different conditions encoun- ciated with haematopoietic cells has which can readily be demonstrated in tered throughout the metastatic cascade. been proposed to assist carcinoma cells vitro: adhesion, matrix dissolution/protein haematogenous dissemination. Throm- olysis and migration [6]. Other molecules involved in adhesion bospondin may mediate adhesion bet- Epithelial cells are normally polarized and Other adhesion molecules implicated in ween circulating tumour cells, platelets firmly attached to each other via desmo- cancer progression include selectins such and endothelial cells, promoting emboliza- somes, tight junctions and intercellular as sialyl Lex and members of the tion (vessel obstruction) and arrest. adhesion molecules such as E-cadherin, immunoglobulin superfamily, including Tumour cells then gain access to the sub- and also bound to the basement mem- intercellular adhesion molecules (ICAM-1, endothelial basement membrane when brane via other adhesion molecules ICAM-2, VECAM and PECAM). The latter endothelial cells retract in response to including integrins. Changes in cell-cell are upregulated on activated endothelial these emboli, and can adhere to exposed and cell-matrix adhesive interactions are cells, and can interact with integrins on proteins. Synthetic peptides containing common in invasive cancer (Cell-cell com- leukocytes and circulating tumour cells, sequences of amino acids which compete munication, p109). Indeed, E-cadherin assisting their arrest and extravasation. with binding to laminin or fibronectin can may be designated a tumour suppressor CD44 is another adhesion molecule uti- inhibit colonization of the lung by intra- gene, since its loss or functional inactivation is one of the most common characteristics of metastatic cancer, and its re- introduction into cells can reverse the malignant phenotype. The adenomatous polyposis coli gene (APC), which is mutated in many inherited and sporadic colon cancers, normally regulates the expression of β-catenin, a protein which interacts with E-cadherin. Mutations in APC (or β-catenin) increase cellular levels of the latter and facilitate interactions with transcription factors such as T-cell factor/lymphoid enhancer factor (TCF/LEF) which drive the expression of genes involved in inhibiting apoptosis and stimulating cell proliferation. Other genes commonly lost in cancers (e.g. DCC, Deleted in Colon Carcinoma) also encode adhesion molecules.

Integrins Integrins are heterodimeric proteins that mediate adhesion between cells and the extracellular matrix or other cellular elements. Ligand specificity is determined by the subunit composition; many integrins bind multiple substrates and others are more selective. Far from being an inert “glue”, they are capable of transmitting important signals regulating cell survival, differentiation and migration [7]. Many dif- ferences in integrin expression between benign and malignant cells have been doc- Fig. 3.44 MRI scan showing skeletal metastases in a patient with a primary prostatic carcinoma (front umented, but the patterns are complex. In and back views). Some of the larger metastases are marked by arrows. Note the numerous metastases addition, their expression and binding in the ribs and in the spine.

Invasion and metastasis 121 WCR-S3.Q (composition) 27/01/03 9:28 Page 122

venously injected cells in experimental laminin and also activates gelatinases. models. Thus, upregulation of these enzymes in cancers leads to proteolytic cascades and RHO potential for invasion of the basement The RHO gene family of small GTP- membrane and stroma. hydrolysing proteins contains several Metalloproteinases also contribute to members known to be involved in cell tumour growth and metastasis by other migration via regulation of actomyosin- means [9]. During angiogenesis, “inva- based cytoskeletal filament contraction sion” of capillary sprouts requires local and the turnover of sites of adhesion. proteolysis (mediated in part by upregu- Overexpression of RhoC alone in lated MMP-2 and MMP-9 together with melanoma cells is sufficent to induce a uPA) and in addition MMP-9 has been highly metastatic phenotype [8]. implicated in the “angiogenic switch” by releasing VEGF from sequestration in the Enzyme functions in invasion and extracellular matrix [10]. Furthermore, metastasis these proteases can contribute to the Invasive tumour cells show increased sustained growth of tumours by the expression of many enzymes due to ectodomain cleavage of membrane- upregulation of genes, enhanced activa- bound pro-forms of growth factors, and tion of pro-enzymes or reduced expres- the release of peptides which are mito- sion of inhibitors such as tissue inhibitors genic and chemotactic for tumour cells. of metalloproteinases (TIMPs). In addition, tumour cells may also induce expres- Heparanase sion of enzymes by neighbouring host Apart from the structural proteins cleaved cells and “hijack” these to potentiate by metalloproteinases in the basement invasion. membrane and extracellular matrix, the other major components are glycos- Matrix metalloproteinases aminoglycans, predominantly heparan Fig. 3.45 Location of metastases at autopsy for One important group is the matrix metal- sulfate proteoglycan (HSPG). Heparanase some common cancers, indicating that the site of loproteinases (MMP). Different cancers is an important enzyme which degrades metastasis is not random. may show different patterns of expres- the heparan sulfate side-chains of HSPGs sion; for instance squamous carcinomas and, like the proteases described above, frequently have high levels of gelatinase B not only assists in the breakdown of extra- (MMP-9), stromelysins 1-3 (MMP-3, cellular matrix and basement membrane, Primary tumour Site of metastasis MMP-10 and MMP-11, normally stromal but is also involved in the regulation of enzymes, but also expressed by these growth factor and cytokine activity. Basic carcinomas) and matrilysin (MMP-7). fibroblast growth factor (bFGF, another Bronchial cancer Adrenal Adenocarcinomas such as breast may potent mitogen and chemotactic factor (often bilateral) have increased levels of gelatinase A for endothelial cells) and other heparin- (MMP-2) and colon carcinomas common- binding growth factors are sequestered by Breast ductal Liver ly overexpress MMP-7. In addition, MT1- heparan sulfate, providing a localized carcinoma MMP, which activates MMP-2, is often depot available for release by heparanase. Breast lobular Diffuse peritoneal upregulated in tumour and/or neighbour- Similarly, uPA and tissue plasminogen carcinoma seeding ing host tissues. The major substrate of activator (tPA) can be released from Breast Bone, ovary the gelatinases is collagen IV, a major heparan sulfate by heparanase, further component of the basement membrane, potentiating proteolytic and mitogenic Lung Brain whereas the stromelysins prefer laminin, cascades. Ocular melanoma Liver fibronectin and proteoglycans, and can also activate procollagenase (MMP-1), Tissue-specific growth factors Prostate Bone which in turn degrades the fibrillar colla- Finally, it is possible that release of tissue- gens of the interstitial tissues. Urokinase specific growth factors may play a role in Melanoma Brain plasminogen activator (uPA) is also fre- organ selectivity of metastasis. For exam- Table 3.6 Some sites of metastasis which are not quently upregulated in cancer. It controls ple, colorectal carcinoma cells over- explicable by circulatory anatomy. the synthesis of plasmin, which degrades expressing EGFR have a predilection for

122 Mechanisms of tumour development WCR-S3.Q (composition) 27/01/03 9:28 Page 123

Target Example of agent Comments

Adhesion/attachment RGD-toxin constructs and RGD-targeted Have not reached clinical trials gene therapy Anti-avfl3 monoclonal antibody Cytostasis in patients; anti-tumour and (Vitaxin, Medi522) anti-angiogenic in animal models Proteolysis Matrix metalloproteinase inhibitors Cytostatic in patients; rare occurence of tumour partial regressions; stromal fibrosis; activity seen in multiple animal models and in combination with chemotherapy; new agents with varied MMP specificity under development Motility No selective agents Signal pathways Squalamine (NHE-3 inhibitor) Selective for endothelial cells

PDGFR, KDR and EGFR small molecule Active in vitro in animal models; preclinical inhibitors activity in combinations; phase I trials completed for several agents, some tumour stabilization or regression Anti-EGFR monoclonal antibody Neutralizing antibody; active in vitro in (C225) animal models; phase I trials ongoing Anti-VEGF antibody Blocking antibody; active in vitro in animal models; preclinical activity alone and in combination; phase I-III trials ongoing CAI (non-voltage-gated Ca++ uptake Active in vitro in animal models; preclinical inhibitor) activity in combinations; phase I trials of single agents and combinations, some tumour stabilization or regression Extracellular matrix Pirfenidone Suppresses stromal/inflammatory cell Remodelling by stromal expression of TGF-β Phase I trials for pulmonary fibrosis

Table 3.7 Therapeutic agents directed towards stroma-tumour interactions.

growth in the liver where there are high transforming growth factor alpha, epider- nating cells. Most metastases occur in concentrations of its ligands. All of these mal growth factor and platelet-derived the first capillary bed or lymph node require proteolytic cleavage for activation. growth factor, can induce chemotactic encountered. The number of involved Other enzymes which have been implicat- responses in tumour cells expressing the nodes is a key prognostic factor for many ed in metastasis include the cysteine pro- cognate receptors. Scatter factor (also cancers, and this has led to efforts to teinases, notably cathepsins B and D. For known as hepatocyte growth factor, HGF) identify “sentinel” lymph nodes in order most of the enzymes described, there are is a potent host-derived motility factor, to improve predictions of cancer spread. active research programmes seeking and tumour cells themselves secrete a Lymphatic channels present less of a selective inhibitors (some of which have variety of autocrine motility factors includ- challenge to tumour cell entry than capil- reached phase II and III clinical trials) to ing autotaxin and neuroleukin/phospho- laries since they have scanty basement prevent or treat metastatic disease. hexose isomerase. membrane. Once in the lymphatics, Motility, coupled with proteolysis, is the tumour cells are carried to the subcapsu- basis of tumour cell invasion, and is also Organ preference of metastases lar sinus of draining nodes, where they important during intravasation and The organ distribution of metastases may arrest and grow, succumb to host extravasation of blood and lymphatic ves- depends on the type and location of the defences, or leave the node via the effer- sels. Many motility factors have been primary tumour, with 50-60% of the sec- ent lymphatics. The propensity for a described which may be tumour- or host- ondary sites being dictated by the tumour cell to generate a lymphatic derived. Many growth factors, such as anatomical route followed by the dissemi- metastasis may depend upon its ability to

Invasion and metastasis 123 WCR-S3.Q (composition) 27/01/03 9:28 Page 124

TNM CLASSIFICATION OF T = primary tumour MALIGNANT TUMOURS TX Primary tumour cannot be assessed T0 No evidence of primary tumour Tis Carcinoma in situ The TNM system for the classification of T1 Tumour invades submucosa malignant tumours (http://tnm.uicc.org/) T2 Tumour invades muscularis propria is a form of clinical shorthand used to T3 Tumour invades through muscularis propria into subserosa or into non- describe the anatomic extent (staging) of peritonealized pericolic or perirectal tissues a cancer in terms of: T4 Tumour directly invades other organs or structures and/or T - the primary tumour perforates visceral peritoneum N - regional lymph nodes M - distant metastases N = regional lymph nodes NX Regional lymph nodes cannot be assessed These components are given a number N0 No regional lymph node metastasis that reflects the absence or presence and N1 Metastasis in 1 to 3 regional lymph nodes extent of the disease. For example, a N2 Metastasis in 4 or more regional lymph nodes tumour of the colon that is classified as T2N1M0 would have extended into the M = distant metastasis colon’s muscular wall, spread to 1 to 3 MX Distant metastasis cannot be assessed regional lymph nodes but without evi- M0 No distant metastasis dence of distant metastasis. Evaluation M1 Distant metastasis by the TNM system can therefore help in Table 3.8 TNM classification of cancer of the colon and rectum. the planning of treatment by the oncolo- gist and in monitoring the efficacy of this treatment, as well as giving some indica- single international TNM classification Uniform Use, 2nd Edition, Wiley 2001) tion of prognosis. Moreover, the use of a should be formulated. This is achieved via with rules and explanations. standardized system facilitates the dis- meetings of experts that update existing The challenge for the future is the incor- semination of information in the clinical classifications, as well as develop new poration into TNM of information from community. ones. The present TNM edition (Eds. Sobin new diagnostic and imaging technologies The TNM system was developed by Pierre LH and Wittekind Ch, TNM Classification of (such as endoscopic ultrasound, magnet- Denoix (President of the UICC, 1973-1978) Malignant Tumours, 6th Edition, Wiley, ic resonance imaging, sentinel node bio- between 1943 and 1952 (Sobin LH, TNM – 2002) contains guidelines for classification psy, immunohistochemistry and poly- principles, history and relation to other and staging that correspond exactly with merase chain reaction). There is an prognostic factors. Cancer, 91:1589-92, those of the 6th edition of the AJCC Cancer expanding array of known and potential 2001). In 1968, a series of brochures pub- Staging Manual (2002). TNM, now the most prognostic factors (Eds. Gospodarowicz M lished by UICC describing the classifica- widely used system to classify tumour et al., Prognostic Factors in Cancer, Wiley, tion of cancers at 23 body sites were com- spread, is published in 12 languages and is 2001) with which TNM could be integrat- bined to produce the Livre de Poche, accompanied by an illustrated TNM Atlas ed to form a comprehensive prognostic which has been subject to regular re-edi- (Eds. Hermanek P et al., 4th Edition, system. Such integration could potential- tion, enlargement and revision over sub- Springer-Verlag, 1997), a TNM Mobile ly be exploited to enhance the prediction sequent years. In order to prevent Edition (Wiley, 2002) and a TNM supple- of prognosis, and individualize cancer unwanted variations in the classification ment (Eds. Wittekind Ch et al., TNM patient treatment. by its users, in 1982 it was agreed that a Supplement 2001. A Commentary on

adhere to reticular fibres in the subcapsu- there; colon carcinoma cells enter the that specific cancer cells (the seed) had lar sinus. These fibres contain laminin, portal circulation which delivers cells to an affinity for certain organs (the soil). fibronectin and collagen IV, and different the liver, and so on (Fig. 3.45). However, a In experimental systems there are many integrins expressed by different tumour non-random element in metastatic pat- examples showing that primary tumours cells may be responsible for adhesion to terns has long been recognized (Table are heterogeneous, and that cloned cells these structures and to the lymphatic 3.6). Stephen Paget developed the “seed can vary in their ability to metastasize to endothelial cells [11]. and soil” hypothesis in 1889, based on his different sites. Some of the patterns Sarcomas tend to metastasize to lungs observations from autopsies of over 700 relate to the ability of malignant cells to because the venous drainage returns women with breast cancer. He proposed adhere to the endothelial cells in target

124 Mechanisms of tumour development WCR-S3.Q (composition) 6/02/03 9:05 Page 125

organs, and to respond to local growth stages, both in experimental models and chain reaction (PCR), individual tumour factors once they have extravasated. It in man, many tumour cells reach distant cells (or specific genetic markers) can be used to be thought that escape from the sites but may remain dormant, either due found in blood, nodes, bone marrow, body primary tumour and survival in the circu- to lack of appropriate growth factors, or fluids etc, but the significance of “posi- lation were the major rate-limiting steps their failure to induce neoangiogenesis. tive” results, and whether they can be for successful metastasis. However, while Indeed, using sensitive assays such as used to predict subsequent overt metas- there is a good deal of attrition at these immunocytochemistry and polymerase tases is not yet established.

REFERENCES WEBSITE 1. Yoshida BA, Sokoloff MM, Welch DR, Rinker-Schaeffer 7. Berman AE, Kozlova NI (2000) Integrins: structure and The Metastasis Research Society (UK): CW (2000) Metastasis-suppressor genes: a review and per- functions. Membr Cell Biol, 13: 207-244. http://www.metastasis.icr.ac.uk spective on an emerging field. J Natl Cancer Inst, 92: 1717- 1730. 8. Ridley A (2000) Molecular switches in metastasis. Nature, 406: 466-467. 2. Simone NL, Paweletz CP, Charboneau L, Petricoin EF, Liotta LA (2000) Laser capture microdissection: Beyond 9. McCawley LJ, Matrisian LM (2000) Matrix metallopro- functional genomics to proteomics. Mol Diagn, 5: 301-307. teinases: multifunctional contributors to tumor progres- 3. Fidler IJ (2000) Angiogenesis and cancer metastasis. sion. Mol Med Today, 6: 149-156. Cancer J Sci Am, 6 Suppl 2: S134-S141. 10. Bergers G, Brekken R, McMahon G, Vu TH, Itoh T, 4. Eccles SA (2000) Cell biology of lymphatic metastasis. Tamaki K, Tanzawa K, Thorpe P, Itohara S, Werb Z, Hanahan The potential role of c-erbB oncogene signalling. Recent D (2000) Matrix metalloproteinase-9 triggers the angio- Results Cancer Res, 157: 41-54. genic switch during carcinogenesis. Nat Cell Biol, 2: 737- 5. Fidler IJ (1999) Critical determinants of cancer metas- 744. tasis: rationale for therapy. Cancer Chemother Pharmacol, 43 Suppl: S3-10. 11. Brodt P (1991) Adhesion mechanisms in lymphatic metastasis. Cancer Metastasis Rev, 10: 23-32. 6. Stracke ML, Liotta LA (1992) Multi-step cascade of tumor cell metastasis. In Vivo, 6: 309-316.

Invasion and metastasis 125