University of Connecticut Health Center, Farmington, Connecticut 06030-1850
Address all correspondence and requests for reprints to: Carl D. Malchoff, M.D., Ph.D., Department of Medicine, University of Connecticut Health Center, 263 Farmington Avenue, Farmington, Connecticut 06030-1850. E-mail: . malchoff{at}nso2.uchc.edu
Nonmedullary thyroid carcinoma (NMTC) includes papillary thyroid carcinoma (PTC), follicular thyroid carcinoma (FTC), the Hurthle cell variant of FTC, insular thyroid carcinoma, and anaplastic thyroid carcinoma. PTC is the most common and occurs with an incidence of about 1 per 20,000 per year (1). Although most NMTCs are sporadic, a familial predisposition for PTC represents an important variant. In 1997, Fagin (2) reviewed the evidence for a familial predisposition to PTC in JCEM. He appropriately concluded that there was strong evidence for a familial predisposition to NMTC as a component of known familial tumor syndromes, but that familial NMTC (fNMTC) as a distinct inherited tumor syndrome had not been unequivocally established. It is still fair to conclude that fNMTC has not been unequivocally established, because this will require the identification of a susceptibility gene for this disorder. Individuals with germline mutations of a NMTC susceptibility gene are at risk for developing NMTC. Recent observations provide further evidence for a familial predisposition to PTC and suggest that there may be three or more PTC susceptibility genes. The identification of these genes will unequivocally establish familial PTC (fPTC) as a distinct entity, provide an important diagnostic tool for the practicing physician, and clarify the molecular mechanisms of thyroid tumorigenesis.
A familial predisposition to medullary thyroid carcinoma (MTC) is well recognized, and the genetic abnormalities of multiple endocrine neoplasia (MEN) types 2A and 2B have been characterized. Genetic testing has transformed the clinical approach to these patients. In contrast, a familial predisposition to nonmedullary thyroid neoplasia confronts the traditional teaching that NMTC is sporadic and not familial. However, epidemiological studies suggest that about 5% of all PTC may have a familial component (3). The results of the human genome project and powerful genetic techniques can be applied to fPTC to identify the susceptibility genes. Mutations of these genes are the earliest gene changes that predispose to development of this neoplasm.
fNMTC can be divided clinically into two groups (Table 1). In the first group, PTC is the predominant clinical feature of a familial tumor syndrome. In the second group of disorders, NMTC is a relatively infrequent component of a familial tumor syndrome. There are no known familial syndromes characterized by a predominance of follicular, insular, or anaplastic thyroid carcinoma. A related group of interest is comprised of the familial syndromes that predispose to nodular goiter. This review will discuss the evidence that supports a distinct familial tumor syndrome(s) characterized clinically by a predominance of PTC and distill the clinically relevant information from these studies. It will briefly review the second group of inherited tumor syndromes characterized by NMTC as a relatively infrequent component. Clinical care suggestions are derived from this new information (Table 2
).
|
|
A familial predisposition to PTC is supported by epidemiological studies, pathological examination of tumor tissue, and both the clinical and genetic analysis of large kindreds. Epidemiological studies have repeatedly observed a 4- to 10-fold increased PTC risk in first-degree relatives of PTC subjects (4, 5, 6, 7, 8). This observation is consistent with a familial predisposition to PTC, but could also be explained by environmental factors or possibly by ascertainment bias. PTC is multifocal in about 2060% of cases (3). This observation could be explained by intrathyroidal spread of the primary PTC or by a predisposing factor that could be either environmental or inherited. It is attractive to hypothesize that a genetic predisposition to PTC accounts for some of the epidemiological and pathologic findings.
If fPTC does exist, then it may be possible to identify large kindreds enriched in PTC. Such kindreds will provide clues to the clinical features of fPTC and will serve as the necessary substrate for positional cloning of the susceptibility gene(s). A number of kindreds have been reported (9, 10, 11, 12, 13, 14, 15, 16), and their clinical features have been summarized (3). Inheritance is autosomal dominant with partial penetrance, women are affected more than twice as frequently as men, and the modal age of diagnosis is 3039 yr. In some, but not all kindreds, PTC is more aggressive than is expected for sporadic PTC (3, 17, 18).
As anticipated, the analysis of fPTC with powerful genetic techniques is uncovering the molecular basis of this disorder. Initially, genetic linkage studies confirmed the clinical impression that fPTC was not a phenotypic variant of established familial tumor syndromes such as familial adenomatous polyposis (FAP), MEN type 1, and the Cowden disease (19, 20). Subsequent linkage studies divided fPTC into at least three genetically distinct disorders. A syndrome of fPTC enriched with papillary renal neoplasia (PRN) is referred to as fPTC/PRN and has been mapped to 1q21 (19). A familial syndrome characterized by PTC alone has been mapped to 2q21 (21). Two different studies report genetic linkage of a large PTC kindred to 19p13.2 (22, 23). Interestingly, contrasting clinical features characterize these kindreds. The kindred described by Canzian et al. (22) is noted to have thyroid tumors with cell oxyphilia (TCO). All of the thyroid neoplasms were oxyphilic on pathologic examination, and many were benign. Of nine affected subjects, three had PTC, and six had benign multinodular goiters. In contrast, thyroid neoplasms from the kindred described by Bevan et al. (23) did not demonstrate the pathologic changes of cell oxyphilia, all six affected subjects had PTC, and there were no benign nodules. It is possible that these two clinically distinct kindreds represent genetically distinct disorders with susceptibility genes located close together on chromosome 19.
Candidate susceptibility genes are those genes with chromosomal locations within the linkage region and for which there is reason to believe that they predispose to thyroid neoplasia. In our own studies of the fPTC/PRN syndrome that has been mapped to 1q21, there are three genes that are obvious candidate susceptibility genes. NRAS has been implicated in multiple malignancies (24, 25), PRCC is the N terminus of a fusion gene found in some sporadic papillary renal carcinomas (26), and NTRK1 is a tyrosine kinase that is rearranged in about 15% of sporadic PTC (27). However, preliminary studies indicate that none of these is the fPTC/PRN susceptibility gene (28). No mutations were found in either the coding region of PRCC or the critical coding regions of NRAS. NTRK1 is expressed in neither normal thyroid tissue nor the normal thyroid tissue of patients with the fPTC/PRN syndrome. Because the product of this gene is not found in the normal thyroid of fPTC/PRN subjects, inherited mutations of NTRK1 are very unlikely to affect cell division in the thyroid. Because the most obvious candidates are not the fPTC/PRN susceptibility genes, the identification of this gene will require further positional cloning strategies as outlined by Eng (29) in her discussion of fPTC. In summary, the chromosomal locations of at least three and possibly four syndromes with a preponderance of fPTC have been determined. These findings strengthen the argument for a familial form of PTC, but final confirmation requires the identification of the susceptibility gene(s).
After linkage has been established, it is possible to evaluate the clinical characteristics of fPTC syndromes with greater accuracy. All subjects carrying the chromosomal region with the susceptibility gene will carry the same variable nucleotide repeat polymorphisms, so that they can be distinguished from genetically unaffected individuals that have different variable nucleotide repeat polymorphisms. Individuals without PTC who carry the susceptibility gene can be distinguished from individuals without PTC who do not carry the susceptibility gene. These analyses confirm that fPTC inheritance is autosomal dominant with partial penetrance, and that women are affected more frequently than men. Most familial tumor syndromes are associated with an increased risk of neoplasia in more than one tissue type. This is expected, because most susceptibility genes are expressed in more than one tissue type and participate in basic mechanisms that regulate cell division or cell death. Therefore, it is anticipated that fPTC syndromes will have an increased incidence of neoplasia in nonthyroidal tissues and that review of those individuals carrying the susceptibility gene will uncover associated neoplasms. In our investigations, a large kindred with fPTC is enriched in PRN. Two subjects carrying the same chromosome 1q21 haplotype as did the PTC subjects developed PRN. One of the PRN subjects was affected with PTC, and the other was likely an obligate carrier for the PTC susceptibility gene. Immunohistochemical analysis of these renal tumors for thyroglobulin was negative, indicating that the PRN did not represent PTC metastatic to the kidney (19, 20). Because PRN is quite uncommon in the general population, it is unlikely that two subjects in the same PTC kindred will develop PRN by chance alone. One of the PRN subjects had multifocal papillary renal adenomas, suggesting an underlying genetic predisposition to PRN in the kidney. Finally, both linkage analysis and sequence analysis of the MET proto-oncogene demonstrated that the fPTC/PRN syndrome is not a variant of familial papillary renal carcinoma caused by inherited activating mutations of MET proto-oncogene (19). Therefore, multiple lines of evidence indicate that PRN is a component of this PTC syndrome and that the fPTC/PRN syndrome is clinically and genetically distinct from other familial tumor syndromes. This finding is useful in identifying other kindreds that are likely to carry mutations in the PTC/PRN susceptibility gene. We have observed a number of other neoplasms in the fPTC/PRN kindred, but none occur with great enough frequency to prove that they are a component of this familial tumor syndrome (19, 20). In contrast, the other fPTC syndromes have not been reported to have an increased frequency of neoplasia in nonthyroidal tissues. Therefore, these syndromes are likely to be caused by susceptibility genes with either expression or neoplasia predisposition limited to the thyroid.
The possible associations of fPTC with benign thyroid nodules, colon cancer, and breast cancer remain to be clarified. These nonthyroidal neoplasms are relatively common in the general population. Therefore, large numbers of genetically affected subjects must be investigated to determine whether these neoplasms occur at a frequency greater than that expected by chance alone. The initial clinical reviews suggested that benign thyroid neoplasms occurred with greater frequency in family members of fPTC subjects (3). An underlying neoplasia susceptibility gene could predispose to benign or malignant neoplasia, depending upon what other mutations of tumor suppressor genes and oncogenes accumulate in the parent cell of the clonal neoplasm. Similar examples are found in other familial tumor syndromes. However, because benign thyroid nodules are common in the normal population, convincing evidence of an increased frequency of benign thyroid nodules in the fPTC syndromes awaits the clinical evaluation of more genetically affected individuals. The exception to this is the TCO syndrome in which there were twice as many subjects with benign nodules as there were with PTC. The pathological finding of cell oxyphilia was the common characteristic of all these neoplasms (22). Early clinical studies suggested an increased incidence of colon cancer (10). In the fPTC syndromes with linkage to specific chromosomal regions, an increased incidence of colon cancer has not been observed. However, a small increased risk may have been missed. Epidemiological studies have investigated the association of NMTC and breast cancer in the same individual and in first-degree relatives of PTC subjects. Although there are conflicting reports, the most consistent finding is an increased incidence of premenopausal breast carcinoma associated in PTC subjects (30). Because premenopausal breast carcinoma is less common than postmenopausal breast carcinoma, a predisposing factor common to PTC and breast carcinoma may be more easily detected in premenopausal breast carcinoma. One possible explanation for this association is that the I131 used to treat thyroid cancer predisposes to the development of breast cancer. However, it is also possible that there is a common environmental or genetic factor that predisposes to both malignancies. Future studies of the fPTC syndromes will determine whether there is a genetic predisposition common to PTC and breast carcinoma. In summary, an association of PTC with benign thyroid nodules seems likely, but further evaluation is necessary. An association of PTC with breast carcinoma and colon carcinoma remains to be confirmed.
Familial syndromes with a predominance of benign thyroid nodules (Table 1)
Disorders potentially related to fNMTC include familial nodular goiter syndromes. Although these thyroid neoplasms are benign, these disorders may provide clues to thyroid tumorigenesis, because the nodules represent abnormal thyroid growth. One familial goiter syndrome has been mapped to 14q (31) and another to Xp22 (32). The susceptibility genes are not known, and there is no increased incidence of NMTC in the kindreds that were investigated. A kindred with goiter and FTC has been reported, but no genetic studies have been undertaken (33). Thyroid nodules associated with either hyperthyroidism or hypothyroidism occur in the setting of mutations in genes that are components of the pathways involved in thyroid hormone production. Activating mutations of the thyroid-stimulating hormone receptor (34) may be inherited and cause hyperthyroidism and thyroid nodules. Numerous inherited syndromes are associated with defects in thyroid hormone production and hypothyroidism. The elevated TSH is presumably the cause of thyroid growth. These will not be reviewed here.
Familial tumor syndromes with NMTC as a relatively infrequent component (Table 1)
NMTC is a relatively infrequent component of both FAP and the Cowden disease, but not a component of the Li-Fraumeni syndrome. FAP is characterized by the familial predisposition to multiple colonic polyps that progress to colon cancer. It is caused by inherited mutations of APC, a tumor suppressor gene. Extraintestinal clinical manifestations include pigmented retinal lesions, jaw cysts, sebaceous cysts, osteomata, and PTC. The prevalence of PTC in FAP individuals is about 510 times that of the normal population, and its prevalence is about 2% in genetically affected individuals (35). Because the APC gene is a tumor suppressor gene, it is anticipated that loss of the normal allele should accompany the development of PTC in this syndrome as it does the development of colonic polyps. Interestingly, tumor-specific loss of heterozygosity and tumor-specific APC mutations were not found in six PTC associated with FAP (36). Possibly, the normal allele is inactivated by a different mechanism, or a single abnormal allele produces a dominant effect in the thyroid follicular cells. Cowden disease is characterized clinically by a familial predisposition to hamartomas, breast carcinoma, and to a lesser frequency, NMTC. The thyroid neoplasm may be FTC or PTC. Cowden disease is caused by inherited mutations of the PTEN tumor suppressor gene (37). The Li-Fraumeni syndrome is characterized clinically by multiple neoplasms including rhabdomyosarcomas, soft tissue sarcomas, breast carcinoma, adrenocortical carcinoma, and leukemia. It is caused by inherited mutations of the p53 tumor suppressor gene (38). Sporadic mutations of the p53 gene were found in five of six anaplastic thyroid carcinomas (39). In contrast to expectations, anaplastic thyroid carcinoma is not a component of the Li-Fraumeni syndrome. It is not understood why sporadic p53 gene mutations in the thyroid follicular cell should lead to anaplastic thyroid carcinoma, whereas inherited mutations do not. In summary, an increased prevalence of NMTC is found in FAP and Cowden syndrome, but not the Li-Fraumeni syndrome.
Clinical suggestions (Table 2)
Some clinical suggestions can be derived from these studies that apply powerful genetic techniques to the analysis of fPTC. Individuals with FAP and the Cowden disease should receive genetic counseling and probably should be screened for NMTC with a yearly physical examination of the thyroid. Guidelines for these disorders are published and discussed elsewhere (40, 41). A family history should be extracted from patients affected with PTC. The coexistence of both PTC and PRN in a single individual or a PTC-affected individual with a first-degree relative with documented PTC or PRN should suggest a familial PTC syndrome.
If an fPTC syndrome is identified, then first-degree relatives of affected subjects should be screened. The screening method remains to be clarified. Prophylactic thyroidectomy, which is indicated in MEN types 2A and 2B, is not appropriate in the fPTC syndromes, because PTC is generally less aggressive than MTC and the absolute certainty of affection status cannot be established by linkage analysis. First-degree relatives of affected fPTC kindred members can be evaluated with yearly physical examination of the thyroid, probably starting about age 20 yr. Some clinicians employ thyroid ultrasound as well. Ultrasound carries greater sensitivity, but also is more likely to identify unrelated and clinically unimportant abnormalities. Members of those very rare kindreds with aggressive PTC probably do warrant evaluation with thyroid ultrasound examination yearly. The etiology of thyroid nodules should be determined by cytologic or pathologic analysis. The role of renal ultrasound to screen for PRN has a limited role, because an increased incidence of PRN is not found in most fPTC kindreds. In those kindreds enriched in PRN, the frequency of PRN is less than that of PTC. Screening should be started later in life and should be performed less frequently. Genetic screening is a research tool and is appropriate only for members of those exceedingly rare large kindreds with multiple affected subjects in which linkage has been established. Caution should be taken when sharing research information with genetically affected and unaffected subjects, because linkage studies are based on statistical probabilities, and the susceptibility genes are still unknown.
Summary and conclusions
Unequivocal evidence of fPTC as a distinct syndrome awaits the identification of the susceptibility genes, as suggested by Fagin (2). However, the clinical and genetic evidence is sufficiently compelling that clinical insights can be derived from these studies. A familial susceptibility occurs in about 5% of PTC. There are at least three familial PTC syndromes characterized by a predominance of PTC, and inheritance is autosomal dominant with partial penetrance. Clinical recommendations are suggested for the identification of fPTC kindreds and for screening the first-degree relatives of affected kindred members.
Acknowledgments
Footnotes
Abbreviations: FAP, Familial adenomatous polyposis; fNMTC, familial NMTC; fPTC, familial PTC; FTC, follicular thyroid carcinoma; MEN, multiple endocrine neoplasia; MTC, medullary thyroid carcinoma; NMTC, nonmedullary thyroid carcinoma; PRN, papillary renal neoplasia; PTC, papillary thyroid carcinoma; TCO, thyroid tumors with cell oxyphilia.
Received March 19, 2002.
Accepted April 8, 2002.
References