Article Text

Download PDFPDF
Neuroendocrine cells associated with neuroendocrine carcinoma of the breast: nature and significance
  1. Tomonori Kawasaki1,
  2. Kunio Mochizuki1,
  3. Hideko Yamauchi2,
  4. Shingo Inoue3,
  5. Tetsuo Kondo1,
  6. Naoki Oishi1,
  7. Tadao Nakazawa1,
  8. Tetsu Yamane1,
  9. Yoshihito Koshimizu1,
  10. Hiroko Tsunoda2,
  11. Hiroshi Yagata2,
  12. Masayuki Inoue3,
  13. Ayako Inoue3,
  14. Takanori Maruyama3,
  15. Hideki Fujii3,
  16. Ryohei Katoh1
  1. 1Department of Pathology, University of Yamanashi, Yamanashi, Japan
  2. 2Breast Center, St Luke's International Hospital, Tokyo, Japan
  3. 3First Department of Surgery, University of Yamanashi, Yamanashi, Japan
  1. Correspondence to Dr Tomonori Kawasaki, Department of Pathology, University of Yamanashi, 1110 Shimokato, Chuo, Yamanashi 409-3898, Japan; tomonori{at}


Background The developmental mechanisms of breast neuroendocrine carcinoma (B-NEC) have not been sufficiently analysed and are not well understood.

Aims To investigate NE cells in the background tissues surrounding B-NECs.

Methods Three cases (four breasts) having many NE cells in the background tissues of multifocal B-NECs were identified at the University of Yamanashi Hospital and St Luke's International Hospital, Japan. These patients were, respectively, 28-, 31- and 38-year-old women with no familial history of NE tumour. The totally-resected breasts were serially studied by immunohistochemistry for specific NE markers (chromogranin A/synaptophysin) and the morphologies and/or localisation of NE cells were investigated.

Results Immunohistochemical examination showed extensively-distributed NE cells in the background mammary ducts/lobules of the NECs in all breasts. These NE cells were classifiable into three emerging patterns: isolated/scattered, clustered and circumferential. Their distributions were intermingled and were not clearly related to B-NEC foci. NE cells were morphologically polygonal, oval or columnar with sometimes eosinophilic and/or fine-granular cytoplasm and round-to-ovoid nuclei lacking atypia. Some cells were located between epithelial and myoepithelial cells. Apical snouts were occasionally observed in NE cells forming luminal structures.

Conclusions Benign-appearing NE cells in the parenchyma of a breast with NEC could be regarded as hyperplastic from their emerging patterns and distribution; this NE cell hyperplasia may be associated with the histogenesis of B-NEC as a precancerous condition. These observations might raise questions about the treatment for B-NEC.

  • Breast
  • neuroendocrine cells
  • neuroendocrine cell hyperplasia
  • neuroendocrine carcinoma (NEC)
  • neuroendocrine ductal carcinoma in situ (NE-DCIS)
  • breast cancer
  • neuroendocrine tumours
  • breast pathology

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The WHO defines neuroendocrine tumours (NETs) of the breast as primary neuroendocrine carcinomas (NECs) exhibiting morphological features similar to NETs of both the gastrointestinal tract and the lung, and in which greater than 50% of the cell population expresses NE markers (chromogranin A and synaptophysin).1 ,2 Breast NEC (B-NEC) is classified as a special tumour entity representing only about 2–5% of mammary cancers, with most patients in the sixth or seventh decade of life.1–6 However, the natural history of B-NEC and the mechanisms underlying its development have yet to be sufficiently investigated and established.

Neuroendocrine (NE) cells are distributed in non-endocrine organs such as the gastrointestinal tract and lung (bronchus) as well as in endocrine organs.7–9 Hyperplastic conditions of NE cells have been described in several organs, such as enterochromaffin-like (ECL) cell hyperplasia of the stomach,10–12 diffuse idiopathic pulmonary NE cell hyperplasia of the lung13–15 and C-cell hyperplasia of the thyroid.16–18 NE cell hyperplasia can be associated with the NET development. On the other hand, whether normal NE cells exist in the breast is still controversial3; some investigators observed a few NE cells in normal breast tissue and/or surrounding B-NECs,19–21 while others failed to detect NE cells in mammary tissues.22–24 In addition, we could find no previous reports on hyperplastic NE cell proliferations in the breast, although the pre-invasive form of B-NECs has been established as NE ductal carcinoma in situ (NE-DCIS).25–29 Accordingly, it has been suggested that B-NECs do not originate from pre-existing and/or hyperplastic NE cells, but rather result from a divergent differentiation event in breast cancer.3 This is the first report to describe extensively distributed NE cells, showing various emerging patterns (isolated/scattered, clustered and circumferential), in the background tissues of B-NECs and to discuss their histogenesis and nature as potential precancerous cells.

Materials and methods

Table 1 presents a clinical summary of three cases with B-NEC accompanied by background NE cells. The patients were, respectively, 28-, 31- and 38-year-old, premenopausal, Japanese women with the clinical symptom of bilateral or unilateral bloody nipple discharge. They had no medical or familial history of breast disease. Blood test results were unremarkable. After preoperatively diagnosing cancer through ultrasound-guided biopsy or cytology, all three patients gave informed consent and received a bilateral (one case) or unilateral (two cases) total mastectomy at the University of Yamanashi Hospital or at St Luke's International Hospital, Japan.

Table 1

Summary of three patients having breast neuroendocrine carcinomas (B-NECs) with NE cells in the background tissue

In all cases, wholly-excised breast tissues were fixed in 10% buffered formalin and embedded in paraffin. To examine the specimens histologically, 3 μm sections were cut from paraffin blocks, then H&E staining and immunohistochemistry were performed. At least two pathological specialists (TK and RK) assessed the histological and immunohistochemical findings. We diagnosed B-NECs, including its pre-invasive counterpart (ie, NE-DCIS), using the pathological criteria and classifications defined by the WHO and the 2001 study by Sapino et al. 1 ,2 That is to say, we classified breast cancers in which >50% of the cell population expressed reliable NE markers (chromogranin A and/or synaptophysin) as NEC.1–4 26–31

For the immunohistochemical analyses, we used the dextran polymer method (Dako, Glostrup, Denmark) with primary antibodies against chromogranin A (rabbit polyclonal, 1:1000; Dako) and synaptophysin (rabbit polyclonal, 1:50; Dako). For antigen retrieval, we applied heat treatments using a water bath (97°C, 45 min) before the primary antibody reactions. Appropriate positive and negative controls were included.

Thirty-two mastectomy specimens harbouring a mammary non-NE carcinoma were also examined in the same fashion. The histological type of these cancers was DCIS in 10 cases, lobular carcinoma in situ in one, invasive ductal carcinoma in 19 and invasive lobular carcinoma in two.


As shown in table 1, our histological examination for multiple cut sections from the breast tissues taken by total mastectomy showed that case 1 had two foci of NE-DCIS in the right breast and multiple foci of NE-DCIS (figure 1) with partial microinvasion in the left breast, case 2 had two foci of NE-DCIS in the right breast, and case 3 had two foci of NE-DCIS with a 2.5 mm invasive lesion in the right breast. These tumours showed characteristic NE features such as solid growth of cancer cells frequently having a well developed vascular network, fine-granular, eosinophilic cytoplasm and fine and/or granular nuclear chromatin with diffuse immunopositivities for chromogranin A and synaptophysin (figure 1).

Figure 1

Breast cancer in case 1 showing a neuroendocrine ductal carcinoma in situ. (A) Intraductal solid proliferation of cancer cells with capillary stroma (H&E, magnification ×200). (B,C) Immunohistochemically, cancer cells were diffusely positive for chromogranin A (B) and synaptophysin (C) (magnification ×200).

Interestingly, non-neoplastic-looking NE cells demonstrating chromogranin A and/or synaptophysin were widely identified in background mammary ducts and lobules in these three cases. Careful observation using serial sections stained by immunohistochemistry and H&E revealed that these NE cells were polygonal, ovoid or columnar with occasionally eosinophilic and/or fine-granular cytoplasm and round to ovoid nuclei without atypism (figure 2). Some NE cells were located between the basal myoepithelial and the luminal epithelial cells, and apical snouts were sometimes recognised at the surface of NE cells forming luminal structures (figures 2 and 3).

Figure 2

Neuroendocrine cells in the background breast tissue of a neuroendocrine carcinoma. Immunostaining for synaptophysin (A) and H&E (B) (magnification ×400). Serial sections demonstrate that synaptophysin-positive cells (A) in a mammary lobule correspond to the polygonal-, ovoid- or columnar-shaped cells (arrows) having somewhat eosinophilic and granular cytoplasm and round-to-oval nuclei without atypia when stained with H&E (B).

Figure 3

Emerging patterns of neuroendocrine cells in the background breast tissue of a neuroendocrine carcinoma. (A–C) Isolated/scattered pattern. (D–F) Clustered pattern. (G–I) Circumferential pattern. H&E (A, D and G) and immunostaining for chromogranin A (B, E and H) or synaptophysin (C, F and I) (magnification ×400).

In terms of an emerging pattern, these NE cells could be classified into three types: isolated/scattered, clustered and circumferential patterns (figure 3). Histological distributions of each emerging pattern of NE cells accompanied by NE-DCIS and microinvasive NEC were mapped on the cut surfaces of bilateral breast tissues in case 1 (figures 4 and 5). We found all three types intermingled and diffusely distributed in the mammary parenchyma. The isolated/scattered type was seen more frequently. Looking at the NE markers, we saw that the frequency of each emerging pattern was different between chromogranin A and synaptophysin. The circumferential and clustered types were more frequently observed with chromogranin A than synaptophysin, while there was no difference between these markers in the frequency of the isolated/scattered pattern (figures 4 and 5). Cases 2 and 3 had findings similar to those described in case 1.

Figure 4

Distribution of each emerging pattern of neuroendocrine cells positive for chromogranin A and neuroendocrine carcinomas in wholly resected mammary tissues (case 1).

Figure 5

Distribution of each emerging pattern of neuroendocrine cells positive for synaptophysin and neuroendocrine carcinomas in wholly resected mammary tissues (case 1).

In contrast, we failed to detect such NE cells and patterns in the background tissues of 32 non-NE breast carcinomas.


It is well known that NE cells are widely distributed not only in endocrine organs but also in non-endocrine organs such as the gastrointestinal tract and lung. However, whether NE cells exist in normal mammary tissues is controversial3; some investigators reported observing very small numbers of NE cells,19–21 while others failed to detect them.22–24 Interestingly, in the present cases, we observed extensive numbers of NE cells with unremarkable cytological findings in the background tissues of B-NECs. To the best of our knowledge, this finding has not previously been described in the English-speaking literature.

In the urinary bladder, single NE cells lacking nuclear atypia can be found between the epithelial cells in some conditions such as cystitis glandularis or cystica, von Brunn's nests and urachal remnants.32 ,33 In the gallbladder and bile ducts, chronic inflammation could cause metaplastic change of the biliary epithelial cells to NE cells.34 ,35 It is therefore possible that the isolated NE cells found in mammary tissues could histogenetically appear in a metaplastic process of duct-lobular epithelium and/or NE differentiation of multipotential stem cells, whereas the stimuli leading to NE differentiation are not specified.

The bland-appearing NE cells in our breast cases were not just isolated and scattered, but were also in clustered and circumferential emerging patterns. It would be difficult to explain these varied findings of the NE cells as only a metaplastic process. These findings are morphologically suggestive of a hyperplastic condition of isolated NE cells. Hyperplasia of NE cells has been described in other organs: ECL cell hyperplasia of the stomach,10–12 diffuse idiopathic pulmonary NE cell hyperplasia (DIPNECH) of the lung,13–15 and C-cell hyperplasia of the thyroid.16–18 These conditions produce a generalised proliferation of scattered single cells, linear proliferations, small nodules and/or adenomatoid growths of normal-looking NE cells. These histological findings of NE cells in other organs are similar to the NE cells we found in our mammary cases. Therefore, it is conceivable that the mammary NE cells found in our current study are the result of a hyperplastic process of differentiated and/or metaplastic NE cells.

The precancerous or neoplastic nature of the hyperplastic condition of NE cells should be taken into account. It has been demonstrated that gastric ECL cell hyperplasia, lung DIPNECH and thyroid C-cell hyperplasia can be relevant to the development of NETs, which tend to have multiple occurrences.10–18 ECL cell NETs of the stomach arising in hypergastrinaemic conditions develop through a sequence of hyperplasia–dysplasia–neoplasia, and linear hyperplasia of ECL (NE) cells, or more advanced changes, represent a risk factor for the development of NETs.10–12 C-cell hyperplasia of the thyroid, when in a nodular form, is considered a definite precursor or cancerous lesion in the hereditary type of medullary carcinoma.16–18 In addition, in the current study, there was no evidence of hyperplastic NE changes in the cells of background breast tissues from 32 non-NE cancers. It is thus reasonable to speculate that mammary NE cells, especially those showing clustered and circumferential patterns, found in our NEC cases could have a premalignant nature, although further molecular and/or genetic analyses of these NE cells are essential for clarifying their true nature.

Cross et al and Tsang et al stated that the pagetoid spread of cancer cells into adjacent ducts and ductules was a common feature of NE-DCIS.25 ,26 Some NE cells located between the epithelial and myoepithelial layers in the current cases may have a histological resemblance to NE cancer cells showing a pagetoid spread. However, our NE cells, with their various emerging patterns, were found throughout the wholly-resected mammary glands and were not associated with cancerous lesions in their distribution. Furthermore, these cells had normal-like and/or benign morphologies, and some had an apical snout. Therefore, it is impossible to interpret widespread NE cells in our patients as merely the pagetoid spread of NE cancer cells.

Hereditary thyroid medullary carcinoma becomes clinically apparent in a young age group (mean age 35 years), is often multiple and bilateral, and is invariably accompanied by C-cell hyperplasia in the residual gland.16–18 The primary treatment for this type of NEC is surgical in the form of total thyroidectomy because of the multicentricity and C-cell hyperplasia. Intriguingly, the current breast cancer cases had similar clinicopathological features, that is, younger patients (in their 20s or 30s), multiple and/or bilateral occurrence, and NE cell hyperplasia in the background mammary tissue, although we failed to demonstrate a familial history of breast cancer or endocrine neoplasm in our cases.

In conclusion, NE cell hyperplasia, that is, isolated/scattered, clustered and/or circumferential proliferations of normal-looking NE cells, could appear in the breast tissues and be associated with the development of B-NECs. This novel observation of NE cells and patterns in non-neoplastic breast epithelia has implications for understanding of B-NEC carcinogenesis as well as for future therapies.

Preliminary results from this article were published as a poster presentation at the 101st Annual Meeting of the United States and Canadian Academy of Pathology (Vancouver, BC, Canada, March 2012).36

Take-home messages

  • This is the first description of widely distributed neuroendocrine (NE) cells showing various emerging patterns (isolated/scattered, clustered and circumferential) in background tissues of breast neuroendocrine carcinomas (B-NECs).

  • These novel findings can be morphologically regarded as representing ‘NE cell hyperplasia’ and are possibly relevant to B-NEC development as a precancerous lesion.

  • Treatment for B-NECs accompanied by this hyperplastic condition should be considered.


The authors thank Ms M Ogawa, Dr K Suzuki, Ms M Yoda, Ms M Ito, Mr K Sakamoto, Ms K Nakazawa, Mr Y Ishii, Mr T Yuminamochi, Ms K Kono and Dr P Zaber for technical support, and Dr G Sakamoto and Dr S Murata for helpful discussions.



  • Funding This work was supported by a Grant-in-Aid for Young Scientists (B) (No. 21790347 and 23790394) from the Japanese Ministry of Education, Culture, Sports, Science and Technology and the Yamanashi University Characteristic Prior Research Fund 2008, 2009 and 2010.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was conducted with the approval of the Ethics Committee, University of Yamanashi and St Luke's International Hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.