Chemical

A variety of chemicals have been identified as carcinogens in man, and others are suspected to be on the basis of their carcinogenic effects in experimental animals. There is no common structural link between the different types of chemical carcinogen, but they appear to have in common the ability to modify the structure of DNA: for example, by forming adducts or by adding alkyl groups.

Many chemical carcinogens are procarcinogens which require metabolic conversion to their active form by enzymes. If the enzyme required is present in all cell types, the carcinogenic effect is likely to occur at the site of exposure. However, some carcinogens require metabolism in another tissue, which influences where they exert their carcinogenic effects. This is well illustrated by the aromatic amine β-naphthylamine, which requires metabolism by the liver before being active, and as a result causes neoplasms of the bladder where it is concentrated during excretion.

The major classes of chemical carcinogens currently known are as follows.

Radiation

Electromagnetic radiation of wavelengths shorter than the visible spectrum can cause damage to DNA that can result in neoplasia. Ultraviolet light is a significant carcinogen because of the high levels of exposure that can occur during daily life, whereas ionising radi-ation (such as x-rays and gamma-rays) is significantly carcinogenic because of the high levels of energy it possesses.

Viruses

A growing number of viruses have been implicated in the development of neoplasms. There are manyexamples of virally induced neoplasms in animals, study of which has done much to promote our understanding of the molecular genetics of neoplasia. Virally induced neoplasms in humans are (as far as we are aware) rather less common. The most ubiquitous oncogenic viruses are the human papilloma viruses; other viruses with well-established carcinogenic effects are the Epstein-Barr virus and the hepatitis B virus (see Table 5.2).

Table 5.2 Oncogenic viruses

Mechanisms of viral carcinogenesis

DNA viruses can be carcinogenic either through integration into the host genome in such a way that interferes with the function of growth-controlling genes, or through their ability to produce proteins that interfere with growth-regulating factors. For example, human papilloma viruses produce proteins that inhibit the function of the p53 and Rb1 gene products (see section on genetics). The best-known examples of oncogenic DNA viruses are the Epstein-Barr virus, which is strongly associated with Burkitt’s lymphoma and nasopharyngeal carcinoma, the hepatitis B virus, which is associated with hepatocellular (liver) carcinoma, and the human papilloma virus (HPV). HPV is associated with neoplasia of a number of different surface epithelia. It is responsible for the common viral wart of the skin and is also the main cause of carcinoma of the cervix, its precursor condition cer-vical intraepithelial neoplasia (CIN), and other forms of analogous intraepithelial neoplasia such as anal intraepithelial neoplasia (AIN). There are many different types of HPV. Individual types have preferred target tissues and have differing oncogenic potential. For example, many different HPV types infect the cervix, but only a small number of types (particularly types 16 and 18) are associated with the development of cervical carcinoma.

Oncogenic RNA viruses are retroviruses which integrate their genetic material into the host genome using the enzyme reverse transcriptase. Although there are many examples of oncogenic retroviruses causingneoplasms in animals, this is rare in man. The best-known examples are Human T-Lymphotrophic Virus-1 (HTLV-1) which causes a form of lymphoma/leukaemia which is endemic in Japan and the Caribbean, and the human immunodeficiency virus (HIV). However, HIV probably does not have a direct carcinogenic effect; the neoplasms that are associated with HIV infection probably arise as a consequence of immunosuppression and may actually be caused by other types of virus. Thus, HIV infection may act as a cofactor for oncogenesis by other viruses. There are other examples of this phenomenon, such as the Epstein-Barr virus requiring malaria infection as a cofactor in the development of Burkitt’s lymphoma.

Other associations between viruses and neoplasms are being described – for example, herpes virus 8 and Kaposi’s sarcoma and myeloma – and it seems likely that further causative associations will be established in the future, particularly in neoplasms of the lymphoreticular system. The sequence of events by which viruses can cause neoplasia is outlined in Fig. 5.1.

Fig. 5.1 Viral carcinogenesis.

Other non-biological factors

Other biological factors

Age

Neoplastic disease is primarily a disease of old age. Although neoplasms can occur at any age, even in utero, neoplasms of almost all types become far more common after the age of 50. This presumably reflects the cumulative effects of exposure to carcinogens over an individual’s lifespan. A major reason for the continuing increase in the incidence of neoplastic disease is the increasing life expectancy of most populations.

Individual types of neoplasm have their own typical age distribution. For example, fibroadenoma of the breast usually occurs in women in their second, third and fourth decades, whereas carcinoma of the breast becomes more common after the menopause. Other types of neoplasm, for example, neuroblastoma of the adrenal, are restricted to children and are almost unknown in adults.

It is a general rule that familial neoplasms – that is, those occurring in individuals who have a genetic predisposition to them (see below under genetic factors) –occur at a younger age than sporadic neoplasms.

Race

Different races are subject to different profiles of neoplastic disease. This is almost entirely due to differences in lifestyle. For example, the commonest fatal neoplasm in the UK and the USA is carcinoma of the bronchus, which is caused largely by tobacco smoking. The commonest fatal neoplasm worldwide is hepatocellular carcinoma, which in Africa and South-East Asia is related to exposure to dietary carcinogens and viral hepatitis. Immigrant groups tend to eventually assume the disease profile of their adopted countries. There are, however, occasional examples of genetically determined racial differences, such as a high frequency of familial breast cancer in Ashkenazi Jews.

Endocrine status

Gender influences the risk of developing many types of neoplasm. This is generally related to differences in hormonal status, although lifestyle differences can play a part. For example, the far higher incidence of neoplasms of the breast in females than males is probably mainly due to endocrine influences, whereas, in the past, bronchogenic carcinoma was more common in men than women because of differences in the frequency of tobacco smoking between the sexes. There are, however, many examples where the influence of gender is not understood: for example, the higher frequency of osteosarcoma in males.

Diet

The risks of developing neoplasia as a result of dietary contaminants are well illustrated by the example of aflatoxin-induced hepatocellular carcinoma. Other dietary factors may also modify the risk of developing certain neoplasms, for example, there is a link between high levels of dietary fat and breast carcinoma. The risk of colorectal carcinoma seems to be associated with diet, but is probably multi-factorial. Dietary fiber, fruit and vegetable consumption seem to be protective and red meat consumption seems to be deleterious, but it has been difficult to consistently show an independent effect for any of these factors. There is an increasing level of public interest in the importance of diet in causing or preventing cancers of many types. This has led to much interest in the media and even governmental public health campaigns, although the level of scientific evidence behind the benefits of any individual dietary manipulations is often dubious at best and imaginary at worst.

Genetic factors

Changes in the structure and function of a cell’s genetic material are central to the development of neoplasia, and more than one such change is required in an individual cell before neoplasia can occur. If all of an individual’s cells already have an abnormality in a relevant gene as a result of that individual’s inherited genetic make-up (a germ-line mutation), then fewer subsequent changes are required for neoplasia to occur. This is well illustrated by the rare familial retinoblastoma syndrome (Fig. 5.2). If both alleles of the retinoblastoma (Rb1) gene in an individual retinal cell are non-functional, retinoblastoma can develop from that cell. Sporadic retinoblastoma is a rare tumour because it is unusual for both retinoblastoma alleles in an individual cell to acquire mutations that inhibit their function. However, if one allele is already non-functional because it was inherited in a defective form, then the chances of retinoblastoma developing as a result of a subsequent mutation of the other allele are very high. This also demonstrates that the ‘retinoblastoma’ gene is a tumour suppressor gene. This is a common property of the genes that are abnormal in the various familial cancer syndromes.

Fig. 5.2 The genetics of retinoblastoma.

Another characteristic that the retinoblastoma syndrome shows, that is common in familial cancer syndromes, is that it affects more than one tissue. If individuals with the retinoblastoma syndrome survive the development of retinoblastomas (which are usually bilateral) early in childhood, they have a very high incidence of osteosarcoma during adolescence. The retinoblastoma syndrome is used here as an illustration because its genetics are simple and well characterised. However, there are a number of other familial cancer syndromes, many of which, such as familial polyposis coli, are more common. Increasingly, these syndromes are being identified with mutations in genes that are involved in DNA repair such as the BRCA1 gene associated with familial breast cancer. The best known examples are given in Table 5.3.

Table 5.3 Examples of familial cancer syndromes

Immune response

Some neoplasms attract large numbers of inflammatory cells, usually lymphocytes, into their substance, and there is evidence that in some tumours this may convey a better prognosis. These observations have led to the development of the major research subspecialty of tumour immunology, but, at present, treating neoplasms by stimulating the host immune response islittle more than a theoretical concept.

However, evidence has been put forward that the immune system can detect and mount a response against neoplasms, principally via NK cells. The activity of these cells can be stimulated by lymphokines such as interleukin 2, and there is some evidence that factors like interleukin 2 could have therapeutic efficacy against some neoplasms. A more convincing example of an immunological treatment that can suppress the development of a neoplasm is the effect of BCG treatment on carcinoma-in-situ of the bladder, although whether this is due to a specific immunological reaction or shedding of unstable transformed urothelium in response to a non-specific inflammatory response is not clear.

The host immune response has another important indirect effect on the development of neoplasms. There is a strong positive link between immunodeficiency and the development of neoplasia. This is illustrated by the frequency of development of lymphomas and Kaposi’s sarcoma in the acquired immune deficiency syndrome (AIDS), and cutaneous and anogenital squamous carcinomas in organ transplant recipients taking immunosuppressive therapy. In these settings the increased risk of neoplasia seems to be due to an inadequate immune response to oncogenic viruses such as HHV-8 in the case of Kaposi’s sarcoma and HPV in the case of transplant-associated squamous carcinomas.

Malignant change in benign neoplasms

The majority of benign neoplasms do not alter in any way, but some benign neoplasms have the ability to progress to become malignant neoplasms. Probably the best-characterised example of this phenomenon is the adenoma-carcinoma sequence in the colon. Adenomatous polyps of the colon are more numerous than colonic carcinomas, but all adenomatous polyps have the potential to develop into carcinomas, and many (but not all) carcinomas originate from adenomatous polyps. The polyps most likely to undergo malignant change show the greatest degree of histological dysplasia and a sequence of genetic changes that leads to the development of colorectal carcinoma from normal epithelium via adenomatous polyps has now been described (Fig. 5.3).

Fig. 5.3 The sequence of genetic alterations in the colorectal adenoma-carcinoma sequence. APC = adenomatous polyposis coli gene; MCC = mutated in colon cancer gene; DCC = deleted in colon cancer gene; Ras = a cellular oncogene involved in growth factors signal transduction; p53 = a tumour suppressor gene.

Metaplasia-dysplasia sequence

Neoplastic transformation of cells occurs in cells undergoing proliferation, and is particularly likely to occur if the cells are also undergoing metaplasia (defined and described in the previous chapter). Neoplastic transformation of metaplastic epithelium usually follows a predictable and histologically identifiable sequence of low grade dysplasia progressing to high grade dysplasia/in-situ malignancy to invasive malignancy as additional genetic abnormalities are acquired in the neoplastic population. This progression is very well demonstrated in the cervix (Fig. 5.4). Other examples of the metaplasia–dysplasia sequence are shown in Table 5.4.

Fig. 5.4 The metaplasia-dysplasia-carcinoma sequence in cervical carcinoma.

Table 5.4 Examples of the metaplasia-dysplasia sequence

Premalignant conditions

These are usually conditions characterised by high cell turnover over a sustained period of time, usually resulting from a destructive form of chronic inflammation. Congenital abnormalities can also be premalignant conditions: for example, maldescent of the testis is associated with an increased risk of testicular neoplasia in later life. Some examples of premalignant conditions are given in Table 5.5.

Table 5.5 Examples of premalignant conditions

Chromosomal abnormalities

Very crude DNA abnormalities may manifest themselves as visible changes in chromosomes isolated from neoplastic cells. These abnormalities can occur in a number of forms:

The DNA of many neoplasms, particularly malignant ones, is inherently unstable, and random chromosomal abnormalities are common. However, there are a number of chromosomal abnormalities that are consistently found in certain tumour types, the best known being the ‘Philadelphia chromosome’ (a reciprocal, balanced translocation between chromosomes 9 and 22). At a purely descriptive level, these can be useful for diagnosis, particularly in groups of tumours in which the cells are morphologically similar, such as leukaemias and the ‘small round cell tumours’ of childhood such as neuroblastoma and alveolar rhabdomyosarcoma. Detailed molecular study of these chromosomal abnormalities has yielded some insight into the pathogenesis of some of the neoplasms with specific chromosomal abnormalities. A good example of this is the translocation between chromosomes 14 and 18 that occurs in follicular (low grade) B cell non-Hodgkin’s lymphomas. This translocation results in the bcl-2 gene coming under the control of the immunoglobulin heavy chain gene promoter. As B lymphocytes constitutively express their immunoglobulin genes, this results in inappropriate over-expression of the bcl-2 gene and thus overproduction of the bcl-2 protein. As bcl-2 is an anti-apoptotic protein, this results in the ‘immortalisation’ of the neoplastic B lymphocytes.