Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Pulmonary neuroendocrine cells (PNECs) are abundant in mammalian fetal lung (1, 2), in which they play a significant role in normal lung growth and maturation (3, 4). PNEC hyperplasia is associated with chronic lung diseases in humans (5), in whom it is thought to play a role in reparative responses to lung injury and may be a precursor to lung cancer (5). Previously, we characterized PNEC hyperplasia occurring in hamsters during 20 wk of combination treatment with a diethylnitrosamine (DEN) and continuous 65% hyperoxia (DEN/O₂) (9). DEN/O₂-treated hamsters displayed marked PNEC hyperplasia that was 5- to 10-fold more intense than that achieved in hamsters treated with DEN or hyperoxia alone. The induction of PNEC differentiation in DEN/O₂-treated hamsters occurred after 8 to 12 wk of treatment and was followed by spontaneous regression of PNEC clusters between 14 and 20 wk despite continued DEN/O₂ treatment. It appeared that the PNEC hyperplasia was a transient response to preneoplastic lung injury. We hypothesized that: (1) the downregulation of PNEC numbers might be mediated via degradation of endogenous bioactive peptides mediating PNEC differentiation; and (2) these changes could be due to altered cell proliferation and/or apoptosis, together with PNEC differentiation.

The present study tested these hypotheses in the hamster DEN/O₂ lung injury model using SCH32615, a long-lived CD10/neutral endopeptidase 24.11 (CD10/NEP) inhibitor, during 16 wk of DEN/O₂ treatment. Lung tissues from the two groups of DEN/O₂-treated hamsters were then analyzed using immunohistochemistry for calcitonin gene-related peptide (CGRP), the most sensitive marker of pulmonary PNECs in rodents, and proliferating cell nuclear antigen (PCNA) to determine whether increased proliferation of PNECs was occurring with CD10/NEP inhibition. In situ visualization of apoptosis in PNECs was carried out in the same tissue sections using the highly sensitive and specific Apoptag (terminal deoxynucleotidyl transferase- mediated dUTP nucleotide nick-end labeling [TUNEL]) fluorescence method. We then utilized morphometric analyses to determine the numbers of CGRP-positive PNECs with and without apoptosis in the two groups of DEN/O₂-treated hamsters. Although the number of animals we could treat was only three to five per group per time point, due to limited availability of SCH32615, this investigation led to several novel observations that shed light on cellular mechanisms of PNEC differentiation and regression.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Care and Treatment of Hamsters

Outbred female Syrian golden hamsters (virus- and pathogen-free) were received from Charles River Laboratories (Wilmington, MA) at 6 wk of age and were rested for 2 wk before beginning DEN and exposure to continuous hyperoxia. Hamsters were maintained under standard laboratory conditions in compliance with Interdisciplinary Principles and Guidelines for the Use of Animals in Research, Testing, and Education as described previously (9, 10). Animals were exposed to 65% oxygen in a vented Plexiglas chamber by ventilation with a mixture of 100% O₂ and compressed air at a flow rate of 20 to 25 liters/min, with relative humidity maintained at 45 to 60%. This same chamber was used successfully by another group for hyperoxia studies (11). Oxygen levels were continuously monitored with a digital oxygen analyzer. The animals were removed from this environment only for brief (less than 15 min) intervals required for subcutaneous injections with DEN. Oxygen-exposed hamsters were treated with DEN (20 mg/kg in 0.1 to 0.4 ml of phosphate-buffered saline ([PBS]) (Sigma Chemical, St. Louis, MO) on Tuesdays and Thursdays at 7:30 A.M. Other controls included hamsters receiving injections of DEN at ambient oxygen levels and hamsters from 8 to 24 wk of age maintained without either DEN, 4-(methylnitrosamine)-1-(3-pyridyl)-1-butanone, or hyperoxia. At the time of harvest, animals were rapidly asphyxiated with carbon dioxide, and autopsy of the thoracic and abdominal viscera was carried out. Lungs were fixed immediately in 4% paraformaldehyde for immunochemistry and histology.

Antisera

Rabbit polyclonal anti-CGRP antiserum (Peninsula Laboratories, Inc., Belmont, CA) was used at a 1:5,000 dilution. As a negative control, anti-CGRP antiserum was preabsorbed with rat CGRP at a concentration of 5 µg/ml of diluted antiserum. Monoclonal anti-PCNA antibody was obtained from Dako Corp. (Carpinteria, CA) and used as previously described (12). Rabbit anti-protein gene product (PGP) 9.5 without and with antigen preabsorption (Ultraclone, Isle of Wight, UK) was used at 1:1,000 dilution. Biotinylated goat antirabbit immunoglobulin G was obtained from Vector Laboratories, Inc. (Burlingame, CA), as were the standard avidin-biotin (ABC) immunoperoxidase reagents. Sections were counterstained using 2% aqueous methyl green (Sigma).

Immunoperoxidase Analyses

Immunoperoxidase analyses were carried out using the standard ABC kit as described previously (9). In brief, tissues were fixed for 16 to 18 h in 4% paraformaldehyde, routinely processed into paraffin, and cut at 5 µm onto coated slides. Tissues were permeabilized with 0.3% Triton X-100 and blocked for 20 min with 1:10 dilution of normal goat serum in PBS before overnight incubation at 4°C with primary antisera (dilutions as given previously). A 1:200 dilution of biotinylated secondary antibody was applied for 30 min, endogenous peroxidase blocked for 30 min in 0.3% hydrogen peroxide (H₂O₂)/PBS, and the standard ABC reagent applied for 45 min. Slides were developed for 5 min in diaminobenzidine (0.25 gm/100 ml in PBS with 0.03% H₂O₂) and counterstained with 2% aqueous methyl green before being reviewed.

Apoptosis Assay

Two assays for in situ apoptosis were used: (1) morphologic evidence of apoptotic bodies using DAPI nuclear staining; and (2) TUNEL analysis for DNA fragmentation in tissue sections, using the Apoptag fluorescence protocol on hamster lung sections that had previously been immunostained for CGRP. Fluorescein isothiocyanate-digoxigenin nucleotide labeling of 3'-OH DNA ends was carried out according to the manufacturer's specifications, with the proteinase K incubation for 30 min (Apoptag; Oncor). Specimens were analyzed by brightfield microscopy to visualize cells containing both nuclei and CGRP-positive cytoplasm, at 360 nm to visualize DAPI (blue) nuclear fluorescence and at 530 nm for fluorescein.

Morphometric Analyses

To compare slides from different animals, all foci positively stained for CGRP were counted and normalized by expression as a function of the total length of bronchial plus bronchiolar epithelium present on each slide (normalized count = number of foci present divided by the length of airway epithelium). The positive foci represent both single PNECs and clusters of PNECs (two or more clustered PNECs). We refer to the enlarged clusters of CGRP-positive PNECs as PNEC clusters rather than as neuroepithelial bodies (NEB) because normal NEB are innervated and we did not observe nerve fibers innervating the PNEC clusters (the PNECs and the nerve fibers should immunostain for CGRP and/or PGP 9.5). The length of airway epithelium present was quantitated using a micrometer grid with the ×10 objective and ×10 eyepiece, where 1 unit length equals 100 µm. At least three complete lung cross-sections more than 200 µm apart were counted for each animal.

Morphometry for PNEC proliferation was carried out by counting the number of PCNA-positive nuclei per PNEC cluster in thin (3-µm) sections adjacent to a serial CGRP-immunostained section. The ×40 power objective was used for these analyses, with all or at least 11 PNEC clusters counted. The percentage of CGRP-positive PNECs with PCNA-immunostained nuclei was used as an estimate of PNEC proliferation.

Morphometry for PNEC apoptosis was carried out by counting the number of Apoptag fluorescence-positive nuclei in lung sections that had been previously stained for CGRP immunoreactivity. The ×100 power oil objective was used for these analyses, with all or at least 11 PNEC clusters counted. The percentage of total CGRP-positive PNECs with Apoptag fluorescence-positive nuclei was used to quantitate PNEC apoptosis as described previously (13).

Statistical Analyses

Numerical data were analyzed using the unpaired Student's t test, with values expressed as means ± 1 standard error of the mean (SE).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Immunoperoxidase Analyses and Morphometry of PNECs

The results of immunoperoxidase analyses for CGRP-positive PNECs are given in Figures 1 and 2. In our previous studies we demonstrated that CGRP immunostaining is the most sensitive method for detection of hamster PNECs (9). In the present 16-wk study, DEN/O₂-treated hamsters had up to a 10-fold increase in the number of CGRP immunopositive cells per centimeter of airway epithelium in hamster lung compared with lung from untreated control hamsters (Figure 2A), which was evaluated as the number of nuclei per CGRP-positive PNEC cluster (as shown in Figure 1A) times the number of PNEC foci per centimeter of airway epithelium (as indicated by the arrows in Figures 3A and 3D). The antigen specificity of CGRP immunostaining was confirmed by pretreating anti-CGRP antiserum with CGRP (5 µg/ml), which abolished immunostaining (Figure 1B). The results of morphometric analyses are given in Figure 2, and representative photomicrographs are shown in Figure 3. Compared with untreated controls, there was a comparable 3- to 4-fold increase in the number of CGRP-immunopositive foci per centimeter of airway epithelium in hamster lung after 4 wk of treatment with DEN/O₂ with or without SCH32615 (Figure 2A). After 8 to 16 wk of treatment, DEN/O₂-treated hamsters with CD10/NEP inhibition had ~ 6 to 30% more CGRP-positive foci per centimeter of airway epithelium (Figures 3A and 3B) than did DEN/O₂-treated animals without CD10/ NEP inhibition (Figures 3C and 3D), but in this small cohort of three to four animals per group this difference was not statistically significant (Figure 2A). Both groups had a decline in the number of foci per centimeter after 12 wk of treatment, followed by a rise to peak numbers of foci (~ 5- to 6-fold increase compared with untreated controls) after 16 wk of treatment. After 8 to 12 wk of treatment, there were significantly more CGRP-positive cells per focus in DEN/O₂-treated animals with CD10/NEP inhibition (Figure 3B), which had about 1.5 times as many cells per focus as compared with similarly treated animals without CD10/ NEP inhibition (Figures 2B, 3C, and 3D). After 8 to 12 wk of treatment, this increased number of cells per focus was reflected in a similar trend toward increased total numbers of CGRP-positive PNECs per centimeter of airway epithelium in DEN/O₂-treated animals given SCH32615 (Figure 2C). Hamsters treated for 12 wk with SCH32615 plus DEN/O₂ had only a minimal change in numbers of PNECs per centimeter of airway compared with animals similarly treated for 8 wk. In contrast, there was a 50% decline in total number of PNECs per centimeter in DEN/O₂-treated animals without CD10/NEP inhibition over the same time interval (Figure 2C). Most of this latter decrease in the number of total PNECs reflects decreased numbers of PNEC foci (Figure 2A), consistent with spontaneous regression. As shown in Figure 2, at this 12-wk time point there were significantly more CGRP-positive cells per cluster and total CGRP-positive cells per centimeter of airway in animals with CD10/NEP inhibition as compared with animals without CD10/NEP inhibition. After 16 wk of treatment with DEN/O₂, there was no longer any significant difference in the relative numbers of CGRP-positive cells between animals with and without CD10/NEP inhibition (Figures 2C and 3C). Of note, there was considerably greater variability in numbers of PNECs between hamsters even within each experimental group at the 16-wk time point (Figure 2C). In all hamsters, the same PNEC clusters were immunostained by both CGRP and PGP 9.5, and thus the relative numbers of PGP 9.5-positive PNECs changed in parallel with the CGRP-positive cells (data not shown). These data agree with our earlier extensive morphometric comparisons of numbers of PNECs immunostained for CGRP, serotonin, and calcitonin, which followed the same kinetics of appearance and disappearance in DEN/O₂-treated hamsters (9).

View larger version (95K): [in this window] [in a new window]   Figure 1.   Antigen specificity of CGRP immunostaining in hamster lung. CGRP immunostaining was carried out using sections of hamster lung after 8 wk of CD10/NEP inhibition together with DEN/O2 treatment. (A) A large NEB is visible, containing prominent CGRP-positive PNECs. (B) A thin serial section of the section shown in (A) was run in parallel using antigen-preabsorbed CGRP antiserum. Note the loss of specific immunoreactivity (original magnification: ×1,000). L = airway lumen; arrows indicate the width of the PNEC cluster along the basement membrane.

View larger version (16K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   Figure 2.   Morphometric analyses of PNECs in hamsters treated for 4 to 16 wk with DEN/O2 with and without CD10/NEP inhibition. Serial sections of hamster lung were immunostained for CGRP. The number of CGRP-positive PNECs per centimeter of airway epithelium (C) was calculated by multiplying (A) the number of CGRP-positive PNEC foci per centimeter of airway epithelium by (B) the mean number of nuclei per CGRP-positive PNEC cluster. The number of weeks of treatment with DEN/O2 ± SCH32615 is given on the x axis. P values are shown for the comparison of animals given SCH32615 (solid squares) versus those without SCH32615 (open circles). P values: B, *P < 0.001, **P < 0.04; C, *P < 0.04, P = 0.077. Values in A were not significantly different between the two groups at any time point.

View larger version (159K): [in this window] [in a new window]   Figure 3.   Comparison of PNEC hyperplasia in hamsters treated with DEN/O2 with and without SCH32615. CGRP immunostaining was carried out using lung tissue sections from DEN/O2-treated hamsters with (A and B) or without (C and D) CD10/ NEP inhibition for 8 to 16 wk (original magnification × 300). (A) After 8 wk of treatment, PNEC foci from the group treated with SCH32615 contained more nucleated cells on average (refer to morphometric data in Figure 2B) than did the group receiving DEN/O2 alone (C and D). (B) After 16 wk of treatment, the PNEC foci from both groups are similar in size. (C, D) After 8 wk of DEN/O2 treatment without CD10/NEP, CGRP-positive PNECs occurring in clusters in hamster lung appear to be separated by CGRP-negative cells. The PNECs in the CGRP-positive clusters tend to be more compact in animals with CD10/NEP inhibition (A and B) than in hamsters without CD10/NEP inhibition (C and D).

It is interesting that the PNEC clusters in hamsters treated with SCH32615 demonstrated a more compact morphology (Figures 3A and 3B). In contrast, there were many CGRP-negative cells interspersed between the CGRP- positive cells within each PNEC cluster in hamsters without CD10/NEP inhibition (Figures 3C and 3D). Most of these CGRP-negative cells were found to be positive for Apoptag fluorescence, suggesting that these cells might be PNECs that have lost their membrane integrity and cytoplasmic contents.

In Situ Analyses for Apoptosis and Proliferation of CGRP-Positive PNECs

To determine whether a decrease in apoptosis could be one mechanism contributing to the relative increase in PNECs in animals with CD10/NEP inhibition, sections of hamster lung tissue were immunostained for CGRP before in situ analysis for DNA fragmentation (evaluation for nuclear fragmentation on DAPI nuclear staining and also Apoptag nuclear fluorescence). Photomicrographs of representative PNEC clusters are shown in Figure 4, with brightfield examination of CGRP-positive cells given on the left, DAPI nuclear staining shown in the middle, and the same field viewed at 530 nm for Apoptag fluorescence given on the right. To be counted as apoptotic, a nucleus had to demonstrate both the presence of fragmented nuclear forms, or apoptotic bodies, and easily visible Apoptag fluorescence. There was a tendency for apoptotic nuclei to be located at the periphery of PNEC clusters (indicated by arrows in Figures 4C and 4F). The highest proportion of apoptotic cells occurred in PNEC clusters with weaker CGRP-positive immunostaining that appeared to be fading, similar to PNEC clusters originally observed in DEN/O₂-treated hamsters after 16 to 20 wk of treatment, during the time period of spontaneous PNEC regression (9). The results of quantitative morphometric analyses are summarized in Table 1 and Figure 5. There was a trend toward decreased PNEC apoptosis in CD10/ NEP-inhibited animals at all time points, which attained marginal statistical significance after 16 wk (Table 1 and Figure 5A). There was a direct linear correlation between numbers of PNECs per centimeter of airway epithelium and the percentage of apoptotic PNECs in both groups of hamsters (Figures 5B and 5C). This linear relationship was most significant for the CD10/NEP-inhibited group (Figure 5B), in which the leftward shift of the graph, compared with animals without CD10/NEP inhibition (Figure 5C), indicates an overall lower prevalence of PNEC apoptosis.

View larger version (67K): [in this window] [in a new window]   Figure 4.   In situ apoptosis of PNECs in DEN/O2-treated hamsters without and with CD10/NEP inhibition. TUNEL analyses were carried out using tissue sections that had previously been immunostained for CGRP to demonstrate hamster PNECs. Photomicrographs of representative sections are given as follows (original magnification ×500): Left column (A and D): Brightfield microscopy of tissue section with CGRP immunostaining demonstrating nucleated neuroendocrine cells. Center column (B and E): DAPI nuclear staining viewed at 360 nm (blue). Right column (C and F ): Apoptag fluorescence of the same field as on left, viewed at 530 nm (fluorescein is yellow-green). A-C: PNEC cluster in a hamster treated with DEN/O2 alone for 8 wk. D-F: Lung from a hamster treated with DEN/O2 and SCH32615 for 8 wk. There was prominent Apoptag fluorescence in about half of the PNEC clusters, including apoptotic bodies that are identifiable by both positive DAPI nuclear staining (B and E) and Apoptag-positive fluorescence (C and F: arrows) in the PNEC clusters. The apoptotic cells occur frequently at the periphery of the PNEC clusters. Note that background cytoplasmic fluorescence of non-neuroendocrine cells does not obscure the neuroendocrine cells in these thin (3-µm) sections. Also, a negative nuclear image is apparent in most of these non-neuroendocrine cells. However, we did not evaluate apoptosis in non-neuroendocrine cells because of the cytoplasmic background fluorescence. Background staining of nerve fibers, connective tissue, and erythrocytes also did not interfere with evaluation of the neuroendocrine cells.

View larger version (15K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Figure 5.   Morphometric analyses of apoptotic PNECs in DEN/O2-treated hamsters with and without CD10/NEP inhibition. Apoptag analyses were carried out using tissue sections that had previously been immunostained for CGRP to demonstrate hamster PNECs, as shown in Figure 4. Numbers of PNECs, both total nuclei (counted on brightfield and DAPI analyses) and the number of nuclei with one or more apoptotic bodies were counted on the same slides for at least 20 random PNEC clusters per slide, or for all PNEC clusters present if there were less than 20. P values are given in Table 1.

To determine whether CD10/NEP inhibition also alters the proportion of proliferating PNECs during the early time period, immunoperoxidase staining of PNECs for PCNA was carried out using thin (3-µm) immediately serial sections of hamster lung. The results of these analyses are given in Figure 6. At the 4-wk time point there were, on average, about twice as many PNECs immunostained for PCNA. However, the standard error was wide because the total number of PNEC clusters per complete lung cross-section of each hamster was very small (two to six), and there were only three animals in each experimental group. It appears likely the lack of statistical significance is largely due to the small number of available samples. By 8 wk this trend became more meaningful, with P = 0.10 (11 PNEC clusters counted per animal). It was only at 12 wk that there was a statistical difference in PCNA-labeling of PNECs between the animals with and without CD10/NEP inhibition, with about a 3-fold increase in PNEC proliferation in animals with CD10/NEP inhibition (P < 0.001).

View larger version (31K): [in this window] [in a new window]   Figure 6.   Morphometric analyses of proliferating PNECs in DEN/O2-treated hamsters with and without CD10/NEP inhibition. Thin (3-µm) immediately serial sections of hamster lung were used for CGRP immunostaining to identify clusters of PNECs for PCNA immunostaining of nuclei within the PNEC clusters. At the 4-wk time point, all PNEC clusters were counted (only two to six clusters per animal on a complete lung cross-section); and at the 8- to 12-wk time points, 11 PNEC clusters were counted per animal. P values are given in the figure.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present study demonstrated more rapid induction of PNEC hyperplasia by DEN/O₂ in hamsters with CD10/ NEP inhibition. The precise molecular and cellular mechanisms for this rapid induction of PNEC hyperplasia are not known. It is likely that the increased number of PNECs with CD10/NEP inhibition is due to higher levels of endogenous bioactive peptides capable of inducing PNEC proliferation, differentiation, and/or apoptosis. The early increase in PNEC numbers after 8 to 12 wk with CD10/ NEP inhibition appears to be primarily due to promotion of PNEC differentiation from undifferentiated epithelial precursor cells. Although significant differences in PNEC proliferation were not observed until 12 wk of treatment, a trend was present after 8 wk (P = 0.10). It is unlikely that differences in PNEC apoptosis played an immediate role in promoting enhanced PNEC hyperplasia, because CD10/ NEP inhibition was not associated with significantly decreased PNEC apoptosis until 16 wk.

The present investigation was based on our earlier observation that CD10/NEP gene expression is induced in hamsters treated with DEN/O₂ in parallel with the occurrence of transient PNEC hyperplasia, suggesting that the enzyme might be at least partly responsible for the observed spontaneous downregulation of PNECs. To test this hypothesis, we treated a different group of outbred Syrian hamsters with CD10/NEP inhibition during preneoplastic lung injury induced by DEN/O₂. In this second series of DEN/O₂-treated animals, the control hamsters (without CD10/NEP inhibition) demonstrated slightly different kinetics of PNEC hyperplasia than were previously observed (9, 10), with a first submaximal peak in PNEC numbers attained at 8 wk (rather than at 12 wk). In this control DEN/O₂-treated group, only partial spontaneous regression of PNECs was observed after 12 wk of treatment, with fewer PNEC foci per centimeter of airway epithelium and fewer PNECs per focus. Thereafter, animals treated for 16 wk achieved peak numbers of PNECs. CD10/NEP inhibition in a subgroup of these DEN/O₂-treated hamsters led to more rapid induction of PNEC hyperplasia, with peak numbers of PNECs attained after 8 wk of treatment, and diminished the spontaneous regression of PNECs, which is most clearly evident from the almost invariant total number of PNECs per centimeter of airway after 8 to 12 wk of treatment.

Possible bioactive peptides involved in PNEC differentiation from undifferentiated epithelial precursor cells include bombesin-like peptides (BLP) such as gastrin-releasing peptide (GRP), the major known pulmonary BLP. For instance, CD10/NEP has been demonstrated to hydrolyze BLP (14, 15). In previous investigations we observed increased fetal lung cell proliferation in human fetal lung explants, and type II cell differentiation in mice in which CD10/NEP was inhibited by SCH32615; these effects were completely blocked by specific BLP receptor antagonists, supporting a major role for endogenous BLP (16, 17). Speirs and colleagues (18) observed that GRP treatment of cultured PNECs isolated from fetal rabbit lungs resulted in a significant increase in numbers of serotonin-positive PNECs but without increased thymidine labeling of PNECs, supporting the hypothesis that GRP can induce cell differentiation without mitogenesis of normal PNECs.

Higher levels of bioactive peptide growth factors such as BLP in lungs with CD10/NEP inhibition could override apoptotic signals induced by cytokines such as tumor necrosis factor and/or macrophage-stimulating protein, which are also induced by DEN/O₂ (13, 19). Multiple growth factors have been demonstrated to reverse apoptotic signals (20, 21), acting via tyrosine kinase receptors and/or via signal transduction pathways involving phosphatidylinositol-3 kinase. In the current study, we observed the greatest proportion of apoptotic PNECs within PNEC clusters partially immunostained for CGRP (with a "fading" appearance). We cannot rule out that this fading appearance represents secretion of CGRP by the PNECs. However, the same cells with weak CGRP staining were also usually positive for Apoptag fluorescence and negative or weakly positive for PGP 9.5, suggesting a loss of cell surface membrane integrity during apoptosis. Such "fading" PNECs frequently occur in clusters during spontaneous regression (9), suggesting that the regression represents PNEC apoptosis. Potential molecular mechanisms regulating PNEC apoptosis have not been explored. Most known regulators of apoptosis, such as Bcl-2, are not accessible to extracellular proteases such as CD10/NEP (21). Although intracellular proteases can promote apoptosis (22), cell-surface endopeptidases such as CD10/NEP have not been previously demonstrated to modulate apoptosis.

These data support the concept that decreased apoptosis is not the major cellular mechanism responsible for this rapid early PNEC hyperplasia. The time course of augmented PNEC hyperplasia in hamsters with CD10/NEP inhibition cannot be explained on the basis of PNEC apoptosis alone. After 4 to 8 wk of treatment, hamsters given DEN/O₂ plus SCH32615 demonstrated a 3-fold increase in PNEC numbers, compared with only a 50% increase in PNEC numbers in animals given DEN/O₂ alone. During the same 4-wk period there was only a 10% decrease in the prevalence of PNEC apoptosis in the SCH32615-treated group. It should be noted that our statistical analyses at the 4-wk time point were especially limited by the paucity of PNECs during the first month of DEN/O₂ treatment, with or without CD10/NEP inhibition. However, even at the 8- and 12-wk time points, considering the small number of animals in each experimental group, it appeared that decreased apoptosis was likely to play a contributing role. The decrease in the proportion of apoptotic PNECs with CD10/NEP inhibition was not statistically significant at 4, 8, or 12 wk.

In contrast, PNEC proliferation could play a role in promoting the early increase in PNEC numbers. On average, about twice as many PNECs immunostained for PCNA at 4 wk (although the low sample size at 4 wk and an insignificant P value at this time point must be considered). By 8 wk this trend was more significant (P = 0.10), and at 12 wk there was high statistical significance (P = 0.001). These data are consistent with our previous studies of cultured human non-neuroendocrine airway epithelial cells in which we demonstrated that cell populations with the lowest levels of CD10/NEP expression had the highest percentage of actively dividing cells (23); conversely, these cells grew more rapidly when cell surface CD10/NEP was inhibited.

In summary, the predominant cellular mechanisms for rapid early PNEC hyperplasia induced by DEN/O₂ and promoted by CD10/NEP inhibition are: first, cell differentiation (at 4 to 8 wk); second, cell proliferation (at 8 to 12 wk); and, lastly, decreased PNEC apoptosis (at 16 wk). It is of particular note that PNEC tumors were not observed in any of the hamsters in our second series, consistent with our earlier experiments (9). The molecular triggers for stimulating PNEC proliferation and differentiation, and the role of apoptosis in regulating spontaneous regression of PNEC hyperplasia, remain to be determined.