The proband underwent surgery for an atypical carcinoid (AC) (eNET, G2) of the lung in the right lower lobe in his late 20s. Five years later, he underwent surgery again for the resection of an adrenocortical adenoma with myelolipoma in the right adrenal gland and an angiofibroma of the skin. Then, the following year, he was diagnosed with a pituitary adenoma with hyperprolactinemia and underwent surgery for the resection of a nodular goiter of the right thyroid gland and a lipoma in the right side of the neck. In his late 30s, he was found to exhibit multiple duodenal ulcers, hypersecretion of gastric juice, and hypergastrinemia. He was treated with octreotide, which significantly decreased gastrin plasma levels from 1300 to 353 pg/mL (normal range: 37–172 pg/mL). However, after several months, his plasma gastrin levels increased again despite treatment with octreotide. Several biopsies of the duodenal mucosa conducted by gastroduodenal endoscopy failed to show a gastrin-producing tumor at that time. Thereafter, his pancreatic body and tail were resected based on the suspicion of gastrinoma; however, microscopic examination revealed several microadenomas, including glucagon- and pancreatic polypeptide-producing adenomas, and the absence of gastrin-producing adenomas, in the pancreas. At approximately 40 years of age, he was found to exhibit urothelial lithiasis, hydronephrosis, hypercalcemia, and high parathyroid hormone (PTH) levels. In his early 50s, his intact PTH, prolactin, and gastrin plasma levels were 161 (normal range: 10–65), 79.4 (normal range: 3.6–12.8), and 1157 pg/mL, respectively.

He was subsequently transferred to another hospital due to difficulty of clinical treatments and was comprehensively examined and retreated; in addition, he underwent a parathyroidectomy due to hyperparathyroidism. Tumors of the upper lobe of the right lung with multiple lymph node metastases in the right pulmonary hilar, upper posterior mediastinal, and around the pancreatic head regions were observed ( Figure 1 ). Fluorodeoxyglucose positron emission tomography (FDG-PET) revealed similar lesions measuring 69 × 86 mm in his ninth rib and mediastinum. Using venous sampling following selective intra-arterial calcium injection, hotspots for gastrin were found in his duodenal mucosa. Although multiple liver tumors were confirmed in the patient by imaging, biopsies could not be performed owing to poor physical condition. However, venous sampling after calcium infusion confirmed that plasma gastrin levels were not elevated, suggesting that the liver tumors were likely not metastases of gastrin-producing tumors. Furthermore, gastrin-producing tumors were not detected by gastroduodenal biopsy. Consequently, the duodenal regions of the patient were not resected. Transcatheter arterial chemoembolization using mitomycin C and degradable starch microspheres was performed on the patient thrice in 10 months to treat multiple liver tumors. In the same year, he was diagnosed with a growth hormone- and prolactin-producing pituitary adenoma. Octreotide long-acting repeatable was administered to the patient to manage hypergastrinemia and elevated growth hormone levels, as well as to control NETs. An anti-prolactin secretion drug (cabergoline) was administered to him to treat hyperprolactinemia. Computed tomography images revealed that the liver tumor in segment S7 had shrunk 10 months following the first administration of octreotide and transcatheter arterial chemoembolization; however, the number of tumors in the liver increased. FDG-PET also revealed similar metastatic lesions in the posterior mediastinum, right lung, and right ninth rib, with elevated plasma serotonin levels (230 [normal range: 1.8–6.1] ng/mL), further confirming AC metastasis. The patient was subsequently monitored at home, but unexpectedly died in his early 50s due to an unidentified cause. An autopsy was not performed. The disease duration was 23 years.

For Patient II-2, urothelial lithiasis in the ureter and hyperparathyroidism were diagnosed in his early 30s. He underwent parathyroidectomy of two glands, the examination of which revealed chief cell hyperplasia; following this, hyperparathyroidism was controlled. The following year, he underwent surgery for the removal of a prolactinoma of the pituitary gland. He experienced ZES with duodenal ulcers and continued diarrhea in his late 30s. Although gastroduodenal endoscopy with biopsy did not reveal duodenal gastrinoma in the patient, he underwent pancreaticoduodenectomy two years later. He was cured of hypergastrinemia following pancreaticoduodenectomy. The patient developed neither pulmonary carcinoid nor tumor/tumorous lesions in the thyroid gland, adrenal gland, and soft tissues. Presently, he is alive and in his early 50s.

For this patient, multiple lipomas in his left shoulder and right thigh were resected in his late 60s. He was diagnosed with a prolactinoma of the pituitary gland and treated with cabergoline. Then, he was diagnosed with parathyroid hyperplasia and adrenocortical adenoma one year later. Two months after this, he was observed to exhibit gastrin hypersecretion and polyposis in the stomach and duodenum, for which gastroduodenal endoscopy with biopsy was subsequently performed. However, histological diagnosis revealed no gastroduodenal NET. He was diagnosed with ZES with hypergastrinemia in his early seventies. Surgeries for lesions in his pancreas and parathyroid glands were not performed because he did not consent. In the same year, he experienced acute myocardial infarction. Two years later, cytology confirmed the presence of an atypical carcinoid in the lower lobe of his right lung. Furthermore, in that year, he underwent surgery for a thoracic aortic aneurysm. He died of respiratory distress caused by the atypical carcinoid in the same year he underwent thoracic surgery. Autopsy was not performed. Table 1 summarizes the tumor locations, tumor sizes, and histological findings of the patients ( Table 1 ).

Tumors were fixed in 10%-buffered formalin and embedded in paraffin. After reviewing all sections stained with hematoxylin–eosin, suitable sections were selected for IHC. IHC for gastrin, chromogranin A (CgA), synaptophysin, CD56, and Ki67 was routinely performed for all tumors of family members with MEN1. Moreover, we used antibodies to evaluate serotonin IHC, calcitonin to characterize lung tumors, insulin, glucagon, somatostatin, and pancreatic polypeptide to characterize pancreatic tumors, and CgA to differentiate parathyroid tumors from the thyroid gland. Menin expression was evaluated in tumors of the parathyroid gland, duodenum, pancreas, and lungs by IHC, which was performed on a VENTANA BenchMark ULTRA Slide Staining System (Roche Diagnostics, Indianapolis, IN, USA) using primary antibodies against CgA (DAKO; DAK-A3), synaptophysin (Nichirei, Tokyo, Japan; clone 27G12), CD56 (Nichirei clone MRQ42), Ki67 (MIB1) (Ventana; clone 30-9), somatostatin (Nichirei, polyclonal), gastrin (Dako, polyclonal), insulin (BioGenex, San Ramon, CA, USA, polyclonal), glucagon (Dako, polyclonal), and pancreatic polypeptide (Dako. polyclonal). Appropriate positive and negative controls were included in the experiment. Only menin (Santa Cruz Biotechnology, Inc.; Clone B-9, 1:200) was stained using the Histofine SAB-PO kit (Nichirei, Tokyo, Japan). Primary antibody against menin, which is mouse monoclonal antibody against amino acids 1–300 of the human menin protein, was incubated at 4°C overnight and secondary antibody was subsequently added at 18°C. We used placental and adult testis tissues as positive controls for menin (12, 13) and phosphate-buffered saline for negative stains.

To count menin-labelled cells, two of the most highly labelled areas (hot spots) were photographed at a magnification of ×400 and the cells were counted using a digital image analyzer (Lumina Vision, Mitani Corp, Tokyo, Japan). The procedure for counting menin-labelled cells was similar to that used for counting Ki67-labelled cells with the same instrument. Menin-positive and menin-negative cells (non-labelled cells) were counted and their ratio in each tumor determined. Normal pancreatic acinar cells were evaluated as positive controls ( Table 2 ).

NETs or tumor precursor lesions (hyperplasia or minimal NETs) in the family members exhibited a similar histological pattern, and comprised cells with hyperchromatic nuclei and eosinophilic cytoplasm with a high nuclear–cytoplasmic ratio, and were arranged in trabecular, insular, glandular, or solid patterns. The tumor cells were characterized based on their IHC reactivity to cytokeratin (AE1/AE3 or CAM 5.2) and neuroendocrine markers, including CgA, synaptophysin, and CD56. Menin expression was observed in the nuclei of the human placental and testis cells, which served as positive controls ( Figures 2A, B ).

The bronchial tumor of the patient comprised small cells arranged in a trabecular pattern, occupying the bronchial lumen with lymph vessel invasion. The tumor cells were strongly positive for CgA and synaptophysin, with a Ki67 labeling index of 3.0%. They were positive for serotonin but negative for calcitonin and gastrin. The tumor was diagnosed as AC (NET G2) of the bronchus. Menin IHC revealed nuclear staining in normal bronchial and alveolar cells. Tumor cells of the bronchial NET G2 displayed a mixture of menin-positive and menin-negative cells ( Figures 2C–I ).

The adrenocortical tumor of the patient, which measured 5.0 cm × 3.5 cm, comprised adrenocortical adenoma mixed with adipose tissue and bone marrow cells (myelolipoma). Menin expression was observed in the nuclei of normal adrenal cortex cells; however, the adrenocortical adenoma was negative for menin ( Figures 2J–L ).

The nodular goiter of the thyroid gland showed strong menin IHC in follicular cells ( Figure 2M ). The parathyroid tumor exhibited a nodular growth pattern with sclerosing stroma. The tumor cells were confirmed positive for CgA for differential diagnosis from thyroid nodules. The tumor demonstrated mixed menin-positive and menin-negative cells ( Figures 2N, O ).

Tumor lesions were not observed in gross pancreatic NETs; however, we identified 2–4 micro-NETs measuring 0.9–7.0 mm in diameter on each glass slide. These micro-NETs displayed varying degrees of menin immunostaining. For example, one NET with solid growth was completely negative for menin IHC, while the glandular-type NET exhibited positive staining for menin. Another pancreatic NET with a glandular structure showed gradation of menin IHC from the periphery to the center of the tumor nodule. Cells at the periphery of the NET were positive but weakly stained as compared to pancreatic acinar cells. Higher magnification revealed that most tumor cells in the center were negative for menin ( Figures 2P–W ).

The posterior mediastinal tumor was examined only by biopsy and was diagnosed as NET G2, and was positive for CgA and serotonin but negative for gastrin ( Figures 2X, Y ). Its Ki67 labeling index was 4%.

We identified micro-NETs measuring below 1.5 mm in diameter in the duodenum of the patient. These lesions were strongly positive for gastrin. Menin IHC revealed positive staining in the micro-NETs with lower intensity as compared to that in the adjacent duodenal mucosa. A micro-NET with a diameter of 0.7 mm was found in the lower bile duct of the patient and showed strong positivity for gastrin and menin. Another NET in the duodenum exhibited glandular and insular structures comprising cells positive for somatostatin, and showed only focal positivity for gastrin and a mixture of positive and negative cells for menin ( Figures 3A–K ).

In the pancreas of the patient, one to two micro-NETs measuring 2.0–8.0 mm in diameter were identified in each paraffin section. Varied menin immunostaining was observed in the pancreatic NETs. A NET with sclerosing stroma and other micro-NETs with a diameter of 2.0 mm exhibited positive but weaker menin staining as compared to acinar cells. The other micro-NETs were almost negative for menin. Some NETs exhibited clusters of clear cells and showed foci of positive cells for glucagon or pancreatic polypeptide ( Figures 3L–T ). Gastrin was totally negative in pancreatic NETs.

As the patient refused surgery for the tumors except lipomas, most were only diagnosed clinically using imaging and laboratory data. Only bronchial brushing was performed; the cytology specimen was examined and found to be similar to that of Case 1 ( Figures 3U, V ), leading to a bronchial NET G2 diagnosis.

The ratio of menin-positive cells to menin-negative cells was found to be 90% for normal pancreatic acinar cells, 35% for duodenal NETs, 12% for parathyroid adenomas, 11% for pulmonary NETs, 9% for pancreatic NETs, and 0.01% for adrenocortical adenoma (myelolipoma).

A Japanese family comprising a father and his two sons suffered from MEN1 with ZES, with a germline alteration in exon 10—c.1714-1715 del TC of MEN1. They had full sets of MEN1 tumors and tumor-like lesions. Only one patient who underwent pancreaticoduodenectomy survived, with the other two dying of the disease. However, it is unclear whether the cause of their death was related to ZES treatment. Menin immunohistochemistry and its significance in tumorigenesis and pathologic diagnosis is of significant research interest.

We focused on the localization of gastrin-producing tumors in Cases 1 and 2. Conventional imaging techniques, such as CT and MRI, were used to localize non-functional/functional neuroendocrine neoplasms in each patient; however, no tumors were detected. MRI and CT are of limited use for the localization of gastrinomas as they are often small, multiple, and located in the duodenal submucosa (14). The patient in Case 1 experienced hypergastrinemia and duodenal ulcers for an extended period and underwent partial pancreatectomy. However, gastrin-producing cells were not detected in the multiple pancreatic NETs of the patient. In contrast, in Case 2, hypergastrinemia improved in the patient following pancreaticoduodenectomy, and histological examination revealed multiple micro-NETs in his duodenum and lower bile duct with lymph node metastases. No NETs in other locations had gastrin-positive cells. Duodenal micro-gastrinomas are the primary cause of ZES in most patients with MEN1, highlighting the importance of pancreaticoduodenectomy in treating ZES (14–16). Most gastrinomas in patients with MEN1 are located in the duodenum rather than the pancreas (14). However, most pancreatic gastrinomas are sporadic (17). The distinction between duodenal gastrinomas/somatostatinomas and hyperplastic lesions is explained by the fact that a loss of heterozygosity (LOH) of MEN1 and/or centromere 11 was demonstrated in approximately 50% of MEN1-associated duodenal NETs; however, hyperplasia or precursor lesions consistently lacked LOH on chromosome 11q13 (18). Further, gastrin microfoci in MEN1 have been described as being biologically aggressive gastrin-producing tumors (18), and it is still difficult to make a clear distinction. Moreover, duodenal NETs are typically considerably small, and endoscopic duodenal and gastric biopsy did not show submucosal tumors in Case 1, despite multiple attempts. Tumor metastases in locoregional lymph nodes are usually larger than the primary foci in the duodenum. This is called primary lymph node gastrinoma, in which only micro-gastrinomas <1.0 mm in diameter in the duodenum metastasize to the lymph node (19). Therefore, in patients with MEN1, screening for gastrinomas should focus on the proximal part of the duodenum, and sampling of other tissues, especially the pancreas, is not required (19).

In these cases, despite the fact that the patients suffered from multiple tumors and persistent ZES, the primary cause of death in Cases 1 and 3 was the presence of bronchial NETs G2. In Case 1, a bronchial NET G2 in the right lower pulmonary lobe was surgically resected when the patient was in his late 20s, without lymph node metastasis. A similar NET G2 tumor appeared in his right upper pulmonary lobe with multiple lymph node, rib, mediastinum, and liver metastases when he was in his early 50s, approximately 22 years after the first surgery. We were uncertain of whether the bronchial NET in his right upper pulmonary lobe was a metastasis of the first NET in his right lower pulmonary lobe or an independent multicentric NET due to the long time interval between the first surgery and appearance of the second NET.

The prevalence of bronchial NETs in patients with MEN1 is low (2%–8%), and multiplicity is relatively common (20). Determining whether a tumor is primary or metastatic in a ZES context can be challenging. In Case 1, the patient had multiple NETs in the bronchus, duodenum, pancreas, parathyroid gland, and pituitary gland, all of which had a similar histology with neuroendocrine features, especially those in the duodenum and pancreas. However, a metastatic NET may be differentiated from a primary one using biomarkers such as some peptide hormones produced by tumors, including serotonin, gastrin, insulin, glucagon, pancreatic polypeptide, somatostatin, parathormone, prolactin, and growth hormone, depending on the organ. In Case 1, bronchial NETs in his lower and upper lung lobes and mediastinal tumors were immunohistochemically positive for serotonin, with simultaneously elevated serotonin plasma levels. This confirmed that the mediastinal tumor cells were the same as those of the bronchial NETs in his right lower lobe. In addition, the mediastinal tumor was located in the posterior mediastinum and not in the anterior mediastinum, ruling out the possibility of it being a thymic NET. Cushing and carcinoid syndromes are the main neuroendocrine syndromes of bronchial ACs (21). Carcinoid syndrome is diagnosed based on the secretion of serotonin alongside histamine and bombesin (21). The prevalence of bronchial NETs among patients with MEN1 is 4.7–31.3%. Histopathological and radiological examinations, as well as patient outcomes, revealed a relatively indolent course and good prognosis (22). However, a cohort study carried out in France (the “Groupe d’étude des Tumeurs Endocrines” study) reported that of 1023 symptomatic patients with MEN1, bronchial NETs were identified in 51 of them (4.8%). Histological examination revealed 27 (53%) typical carcinoids (TCs), 16 (31%) ACs, 2 (4%) large cell neuroendocrine carcinomas, 3 (6%) small cell neuroendocrine carcinomas, and 3 (6%) AC-associated TCs. Moreover, nodal and distant metastases were observed in 37% and 16% of cases, respectively, in the cohort study (23). TC is a low-grade malignancy defined as NET G1. In contrast, AC is defined as NET G2, is more aggressive, and tends to metastasize to local lymph nodes, liver, bone, or brain (21). In the cases described, the proband and his father died of respiratory insufficiency caused by AC of the lungs and its metastases or multiplicity.

Menin expression was observed in the nuclei of normal human tissue cells. Loss of menin expression as determined through IHC has been reported in MEN1-related tumors in retroperitoneal NETs (24), multifocal pancreatic NETs (25), adrenocortical carcinomas (26), and thymic NETs (27). However, these previous studies did not address the usefulness of menin IHC in MEN1 diagnosis. We observed that menin IHC patterns were inconsistent between cases; thus, its application in surgical diagnosis is yet to be elucidated.

In patients with MEN1, all somatic cells harbor a germline mutation of MEN1. A loss of heterozygosity (LOH) of MEN1 and/or centromere 11 has been demonstrated in approximately 50% of MEN1-associated duodenal NETs (28). Allelic loss has been detected in tumors as small as 0.3 mm in diameter. As opposed to tumors, hyperplastic gastrin cells consistently lack LOH on chromosome 11q13 (28). In our study, IHC staining in duodenal gastrinomas varied from strong to weakly positive in micro-NETs of 0.7–8.0 mm in diameter. Two conceptualized issues exist in the pathological classification of proliferative lesions such as hyperplasias and neoplasias in endocrine organs. However, hyperplasias and neoplasias for small lesions, especially multiple nodules in an organ, are yet to be defined. Furthermore, there are no distinct morphological criteria to differentiate hyperplasias and neoplasias. Menin-negative IHC has been suggested as a possible indicator for MEN1-related tumors in previous studies (24–27, 29). Our results demonstrated that microadenomas in MEN1-related tumors exhibited a mixture of menin-positive and menin-negative cells.

To avoid making subjective judgement, we used an imaging analyzer to observe menin IHC. Specimens were compared to positive and negative controls. The menin-positive to menin-negative cell ratio was different for each case i.e., 90% in normal pancreatic acini for the positive control, 35% in duodenal NETs, 12% in parathyroid adenomas, and 0.01% in adrenocortical adenomas. We hypothesized the relationship between tumorigenesis and loss of immunoreactive menin in pancreaticoduodenal small NETs in MEN1 as follows: “At the early stage of small NET development, tumor cells exhibit features of neuroendocrine cells, as indicated by the strong positivity for CgA, gastrin, or somatostatin, without loss of menin. During tumor growth, the menin-positive to menin-negative cell ratio becomes smaller i.e., approximately 90% with low menin intensity (early stage), 90–15% (transitional stage), and below 15% (final stage). Mature NETs are occupied by menin-negative cells despite the presence of residual menin-positive cells (up to 10%) at the periphery of the nodules (final stage)” ( Figure 4 ). This hypothesis could explain the presence of varying menin IHC patterns, especially in pancreaticoduodenal small NETs in a context of ZES. This hypothesis is based on histological data and cannot be compared to the theory of the second hit of Knudson classically accepted for tumorigenesis due to tumor suppressor genes (30).

Menin IHC demonstrated positivity in the nuclei of thyroid follicular cells, with the lesion being compatible with nodular goiter but not follicular neoplasm. Lodewijk et al. (29) examined menin expression in non-tumorous lesions in patients with MEN1 and found that menin was expressed in all non-tumorous thyroid lesions. Furthermore, the prevalence of thyroid incidentalomas in patients with MEN1 was found to be the same as that in a matched reference group of patients without MEN1 (29). Thus, thyroid lesions are not considered MEN1-associated lesions. The adrenal myelolipoma demonstrated loss of menin in its adenomatous cortical component; this was not the case with the myelolipoma in Case 1. Previously reported cases of MEN1-associated adrenal myelolipoma did not exhibit LOH, suggesting that menin is not a contributing factor to myelolipoma (31).

In conclusion, pancreaticoduodenectomy is essential for ZES treatment, and bronchial NET is one of the important prognostic factors for patients with MEN1. Menin immunohistochemistry is not necessarily useful for diagnostic pathology on MEN1-related tumors especially for tumors of early stage due to the variability of immunostaining. MEN1 tumorigenesis was hypothesized based on menin IHC; however, further investigations are necessary.