Metabolic Diseases Branch (S.J.M.), National Institute of Diabetes and Digestive and Kidney Diseases; and Pediatric and Reproductive Endocrinology Branch (L.K.N.), National Institute of Child Health and Human Development, Bethesda, Maryland 20892
Address all correspondence and requests for reprints to: Stephen J. Marx, M.D., Building 10, Room 9C-101, NIH, Bethesda, Maryland 20892-1802. E-mail: StephenM{at}intra.niddk.nih.gov
Development and growth of tumors in the anterior pituitary are incompletely understood (1, 2, 3). However, differences in tumor behavior have long been recognized if pituitary tumors were grouped according to the hormones oversecreted. For example, ACTH-secreting tumors are the smallest, and PRL- or GH-secreting tumors are typically larger (4, 5, 6). In contrast, tumor behavior has been assumed to be indistinguishable if pituitary tumors were grouped according to the presence or absence of MEN type 1 (MEN1 1) (7, 8). In this issue, Verges et al. report the first systematic analysis of pituitary tumors in patients with MEN1 vs. patients without (9). The frequency distribution of the oversecreted hormones was similar in both tumor groups. However, the pituitary tumors in MEN1 were more aggressive as assessed by size and by responsivity to treatments. The tumor size difference was evidenced by a doubled and high prevalence of macroadenoma (85% in tumors with MEN1 vs. 42% without MEN1). Greater resistance to treatment was deduced because only 42% of tumors in MEN1 cases suppressed to a normal hormone secretion after treatments, compared with 90% without MEN1. These two differences, together labeled as greater pituitary tumor aggressiveness in MEN1, are unexpected, have clinical implications, and raise interesting issues about tumorigenesis mechanisms.
Could these novel findings about tumor aggressiveness reflect study biases? The MEN1 group was very large (136 cases with pituitary tumor in MEN1), and those data were obtained retrospectively from a questionnaire-based registry. Although the large number of tumors helped to assure broad tumor representation, survey responses are one possible source of bias. Although the vast majority of all pituitary tumors occur without MEN1 syndrome, patient numbers in the study groups were balanced toward more MEN1 cases. Both groups were well matched for important demographic variables, but differences in evaluation and reporting between the two groups remain another possible source of bias. Diagnostic criteria are another possible source of bias. Hormone oversecretion was partially undefined, and there was ambiguity about whether a tumor was diagnosed if a hormone oversecreting process was not seen as a mass on imaging (9). Because of the unresolved issues, findings of Verges et al. (9) justify confirmation and clarification. The following discussion assumes that the findings of Verges et al. (9) are robust.
Molecular and genetic concepts of stepwise tumor development may clarify several important clinical differences between tumors from cases with MEN1 vs. without. The most frequently used synthesis is Knudsons two-hit model of tumorigenesis (10). This was originally based on epidemiologic observations about retinoblastoma, both its early age at onset and its bilaterality in hereditary cases. The two-hit model has evolved to state that a tumor can begin after two separate mutations of a critical gene have occurred in one cell. For oncogenesis through a tumor suppressor like the MEN1 gene, the two hits are inactivating mutations of both MEN1 alleles, thereby removing, at this genetic locus, all inhibition of cell accumulation. In a hereditary setting like MEN1 syndrome, one hit (the first hit) is inherited in each cell of the body through germline transmission of an inactivating mutation in one copy of the MEN1 gene. Thus, only one postnatal hit (the second hit) must occur at the remaining normal copy of the MEN1 gene in one of many susceptible cells (for example, in one of many anterior pituitary cells in MEN1 syndrome); that cell can then begin expansion into a clonal tumor. On the background of these inherited first mutations in every cell, a second mutation can, by random chance, occur relatively early (i.e. cause tumor onset at relatively early age) and even in more than one susceptible cell (i.e. cause tumor multiplicity). In analogous fashion, the nonhereditary tumor of the same tissue type (such as isolated parathyroid adenoma) may begin after one susceptible cell has had postnatally two separate mutations, thereby inactivating both copies of the same critical gene. Two rare events (a postnatal first hit and then a second hit to the same single cell) are much less likely than one moderately rare event (a postnatal second hit in one of many susceptible cells). This helps explain why the nonhereditary tumor typically begins later and in only one location.
The first hits at the MEN1 gene in hereditary or sporadic tumor usually are structurally similar, a point mutation or another small mutation. In analogous manner, the second hits in hereditary or sporadic tumor can be unlike the first hits but still structurally similar to each other, generally a complete chromosomal deletion or a large subchromosomal deletion (11). Similar tumorigenesis models also can account for stepwise accumulation of hits at multiple genes as in polyposis of the colon (12, 13). They also can account for tumorigenesis by gain of function mutations such as with activating mutation of the MET or RET oncogenes (10, 14). Furthermore, these models encompass stepwise changes other than mutation, that is to say epigenetic events like promoter methylation. Because both hereditary and nonhereditary tumor of one tissue type can have a similar or even identical collection of mutations, these models can also explain why hereditary and nonhereditary tumor types in one tissue can behave similarly, albeit beginning at differing ages. One difference of unknown importance is that tumor stromal cells in MEN1 each have one inactive MEN1 allele whereas none are thought to have this in sporadic tumor.
The most frequent endocrine tumors in MEN1 display several clinical features typical for the two-hit tumorigenesis model (earlier patient age at onset of hereditary tumor, multiplicity of hereditary tumors, and similar behavior of tumors whether or not hereditary).
Earlier patient age at onset of tumor
The most common MEN1-related endocrine tumors (parathyroid adenoma and gastrinoma) are expressed at an earlier age with MEN1 than without. Hyperparathyroidism onset age is earlier by 20 yr in MEN1 (9, 11), and Zollinger-Ellison syndrome begins 10 yr earlier (15).
Tumor multiplicity
The commonest endocrine tumors in MEN1 are multiple (parathyroid tumors, duodenal, or pancreatic gastrinoma, and other pancreatic islet tumors) (11). In paired organs, i.e. the parathyroids, this results in bilateral tumors.
Similar behavior of tumor
The most frequent endocrine tumors in MEN1 seem similar in aggressiveness to their counterparts without MEN1. Parathyroid tumors are similar in size with MEN1 vs. without it (16). Also, they are rarely malignant in MEN1, just as the parathyroid tumors without MEN1 are rarely malignant (11). On the other hand, gastrinoma in MEN1 is frequently malignant and sometimes metastatic and, thereby, similar to gastrinoma without MEN1 (17, 18). Suboptimal outcomes of surgery in MEN1 would seem to be exceptions to similarity in tumor behavior. For example, postoperative recurrence of hyperparathyroidism is more frequent with MEN1 (11). Rather than greater aggressiveness of tumor, this probably reflects the usual surgical strategy of subtotal parathyroidectomy in MEN1, deliberately leaving a remnant of hyperfunctional parathyroid tissue in place with associated regrowth potential (11). Surgical cure is particularly rare for gastrinoma in MEN1 (19). The cause is uncertain but may be the multiplicity of undetected gastrinomas in MEN1.
According to the two-hit model of tumorigenesis, pituitary tumors (like parathyroid tumor or gastrinoma) should begin at earlier ages with MEN1 than without, should be multiple, and should show behavior similar to that of tumors without MEN1. The Verges et al. findings about pituitary tumor in MEN1 seem to be an exception to one or more of the above predictions from the two-hit model (9). Onset age for pituitary tumor in MEN1 has been examined in several uncontrolled series. It has been 3342 yr in modest sized series of MEN1 (20, 21, 22, 23) and in Verges MEN1 group (38 yr) (9). It has been no later in studies of pituitary tumor without MEN1 (5, 6). Pituitary tumors do not allow a meaningful comparison for multiplicity; multiplicity of pituitary tumors is rarely recognized, including rarely in MEN1 (24). Postoperative recurrence in all pituitary tumors is frequent (25). Most importantly and in contrast to the prediction of the two-hit model, pituitary tumor aggressiveness was strikingly more abnormal with MEN1 than without (9).
What molecular or genetic explanations might account for the unexpected aggressiveness of pituitary tumor with MEN1? One possible explanation is a difference in the tumor environment during its development. Younger patient age has been associated with more aggressive tumors of the pituitary in sporadic cases (26, 27). However, considering that patient age at pituitary tumor onset is similar with or without MEN1 (see above), this explanation is unlikely. A second possible explanation is that a selective type of MEN1 germline mutation might contribute to aggressiveness of pituitary tumors, analogous to the RET mutation at codon 918 as a cause of unusually aggressive thyroidal C cell cancer (28). To date, studies of germline or somatic MEN1 mutation have found no relation with familial MEN1 phenotype or with aggressiveness of sporadic tumors (11, 29), and Verges et al. do not comment on their germline mutation data regarding this unlikely relation (9). It seems more likely that the greater aggressiveness of pituitary tumor in cases with MEN1 syndrome reflects the certainty of their development through biallelic MEN1 gene mutation.
Most sporadic pituitary tumors, including those in MEN1, are thought to be clonal in origin (1, 30, 31). This implies that the clonal expansion developed after mutation in one or several critical genes. Consistent with the two-hit model, all or most of the pituitary tumors in MEN1 cases have inactivation of both MEN1 alleles (30). Whereas MEN1 biallelic inactivation also occurs with 1035% of sporadic parathyroid adenoma, sporadic gastrinoma, sporadic insulinoma, or sporadic bronchial carcinoid (11), sporadic pituitary tumor consistently shows a lower frequency of MEN1 biallelic mutation (11, 32). However, mutations in genes other than the MEN1 gene have been implicated in most sporadic pituitary tumors. The known gene most frequently mutated in sporadic pituitary tumor is GNAS1, which encodes the subunit of the stimulatory GTP-binding heterotrimer. This is activated by mutation in up to 40% of somatotropinomas (31). The PKAR1A gene at 17q is an unexplored candidate gene, previously implicated in hereditary somatotropinoma as a component of Carney complex (33). Several other unidentified genes have been suggested to contribute to sporadic pituitary tumors. This is based upon subchromosomal loci identified with loss of heterozygosity, fluorescent in situ hybridization, and comparative genomic hybridization (31). In fact, mutation of unidentified genes at chromosomal loci other than the MEN1 locus seems even more frequent in sporadic pituitary tumor than mutation of the known genes. To summarize the published data, the pituitary tumor of MEN1 cases develops after biallelic MEN1 gene mutation. In contrast, sporadic pituitary tumor rarely begins with MEN1 biallelic mutation. The critical genes mutated in most sporadic pituitary tumors are not yet known, but development of sporadic pituitary tumor with mutations of genes other than MEN1 could well be the basis for their lower aggressiveness than that in pituitary tumor of MEN1.
There are precedents for differences in behavior of tumors of the same tissue but with mutation of different genes or even different codons of one gene. Good examples can be found among endocrine tumors (Table 1). With regard to parathyroid tumors, 15% of the tumors are malignant in the hyperparathyroidism-jaw tumor syndrome compared with less than 1% without this syndrome (34, 35). With regard to the adrenal cortex, tumors associated with germline P53 mutation are often malignant, and this may be even more frequent with the Arg337His mutation of P53 (36, 37). With regard to adrenal medullary tumors, major differences in tissue enzymes, biochemical phenotype of tumor, and clinical expressions are dependent on mutation in the VHL vs. RET gene (38). With regard to C cell neoplasia, three codon clusters of germline mutation of the RET gene correlate with three classes of thyroid tumor aggressiveness and with the following three syndromes: MEN type 2a, MEN type 2b (most aggressive), and familial medullary thyroid cancer (least aggressive) (28).
|
Whatever their underlying molecular and genetic mechanisms, the findings of Verges et al. (9) have implications for management of MEN1 cases. The findings suggest increased attention to three interrelated issues: MEN1 carrier ascertainment, surveillance for pituitary tumors in MEN1 carriers, and treatment of pituitary tumors in MEN1. Ascertainment of the MEN1 carrier state (most sensitively by sequence testing the MEN1 gene in germline DNA) should antedate systematic surveillance for tumors. And it should be offered at ages when effective tumor interventions are possible (9, 40). Early age at tumor onset coupled with aggressiveness of the pituitary tumors in MEN1 seem to justify both early and periodic surveillance for pituitary tumors and early treatment. Similar and more detailed recommendations about ascertainment of the MEN1 carrier status and surveillance of pituitary tumor status are in the recent MEN1 guidelines (41). The need for modified treatment programs for pituitary tumors in MEN1 should be assessed. Because pharmacologic suppression is generally less effective for pituitary macroadenomas than for microadenomas, pituitary radiation and/or surgery may attain a greater role in MEN1. This may be a complex problem because Verges et al. report suboptimal responses to treatment among MEN1 cases also with milder grades of pituitary tumor (9). Finally, it will be important to determine whether the expected benefits come from earlier ascertainment for MEN1 carriers, from more rigorous surveillance to identify pituitary tumor in MEN1 carriers, and from earlier treatment for pituitary tumors in MEN1. Similar questions related to carrier screening, tumor surveillance, and intervention are important for the management of many hereditary tumors, whether or not hormone-secreting and whether or not malignant (42).
Footnotes
Abbreviations: MEN1, MEN type 1.
1 Syndromes are indicated in capitals; genes are in italics. In particular, MEN1 is a syndrome, whereas MEN1 is a gene.
Received November 5, 2001.
Accepted November 27, 2001.
References