Genes and cancer: Genetic counselling and clinical management

Single base pair substitutions and point mutations

DNA replication is the result of an accurate yet accident-prone process. The final accuracy of DNA replication depends on the fidelity of the initial process and the ability of subsequent repair processes to correct any mistakes. Point mutations lead to single changes in the DNA sequence. These single mutations can have very different effects on the reading of a protein, as demonstrated in Figs. 18.10 and 18.11 . If this mutation is in a regulatory portion of a gene, loss or alteration of regulation of gene expression can occur. If the mutation is in a coding portion of the gene, an altered protein can be formed.

One of the most common mutations as the result of substitution occurs in CpG dinucleotides. (CpG is shorthand for 5’—C—phosphate—G—3’ , that is, cytosine and guanine separated by only one phosphate group; phosphate links any two nucleosides together in DNA. The CpG notation is used to distinguish this singlestranded linear sequence from the CG base-pairing of cytosine and guanine for double-stranded sequences.) Data show that these transversions account for about 30% of point mutations. The fidelity of this process is related to the efficiency and accuracy of the DNA polymerase responsible for replication. Other types of these small nucleotide substitutions that can also occur are messenger ribonucleic acid (mRNA) splice junction mutations, translational mutations, and single deletions. Alternatively, mismatches can occur by slipped mispairing, which leads to single base pair mutagenesis and deletion.

Deletion or insertion of a single nucleotide may be secondary to DNA mispairing during replication. This type of mutation often occurs in runs of identical base pairs and is the result of slippage of a single base pair ( Fig. 18.14 ). This mechanism is similar to larger deletions caused by this same mechanism. The significance of this type of mutation to human malignant neoplasia is as yet undefined. As more information becomes available about MMR, we should have a better estimate of the significance of this type of mutation to cancer formation. However, this slippage is common and similar to the errors found in individuals who have an inherited cancer syndrome, hereditary nonpolyposis colon cancer (HNPCC), or Lynch syndrome.

CpG dinucleotides are very susceptible to mutations. Methylation of cytosine results in a high level of mutation because of its propensity to undergo deamination. This then forms thymidine and, therefore can result in a change of gene expression. The evolution of the human genome seems to indicate a progressive loss of CpG islands because of this transition. The frequency of these islands is about 20% to 30% of the predicted frequency, probably because of the progressive loss of CpG dinucleotides. This excess of transition from cytosine to thymine (C to T) transitions was first reported by Vogel and Rohrborn in a study of the mutations responsible for variants of hemoglobin. Others now also have shown that CpG islands are associated with a high rate of transition, supporting the thought that these mutations play a significant role in human disease ( Fig. 18.15 ).

Larger deletions

Gene deletions are responsible for more than 150 inherited diseases. Deletions can range from a few base pairs to several hundred kilobases and are often classified by size. These are also nonrandom in that certain sequences appear to be more prone to deletion than others are. This nonrandomness is apparent when one evaluates genetic disease. Spontaneous deletions often occur in the same sequences of various genes, indicating that some sequences of DNA are more prone to deletion than others.

Deletions occur when there is homologous but unequal recombination between gene sequences. Similar sequences in the human genome can cross over during mitosis or meiosis, resulting in a shortened portion of the gene sequence. Long areas of homology (homology boxes) are thought to be the most likely to have this type of mutation. Repetitive DNA sequences are particularly susceptible to deletions because this can allow slippage. This mechanism is called homolog or unequal recombination between repetitive sequence elements. The most common repetitive sequence is the Alu repeat (up to 106 Alu sequences are in the human genome). Alu repeats are characterized by an average spacing of 4 kilobases and 300-base pair length separated by a short A-rich region. The Alu element is 70% to 90% homologous to the consensus sequence. The 3´ homology (homology boxes) is thought to be the most likely to have this type of mutation. Repetitive DNA sequences are particularly prone to mutation.

There are more than 400 gross gene deletions recorded in the HGMD and 3000 or so smaller gene deletions. Almost 2500 gene deletions of 20 base pairs or fewer have been identified. These can be the result of deletion of repeats or a part of the excision repair process. Deletion is not random; particular sequences are more prone to deletion and often contain regions involving guanine or other repeats. The application of the techniques of molecular biology, such as gene cloning, in situ hybridization, restriction endonuclease mapping of genetic sequences, and polymerase chain reaction analysis of gene transcription, has led to the conclusion that a given chromosome abnormality may be associated with a variety of neoplasms and that a given oncogene may be activated in a variety of human cancers. The most common defects in solid tumors are deletions in specific gene sequences, observed as a loss of part of a banding region or the loss of heterozygosity of a specific genetic allele. Deletion of genetic material in a cancer cell suggests loss of function that regulates cell proliferation and differentiation. Many human solid tumors have been shown to have some type of chromosome deletion. The Tp53 tumor suppressor gene–containing region of chromosome 17p is deleted or mutated in a wide variety of human cancers. The fact that there is such commonality among cancer cell types in the loss of chromosome material suggests that these regions contain genes coding for regulating functions of a wide variety of cell types. Induction of the malignant neoplastic process must involve at least two hits when a tumor suppressor gene is involved. This implies that two mutations or deletions in each of the two alleles are necessary before the effect can be expressed. This is the two-hit theory proposed by Knudson. The possibilities for the two hits necessary to express a recessive gene can be seen in Fig. 18.16 . This explained why some people could inherit a predisposition for a malignant disease and raised the question of whether sporadic cancers could have a genetic basis. In a genetically predisposed cell, the remaining single normal allele may be sufficient to maintain normal growth and regulation; a second deletion or mutation is required to inactivate the remaining normal allele.

Insertions

Most insertions appear to be small, between one and a few base pairs. This phenomenon has been studied extensively, and these insertions often appear to be nonrandom in character. Predisposing factors lead to these insertions. These insertions can be due to slipped mispairings, or they may be mediated by inverted repeat sequences or by symmetric elements. These sorts of DNA sequences are particularly prone to mutation.

Less common are large insertions. The largest insertion known is the 220-kilobase segment inserted into the DMD gene. Alu and long interspersed nucleotide elements (LINE) sequences are especially susceptible to insertion and deletion. Mutations are nonrandom, and some DNA segments are more susceptible to mutation than others are. For example, the myc gene is especially susceptible to insertions; a 50-base pair LINE element has been noted to be inserted in this gene in breast cancer cell lines. The fact that insertional mutations may not be random is underscored by the finding of Fearon and Vogelstein, who reported 10 independent examples of insertions within the same 170-base pair intronic region of the DCC gene involved in the progression of colorectal cancers.

Duplications

Duplication of small parts of whole exons, or even of whole chromosomes, has played a large part in the development of cancer. This phenomenon also has a large role in the evolution of the human genome. Again, the paradox is that mutations are sometimes good and sometimes bad. Duplication of larger segments of DNA may lead to gene amplification. Gene amplifications can occur in many ways. A gene that is usually present in a single copy in the normal cell may be duplicated or may undergo a small increase in the copy number. Another type of amplification may involve a 10- to 100-fold increase in copies of a genetic locus containing key regulatory genes. A third type of amplification can result in the duplication of whole chromosomes. One can envision how amplification has been important over the course of evolution. We have 23 pairs of chromosomes, and all other organisms have fewer chromosomes. This is due to millions of years of evolution resulting from innumerable insertions and duplications.

Inversions

Inversions are perhaps the least common type of mutation. In this mutation, a long segment of DNA forms a loop and alters its direction; the 3´ end now becomes the 5´ end. These segments can obviously be large or quite small. The most important inversion event in human disease involves the factor VIII gene, causing severe hemophilia A. An MSH2 inversion has been reported in exons 1 to 7 in Lynch syndrome and is one of the few cancer-associated inversions.

Translocations

More than 100 common translocations have been observed in malignant cells. Many of these occur consistently in certain specific cancer types, which argues that these areas are involved in the malignant process of the cell. Some of the translocations may be secondary events in the evolution of more aggressive phenotypic changes. The inherent genetic instability of malignant cells leads to further karyotypic abnormalities as the disease progresses, reflecting additional genetic alterations that increase growth potential. Evidence that malignant transformation does not usually result from a single translocation event comes from the study of patients with ataxia-telangiectasia, who are at high risk for leukemia. These patients have lymphocytes with a characteristic translocation present for many years before a malignant change develops.

A common finding in cancers is some form of genetic instability. This is most clearly manifested in Lynch syndrome (or HNPCC). Lynch syndrome is characterized by genetic instability at microsatellite sequences and has been linked to several MMR genes. Mutation rates in cells that are positive for MMR defects are orders of magnitude higher than MMR intact cells.

Cancers in patients with hereditary cancer syndromes may serve as a model for sporadic disease because mutations that occur in the sporadic tumors are often very similar. The same genes and sequences seem to be targeted. For instance, Krawczak and Cooper demonstrated that the bulk of single base substitutions in the Tp53 gene in sporadic cancers strongly resemble those found in inherited cancers.

Oncogenes

The past 25 years have contributed significantly to our understanding of the molecular mechanism of cancer and have identified three types of molecular aberrations that can lead to cancer: dominant transforming genes called oncogenes, recessive transforming genes called tumor suppressor genes, and genes responsible for repairing DNA errors (MMR genes). Many oncogenes have been isolated as forms of proto-oncogenes acquired by RNA tumor viruses. We have known for many years that viruses can cause malignant tumors in animals. The link noted in animals spurred a great deal of research aimed at identifying the cancer-causing genes carried by the viruses and finding the human genes that were affected. These investigations surprisingly revealed that the genes implicated in malignant disease were often altered forms of viral genes. These were probably acquired when the virus infected the animal and then moved on to other animals. At other times, the viruses activated host genes that were normally quiescent. The normal versions of these pirated and activated genes, now termed proto-oncogenes, carried codes specifying the composition of proteins that encourage and stimulate cell replication. These growth-promoting genes come in various forms. Some specify the amino acid sequences of receptors that are found on the cell surface and bind to molecules known as growth factors. When bound by such factors, the receptor issues an intracellular signal that ultimately causes cells to replicate. Other growth-promoting genes code for proteins that lie inside the cell and govern the propagation of intracellular growth signals. A third group encodes proteins that control cell division and are under nuclear control. These oncogenes can be activated via several mechanisms. The oncogene can be amplified, and many copies of the gene can become activated. Occasionally, the gene may be translocated to another chromosome, where, under the influence of another promoter, it promotes uncontrolled growth.

The three classes of oncogenes are controlled differently. The first comprises a group of peptide growth factors and their receptors, such as epidermal growth factor receptor (EDGFR) and platelet-derived growth factor (PDGFR). These peptide growth factor receptors likely serve as stimulatory cofactors rather than the actual force that drives the malignant process. Targeted therapy is becoming a reality as our understanding of these cancer-promoting entities increases. Unique molecules are aimed at gene products or proteins activated by these factors. Another class of oncogenes comes from non–membrane-traversing extranuclear growth factors. These include G-proteins and the ras gene family. Finally, there are oncogenes of the nuclear regulatory variety such as myc. A schematic of oncogenes representing their function can be seen in Fig. 18.19 . A more detailed description of targeted therapy can be found in Chapter 16 .

The discovery that viral genes had human counterparts introduced the intriguing possibility that human cancers, including the majority not caused by viruses, might originate from mutations that convert useful proto-oncogenes into harmful genes—that is, oncogenes. Consistent with this notion, studies indicated that alteration of just one copy, or allele, of these proto-oncogenes was enough to transform and render cancerous some types of cells growing in culture. Such dominant mutations cause cells to overproduce a normal protein or make an aberrant form that is overactive. In either case, the result is that stimulatory signals increase within the cell even when no such signals come from the outside.

It is ironic that research on animal RNA tumor viruses (retroviruses) having no ability to cause human cancer provided the first key to uncovering the identity of some oncogenes and the verification that they actually exist. These retroviruses, which infected chickens, rodents, cats, and monkeys, were extremely tumorigenic in that infected animals often showed tumors on initial exposure to the virus. One of these viruses, the Rous sarcoma virus of a chicken, was found to carry a specific gene that is used to transform infected cells into a malignant state. This type of transforming oncogene was termed a viral oncogene. A single oncogene carried onto a chicken cell by a Rous sarcoma virus was able to derail and redirect the entire metabolism of the cell, forcing it to grow in a malignant fashion.

In 1976, Varmus and Bishop found that the oncogene in the Rous sarcoma virus was not a viral gene at all; instead, it arose directly from a preexisting cellular gene that had been captured by an ancestor of the Rous sarcoma virus. After it was captured, this gene was used by the Rous sarcoma virus to transform mammalian cells. The early ancestor of the Rous sarcoma virus was capable of replicating in infected cells but was unable to transform them; it instantly gained tumorigenic potency by kidnapping this normal cellular gene, the proto-oncogene. In the end, this work by Varmus and Bishop revealed much more about the cell than it did about the Rous sarcoma virus because it pointed to the existence of a gene residing in a normal mammalian genome that possessed potent transforming ability when appropriately activated, in this case by a retrovirus. Here was clear proof of the presence of at least one gene in the normal cellular DNA that might serve as one of the target genes activated by nonviral carcinogens of the retrovirus; such nonviral agents might also activate this proto-oncogene, converting it to a powerful oncogene. A cell carrying such a mutant gene might, in turn, respond to this damage by launching a program of deregulated growth and thus become a cancer cell.

The information uncovered concerning retroviruses and oncogenes provided little immediate comfort to investigators interested in the origins of human cancer. Retroviruses, similar to Rous sarcoma virus, never infected humans and therefore could not act to mobilize human proto-oncogenes. The possible connection came from a notion that many such proto-oncogenes could be activated through an alternative route. Changes in the DNA sequence created by chemical or physical rearrangement might substitute for the virus. Mutations induced by these chemical or physical conditions in the genomes of target cells in one or another tissue might be as effective as retroviruses in activating latent carcinogenic potential in the proto-oncogenes. By the early 1980s, these suspicions were validated: mutated genes (proto-oncogenes) were found in human tumor genomes. In each case, a change in the sequence of the gene was identified as being responsible for converting a proto-oncogene into an active oncogene. For example, a ras oncogene in one human bladder carcinoma was found to have arisen through a single base change that altered the DNA sequence of a precursor proto-oncogene; myc oncogenes arose through gene amplification in various malignant neoplasms.

Next, researchers tried to ascertain how these genes succeeded in transforming cells. A simple theme emerged that makes it possible to understand and explain the mechanism of action of most, if not all, of the oncogenes. This comes through the understanding of how cells regulate their own growth. The growth and division of a normal cell residing in a particular tissue are controlled largely by its surroundings. A normal cell rarely, if ever, decides its own rate of proliferation; rather, it responds to the signals or messages from surrounding cells. These messages, which may carry growth-stimulatory or growth-inhibitory information, are conveyed by growth factors released by surrounding cells, traverse the intercellular space, and bind to the cell surfaces. These cells then respond to these growth signals by activating their synthetic machinery, copying their DNA, and dividing. A normal cell will never commit itself to such a growth program without being simulated by these external signals. Each cell possesses complex machinery that enables it to receive these signals, process them, and launch a growth program. The machinery consists of an array of proteins that function to acquire growth-activating signals and transmit them throughout the cell. These proteins include the following:

• 1.

  Cell surface receptors that recognize the presence of growth factors in the extracellular space and transmit signals into the cell’s interior

• 2.

  Cytoplasmic signal transducers that become activated by these receptors and then pass signals farther into the cell

• 3.

  Nuclear transcription factors that are activated by the cytoplasmic signal transducers and in turn respond by activating entire banks of cellular genes

These activated gene banks together orchestrate the cell’s growth program; they detail events that, acting in concert, enable the cell to grow and divide. Proto-oncogenes encode many of the proteins in this complex signaling circuitry that enable a normal cell to respond to exogenous growth factors. Oncogenes participate in this signaling circuitry by selecting aberrantly functioning versions of the components of this circuitry. Oncogene proteins succeed in activating these signal circuits even in the absence of stimulation by extracellular growth factors. In doing so, they force a cell to grow even when its surroundings do not contain some of the clues that are normally required to provoke growth.

Other researchers suggest that mutations in at least two proto-oncogenes have to be present and that only certain combinations of mutations lead to malignant change. These findings suggest that individual oncogenes, although powerful controllers of cell metabolism, are not capable of causing malignant neoplasms by themselves. Thus, oncogenes cannot explain most cancers by themselves. This view was strengthened by the discovery of more than a dozen different oncogenes in human tumors ( Table 18.2 ). However, on careful evaluation, only about 20% of tumors turned out to carry expected alterations. None of the tumors had pairs of cooperative alterations sometimes found in cultured cells. It also appeared that the inherited mutations responsible for predisposing people to cancer were not oncogenes. The concept of a recessive-type antioncogene was conceived. This type of gene is called a tumor suppressor gene. This class of gene appears to be equally important in the development of cancer.

Tumor suppressor genes

The direct identification of tumor suppressor genes has been more difficult than the identification of oncogenes. Tumor suppressor genes were first conceived of in a theoretical sense long before any were actually identified. Harris was the first to demonstrate that the malignant characteristics of a cell could be suppressed when malignant cells were fused with nonmalignant cells. Furthermore, loss of portions of chromosomes can be associated with malignant transformation, and transfer of a normal chromosome can reverse these traits into the malignant cells. These data suggested a recessive type of control in normal cells. A schematic of a tumor suppressor gene can be seen in Fig. 18.20 .

The first tumor suppressor gene cloned was the RB gene; it is the defective gene in retinoblastoma. Knudson extensively studied the epidemiology of retinoblastoma. Most cases are sporadic, but some cases appear to be clustered around families. He noted that cases of familial retinoblastoma were more likely to be bilateral or multifocal. Based on these data, he developed the two-hit hypothesis that bears his name (see Fig. 18.16 ). That is to say that two mutagenic events are required for retinoblastoma to develop. In patients with inherited disease, the first hit was inherited in one allele of the RB gene in the germline. Although one abnormality was not enough for development of the disease, there was an increased likelihood that the second allele would be damaged during eye development. The hypothesis of two mutational events developed by Knudson applies specifically to tumor suppressor genes. This is the basis of most inherited gene models.

Tumor suppressor genes have proved difficult to identify, and many techniques have been used to locate such genes within the human genome. Cytogenetic studies provide some clues to the location of tumor suppressor genes. A significant percentage of patients with retinoblastoma or some patients with Wilms tumor have large deletions in 13q14 or 11p13, respectively. These are the regions of the respective tumor suppressor genes responsible for each malignant neoplasm. Significant loss on 5q is associated with adenomatous polyps of the colon. This can be measured on a gel during electrophoresis and is called loss of heterozygosity. Although recurrent loss of a chromosome region associated with a specific malignant neoplasm is highly suggestive that this region is important in the development of the cancer, it is not sufficient to prove that there is a tumor suppressor gene in this region. Fig. 18.16 shows the possible combinations of loss that would lead to the expression of a tumor suppressor gene. Use of genome-wide linkage scans can further localize the presence of such a gene. Much smaller areas of deletion can be detected by molecular techniques such as pulsed-field electrophoresis. Now we can also easily and cheaply sequence long segments of the genome as these technologies improve.

When scientists were first trying to identify the regions of loss, they used DNA probes for specific regions of chromosomes. If allelic loss is noted, this is called loss of heterozygosity; if this loss is disproportionate relative to the general population, this suggests a tumor suppressor gene in the region specific for the probe. An example of loss of heterozygosity can be seen in Fig. 18.16 . Here, a region on 10q in endometrial cancers indicates the existence of a possible tumor suppressor gene. Using various markers, a precise location of the tumor suppressor gene can be identified. A contig of the region must then be built and the region sequenced to look for specific genes. Many regions of loss have been identified by this technique. There are many more candidate regions than actual tumor suppressor genes.

The tumor suppressor gene Tp53 is a gene associated with many cancers that was identified by searching for loss of heterozygosity. Several other tumor suppressor genes have been identified since. This is almost certainly only the tip of the iceberg, and many more tumor suppressor genes are likely to be discovered as investigators learn more about cancer cell genetics and as the potential of the Human Genome Project is realized, and new genetic testing technologies continue to bring genetic testing to the forefront of medicine. Several generalizations can be made, and a comparison with the activation of oncogenes is constructive. Mutations in oncogenes are gain of function events and lead to increased cell proliferation and decreased cell differentiation. Oncogene mutations do not appear to be inherited through the germline. In contrast, tumor suppressor gene inactivations are loss of function events, usually requiring a mutational event in one allele followed by a loss or inactivation of the other allele. This loss of gene function leads to loss of cellular control and unchecked growth. Tumor suppressor genes are recessive, and mutations may be inherited as a germline mutation. Somatic mutations may occur in both types of genes and accumulate throughout life. Comparison of oncogenes and tumor suppressor genes is made in Table 18.3 . Common tumor suppressor genes are listed in Table 18.4 .

Apoptosis

Apoptosis means programmed cell death and refers to the “intentional” induction of cell death. Apoptosis is important in the growth and development of an organism because as an organism matures and differentiates, cells must die to give way to more differentiated and specialized cells. If a cell does not die and becomes immortal, a tumor or cancer can result. Apoptosis was first described in the 1970s, but only recently have scientists begun to realize the importance of this phenomenon in organism development, differentiation, and cancer formation. Excitement about this process has been driven by the finding that apoptosis is controlled at the molecular level by genes associated with malignant change (i.e., oncogenes, proto-oncogenes, and tumor suppressor genes). It is also clear that many of these same controlling elements and pathways are active during development. Many believe that understanding the control of the apoptotic process is essential to understanding the control of the developing organism and control of senescence. Loss of these aspects of cellular control could lead to cancer.

Apoptosis is a distinct mode of cell death that is responsible for the deletion of cells in normal tissues; of note, it also occurs in pathologic conditions. The process of apoptosis is characterized histologically by cell contraction, blebbing of the cell membrane, and condensation of the nucleus. Apoptotic bodies are formed that contain intact organelles; surrounding cells eventually phagocytose these. There is no associated inflammatory response. Cell swelling and disintegration of the cell components characterize necrotic cell death, and there is a marked inflammatory response. On a molecular level, apoptotic death causes cellular enzymes to autodigest the genome into little fragments, and this can be identified on polyacrylamide gels by multiple regular fragments seen as a DNA ladder.

Apoptosis is important in the development of the normal organism, and it is important in the development and growth of cancers. It occurs spontaneously in malignant tumors, often markedly retarding their growth, and it is increased in tumors responding to radiation, chemotherapy, heat, and hormone damage. In cancer, apoptosis appears to be a mechanism for deleting cells from the population that have sustained carcinogenic DNA damage; however, when apoptosis of such cells is blocked or inhibited by mutations in genes that help control this process, such as BCL2 or Tp53, these cells are suddenly free to continue replicating and propagating their mutations. This genetic instability may be an early step in the development of cancers. Many of the current cancer treatments, such as radiotherapy and chemotherapy, kill cells by the production of DNA damage. Mutations in BCL2 and Tp53 may then influence the effectiveness of these therapies through their ability to inhibit cell death.

The primary importance of apoptosis related to cancer and cancer treatment lies in its being a regulated phenomenon subject to stimulation and inhibition. Although little is known about how to establish therapeutic agents to affect its initiation, it seems reasonable to suggest that greater understanding of the process of apoptosis might lead to the development of improved treatment possibilities. Inhibitory mechanisms such as BCL2 proto-oncogene expression may be implicated in the development of resistance to therapeutic agents and may contribute to tumor growth and perhaps to oncogenesis by allowing the inappropriate survival of cells with DNA abnormalities. It is likely that other inhibitory mechanisms will be identified, and a better understanding of the apoptotic process may lead to novel treatment regimens by allowing us to control cell death. Apoptosis is not simply a description of cell death, nor is it a spurious trend in the biology literature. It is a fundamental process and is controlled at the molecular level; as such, it can be understood and manipulated. Fig. 18.21 demonstrates how apoptosis may occur.

Mismatch repair defects

A number of mechanisms are involved in the correction of DNA changes that result from exogenous or endogenous mutagenic agents. These mechanisms include base excision repair, nucleotide excision repair, and DNA MMR. The MMR system repairs mistakes that are made in the course of normal cellular division. These mutations often occur in regions of multiple repeats because these regions appear to be particularly prone to DNA slippage. One can intuitively see how this can happen by looking at Fig. 18.22 . Six genes must work in concert to repair DNA damage. These genes and an example of how we think they work can be seen in Fig. 18.23 .

A paradox of human tumor genesis is that the rates of mutation in normal cells are too low to account for the transformation of benign to malignant. The instability of some colon cancers was the first strong evidence for the so-called mutator phenotype. If a cell cannot repair DNA damage that occurs during normal cell division, mutations can continue to accumulate; eventually, enough mutations occur in genes critical for cell control and cancer results. Many studies have shown that tumors, which have microsatellite instability (MSI), have defects in MMR. Key metabolic genes usually do not have large areas of microsatellite repeats that are particularly prone to mismatch problems. The mutation rate in nonrepetitive sequences is also elevated 100 to 10,000 times above that in normal cells with normal MMR. Although microsatellite sequences are largely absent from functional genes, mononuclear repeats in some key genes can lead to their inactivation. Examples of such genes include TGF-BR2H (A ₁₀ ), hMSH3 (A ₈ ), hMSH6/GTBP (C ₈ ), IGF2R (G ₈ ), and BAX (G ₈ ). Paradoxically, instability of chromosome numbers is often observed in tumors without MMR defects; tumors with MMR defects often have stability of chromosome number.

MMR proteins interact to form complexes that recognize mismatched DNA base pairs and abnormal loops of DNA. The repair complexes initiated by MMR genes trigger excision and repair of DNA mutations. This system is critically important in ensuring the fidelity of the replicative process and is highly conserved throughout all living organisms. Areas in the genome that contain simple repetitive DNA base sequences called microsatellites (e.g., CACACACA or AAAAAAA) are especially prone to mutation. If the MMR system does not function normally, these errors will go unrecognized and continue to accumulate. Cells in which both copies of one of the MMR genes have been inactivated exhibit the so-called mutator phenotype; this allows cells to accumulate mutations as they divide. These random mutations eventually accumulate in a series of genes that are important in controlling cellular growth; the result is loss of growth control and transformation to cancer.

Cells with a defective MMR system exhibit a phenomenon called MSI. This phenomenon occurs as DNA mismatches cause shortening or lengthening of repetitive DNA sequences, and these mismatches go unchecked. This results in generation of alleles in the cancer that contain a greater or lesser number of repeats than are present in normal cells from that individual. Laboratory testing for MSI uses an established consensus panel of five microsatellite markers (D2S123, D5S346, D17S250, Bat 25, and Bat 26). Fig. 18.24 shows an example of MSI detection when mismatch-prone regions are compared in blood and tumor.

Germline mutations in MMR genes ( MLH1, MSH2, MSH6, PMS2 ) may present in many families as Lynch syndrome. This syndrome is associated with increased risk for colon cancer, endometrial cancer, ovarian cancer, stomach cancer, small bowel, renal pelvis, biliary tract, and brain cancer. Mutations in one of the Lynch syndrome genes account for about 3% to 5% of all colon cancers and approximately 2% to 3% of all newly diagnosed endometrial cancers.

The Amsterdam II criteria, used to identify Lynch syndrome based on the pattern of cancer in a family, is defined as follows:

• 1.

  Three or more relatives with Lynch syndrome–associated cancer; one must be a first-degree relative to the other two.

• 2.

  At least two successive generations must be involved.

• 3.

  At least one of the cancers must have been diagnosed before the age of 50 years.

Seventy percent of patients that meet Amsterdam II criteria have a germline mutation segregated around an MMR gene: hMLH1 (49%), hMLH2 (45%), and hPMS2 (6%). Approximately 25% of endometrial cancers have an MMR defect; approximately one-fourth of these are germline mutations, and the rest are largely the result of somatic MLH1 inactivation resulting from hypermethylation of the promoter region. MMR will be increasingly investigated and more clearly defined during the next decade.

Telomerase

Normal cells divide; this process is repeated many times as an organism grows and matures. Normal cells stop dividing and terminally differentiate, but cancer cells continue to divide. Cells from younger animals divide more times than do cells from older animals. One difference between normal cells and tumor cells is at the end of their chromosomes. Telomeres, specialized structures at the ends of chromosomes, act as protective caps. In humans, telomeres are made up of 5000 to 15,000 base pairs of TTAGGG repeats. These terminal structures protect the chromosome ends from exonuclease digestion, prevent aberrant chromosome recombination, and form specific complexes that bind proteins.

Normal cells lose about 50 to 100 base pairs from the end of each chromosome every time the cell divides. When a telomere loses a critical number of base pairs, it triggers a signal for the cell to stop dividing and for senescence. Cells have developed several mechanisms to get around this terminal trigger. Some of these are uncommon, such as complex recombination and retrotransport techniques. The most common mechanism is the development of an enzyme complex called telomerase, which adds back telomere sequences lost during replication. Cells that have significant telomerase activity are often immortal cells such as cancer cells and germ cells.

Normal cells lack telomerase activity. Telomerase becomes reactivated in most cancer cells. It is the most prevalent cancer marker known. Therefore, telomerase may be used as a generic cancer marker and as a possible treatment.

Clinically relevant hereditary syndromes

The advent of linkage analysis and the rapidity of DNA sequencing with other molecular technologies have made it possible to identify many inherited disorders that we could not identify a few short decades ago. Gynecologic malignancies associated with autosomal dominant inherited cancer predisposition that have been identified thus far are as follows:

• Ovarian cancer

  • Epithelial

    • Hereditary breast and ovarian cancer (HBOC): all epithelial histologies; rarely mucinous; most commonly high-grade serous

    • Lynch syndrome: all epithelial histologies, typically endometrioid

  • Epithelial: borderline tumors

    • Rarely associated with hereditary cancer syndromes

  • Germ cell

    • Not believed to be a part of HBOC

  • Sex cord–stomal tumors

    • Granulosa cell tumors

      • Peutz-Jeghers syndrome (PJS)

    • Sex cord tumors with annular tubules

      • PJS

  • Sertoli-Leydig and juvenile granulosa cell tumors

    • Pleuropulmonary blastoma (PPB) syndrome

• Fallopian tube or primary peritoneal cancer

  • HBOC

  • Lynch syndrome

• Uterine cancer

  • Epithelial (majority of uterine cancers)

    • Lynch syndrome (typically endometrioid, less commonly clear cell; rarely uterine papillary serous (UPSC): not typically thought to be associated with HBOC (reports of UPSC associated with HBOC)

    • Mesenchymal (endometrial stromal sarcoma, leiomyosarcoma, other nonspecific sarcomas) (rare)

    • Rarely associated with hereditary cancer syndromes

    • Uterine soft tissue sarcoma

    • Li-Fraumeni syndrome

  • Mixed: Malignant mixed Mullerian tumor (MMMT) or carcinosarcoma, adenosarcoma

    • Rarely associated with hereditary cancer syndromes; seen in some cases of Lynch syndrome

• Cervix

  • Rarely associated with hereditary cancer syndromes

  • Cervix adenoma malignum associated with PJS

• Vaginal or vulvar cancer

  • No known association with hereditary cancer syndromes

• Gestational trophoblastic disease

  • No known association with hereditary cancer syndromes

• Hereditary cancer syndromes that include increased risk for developing gynecologic cancers.

Commercial clinical genetic testing is available for hereditary cancer risk and has undergone rapid changes in the last several years. Massively parallel sequencing, also called next-generation sequencing (NGS), technology allows for rapid testing of large amounts of DNA sequence. NGS is quicker and less expensive than previous genetic testing tools. It is becoming routine to order a single test that simultaneously evaluates multiple genes, thus testing for multiple hereditary cancer syndromes at once. The Supreme Court Decision in June 2013, overturning the ruling on allowing gene patents, particularly BRCA1 and BRCA2 , brought an immediate change in the availability of multigene panel testing through multiple different commercial laboratories.

Our understanding of the tumor spectrum associated with the different hereditary cancer predisposition genes, the risks for cancer, and targeted therapy options is growing exponentially. An increasing number of individuals are seeking the advice of physicians and genetic counselors regarding options for genetic testing hereditary risk for cancer. It is important to understand hereditary cancer syndromes associated with an increased risk for developing gynecologic cancers.

As genetic testing has become much more accessible, it is possible that in the clinical setting, a patient may be coming to see the provider for a previously known pathogenic mutation. There are additional considerations when interpreting genetic test results and managing patients who have a “positive” genetic test result. It is important to review a physical copy of the genetic test results. Additional or confirmatory genetic testing should be offered in certain cases based on the following considerations:

• First and foremost, it is important to confirm that the testing performed was through a laboratory that is Clinical Laboratory Improvement Amendments (CLIA) and College of American Pathologists (CAP) certified. (Some states may have additional requirements.)

• Direct to consumer testing is becoming more popular in the general population. Such tests usually employ single nucleotide polymorphism testing using microarrays. These tests are not clinically validated.

• Patients may use raw data from such testing and have it analyzed through third-party software for interpretation. As the raw data is unreliable and not validated, interpretations based on such data may be incorrect

• Patients may have had limited testing from a CLIA/CAP certified laboratory that did not include all the candidate genes.

• Patients may have had previous negative tumor testing. Somatic testing tries to detect targetable variants, a mutation in a significant gene may not be reported as it is deemed clinically insignificant for the purpose of the test ordered.

• Some laboratories offer paired germline and somatic genetic tests. Such testing may have incomplete coverage. Although germline mutations may be detected on such testing, it is reasonable to confirm the presence of these mutations by additional specific germline testing. This may especially be a consideration in individuals who present with a mutation that is inconsistent with their personal and/or family history.

• Patients may have participated in research studies that identified a significant pathogenic mutation. It is equally important to review the clinical validity of such reports and review the findings with a genetics professional

Genetic tests are unique in that they affect the individual and may also affect the offspring and other relatives. Thus, any patient offered genetic testing needs counseling regarding the impact of the results of such testing on their future life and other members of the family. The process of genetic testing has thus been separated into three parts, viz. pre-test counseling, genetic testing, post-test counseling.

Pre-test counseling

• Assessment of patient’s personal history of cancer—This includes evaluating the age at diagnosis, presence of any other tumors, the histologic type and stage of tumor. Any prior treatments, surgeries and biopsies should also be noted.

• Evaluation of patient knowledge of genetics and concern regarding cancer risk

• Family history assessment—Providers should take a detailed three generation family history from patients who are interested in genetic testing. All diagnoses of cancer and the age at diagnosis should be noted while taking such a history.

• Ancestry—Patients with Ashkenazi Jewish ancestry and a diagnosis or family history of breast and/or ovarian cancer are at a higher risk for having BRCA1/2 mutations. Direct to consumer ancestry testing results should be taken into consideration.

• Assessment of patient’s past medical and surgical history—As some genetic syndromes may present with additional findings (dermatologic findings, benign tumors in other areas, personal history of colon polyps, physical findings). Reproductive history, prior environmental exposures, and history of previous risk reducing surgeries should also be taken into consideration

• Risk assessment—Using the family history and the patient’s personal medical history, a differential diagnosis can be generated. It is important to counsel the patient regarding the genetic syndrome that is suspected and the chance of finding a pathogenic mutation significant for that syndrome.

• Psychological assessment and support—A diagnosis of cancer or the risk for a future diagnosis of cancer can bring up anxiety and other emotional responses in the patient. Such reactions are normal and even expected. The possibility of having an emotional response to genetic test results must be discussed with the patient. Assessing whether an unaffected patient, who is testing to define cancer risk, understands the psychological impact of such testing is important.

• Possible test results—There are three types of results possible when pursuing a genetic test.

  • Positive result—This indicates the presence of a pathogenic or likely pathogenic mutation in a cancer predisposition gene. Identification of such a mutation may lead to further testing, screening, and changes in treatments. As positive results have an impact on other family members, patients should be informed regarding the importance of sharing these results with other family members.

  • Variant of uncertain significance (VUS) result—As the name suggests, a VUS is a change identified in the gene which has no established significance. This means that there is limited evidence to suggest that this change impacts the gene or consequent protein function. If a panel includes preliminary genes, the chances of identifying a VUS are higher as these genes have not been characterized fully. Most, greater than 85%, of VUS results are reclassified as benign polymorphisms, or genetic changes that do not impact disease risk

  • Negative result—This indicates that the test did not identify any clinically significant mutations. Patients need to be informed regarding their treatment and surveillance options in the absence of a hereditary predisposition.

• Informed consent—Informed consent should be obtained from the patient and documented in the patient’s medical record. As discussed previously, this includes a discussion of the purpose of genetic testing, possible test results, and the implications of these results. In addition, patients should also be informed regarding the accuracy of the test, the possibility of false-negative or false-positive results, and the changes in management, including but not limited to prophylactic surgeries and contraindication for certain therapies (such as radiation therapy).

• Cost and billing—A lot of patients are more concerned regarding the cost of genetic testing. Though genetic testing has seen an increase in uptake, many patients, if not most, think of genetic tests as cutting-edge science that is expensive. Genetic tests in the oncology realm are available for as low as $250 out of pocket (without insurance). Providers can refer patients back to a billing specialist within the institution or with the testing lab.

The actual process of ordering a germline genetic test is relatively straightforward. If a patient has an active or recent hematologic malignancy or has had an allogenic bone marrow transplant, germline genetic testing cannot be performed using blood, saliva, or buccal samples. Extracted DNA from fibroblast cultures can be used for these patients.

In an ideal world, it is recommended that the affected individual with the youngest age at diagnosis should be tested first in families with extensive histories of cancer. However, we live in a world that is less than ideal. Historically unaffected patients would be asked to communicate the importance of genetic testing to their affected relatives. Such strategies are fraught with negative outcomes, as affected relatives may not be interested in testing or may not have the financial means to do so. Sometimes affected individuals are deceased and/or not available for testing. In such instances, it is important to discuss the implications of testing an unaffected individual but continue with testing for that patient. There is difficulty associated with interpreting test results when an unaffected family member is the first to be tested. The presence of a negative test result in an unaffected relative does not eliminate the possibility of a gene mutation in the family. It is unknown whether the negative test result occurred because our technique cannot detect the mutation in the family or because the unaffected person did not inherit a detectable mutation. There is still a chance that other family members may have a mutation in a cancer predisposition gene.

Ordering genetic tests

• Purpose of testing

  • If the goal is to identify chemotherapy targets, somatic testing is recommended

  • In the presence of a significant family history or clinical presentation indicative of a specific hereditary syndrome, single-gene testing or targeted-panel testing should be considered.

• Genes included in the panel

  • A patient may have a family history that may be indicative of more than one genetic syndrome.

  • Some panels include moderate-risk genes or preliminary evidence genes that have no established disease association or will not lead to change in management.

• Candidate for genetic testing

  • When ordering genetic testing for an unaffected individual for a family history of cancer, ordering a broader panel is advised

  • Broader panels may also make sense for patients who present with limited family history information (small family, limited number of female relatives, adopted out)

• Timing of genetic testing

  • If genetic testing is being ordered to advise treatment decisions (specifically, regarding radiation therapy), some laboratories offer rapid turnaround times for limited number of genes.

  • Initiating a referral for genetic counseling or performing genetic testing immediately after diagnosis can help unaffected family members take steps to manage their risk for cancer especially as the overall survival rate for patients with ovarian cancer is less than 50%. Affected patients may succumb to the disease and cannot be tested.

• Choice of testing laboratory

  • Some laboratories offer to reflex limited panels to include a broader set of genes at no extra cost

  • Different laboratories have different “extras” or additional testing strategies. Currently, some laboratories offer paired somatic and germline testing with targeted mutational analysis on the somatic level to identify targets for chemotherapy.

  • Recently, certain laboratories have started offering RNA sequencing. This technology is limited and is not available for the analysis of all genes included in the panel.

  • Laboratories that have been performing genetic testing for many years have a gigantic clinical variant database that can impact variant interpretation. In fact, all genetic testing laboratories, though they follow basic standards for variant interpretation, rely heavily on their internal databases and additional available technologies.

  • Cost of genetic testing, insurance coverage, and availability of financial assistance are important considerations when choosing a laboratory for genetic testing.

  • Some laboratories offer free testing using single-site analysis for at-risk family members.

Post-test counseling

Post-test counseling involves a thorough discussion of the genetic test results in the context of the patient’s personal and family history. Changes in medical management and screening should be discussed with the patient at length. Some genetic syndromes have multi-system presentations; therefore, referrals to other providers and specialties should be made when appropriate. Patients may also qualify for clinical trials and novel therapeutic options based on their genetic results.

One important consideration for post-test counseling is testing for family members. Testing of at-risk family members is known as cascade testing. Some laboratories offer free testing for family members within a time frame. Thus, patients can be encouraged to bring family members who are most at risk (siblings, adult children, parents) to post-test counseling visits. In the event of a positive result, family members who are present at the visit can be consented for genetic testing at the time of result disclosure.

It is probable that patients with a known pathogenic variant are hesitant to share that information with other family members. They may not have contact information for those family members or may not be on speaking terms with some of their family. Patients should be counseled regarding steps that unaffected family members may be able to take based on such important genetic information. Providers can offer the option of sending anonymous letters to inform family members regarding the presence of a hereditary predisposition to cancer in the family. Such letter templates are available online and can be modified based on institutional standards.

Family members should not be tested for VUSs identified on genetic testing unless such testing will help elucidate the significance of the identified variant. Testing laboratories usually offer insight regarding the need for testing other family members to help classify the identified variant.

Hereditary cancer predisposition syndromes confer a risk for malignancy when a mutation is identified on one copy of the gene. Biallelic mutations in some genes (both copies of the gene are rendered non-functional by mutations) are associated with risk for autosomal recessive disease. These genes and the specific risks will be discussed further in this chapter. It is important to discuss this risk for future pregnancies. Patients of reproductive age and those planning future pregnancies should be counseled regarding this risk. Partners of these patients should be offered genetic counseling and testing.

Hereditary breast and ovarian cancer ( BRCA1 and BRCA2 ; other DNA repair pathway genes)

HBOC syndrome is the best known and most well characterized of the hereditary cancer syndromes. BRCA1 and BRCA2 were identified in the early 1990s (1994 for BRCA1 and 1995 for BRCA2 ). Genetic testing for the predisposition to breast and ovarian cancer by testing for mutations in the BRCA1 or BRCA2 genes became clinically available in 1996. The BRCA1 and BRCA2 genes are located on chromosomes 17q and 13q, respectively ( Figs. 18.25 and 18.26 ). Both genes are quite large, containing at least 20 exons (coding regions), and are of at least 7000 base pairs (see Fig. 18.26 ). Genetic testing for BRCA1 and BRCA2 has changed significantly since 1996; the initial testing in 1996 only included sequencing for BRCA1 and BRCA2 . This testing was later updated to include five specific BRCA1 rearrangements; deletions or duplications not detected through gene sequencing. In 2006, the addition of deletion and duplication (also called large rearrangement testing) for both BRCA1 and BRCA2 was automatically run for individuals from high-risk families who met certain family history criteria or as an optional add-on for additional cost. Current testing includes sequencing and deletion and duplication analysis automatically for both genes in every test, regardless of personal or family history. Individuals with ovarian cancer who had genetic testing for BRCA1 and BRCA2 prior to 2006 should be offered updated genetic testing, especially if no mutations were identified.

Founder mutations are mutations that are common or even unique to a specific patient population and occur in populations of patients who have been genetically isolated for many generations. The population may have shrunk because of a disaster, but the mutation was preserved in a few individuals. When the population expands, the frequency of that specific mutation can be quite high within a population ( Fig. 18.27 ). One example of founder mutations occurs in BRCA1 and BRCA2 ( Fig. 18.28 ). Approximately 2% of individuals of Ashkenazi (Eastern European) Jewish ancestry have one of three founder BRCA mutation: 187delAG ( BRCA1 ), 5385insC ( BRCA1 ), and 6174delT ( BRCA2 ). Historically, targeted testing for the specific founder mutations was initiated first with reflex testing to full BRCA1 and BRCA2 gene testing if none of the three mutations were identified. The adoption of NGS tools for genetic testing is changing this process, and the specific founder mutation testing is skipped in favor of proceeding directly to multigene testing. These founder mutations are identified with the NGS technology.

Cancer risks for individuals with BRCA1 or BRCA2 mutation are presented as lifetime risk ranges. These are typically broad risk ranges because the actual reported risk varies depending on the population studied in any given report. Reports of lifetime risk for ovarian cancer ranges from 8% to 62%. The age of onset for ovarian cancer is not necessarily an indication of hereditary risk. Reported lifetime risk for breast cancer range from 41% to 90%. There is limited data to provide individualized risks or mutation-specific risks. Unlike ovarian cancer, early age of onset for breast cancer is a hallmark of hereditary risk, but later ages of onset for breast cancer should not be discounted. This should be considered in the context of the cancer family history when determining concern for inherited risk. Additionally, triple-negative breast cancer seems to have a greater association with BRCA1 mutations.

Additional genes in the DNA repair BRCA-FA (Fanconi anemia) pathway were reported to confer increased risk for breast and ovarian cancer beginning around 2010. Genes in the DNA repair BRCA-FA (Fanconi anemia) pathway include BRCA1, BRCA2, CHEK2, BRIP1, RAD50, RAD51C, RAD51D, PALB2, BARD1, MRE11, and NBN. Current information suggests a moderately increased risk for cancer for individuals with mutations in these genes as compared with the higher risks reported with mutations in BRCA1 or BRCA2 . Some of these genes within the BRCA-FA DNA repair pathway may turn out to be associated with a more significant risk for developing cancer than others, and clinical availability of testing these genes will also change over time.

A 2011 study of 360 women with primary ovarian, peritoneal, or fallopian tube carcinoma not selected for age or family history found 24% carried pathogenic mutations in 12 different genes, including BRCA1, BRCA2 BARD1, BRIP1, CHEK2, MRE11A, MSH6, NBN, PALB2, RAD50, RAD51C, and TP53 . The majority were attributed to mutation in BRCA1 or BRCA2 . The remaining portion of inherited risk is due to mutations in one of the Lynch syndrome genes or one of the newly identified BRCA-FA DNA repair pathway genes. The overall contribution of each gene to the risk of ovarian cancer can be seen in Fig. 18.29 . A 2015 study of 1915 women with primary ovarian, peritoneal, or fallopian tube carcinoma not selected for age or family history found 18% carried a pathogenic germline mutation and suggests that 11 genes comprise the majority of hereditary risk for ovarian cancer: BRCA1, BRCA2, BRIP1, RAD51C, RAD51D, PALB2, BARD1, MLH1, MSH2, MSH6, and PMS2 . Data continue to emerge to sort out the significance of these and other BRCA/FA pathway genes. Data are quickly evolving regarding the frequency of these mutations associated with specific cancer types, the cumulative risk for cancer, and the most appropriate medical management recommendations for cancer detection or prevention for women with hereditary mutations in these genes.

Cumulative risk is the concept of total risk over time. For individuals with a mutation in a gene involved in a DNA repair pathway, mistakes cannot be adequately repaired that occur with normal DNA replication in cellular division as well as with environmental damage. Therefore, there is a greater risk for developing cancer over time compared with an individual at average risk, and the development of cancer tends to occur at earlier ages. Cumulative risks of breast and ovarian cancer associated with mutations in BRCA1 and BRCA2 are shown in Fig. 18.30 .

Pathologic characteristics of tumors have been studied as a method of determining which patients should be referred for hereditary cancer risk assessment. For example, there is a preponderance of triple-negative (ER-/PR-/her2-) breast cancers associated with BRCA1 , so that historically, women with breast cancer under the age of 50, those with triple-negative breast cancer, and those with an extensive family history were advised to undergo BRCA1/2 testing. However, guidelines have shifted to broader inclusion criteria, in that some societies recommend panel testing for all women with breast cancer so as not to miss more moderate-risk penetrance gene mutations. High-grade serous ovarian cancer is most commonly associated with mutations in BRCA1 or BRCA2 . However, there are reports of mutations identified in individuals with any of the histologic subtypes. Mucinous ovarian cancers are rarely associated with mutations in BRCA1 or BRCA2 . Current recommendations by the Society of Gynecologic Oncologists (SGO) and National Comprehensive Cancer Network (NCCN) suggest referral of all patients with epithelial ovarian cancer for genetic counseling and genetic testing.

Identification of germline BRCA mutations and the other BRCA-FA DNA repair genes is also becoming more important in treatment decisions. Patients with BRCA1 or BRCA2 mutation–associated cancers have longer recurrence free and overall survival compared with patients with no identifiable germline mutations, and this extends to moderate-risk genes within this DNA repair pathway. We can now profile ovarian cancers into two categories: homologous recombination deficient (HRD) versus proficient and predict response not only to carboplatin-containing chemotherapy but also poly (ADP-ribose) polymerase inhibitors (PARP inhibitors). Loss of heterozygosity is another marker of DNA repair stability of the ovarian cancers. Cancers with homologous recombination deficiency have been found to be more susceptible to chemotherapeutic agents that cause DNA damage, as these cancer cells do not have the machinery to repair double-stranded DNA repair. These tumors then upregulate PARP to aid in single-stranded DNA repair, but the addition of a PARP inhibitor exploits the breakdown of this DNA repair pathway, and these tumors then exhibit synthetic lethality. Synthetic lethality arises when a combination of deficiencies in the expression of two or more genes leads to cell death. There are currently three PARP inhibitors FDA-approved for clinical use in patients with ovarian cancer. The degree to which these target different ovarian cancers based on germline and somatic mutation status is covered in more detail in Chapter 15 . In a similar fashion to MMR deficiency testing for endometrial cancer, some now argue that all ovarian cancers should be profiled somatically first, then triaged to germline testing only if a somatic mutation in an HRD gene is identified. Caveats here are that large-gene rearrangements may not be adequately found by next-generation sequencing, and the cost of and access to somatic testing may be prohibitive to some.

Lynch syndrome (hereditary nonpolyposis colon cancer)

In 1895, a pathologist at the University of Michigan, Aldred Scott Warthin, had a seamstress who had a family with a preponderance of gastric and endometrial cancers. This family became known as family G and is part of the basis for the discovery of Lynch syndrome or HNPCC. Warthin described this family in a report in the Archives of Internal Medicine in 1913 ( Fig. 18.31 ). The family pedigree was updated in 1925 but was largely forgotten until Henry Lynch and colleagues discovered two other families with similar pedigrees in Iowa and Nebraska, which they reported in the 1970s. This was proposed as the “cancer family syndrome” (CFS) in 1971, later changed to HNPCC. The current terminology refers to this syndrome as Lynch syndrome, believing that HNPCC fails to adequately acknowledge the importance of extracolonic malignancies such as endometrial, ovarian, upper gastrointestinal (GI) tract, and urinary tract cancers.

Lynch syndrome is inherited in an autosomal dominant pattern and results from a mutation in one of the DNA MMR genes, MLH1, MSH2, EPCAM, MSH6, or PMS2 . These genes code for proteins that proofread DNA and correct mistakes that are made during the normal process of replication. MMR is covered in detail earlier in this chapter.

Mutations in MSH2 and MLH1 account for the majority of cases of Lynch syndrome. Mutations in MSH6, EPCAM, and PMS2 account for a smaller percentage. EPCAM and MSH2 are located on 2p21, MSH6 on chromosome 2p16, MLH1 on chromosome 3p22, and PMS2 on chromosome 7p22 ( Fig. 18.32 ). Large germline deletions of the last exon of EPCAM , located upstream from MSH2 , were reported in 2009 in 5 of 27 families suspected of having Lynch syndrome who previously tested negative for MLH1 or MSH2 . Analysis of the 3´ region of EPCAM is now standard in genetic testing for Lynch syndrome.

Endometrial cancer is the most common gynecologic cancer in the United States and accounts for 6% of all gynecologic cancers in women. This currently accounts for over 65,000 cancers per year. Approximately 5% of endometrial cancers are caused by inherited risk, with the majority of this 5% caused by Lynch syndrome. A list of cancer types associated with Lynch syndrome and a comparison of the increased risk with average risk are depicted in Fig. 18.33 and Table 18.5 . The cumulative lifetime risk of endometrial cancer is between 40% and 60%, and the ovarian cancer risk is approximately 12%. The cumulative risk of colon cancer is 40% to 80%. There appear to be lower cumulative risks associated with MSH6 and PMS2, perhaps with later ages of onset. Up to 5% of all cases of endometrial cancer and colon cancer are caused by inherited mutations in one of the Lynch syndrome genes. The most common histologic subtype of endometrial cancer associated with Lynch syndrome is endometrioid; uterine papillary serous carcinoma is the least likely subtype type to be associated with Lynch syndrome. Approximately 2% to 3% of cases of epithelial ovarian cancer are caused by inherited mutations in one of the Lynch syndrome genes.

Family history criteria used to classify a family as having Lynch syndrome are the Amsterdam I and Amsterdam II criteria. Amsterdam I criteria only included colon cancer. Amsterdam II criteria, shown in Table 18.6 , include other Lynch syndrome–associated cancers. Colon cancers associated with Lynch syndrome have a propensity to occur in the ascending colon, relative to sporadic cases, which are more often distal. Individuals with Lynch syndrome tend to have shorter intervals in formation of adenomatous polyps and transformation to malignancy, as well as earlier ages of diagnosis.

Genetic testing for lynch syndrome

Historically, the recommendation for genetic testing for Lynch syndrome was determined by cancer family history ( Table 18.7 ). In families that fulfill the Amsterdam II criteria, the likelihood that a mutation will be found is quite high; the chance of identifying a mutation in one of the MMR genes is between 50% and 92%. The Bethesda criteria were developed to provide guidelines to order MSI for colon tumors ( Table 18.8 ). There are no similar guidelines for endometrial cancer for guidance on when to order tumor genetic screening.

TABLE 18.7

Criteria for Referral to a Genetic Counselor for Suspected Lynch Syndrome

Bethesda guidelines Amsterdam Criteria fulfilled Persons with two relatives with Lynch syndrome Colorectal cancer and first-degree relative with a Lynch cancer before 45 years or a first-degree relative with an adenoma before 40 years of age Colorectal cancer or endometrial cancer before 45 years of age Right-sided, undifferentiated colon cancer before 45 years of age Signet ring colon cancer before 45 years of age Adenoma before 40 years of age

TABLE 18.8

Revised Bethesda Guidelines for Testing Colorectal Tumors for Microsatellite Instability

Tumors from individuals should be tested for MSI in the following situations: 1. Colorectal cancer diagnosed in a patient who is younger than 50 years of age 2. Presence of synchronous, metachronous colorectal, or other HNPCC-associated tumors a regardless of age 3. Colorectal cancer with the MSI-H histology diagnosed in a patient who is younger than 60 years of age 4. Colorectal cancer diagnosed in one or more first-degree relatives with an HNPCC-related tumor, with one of the cancers being diagnosed before age 50 years 5. Colorectal cancer diagnosed in two or more first- or second-degree relatives with HNPCC-related tumors, regardless of age

MSI, Microsatellite instability.

a Hereditary nonpolyposis colorectal cancer (HNPCC)–related tumors.

Recent collaborative efforts have focused on the adoption of universal tumor screening (UTS) for all newly diagnosed endometrial and colon cancers to include immunohistochemistry (IHC) and MSI. The Lynch Syndrome Screening Network ( http://www.lynchscreening.net ) was formed in 2011 to enable ongoing cooperative efforts toward UTS for all patients with newly diagnosed colon and endometrial tumors. Updated information regarding implementation of a universal screening program is available on their website. A 2014 consensus statement from US Multi-Society Task Force recommends that all newly diagnosed colorectal cancers should be tested for MMR deficiency either by IHC testing for MLH1, MSH2, MSH6, and PMS2 and/or MSI testing. Colon tumors that have absent staining of MLH1 or MSI - high should undergo BRAF testing of MLH1 promoter hypermethylation to determine if the abnormal MMR screen is most likely attributable to somatic inactivation of MLH1 . Similar guidelines have been proposed for all endometrial cancers. However, BRAF mutations are not commonly seen in endometrial cancers, so reflex testing should be for MLH1 promoter hypermethylation if IHC shows absent staining for MLH1 . Approximately 20% to 25% of all endometrial cancers exhibit MSI as a result of epigenetic inactivation of MLH1 caused by hypermethylation of the promoter.

NCCN guidelines version 2.2015 for Lynch syndrome provide a detailed grid of tumor testing results and additional testing strategies based on those results. A typical screening algorithm is seen in Fig. 18.34 . Recent literature suggests 70% of women with Lynch syndrome who develop endometrial cancer do not met any published criteria for Lynch syndrome. MSI and/or IHC testing is recommended in patients who have cancers suggestive of Lynch syndrome unless there is a previously known mutation. It is important to note, however, that 25% of sporadic endometrial cancers exhibit MSI.

Education regarding the importance of reporting abnormal uterine bleeding and postmenopausal bleeding is important in women with Lynch syndrome. Prophylactic hysterectomies may help reduce the incidence of endometrial cancer. The timing for surgery can be determined on a case-to-case basis. Consider screening these women by performing endometrial biopsies every 1 to 2 years beginning from age 30 to 35. Similarly, prophylactic bilateral salpingo-oophorectomies may be considered in women who are concerned about their ovarian cancer risk. There are no specific guidelines regarding the timing of this surgery.

Biallelic mutations in either MLH1, MSH2, MSH6, or PMS2 genes cause a condition called constitutional mismatch repair deficiency. This is a childhood cancer predisposition syndrome. The cancer risks associated with this syndrome are broad and include risk for tumors in brain, intestinal tract, and hematological malignancies. Partners of patients who test positive for Lynch syndrome should be referred for genetic counseling and testing. Muir-Torre syndrome and Turcot syndrome are clinical variants of Lynch syndrome. Individuals with Muir-Torre syndrome have higher risks for developing skin lesions such as sebaceous adenomas, carcinoma, epitheliomas and keratocanthomas. Individuals with Turcot syndrome have a 1% to 2% risk for developing brain tumors such as glioblastomas.

As many as 7 of 10 individuals with Lynch syndrome who develop endometrial cancer do not meet any published criteria for Lynch syndrome, and many of these cancers are diagnosed after age 50. Multiple studies evaluating the feasibility of implementing an IHC-based screening protocol for Lynch syndrome in endometrial cancer show this detects Lynch syndrome that would have been previously missed and increases the referral and uptake for genetic counseling and genetic testing.

If there is no tumor tissue available, then personal and family history of cancer can guide recommendations for genetic testing. Referral of the endometrial cancer patient is discussed later.

With the increased adoption of NGS, genetic testing usually includes multiple genes tested simultaneously. This is true for Lynch syndrome as well. Genetic testing based on cancer family history consistent with Amsterdam II should include all five Lynch syndrome genes. Given the overlap of hereditary cancer syndromes, including colon cancer and other cancer types, testing may include multiple hereditary cancer syndromes, including Lynch syndrome. Genetic testing based on abnormal IHC findings typically involves two or three of the Lynch syndrome genes because of the formation of heterodimers. There have been reports of IHC showing loss of just a single protein, yet the germline mutation is found in the companion genes

A typical pedigree of a family meeting the criteria for a Lynch syndrome family can be seen in Fig. 18.35 . An example of a pedigree of a family that does not raise suspicion for Lynch syndrome but was identified through abnormal IHC and subsequent genetic testing is shown in Fig. 18.24 .