Adrenocortical Cancer in Carney Complex: A Paradigm of Endocrine Tumor Progression or an Association of Genetic Predisposing Factors?

Carney complex (CNC) was described in 1985 as “the complex of myxomas, spotty pigmentation, and endocrine overactivity” (1). The most frequent manifestations of CNC are lentiginosis, myxomas, and ACTH-independent Cushing's syndrome due to primary pigmented nodular adrenocortical disease (PPNAD). PPNAD is a unique cause of adrenal Cushing syndrome in which bilateral pigmented micronodules are typically visible at macroscopic examination in the adrenal cortex. It is mostly the investigation of this rare type of adrenal tumors that led to the identification of CNC by Dr. J. A. Carney at the Mayo Clinic (2). Since the first description of CNC, an autosomal dominant inheritance was postulated. In 2000, the identification of the CNC1 gene (PRKAR1A), located at 17q22-24, was reported by Stratakis and colleagues (3) at the National Institutes of Health. This gene encodes the regulatory subunit 1A of the protein kinase A, a central component of the cAMP signaling pathway. Because cAMP signaling is activated by various molecular and cellular defects in endocrine tumors, it appeared quickly that the identification of CNC1 was a major step for the understanding of endocrine tumorigenesis. Germline PRKAR1A mutations are heterozygous inactivating mutations. Allelic loss of the wild-type allele is frequently observed in tumors from CNC patients. Therefore, PRKAR1A is considered as a tumor suppressor gene. So far only PPNAD, which is a typical benign secreting tumor with a very slow growth, has been reported in adrenal tumors of CNC patients. In this issue of the JCEM, two cases of adrenocortical cancer (ACC) are reported in CNC patients with germline PRKAR1A mutations. The paper by Anselmo et al. (20) describes a large Azorean family with at least 10 CNC patients and a germline heterozygous PRKAR1A point mutation (S147G). The index case was a 28-yr-old female who presented with a right 12-cm ACC with liver metastasis arising in the background of bilateral PPNAD. Clinical and hormonal investigations showed cortisol and androgen excess. Three other family members presented with PPNAD and a typical paradoxical increased cortisol secretion after dexamethasone suppression test. The other frequent CNC manifestation in this family was lentiginosis (facial pigmented macules typically located on the lips and eyelids in CNC). The paper by Morin et al. (21) describes a 22-yr-old female with cardiac myxoma and lentiginosis. Her mother also suffered from cardiac myxoma at age 23 and had lentiginosis. A germline heterozygous PRKAR1A mutation that creates a premature stop codon in exon 2 of the gene (p.Lys32Argfs*12) was found in the index case. Mutant PRKAR1A mRNA with a premature stop codon are usually degraded by a mechanism of nonsense-mediated mRNA decay, inactivating the mutant allele. This applies to the majority (about 80%) of the PRKAR1A mutations that have been reported (4). This 22-yr-old female patient was diagnosed with a cortisol- and androgen-secreting 8.5-cm tumor of the left adrenal that was removed by laparoscopy. Pigmented micronodules typical of PPNAD were observed at pathological examination in the adrenal adjacent to the ACC. Three months after surgery, local recurrence was observed. The patient was then reoperated and treated with mitotane. Twenty-nine months later, liver and pulmonary metastasis were progressing. In these two cases, the clinical characteristics of the unilateral adrenocortical tumor that developed in the background of PPNAD are quite typical of ACC. The cosecretion of androgen and cortisol is observed in the majority of secreting ACC (5), whereas benign secreting tumors cause cortisol excess only. The large size of the tumors is also suggestive of ACC. In both cases, the rapid occurrence of metastasis or local recurrence that was proven by pathological investigation to be of adrenal origin is demonstrative of the malignant nature of the adrenal tumor. Because the adjacent adrenal cortex showed pigmented micronodules, it is clear that the ACC developed in the background of PPNAD. This is compatible with a model of tumor progression from adrenal hyperplasia to benign nodule, and then cancer. This multistep tumorigenesis is a general scheme usually accepted for colon cancer. Some rare cases of ACC that have apparently developed in the background of a benign tumor have been reported previously (6, 7). Considering that most adrenal incidentalomas are benign adenomas, the prevalence of adrenocortical adenomas might be higher than 1%, whereas the prevalence of ACC is estimated between 4 and 12 per million (8). Therefore, if this multistep progression was a general rule, the epidemiology tells us that the probability of a benign adrenocortical lesion to acquire the full set of molecular alterations needed to become a malignant tumor would be very low. The study of familial tumor syndromes responsible for ACC has been, before the genomic contemporary area, the most productive way to elucidate the molecular genetics of adrenocortical tumors (9). The study of the Li-Fraumeni syndrome due to germline mutations of the tumor suppressor gene TP53 led to the identification of somatic TP53 mutations in sporadic ACC. The study of the Beckwith-Wiedemann syndrome due to overexpression of the growth factor IGF-II led to the report of IGF-II overexpression in sporadic ACC. However, in these two inherited neoplasia syndromes the characteristics of the observed adrenal tumors are quite homogenous, and the progression from a benign to a malignant tumor is usually not reported. Is CNC a new example of a genetic syndrome that would help to elucidate the molecular genetics of ACC ? In these two CNC patients, it is tempting to speculate that the development of ACC was initiated by the inactivating mutation of PRKAR1A. Indeed, loss of heterozygosity (LOH) at the 17q22-24 PRKAR1A locus is frequent in ACC (10), but so far somatic PRKAR1A mutations have only been reported in secreting adrenocortical adenomas. In the ACC from the CNC patient from the Azores investigated by Stratakis and colleagues (20), no LOH at the PRKAR1A locus could be observed. However, the S147G germline PRKAR1A mutation of this family belongs to the minority of mutations that give rise to an expressed mutant protein. These mutations are less often associated with LOH of the wild-type allele and might exert a dominant-negative effect (11). Interestingly, the expressed mutations are often associated with more severe forms of CNC, including malignant tumors (4, 12). Increased protein kinase A activity was demonstrated for the S147G-expressed PRKAR1A mutant. This mutant stimulates phosphorylation of the transcription factor cAMP response element-binding protein. However, the decrease of this transcription factor, instead of its overactivation, has been reported in ACC (13). In PPNAD, by contrast with the lack of previous report of ACC, benign macronodules (between 1 and 3.5 cm) have been described. When they are unilateral, they can even be misdiagnosed as unilateral adrenal adenomas before the diagnosis of PPNAD or CNC (14). Molecular analysis of these macronodules can demonstrate the occurrence of a somatic alteration, as β-catenin (CTNNB1) or PRKAR1A-inactivating mutations (7, 14). The multistep tumorigenesis model suggests that the occurrence of this somatic genetic defect restricted to the macronodule might participate in its development. Furthermore, β-catenin mutations are found in ACC, and β-catenin nuclear accumulation, as shown by immunohistochemistry, is associated with more aggressive tumors (15). Interestingly, in the case reported by Anselmo et al. (20), a cytoplasmic accumulation of β-catenin is found, suggesting activation of the pathway. An alternative hypothesis for the development of ACC in these two CNC patients would be the coexistence of additional germline factors predisposing to malignancies. We have previously shown that in CNC patients with PRKAR1A mutation, the development of PPNAD or testicular tumors is associated with the presence of germline variants of the phosphodiesterase PDE11A4 (16). In this situation, the effect of the mutation of the causative gene PRKAR1A is potentiated by the variant of the second “modifyer” gene. In the patient reported by Morin et al. (21), a germline variant of the tumor suppressor gene TP53 was observed. This Pro72Arg variant has been previously associated, as discussed by the authors, with various malignancies. Transgenic mice heterozygous for inactivation of both PRKAR1A and TP53 [Prkar1a(+/−) Trp53(+/−)] develop more malignancies than the single heterozygous animals (17). They also present an increased susceptibility to develop chemically induced skin tumors. Hopefully, most tumors observed in patients with PRKAR1A mutations are benign. This fits well with the concept that activation of the cAMP pathway is associated with the development of benign secreting endocrine tumors. However, it is possible that the association of PRKAR1A mutation with other genetic factors that play a role in the susceptibility to develop malignancy would stimulate cancer development. Interestingly, in the family reported by Anselmo et al. (20), several malignancies (lung adenocarcinoma, colon cancer, melanoma …) have been observed in patients that were not all carriers of the PRKAR1A mutation. It is likely that predisposing genetic factors might participate in these malignancies and could also have been present in the CNC patient that developed an ACC. Now that we know the possibility of ACC development in CNC patients, what would be the recommendation for patient management? The rarity of CNC and ACC precludes statistical demonstration of their association. Now one has to consider the association as plausible, regardless of the molecular explanation. CNC patients are routinely investigated for cortisol excess. Adrenal imaging is also done quite systematically, especially in the case of steroid excess. In an international series of 353 CNC patients, no ACC has been clearly reported, despite 212 of the patients having PPNAD (12). Since 2008, a prospective screening program of CNC patients is being done at the national level in France with systematic endocrine and imaging (computed tomography scan) investigations (www.ClinicalTrials.gov identifier: NCT00668291). To date, no ACC has been reported in the analysis of the first 68 enrolled patients (J. Bertherat, personal communication). This reassures us that ACC is quite rare in CNC, even if it could be considered as part of the syndrome. However, one can speculate that early adrenal surgery in patients with PPNAD because of Cushing's syndrome prevents the later development of the malignant tumor. For patient counseling, familial anamnesis of cancer might be important information on the basis of these two case reports. Because the two patients reported in this issue of the JCEM presented steroid excess, it is likely that the current advice of annual hormonal investigations for steroid excess would be sensitive enough to detect an ACC. When Cushing's syndrome is demonstrated, adrenal imaging is done. Abdominal/adrenal imaging can also be done systematically in CNC workup, but at present there is no clear recommendation on its frequency. In patients with a macronodule larger than 3 to 4 cm, especially with androgen excess, one could discuss the diagnosis of ACC. In these patients, adrenal expert surgeons should take into consideration the debate over the use of open surgery vs. laparoscopy for small ACC (18, 19). This is important to consider because PPNAD is usually operated by laparoscopy. Disclosure Summary: The author has nothing to declare. Adrenocortical cancer Carney complex loss of heterozygosity primary pigmented nodular adrenocortical disease.