Genetic predisposition to cancer

Genetic predisposition to cancer

CANCER TRIALS AND SERVICES Genetic predisposition to cancer What’s new? C D Gareth Evans C New gene-based treatments, such as poly (ADP-ribose) po...

124KB Sizes 0 Downloads 18 Views

CANCER TRIALS AND SERVICES

Genetic predisposition to cancer

What’s new? C

D Gareth Evans C

New gene-based treatments, such as poly (ADP-ribose) polymerase (PARP) inhibitors Better risk prediction, particularly in common cancers like breast and prostate, using multiple validated common single nucleotide polymorphisms (SNPs)

Abstract Over the last 24 years there has been a burgeoning development of genetic risk assessment and ‘family history’ clinics to deal with the ever-increasing demand from individuals at increased risk of cancer by virtue of their family history. Risk of inherited cancer can be divided into known syndromes, such as familial polyposis, and increased risk of common cancers due to family history alone. Risk of cancer can be assessed in three categories: average, moderate and high. Individuals at high risk and those at risk of syndromes will generally be referred to a regional genetics centre. Moderate-risk individuals may benefit from early surveillance in secondary care particularly for breast and colorectal cancer. Average-risk individuals can be reassured in primary care. Newer surveillance techniques such as MRI screening are now being approved in high-risk categories. Genetic testing for a minority of high-risk individuals is now in routine practice and surgical management options have gained validity. Much research is still necessary to improve early detection and develop non-surgical means of prevention.

which needs to be inactivated to enable growth (like the brakes on a car) and the oncogene, which requires activation to promote growth (like the accelerator pedal of a car being stuck down). The great majority of these genetic events are acquired, through replication error (simple copying of DNA during cell division), following exposure to external agents (radiation, chemicals, viruses), or as a result of epigenetic factors such as ageing, which increase gene silencing through methylation. Nonetheless, recent evidence has confirmed that predisposition to cancer involves a polygenic pattern with multiple common gene variations associated with small elevations in risk.

Inherited cancer Occasionally, mutations in tumour suppressor genes can be inherited rather than acquired. Identifying the genes that cause hereditary disease has given an insight into many cancers. The role of cancer-predisposing genes in the causation of sporadic cancer is still the subject of much research, and we can still learn from the more obscure cancer-prone syndromes. Broadly, predisposition can be subdivided into rare genetic syndromes, which have malignancy as a high-risk side effect, and a larger group that cannot be easily identified clinically, which has a strong family history of one or more common malignancies. Identifying patients with a genetic predisposition to cancer may also be important in treatment selection, as the genetic abnormality underlying their cancer may predispose them to enhanced treatment-related toxicity, especially if DNA damage and repair pathways are affected.

Keywords APC; BRCA1; BRCA2; breast cancer; familial polyposis; oncogene; tumour suppressor gene

There has been increasing evidence of familial predisposition to cancer since the classic model of hereditary retinoblastoma was outlined.1 The idea that some cancers were due to hereditary factors has long been held by more than just a few die-hard clinicians. The earliest reports of cancer families date back nearly 200 years to a large cluster of breast cancer in the wife and family of a French physician named Broca, and a cluster of gastric cancer in Napoleon’s family.2 Despite the pioneering work of clinicians and researchers such as Henry Lynch and Mary-Claire King in the USA in the 1960se1980s, demonstrating the hereditary nature of at least a proportion of cancers, such as those affecting the breast and colon, the hereditary element was not proven until the advent of molecular biology, when abnormalities were demonstrated in cancer-predisposing genes. It is, therefore, only in the last 21 years that the hereditary nature of a small proportion of certain common cancers has been proven.

Retinoblastoma Retinoblastoma has been the model from which much of our current knowledge of tumour suppressor genes was fashioned. A familial tendency to this early childhood eye malignancy was recognized in the 19th century. About 50% of cases are due to the inheritance of a gene defect in one copy of the retinoblastoma gene (RB on chromosome 13), and over 90% of individuals who carry a mutation will develop retinoblastoma, usually bilaterally. In 1971, Knudson proposed that tumour development requires mutational events in both copies of the gene.1 Those cases that inherited a mutated copy need only one further mutation and are far more likely to develop the malignancy, which occurs at a younger age and is usually bilateral. The sporadic cases require two mutations (‘hits’) in a retinal cell as opposed to one in the familial case (Figure 1), so bilateral tumours are incredibly unlikely to occur and the unilateral tumours present later. This

Molecular basis of cancer That cancer is ‘genetic’ at the cellular level is now beyond dispute. All tumours result from mutations or other disruptions of essentially two types of gene: the tumour suppressor gene,

D Gareth Evans MD FRCP is Consultant in Genetic Medicine at the University Department of Genetic Medicine, St Mary’s Hospital, Manchester, UK. Competing interests: none declared.

MEDICINE 40:1

29

Ó 2011 Elsevier Ltd. All rights reserved.

CANCER TRIALS AND SERVICES

Ideagram of the ‘two hit’ hypothesis

Examples of autosomal dominant and recessive syndromes predisposing to cancer, their chromosomal location and protein product

The first hit is usually a mutation (represented by a cross) which causes disruption of the protein product. The second hit is often loss of the whole gene by deletion of part or all of the chromosome on which the gene resides. INHERITED

normal

Location (chromosomal)

Protein

Autosomal dominant FAP NF1 NF2 von HippeleLindau MEN1 MEN2 Tuberous sclerosis

5q 17q 22q 3p 11q 10q 9q (TSC1) 11q (TSC2) 16q 18q and other(s) 19p and other(s) 10q 17q

APC Neurofibromin Merlin/schwannomin pVHL Menin RET PTCH Hamartin Tuberin pDPC/SMAD4 pSTK11/LKB1 PTEN Not found

15q 11q Seven types

Many, including BRCA2 pBLM pATM Two types

1q 11q

pLYST OCA1, OCA2

Xq Xp

BLk CD43

SPORADIC

mutated

deleted

Juvenile polyposis PeutzeJeghers Cowden Tylosis Autosomal recessive Fanconi anaemia 14 loci

deleted

Bloom syndrome Ataxia telangiectasia Xeroderma pigmentosa ChediakeHigashi Albinism X-linked Bruton WiskotteAldrich

Figure 1

hypothesis, which has since been validated in other conditions, now bears the conceiver’s name.

The route to discovering cancer genes The discovery of retinoblastoma cases with constitutional deletions of chromosome 13 visible under the microscope concentrated research on that region.3 One of the genes removed by the deletion in these cases, esterase D, then acted as a genetic marker for further studies. The gene for retinoblastoma, the first hereditary cancer gene, was eventually localized and identified by this gene linkage analysis in 1986.4 The same approaches to research using chromosome studies of individuals, their tumours and genetic linkage have led to the discovery of nearly all the high-risk genes that cause cancer predisposition (Table 1). Indeed most of these genes and their gene products (proteins) were found in a heady 6-year period between 1989 and 1995. The last 16 years has seen progress towards the understanding of how these genes function through their protein products and the development of the first line of gene-based treatments, such as trastuzumab, imatinib and a new breed of synthetic lethal drugs, the poly (ADP-ribose) polymerase (PARP) inhibitors, to treat BRCA1/2-related cancers. However, much of the remaining inherited component has been unpicked in the last 4 years by genome-wide association studies to find the lower-risk genetic components.5

Table 1

of most interest as they are likely to represent the inheritance of a faulty copy of a tumour suppressor gene, which predisposes the individual to common cancers. Although the conditions are generally uncommon, the tumour suppressor genes involved are likely to play a fundamental role in the genesis of tumours, which affect 35e40% of humans in the developed world in their lifetime. The identification of those causing genetic syndromes is likely to lead to more specific treatment using genetic-based therapies, as well as earlier identification, monitoring and, most hopeful of all, prevention of common cancers. Familial adenomatous polyposis (FAP) FAP is the model condition by which researchers have hoped to transpose knowledge of a rare genetic disease to a commonly occurring cancer. FAP is an autosomal dominant condition characterized by the development of hundreds to thousands of adenomatous polyps in the colon and rectum, usually by 30 years of age. Untreated, this leads to the almost inevitable development of a colorectal cancer by the age of 60 years. The condition may be associated with osteomas and epidermal cysts, as well as increased risks of other malignancies such as duodenal cancer, hepatoblastoma, glioma and thyroid cancer. The gene for FAP (APC ) was localized to 5q21eq22 by genetic linkage and identified in 1991. FAP was one of the first conditions in which a clear-cut correlation

Genetic syndromes These are usually readily identifiable by a clinical phenotype (group of associated clinical features) or by laboratory tests. The syndromes may be autosomal dominant or recessive or X-linked (Tables 1 and 2). Of these, it is the dominant conditions that are

MEDICINE 40:1

Name of disease

30

Ó 2011 Elsevier Ltd. All rights reserved.

CANCER TRIALS AND SERVICES

Chromosomal location and implications of various dominant cancer syndromes Disease

Location

Tumours

Probable earliest tumour

Risk in lifetime (%)

Start of screening

FAP NF1 NF2 vHL

5q 17q 22q 3p

Adenomas, bowel cancer Neurofibroma, glioma, sarcoma Schwannomas, meningiomas, ependymoma Haemangioblastoma, renal carcinoma

First year?, 4 years, 7 years First year, first year First year, first year, first year 1e2 years, 20 years

100, 99 100, 12 100, 60, 10 90, 70

MEN 1 MEN 2a

11q 10q

5 years 3 years

95 80

MEN 2b Gorlin Cowden LFS

10q 9q 10q 17p

1 year 5 years, 1 year 30 years First year

100 90, 5 30 95

Birth Birth 35 years First year

BRCA1 BRCA2 HNPCC

17q 13q 2p, 3p 2q, 7q

Parathyroid, insulinoma, gastrinoma Medullary thyroid, cancer, parathyroid, phaeochromocytoma As in MEN 2a, except parathyroid Basal cell carcinoma, medulloblastoma Breast, thyroid Sarcoma (bone/soft tissue), adrenal, breast cancer, gliomas Breast/ovary/colon/prostate carcinoma Breast/ovary/prostate carcinoma, male breast Colorectum, ovary, endometrium, ureter, gastric, pancreas

10e16 years Birth Birth 5 years, 15 years 5 years 3e4 years

>16 years >16 years >16 years

80e90 80e90, 10 80

30 years 30 years 25 years

Table 2

between genotype (the genetic change in APC ) and phenotype (the clinical picture) emerged. Patients with mutations in the early part of the gene (50 :exons 2e5) had a very mild clinical picture with late onset of polyps, whereas those with mutations from exon 9 through to codon 1450 of exon 15 had a classical disease course with nearly all patients manifesting the typical congenital retinal pigmentation. However, those with mutations beyond codon 1450 showed typical features of Gardner’s syndrome (osteomas, cysts and desmoid disease) without retinal signs, but also mild polyp disease.6 The APC gene is of fundamental importance in the genesis of the majority of ‘sporadic’ colorectal cancer, although the loss of both functioning copies of the gene is acquired, rather than one inherited and one acquired, as seen in FAP. Use of genetic registers and genetic testing to target screening along with appropriately timed removal of the colon has led to an improvement in life expectancy in FAP of 15e30 years.7

Gorlin’s syndrome is characterized by the development of multiple jaw keratocysts and basal cell carcinomas as well as a 5% risk of childhood medulloblastoma. The multiple endocrine neoplasias are further conditions that predispose to benign tumours and at least one malignancy. In MEN1, the organs affected are the parathyroid glands, pituitary and pancreas. In MEN2, there is the association of medullary carcinoma of the thyroid and phaeochromocytoma. All of these conditions have benefited from the identification of the causative gene and the targeted use of screening that, in most of the conditions, improves life expectancy. This may mean intervention in childhood, particularly in MEN2, to remove the thyroid gland preventatively.

Common cancer predisposition Recent years have seen an enormous improvement in our understanding of the mechanisms of carcinogenesis. Most cancers require a number of genetic changes in a cell before an invasive tumour results. Few are likely to be caused purely by the loss of two copies of a single tumour suppressor gene, as in retinoblastoma, and the number of changes probably varies between four and 10. A combination of loss of function of tumour suppressor genes and activation of oncogenes is usually involved. The particular combination and order may alter the histological as well as invasive nature of the cancer. There is clear evidence that a minority of people who develop common cancers have inherited a faulty gene, which puts them at highrisk of malignancy, but this is usually not recognized as a syndrome apart from in the family history. Adenocarcinomas are more likely than carcinomas of squamous epithelium to have a strong hereditary component, with 4e13% of all breast, ovarian and colon cancer resulting from an inherited high-risk

Other dominant tumour syndromes A number of other important dominantly inherited conditions are now well outlined clinically and genetically. von Hippele Lindau (vHL) syndrome predisposes to retinal angiomas, haemangioblastomas of the cerebellum, renal cell carcinoma, phaeochromocytoma and renal, pancreatic and epididymal cysts. The neurofibromatoses, which consist of at least three subtypes, NF1, NF2 and schwannomatosis, give rise to an increased risk of mainly benign tumours of the nervous system. The chief malignancy risk is in NF1, where the predominant tumour, the neurofibroma, is associated with at least a 10% lifetime risk to each individual of developing malignant peripheral nerve sheath tumours that are usually fatal.8 The second type, NF2, is largely associated with schwannomas and meningiomas with most individuals being deafened by bilateral 8th nerve involvement.

MEDICINE 40:1

31

Ó 2011 Elsevier Ltd. All rights reserved.

CANCER TRIALS AND SERVICES

Hereditary conditions predisposing to breast cancer Gene

Other tumour susceptibility

Population frequency (%)

Proportion of breast cancer (%)

Proportion of HPHBC (%)

Proportion of familial breast cancer risk (%)

Lifetime risk in women (RR) (%)

BRCA1 BRCA2 TP53 LFS PTEN Cowden’s CHEK2 ATM

Ovary/prostate, colorectal Ovary/prostate, pancreas Sarcoma, glioma, adrenal Thyroid, colorectal Colorectal, prostate HoZ (AR) lymphoma, leukaemia Colorectal HoZ-Fanconi (AR) HoZ-Fanconi (AR) HoZ-Fanconi (AR)

0.1 0.1 0.0025 0.0005 0.5 0.5

1.5 1.5 0.02 0.004 0.5 0.5

40 40 2 0.3 0 0

5e10 5e10 0.1 0.02 2 2

50e85 40e85 80e90 25e50 18e20 (2.0) 20

0.001 0.1 0.1 0.001 25e46 100 for any

0.001 0.1 0.1 0.1 0.5 20

0.6 0 0 0 0 83

0.04 0.4 0.4 0.2 20 40

50 20 (2.0) 20 (2.0) 15e20 11e13 (1.1e1.4)

STK11 BRIP1 PALB2 RAD51C 19 SNPs Ref. 5 Totals

AR, autosomal recessive; HoZ, homozygous; HPHBC, highly penetrant hereditary breast cancer (e.g. more than three affected relatives); SNPs, single nucleotide polymorphisms.

Table 3

outcomes from the use of PARP inhibitors in BRCA1/2-related tumours.16

gene defect, although twin studies show that this figure rises to around 30% for any inherited component.9 In addition to family history, cancers are more likely to be hereditary if they are early-onset, bilateral or associated with other related primaries. The discovery of inherited mutations in the TP53 gene10 (possibly the most important gene in cancer) in rare families with a horrific pattern of malignancy was the first major discovery in this area. LieFraumeni syndrome (LFS) predisposes to sarcomas, breast cancer, glioma and other tumours in children and adults. The risk of cancer approaches 100% for female mutation carriers by 50 years of age. Whilst the condition is rare, affecting around one in 40,000 people, it is as common as BRCA1 and BRCA2 combined in very early-onset apparently isolated breast cancer.11

Colorectal cancer The other main tumour site in which high-risk dominant genes play an important role is the large bowel. Approximately 2e3% of colorectal cancer is due to inherited mutations in one of five DNA mismatch repair (MMR) genes (a similar proportion to BRCA1/2 in breast cancer) that cause hereditary non-polyposis colorectal cancer (HNPCC). MMR genes are important in about 13% of colorectal cancer and inheriting a mutation confers about a 30e80% lifetime risk. However, mutations also enhance risks of endometrial, ovarian, gastric and upper urinary tract cancers. Identifying high-risk individuals can also save lives because bowel cancer can be prevented by regular colonoscopy.

Breast/ovarian cancer Breast cancer can occur as part of high-penetrance predisposition such as LFS, and also BRCA1/2, but may also be contributed to by mutations in genes such as ATM, CHEK2 and PTEN, the latter three giving lifetime risks of 20e40% (Table 3). Unsurprisingly, the greatest interest has focused on BRCA1/2, mutations which are each carried by approximately 0.1% of the population (this rises to 2.5% in the Ashkenazi Jewish). This is because of the high lifetime risk of 40e85% of breast cancer with an associated 20e60% risk of ovarian cancer. Many women who test positive for mutations are now opting for riskreducing surgery, and this can reduce risks of both cancers by over 90%, with oophorectomy conferring protection against breast cancer.12e14 The advent of more reliable surveillance with MRI15 and better treatments may reduce the use of mastectomy in future. In the last few years much better risk prediction is possible using information from mammographic density and multiple validated common SNPs (single nucleotide polymorphisms).5 There is also huge promise of improved

MEDICINE 40:1

Where are the remaining genes? Whereas the great majority of high-risk predisposition has been elucidated, there remains a need to discover the remaining polygenic element. It may be several years before ‘generic’ genetic tests for cancer, which test in excess of 100 genes, are developed. In the meantime, there is still benefit to families from identifying the single high-risk genes. Indications on how to assess risks and who to refer for screening and genetic testing are available in more detailed texts.2,16,17

Conclusion The last 21 years has seen an enormous advance in our understanding of cancer and its familial elements. Gene tests are now available in most high-risk conditions and generic tests may soon also be developed. Our increasing knowledge will also give rise to better targeting of surveillance and hopefully more gene-based treatments. A 32

Ó 2011 Elsevier Ltd. All rights reserved.

CANCER TRIALS AND SERVICES

REFERENCES 1 Knudson AG. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci USA 1971; 68: 820e3. 2 Hodgson S, Foulkes WD, eds. Inherited susceptibility to cancer. Cambridge: Cambridge University Press, 1998. 3 Knudson AG, Meadows AT, Nichols WW, Hill R. Chromosomal deletion and retinoblastoma. N Engl J Med 1976; 295: 1120e3. 4 Friend SH, Bernards R, Rogelj S, et al. A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature 1986; 323: 643e6. 5 Turnbull C, Ahmed S, Morrison J, et al. Genome-wide association study identifies five new breast cancer susceptibility loci. Nat Genet 2010; 42: 504e7. 6 Davies DR, Armstrong JG, Thakker N, et al. Severe Gardner’s syndrome in families with mutations restricted to a specific region of the APC gene. Am J Hum Genet 1995; 57: 1151e8. 7 Mallinson EK, Newton KF, Bowen J, et al. The impact of screening and genetic registration on mortality and colorectal cancer incidence in familial adenomatous polyposis. Gut 2010; 59: 1378e82. 8 Evans DGR, Baser ME, McGaughran J, Sharif S, Donnelly B, Moran A. Malignant peripheral nerve sheath tumours in neurofibromatosis 1. J Med Genet 2002; 39: 311e4. 9 Peto J, Mack TM. High constant incidence in twins and other relatives of women with breast cancer. Nat Genet 2000; 26: 411e4. 10 Malkin D, Li FP, Strong LC, et al. Germline TP53 mutations in cancer families. Science 1990; 250: 1233e8. 11 Lalloo F, Varley J, Ellis D, O’Dair L, Pharoah P, Evans DGR, the early onset breast cancer study group. Family history is predictive of pathogenic mutations in BRCA1, BRCA2 and TP53 with high penetrance in a population based study of very early onset breast cancer. Lancet 2003; 361: 1101e2. 12 Hartmann LC, Schaid DJ, Woods JE, et al. Efficacy of bilateral prophylactic mastectomy in women with a family history of breast cancer. N Engl J Med 1999; 340: 77e84.

MEDICINE 40:1

13 Rebbeck TR, Friebel T, Lynch HT, et al. Bilateral prophylactic mastectomy reduces breast cancer risk in BRCA1 and BRCA2 mutation carriers: the PROSE study group. J Clin Oncol 2004; 22: 1055e62. 14 Rebbeck TR, Lynch HT, Neuhausen SL, et al. Reduction in cancer risk after bilateral prophylactic oophorectomy in BRCA1 and BRCA2 mutation carriers. N Engl J Med 2002; 346: 1616e22. 15 Leach MO, Boggis CR, Dixon AK, et al. Screening with magnetic resonance imaging and mammography of a UK population at high familial risk of breast cancer: a prospective multicentre cohort study (MARIBS). Lancet 2005; 365: 1769e78. 16 Stebbing J, Ellis P, Tutt A. PARP inhibitors in BRCA1-/BRCA2associated and triple-negative breast cancers. Future Oncol 2010 Apr; 6: 485e6. 17 McIntosh A, Shaw C, Evans G, et al. Clinical guidelines and evidence review for the classification and care of women at risk of familial breast cancer. London: National Collaborating Centre for Primary Care/University of Sheffield. NICE guideline CG014, 2004 (updated 2006). Also available at: http://www.nice.org.uk.

FURTHER READING Eeles R, Ponder B, Easton D, Eng C, Horwich A. Genetic predisposition to cancer. 2nd edn. London: Chapman and Hall, 2004. Hodgson S, Foulkes WD, eds. Inherited susceptibility to cancer. Cambridge: Cambridge University Press, 1998. Lalloo F, Kerr B, Friedman J, Evans DGR, eds. Risk assessment and management in cancer genetics. Oxford: Oxford University Press, 2005. McIntosh A, Shaw C, Evans G, et al. Clinical guidelines and evidence review for the classification and care of women at risk of familial breast cancer. NICE guideline CG014. London: National Collaborating Centre for Primary Care/University of Sheffield, 2004. updated 2006. Also available at: http://www.nice.org.uk.

33

Ó 2011 Elsevier Ltd. All rights reserved.