Cancer Genetics [M.Tevfik DORAK]

Genetics Evolution HLA MHC Epidemiology Genetic Epidemiology Population Genetics Glossary Homepage

CANCER GENETICS

M.Tevfik Dorak, M.D., Ph.D.

IARC: Most types of cancer NOT due to “bad luck” referring to Tomasetti & Vogelstein, 2015

Reviews on Genetic Predisposition to Cancer: Dong et al, 2008; Turnbull et al, 2005; Frank SA, 2004; Krepischi, 2012; Rahman, 2014 (Book: Eeeles et al, 2004)

Reviews on GWAS and Cancer Susceptibility: Easton & Eeles, 2008; Stadler et al, 2010; Ioannidis et al, 2010; Varghese & Easton, 2010; Chung, 2010

Taby & Issa, 2010: Cancer Epigenetics Dorak MT: Cancer Immune Surveillance Presentation

WHO Report: Genetics in Prevention and Treatment of Cancer

Cancer has some genetic component in its etiology. Environmental component, however, appears to be stronger (Lichtenstein 2000; Czene, 2002; Le Marchand, 2005; Hemminki, 2005 & 2006). The genetic component is evident from increased risk in family members of cancer patients, twin studies, the presence of familial forms of cancer, experimental studies and classical/molecular genetic studies. In the majority of human cancers, however, there is no clear-cut mode of inheritance. This and other evidence suggest that cancer is a multigenic (oncogenes, tumor suppressor genes, MHC genes), multifactorial (radiation, chemicals, hormones, viruses, ultraviolet, diet, etc) and multistep (transformation, promotion, overt cancer) process. In sporadic tumors, there is no inherited genetic abnormality predisposing to tumor. Strong genetic predisposition occurs when a mutation is present in the heritable genetic material (germ-line). About 5-10% of cancers result from germ-line mutations and most are evident as cancer family syndromes or hereditary cancers (Garber & Offit, 2005). In majority of cancers, genetic component is not strong and there is no inheritance pattern or familial aggregation. Therefore, it would be wrong to think of cancer as a genetic disorder in general despite that cancer cells contain innumerable somatic mutations in genetic material.

Knudson proposed a two-hit model for the development of cancer in 1971 from the evidence in familial retinoblastoma cases. This simplified model suggests two separate genetic events in the development of retinoblastoma. The first one is inherited in the familial form and the time required for the second hit to occur is shorter than two hits to occur in a person lacking the first germ-line mutation. This is why the familial form occurs at an earlier age and is usually bilateral. In general, inherited cancer occurs earlier and is more severe, whereas, sporadic cancer has later onset. (The original Knudson's two-hit hypothesis based on bi-allelic gene inactivation has now been extended to include transcriptional silencing by epigenetic changes (such as DNA methylation of promoters) that can disable tumor-suppressor genes (Jones & Laird, 2000), and haploinsufficiency or dominant-negative mutation that can inactivate the whole gene function. See Balmain, 2003; Paige, 2003; Payne & Kemp, 2005; Taby & Issa, 2010).

The immune system is one of the factors influencing susceptibility to cancer. The evidence for this is that in immune deficiencies, there is an increased risk for cancer. Immunogenicity of cancer is suggested by the presence of tumor-specific cytotoxic T lymphocytes (CTL) and tumor infiltrating lymphocytes (TIL). Immunogenicity is impaired by the loss of MHC expression. The immune surveillance theory proposes that oncogenic transformation of normal cells is a frequent occurrence but the immune system clears them as they emerge. The cell surface expression of MHC molecules is a major regulator of this function (Meruelo 1979, 1980; Tanaka, 1985, 1986). The immune surveillance notion has been well supported (Dunn, 2002; 2003; Nakachi, 2004; Zitvogel, 2006; Mittal, 2014), and particularly the role played by natural killer (NK) cells has been shown repeatedly (Nakachi, 2004; Hayashi, 2006; Hoglund, 2006). A Japanaese study (the Saitama prospective cohort study) has reported the correlation between NKG2D haplotype variants and NK cell cytotoxic activity (Hayashi, 2006). This activity that appears to be linked to NKG2D genetic variation is inversely correlated with cancer development risk (Imai, 2000). The immune surveillance theory implies that homozygosity for the MHC molecules would be deleterious. Several modes of cancer treatment are based on the role played by the immune system (in vitro activation of CTLs and TILs, vaccination with tumor-derived peptides and even gene therapy using TILs).

Cancer results from uncontrolled cellular proliferation and immortality (lack of apoptosis, which is a mechanism of the organism that regulates embryogenesis and development, maintains homeostasis of the immune system and removes potentially hazardous cells. A dysregulation of apoptosis signaling may disturb the balance of cell survival and cell death). Cell division is regulated by growth factors, their cell surface receptors, membrane tyrosine kinase, signaling molecules (GTP-binding proteins), nuclear/transcription factors (the signal transduction system) and growth regulatory (inhibiting) factors. The genes with growth promoting activity are generally called oncogenes (dominant) and those with growth inhibitory activities are tumor suppressor genes (recessive). Proto-oncogenes may encode surface membrane proteins (HER2/NEU/ERBB2), signal transduction pathway molecules (HRAS) or transcription factors. Tumor suppressor genes encode cell cycle regulators, adhesion molecules (FAP1), DNA repair enzymes (MSH2, MLH1) or signal transduction pathway molecules. Cancer promoting genes are activated during carcinogenesis due to an extremely increased rate of somatic mutation which results from combined effects of genetic and environmental factors (Bielas, 2006).

Germ-line versus somatic mutations

Germ-line mutations are those people are born with and those that exist in all cells (with few exceptions of mosaicism). High penetrance germ-line mutations cause hereditary cancers. In this group, more than 100 genes have been well characterized to have them (Rahman, 2014). These genes are called cancer predisposition genes, which may be proto-oncogenes or tumor suppressor genes, and the list includes TP53, BRCA1, BRCA2, RB1, CDKN2A, CDKN1B, MSH2, MSH6, MLH1 and ATM. The high penetrance mutations in these genes exert a very strong effect on gene function, which may be a gain-of-function type converting a proto-oncogene to an oncogene, or a loss-of-function type in a tumor suppressor gene. Hereditary cancers caused by these mutations are rare (<5% of total cancer burden), characterized by earlier-onset compared with their sporadic (non-genetically caused) counterparts and a more aggressive clinical course. A typical feature of these cancers is the family history and a Mendelian inheritance pattern (autosomal/X-linked dominant or recessive). Other common germ-line polymorphisms that are found associated with cancer susceptibility in genetic association studies do NOT cause cancer; they are neither necessary nor sufficient for cancer development; and they are low-penetrance variants. These polymorphisms are not disease-causing mutations and only correlate with a slightly increased risk for cancer development. Neither the presence of cancer predisposition genes nor the risk associations of these polymorphism make cancer a genetic disease. A large majority of cancers belong to the multifactorial (complex) disease category.

On the other hand, each and every cancer cell contains thousands of mutations that are not seen in their non-neoplastic counterparts (Lawrence, 2014). These are called somatic mutations that have occurred in the cancer cells after the initial transformation event. These mutations are due to genomic instability that develops within the transformed cell and promote progression to cancer. In the development of somatic mutations, both genome and epigenome are involved (Shen & Laird, 2013). COSMIC (Catalogue of Somatic Mutations in Cancer) database lists genes known to contain somatic mutations. Of the nearly 500 genes listed in the database, around 10% of them are known to be cancer predisposition genes. There is, therefore, some overlap between germ-line cancer predisposition genes and genes that are somatically mutated in cancer, but the two concepts are substantially different. A large genomics study compared the characteristics of somatic mutations in cancer cells with inherited germ-line polymorphisms and found that 99% of somatic mutations involving single nucleotides occur in non-coding regions, and somatic cancer mutations are enriched for functionally deleterious ones (Khurana, 2013). The Authors concluded that somatic variants in the noncoding genomic elements (most commonly transcription factor motif breaking) under strongest selection are the most likely to be cancer drivers.

Genes involved in oncogenesis

The human genome contains more than 50 genes (cellular oncogenes, c-onc) that are similar to the genes carried by carcinogenic retroviruses (viral oncogenes, v-onc). The cellular counterparts of v-onc are normal cellular genes coding for proteins playing roles in normal cell growth and division. They are called proto-oncogenes, but when activated they become oncogenes and show their oncogenic effects. Examples of oncogenes include sis which encodes a chain of platelet derived growth factor (PDGF); int-2 have similarities to fibroblast growth factor (FGF); c-erb-B (on chromosome 7p) which encodes a truncated form of the receptor for epidermal growth factor (EGFR); mutant ras genes have a reduced capacity to terminate a growth stimulating signal; abl and src have tyrosine kinase activity; c-myb and c-myc are stimulators of cell cycle; and bcl-2 blocks apoptosis and promotes cell survival. One of the earliest phenomena in tumor formation is genomic instability. It is due to defects in DNA repair and cell cycle controls. This can happen by gain-of-function mutations in proto-oncogenes or loss-of-function mutations in tumor suppressor genes.

a. Proto-oncogenes may be activated to act like oncogenes by the following events:

1. Amplification: Int-2 and c-erb-B2 in breast cancer; N-myc in neuroblastoma.

2. Insertional mutagenesis: c-myc (nuclear DNA-binding phosphoproteins) activation by EBV in Burkitt lymphoma (Int3 activation by MMTV in mouse mammary tumors). c-myc may be activated by rearrangement, amplification or overexpression.

3. Chromosomal translocations: In Burkitt lymphoma, c-myc (8q24), in 85% of follicular lymphoma, bcl‑2 (18q21), and in mantle-cell lymphoma bcl-1 (11q13) translocate to the immunoglobulin heavy chain gene region on chromosome 14q32 (subsequently these oncogenes are over expressed); in chronic myeloid leukemia, the bcr gene moves next to c-abl [t (9;22) (q34; q11)] resulting in the expression of a fusion protein. In M3 AML (acute promyelocytic leukemia), RARa / PML genes are rearranged by translocation t(15,17).

4. Mutations in coding sequences: Activating point mutations occur in ras genes in about 30% of human cancers. ras gene family consists of N-ras, H-ras and K-ras whose products are involved in intracellular transduction of external stimulation of growth factor receptors. Mutated RAS proteins are expressed but functionally appear to have lost the ability to be inactivated. In general, coding region mutations changing the activity of a gene may be missense, nonsense or frameshift type.

Gene expression and function can also be affected by non-DNA base pair changes, namely epigenetic effects:

5. Demethylation or hypomethylation: bcl-2 overexpression in chronic lymphoid leukemia; c-erb-B1 (epidermal growth factor receptor) overexpression in breast cancer.

A number of genes involved in regulation of the cell cycle, cyclins, may also act as oncogenes when mutated (for example, cyclin D2 in mantle cell lymphoma and cyclin D1 and E in breast cancer). On the other hand, germ-line mutation of a cyclin-dependent kinase inhibitor CDKN2A (p16), a tumor suppressor gene, has been implicated in hereditary melanoma.

b. Tumor suppressor genes (TSGs) act like recessive anti-oncogenes. Theoretically, both copies of a TSG must be lost or inactivated for oncogenesis through loss-of-heterozygosity events (unlike a single mutation in proto-oncogenes). There are, however, exceptions to this expectation. The best-known TSG is TP53 (on chromosome 17p) and mutation of a single copy of the two copies is enough for the deleterious effect. This is because mutant TP53 protein monomers are more stable than the normal TP53 proteins and can form complexes with the wild type TP53 acting in a dominant-negative manner to inactivate it. Therefore, one mutated copy is enough for the loss of whole TP53 function. The tumor suppressor gene TP53 is the guardian of the genome because it prevents progress through the cell cycle when there is something wrong to allow the damage or fault to be repaired. Because mutation of a single allele (single hit) is enough for neoplastic transformation, TP53 mutations are the most frequent genetic abnormalities found in human cancers (over 50% in bladder, breast, colon and lung cancers). Different mutations differentially occur in specific cancers (like the codon 249 mutation in aflatoxin-related liver cancer). Inherited (germ-line) TP53 mutation is the cause of a well-known inherited cancer syndrome (Li-Fraumeni syndrome; childhood sarcomas, early-onset familial breast cancer and other neoplasms). Most familial cancers are related to defects in TSGs.

Other TSGs are the retinoblastoma (RB1) gene on chromosome 13q, the deleted in colorectal cancer (DCC) gene on chromosome 18q, and the Wilms tumor gene (WT1) on chromosome 11p13. In these cases, the loss of both alleles is required for neoplastic transformation (via mutation or deletion). The RB1 gene encodes a nuclear protein that is involved in the regulation of the cell cycle (it suppresses growth). Various viral proteins interact with the RB protein and inhibit its action (adenoviral E1A, HPV-E7 protein and SV40 large T antigen). The DCC gene is involved in cell adhesion and is deleted in over 70% of colorectal cancer cases. WT1 is a transcription factor for normal kidney and gonadal development. Like most TSGs, TP53 and RB1 are also transcription factors.

Alterations in the patterns of DNA methylation are a common genomic alteration in human cancer. Abnormal methylation of CpG islands in the promoters of TSGs (such as RB1, von-Hippel Lindau gene VHL and CDKN2A) can contribute to their functional inactivation as one of the two hits (for a review of DNA methylation in neoplasia, see Rountree, 2001). The more conventional way of TSG inactivation is its loss due to deletion or mutation. The loss-of-heterozygosity events may be genomic deletions that discard the normal copies of TSGs. Inherited abnormalities of TSGs are associated with familial cancer syndromes that cause a variety of cancers at an early age. It is in fact a hallmark of familial cancers that they occur at earlier ages than sporadic forms of same cancers (Brandt, 2008; Lindor, 2008; Mai, 2009; Giambartolomei, 2009; Phuong, 2009).

The following summary outlines the role played by oncogenes and tumor suppressor genes in cancer development (Stanford Health Care - Understanding Cancer Genetics):

- Mutations in proto-oncogenes are usually acquired. The mutations in the RET proto-oncogene can be inherited and cause multiple endocrine neoplasia type II.

- Having a mutation in one copy of a proto-oncogene is usually enough to cause a gain-of-function to change in cell growth and transformation of a healthy cell. Thus, oncogenes are said to be "dominant" at the cellular level.

- Both copies of a tumor suppressor gene need to be mutated to cause a loss-of-function to change in cell growth and transformation of a healthy cell. Thus, tumor suppressor genes are said to be "recessive" at the cellular level.

- Like proto-oncogenes, mutations in tumor suppressor genes are usually acquired. Mutations may also occur as the result of aging and/or environmental assaults.

- Germline tumor suppressor gene mutations are also possible. When a child is born with en existing mutation in one copy of the gene, it takes shorter to acquire a mutation in the second copy for loss-of-function, hence the earlier onset of inherited forms of cancers.

- Most of the genes associated with hereditary cancer types are tumor suppressor genes (with acquired mutations).

c. DNA repair: DNA repair involves base-excision repair (BER), nucleotide excision repair (NER) and mismatch repair (MSH). While oncogenes act as accelerators of growth during G1 phase of cell cycle, and suppressor genes act as stop signals during S phase, DNA repairs gene are not directly involved in cell growth. Their role is in DNA mismatch repair during replication of DNA just before the chromosomes condense in G2 phase for mitosis. These are DNA damage response genes and their defects result in predisposition to a range of tumor types (usually skin or colon cancer and hematological malignancies). Mutations in DNA mismatch repair genes represent more of a predisposition state than a transforming event. BRCA1 and BRCA2 are DNA repair genes. Mutations in critical oncogenes and TSGs are more likely to occur in repair-deficient cells. One mechanism for inactivation of one of the DNA repair genes (MLH1) is silencing via promoter methylation (as in sporadic colon cancer). The best-known DNA repair genes MSH2 and MLH1 are involved in hereditary non-polyposis colon cancer (HNPCC1 and HNPCC2, respectively) when mutated. PMS1 is mutated in HNPCC3, PMS2 in HNPCC4, MSH6 (mutS homolog 6) in HNPCC5 (HNPCC6 is due to mutations in TGFBR2). DNA repair genes are referred to as 'caretakers' while the genes involved in cell cycle control are called 'gatekeepers' (See Kinzler & Vogelstein, 1997). Microsatellite instability (MSI) is a hallmark of DNA mismatch repair defect (Duval & Hamelin, 2002). For a review of polymorphisms of DNA repair genes and associations with cancer risk, see Wood et al, 2001 and Goode et al, 2002. See also DNA Repair Mechanisms in the Molecular Biology Web Book, and DNA Damage and Repair and Their Roles in Carcinogenesis in Molecular Cell Biology. For a review, see Jefford & Irminger-Finger, 2006.

d. Growth factors: They regulate cellular proliferation through receptor mediated autocrine or paracrine mechanisms. Transforming Growth Factor-a (TGFa) is one of them (ligand for epidermal growth factor receptor-EGFR). It is over expressed in 50% of invasive breast cancer cases. Insulin-like Growth Factor-1 (IGF1) is a physiologic mediator and stimulator of normal cell growth. Most breast cancers express receptors for it.

e. Other genes: Ataxia-telangiectasia (ATM) gene confers increased risk for leukemia and lymphoma in homozygous subjects for its mutations. BRCA1 and BRCA2 are involved in familial breast cancer predisposition (hereditary form is no more than 5% of all breast cancers). More recently, microRNAs have gained importance in cancer genetics and their (or their binding site) polymorphisms are becoming useful markers of cancer genetic risk in molecular epidemiology (Chen, 2008, Iuliano, 2013; see also the microRNA database miRBase that lists known human miRNAs).

Genetic prediction

Having a gene mutation/variant may be necessary, but most of the time is not sufficient for developing a disease. Penetrance is defined as the proportion of individuals with a specific genetic alteration who will express the associated trait. While mutations causing Mendelian disorders have high penetrance, genetic variants modifying risk for complex diseases have low penetrance. Thus, genetic tests are not always perfect predictors of health risks. The BRCA1 mutations, for example, cause familial breast cancer in 60% of women by the age of 60 (45-90% risk for life-time). The detection of a mutation in this gene (>100kb) does not necessarily indicate an absolute risk. Examples of high penetrance are polyposis-associated colon cancer and RET (MEN2A)-associated medullary thyroid cancer (penetrance close to 100%).

Pharmacogenetic studies deal with the polymorphisms in xenobiotic enzyme genes. One of them, CYP1A1 is involved in the activation of polycyclic aromatic hydrocarbons (PAH). The susceptibility allele of CYP1A1 increases the risk of lung cancer in smokers, whereas, those lacking the PAH activating allele are relatively protected from the carcinogenic effects of smoking. A striking example of gene and environment interaction has been shown in mesothelioma cases in central Turkey (Carbone, 2007). There is also parental imprinting effect in inherited cancer predisposition. Paraganglionoma occurs only if the mutation on chromosome 11 is inherited from the father. In neurofibromatosis type 2, however, children of affected females show earlier and more severe symptoms than children of affected males.

Genetic anticipation refers to the younger age or increased severity of a disease in successive generations. This phenomenon is better known in some non-malignant diseases caused by unstable triplet repeats (as in Huntington disease and congenital myotonic dystrophy), but is also rarely seen in familial leukemia and ovarian cancer. Like anticipation and imprinting, sex influence and mosaicism also cause a non-Mendelian inheritance pattern in familial cancers. Mosaicism refers to the presence of the normal (wild type) genotype as well as the abnormal (mutant) genotype in the germ-line of an individual. In this case, the parent will not have the phenotype but the offspring will. Germ-line mosaicism has been observed in retinoblastoma.

One important issue about genetic prediction of cancer is genetic heterogeneity. There may be a number of distinct genotypes associated with the same phenotype. Absence of a known genotype causing a particular cancer may not mean the lack of genetic predisposition to that cancer. Similar to the effect of genetic heterogeneity, phenocopies also confound pedigree analysis in cancer families. Phenocopy is a trait that appears to be identical to a genetic trait but that does not have a genetic basis. Sporadic form of a tumor in a cancer family may make the interpretation of the pedigree difficult.

Genetic therapy in cancer

Antisense therapy (for HPV in cervical cancer); targeted cytotoxic treatment against fusion proteins; somatic gene therapy (using carrier vectors incorporated with a toxin, an enzyme activating the cytotoxic agents; using DNA/liposome complex containing a foreign MHC antigen; or TILs infected with retroviruses carrying TNF or IL-2); immunization with tumor-specific proteins (TP53, fusion proteins) or idiotypic antibodies in B-cell malignancies have been used with variable success in clinical trials, but not yet established as routine treatment options for widespread use.

BOOK: Lynch HT & Hirayama T: Genetic Epidemiology of Cancer, CRC, 1989

NCI Cancer Genome Anatomy Project NCI Cancer Genetics Overview

Wellcome Trust Sanger Institute: Cancer Genome Project

HuGE Navigator > Cancer GAMAdb GWAS Catalogue

NCI Division of Cancer Epidemiology and Genetics Cancer Genetic Markers of Susceptibility (CGEMS)

Breast and Prostate Cancer and Hormone-Related Gene Variant Study by BPC3

American Cancer Society: Cancer Facts and Figures

NCBI Bookshelf: Cancer Medicine 6th Edition NCI: Understanding Cancer Series

Concise Handbook of Familial Cancer Susceptibility Syndromes

What Makes a Cancer Cell a Cancer Cell? in Cancer Medicine

Oncogene Activation and Loss of Tumor-Suppressor Function & Genetic Basis of Cancer Syndromes in Cancer Medicine

Genetic Predisposition to Cancer in Cancer Medicine