Overview

In the United States, one woman in eight will develop breast cancer in her lifetime, but some families will be affected at even higher rates. The breast-ovarian cancer syndrome is characterized by the clustering of breast cancer alone or breast and ovarian cancer in a single family. Historically, the hereditary breast-ovarian cancer (HBOC) syndrome was recognized in families with an autosomal dominant mode of genetic transmission in the early 1970s.^{12, 13}

Familial breast cancer accounts for 5% to 10% of all breast cancer, and a substantial number of these cases can be linked to mutations in the genes BRCA1 or BRCA2.^{14, 15} Familial cases of ovarian cancer comprise up to 8% of all ovarian cancer, and recent reports suggest that most, if not all, of these may be accounted for by mutations in BRCA1 or BRCA2.¹⁶ The transmission of the heritable susceptibility follows an autosomal dominant pattern. It is estimated that BRCA1 or BRCA2 carriers have a 50% to 85% chance of developing breast cancer by age 70 years and a 16% to 60% risk of developing ovarian cancer.^{17, 18} Initial reports suggesting that BRCA1 mutations are associated with a higher than expected incidence of prostate cancer have not been confirmed, but there is evidence from several studies for an increased risk of colon cancer (relative risk [RR] 2), pancreatic cancer (RR 2 to 3), gastric cancer (RR 3 to 7), and fallopian tube cancer (RR 50 to 100).^{19, 20} BRCA2 mutations are associated with an increased incidence of male breast cancer (RR 50 to 100),²¹ as well as with an increased risk of prostate cancer (RR 5), pancreatic cancer (RR 4), gall bladder and bile duct cancer (RR 5), stomach cancer (RR 3), and malignant melanoma (RR 3).²²

BRCA1 was first linked to chromosome 17q21^{23, 24} and subsequently isolated in 1994.²⁵ The existence of breast cancer families without associated mutations in BRCA1 led to the discovery of a second breast cancer susceptibility gene, BRCA2, on chromosome 13q12.^{26, 27} The scientific and clinical impact of the discovery of these two genes is discussed below.

BRCA1 and BRCA2: Gene Structure and Function

BRCA1 and BRCA2 encode large proteins of 1,863 and 3,350 amino acids, respectively. They are both complex genes made up of more than 20 exons. Neither gene bears significant homology with other known genes, with the exception of the BRCT domain in the C-terminus of BRCA1, a domain found in more than 40 other genes associated with response to deoxyribonucleic acid (DNA) damage.²⁸ It is clear that both BRCA1 and BRCA2 are important components of the pathway that protects cells from the effects of DNA damage (reviewed by Venkitaraman²⁹). The majority of mutations identified thus far lead to protein truncation, and it is believed that cancer then develops when the second copy is lost. Therefore, it is thought that BRCA1 and BRCA2 behave like classic tumor-suppressor genes, with the loss of one copy predisposing the carrier to the development of the characteristic cancers of this classic cancer syndrome.

Several lines of investigation have implicated BRCA1 and BRCA2 in DNA damage-response pathways. An association between BRCA1 and p53 and the subsequent enhancement of p53 activity incriminates BRCA1 in p21-mediated cell-cycle arrest following DNA damage.³⁰ In addition, BRCA1 forms complexes with both BRCA2 ³¹ and Rad51,³² the human homolog of the Escherichia coli gene RecA, which is essential to normal recombination and genome stability. Colocalization of BRCA1, BRCA2, and Rad51 in “nuclear dot” structures is seen to disappear following treatment of cells with DNA-damaging agents that cause double-stranded chromosome breaks.³³ In addition, mouse embryonic stem cells that are homozygous for BRCA1 null and BRCA2 truncations are hypersensitive to ionizing radiation and other forms of oxidative damage.³⁴

BRCA1 may also play a role in transcription regulation, cell-cycle control, and development. Both BRCA1 and BRCA2 are nuclear proteins.^35–38 BRCA1 associates with the ribonucleic acid (RNA) polymerase holoenzyme³⁹ and binds CREB,⁴⁰ implicating it in transcriptional regulation. BRCA1 protein levels are also altered by variations in hormone levels.⁴¹ During the cell cycle, BRCA1 and BRCA2 messenger ribonucleic acid (mRNA) levels increase from low levels at the start of G1 to maximum levels at the G1/S transition in parallel to cyclin A levels. Phosphorylation of BRCA1 also occurs in a cell-cycle-dependent manner, suggesting that it may play a role in the G1/S checkpoint control.⁴² Mice with homozygous deletion of BRCA1 or BRCA2 die as embryos, revealing a key role for BRCA1 and BRCA2 in development.^43–45 However, mice homozygous for a partial deletion of BRCA2 are viable but show growth retardation and the development of thymic lymphomas.^{46, 47} Recently, homozygous deletion of BRCA1 in mouse mammary tissue was found to lead to the development of tumors.⁴⁸ This advance in the mouse model will likely yield important insights into the function of BRCA1 in the near future.

Clinical Aspects of BRCA1 and BRCA2

It is estimated that a family history of breast cancer in the general population confers a 1.7 relative risk of developing the disease.^49–51 The majority of these families will not have a mutation in BRCA1 or BRCA2. However, in a woman who develops breast cancer prior to age 40 years, the chance that she carries a mutation in BRCA1 might be as high as 10%, or even 20%, if she is Ashkenazi Jewish.^{15, 52–54} This may not be true for BRCA2, which may be associated with a slightly older age of diagnosis.⁵⁵

Of the women seen in American cancer risk evaluation clinics, BRCA1 accounts for up to 30% of breast-ovarian cancer families, whereas BRCA2 accounts for 10% to 20%. Together, BRCA1 and BRCA2 account for most, if not all, familial ovarian cancer.¹⁶ The prevalence of certain BRCA1 and BRCA2 mutations is occasionally higher than expected because of the increased transmission of a few mutations (“founder effect”) in several subpopulations. Characteristic founder mutations for BRCA1 and BRCA2 have been identified in families of Ashkenazi Jewish descent and in Iceland, Finland, Hungary, Russia, France, Holland, Belgium, Sweden, Denmark, and Norway. Most of what we know about BRCA1 and BRCA2 has come from studies of whites in the United States and Europe. Significantly less is known about the rates of BRCA1 and BRCA2 mutations in families from other ethnic or racial backgrounds.

BRCA1-associated tumors are more likely to be aneuploid, have a high S-phase component,⁵⁶ be high grade,^{57, 58} and be estrogen and progesterone receptor negative.⁵⁹ Whether these characteristics result in poorer survival is difficult to study because of multiple confounding variables, although some studies suggest that this may be the case.^{60, 61} Evidence from the Breast Cancer Linkage Consortium reveals that unlike BRCA1, BRCA2 tumors are largely indistinguishable from sporadic tumors in their mitotic rate or degree of pleomorphism.⁵⁶

Risk Evaluation and Screening

The first and most important step in any cancer-risk evaluation program is the recording of an accurate history and cancer pedigree. Where possible, it is recommended that all cancers be confirmed by documentation from a hospital record. This is especially true in the case of ovarian cancer, which may be mistaken for stomach or liver cancer as the information is passed through the family. The pedigree helps in identifying the type of cancers and the transmission pattern. Once an accurate pedigree is attained, recommendations can be made regarding the advisability of increased cancer surveillance and DNA testing.

There are several useful models that assist the provider in the evaluation of a patient's risk for the development of breast cancer. These models are based on documented breast cancer risk factors. Other models are available that estimate the probability that the pattern of cancers in a family is the result of a mutation in BRCA1 or BRCA2. These models take into account, to varying degrees, the number, age at diagnosis, and relationship of others in the family who have had breast or ovarian cancer. Although every model has its strengths and weaknesses, the appropriate use of different models can help give a range of probabilities that the patient may, indeed, carry a mutation in one of the two known susceptibility genes.

Because of the expense and difficulty in testing for mutations in BRCA1 and BRCA2, an effort is made to target the testing to the mutations that are thought to be likely in the tested individual where possible. For example, families of Ashkenazi Jewish descent can be first screened relatively quickly and inexpensively for the three founder mutations: 185delAG, 5382insC in BRCA1, and 6174delT in BRCA2. These three mutations are thought to account for more than 90% of mutations in this population.^{18, 62} Where possible, testing is performed on an individual who has been affected with breast or ovarian cancer, thereby increasing the chance that if a genetic alteration exists, it will be identified. Once a mutation is identified in a family, testing of that one mutation can be carried out in the other high-risk individuals. The exception to this approach is in the Ashkenazi Jewish population, where the risk of having more than one founder mutation is sufficient to always warrant screening for all three.

For those who desire genetic testing, adequate genetic counseling is required prior to and following the acquisition of any testing results. Once testing has been obtained, those with positive test results have many treatment options available to them including various screening methods, chemoprevention, and prophylactic surgery. Ideally, screening identifies most breast cancer as stage I or II disease; some patients may desire to undergo prophylactic mastectomy. Unlike breast cancer, ovarian cancer screening is not effective; prophylactic bilateral oophorectomy is the procedure of choice, as it reduces ovarian cancer risk in mutation carriers by 95%.⁶³ In addition, bilateral oophorectomy reduces breast cancer risk by almost 50%.⁶³ However, because the risk of ovarian cancer increases in the late forties and early fifties, women may wait until after childbearing is complete before choosing to undergo the surgery. In a family where no mutation has been identified, a negative result on testing is uninformative, whereas a negative test on an individual from a family with a known mutation defines the woman's chance of acquiring breast cancer as that of the general population.

With the identification of BRCA1 and BRCA2, the care of people who come from families with mutations in these genes has been directly impacted. With further characterization of the genes, further insights into the mechanisms that underlie the development of these cancers and their potential for cancer prevention and cure may be gained.

Overview

App-roximately 148,300 new cases of colorectal cancer (CRC) were diagnosed in the United States during 2002.¹¹ Known hereditary syndromes accounted for at least 5% of those, and it is likely that more subtle susceptibility factors contributed to the pathogenesis of many more. A large prospective study⁶⁴ recently confirmed what more than a dozen retrospective studies had asserted; that is, that a family history of CRC confers increased risk of the disease. In 1985, analysis of a large colon cancer-prone Utah kindred pointed to the existence of a partially penetrant gene, inherited in an autosomal dominant fashion.⁶⁵ The missense mutation in the APC gene described by Laken and colleagues (I1307K)⁶⁶ is an example of such a mutation. Other low-penetrance mutations may prove to be important contributors to colon cancer. This section reviews four inherited syndromes that feature increased risk for colon cancer: familial adenomatous polyposis, hereditary nonpolyposis colorectal cancer (Lynch syndrome), familial juvenile polyposis, and Peutz-Jeghers syndrome (PJS).

Familial Adenomatous Polyposis

Bulow⁶⁷ cites an 1861 clinical patient report by Luschka⁶⁸ that may have been FAP. The first definitive report of a patient with multiple colonic polyposis was published in 1881 by Sklifosovski.⁶⁹ The first familial example was that of a brother and sister with FAP reported by Cripps¹ in 1882. Additional cases of CRC in patients with FAP were reported by Smith in 1887.² The histologic changes from adenoma to adenocarcinoma were first described by Handford.⁷⁰ Lockhart-Mummery⁷¹ was the first to report that the relevant hereditary factor in this disease is not the cancer per se, but rather the presence of multiple adenomas, which have a tendency to undergo malignant transformation. This sentinel report, emanating from the histories of three FAP families, formed the basis for the now renowned St. Mark's Hospital Polyposis Registry in London, England.

FAP is inherited in an autosomal dominant pattern. The incidence is 1 in 6,000 to 1 in 13,000; Powell and colleagues estimate that more than 50,000 families in the United States could benefit from genetic counseling for this disorder.⁷² Affected individuals carry germ line mutations of the APC gene on chromosome 5q21–q22.^{6, 7, 73–75} Approximately one-third of FAP patients have no family history and probably represent new mutations. Many mutations of APC have been described; 80% of them are truncating. There is some correlation between the position of the truncating mutation and phenotype.

APC is a large gene, with 15 exons encoding a 2,843-residue protein. The gene product participates in several functions, including differentiation, proliferation, apoptosis, adhesion, migration, and chromosomal segregation.⁷⁶ First, APC may negatively regulate the Wnt-1 signaling pathway by binding to β-catenin.^{77, 78} Mutated APC product, unable to bind, allows cytoplasmic and nuclear accumulation of β-catenin, with persistent activation of downstream transcription and growth factors. APC protein also colocalizes with the microtubule cytoskeleton, leading to speculation that APC is involved in cell migration and cell adhesion.⁷⁹ Finally, APC might be involved in cell cycle regulation.⁸⁰

Affected individuals have multiple colonic adenomas and, if untreated, will inevitably develop CRC. Adenomas arise in the mid to late teens; 95% of mutation carriers have adenomas by age 35 years. More than 90% of FAP patients develop duodenal adenomas, but duodenal carcinoma supervenes in only about 5%. Gastric polyposis is seen in at least 50% of affected patients; most of the polyps are fundic gland polyps, but gastric adenomas do occur. Although fundic gland polyps are considered to be benign, there are reports of dysplasia in fundic gland polyps and reports of carcinoma arising adjacent to, or in continuity with, fundic gland polyps.⁸¹ Abraham and colleagues⁸² showed that FAP-associated fundic gland polyps often have sporadic APC alterations, which sporadic fundic gland polyps lack, suggesting that the polyps arising in the hereditary setting are neoplastic. Gastric carcinoma risk, however, is not appreciably elevated in Western FAP patients. Japanese and Korean families with FAP have a three- to fourfold excess risk for gastric carcinoma.⁸³

Extraintestinal manifestations of FAP include desmoid tumor, hepatoblastoma, thyroid carcinoma, medulloblastoma, and a litany of benign lesions: sebaceous or epidermoid cysts, lipomas, osteomas, supernumerary teeth, congenital hypertrophy of retinal pigment epithelium, and juvenile nasopharyngeal angiofibromas.

The attenuated variant of FAP (AFAP) presents a particularly difficult diagnostic challenge.^84–86 Adenomas can be sparse (one or two in some patients, dozens in others) and are often right-sided. The adenomas appear at a later age (35 to 40 years) than in classic FAP, as do colon carcinomas (55 years). Upper gastrointestinal manifestations (fundic gland polyps and duodenal adenomas) are seen with the same frequency as in classic FAP. The first AFAP families to be described had mutations near the proximal end of the APC gene; the phenotype has since been seen in families with mutations at the extreme distal end as well.

Genetic testing is available for FAP. If a mutation is found in an affected family member, other at-risk family members can easily be tested for the mutations. Genetic testing is not recommended for children younger than 10 years old. If no mutation is found in an affected family member, the test is considered uninformative; a negative result in that setting does not rule out FAP.

Two recent studies compared the phenotype of FAP families with and without documented mutations. Moisio and colleagues⁸⁷ identified 38 different germ line mutations in APC in 47 of 65 (72%) Finnish FAP kindreds. Families without detectable APC mutations differed from mutation-positive families by having an older mean age of polyposis diagnosis (38.6 years [48 individuals] vs 30.0 years [140 individuals]; p = 0.001) and a higher proportion of kindreds without extracolonic disease (6 of 18 vs 5 of 47; p = 0.04). Heinimann and colleagues⁸⁸ studied 36 FAP families, 72% of which were positive for the APC germ line mutation. The mean age at diagnosis of colonic adenomas was 35.2 years in APC-positive families vs 45.3 years in APC-negative families. Gastric polyps were found in 14 patients, all of whom were APC positive. Finally, patients who were APC negative showed a lower number of colonic adenomas at diagnosis and fewer extracolonic manifestations. These authors concluded that patients from FAP families who lacked APC germ line mutations presented with a notably milder disease phenotype when compared to the APC-positive families.

Randomized, controlled trials demonstrating the efficacy of screening and management regimens have not been performed and are not likely to be. Nevertheless, the following recommendations are generally agreed upon: affected or at-risk individuals should have annual flexible sigmoidoscopy beginning by age 12 years. If the family history suggests AFAP, then colonoscopy is required but can begin later (age 20 years). Prophylactic colectomy should be considered once adenomas have appeared. (It may be that prophylactic colectomy can be postponed in AFAP; we are following several mutation carriers who develop one or two new adenomas every year but who have not developed carcinoma.) Endoscopic surveillance of the rectum and anus should be continued after colectomy. A baseline upper endoscopy is advisable by age 20 years, with follow-up examinations every 2 to 3 years unless symptoms occur. Annual thyroid palpation is suggested, and children at risk for FAP should have serum α-fetoprotein testing every 6 months until 6 years of age in the interest of early detection of hepatoblastoma.

When duodenal adenomas are discovered, they should be removed endoscopically if feasible. Often, however, the adenomas are too numerous to remove; in that case, annual surveillance with biopsy of grossly suspicious lesions is a prudent course. It is hoped that chemoprevention will help control duodenal adenomas, but results thus far have not been encouraging (reviewed by Hawk and colleagues⁸⁹).

Desmoid tumors can be a difficult management problem. These locally aggressive soft-tissue tumors typically arise in the abdominal wall or bowel mesentery. Relentless recurrences are the rule.^{90, 91} The difficulty of surgical cure, complete with the fact that surgery may initiate the pathogenesis of desmoids, has led to the recommendation that only symptomatic desmoid tumors should be surgically resected.⁹²

I1307K Ashkenazi Mutation

First described by Laken and colleagues,⁶⁶ the I1307K mutation/polymorphism is carried by approximately 6% of Ashkenazi Jews and by a lower proportion of other Jews; it has not been seen in non-Jews.^{66, 93, 94} Carriers have an approximately twofold risk of colorectal cancer when compared to noncarriers.⁶⁶ Rozen and colleagues⁹⁵ studied the I1307K APC gene variant in 718 Israeli Jews wherein “…I1307K occurred in 6.2% of Ashkenazi participants, in 1.5% of non-Ashkenazi control participants (p = 0.02), and in 10% of Ashkenazim with familial neoplasia (relative risk 1.73 [not significant when compared with controls]; 95% confidence interval, 0.7 to 3.2). Colorectal neoplasia was detected in carriers at a younger age (p < 0.05) without excess risk for multiple colorectal neoplasia or noncolorectal neoplasia. I1307K attributable risk for colorectal neoplasia was 0.5% to 0.6%. Compared with noncarriers, both Ashkenazi and non-Ashkenazi I1307K carriers had similar flanking polymorphic markers (p < 0.01).”

The conclusion from this study was that this low penetrant genetic variant poses an approximate 1.7 relative risk for neoplasia in mutation carriers who have familial carcinoma and that it appears to be clinically equivalent to “… obtaining a family history of sporadic colorectal neoplasia and promoting early screening. I1307K is a founder genetic variant in Jews of different ethnic origin, namely Ashkenazim, but it explains only partially their higher instance of colorectal carcinoma.”

Familial Juvenile Polyposis Coli

Sporadic juvenile polyps are relatively common (one autopsy study of patients under 21 years reported a prevalence of 1%).⁹⁶ Usually solitary, the polyps are innocuous and do not confer increased risk for cancer of the colon.⁹⁷ The discovery of multiple juvenile polyps in a patient without a family history of polyposis can create a diagnostic dilemma; various numbers in the range of 5 to 10 have been proposed as an upper limit for sporadic juvenile polyps. The clinical diagnosis of familial juvenile polyposis (FJP) requires histologic confirmation of a juvenile polyp, plus the presence of polyposis (ie, more than 5 or 10 polyps) or a family history of polyposis. Even when there are multiple juvenile polyps and a family history of polyposis, the diagnosis of juvenile polyposis must be made with care, because Cowden syndrome and its allelic cogeners have similar polyps but a very different spectrum of organs at risk for carcinoma (breast and thyroid).⁹⁸

Two variants of FJP have been described: a rare, usually fatal juvenile polyposis of infancy with diarrhea, protein-losing enteropathy, and alopecia and the more common form, with juvenile polyps appearing during childhood. The inheritance pattern is autosomal dominant. Extraintestinal anomalies have been reported in association with juvenile polyps; these include hydrocephalus, thyroglossal duct cyst, tetralogy of Fallot, coarctation of the aorta, idiopathic hypertrophic subaortic stenosis, and malrotation of the gut.

A gene for FJP has been mapped to a locus on chromosome 18q21.⁹⁹ The gene encodes a component of the transforming growth factor (TGF)-β signaling pathway SMAD4.^{100, 101} One can expect that genetic testing will change the way at-risk patients are screened and managed. It will also allow better classification of juvenile polyposis syndromes and help resolve unanswered questions such as these: Is the infantile form of juvenile polyposis allelic with the more common form, or is it an acquired disease akin to a childhood Cronkhite-Canada syndrome? Are the congenital anomalies described in the literature truly a manifestation of FJP, or do they represent missed cases of Cowden syndrome or other genetic conditions?

Juvenile polyps may be found in the colorectum (98%), stomach (14%), duodenum (2%), and jejunum and ileum (7%).^{102, 103} Although the polyps are considered benign, FJP patients are at increased risk for CRC. The cumulative lifetime risk was estimated by Järvinen to be 50%,¹⁰⁴ an estimate that compares remarkably well to a report from the University of Iowa describing a large FJP kindred in which 16 of 29 (55%) affected individuals developed gastrointestinal cancer.¹⁰⁵ Eleven members of the Iowa kindred had colon cancer and six had gastric cancer.

Affected individuals need regular endoscopic surveillance. Scott-Conner and colleagues¹⁰³ recommend upper and lower gastrointestinal endoscopy every 3 years as long as no lesions are detected. Small numbers of polyps can be managed by polypectomy; in these cases, endoscopy should be repeated in 1 year. If there are multiple polyps in the colon, subtotal colectomy is recommended. Multiple gastric polyps, particularly if dysplasia is present, should prompt consideration of gastrectomy.

Children with large numbers of polyps present a difficult management problem. Colectomy is the prudent choice, but we believe regular colonoscopy with removal of the largest polyps until the patient attains puberty is also an option.¹⁰⁵

Peutz-Jeghers Syndrome

PJS is inherited in an autosomal dominant pattern and has been mapped to a locus on chromosome 19p13.3.^{106, 107} The gene responsible for the syndrome encodes a serine threonine kinase, STK11 (also known as LKB1).¹⁰⁶ Abnormal (ie, inactive) forms of this kinase may lead to defective control of cellular growth and differentiation. Not all PJS families can be linked to 19p13.3, leading to speculation that there is a second locus for the syndrome.^{108, 109} The incidence of PJS is unknown; it is thought to be about one-tenth as common as FAP.⁶⁵

The diagnosis of PJS requires histologic confirmation of a hamartomatous, Peutz-Jeghers-type polyp. Because such polyps can be seen in individuals who do not have PJS, clinical diagnosis also requires at least two of the following: small bowel polyposis, family history of PJS, or pigmented macules of buccal mucosa, lips, fingers, and toes.¹¹⁰ Genetic testing for the syndrome is being developed.¹¹¹

Peutz-Jeghers polyps have been found in the entire gastrointestinal tract. The small bowel is the site affected most often, but stomach and colon are involved as well; esophageal polyps are rare. The polyps are almost always multiple but tend to number in the dozens rather than the hundreds. Peutz-Jeghers polyps have also been described in respiratory mucosa and the urinary tract. One member of the Dutch family originally studied by Peutz suffered from severe nasal polyposis and eventually developed a nasal carcinoma.¹¹²

PJS carries increased risk for malignancy. Giardiello and colleagues¹¹⁰ found a relative risk for cancer 18 times that of the general population in 31 PJS patients. Malignancies involved the pancreas (four), breast (two), stomach (two), colon (two), lung (two), and endometrium (one). Spigelman and colleagues¹¹³ reported that 72 retrospectively studied PJS patients were 13 times more likely than the general population to develop a malignancy. The tumors involved the colon, stomach, small intestine, ovary, fallopian tube, thyroid, and lung. Investigators from the Mayo Clinic¹¹⁴ found a relative risk for cancer of 9.9 in 34 PJS patients, with cancers of the colon (7), breast (6), lung (3), and cervix (2) predominating. The Mayo Clinic group found a particularly high incidence of breast and gynecologic cancers, contributing to a relative cancer risk of 18.5 in women, as compared to 6.2 in men.

The Dutch family originally described by Peutz has been updated, and the findings further demonstrate that PJS is not a benign condition.¹¹² Of 22 affected individuals, 7 developed carcinoma (3 colon, 1 stomach, 1 gastrointestinal not otherwise specified, 1 breast, and 1 nasal cavity). All patients with malignancy were dead of their disease before the age of 50 years.

Nearly every female PJS patient will have ovarian involvement by sex cord tumor with annular tubules (SCTAT). The tumors are bilateral in at least two-thirds of cases (in contrast to SCTAT in the sporadic setting, which is almost always unilateral). An unusual form of cervical cancer, minimal deviation adenocarcinoma (adenoma malignum), is also characteristic of PJS. This rare tumor accounts for 1% to 3% of all cervical adenocarcinoma, but in one series, affected 4 of 27 women with PJS.¹¹⁵

The surveillance protocol advocated by the St. Mark's Polyposis Registry¹¹⁶ includes yearly hemoglobin and yearly ultrasound of pelvis in females and pancreas in all patients. Testicular ultrasound should also be done in males with feminizing features. Biannual upper and lower endoscopy with small-bowel x-ray are recommended. Regular mammography and cervical smear are critical surveillance measures. Tomlinson and Houlston¹¹⁷ suggest that upper endoscopy, colonoscopy, and small-bowel x-ray begin in the second decade and that mammography begin at age 25 years.

The gastrointestinal polyps may be associated with bleeding, obstruction, or intussusception. Conservative removal (snare polypectomy) is favored over segmental resection of bowel to avoid development of a short-bowel syndrome.

Eng and Ji,⁹⁸ in discussing the differential diagnosis of juvenile polyposis syndrome (JPS), Cowden syndrome, Bannayan-Ruvalcaba-Riley (BRR) syndrome, and Peutz-Jeghers syndrome, note that there are often very subtle clinical distinctions among them. Hence, the value of emerging evidence on their molecular bases (Table 16-2). For example, germ line mutations in PTEN have been identified in Cowden and BRR that show the two syndromes may at least be allelic, and “…might even be one and the same syndrome along a broad spectrum.” Eng and Ji discuss the problem of phenotypic features that may be shared by the several hamartomatous syndromes, thereby contributing to the complexity of clinical diagnosis, which may not be straightforward. They suggest the referral of such cases to physicians with extensive experience with these disorders will prove helpful.

Table 16-2

Hamartomatous Polyposis Syndromes.

Hereditary Nonpolyposis Colorectal Cancer (Lynch Syndrome)

The history of hereditary nonpolyposis colorectal cancer (HNPCC) dates to an observation of Aldred Warthin, who was a pathologist at the University of Michigan School of Medicine.¹¹⁸ He became deeply moved when his seamstress, in 1895, told him that she would likely die of cancer of the colon, stomach, or her female organs, because of the enormous proclivity to these cancers in her family (unfortunately, just as she had told Warthin, she died at a young age of metastatic endometrial carcinoma). Warthin listened intently, developed her pedigree, and along with other similar cancer-prone families published this work in 1913.¹¹⁹ He updated the family in 1925.¹²⁰ The seamstress's family has since been known as Family G.

Lynch and colleagues⁴ described the natural history and genetics of two large Midwestern kindreds (Families N and M) in 1966. The clinical genetic features in these families were similar to those of Family G.⁴ Dr. A. James French, Warthin's successor as chairman of pathology at the University of Michigan, heard about Lynch's research on Families N and M,⁴ and recalled that Warthin had discovered a similar family (Family G) in 1895. Lynch was then invited by French to take custody of all the detailed documents and pathology specimens that the meticulous Warthin had investigated, cataloged, and published over a span of more than 30 years.^{119, 120} Family G was then updated and published in 1971.¹²¹ This material is discussed in a more detailed review of the history of HNPCC.¹²² Through the use of conversion technology,¹²³ an MSH2 mutation was identified in Family G in the year 2000.

HNPCC is inherited in an autosomal dominant pattern. A germ line mutation in one of the genes responsible for DNA mismatch repair (MMR) can be demonstrated in the 40% to 60% of families meeting clinical criteria for HNPCC. The incidence has not been definitely established, but a large-scale molecular screening study from Finland indicates that HNPCC accounts for approximately 2% of the CRC burden in that country.¹²⁴ Other estimates of its frequency range much higher.¹²⁵

Mutations in six different MMR genes have been identified in HNPCC patients^{126, 127}: MLH1,¹²⁸ located on chromosome 3p21.3; MSH2 ^{129, 130} and MSH6,¹³¹ both located on 2p21; PMS2,¹³² located on 7p22; MLH3,^{133, 134} located on 14q24.3; and possibly PMS1,¹³² located on 2q31–q33. However, only 40% to 60% of HNPCC patients harbor identifiable germ line mutations, most common of which are MSH2 and MLH1 mutations.¹³⁵ Approximately 90% of the identified HNPCC mutations involve MLH1 or MSH2, while mutations in the MSH6 gene account for approximately 10%. MSH6 mutations appear to predispose to an atypical form of HNPCC characterized by an excess of endometrial cancer and a deficit of CRC.¹³⁶ These findings therefore suggest that other genes, including modifier genes, may be of etiologic importance in HNPCC; differing mutation types, environmental factors, and/or chance could also explain the etiology of those 40% to 60% of HNPCC families in which no known cancer-causative germ line mutations have to date been identified.^{125, 127}

All cells of affected individuals carry a nonfunctioning allele of a DNA MMR gene; if the wild-type allele is lost or inactivated, the cell can no longer repair DNA mismatches that inevitably arise during DNA replication. Cells with defective DNA MMR accumulate mutations at a very high rate (as much as 1,000 times that of normal cells). Because DNA mismatches are more likely to occur in DNA microsatellites (areas with multiple repeats of one nucleotide or one pair of nucleotides), defective DNA MMR leads to the phenomenon of microsatellite instability, in which the progeny of the defective cells have varying lengths of a given microsatellite.

The predilection of DNA mismatches for mono- and dinucleotide repeats plays a deciding role in the genetic mutations contributing to carcinogenesis. Nearly all colon cancers with microsatellite instability have mutations in the transforming growth factor β type II receptor (TGF-βIIR) and BAX genes, and those mutations are located in repeating sequences. Mutation of TGF-βIIR leads to escape from the growth inhibitory effects of TGF-β, whereas mutation of BAX interferes with its proapoptotic effect. Thus, carcinogenesis in colon cancers with microsatellite instability may involve mutations in critical genes different from those involved in other colon cancers (APC, K-ras, p53, etc.).

Individuals carrying germ line mutations of hMLH1 or hMSH2 have a lifetime risk for CRC on the order of 80%.¹³⁷ The cancers tend to arise proximal to the splenic flexure (70%). The average age at diagnosis is 45 years. Multiple synchronous and metachronous colon cancers are a feature of the syndrome, with 30% of patients developing a second colon cancer within 10 years if a limited operation (right hemicolectomy or segmental resection) is done for the initial cancer. Even when total abdominal colectomy is performed, the rectum is still at risk; Rodriguez-Bigas and colleagues¹³⁸ reported that 12% of HNPCC patients had rectal cancer within 12 years after colectomy.

The colon cancers arising via the microsatellite instability pathway, whether hereditary or sporadic, often have clinicopathologic clues to their molecular pathogenesis. Such tumors are usually located proximal to the splenic flexure, and may have mucinous or signet ring cell morphology.¹³⁹ Medullary carcinoma, a subtype recently recognized in the World Health Organization classification, is found almost exclusively in tumors with microsatellite instability-high (MSI-H).¹⁴⁰ There is commonly a host lymphoid response, either in the form of lymphoid aggregates at the edge of the tumor or lymphocytes infiltrating the tumor. The latter are the single best marker of MSI status.¹⁴¹ Jass developed an algorithm for identifying MSI-H colon cancer.¹³⁹

When one suspects that a tumor has microsatellite instability, immunohistochemical stains for the protein product of the DNA MMR genes MLH1 and MSH2 can be a useful confirmatory tool. For example, Marcus and colleagues found that 37 of 38 neoplasms known to be MSI-H (by polymerase chain reaction [PCR] analysis) demonstrated the absence of MLH1 or MSH2 expression, and 34 of 34 microsatellite stable (MSS) tumors had intact staining.¹⁴² There are, however, important differences between hereditary and sporadic MSI-H cancers that must be understood. Essentially, all sporadic MSI-H colon cancers are the result of hypermethylation of the MLH1 promotor site. In this situation, no protein is produced and immunohistochemical stains are negative. In the hereditary setting, one allele carries a germ line mutation; when the second allele is lost or inactivated, it is possible that the mutant allele will produce a truncated or otherwise altered protein that stains normally but functions abnormally, thus resulting in a falsely normal immunohistochemical stain. Salahshor and colleagues,¹⁴³ for example, described intact staining for MLH1 in 2 of 15 MSI-H cancers from patients with known mutations of MLH1. In a large series of unselected cases, Lindor and colleagues¹⁴⁴ found that 27 of 818 tumors with intact staining of MLH1 and MSH2 were MSI-H. The authors concluded that immunohistochemistry is a specific (100%) and sensitive (92.3%) screening tool, but that some MSI-H tumors will be missed if only immunohistochemistry is performed. We believe the same general rule applies to testing in the hereditary setting, but would emphasize the possibility of falsely normal staining; if clinical features suggest HNPCC, negative immunohistochemistry should not be taken as definitively ruling out the syndrome.

Extracolonic cancers are common in HNPCC.¹⁴⁵ Endometrial carcinoma heads the list; women with HNPCC have a 20% to 60% lifetime risk.^{137, 146} There is also increased risk for carcinoma of the stomach, ovary, renal pelvis and ureter, small bowel, hepatobiliary tract, and pancreas. Glioblastoma multiforme is seen in some HNPCC families. (Families with colon cancer and brain tumors described under the eponym Turcot syndrome have included FAP families with medulloblastoma and HNPCC families with glioblastoma.¹⁴⁷) Muir-Torre syndrome (benign and malignant sebaceous skin tumors in combination with colon cancer and other internal malignancies) has been shown by mutational analysis to be a variant of HNPCC.¹⁴⁸ Although the relative risk for breast cancer is not increased in HNPCC, breast cancers arising in HNPCC demonstrate microsatellite instability,¹⁴⁹ suggesting that breast cancer is a tumor integral to HNPCC, albeit one with lower penetrance.

The diagnosis of HNPCC is based on pedigree assessment. Still widely used are the Amsterdam criteria,¹⁵⁰ which are (1) at least three relatives with CRC, one of whom must be a first-degree relative of the other two; (2) involvement of at least two generations; and (3) at least one colon cancer diagnosed before age 50 years. FAP should be excluded and tumors should be verified by pathologic examination. These criteria helped standardize reporting of HNPCC but had limitations as a case-finding tool, notably the failure to acknowledge any contribution from extracolonic cancer. The importance of extracolonic cancer was demonstrated by Wijnen and colleagues,¹⁵¹ who found three independent variables predictive of germ line mutations in hMSH2 or hMLH1: fulfillment of the Amsterdam criteria, the presence of endometrial cancer in the kindred, and early age at diagnosis of colon cancer. Table 16-3 lists the new criteria (Amsterdam II) developed by the International Collaborative Group on HNPCC.¹⁵²

Table 16-3

ICG-HNPCC Definition of HNPCC (Lynch Syndrome).

Gene testing is available on a clinical basis but is costly and not as sensitive as testing for FAP. Remember that several genes need to be considered (clinical testing is currently limited to hMLH1 and hMSH2). Hundreds of different mutations have been described,¹²⁷ and although many of them are predicted to be truncating mutations (the easiest to detect), the prevalence of truncating mutations is less than in FAP. Also, it appears that one-third of all pathogenic hMSH2 mutations, at least in the Dutch population, show deletion of almost all of one hMSH2 allele.¹⁵³ These large deletions are difficult to detect by the usual screening and sequencing techniques.

Because of the early age of CRC onset in HNPCC and its penchant for the proximal colon, full colonoscopy should be initiated by age 20 to 25 years in germ line mutation carriers and those at risk based on pedigree analysis. Colonoscopy should be performed at least every 2 years, given the problem of accelerated carcinogenesis of CRC in HNPCC.^{125, 154} We prefer every other year in high-risk patients who have not had DNA testing, and annually in patients with HNPCC germ line mutations or who are obligate gene mutation carriers by their position in the pedigree.

Järvinen and colleagues showed the benefit of colonoscopic screening in HNPCC through a controlled clinical trial extending over 15 years.¹⁵⁵ The incidence of CRC was compared in two cohorts of at-risk members of 22 HNPCC families. CRC developed in 8 screened subjects (6%), compared with 19 controls (16%; p = 0.014). The CRC rate was reduced by 62%. All CRCs in the screened group were local, causing no deaths, compared with nine deaths caused by CRC in the controls. It was concluded that CRC screening at 3-year intervals more than cuts in half the risk of CRC, prevents CRC deaths, and decreases overall mortality by about 65% in HNPCC families. The relatively high incidence of CRC even in the screened subjects (albeit without deaths) argues for shorter screening intervals, for example, 1 year. In addition, Vasen and colleagues¹⁵⁶ discovered five interval cancers in HNPCC patients within 3.5 years following a normal colonoscopy.

In reviewing this subject, Church¹⁵⁷ suggests that interval CRCs develop from normal epithelium within 3 years or from adenomas that were missed. It is important to realize that colonoscopy “miss” rates are as high as 29% for polyps < 5 mm in diameter.¹⁵⁸ Patients should be advised that colonoscopy is not a perfect screening procedure, and the option of prophylactic colectomy should be discussed.^{159, 160}

Subtotal colectomy as a prophylactic measure among HNPCC patients remains controversial, but patients who carry germ line mutations should be offered this option as an alternative to lifetime colonoscopic surveillance. Genetic counseling must be provided so that patients can be in a better position to evaluate the various management strategies. Church¹⁶⁰ and Lynch¹⁵⁹ both suggest that prophylactic surgery should be an option for patients likely to show reduced compliance for colonoscopy. Syngal and colleagues¹⁶¹ examined the life expectancy and quality-adjusted life expectancy benefits resulting from endoscopic surveillance and prophylactic colectomy among carriers of germ line mutations for HNPCC. Both risk-reduction programs showed large gains in life expectancy for mutation carriers, with benefits of 13.5 years for surveillance and 15.6 years for prophylactic proctocolectomy at 25 years of age, compared with no intervention. The benefits of prophylactic colectomy decreased with increasing age.

Women at risk for HNPCC should have annual screening for endometrial cancer beginning at age 25 to 35 years. Endometrial aspiration or transvaginal ultrasound are advised for screening. Prophylactic hysterectomy and oophorectomy can be considered when childbearing is completed.

Those with hematuria should have annual ultrasound and urinalysis with cytologic examination. This screening should begin at age 30 years. Periodic upper endoscopy should be performed in families with gastric cancer, and those of oriental origin.

Genotype-Phenotype Heterogeneity

HNPCC, not unlike other autosomal dominantly inherited disorders, is noteworthy for genotypic and phenotypic heterogeneity.^{125, 136, 162} MSH2 mutations may predispose to the more severe phenotypic cancer phenomenon in HNPCC. An excellent example is the study by Vasen and colleagues¹⁶³ that included 138 families with HNPCC wherein mutations were identified in 79 families (34 with MLH1, 40 with MSH2, 5 with MSH6). These investigators found that the lifetime risk for developing cancer at any anatomic site was significantly higher for MSH2 mutation carriers as opposed to MLH1 mutation carriers (p < 0.01). With respect to specific anatomic sites, findings disclosed that, “…The risk of developing colorectal or endometrial cancer was higher in MSH2 mutation carriers than in MLH1 mutation carriers, but the difference was not significant (p = .13 and p = .057, respectively). MSH2 mutation carriers were found to have a significantly higher risk of developing cancer of the urinary tract (p < .05). The risk of developing cancer of the ovaries, stomach, and brain was also higher in the MSH2 mutation carriers than in the MLH1 mutation carriers, but the difference was not statistically significant.” These findings are highly important. In particular, the observation of a significantly higher risk of developing cancer of the urinary tract in those who were carriers of the MSH2 mutation is pertinent. Research continues dealing with the clinical and molecular genetic features of the Lynch syndrome.¹⁶⁴

Multiple Endocrine Neoplasia Type 1

Multiple endocrine neoplasia type 1 (MEN1) is a rare autosomal dominant disorder associated with several endocrine tumors that cause as much harm by the hormones they oversecrete as by their malignant potential. The tumors most commonly seen include parathyroid adenomas (90%), gastrinomas (40%), insulinomas (10%), and other gastrointestinal endocrine tumors, as well as prolactinomas (20%) and other tumors in the anterior pituitary. A diagnosis of MEN1 is made with two of the three primary classes of tumors in an individual, with familial MEN1 diagnosed with one case and at least one first-degree relative with one of the three component tumor types (reviewed by Brandi and colleagues¹⁸⁴). Also seen in this syndrome are carcinoids, lipomas, and angiofibromas. MEN1 was first recognized as a familial disorder by Wermer¹⁸⁵ (for review see Marx¹⁸⁶).

MEN1 often presents as primary hyperparathyroidism and less commonly as Zollinger-Ellison syndrome (ZES), insulinoma, or a pituitary tumor. The majority of those affected will have hyperparathyroidism, with 50% experiencing symptoms by the age of 25 years.¹⁸⁷ As is the case with other familial cancer syndromes, the presentation may be multifocal, as opposed to the solitary tumors found in most sporadic cases. Management of these tumors can be difficult because of the nature of the hormones they secrete. In addition, these tumors are often small, multiple, and difficult to remove surgically.

The MEN1 gene was discovered by positional cloning following its localization on chromosome 11q13.^{188, 189} MEN1 encodes menin, a protein of unknown function that inhibits growth of ras-transformed fibroblasts.¹⁹⁰ More recently, it was shown that inactivation of menin by antisense RNA antagonizes TGF-β-mediated cell growth inhibition, implicating this pathway in menin-related tumor suppression.¹⁹¹ The majority of mutations documented at this time led to early termination of the protein product, suggesting that it is a classic tumor-suppressor gene.¹⁸⁹ Most, if not all, families with MEN1 carry mutations in this gene. Unlike with MEN2, no genotype-phenotype correlation is known to exist, and the role of MEN1 in sporadic tumors has yet to be elucidated.

Clinical DNA testing exists and can save an individual from a MEN1 family from periodic biochemical testing of serum calcium, parathyroid hormone, and prolactin, which otherwise is begun as an adolescent. However, genetic testing of individuals at this young age is controversial as it is not clear that parathyroidectomy in childhood is indicated. The fact that biochemical testing is readily available and often less expensive than mutational analysis of MEN1 makes the role of clinical DNA testing uncertain at present. As we learn more about the MEN1 gene, and can develop effective treatment strategies based on the knowledge of carrier status, DNA testing may take on a more important role.

Multiple Endocrine Neoplasia Type2

MEN2 families are predisposed to develop the triad of medullary thyroid carcinoma (MTC), pheochromocytoma, and parathyroid hyperplasia with hyperparathyroidism (HPT). MEN2A, which encompasses the majority of the cases, usually presents with MTC, which has been documented in infancy and is the major source of mortality in MEN2. MEN2B families have very early onset, aggressive MTC, as well as HPT and characteristic developmental abnormalities involving hyperplasia of the autonomic nerves of the intestines and characteristic facies caused by the disorganized growth of axons on the lips, conjunctiva, and oral mucosa.

Familial medullary thyroid carcinoma (FMTC), the third manifestation of MEN2, consists of families in which only MTC is present. The MTC in FMTC is often later in onset and has a better prognosis than in MEN2A and MEN2B.

MEN2 is caused by mutations in RET.^{192, 193} RET is located on chromosome 10q11.2, has 21 exons, and encodes a transmembrane receptor tyrosine kinase. Unlike all other known familial cancer syndrome genes, RET is a protooncogene. Mutations that increase RET activity lead to the subsequent development of cancer and somatic overgrowth and those that inactivate RET cause Hirschsprung disease (HD), a congenital absence of sympathetic neurons in the distal colon and rectum resulting in severe bowel dysfunction.¹⁹⁴ Unlike most other familial cancer syndromes, MEN2 has a very strong genotype-phenotype association. Ninety-eight percent of MEN2A and 85% of FMTC families have missense mutations altering a cysteine residue in the extracellular juxtamembrane cysteine-rich region (exons 10 and 11), while a few FMTC families have missense mutations in the intracellular tyrosine kinase domain encoded by exon 13. Ninety-five percent of MEN2B families have a single mutation, M918T (exon 16), also in the intracellular tyrosine kinase domain (reviewed by Eng¹⁹⁵).

Prior to the availability of clinical DNA testing for RET mutations, biochemical screening and imaging of individuals at risk for this syndrome began at the age of 4 or 5 years and continued into early adulthood. However, as it is now known that 95% of MEN2A and MEN2B and 85% of FMTC families had documented mutations in RET,^{196, 197} genetic testing has replaced biochemical screening in the diagnosis and management of MEN2. DNA testing for the few common mutations known to be associated with MEN2 can exclude germ line RET mutations in > 99% of those at risk and patients with sporadic MTC.¹⁹⁸ In a family with a known mutation, DNA testing is 100% accurate and can save a noncarrier from unnecessary and uncomfortable biochemical screening. Because of the mortality and very early onset of MTC in MEN2, prophylactic thyroidectomy in all mutation carriers should be performed by age 5 years in MEN2A, and no later than age 6 months in MEN2B on the basis of positive DNA testing alone.^{184, 199}

Neurofibromatosis Type 1

Neurofibromatosis type 1 (NF1) is inherited in an autosomal dominant pattern. The disease incidence is 1 in 3,000, with one-third to one-half of cases representing new mutations. The implicated gene (NF1) is a large gene (59 exons) located on chromosome 17q11.2. Many different mutations of NF1 have been reported; most (70% to 80%) are truncating mutations. A protein truncation test is available for clinical genetic testing.

Clinical diagnosis of NF1 requires two or more of the following: (1) multiple café-au-lait spots; (2) two neurofibromas or one plexiform neurofibroma; (3) multiple axillary or inguinal freckles; (4) sphenoid wing dysplasia or congenital bowing or thinning of long bone cortex; (5) bilateral optic nerve gliomas; (6) two or more iris hamartomas; (7) a first-degree relative with NF1 by these criteria.²⁰⁰ In addition to the lesions listed above, mutation carriers are at risk for neurofibrosarcoma (3% to 15% of affected individuals), pheochromocytoma, duodenal carcinoid, neuroblastoma, ependymoma, rhabdomyosarcoma, and Wilms tumor (WT). Children with NF1 are at increased risk for juvenile myelomonocytic leukemia (juvenile chronic myelogenous leukemia) and the monosomy 7 syndrome, a childhood myelodysplasia.

The NF1 gene encodes a guanosine triphosphatase-activating protein known as neurofibromin. The gene product negatively regulates signals transduced by ras proteins. Side and colleagues²⁰¹ note that NF1 apparently “…functions as a tumor-suppressor gene in immature myeloid cells but inactivation of both NF1 alleles has not been demonstrated in leukemic cells from patients with neurofibromatosis type 1.”

There are few screening recommendations for patients with NF1. Blood pressure should be monitored twice a year (pheochromocytoma). Neurologic symptoms (headaches, hearing loss, visual change) should be sought and investigated.

Neurofibromatosis Type 2

NF2 is also autosomal dominant. Disease incidence is 1 in 35,000, with about one half of cases representing new mutations. The NF2 gene is on chromosome 22q12.2. Genetic testing is available clinically.

The consensus criteria for NF2 require either (1) bilateral nerve VIII masses or (2) a first-degree relative with NF2 in a patient with one of the following: unilateral nerve VIII mass, a plexiform neurofibroma, two or more neurofibromas, two or more gliomas, two or more meningiomas, posterior subcapsular cataract at a young age, imaging evidence of an intracranial, or a spinal cord tumor.²⁰⁰

Patients are at risk for central nervous system (CNS) tumors, the most common of which are vestibular schwannomas, schwannomas at other sites, meningiomas, and ependymomas. Although these tumors are often benign and slow growing, their location within the CNS may lead to intracranial and intraspinal involvement with a high rate of morbidity and mortality. Affected individuals also typically develop hearing loss (often bilateral), imbalance, tinnitus, facial weakness and headache, and posterior capsular lens opacities. The average age at onset is in the mid twenties.²⁰²

NF2 patients may be clinically subdivided into a severe type and a mild subtype, a classification that is based on the age at onset of symptoms, the number and type of tumors developing, and the duration of disease.²⁰³ The severe type is termed the Wishart form of the disease, which is usually manifested prior to 25 years of age. These patients rarely survive past 50 years of age. Those with the mild subtype, referred to as the Gardner subtype, present with symptoms usually later in life (after 25 years of age), develop a lesser number of more slow-growing tumors, have only bilateral vestibular schwannomas, and generally survive beyond the fifth decade. Ruttledge and colleagues²⁰³ suggest that the majority of familial occurrences of NF2 involve only one form of the disease (ie, Wishart type or Gardner subtype), but they note that Kanter and colleagues²⁰⁴ have shown that some families will manifest both extremes with intermediate cases occurring within these families. Annual neurologic examination, to include audiologic and ophthalmologic tests, is advised.