|Home | About | Journals | Submit | Contact Us | Français|
Neoplasms of the peripheral nerve sheath represent essential clinical manifestations of the syndromes known as the neurofibromatoses. Although involvement of multiple organ systems, including skin, central nervous system and skeleton, may also be conspicuous, peripheral nerve neoplasia is often the most important and frequent cause of morbidity in these patients. Clinical characteristics of neurofibromatosis type 1 (NF1) and neurofibromatosis type 2 (NF2) have been extensively described and studied during the last century, and the identification of mutations in the NF1 and NF2 genes by contemporary molecular techniques have created a separate multidisciplinary field in genetic medicine. In schwannomatosis, the most recent addition to the neurofibromatosis group, peripheral nervous system involvement is the exclusive (or almost exclusive) clinical manifestation. Although the majority of cases of schwannomatosis are sporadic, approximately a third occur in families and a subset of these has recently been associated with germline mutations in the tumor suppressor gene SMARCB1/INI1. Other curious syndromes that involve the peripheral nervous system are associated with predominant endocrine manifestations, and include Carney Complex and MEN2b, secondary to inactivating mutations in the PRKAR1A gene in a subset, and activating mutations in RET respectively. In this review, we provide a concise update on the diagnostic criteria, pathology and molecular pathogenesis of these enigmatic syndromes in relation to peripheral nerve sheath neoplasia.
Neoplasms developing from cell components of the peripheral nerve sheath may arise sporadically, but are a frequent manifestation of specific inherited genetic disorders. Some of these syndromes may afflict multiple organ systems; for example, neurofibromatosis type 1(NF1) (the most frequent of these syndromes) in addition to the development of neurofibromas and malignant peripheral nerve sheath tumors (MPNST) is associated with neoplastic and developmental disorders of the central nervous and musculoskeletal systems. Carney complex is characteristically associated with additional endocrine manifestations. Other syndromes such as schwannomatosis may be almost entirely restricted to peripheral nerve.
Recent scientific developments have expanded greatly our knowledge of the molecular basis for these disorders (Figure 1). These include increased efficiencies of sequencing technologies, allowing parallel screening for multiple mutations in different genes, as well as the development of relevant animal models that recapitulate the key manifestations of these syndromes with surprising accuracy.
In this review we attempt to summarize the most updated criteria for these problematic syndromes. In addition, we provide updated concepts on their pathogenesis, pathology, and on our molecular understanding of the underlying mechanisms. We have focused this review on the neoplastic manifestations in peripheral nerve. Excellent reviews covering additional clinical manifestations of these multiorgan syndromes are available[33, 38, 55, 68, 69, 92]. Detailed coverage of current concepts on the biological properties and pathogenesis of Schwann cell neoplasms are covered in this same journal issue by Carroll S, as well as diagnostic pathology of peripheral nerve sheath tumors by Rodriguez FJ et al.[91
NF1 is the most frequent of the inherited genetic syndromes resulting in peripheral nerve sheath tumors. The incidence of this syndrome is approximately 1 in 2,500–3000 births[32, 45, 53, 87], and afflicts individuals of all races and geographic regions. NF1 results from germline mutations in the gene encoding for neurofibromin (NF1). The NF1 gene was identified by positional cloning in 1990[22, 114], and is located in chromosomal region 17q11.2. Neurofibromin is ubiquitously expressed, but the highest levels are found in the nervous system, both central and peripheral, which is prominently involved in the NF1. Neurofibromin is an important tumor suppressor, a GTPase activating protein that regulates ras; therefore, absence of neurofibromin leads to constitutive activation of ras signaling. Neurofibromin may also regulate cAMP levels, although the effects of altered cAMP levels may be more important in central nervous system manifestations of NF1 deficiency, including optic glioma formation and cognitive deficits in model systems[17, 115].
The diagnosis of NF1 is essentially a clinical one, and diagnostic criteria have been proposed and generally well accepted[1, 34, 44](Table 1). The caveat with these diagnostic criteria schemes, is that some patients present with mosaicism, or findings limited to a specific anatomical region, i.e. so called “segmental” neurofibromatosis. It is possible that an individual can fulfill criteria within a body segment and clinicians need to be aware not to apply the criteria in these situations. A variety of tissues are affected in the form of hyperplasias, hamartomas and neoplasms (Figure 2). Involvement of the skin (café-au-lait spots and axillary freckling), eye (Lisch nodules), bone (dysplasia of sphenoid bone or long bones, kyphoscoliosis), cardiovascular system (cerebral arteriopathy, pulmonary artery stenosis), and central nervous system (learning disabilities) are all important manifestations of the syndrome. However, although it is thought that NF1 penetrance is 100% with appropriate follow-up, extent of organ involvement and clinical manifestations in specific patients, even within families, are variable. Syndromes that present with similar facial and pigmentary features of neurofibromatosis include Noonan syndrome, associated with mutations in various components of the RAS/MAPK signaling pathway, and a recently described NF1-like syndrome associated with SPRED1 mutations[15, 76], although they are not known to be associated with neoplasms of peripheral nerve.
The main neoplasms that develop in the setting of NF1 predominantly involve the peripheral nerve, in particular benign neurofibromas (Figure 2b). Neurofibromas can be of different clinicopathologic types according to the WHO classification[67, 97], including cutaneous (localized or diffuse), intraneural (localized) or plexiform. Histologically, the main cell type in neurofibromas is the Schwann cell, but they also contain a variety of cellular nerve components, including fibroblasts, mast cells, axons and perineurial cells. Cellularity in conventional neurofibromas is low, and they contain a variable myxoid stroma associated with collagen fiber bundles (Figure 2c). When involving a peripheral nerve trunk, they cause a characteristic, diffuse fusiform expansion. Diffuse neurofibromas may be encountered in NF1 patients, although they are not exclusive of the syndrome (Figure 2d). They usually present as large cutaneous plaques that infiltrate dermis and subcutaneous tissues, and entrap cutaneous adnexal structures. Occasionally, they contain pseudo-meissnerian corpuscles, i.e. small structures resembling the specialized Wagner-Meissner sensory bodies(Figure 2e). Diffuse neurofibromas may also arise in deep locations and even involve internal organs.
Plexiform neurofibromas are complex neoplasms that by definition involve many peripheral nerve fascicles or components of a large nerve plexus (Figure 2f). The gross appearance of these tumors is characteristic, which may exhibit a “worm-like” mass tangle. Deep plexiform neurofibromas involving large nerves or nerve plexuses occur almost exclusively in the setting of NF1 syndrome[97, 119]. In addition to be associated with increased morbidity, plexiform neurofibromas have a propensity for malignant degeneration and represent a precursor to MPNST, an important cause of mortality in NF1 patients. However, it should be noted that on rare occasions localized intraneural neurofibromas may also undergo malignant degeneration.
MPNSTs are usually high grade spindle cell neoplasms with various degrees of Schwann cell differentiation, with a high propensity for metastasis (Figure 2g). They occur in 8–13% of patients with NF1. Divergent differentiation (i.e. mesenchymal or epithelial “heterologous” components) may be present, in particular in the setting of NF1, including skeletal muscle, smooth muscle, cartilaginous, osseous, angiosarcomatous or glandular differentiation[27, 89, 117, 118] (Figure 2h,i). Another rare neurofibroma variant that is strictly limited to the NF1 syndrome is the massive soft tissue neurofibroma (Figure 2j,k,l), which can reach an enormous size and lead to distortion of the soft tissues or even a whole extremity. Histologically, it infiltrates diffusely into soft tissue, including skeletal muscle and nerve fascicles, and may contain regions packed with cells containing high nuclear cytoplasmic ratios (Figure 2l), as well as pseudo-meissnerian corpuscles. In spite of its increased and alarming cellularity at the histological level, and negative cosmetic effects, the massive soft tissue neurofibroma usually follows a benign course. However, a plexiform component is present in some instances, and therefore a low potential for malignant degeneration exists.
Internal organs may be involved in NF1, predominantly the gastrointestinal tract, in the form of neurofibromas, ganglioneuromatosis, or gastrointestinal stromal tumors. Glomus tumors have recently been added to the NF1 spectrum, although they rarely involve peripheral nerve.
Numerous advances have been made in recent years regarding the biology of nerve sheath tumors in NF1. Neurofibromas exhibit genetic alterations in S100 positive Schwann cells, which demonstrate loss of heterozygosity in the NF1 gene. However, early observations have demonstrated a variety of cellular components in neurofibromas in addition to Schwann cells, including perineurial cells, fibroblasts as well as hematopoietic components such as mast cells. Mouse knockout experiments have highlighted the importance of the tumor microenvironment in the development of neurofibromas, since haploinsufficient microenvironment cell components are required for neurofibroma development. Recently, NF1 heterozygous multipotent precursor cells, presumably hair follicle-derived, have been isolated from neurofibromas, and may aid in the development of cutaneous tumors.
Timing and precise location of Nf1 inactivation in neurofibroma has been tested in various mouse models, with Nf1 loss in the embryonic period, perinatal period or even adulthood resulting in neurofibromas. Some investigators have proposed nonmyelinating Schwann cell progenitors (p75+) as the cell of origin of plexiform neurofibroma. However, this finding is controversial since Nf1 loss in Schwann cell precursors expressing desert hedgehog also leads to plexiform neurofibromas. In addition, a cutaneous neural stem cell/progenitor, which is neural crest derived but outside of the Schwann cell lineage, may give rise to cutaneous neurofibromas. Malignant peripheral nerve sheath tumors, in addition to NF1 loss, develop alterations in additional tumor suppressors and amplification/overexpression in receptor tyrosine kinases such as the epidermal growth factor receptor[65, 83, 124]. Detailed updates in the biology of these tumors is covered in recent reviews, including a comprehensive one by Carroll S. in the current issue.
NF2 occurs at a lesser frequency than NF1, with an estimated incidence of 1 in 33–40,000 births. NF2 has also been referred historically as “central” neurofibromatosis, a currently discouraged term, which emphasized the predilection of NF2 associated neoplasms for the central nervous system and craniospinal axis, compared with NF1 (Figure 3a,b). The NF2 gene was identified in 1993[93, 111]. The protein product of NF2 is also known as Merlin or schwannomin, another tumor suppressor with numerous cellular functions, particularly associating with cell junction complexes.
The clinical manifestations of NF2 in many instances may be associated with increased morbidity compared with those of NF1, and include a variety of neoplasms (schwannoma, meningioma, ependymoma), as well as cutaneous (Figure 3c) and eye manifestations (posterior subcapsular cataracts, retinal hamartomas, epiretinal membranes). Peripheral nerve manifestations include a polyneuropathy in addition to schwannomas. The development of bilateral vestibular schwannomas is a hallmark of NF2, and a useful diagnostic criterion (Table 2). However, NF2 mosaicism has been extensively reported, and the presence of unilateral vestibular schwannoma in the presence of other NF2 manifestations should raise this possibility. The presence of NF2 mosaicism is particularly an issue in patients without family history (i.e. de novo), where it occurs in 20–30% of patients[61, 77]. Clinical manifestations may also vary in patients with different mutations in the NF2 gene. For example, several studies have demonstrated a more severe NF2 phenotype (earlier onset of symptoms, higher tumor burden, earlier death) in patients with nonsense or frameshift mutations resulting in abnormal protein expression, compared with mutations/large deletions resulting in complete Merlin loss[75, 99], or Merlin retention [75, 99].
Abnormalities in skin pigmentation (i.e. café au lait spots), and development of cutaneous neoplasms occurs to a lesser extent in NF2 than in NF1. A characteristic lesion occurring in NF2 with hamartoma-like features that demonstrates involvement of the brain cortex proper is meningioangiomatosis. This uncommon, epilepsy-associated lesion is characterized by downgrowth of meningothelial or meningothelial-like cells in association with leptomeningeal and superficial cortical vessels. Occasionally they coexist with meningiomas. The clinical manifestations of NF2 have been embodied in contemporary diagnostic criteria [1, 34, 44, 78], which have emphasized clear, objective separation of NF2 from NF1[1, 2, 8, 28, 44](Table 2). Recent interest to incorporate genetic information to clinical classification schemes have been developed in Baser’s criteria (Table 3)[1, 2, 8, 28, 44].
The development of multiple schwannomas is one of the hallmarks of the NF2 syndrome (Figure 3d). As in the sporadic setting, these benign neoplasms are characterized by a compact proliferation of neoplastic Schwann cells (Figure 3e). Unlike neurofibroma, non-neoplastic nerve sheath components and infiltration of peripheral nerve are lacking. Histologically, schwannomas demonstrate a well formed capsule, as well as alternating Antoni A (cellular, collagen rich) and Antoni B (loose, mucopolysaccharide rich) areas (Figure 3f)). Verocay bodies, distinctive palisades in Antoni A areas, are variable in frequency but one of the diagnostic hallmarks (Figure 3g). Additional histologic features of schwannomas include degenerative vascular changes with hyalinization and perivascular hemosiderin. Immunohistochemical stains demonstrate strong expression of S100 and collagen IV (pericellular), while EMA is limited to the peripheral perineurium and neurofilament protein highlights peripheral axons. Plexiform schwannoma is a rare variant with a predilection for cutaneous sites, and characterized by a multinodular pure Schwann cell neoplastic growth. Multiple plexiform schwannomas may be identified in NF2 patients[85, 112], although they are not specific to the syndrome (in contrast to large or multiple plexiform neurofibromas in NF1).
In general, syndrome-associated schwannomas are similar to sporadic examples. Although no formal large pathologic studies are available, some features described at increased frequency in NF2-associated tumors include the formation of whorls (Figure 3h) and multifocal involvement within a single nerve, as well as juxtaposition with meningiomas (Figure 3i). In addition, patchy loss of INI1 immunohistochemical staining (“mosaic” pattern) has been reported in the majority of syndrome associated schwannomas, including both NF2 and schwannomatosis .
Although malignant degeneration in schwannomas, including the NF2 setting, is an extremely rare phenomenon, this has been reported in NF2 patients, in particular after radiation therapy. Also, it should be noted that cutaneous neurofibromas may also occur in NF2. Although, the presence of neurofibromas has been used in the past as a clue to distinguish NF2 from schwannomatosis, neurofibromas have been reported in two patients satisfying clinical criteria for schwannomatosis.
At the molecular level, the neoplastic manifestations of NF2 are explained by initial Merlin loss. A critical function of Merlin is its participation in contact-dependent inhibition, whereby cell proliferation is suppressed with increased cell density and formation of intercellular adhesions. However, Merlin associates with many different proteins in diverse cells types, and additional functions include regulation of cell surface and intracellular signaling cascades[9, 60], as well as actin-cytoskeleton interactions.
Schwannomatosis is the most recent addition to the neurofibromatosis group. The incidence of schwannomatosis in the general population may be similar to NF2, but is probably less well recognized. Schwannomatosis, as NF2, is characterized by the development of multiple schwannomas (Figure 4a,b). Unlike NF2, schwannomas in schwannomatosis spare the vestibular nerves and the majority of the patients lack a family history. Ocular involvement has not been described and the typical NF2 intracutaneous plaque lesions  also do not appear to be part of the schwannomatosis phenotype. The main clinical manifestation and indication for surgery is pain, which in these patients may be debilitating.
Although our understanding of this enigmatic syndrome continues to evolve, practical diagnostic criteria have been developed. Age is a critical factor in the diagnosis of schwannomatosis, and a cutoff of age 30 to document absence of vestibular schwannomas (and therefore NF2) has been proposed (table 4). Some investigators have advocated the use of high resolution MRI studies with special protocols focusing on the internal auditory canals, to exclude small vestibular schwannomas which are diagnostic of NF2 . Other investigators advocate molecular testing, particularly young patients, to exclude NF2 including mosaics.
Schwannomas presenting in the setting of schwannomatosis share many of the histologic features with those occurring in NF2 or sporadically. These features include alternating cellular and loose areas (Antoni A and Antoni B respectively), Verocay bodies, hyalinized vessels and the presence of a capsule. However, features that appear to be overrepresented in schwannomatosis associated tumors include myxoid changes, intraneural growth, and peritumoral nerve edema(Figure 4c).
An important hallmark in our understanding of schwannomatosis was the identification of an inactivating germline mutation in the SMARCB1/INI1 gene in a family with the syndrome . This was a somewhat surprising finding, given that germline mutations in this key tumor suppressor gene have been described in families predisposed to high grade malignant tumors, in particular atypical teratoid rhabdoid tumors. More recently, different somatic (but not germline) NF2 mutations in separate schwannomas, in addition to a SMARCB1/INI1 germline mutation, were described in a family suggesting a “four-hit” mechanism for schwannoma tumorigenesis in some schwannomatosis patients. Indeed, this inactivation of both NF2 and SMARCB1/INI1 appears to be the typical mechanism in schwannomas from SMARCB1/INI1 germline mutated individuals. Similar findings were reported in a multiple meningioma family, although SMARCB1/INI1 mutations seem very rare in multiple meningioma patients. Immunohistochemical testing for INI1 is feasible in pathologic specimens. Curiously, a “mosaic” pattern of immunostaining, where a subset of tumor cells are positive while others are negative, has been described in syndrome associated schwannomas, including both NF2 and schwannomatosis(Figure 4d). This contrasts with INI1 loss in atypical teratoid rhabdoid tumor, where the loss occurs in most, if not all tumor cells. This may correlate with the different genotypes in schwannomatosis patients which are often missense mutations indicating the possibility of hypomorphic mutations.
Although patients with schwannomatosis were felt to be predisposed to multiple schwannomas only, a recent schwannomatosis family with a SMARCB1 mutation and multiple meningiomas has been recently described, adding multiple meningiomas to the schwannomatosis spectrum. Similarly, cutaneous neurofibromas were thought not to be part of the syndrome, but the recent report of two schwannomatosis patients with superficial neurofibromas suggests caution in the interpretation of the schwannomatosis patient with isolated cutaneous neurofibromas.
Carney complex (CNC) is a rare multiple neoplasia syndrome first described by Dr. J. Aidan Carney in 1985 as ‘the complex of myxomas, spotty pigmentation, and endocrine overeactivity’ and inherited in an autosomal dominant manner; Dr. Carney recognized that patients with syndromes that had been previously described as LAMB (lentigines, atrial myxoma, myxoid neurofibroma and ephelide) or NAME (nevi, atrial myxoma, blue nevi) had the same condition [6, 86]. To this date, over 500 patients of diverse ethnicity from all continents have been registered by the NIH-Mayo clinic (USA) and the Cochin Hospital (France); 43% are males and 57% are females; approximately 70% are familial cases .
The diagnosis of CNC is made if two or more major manifestations of the syndrome are present (Table 5) [12, 71, 95, 107]. These must be confirmed by histology, biochemical testing, and imaging. However, one can also make the diagnosis when only one of the major criteria is present if the patient is a carrier of a known inactivating mutation of PRKAR1A . Additionally, a considerable number of clinical and biochemical manifestations listed in Table 4 are suggestive, but not diagnostic, of CNC.
Pigmented skin lesions are reported in the majority of CNC patients (over 80%). They constitute one of the three major criteria of CNC and are diagnostically very important because, easily recognizable and occurring early in life, they may lead to early detection of a potentially life threatening disease. The most common skin lesions are lentigines (they are present in 70–75% of cases). Morphologically, lentigines are flat, poorly circumcised, brown-to-black macules usually measuring less than 0.5 cm in diameter. Histologically, lentigines show hyperpigmentation of the basal cell layer associated to melanocytic hyperplasia and hypertrophy. In contrast, the pigmentation in common freckles is the result of increased melanin production, usually without melanocytic hyperplasia  . Even though lentigines may be the first sign of CNC at birth, they usually do not acquire their typical intensity and characteristic distribution (lips, conjunctiva and inner and outer canthi; vaginal and penile mucosa) until the late prepubertal and early peripubertal period [10, 71]. The second most frequent skin manifestations in CNC patients are blue nevi, and in particular, epithelioid blue nevi[10, 51], followed by cutaneous myxomas [10, 71]. Early identification of cutaneous myxomas is essential since it is estimated that almost 80% of CNC patients with a lifethreatening cardiac myxoma had presented earlier in life with cutaneous myxoma [10, 71, 107].
Other CNC-related skin manifestations include: café-au-lait spots (typically less pigmented than those seen in Mc-Cune Albright syndrome) or depigmented lesions that can also be present at birth or develop during childhood; melanocytic and atypical nevi, and the so called Spitz nevus.
Cardiac myxomas, which can appear as early as in infancy, are responsible for over 50% of the disease-specific mortality of CNC patients. Early detection and regular screening for cardiac myxomas by echocardiography is essential, as these tumors can lead to sudden death by embolism, strokes, or cardiac failure. Cardiac myxomas are the most common, clinically significant non-cutaneous lesions in CNC patients.
Pigmented nodular adrenocortical disease (PPNAD) is the most frequent endocrine manifestation in CNC patients. Adrenocorticotropic hormone-independent Cushing syndrome (CS), is present in 25–30% of CNC patients . PPNAD is named after the macroscopic appearance of the adrenal glands that is characterized by small, cortisol-producing, pigmented micro-nodules (less than 1 cm in diameter) of the adrenal cortex. Diagnosis of CS due to PPNAD is often difficult because hypercortisolism can develop progressively over years and may be periodic: cyclical and atypical CS is more often the rule, rather than the exception among patients with CNC [104, 107]. Pathological investigation reveals that adrenal glands from patients with PPNAD are usually normal in size and weight; it is for this reason, that one out of three patients has essentially normal-appearing adrenal glands on computed tomography (CT-scan). The remaining patients may have visible round micronodules (smaller than 1 cm in diameter) or, rarely, macronodules (larger than 1 cm) within the background of hyperplasia.
Psammomatous melanocytic schwannoma (PMS) was present in 8% of CNC patients studied by Bertherat and al, and four patients died of metastatic PMS . PMS may occur anywhere in the peripheral nervous system, but it is most frequently found in the gastrointestinal tract (esophagus and stomach) and the paraspinal sympathetic chain. Schwannomas in CNC are characterized by their heavy pigmentation (melanin), frequent calcifications, and multicentricity (Figure 5). PMS is unlike any of the schwannomas seen in the other conditions reviewed in this article [18, 19, 107].
In contrast to conventional schwannomas, melanotic schwannomas usually lack Verocay bodies, microcysts, a well formed capsule and thick-walled hyalinized vessels. Psammoma bodies may be focal and only identifiable after extensive search. However, their recognition is important since approximately half of patients with PMS have CNC, and the syndrome association is even higher when multiple tumors are present. In addition these tumors frequently contain large cytoplasmic vacuoles simulating adipose tissue. Another important distinction with conventional schwannoma is that a subset of melanotic schwannomas (up to 15% reported cases) are associated with malignant clinical behavior [25, 36, 54, 94]. Strict criteria of malignancy in melanotic schwannoma are not well developed, although a combination of worrisome histologic features (large vesicular nuclei with macronucleoli, brisk mitotic activity and necrosis) raises concern of aggressive behavior.
The differential diagnosis of melanotic schwannoma is largely restricted to melanocytic tumors ranging from melanocytoma to primary and metastatic melanoma. Melanocytic immunohistochemical markers (e.g. S100, melanA, HMB45) are not useful in this distinction, since they are frequently expressed in both tumor categories. The identification of pericellular basement membrane by histochemical (reticulin) and immunohistochemical (Collagen IV, laminin) stains or by electron microscopy is more specific for melanotic schwannoma (Figures 5 and and6),6), although this important diagnostic feature is not as accentuated as in conventional schwannomas.
Additional tumors and tumor-like manifestations of CNC include growth hormone secreting pituitary hyperplasia and adenomas, large-cell calcifying Sertoli cell tumors (LCCSCT), thyroid gland disease, ovarian cysts [80, 106] and osteochondromyxoma.
Genetic linkage analysis identified two independent loci for CNC, CNC1 located on chromosome 17p22–24 and CNC2 located on chromosome 2p16 [58, 102]. The genetic defect responsible for CNC at locus 2p16 remains unknown. In most cases, CNC is caused by inactivating mutations in the regulatory subunit type 1 alpha gene (PRKAR1A) located at 17q22–24 which encodes the most widely expressed of the protein kinase A (PKA) regulatory subunits . Thus, PRKAR1A is a key component of the cAMP signaling pathway. PRKAR1A’s genomic region is approximately 21 kblong and the open reading frame contains 11 exons that code for a protein that totals 384 amino acids. The peptide consists of a dimerisation/docking domain and two cAMP binding domains (A and B) that are essential for its function within the PKA tetramer (see below).
To date, over 100 disease-causing pathogenic sequence variants, spread all over the coding length of the gene, have been identified. These coding variants show no preference for a particular genomic region or exon. A recent extensive update by Horvath and al., reviewed all the known PRKR1A mutations. The majority of CNC-causing PRKAR1A mutations are base substitutions, small deletions and insertions or rearrangements . Rarely, large PRKAR1A deletions can occur . The phenotype of these depends on their intronic versus exonic sequences that are involved and whether additional genes are involved. Over 70% of CNC patients with a classical phenotype show a PRKAR1A mutation which leads to a premature stop codon and subsequently non-sense mediated mRNA decay (NMD) [10, 107] and, thus, PRKAR1A haploinsufficiency. The mutations that escape NMD are significantly less frequent and lead to the expression of an abnormal and defective PRKAR1A protein [42, 50, 74, 113]. The in vitro effect of a certain number of these expressed mutations has been evaluated and confirmed their pathogenic potential [39, 74].
Until recently, no genotype-phenotype correlations had been found for the stop codon mutations that lead to PRKAR1A haploinsufficiency. This is because most of the PRKAR1A mutations in CNC patients are family or patient specific and only three mutations have been identified in more than three kindreds: c.82t, c.491–492delTG in exon 5 and c.709-7del6 in intron 7 [42, 107]. Recently, Bertherat and al. reported a genotype-phenotype study in 353 CNC patients of which 73% were positive for a PRKAR1A mutation . PRKAR1A mutations were observed in 80% of the familial cases compared to 37% of the sporadic CNC patients. The overall penetrance of CNC in PRKAR1A mutated patients was 97.5%. Patients with PRKAR1A mutations more frequently had myxomas, skin lesions, thyroid and gonadal tumors and these clinical manifestations appeared at an earlier age compared to CNC patients that did not have PRKAR1A mutations or deletions. Acromegaly, cardiac myxoma, lentigines and PMS were more often associated with exonic mutations. An ongoing study, by the same group, is evaluating the clinical implications of the co-occurrence of phosphodiesterase 11A (PDE11A) mutations in CNC patients with PRKAR1A mutations. The PDE11A locus was identified by a genome-wide SNP genotyping study in individuals with adrenocortical hyperplasia leading to Cushing’s syndrome that was not caused by known genetic defects . More CNC-causing genes remain to be identified as over 60% and 25% of respectively sporadic and familial cases are negative for PRKAR1A mutations.
How do PRKAR1A defects cause CNC and its tumors? In its inactive form, cyclic AMP (cAMP)-dependent protein kinase (or PKA) is a holotetramer composed of two regulatory (R) and two catalytic (C) homodimers. The tetramer is dissociated into two regulatory and two free enzymatically active catalytic subunits when intra-cellular cAMP levels increase and two molecules of cAMP bind to each of the regulatory subunits [57, 110]. In turn, free catalytic subunits phosphorylate a variety of cellular target substrate proteins, that regulate a wide range of cellular processes: transcription, metabolism, cell cycle progression and apoptosis [98, 101]. So far, four different regulatory subunits (PRKAR1A, PRKAR1B, PRKAR2A, and PRKAR2B) and four catalytic subunits (PRKACA, PRKACB, PRKACG and PRKX) have been identified. It is the PRKAR1A gene coding for the PKA type I-A regulatory subunit (R-Iα) that is mutated in CNC and PRKAR1A is the only PKA subunit in which mutations have been found to lead to human disease. PRKAR1A haploinsufficiency leads to excess cellular cAMP signaling in affected tissues [58, 88]. Loss of R-Iα leads to an increase in total (but not PKA specific) cAMP-stimulated kinase activity [21, 58].
The role of R1α has been explored in several different cancer tissues and cell lines, including colorectal, breast, renal, and ovarian cancer [14, 35, 47, 66, 73, 79, 116]. The CNC data suggest that PRKAR1A is a tumor suppressor gene, as tumors from CNC patients frequently carry both the germline mutation and LOH of the 17q22–24 PRKAR1A locus. However, a few CNC tumors did not show PRKAR1A LOH [21, 58]. Mouse models have shed some light into the role of PRKAR1A haploinsufficiency in tumor formation in cAMP responsive tissues. Prkar1a knock-out (KO) animals die in utero due to failed cardiac morphogenesis . Heterozygous KO (HetKO) mice develop tumors in cAMP-responsive tissues (non pigmented schwannomas, bone lesions and thyroid neoplasias) but they lack some characteristic CNC lesions, like cardiac and skin myxomas, and pituitary adenomas. However, compared to human CNC tumors, cAMP signaling in HetKO mouse cells is only modestly increased, which may explain the lower susceptibility of the mouse pituitary, heart and skin in tumor formation . A different mouse model, which led to significantly higher PRKAR1A downregulation, showed higher cAMP signaling and an overall more severe phenotype, including endocrine manifestations such as hypercorticosteronemia and thyroid cancer, and an overall shorter lifespan [40, 41]. Tissue-specific, pituitary and cardiac, complete Prkar1a-KO mice, did develop pituitary adenomas and cardiac lesions that resembled human myxomas, respectively [122, 123].
A recent mouse model studied the effect of PRKAR1A haploinsufficiency on tumor formation in the background of other known tumor suppressor gene defects and chemically induced skin papillomas. Interestingly, Prkar1a+/−Trp53+/− and Prkar1a+/−Rb1+/− double heterozygote mice developed more sarcomas and grew more and larger pituitary and thyroid tumors compared to the single HetKO Trp53+/− and Rb1+/− heterozygous mice, respectively. Similarly, HetKO Prkar1a+/− mice developed more papillomas than wild-type animals after chemical induction . Thus, Prkar1a haploinsufficiency augmented the previously described phenotype for the respective mouse models without causing any new tumors in other cAMP-responsive or other, tissues. In the same study, Wnt signaling and cell cycle abnormalities were identified by whole-genome transcriptome profiling as the main pathways activated by abnormal cAMP signaling confirming recent data from human studies which identified somatic beta catenin (CTNNB1) mutations in PRKAR1A-haploinsufficient tumors [37, 108]. These studies showed that PRKAR1A haploinsufficiency maybe an overall relatively weak tumorigenic signal, unless it is associated with other tumor suppressor gene defects, Wnt signaling activation, and cell cycle dysregulation, mainly an increase in the expression of cyclin D1 and the transcription factor E2F1 .
The peripheral nervous system may also be involved in the form of pseudoneoplastic, hamartomatous growths in inherited genetic syndromes. Prominent among these, is the syndrome of multiple endocrine neoplasia 2b (MEN 2b). In MEN 2b the primary manifestations include the development of endocrine neoplasms affecting most importantly the thyroid (medullary carcinoma), as well as parathyroid (hyperplasia) and adrenal gland (pheochromocytoma). Two important conditions affecting the peripheral nervous system in this syndrome are mucosal neuromas and intestinal ganglioneuromatosis, resulting from hypertrophy of autonomic nervous system components (i.e. nerve plexuses and ganglia).
Mucosal neuromas are important from the diagnostic standpoint, since they may be subjected to biopsy, and their correct identification may lead to life saving prevention of more ominous manifestations of the syndrome such as medullary thyroid carcinoma. Histologically, mucosal neuromas represent tortuous, abnormal enlargements of nerves at submucosal sites (Figure 7). It should be noted that multiple neuromas were the subject of a curious report in a patient with a germline PTEN mutation, therefore in rare occasions they may arise independent of MEN 2b. Gastrointestinal ganglioneuromatosis consists of a similar enlargement and tortuosity of submucosal and myenteric plexuses in the intestine. Although, gastrointestinal ganglioneuromatosis may be present in other syndromes such as Cowden, Juvenile polyposis, and NF1, extensive and diffuse involvement of all intestinal layers is relatively specific to MEN 2b.
The main genetic change in MEN 2b is an activating mutation in the RET protooncogene. RET was identified originally as a chimeric oncogene using transfections assays with neoplastic DNA, therefore its name “rearranged during transfections”. The RET gene, located in chromosomal region 10q11.2, encodes a membrane bound receptor tyrosine kinase. The most frequent mutation in MEN 2b is caused by a substitution of the methionine at residue 918 by a threonine in the kinase domain. This leads to increased downstream signaling by several molecular mechanisms, including ligand-independent activation, increased ATP binding affinity, and loss of normal autoinhibition.
The peripheral nerve sheath is an important site for neoplastic development in a variety of inherited or sporadic tumor syndromes. The spectrum of neoplasms and pseudoneoplastic growths is wide, encompassing both benign and malignant tumors, and consequently leading to marked morbidity, and even mortality in patients afflicted by them. Recent advances have started to expand our genetic and molecular understanding, and hopefully will lead to most needed novel, directed therapies for these devastating disorders.
the authors thank Drs Jaishri Blakeley, J. Aidan Carney, Caterina Giannini and Arie Perry who contributed pictures. They also thank Sharon Blackburn at the Johns Hopkins Pathology Photographic Arts Laboratory for graphical technical assistance.