Pulmonary Neuroendocrine Cell Hyperplasia: Identification,Diagnostic Criteria and Incidence in Untreated Ageing Rats of Different Strains

Abstract

Pulmonary Neuroendocrine Cells (PNEC) are found as clusters called neuroepithelial bodies (NEB) or as single cells scattered in the respiratory epithelium. Pulmonary neuroendocrine cell hyperplasia is recorded in humans and experimentally manipulated rodents. The objectives of this work were to identify the optimal immunohistochemical markers for PNEC in the rat for use on paraffin-embedded, formalin-fixed material and to provide the first comparative incidence of PNEC hyperplasia in untreated 2-year-old rats of different strains. Calcitonin-gene related peptide (CGRP) and protein G product 9.5 (PGP9.5) antibodies identified PNEC consistently and selectively. In contrast, PNEC did not express chromogranin-A or S-100. PNEC hyperplasia was defined as foci of PNEC with greater than 40 nuclei, excluding overlying respiratory epithelium and submucosal PNEC. PNEC hyperplasia was observed at low incidence (0–7%) in untreated 2-year-old Sprague-Dawley, Han Wistar and Wistar rats but not Fischer 344 rats. This is the first report of spontaneous PNEC hyperplasia in rats. The cause of this hyperplasia is unknown, but experimental models that induce PNEC hyperplasia by causing bronchiolar cell injury are discussed. PNEC neoplasia in the rat is unreported in the literature and was not observed in animals examined in this study.

Introduction

Pulmonary Neuroendocrine Cells (PNEC) are found as clusters called neuroepithelial bodies (NEB) or as single cells scattered in the respiratory epithelium. They have been detected in all air-breathing vertebrates studied so far. Although there are estimated to be 3500 NEB in the rat lung (Van Genechten et al., 2004), they are often not recognised on conventional haematoxylin and eosin (H&E) stained tissue sections prepared as part of toxicology studies. In the rat, NEBs are located at all levels of intrapulmonary airways from the bronchi to alveoli but predominantly in the bronchioles. NEBs extend from the basement membrane to the airway lumen. Morphologically they resemble known chemoreceptors, such as taste buds and carotid bodies.

PNEC are chemoreceptors sensitive to hypoxia and hypercapnia and may exert paracrine control on proliferation of adjacent cells (Youngson et al., 1993). They are richly innervated from afferent sensory fibres from the vagus and spinal nerves (Brouns et al., 2003). During lung development, PNEC are derived from a different set of precursors than those forming peripheral respiratory epithelial cells (Perl et al., 2002) and are thought to play a role in regulation of branching morphogenesis and cellular growth and maturation. PNEC produce and secrete several neuropeptides, such as bombesin, calcitonin gene-related peptide (CGRP), gastrin-releasing peptide (GRP) and calcitonin. CGRP and GRP are the major neuropeptides produced by PNEC in rodents and primates respectively (Li et al., 1994). Release of these mediators may result in redistribution of pulmonary blood flow, regulation of bronchomotor tone, and modulation of the immune response.

Although the distribution of PNEC in the normal adult rat has been known for a long time (Gosney and Sissons, 1985), there is no published account of spontaneous PNEC hyperplasia in adult rats. Diagnostic criteria for distinguishing NEB from PNEC hyperplasia are poorly defined. Spontaneous PNEC neoplasia is, to the authors’ knowledge, unrecorded in the rat.

There were 3 aims of this work. First, to establish robust immunohistochemical (IHC) markers to identify pulmonary neuroendocrine cells in the rat on formalin-fixed, paraffin embedded material. Second, to develop criteria for PNEC hyperplasia in the rat. Finally, we aimed to investigate spontaneous PNEC hyperplasia in different strains of 2-year-old rats and report the incidence observed by retrospective analysis of animals used in carcinogenicity studies.

Materials and Methods

Animals and Tissues

Animals from 7 previous carcinogenicity studies were examined (1 Wistar, 3 Han Wistar, 2 Fischer 344 and 1 Sprague–Dawley study). These studies are identified by strain and study letter in Table 1. Standard lung sections from 2-year-old untreated rats were obtained from these studies and examined microscopically. Lungs from all rats within each of the 2 control groups per sex, which survived to the terminal sacrifice at 2 years, were examined. The number of animals per sex per group is indicated in the results table.

Histology Procedures

Lungs were fixed in 10% buffered formalin by tracheal inflation and then immersion. Fixed lung was embedded in paraffin wax, sectioned (3–5 μm) and stained with haematoxylin and eosin (H&E). Serial sections were used for immunohistochemistry.

Diagnostic Criteria

The incidence of PNEC hyperplasia was recorded using H&E stained sections. PNEC hyperplasia was defined as a discrete PNEC cluster with greater than 40 nuclei (Ito et al., 1992). This figure excludes epithelial surface cell nuclei and nuclei below the basement membrane within adjacent connective tissue.

Immunohistochemistry

An indirect peroxidase Streptavidin ABC/HRP-DAB system was used with the following antibodies: chromogranin A (rabbit polyclonal, Lab-Vision RB903-PI, 1:250 with no antigen retrieval), S100 (rabbit polyclonal, Dako Z0311, 1:1500, antigen retrieval in pH 9.0 buffer at 97°C for 20 minutes), calcitonin gene-related peptide (CGRP) (Polyclonal Goat, Biogenesis 1720-9007, 1:2000 with no antigen retrieval), PGP 9.5 (13C4/13C4) (Mouse monoclonal, Abcam ab8189,1:50, antigen retrieval microwaved for 15 minutes in pH 6.0 citrate buffer), pan cytokeratin (mouse monoclonal, Sigma C2562,1:500, antigen retrieval microwaved for 15 minutes in pH 6.0 citrate buffer). Sections were dewaxed, incubated for 30 minutes with 0.5% hydrogen peroxide/methanol and antigen retrieval was performed as noted. Nonspecific binding was blocked with normal swine serum (Dako X0901) or normal rabbit serum (Dako 0902) in Tris saline, before incubation with the primary antibody overnight at 4°C.

After washing in Tris saline, sections were incubated for 30 minutes with biotinylated swine anti rabbit (Dako E0353), biotinylated rabbit anti goat (Dako E0466) or rabbit anti-mouse immunoglobulin (Dako E0354), rinsed and the Streptavidin AB complex added. Diaminobenzidine substrate was added to visualise peroxidase activity and sections were counter-stained with Mayer’s haematoxylin.

For each primary antibody, appropriate positive control tissue was included as part of the IHC procedure to demonstrate antigen specificity. In addition, isotype-matched monoclonal or irrelevant polyclonal antibodies were used as negative control antibodies.

Results

Identification of NEB and PNEC Hyperplasia

NEB were identified initially on H&E-stained lung sections. At low power, these clusters of PNEC need to be differentiated from sites of increased epithelial thickness (plane of section artefact) or subepithelial lymphoid infiltrates. However at high power, the following characteristic features of PNEC were clear: indistinct cell borders, vacuolated cytoplasm and nuclei with stippled chromatin.

PNEC hyperplasia was identified as a large NEB-like structure with greater than 40 nuclei. These polypoid projections up to 0.2 mm in diameter extended into the bronchiolar lumen and contained NE cells with indistinct cell borders and vacuolated cytoplasm. Frequently, NE cells surrounded areas of eosinophilic connective tissue (Fig. 1). Occasionally, mitoses were present within foci of PNEC hyperplasia. In contrast, areas of connective tissue and mitoses were not observed in NEB. On H&E-stained sections, the overlying layer of respiratory epithelium was indistinct. Frequently these polypoid foci contained multiple finger-like projections and serial sections were required to demonstrate that they shared a common stalk and constituted a single focus of hyperplasia.

The selection of antibodies for use in immunohistochemistry was based on those commonly used to identify neuroendocrine cells (calcitonin gene-related peptide (CGRP), protein G product 9.5, chromogranin A), nerve (S-100) and respiratory epithelium (pan-cytokeratin). Using immunohistochemistry, solitary NE cells, NEB and foci of PNEC hyperplasia stained positive for CGRP, protein G product 9.5 and pan-cytokeratin (Fig. 2). In contrast, they were negative for chromogranin A and S-100 (results not shown).

However positive control rat tissue for chromogranin A (adrenal medulla) and S-100 (peripheral nerve within lung) were positive, indicating that these antibodies were able to detect their respective antigens within the same immunohistochemical run. Pan-cytokeratin was expressed by the majority of PNEC, in addition to the overlying and adjacent respiratory epithelium (Fig. 2b). In contrast, use of PGP9.5 distinguished the overlying epithelium which was negative, from the positive PNEC within the focus. This marker also enabled PNEC within the underlying submucosa to be recognised (Fig. 2c). CGRP also enabled these cell compartments to be distinguished (Fig. 2d). For the purposes of the diagnostic criterion, PNEC within the submucosa were excluded from the cell count.

Incidence of PNEC Hyperplasia

The incidence of PNEC hyperplasia varied according to strain and sex (Table 1). The highest incidence was observed in Sprague–Dawley and Han Wistar rats. In a single Sprague–Dawley rat study, PNEC hyperplasia was identified in 2% of males and 7% of females. In 3 Han Wistar rat studies, the incidence varied from 0 to 2.6% in males and 0 to 6% in females. Although the highest incidence for this strain was observed in an inhalation study, the significance of this is not clear on the limited data available. In a single Wistar rat study, the incidence was 3.3% in males and 0% in females.

The lowest incidence was seen in control Fischer 344 rats. In this strain, NEB were identified but PNEC hyperplasia was not observed in untreated animals sacrificed at the end of either study. However in a single treated animal on study B and in a decedent untreated animal on study C, PNEC hyperplasia was observed (data not shown), demonstrating that it can occur in this strain.

Discussion

This is the first report of PNEC hyperplasia as a background finding in untreated rats. This histological change can be recognized by careful microscopic observation of standard, H&E-stained lung sections. Previous investigators, examining fewer animals, have reported that untreated old rats do not show PNEC hyperplasia (Elizegi et al., 2001). Others have identified this change at low incidence in untreated young hamsters (Ito et al., 1992). It is the authors’ conclusion that this change may be underreported in toxicological studies in the rat, both as a background and potentially as a treatment-related finding.

The diagnostic criteria for PNEC hyperplasia in different species is not yet standardised. In humans, PNEC hyperplasia is recognised in association with chronic pulmonary inflammation or fibrosis. When these cells extend into the interstitium associated with fibrosis, then it is termed a carcinoid tumourlet. If the lesion exceeds 5 mm in diameter, then it is termed a carcinoid tumour (Kerr, 2001).

In the rat, early work defined a NEB as containing between 5–20 cells (Wasano, 1977). Reports variously use the number of NEB or the number of nuclei within NEB as criteria for hyperplasia. All quantification of NEBs is subject to variations of methods and to a lesser extent, markers. There is, therefore, a need for standardization of terminology and diagnostic criteria for PNEC hyperplasia. This finding is currently not listed in diagnostic guides for toxicological pathology (Schwartz et al., 1994). PNEC hyperplasia is defined in this study as a discrete PNEC focus consisting of greater than 40 nuclei, excluding epithelial surface cell nuclei and PNEC nuclei below the basement membrane within adjacent connective tissue.

This represents a refinement of the criteria previously used in hamster models (Ito et al., 1992). Further morphological features that can be present in the rat are also described. For example, within PNEC foci there are frequently areas of eosinophilic connective tissue. On rare occasions, mitoses can also be observed. To the authors’ knowledge, this is the first example of mitoses being observed within PNEC hyperplasia in the rat. Since these features are variably present, we do not consider that they should be included as requirements of the diagnostic criteria.

Confirming the identity of PNEC cells in NEB or foci of hyperplasia is readily achieved by use of IHC. The present study illustrates that antibodies to CGRP and PGP9.5 are effective markers of PNEC in the rat. CGRP, which is also expressed in C-cells of the thyroid, is a potent vasodilator and may contribute to redistribution of pulmonary blood flow at times of altered inspired oxygen concentrations. Expression of PGP9.5 is highly specific to neurons and to cells of the diffuse neuroendocrine system and their tumors (Lauweryns et al., 1988). This was the optimal marker for PNEC in this study, and the method described here gave very low background staining and clear differentiation of overlying respiratory epithelium from underlying PNEC.

A majority of PNEC also stained positive using a pan-cytokeratin polyclonal antibody. This is an interesting result since cytokeratin intermediate filaments were not present in rabbit NEB in a previous report using an antibody to low molecular weight cytokeratin (Pan et al., 2004). Few markers are available to distinguish solitary PNEC from NEB. The pituitary protein marker 7B2 is reported to selectively label NEB but not solitary PNEC (Seldeslagh and Lauweryns, 1994). Despite being an excellent marker for neuroendocrine cells in the gastrointestinal tract, pancreatic islets and adrenal medulla, chromogranin-A was not expressed by PNEC in this study.

PNEC hyperplasia was observed as a background finding in 2 year old rats of all strains examined, except the Fischer 344. The reasons why this change develops spontaneously are not understood but review of the experimental literature offers some clues. For example, 21-day-old rats treated with chronic hyperoxia results in an increased number of both solitary PNEC and NEB (Shenberger et al., 1997). Unfortunately, the number of cells per NEB is not reported and therefore the criteria for PNEC hyperplasia applied in the current study cannot be used. Infants with bronchopulmonary dysplasia have increased numbers of PNEC, which is considered to be due to oxygen toxicity.

PNEC hyperplasia is known to occur in rats in response to pulmonary hypertension and exposure to silica. Silica induces hyperplasia of rat alveolar NEB but not bronchiolar NEB. This change accompanies marked silicosis with alveolar and bronchiolar epithelial hyperplasia. NEB hyperplasia is not observed in mice and hamsters using this model, suggesting that the rat is predisposed to this lesion (Elizegi et al., 2001). In this study, pro-adrenomedullin N-terminal 20 peptide was reported to be increased in expression in hyperplastic NE cells relative to normal NEB in the rat and is thus a potential marker for reactive NE cells.

In mice, using naphthalene to selectively destroy Clara cells, increased numbers of NEB are detectable during the recovery phase with increased thymidine labelling of CGRP immunoreactive cells (Stevens et al., 1997). Therefore, the spontaneous PNEC hyperplasia observed in aged rats in this study may result from previous bronchiolar injury or local changes in inhaled oxygen tension.

Is PNEC hyperplasia an age-related change? In the authors’ experience, this finding is not observed in young rats. However, it may occur at low incidence and the establishment of clear criteria will enable screening for this change on short term toxicity studies. There is no significant difference in the distribution or absolute number of NEB between neonate and adult rats when compared by airway generation (Avadhanam et al., 1997). However the density of NEB decreases, probably by lung expansion and growth.

In humans, PNEC hyperplasia has been proposed as a pre-neoplastic lesion leading to NE lung tumours (Gould et al., 1983). Clinically, the most important NE carcinoma is small-cell lung cancer, which accounts for 15–20% of all lung cancers; others include non-SCLC with neuroendocrine features (NSCLC-NE), carcinoid and large cell NE carcinoma (Linnoila, 2006). However, PNEC hyperplasia and dysplasia, which are common in tumour-bearing lungs, appear to be a component of the general pathological process, such as inflammation or emphysema, rather than a preneoplastic lesion (Linnoila, 2006). Severely hyperplastic and dysplastic PNECs can also express ectopic hormones, e.g., adreno-corticotropic hormone and vasoactive intestinal polypeptide (Gould et al., 1983). These hormones have not been reported in similar morphological entities in rodents. There are no reported spontaneous cases of PNEC neoplasia in rodents. This suggests that rodents are more resistant to developing these tumours than humans.

Exposure of hamsters to chemical carcinogens such as nitrosamines often leads to marked PNEC hyperplasia, which is reversible, and non-NE adenomas and NSCLC (Sunday et al., 1995). Paracrine effects of PNEC-derived growth factors may contribute to the development of non-NE lung tumours in this hamster model (Sunday et al., 1994).

In human SCLC, mutations or alterations in p53 and Rb-1 tumour suppressor genes are common. Conditional ablation of Rb family proteins leads to PNEC hyperplasia in mice (Wikenheiser-Brokamp, 2004). Transgenic mice with conditional alleles for both Rb-1 and p53 (Rb-1 ^−/−/p53 ^−/−) in lung epithelial cells develop aggressive lung tumours arising from bronchiolar epithelium and with similarities to human SCLC including morphology, high expression of ASH-1, NE markers and extrapulmonary metastases (Meuwissen et al., 2003). Therefore Rb function is critical for regulating NE differentiation in normal and neoplastic lung. However in the absence of genetic manipulation, rodents appear to be resistant to developing NE tumours.

In summary, this is the first report of spontaneous PNEC hyperplasia in the rat. This change can be identified on standard H&E stained lung sections and confirmed if necessary using CGRP or PGP9.5 IHC. Diagnostic criteria for this change are described and the incidence provided in ageing rats of four different strains. Application of these diagnostic criteria will permit collection of further background incidence data and reveal the extent of PNEC hyperplasia in rats of different ages. This knowledge will improve our understanding of the biological behavior of these cells in untreated animals and in those exposed to xenobiotics.