66.1.1

Basic Mechanisms

Understanding the process of neoplastic transformation has been an important research focus. It has been determined that, at its most basic level, cancer is a genetic disease. Mutations in a cell’s genomic material disrupt critical gene functions that normally serve to regulate cell proliferation, programmed cell death (apoptosis), and cell differentiation ( Table 66.1 ). The result is disordered, unregulated cell growth. Laboratory based and clinical observations have identified several different processes that can give rise to neoplasia. They comprise the following: (1) hereditary predisposition to cancer, (2) exposure to carcinogens, (3) chronic inflammatory conditions, and (4) sporadic mutations and epigenetic changes that accumulate during a normal lifetime.

66.1.2

Heredity

A genetic predisposition for gastrointestinal cancers exists in some families due to the inherited transmission of a mutant gene from a parent to a child. Organs primarily affected by classic familial cancer syndromes include the esophagus, stomach, intestine, colon, and the pancreas. In the liver and bile ducts, inherited metabolic disorders can cause the development of hepatocellular and cholangiocarcinomas. Despite the specific differences between the cancers and the genes involved, familial cancer syndromes do share certain features in common. First and foremost, they must have a significantly increased risk for a particular cancer in the absence of any other predisposing factors. Second, they typically develop multiple tumors, and they occur at a younger age than in the sporadic forms of these cancers. Penetrance can be variable, but is typically greater than 50% in affected individuals, and therefore multiple family members are affected across several generations. Lastly, the predisposition for cancer may affect a single organ or involve multiple tissues, but the pattern of involvement is typically conserved within a family kindred.

The study of families with a preponderance of gastrointestinal malignancies has provided novel insights into the mechanisms governing neoplastic transformation. Frequently, genes identified in familial cancer syndromes are found to be mutated or silenced in the more common sporadic forms of the cancers as well ( Table 66.2 ). The major inherited familial colon cancer syndromes, familial adenomatous polyposis (FAP), and Lynch syndrome, together account for approximately 5% of all colon cancers. However, identifying the predisposing inherited mutations illuminated gene pathways whose mutation is critical in the development of the majority of sporadic colon cancers. The gene mutations responsible for the more common familial cancer syndromes have largely been identified ( Table 66.2 ). Together with advances in DNA sequencing technologies, this has led to wholesale changes in the guidelines for testing for inherited colon cancer syndromes.

The molecular events that precede cancer development, even in cells with an inherited predisposition, remain poorly characterized. The presence of the mutation by itself is not transforming; the second normal allele must be inactivated through allelic deletion, mutation, or silenced by gene methylation. Thus, biallelic inactivation of tumor suppressor genes is critical to the genesis of tumors. Yet, somatic alterations in other genes are typically required as well.

Several features of hereditary cancers remain difficult to explain. While every cell in the body harbors the same mutation, the cancers typically develop in a limited set of tissues. This suggests that these tissues may possess vulnerabilities not present in others, but the mechanisms for this have yet to be determined. Moreover, cancer penetrance can vary between family kindreds, possibly due to the actions of unknown modifier genes. Therefore, much remains to be learned about the process of carcinogenesis, even in cells with a genetic predisposing mutation.

66.1.3

Carcinogen Exposure

DNA, the basic agent for transmission of genetic information, is a chemical. As such, its composition can be altered by physical and chemical reactions. Highly reactive chemical carcinogens, such as alkylating agents, can be directly genotoxic. However, most agents must first be metabolized in a way that generates highly reactive electrophilic derivatives, which can in turn react with DNA and are called procarcinogens. The resulting chemical modifications to the DNA base can disrupt normal nucleotide pairing. If not corrected prior to cell division, daughter cells can permanently inherit sequence changes that alter gene function and expression.

Given these risks, cells have evolved a number of mechanisms to protect DNA from environmental mutagens, as well as identify and correct mutations when they occur. Additionally, cells are programmed to undergo apoptosis if the DNA is severely mutated, thus protecting the organism from defective and potentially harmful cells. Despite these protective mechanisms, the gastrointestinal tract is exposed continuously to chemical carcinogens, often due to ingestions. For example, Aspergillus molds contaminate food staples in developing countries. These molds can produce aflatoxin, a potent DNA mutagen, which when ingested and metabolized by hepatic cytochrome P450 enzymes, becomes highly reactive and forms adducts with guanine nucleotides. Significant dietary aflatoxin exposure in individuals with chronic-active hepatitis B infection yields a synergistic increase in hepatocellular cancer rates. Tobacco smoke contains over 40 different chemical mutagens and carcinogens that cause damage to DNA nucleosides. Smoking has been strongly associated with increased rates of several gastrointestinal cancers, including head and neck and esophageal squamous cell cancers. Nitrates, which are a common preservative found in processed meats, can be converted into the genotoxic N-nitrosamines by the actions of gastric bacteria and acid. High nitrate consumption has been associated with increased risk for gastric and esophageal cancers. In summary, carcinogen exposure is an important cause of sporadic gastrointestinal neoplasia.

66.1.4

Inflammation

Chronic inflammatory conditions have long been associated with increased rates of neoplasms. Typically, this risk increases the longer the inflammation is present, but can be reduced if the inflammatory response is suppressed. The causes of chronic inflammation in gastrointestinal tissues are varied and include infections such as Helicobacter pylori (chronic gastritis) or the viral hepatitis B and C (chronic hepatitis), recurrent chemical (gastroesophageal reflux disease—GERD) or enzymatic injury (recurrent pancreatitis), as well as autoimmune processes (Crohn’s disease, ulcerative colitis). Despite differences in the underlying processes, each of these chronic inflammatory conditions is known to increase the risk for cancer in patients with the particular disorder.

There are several mechanisms by which chronic inflammation induces neoplastic transformation. First, activated inflammatory cells elaborate high levels of reactive oxygen and nitrogen species (ROS and RNS). These highly reactive oxygen and nitrogen species can in turn react with and damage DNA, RNA, lipids, and proteins, causing mutations and altered protein function, ultimately leading to neoplastic transformation. Second, the cytokines and chemokines secreted by activated inflammatory cells also promote carcinogenesis. Cytokines, such as TNFα, IFNγ, and IL-6, and chemokines, such as CXCL8 and CXCL12, are secreted by inflammatory cells and in many studies promote the growth and metastasis of tumors ( Fig. 66.2 ). These immune mediators can act in both a paracrine and an autocrine fashion to induce cell proliferation, inhibit apoptosis, promote cell migration and stromal degradation, and enhance new blood-vessel growth (angiogenesis). Third, tumor-derived cytokines and chemokines can also modulate the immune response to inhibit normal surveillance mechanisms that target neoplastic cells and promote leukocyte infiltration, causing degradation of stromal elements and enhanced neoplastic cell migration and metastasis.

Inflammation in gastrointestinal cancers. ROS and reactive nitrogen intermediates (RNI), produced by inflammatory cells, can cause DNA damage that in addition to other mutagens can initiate cancer. Cytokines from inflammatory cells can increase intracellular ROS and RNI in premalignant cells, resulting in epigenetic changes that silence tumor suppressors and promote tumor initiation. miRNA, microRNA; ROS, reactive oxygen species; RNI, reactive oxygen intermediates.

(From Terzic J, Grivennikov S, Karin E, Karin M. Inflammation and colon cancer. Gastroenterology 2010; 138 (6):2101–14 e5.)

Lastly, inflammation and inflammatory cells also release eicosanoids, oxygenated lipids produced by arachidonic acid metabolism. Eicosanoids in general cause vasodilation and increase blood vessel permeabilization, which induces edema and the extravasation of immune cells. Eicosanoids have been associated with promoting transformation by enhancing cell proliferation and motility while reducing apoptosis and promoting angiogenesis. Eicosanoids can also induce cytokine and chemokine synthesis and release, thereby further potentiating the inflammatory response. Critical products of eicosanoid biosynthesis are reactive lipid hydroperoxides. One such reactive lipid is malondialdehyde (MDA). MDA is a potent mutagen as it reacts with DNA nucleosides to form DNA adducts. Similarly, hydroxyperoxyoctadecadenoic acids (HPODEs) are an abundantly produced product of Cox-2 activity. These decompose into the genotoxic lipids 4-hydroxy-2(E)-nonenal (4-HNE) and 4-oxo-2(E)-nonenal (4-ONE). At high levels, these reactive lipids form DNA adducts and damage DNA, altering gene expression and inducing cancer. Due to these many beneficial effects on tumor growth and survival, increased eicosanoid biosynthesis can be frequently seen in many cancers.

66.1.5

Sporadic Carcinogenesis

The majority of gastrointestinal cancers are sporadic in nature that are acquired during an individual’s lifetime. Discerning the mechanisms underlying sporadic GI cancers has been difficult. Typically, sporadic cancers occur later in life. This has led to the hypothesis that such cancers develop from cells that have accumulated a lifetime’s worth of random DNA mutations culminating in neoplastic transformation. While this remains an attractive hypothesis, there are now multiple observations reporting that certain exposures, habits, diets, and backgrounds increase or decrease the risk for certain gastrointestinal cancers. Western high-fat diets are associated with increased risk for colorectal cancer, while high-salt diets elevate the risk for gastric cancer. Obesity, in addition to the many other associated health concerns, also increases the risk for esophageal adenocarcinoma and colorectal cancer. Some practices can potentially be preventative. Several population-based studies have shown that frequent use of aspirin or nonsteroidal antiinflammatory drugs (NSAIDs) can reduce rates of esophageal, gastric, colon, and possibly pancreatic cancer. Lastly, genetic polymorphisms may contribute to this process by modestly increasing susceptibility to dietary and environmental carcinogens, or altering the responsiveness of critical regulatory pathways. In summary, the molecular mechanisms responsible for sporadic neoplasms are likely to be highly complex and vary considerably between individuals.

66.1.6

Gastrointestinal Carcinogenesis Is a Multistep Process

Gastrointestinal cancers are typically preceded by benign dysplastic intermediates; they do not arise directly from normal tissues. These dysplastic lesions can be distinguished morphologically and classified based on specific pathologic criteria. In the colon, the adenoma-carcinoma sequence describes this progression from normal mucosa through dysplastic intermediates to invasive carcinoma and has been well supported by many pathologic, epidemiologic, and animal studies. The earliest distinct lesion in this sequence is the aberrant crypt foci (ACF). ACF are microscopic lesions, first identified in methylene-blue stained colons of azoxymethane-treated mice as crypts which appeared larger and thicker, with an increased luminal diameter and an opening that was often slit-like or serrated. Most human ACF are hyperplastic (65%–95%), however, a significant proportion are dysplastic and similar in many ways to adenomatous polyps.

Similar multistep progression from normal tissue through dysplastic intermediates to cancer has been well described for human esophageal, gastric, and pancreatic cancers. In each case, cancers arise in dysplastic precursor lesions that are grossly or histologically apparent. Current models suggest intestinal-type gastric cancer is preceded first by atrophic gastritis, then intestinal-type metaplasia, before giving rise to the dysplasia and adenomas in which the carcinomas develop. Similarly in the esophagus, Barretťs intestinal-type metaplasia is the earliest recognized tissue abnormality. Precursor lesions to pancreatic cancer have been formally agreed upon and the characteristics necessary for their classification established. These criteria have permitted the classification of pancreatic lesions for both clinical and scientific uses, greatly aiding pancreatic cancer research.

The ability to identify intermediate stages in the progression from normal tissues to cancer would seem to lead naturally to the hypothesis that carcinogenesis is a multistep process involving sequential genetic changes. An important observation in support of the multistep hypothesis is that neoplasms are clonal, with each neoplastic cell derived from a single progenitor. This model proposes that the genetic mutations required for neoplastic transformation are not obtained all at once, but rather in stages. With each step in this process, the transforming cell acquires a new mutation that enhances cell proliferation or survival. By a process of natural selection or evolution, there is the emergence of a cell clone possessing all the features necessary for neoplastic transformation.

Selection is a critical component of this process, as mutational events are random and therefore only rare mutations lead to activation of growth-promoting and cell-survival pathways or inactivate tumor suppressors and apoptotic pathways. It is these mutations that provide a selective growth and survival advantage to that cell and its progeny. This results in the expansion of that cell into a clonal population. Subsequent mutations occur in cells of that clonal population and thereby endow a few rare cells with new advantages. Recent genetic studies have suggested that in some cancers this progression is not “linear” and gradual but instead may be “branched,” and occur more rapidly than otherwise anticipated ( Fig. 66.3 ). While the “linear” model may better describe the emergence of sporadic colon cancer, with the formation of discreet polyps, developing over years, preceding the onset of colon cancer, the latter “branched” model may better describe the pathogenesis of esophageal adenocarcinoma, in which rapid progression to cancer has been more commonly observed.

Linear and branched evolution of cancer. (A) The linear model of disease has dominated medical thinking about early detection for more than a century. In its recent versions, it has postulated that a slow, gradual linear occurrence of molecular abnormalities (1, 2, 3, 4) cause changes in tissues (A, B, C, D) before the onset of cancer. This model predicts that interrupting any event (e.g., B) in the linear pathway will prevent progression. (B) Recent advances in genome technologies have reported that cancers arise by “branched evolution.” In some cases, such as in BE, an early branch leads to a state in which the esophageal metaplasia can remain stable for prolong periods of time, even though it has some genomic alterations (B). However, in other BE, progression is branched. In this case inhibiting one step (e.g., E) will not necessarily block progression, which can proceed through C → D.

(From Reid BJ, Paulson TG, Li X. Genetic insights in Barrett’s esophagus and esophageal adenocarcinoma. Gastroenterology 2015; 149 (5):1142–52 e3.)

66.2.1

Proto-Oncogene Activation

The study of DNA viruses and retroviruses that cause tumors in humans and animals provided significant insight into gain-of-function processes that lead to carcinogenesis. In these studies, a single viral gene was frequently identified that could lead to neoplastic transformation of normal cells, and these genes were called oncogenes. The first example of this was the v-src oncogene of the Rous sarcoma virus. Subsequently, it was learned that normal cells contain genes with sequences very similar to the viral oncogenes. These normal genes were termed “proto-oncogenes” or “cellular oncogenes” and were determined to play roles in promoting cell proliferation and regulating differentiation. Unlike the viral proteins, the activity of these proto-oncogenes is tightly regulated in normal cells. More significantly, studies of human tumors revealed that these proto-oncogenes were frequently mutated in cancers such that they were inappropriately active, much like their viral oncogene counterparts. This would allow cells to maintain sustained proliferative signaling, which is one of the hallmark features of cancer as outlined by Hanahan and Weinberg. The first confirmed human oncogene was mutant HRAS , isolated from a human bladder cancer cell line, but since then there have been numerous additional human oncogenes isolated, and new candidate genes are being identified on an ongoing basis.

Several different kinds of mutations can be predicted to induce proto-oncogene activation, and all have been observed in human cancers. These mutations lead to either (1) inappropriate proto-oncogene expression patterns or (2) loss of normal mechanisms that regulate proto-oncogene function. Gene rearrangements and amplifications, point mutations, as well as insertion/deletion events can foster proto-oncogene activation by increasing its promoter-mediated expression or stabilizing proto-oncogene mRNA and/or protein levels. Alternatively, these events can eliminate or inactivate proto-oncogene regulatory domains, or alter proto-oncogene protein conformation in a way that leaves it in a perpetually active state. These processes can ultimately lead to inappropriate cell proliferation, delayed differentiation, and neoplastic transformation.

Proto-oncogenes typically are proteins with important roles in signaling that are crucial for cell proliferation. They fall into four broad categories: peptide growth factors (for example EGF), receptor and nonreceptor protein tyrosine kinases (for example, EGFR, HER2, and Src family kinases), other signal transducing proteins including serine/threonine kinases, their regulatory subunits, as well as regulatory GTPases (for example, BRAF and KRAS), and nuclear transcription factors (for example, c-myc, β-catenin, and c-jun). Gastrointestinal cancers frequently express several activated oncogenes simultaneously, and they remain exquisitely dependent upon ongoing oncogene signaling. Additionally, interruption of oncogene stimulus not only inhibits growth but frequently also precipitates cancer cell apoptosis and death.

One therapeutic challenge that has arisen is that while every cancer will have several active oncogenes driving proliferation, the identity of the oncogenes involved can vary substantially, even within the same cancer type. For example, KRAS mutations resulting in an active oncogenic KRAS occur in nearly ~ 90% of pancreatic cancers, but only ~ 40% of colorectal cancers, and rarely in upper gastrointestinal cancers. Critical growth-signaling pathways often will contain several proto-oncogenes in epistatic succession. Activating mutations in any of these proto-oncogenes can produce a proliferative effect that would be phenotypically similar if not identical to any of the others. For instance, EGF, EGFR, KRAS, c-jun, and c-myc are linked in the same proproliferative signaling pathway. In the majority of esophageal squamous cell cancers without an activating KRAS mutation, there is in fact increased EGFR expression and activation due to EGFR gene amplification. In summary, one critical feature of cancer cells, sustaining proliferative signaling, is typically provided by gain-of-function mutations of proto-oncogenes, which dissociates their growth-promoting functions from the normal regulatory mechanisms that keep these functions in check.

66.2.2

Tumor Suppressor Inactivation

The existence of genes whose products regulate proliferation and prohibit the progression to neoplasia emerged from the statistical study of the pediatric cancer retinoblastoma. Alfred Knudsen postulated in 1971 that the genetics of the familial and sporadic versions of the disease were related and he hypothesized that retinoblastoma was caused by two mutations in a single gene. In the familial form of the disease, one mutation was inherited in the germline, and the second acquired somatically at a relatively high frequency. In the sporadic form of the cancer, both copies of the gene must be mutated in the same cell by somatic processes. The frequency of this occurring is quite low, hence the much lower incidence of the sporadic retinoblastoma cancer. Based upon these observations, David Comings first postulated that the gene mutated in familial cancer syndromes might encode a tumor-suppressing product. These hypotheses were not fully confirmed until 1986 with the cloning of the retinoblastoma gene ( RB1 ), the first of many additional tumor suppressor genes to be identified. In fact, the tumor suppressor genes themselves are a varied group, including cell-surface receptors, cell-cell adhesion proteins, lipid phosphatases, factors that modulate intracellular signaling pathways, and transcription factors.

Unlike oncogenes, tumor suppressor mutations are typically classic loss-of-function events. Both alleles of a given tumor suppressor must be inactivated or silenced in order for neoplastic progression to proceed. In familial cancer syndromes, there is inheritance of a single mutant tumor suppressor allele, with inactivation of the second normal allele considered to be the initiating event for ultimate neoplastic transformation. Tumor suppressors can be inactivated by several different mechanisms including mutation of the coding sequence, loss of heterozygosity (LOH), or epigenetic gene silencing.

LOH occurs when chromosomal mis-segregation or nonreciprocal translocation during cell division results in the loss of one of two gene alleles. In the case of familial cancer syndromes, one of the two alleles was already nonfunctional due to a germline mutation. Loss then of the remaining normal allele by LOH fully inactivates the tumor suppressor, opening the door for neoplastic transformation. A number of different tools have been employed to identify regions of chromosomes frequently deleted in familial and sporadic gastrointestinal malignancies, which has led to a number of tumor suppressor genes being identified within these regions. An analysis of deletions in chromosome 18q, which is common to colorectal and pancreatic cancers, led to the recognition of SMAD4 (DPC4) as an important tumor suppressor in these tissues. The tumor-suppressor protein p53 is encoded for by the TP53 gene, which is located in a region of chromosome 17p with frequent LOH in many cancers. The APC tumor suppressor gene is located in another hotspot for LOH in colon cancers on chromosome 5q, and in breast, prostate, and gastric cancers the CDH1 tumor suppressor gene locus on chromosome 16q, which codes for E-cadherin, is a frequent target for LOH.

Apart from direct loss-of-function mutations and LOH, an addition mechanism of tumor suppressor gene inactivation is epigenetic-mediated promoter hypermethylation. A number of tumor suppressors have high rates of promoter hypermethylation in gastrointestinal tract tumors, including the DNA repair gene MLH1 (colon, gastric, and esophageal cancers), CDH1 (gastric, colon, and hepatocellular), CDKN2A (gastric, esophageal, pancreatic, and colon), and the mitotic checkpoint gene CHFR (gastric, esophageal, and colon). Additional candidate tumor suppressors have been identified by screens for genes epigenetically silenced in colorectal cancers. The secreted frizzled-related proteins (SFRPs) are a family of secreted Wnt receptor proteins that serve to dampen Wnt signaling and the proliferative response of intestinal epithelial cells. Expression of SFRP1, SFRP2, and SFRP5 is silenced frequently in colorectal cancers due to promoter hypermethylation. This silencing may be an early event in colorectal carcinogenesis as it can be detected in ACF. Additional candidate tumor suppressor genes silenced by epigenetic promoter methylation have also been identified in pancreatic cancer, including ARID1B and SOX15, as well as in gastric cancer including EPB41L3 which encodes for DAL-1.

Two of the most well-known and important tumor suppressor genes include RB1 and TP53 . RB1 , which encodes for the protein pRb, was the first tumor suppressor identified and is a critical regulator of progression through the cell cycle, particularly the transition from the resting G ₁ phase to S phase when DNA replication occurs. In normal cells, pRb acts to inhibit a cell’s exit from G ₁ by binding several members of the E2F transcription factor family and silencing the expression of E2F target genes. E2F transcriptional targets include a number of genes whose products are necessary for nucleotide synthesis and DNA replication, both of which are required S phase activities. Phosphorylation of pRb by cyclin-dependent kinases (CDKs) inactivates it by disrupting its interaction with E2F. CDK function itself is regulated tightly by posttranslational mechanisms as well as the presence of two families of CDK inhibitors (CDKIs), the Cip/Kip and INK4a families of proteins. Loss of pRb function permits uninhibited progression through the G ₁ phase of the cell cycle and generally unrestricted cell division. Due to its central role regulating progression through the cell cycle, all mitogen and growth-inhibitory signaling pathways ultimately converge on pRb to regulate its function, thus making it an attractive target for modulation by neoplastic processes.

While RB1 itself is not frequently mutated in gastrointestinal cancers, cancer cells have many alternative methods they employ to inhibit pRb function. A number of viral oncoproteins directly bind to and inactivate pRb, including adenovirus E1A, SV40 T-antigen, and the E7 proteins of the human papilloma viruses. Moreover, the actions of a number of human oncogenes either induce CDK activity or limit the expression of the CDKIs, thereby leading to pRb inactivation. Alternatively, loss of the Cip/Kip and INK4a CDKI tumor suppressors p21, p27, and p16 by mutations or gene silencing, as frequently occurs in gastrointestinal malignancies, also contributes to pRb inactivation by permitting unopposed CDK function.

Another tumor suppressor with a pivotal role in preventing neoplastic transformation is p53, which is encoded for by the TP53 gene. The importance of p53 rivals that of retinoblastoma in terms of the frequency with which it is inactivated and the breadth of human cancers it is involved in. However, unlike RB1 , which is infrequently mutated in gastrointestinal cancers, loss-of-function TP53 mutations are very common in many types of gastrointestinal cancer. p53 was initially identified in a complex with SV40 T-antigen. Shortly thereafter, it was observed that when p53 was coexpressed with a mutant, oncogenic Ras in primary rat embryonic fibroblasts, the cells acquired neoplastic features. This led to the early conclusion that p53 was an oncogene that promoted cell immortalization. It took an additional five years and several dozen other investigations before p53’s true role as a tumor suppressor was fully revealed. Investigators then began finding p53 were frequently targeted by mutations or gene deletion in a variety of cancers, including gastrointestinal cancers. Moreover, germline mutations in TP53 give rise to the human condition Li-Fraumeni syndrome, which increases the risks of multiple cancers including colorectal cancer. Some p53 mutations not only inactivate its function, but because p53 forms oligomeric complexes, also inhibit the function of wild-type p53 in a dominant-negative fashion. The existence of these dominant-negative mutants undoubtedly contributed to the early confusion over p53’s role in carcinogenesis.

p53 is a sequence-specific transcription factor that is activated when the cell is stressed or damaged. Upon activation, p53 accumulates in the nucleus where it binds its cognate responsive element and induces the expression of targeted genes. The protein products of the targeted genes are the effectors of p53, inducing either cell growth arrest or apoptosis depending upon the cellular context in which they are expressed. In the absence of stress or damage, p53 levels and activity are kept in check by Mdm2, coded for by the proto-oncogene MDM2 ( Fig. 66.4 ). Mdm2 binds p53 tightly and blocks its transcriptional activation domain, inhibiting p53 function. Furthermore, Mdm2 also serves as an E3 ubiquitin ligase, attaching ubiquitin moieties to p53 that target it for proteosomal degradation. Thus, Mdm2 is a potent inhibitor of p53 tumor-suppressor function. It is not surprising, then, that p53 function is abrogated in some cancers by Mdm2 gene amplification or overexpression, including in colon cancer and pancreatic cancer. Moreover, Mdm2 is itself a p53 target gene, induced when p53 is activated to negatively feedback and suppress p53 levels, and for this reason, Mdm2 must be inhibited in order to activate fully p53. The importance of Mdm2 in p53 function, as well as new p53-independent roles of Mdm2, has led to the investigation of Mdm2 inhibitors as potential therapeutic options in gastrointestinal cancer.

Regulation of p53 function by MDM2. MDM2 inhibits the tumor suppressor roles of p53 through multiple mechanisms including direct binding of p53 as well as ubiquitination. Active p53 is also involved in negative feedback as MDM2 is a transcriptional target of p53. Ub, ubiquitin.

There are three independent mechanisms by which p53 can be activated. The first is direct DNA damage, especially DNA double-strand breaks. These frequently occur with telomere erosion or ionizing radiation, and they are potent stimulators of the protein kinases ATM (ataxia telangiectasia mutated) and Chk2. These kinases phosphorylate p53, disrupting p53 binding to Mdm2, thereby leading to stabilization of the p53 protein and enhancing transcriptional activity. Additionally, they phosphorylate the Mdm2 protein, inhibiting Mdm2 function and further increasing p53 activity. In mice, Atm deletion in a pancreatic cancer model accelerated pancreatic cancer formation, and in humans, studies have demonstrated that pathogenic germline ATM mutations are associated with an increased risk of pancreatic cancer, and more recently associated with increased gastric cancer risk as well. A second mechanism for inducing p53 function is triggered by aberrant growth signals and oncogene activity, which stimulate the expression of the tumor-suppressor p14ARF, which can sequester Mdm2 away from the p53 protein. Finally, a wide array of cell stressors can activate p53 as well through mechanisms that do not include ATM, Chk2, or p14ARF. Regardless of the mechanism of activation, these pathways lead to a common endpoint with stabilization of p53 and disruption of the Mdm2-p53 interaction.

Once activated, p53 binds DNA and activates the transcription of different genes that typically contribute to tumor suppression by inhibiting cell-cycle progression, promoting apoptosis, enhancing DNA repair, and blocking tumor-induced angiogenesis. One of the earliest effects of p53, and the best understood, is the ability of p53 to block progression of cells through the cell cycle. Most important to this effect is p53-mediated induction of the CDKI p21 ^WAF1/CIP1 . p21 ^WAF1/CIP1 is a potent inhibitor of CDK activity, which is required for orderly progression through the cell cycle. Inhibition of CDK activity by p21 ^WAF1/CIP1 blocks progression of cells at two critical points, the G ₁ -to-S and G ₂ -to-M transitions. p53 can also lead to the induction of programmed cell death through induction of proteins such as BAX, NOXA, and P53AIP1. Additionally, p53 induces the expression of death-signal receptors like Fas, DR5, and PIDD, which can respond to extracellular signals and initiate apoptotic cell death. Lastly, p53 can directly stimulate production of highly toxic reactive oxygen species (ROS) by the mitochondria which, if severe enough, can hasten cell death. In summary, inactivation of tumor suppressors, such as p53, is critical for promotion of gastrointestinal carcinogenesis.

66.2.3

Enabling Unlimited Replication

Neoplastic cells are “immortalized,” and unlike normal cells, they can undergo an infinite number of cell divisions. By contrast, normal human cells can only undergo a set number of cell divisions before permanently withdrawing from the cell cycle. This limit on the number of cell divisions is known as the “Hayflick Limit” and it represents a powerful tumor suppressing function inherent in normal human cells. The basis for the Hayflick Limit is the gradual erosion, with each cell division cycle, of the structures at chromosomal ends known as telomeres.

Telomeres are nucleoprotein complexes at chromosome ends that function to maintain chromosomal integrity. They are comprised of a number of double-stranded TTAGGG DNA repeats, a 3′ single-strand DNA overhang, and associated binding proteins. Telomeres evolved to solve two problems inherent in the maintenance and replication of the linear chromosomal DNA. First, DNA repair mechanisms recognize double-stranded breaks (DSBs) and act to repair them when detected, and chromosomal ends would naturally trigger this response unless otherwise protected. Second, DNA polymerase cannot effectively replicate the template strand at the 5′ end of the chromosome without a nucleotide primer. Telomeres act to disguise the chromosomal ends and prevent their recognition as DSBs. Telomeres also provide priming to permit faithful chromosomal replication by the DNA polymerase. The repetitive telomere sequence is synthesized at chromosomal ends by the actions of the enzyme telomerase reverse transcriptase (TERT) and the RNA component encoded by the telomerase RNA component (TERC). In human cells, telomerase activity correlates well with TERT expression levels, and in adults TERT expression is normally limited to activated lymphocytes and a subset of hematopoietic and epithelial stem cells. Limiting TERT expression yields an important benefit for human cells—telomere length functions as a potent tumor suppressor. In the absence of telomerase activity, telomere length is not maintained with cell division but is instead slowly eroded, which can promote genomic instability (as discussed in more detail in Section 66.3 ), and which can ultimately lead to cell cycle arrest or cell death.

Human cancer cells develop mechanisms to maintain their telomere integrity, although their telomere length is generally not as great as that seen in normal tissues. The majority of cancers have reactivated TERT expression to levels sufficient for telomere maintenance, and typically this occurs as a late step in the transformation process. The remaining cancers that do not reactivate TERT have acquired an alternative means of telomere biosynthesis, termed ALT (alternative lengthening of telomeres). This mechanism employs a recombination-based interchromosomal exchange to reconstitute the telomere with successive cell division cycles. In summary, telomere length serves as an additional critical barrier to neoplastic transformation of normal cells. To become immortalized and proliferate indefinitely, cancer cells must acquire the means to protect their chromosome ends, and this is most commonly done by reactivation of TERT expression.

66.3.1

Detecting Genomic Instability in Gastrointestinal Cancer

Recent advances have made the process of sequencing cancer genomes far less expensive and much more efficient, which has led to genome sequencing being increasingly utilized to study nearly all types of malignancies, and gastrointestinal cancers are no exception. Large sequencing efforts, including those from The Cancer Genome Atlas (TCGA), in colon cancer, gastric cancer, and pancreatic cancer, have further characterized the prevalence of genomic instability in these cancers as well as the importance of this instability in gastrointestinal carcinogenesis. For example, in colorectal cancer, it has been reported that 16% of these cancers are hypermutated. This group of hypermutated tumors was found to have several mechanisms accounting for this increased mutational rate including MSI as well as polymerase proofreading dysfunction. In contrast, 20% of gastric cancers have stable genomes. Of the genomically unstable gastric cancers, 50% demonstrated CIN, 22% demonstrated MSI, and the remaining cancers were found in an EBV-positive subgroup that had significant DNA hypermethylation. In pancreatic cancer, whole-genome sequencing studies have highlighted the significant presence of chromosomal rearrangements and structural variation that is a hallmark of CIN in these cancers. However, pancreatic cancers tend to have less MSI than what is seen in colon and gastric cancer. The reasons for this are unknown and are an active area of research. Given the prevalence of significant genomic instability within multiple different gastrointestinal cancers, it is important to understand the mechanisms that commonly lead to the onset of this instability.

66.3.2

Chromosomal Instability (CIN)

CIN refers to the process by which either whole chromosomes or parts of chromosomes are duplicated or deleted, thus leading to both numerical and structural chromosomal changes. Numerical aneuploidy, which is a form of CIN that is marked by gains or losses of whole chromosomes, is present in the majority of human gastrointestinal tumors. Numerical aneuploidy frequently results from chromosomal mis-segregation during mitosis. It is important to consider though that all aneuploidy is not indicative of CIN, and instead CIN aneuploidy refers to aneuploidy resulting from increased rates of chromosomal mis-segregation. However, CIN gastrointestinal tumors also possess structural aneuploidy as well, which results from more focal chromosomal changes including chromosomal amplifications, deletions, inversions, and translocations. Structural aneuploidy can often result in the loss of heterozygosity (LOH), which represents the loss of one allele of a gene at a specific locus, and which frequently involves tumor suppressor genes. Common sites for LOH in gastrointestinal malignancies include 5q ( APC ), 17p ( TP53 ), and 18p ( DPC4 / SMAD4 ). CIN tumors can also have chromosomal abnormalities that lead to gene amplifications, commonly of growth-promoting cellular proto-oncogenes such as the EGF receptor or cyclin D1. This represents another mechanism whereby CIN can promote cell transformation and tumor growth. There are multiple mechanisms that can result in CIN in gastrointestinal cancers including defects associated with chromosomal segregation, the DNA damage response pathway, and telomere stability, and these mechanisms will each be discussed in more detail.

The process of chromosomal segregation during mitosis is extremely complex, and defects in this pathway, including impairment of the mitotic spindle checkpoints, ineffective chromatin cohesion, abnormal microtubule attachments, increased centrosome number, and increased microtubule dynamics, can lead to the development of CIN. The mitotic spindle, which equally segregates chromosomes between daughter cells during cell division, is monitored by spindle checkpoints. Impairment of the mitotic spindle assembly checkpoint (SAC) can lead to mis-segregation and aneuploidy through premature separation of sister chromatids. Evidence of this was seen through deletion of one allele of the essential SAC gene MAD2 , which changed a chromosomally stable colon cancer cell line to a CIN cell line. Despite this, there is little evidence of this pathway playing a major role in the majority of CIN gastrointestinal tumors as mutations in the mitotic SAC genes are quite rare. However, both downregulation and upregulation of SAC genes can lead to SAC impairment and subsequent CIN. For example, it was demonstrated that HMGA1, which can be overexpressed in colon cancer, enhances the transcription of multiple components of the SAC and leads to development of CIN.

CIN can also be mediated by defects in sister chromatid cohesion. STAG2 , which is a gene encoding a critical cohesion subunit that is necessary for proper sister chromatid cohesion, has been described as being mutated in a subset of colon cancer. Additionally, abnormal stability of the microtubule-kinetochore attachment can also lead to the development of CIN. APC , which is frequently mutated in colon cancer, has been shown to be involved in the interaction between microtubules and kinetochores. Furthermore, loss of APC can lead to stabilization of the kinetochore attachments, thus facilitating chromosomal mis-segregation and development of CIN. Other mechanisms involving chromosomal segregation that can lead to the development of CIN include an increase in the number of centrosomes over the standard number of two, as well as an increase in the rates of microtubule plus end assembly; however, the role of these processes in gastrointestinal cancer pathogenesis is currently unclear.

Defects in the DNA damage response can also lead to CIN as genes involved in this pathway are known to play important roles in tumorigenesis including BRCA1 , BRCA2 , ATM , and TP53. Each of these gene products responds to DNA damage, in particular, double-stranded DNA breaks, by inhibiting further progression through the cell cycle and induction of DNA repair, or if necessary, the induction of apoptosis if the damage is severe. Inactivation of these checkpoint controls yields cells more tolerant of severe DNA damage and permissive of neoplastic transformation and CIN events, as evidenced by the increased risk for cancer in families carrying germline mutations in these genes. For example, germline BRCA1 and BRCA2 mutations have been associated with an increased risk of pancreatic cancer, whereas germline TP53 mutations in Li-Fraumeni syndrome have been associated with an increased risk of colorectal cancer.

Finally, telomere length is a critical barrier to neoplastic transformation and cell immortalization. Telomeres serve to indicate the natural ends of chromosomes and distinguish them from a DNA double-strand break (DSB). Telomeres can be maintained by the enzyme telomerase or by the telomerase independent mechanism termed ALT. Because most cells do not utilize a telomere maintenance mechanism, their telomeres are gradually eroded with each successive generation, ultimately leading to loss of telomere protection of chromosome ends. In normal cells with eroded telomeres, the DSB repair mechanism is activated, and critical cell-cycle regulators like p53 and/or p16 INK4 induce cell senescence or apoptosis. In the case where these tumor suppressors are inactivated, continued telomere erosion can lead to telomere crisis due to extremely short telomeres. Telomere crisis can lead to end-to-end chromosomal fusions leading to dicentric chromosomes, which then initiates an ongoing CIN via breakage-fusion-bridge cycles. Dicentric chromosomes pulled in opposite directions during mitosis can break at random points, resulting in nonreciprocal translocations. Moreover, the broken ends reactivate DSB repair and are fused again, possibly with another chromosome. The process is repeated with each cell division, resulting in massive genomic instability. Telomere shortening to promote CIN is important for the initial steps in carcinogenesis, however cancer cells ultimately need to utilize a method of telomere maintenance, such as telomerase or ALT, to ensure telomere maintenance during subsequent proliferation.

This model is supported by many observations both in vitro and in vivo . Comparative genome hybridization (CGH) studies and histological studies reported increased rates of chromosomal rearrangements, nonreciprocal translocations, and anaphase bridging observed in adenomatous polyps. Clinical-pathologic studies reveal that telomerase activity is low, and telomere length is shortened in small- to moderate-sized human colon polyps, whereas telomere length is stabilized and telomerase activity is increased in large polyps and colon cancers, and telomere shortening in colonic adenomas is accompanied by CIN. In all PanIN lesions, telomere length is considerably shortened, even in the earliest PanIN-1 lesions. In addition, the degree of cytogenetic abnormalities correlated inversely with telomere length, with the greatest number and size of abnormalities arising in tumors with the shortest telomeres. Telomeres were also found to be shortened in early gastric cancer as well and shortened telomeres may serve as a risk factor for development of gastric cancer. However, as the stage of gastric cancer increased, it was found that so did telomere length.

Genetic studies in mice and yeast confirm these observations. In mice, disruption of the Terc gene yielded cells with shortened telomeres and a high frequency of end-to-end chromosome fusions. The chromosomes that participated in the fusions were those with the shortest telomeres in the population. These mice also experienced premature aging in many organs and an increase in rates of neoplasia, primarily lymphomas, teratocarcinomas, and hepatomas. Breeding these mice with disruptions in their p53 gene but not, surprisingly, in p16Ink4a, leads to a dramatic increase in epithelial neoplasms including breast, squamous cell, and gastrointestinal cancers. Analysis of these tumors found CIN with features nearly identical to human cancers, including chromosomal bridging, nonreciprocal translocations, chromosomal losses, regional amplifications, and deletions. Based on these many different observations as well as the requirement of telomerase or ALT activation for tumor progression, telomerase dysfunction appears to be an important contributor to the CIN cancer phenotype and a source for much of the genetic instability associated with human cancer.

66.3.3

Microsatellite Instability (MSI)

MSI is perhaps the best understood of the genetic instabilities involved in GI cancer progression. The mismatch- repair (MMR) system is a multiprotein complex of MutS (MSH2, MSH3, MSH6) and MutL (MLH1, PMS1, PMS2) homologues that recognizes and signals for repair of the base-base mismatches and short insertion/deletion mispairings, which spontaneously occur with DNA replication ( Fig. 66.5 ). Short-repetitive DNA segments, known as microsatellites, are especially vulnerable to this type of error. In a certain proportion of colonic (~ 15%), gastric (~ 20%), and pancreatic (~ 5%) cancers, the MMR system is inactivated leading to an increased frequency of mutations in microsatellites. While most microsatellite sequences are contained in noncoding regions of the DNA, some genes contain repetitive sequences in their coding regions and are therefore highly susceptible to inactivating mutations in MMR-deficient cells. Workshops and panels of experts have subsequently issued guidelines for the identification of MSI cancers ; a consensus panel of five mononucleotide and dinucleotide markers are used to define MSI neoplasms. Instability in two or more markers is designated as MSI-High (MSI-H), with a high-degree of MSI. Instability in one marker is termed MSI-Low (MSI-L), and absence of instability is termed microsatellite-stable (MSS). MSI can arise in either familial or sporadic GI cancers, although the mechanisms responsible for each are quite different.

The DNA MMR system functions through a series of steps. (A) MSH2–MSH6 (MutSα) recognizes single base-pair mismatches, in which the DNA polymerase has matched the wrong base (G) with the T on the template (shown on the left side), and creates a sliding clamp around the DNA. This step that requires the exchange of adenosine triphosphate (ATP) for adenosine diphosphate (ADP) (by MSH2, but not MSH6 or MSH3). The complex diffuses away from the mismatch site, which is then bound by the MLH1-PMS2 (MutLα) complex ( right ). This “matchmaker” complex moves along the new DNA chain until it encounters the DNA polymerase complex. (B) The DNA MMR protein sliding clamp interacts with exonuclease-1, proliferating cell nuclear antigen (PCNA), and DNA polymerase. This complex excises the daughter strand back to the site of the mismatch (shown on the left side). Eventually, the complex falls off the DNA and resynthesis occurs, correcting the error. (C) Variations on the DNA MMR theme. Whereas MSH2–MSH6 recognizes single pair mismatches and small IDLs, MSH2–MSH3 (MutS β ) complements this by also recognizing larger IDLs (shown on the left side). The right side shows the possible interactions with different MutL dimers, as MLH1 can dimerize with PMS2, PMS1, or MLH3. The preferred interaction with MSH2–MSH3 is MLH1–MLH3 (MutL γ ), but the precise roles of the other MutL heterodimers in this reaction are not entirely understood.

(From Boland CR, Goel A. Microsatellite instability in colorectal cancer. Gastroenterology 2010; 138 (6):2073–87 e3.)

The prototypical familial MSI cancer-causing syndrome is Lynch syndrome. MSI was first recognized to be an important feature of Lynch syndrome-associated cancers in 1993. Lynch syndrome families, which together constitute about 3%–5% of all colorectal cancers, frequently have germline mutations in MLH1 and MSH2. These mutations do not cluster within specific domains and are distributed throughout the MLH1 and MSH2 genes. Other germline mutations that have been associated with Lynch syndrome include mutations in PMS2 and MSH6 , which are associated with a lower risk of colorectal cancer as well as a later age of onset, as well deletion of EPCAM , which leads to decreased MSH2 expression through promoter hypermethylation. Germline mutations have also been documented in PMS1 and MLH3 ; however, their association with Lynch syndrome at the current time remains uncertain. As with all familial cancer syndromes, individuals with predisposing germline mutations in one of their MMR genes frequently acquire a somatic “second hit” mutation that inactivates the remaining normal allele and initiates neoplastic transformation. Generally, this occurs by LOH or somatic mutation of the normal allele. Epigenetic inactivation of the normal allele has also been described. Once the MMR gene is effectively inactivated in both alleles, MMR function is compromised and sequence errors begin to accumulate in the genomes of daughter cells, thus facilitating tumorigenesis.

Sporadic colon cancers with MSI constitute about 10%–15% of all colon cancers. Similar to Lynch syndrome-associated colon neoplasms, sporadic MSI colon cancers tend to occur proximally, have a greater mucinous component, be poorly differentiated and diploid, and are associated with an improved survival for patients. Somatic mutations in MLH1, MSH2 , MSH6 , and PMS2 have been reported as a cause of sporadic MSI cancers. Further studies with sporadic MSI colon cancers revealed that often these cancers had wild-type coding sequences for MLH1 ; however, MLH1 protein levels were markedly diminished. The loss of the MLH1 protein was found to be due to inactivation of the MLH1 promoter by an epigenetic mechanism, which will be discussed in more detail in Section 66.3.6 . Thus, while the basic mechanisms driving MSI in Lynch syndrome and sporadic colon cancer are similar, the underlying genetic events can be quite different.

Nearly all mutations selected for in MSI cancers tend to lead to inactivation of target genes. In the colon, the majority of Lynch syndrome cancers acquire frameshift mutations in APC or β-catenin. These mutations either inactivate APC or activate β-catenin by mutation of a regulatory domain, and in both cases, this leads to increased β-catenin signaling. The TGF-βRII and insulin-like growth factor-II receptors (IGFIIR) are also frequently inactivated by mutations in MSI neoplasms. Although the p53 tumor suppressor is not frequently targeted in Lynch syndrome colon cancers, p53-mediated apoptosis is significantly inhibited by BAX inactivation in 55% of MSI cancers. Other genes whose products are important for control of cell proliferation, DNA replication and repair, or apoptosis are frequently mutated in MSI cancers, including MED1, ATM, the DNA helicase BLM, the transcription factor E2F4, the candidate tumor-suppressor RIZ (retinoblastoma protein interacting zinc finger) gene product, as well as caspase-5. In summary, inhibition of DNA mismatch repair promotes the accumulation of frameshift mutations that inactivate tumor suppressors, promote growth, and accelerate malignant transformation. This likely contributes to further diminished DNA repair and therefore increased mutagenesis during the adenoma-carcinoma sequence.

66.3.4

Base Excision Repair (BER)

DNA nucleotide bases are fairly stable chemically although in the context of the cellular environment, especially in stressed cells, these bases can be damaged by oxidative damage, deaminations, depurinations, and alkylations. All of these modifications can alter the noncovalent interactions between nucleotides, and if not corrected can lead to base mispairing and transition or transversion mutations during DNA replication. If these single base mutations occur in critical coding sequences, they can result in nonsense mutations, which often lead to premature protein truncation and subsequent inactivation, or missense mutations that can be inactivating (i.e., APC) or activating (i.e., KRAS). The BER pathway is responsible for identifying these chemically modified bases and initiating their repair, and defects in this pathway can contribute to GI cancer pathogenesis. While it has been observed that the CIN- and MSI-mediated genetic instabilities are mutually exclusive, BER defects can be observed in CIN + and MSI + colon cancers, as well as a small subgroup of cancers that are chromosomal stable and MSS.

One component of the BER pathway that has been clearly linked to gastrointestinal carcinogenesis is the defects in the DNA glycosylase MUTYH. MUTYH is encoded for by the MUTYH gene, and biallelic pathogenic mutations of MUTYH have been associated with the autosomal recessive colonic polyposis syndrome, MUTYH-associated polyposis (MAP), which leads to a significantly increased risk of colon cancer. This syndrome is characterized by numerous colonic adenomas, and MAP-associated colon cancers are diploid or near diploid and are typically stable genetically at the chromosomal and microsatellite level (CIN-/MSS). In addition to colon cancer, biallelic MUTYH mutations also increase the risk for duodenal adenomas. There has been additional evidence that MUTYH mutations, including monoallelic mutations, may be associated with an increased risk of gastric cancer as well. Whereas biallelic MUTYH mutation leads to increased risk of colon cancer, somatic mutations in this gene have not been commonly found in sporadic colon cancer.

MUTYH functions in the repair of the stable guanine adduct 8-oxo-7,8-dihydroxy-2′-deoxyguanuisine (OG), which can form after oxidative damage of DNA ( Fig. 66.6 ). In the absence of repair, OG can mimic a thymidine base, which can result in high levels of G:C to T:A transversion mutations. MUTYH works in conjunction with OGG1 to identify and excise these oxidized guanines to permit their repair. In the absence of properly functioning MUTYH, it has been reported that APC is inactivated by premature nonsense codons in up to half of the MAP cancers, all due to G->T transversions, while LOH at the APC allele is quite infrequent. Moreover, the majority of MAP colon cancers possessed activating missense mutations in codon 12 of KRAS caused by G->T mutation at a single guanine base, leading to inappropriate KRAS protein signaling. Conversely, the critical tumor suppressors TP53, SMAD4 , and TGF-βII receptor genes are not frequently mutated in MAP colon cancers.

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related posts:

Gastrointestinal Hormones Cell Death Enteric Neural Regulation of Mucosal Secretion Enteric Nervous System Structure and Neurochemistry Related to Function and Neuropathology Bile Formation and the Enterohepatic Circulation The Vitamin D System: Biological and Molecular Actions in the Intestine and Colon

Stay updated, free articles. Join our Telegram channel

Join