Search

Chronic mucus hypersecretion (CMH), a common feature of various obstructive pulmonary diseases, is caused by a variety of airway irritants. Bronchial mucous cell metaplasia (MCM), a histological correlate of CMH, can be induced in hamster airways by a number of different irritants. Previous studies with the hamster model suggest that the secretory cell response to different agents is not stereotyped but can vary in the type of mucus glycoconjugate produced. The present ultrastructural study was conducted therefore to provide quantitative evidence of phenotypic variation in mucous cells induced independently by exposure to the metaplastic agents elastase and acid. HPA-gold lectin cytochemistry revealed an increase in N-acetyl galactosamine at the cell surface and secretory granules of mucous cells in elastase-treated vs. acid-treated animals. Although there was no quantitative difference between the acid-treated and untreated groups, a difference in the pattern of binding within granules indicated variation in the secretory product. Because mucus glycoconjugates serve as attachment sites for specific pathogens, phenotypically distinct mucous cells may promote differential microbial colonization. In humans therefore, variation in the severity and progression of CMH may be due in part to secretory cell susceptibility and response to different pathogenic stimuli.

The bronchus is the only region of the hamster conducting airways to develop secretory cell metaplasia after an intratracheal instillation of human neutrophil elastase (HNE). We tested the hypothesis that this pathological change occurs because of cellular uptake of the enzyme that is specific to this region. HNE, dissolved in saline, was instilled into the trachea of hamsters, that were sacrificed 5, 15, 30 or 60 min later for immunocytochemical localization of the enzyme. Saline-treated animals served as controls. By light microscopy, HNE was evident only in the lumen and upon the epithelial surface in all airways, at all time points. Saline control tissues were negative. Electron microscopic immunogold staining revealed HNE within luminal macrophages and associated with mucus and, to a limited extent, upon the apical cell surface both in trachea and bronchus. A small amount of HNE staining occurred in the intercellular space and lamina propria of bronchi. Cytoplasmic gold particles were sparse both in treated and control animals. We conclude that instilled neutrophil elastase is excluded from the epithelial cytoplasm regardless of region. We thus reject the hypothesis of airway cellular uptake of HNE and suggest that stimulation of bronchial secretory cells to accumulate mucin granules is initiated at the cell surface, possibly by unmasking or altering region-specific receptors involved in signal transduction pathways governing mucin granule synthesis.

A single intratracheal instillation of human neutrophil elastase (HNE) into hamsters causes granule discharge from bronchial secretory cells followed by marked accumulation of granules, visible by light microscopy at 21 days and persisting through 18 months. To determine whether persistence of this secretory cell metaplasia (SCM) is due to inability of the metaplastic secretory cells to secrete their granules, hamsters having HNE-induced SCM were challenged with the potent secretagogue HNE. Four groups of 10 hamsters each received 300 micrograms HNE intratracheally. Twenty-one days later, hamsters were intratracheally treated with HNE or saline; the groups were designated HNE-HNE and HNE-SAL, respectively. Hamsters were killed 2 h or 21 days following the second treatment. Using light microscopy, nucleated epithelial cells were counted in plastic sections of the left main intrapulmonary bronchus. Cells were classified as ciliated (C), basal (B), indeterminate (IN), or secretory. Secretory cells were subcategorized as S0 (0 granules), S1 (1-4 granules), S2 (> or = 5 granules with intervening cytoplasm), and S3 (abundant granules completely filling the cytoplasm). At 2 h, S3 cell frequency in the HNE-HNE group was 13.0 +/- 2.2 (% mean +/- SE), significantly lower than in the 2 h HNE-SAL group (31.1 +/- 4.5). Concomitantly, higher cell frequencies were seen in the other secretory categories of the HNE-HNE group compared to the HNE-SAL group; S2 17.1 +/- 1.9 compared to 9.4 +/- 1.9, S1 2.4 +/- 0.4 compared to 1.1 +/- 0.5, and S0 2.4 +/- 0.5 compared to 1.1 +/- 0.5, respectively. The S3 cell frequency of the 21-day HNE-HNE group was 25.4 +/- 4.7, increased significantly compared to the 2 h HNE-HNE group; this change was concomitant with significant decrease in the frequency of the S0 secretory cells. Cell frequencies of C, B, and IN were not affected by treatment or time. It is concluded that metaplastic secretory cells discharge their granules in response to HNE; SCM returns to its original state after HNE rechallenge; persistent SCM is not due to the inability of metaplastic secretory cells to discharge their granules.

Alpha-1-protease inhibitor is susceptible to oxidative impairment by the neutrophil myeloperoxidase (MPO) system. The purpose of this study was to assess the effect of the MPO oxidant system on elastase-induced emphysema in the hamster. Intratracheal instillation of 200 micrograms of human neutrophil elastase (HNE) induced a significant secretory cell metaplasia (SCM) and airspace enlargement [23% increase in mean linear intercept (MLI) as compared with control values]. Instillation of MPO system components [0.6 international units (U) of MPO, 5.5 U of glucose oxidase and glucose (0.02 M)] along with 200 micrograms HNE failed to enhance the severity of the SCM or emphysema induced by HNE alone. A second experiment was carried out using 50 micrograms of porcine pancreatic elastase (PPE) to induce emphysema. PPE produced a significant 45% increase in MLI, but the MPO system combined with PPE failed to enhance the emphysema induced by PPE alone. The MPO system alone had no measurable effect on airspace size or SCM. In vitro studies showed that PPE was partially inactivated by the MPO system; a 56% loss of elastolytic activity occurred during a 6-min incubation of PPE with the MPO system. This may explain why the MPO system did not exacerbate PPE-induced injury, but it does not explain the lack of enhancement for HNE. A 6-minute incubation of HNE with the MPO system resulted in a nonsignificant 10% decrease of elastolytic activity.(ABSTRACT TRUNCATED AT 250 WORDS)

Hamster airway epithelial secretory cells were investigated by light and electron microscopic cytochemistry to study possible changes in their carbohydrate content induced by human neutrophil elastase (HNE), an agent known to cause replacement of Clara cells by mucous cells in hamster bronchi. Characterization of secretory cell carbohydrates by the AB/PAS, PA-TCH-SP, HID-TCH-SP, and LID-TCH-SP sequences indicated the existence of periodate-reactive acidic glycoconjugates, but the absence of sulfated or carboxylated glycoconjugates in both treated and control animals. Differences were seen in the quality and quantity of historeactive carbohydrates throughout various regions in the lower respiratory tract. This was especially evident in the HNE-treated animals. It is concluded that the HNE-induced expression of the mucous cell phenotype is associated with an increase in the amount of neutral and acidic nonsulfated and noncarboxylated polysaccharides stored in the secretory granules of these cells.

An intratracheal instillation of human neutrophil elastase (HNE) causes accumulation of an excess number of secretory granules in the epithelial secretory cells lining the hamster bronchus. This chronic lesion, which we refer to as secretory cell metaplasia (SCM), is not seen in the trachea or bronchioles. Because luminal cell surface lectin binding is much higher in the trachea than in the bronchus, we concluded that tracheal resistance may be due to a protective glycoconjugate coat. In the present ultrastructural study, we analyzed the lectin-binding capability of bronchiolar epithelial cells to determine whether their luminal cell surface glycoconjugate layer is similar to tracheal epithelial cells. None of the six ferritin-conjugated lectins showed higher binding in bronchioles compared to the bronchus, suggesting that a high level of surface oligosaccharides is not necessary for resistance to the metaplastic effects of HNE. HNE caused a significant reduction in bronchiolar surface binding of the gold-labeled, secretory cell-specific lectin, Helix pomatia agglutinin. The principal granulated secretory cell type in bronchioles was ultrastructurally similar to a form of bronchial Clara cell that converts to a mucous cell phenotype in response to HNE. The results suggest that absence of bronchiolar SCM is not attributable to a protective layer of cell surface oligosaccharides, a lack of cellular contact by HNE, or the presence of a morphologically distinct population of epithelial cells in bronchioles.

Hamsters exposed to an intratracheal instillation of human neutrophil elastase (HNE) accumulate an abnormally high number of secretory granules in bronchial but not tracheal epithelial cells. We employed lectin cytochemistry to investigate possible differences in the epithelial cell surface glycoconjugate layer in trachea compared to bronchus which might explain the regional dissimilarity in response to HNE. Portions of glutaraldehyde-fixed trachea and bronchi were incubated in one of several ferritin-labeled lectins prior to embedding for transmission electron microscopy. Lectins from Ricinus communis, Helix pomatia, and Triticum vulgaris bound to the surface of tracheal secretory cells in moderate to profuse amounts, while most bronchial secretory cells showed little or no label with these lectins. Gold-labeled Helix pomatia agglutinin (HPA), a lectin specific for secretory cells, showed a decrease in surface binding to all tracheal secretory cell types within 2 h of HNE instillation, compared to saline controls. In contrast, the majority of bronchial secretory cells showed an HNE-induced increase in surface label from extremely low levels in saline controls. The low levels of lectin binding to bronchial cells, in contrast to the trachea, may indicate the lack of a protective surface glycoconjugate coat, thus explaining the vulnerability of these cells to HNE. The rise in number of accessible HPA binding sites on the surface of bronchial secretory cells exposed to HNE may represent an important event in the pathologic accumulation of secretory granules by these cells.

A study was undertaken to determine whether emphysema and airway secretory cell metaplasia, induced in hamsters by intratracheal treatment with human neutrophil elastase (HNE), could be moderated by pretreatment with human alpha 1-protease inhibitor (API). API (4.9 mg) was given intratracheally to hamsters 1 h before 0.3 mg HNE. Eight weeks later, lung volumes and pressure-volume relationships were measured in the anaesthetized animals. Mean linear intercepts and secretory cell indices were measured in lung sections. API given 1 h before HNE moderated the development of bronchial secretory cell metaplasia. The severity of emphysema was reduced by 75%. Clearance studies indicated that 80% of the functional activity of instilled API could be lavaged from the lungs after 1 h, indicating a 4 h half-life in the lavageable compartment of the lungs. We calculate that for 50% protection from emphysema the molar ratio of lavageable API to HNE at the time of HNE instillation was 4.8 as compared with 0.78 for 50% inhibition of elastolytic activity in vitro, indicating that API is only 16% as efficient in vivo as compared with its in vitro HNE inhibitory effectiveness. Nevertheless, we conclude that human API given intratracheally is efficacious against HNE-induced emphysema and secretory cell metaplasia.

Studies were undertaken to evaluate the in vitro properties of recombinant human secretory leukocyte-protease inhibitor (rSLPI) that had been made in Escherichia coli in an inactive form and refolded, and to determine whether emphysema and bronchial secretory cell metaplasia, induced in hamsters by intratracheal treatment with human neutrophil elastase (HNE), could be amelio-rated by prior intratracheal instillation of rSLPI. Chromatographic studies indicated that 3H-rSLPI formed a 1:1 complex with HNE. Blockage of the active site of HNE by a covalently bound tetrapeptide chloromethyl ketone reduced complex formation with 3H-rSLPI by more than 98%. Incubation of 3H-rSLPI-HNE complex with alpha 1-protease inhibitor for 3 hours at 37 degrees C decreased the amount of complex compared with incubation in the presence of bovine serum albumin (70% vs 27% dissociated). The calculated dissociation rate constant was 1.1 x 10(-4) sec-1, indicating a 1.8 hour dissociation half-life. Dissociated 3H-rSLPI retained its ability to recombine with HNE. rSLPI was as effective at inhibiting HNE released from stimulated neutrophils as 3H-rSLPI was at inhibiting purified HNE. Intratracheal pretreatment of hamsters with 3000 micrograms of rSLPI as long as 8 hours before the intratracheal instillation of 250 micrograms of HNE, resulted in significant protection against induction of emphysema and secretory cell metaplasia. One and 4 hours after instillation of rSLPI, 59% and 44%, respectively, of the initial functional activity was recovered in lung lavage supernatant, indicating a half-life of approximately 2 hours.(ABSTRACT TRUNCATED AT 250 WORDS)

The authors wished to determine whether secretory cell metaplasia (SCM) induced in the bronchi of hamsters by human neutrophil elastase (HNE) was enzymatically mediated. We also wished to determine whether SCM could be induced by a proteolytic enzyme devoid of elastolytic activity. Accordingly, groups of weight-matched hamsters were given a single intratracheal instillation of 0.5 ml of saline solution containing one of the following: 300 micrograms of HNE purified from blood neutrophils, n = 14; 300 micrograms of HNE inactivated with Suc-Ala-Ala-Pro-Val chloromethyl ketone (CMK), n = 10; 500 micrograms of porcine pancreatic trypsin treated with CMK to eliminate residual active elastase, n = 10; 500 micrograms of trypsin inactivated by tosyl lysine chloromethyl ketone, n = 10; 2 micrograms CMK, n = 10; and saline alone, n = 10. Seven untreated animals served as additional controls. Twenty-one days post treatment, 5-6 micron paraffin-embedded sections, from the left lung hilar region, stained by Alcian blue and periodic acid-Schiff reaction were graded on a five-point scale for determination of the secretory cell index, which reflects SCM. Both elastase and trypsin produced severe SCM: mean +/- SEM secretory cell indices were 2.96 +/- 0.11 and 2.72 +/- 0.19, respectively, compared with values of 0.90 +/- 0.35 for the untreated group and 0.93 +/- 0.46 for the saline group (p less than .05). The secretory cell indices of the groups treated with inactivated elastase or trypsin were comparable to those of the saline-treated and untreated groups.(ABSTRACT TRUNCATED AT 250 WORDS)

A single intratracheal instillation of pancreatic elastase in hamsters induces a lesion resembling human panacinar emphysema. This paper reports the occurrence of irreversible goblet cell metaplasia in the bronchial epithelium of hamsters similarly exposed to elastase. The goblet cell change was dose related; a dose of 0.1 mg/100 g body wt or less at 16 days, produced slight or moderate goblet cell metaplasia in fewer than half the animals, whereas 84% of animals treated with a dose between 0.2 and 0.5 mg/100 g body wt developed goblet cell metaplastic lesions, more than half of which were considered to be severe. The percentage of goblet cells in the epithelium of elastase-treated hamsters (32.5) was significantly higher (P less than 0.005) than that of unexposed (12.2) and saline-exposed controls (18.7). All hamsters examined 6 and 12 mo after elastase treatment showed the lesion. The pathogenesis of the changes is unclear but the possibility is raised that the bronchial changes may be due to disturbance of an endogenous protease-antiprotease system. The findings suggest the hypothesis that, under appropriate circumstances, a single pulmonary insult in man could lead to a permanent lung injury demonstrating the anatomic lesions of both chronic bronchitis and panacinar emphysema.

To determine whether purified human neutrophil cathepsin G (Cat-G) can act by itself or in concert with purified human neutrophil elastase (HNE) in the induction of emphysema and bronchial secretory cell metaplasia (SCM), we gave golden Syrian hamsters 100 micrograms of HNE alone or in combination with either 100 or 200 micrograms of Cat-G. Other groups of animals received intratracheal doses of up to 600 micrograms of Cat-G alone. The severity of emphysema was determined from measurements of lung volumes, compliance, forced expiratory flow, and the mean linear intercept. The severity of SCM in the main airways was graded on sections stained by the alcian blue and periodic acid-Schiff reaction. The Cat-G was a weak inducer of SCM; significant SCM was produced by 400 and 600 micrograms but not by 100 or 200 micrograms or 200 micrograms of Cat-G. The Cat-G (100 and 200 micrograms) did not potentiate the SCM induced by 100 micrograms of HNE. The Cat-G alone did not produce emphysema, and neither 100 nor 200 micrograms of Cat-G potentiated the mild emphysema induced by 100 micrograms of HNE. These results were not consonant with a report that Cat-G and HNE were synergistic in solubilizing human lung elastin. We therefore measured the ability of Cat-G and HNE to solubilize several radiolabeled elastins. The combination of Cat-G and HNE did not solubilize significantly more hamster lung elastin (23%) than the sum of their individual activities.(ABSTRACT TRUNCATED AT 250 WORDS)

To explore the time course and the mechanism of development of bronchial secretory cell metaplasia (SCM) induced by human neutrophil elastase (HNE), anesthetized hamsters were injected intratracheally with 300 micrograms highly purified HNE in 0.5 ml saline solution; saline-injected and untreated animals served as controls. At 3, 8, 16, and 21 days after treatment, animals were killed and their lungs fixed by vascular perfusion. Samples from the hilar region of the left lung, containing the main axial airway and its proximal branches, were embedded in Epon-Araldite, and 1 micron sections were stained with methylene blue. Epithelial cells with a luminal border were categorized into three cell types and expressed as a percent of total cells counted (mean 1900 per animal); cells containing at least three mucin granules were classified as secretory, ciliated cells displayed cilia or basal bodies, and cells with none of these characteristics were classified as indeterminate. Percentages of the three cell types in saline-treated animals, at all time points, were comparable to those in the untreated controls. With HNE treatment the secretory cell percentages were higher at 16 days (mean +/- SEM, 36.4% +/- 3.2%) and at 21 days (35.7% +/- 2.9%) than in the untreated animals (18.2% +/- 1.8%, P less than 0.05). The percentage of the indeterminate cells in the HNE group was decreased at days 8, 16, and 21 (7.2% +/- 1.6%, 5.0% +/- 1.4%, and 8.2% +/- 2.5%, respectively) compared with that in the untreated group (21.7% +/- 2.5%, P less than 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)

Purified human neutrophil elastase (HNE) and a crude extract of human neutrophils (EXT) were administered intratracheally to hamsters, and their effects were determined 21 and 56 days later by measurement of lung statics and dynamics and by light microscopy and morphometry of the lungs. A dose of 450 micrograms of HNE (HNE 450) produced focal emphysema of moderate severity. The combination of HNE 450 and EXT (HNE 450 + EXT) produced no more severe emphysema than HNE 450 alone; EXT alone did not produce emphysema. The HNE 450, HNE 450 + EXT, and EXT all produced persistent secretory cell metaplasia in large intrapulmonary airways. In a second experiment, 40 micrograms of HNE (elastolytic activity equivalent to that in EXT), designated HNE 40, did not produce secretory cell metaplasia. Neither EXT nor EXT treated with the elastase inhibitor suc-ala-ala-pro-val chloromethyl ketone (EXT + CMK), which had a residual elastolytic activity equivalent to 15 micrograms of HNE, produced emphysema; both preparations caused bronchial secretory cell metaplasia 21 days after treatment. Values of maximal expiratory flow, measured 21 days after treatment, were reduced at points between 20 and 50% of vital capacity for HNE 450 and between 30 and 80% of vital capacity for HNE 450 + EXT; maximal expiratory flows were not significantly different from saline controls for HNE 40, EXT, or EXT + CMK.(ABSTRACT TRUNCATED AT 250 WORDS)

Human neutrophil elastase (HNE) causes secretory granule discharge and conversion of many Clara cells to mucous cells in hamster bronchi. We investigated whether the trachea responds to HNE in a similar manner because of its abundance of Clara cells. By light microscopy, the tracheal epithelium of animals exposed to a single intratracheal injection of HNE was normal at 21 days, although bronchial secretory-cell metaplasia (SCM) was present. An ultrastructural differential cell count showed no increase in the proportion of granulated secretory cells in HNE-treated animals at 8 and 21 days postinjection compared to saline or untreated controls. At 2 h, the percentage of granulated secretory cells was lower and that of granulated secretory cells was higher in HNE-treated animals than in controls. The HNE-treated animals had fewer secretory granules per cell profile and more surface undulation than controls. By 1 day, the differential cell count and number of granules per cell profile were normal. Saline did not affect the differential cell count or granule number at any time. Ultrastructural study of untreated trachea disclosed the same three types of Clara cell that are found in the bronchus, but their frequencies, with one exception, are significantly different in the two regions. We conclude that HNE acts as a secretagogue in both trachea and bronchus but that an amount of enzyme sufficient to cause bronchial SCM does not induce a similar lesion in trachea. Heterogeneity of Clara cell types in hamster airways may explain the regional variation in secretory-cell modulation by HNE.

The central intrapulmonary bronchi of hamsters were examined by transmission electron microscopy at varying times following intratracheal instillation of human neutrophil elastase (HNE) or its vehicle, saline. Two hours after HNE treatment, there was a marked irregularity of the surfaces of many nonciliated epithelial cells; a differential count of transepithelial cells (those with both a basal lamina and luminal border) demonstrated a significant decrease in the proportion of granule-containing (granulated) secretory cells and a corresponding increase in nongranulated secretory cells. By 3 days after HNE injection, the differential count had returned to control levels and cell surface alterations were less evident. By 8 days, the proportion of granulated secretory cells had significantly increased, while that of nongranulated secretory cells had decreased. Many Clara cells developed the characteristics of mucous cells so that mucous cells constituted 57% of the secretory cells compared to 14% for the saline controls. The mucous cells contained an increased number of mucous granules including bizarre forms never seen in controls. By day 16, the average mucous cell proportion had increased to 75%; the mucous cells were larger and contained many more secretory granules than at day 8. At no time was there evidence of overt cell injury or alteration of extracellular connective tissue due to HNE. Basal and pseudobasal cells, distinguished by the presence or absence of hemidesmosomes, did not change as a percentage of total nucleated epithelial cells. Saline had no effect on the differential cell count compared to untreated values. Our results indicate a strong likelihood that HNE causes early discharge of secretory granules and alters the phenotypic expression of Clara cells so that they produce abundant, often abnormal mucous granules. The mechanism of HNE-induced disturbance of epithelial homeostasis is unknown, but the early irregularity of nonciliated epithelial cell surfaces may signify an important event in the evolution of the resultant lesion.

It is known that human neutrophil elastase (HNE) treatment of hamster tracheal explants causes the release of glycoconjugates, most of which appear to have the characteristics of mucus glycoproteins. This study was designed to determine the origin of HNE-induced glycoconjugate release from 1-day-old cultures of adult hamster trachea. After confirming that HNE treatment released glycoconjugates from cultures labeled with tritiated glucosamine, light microscopic autoradiograms and electron micrographs were prepared. Untreated cultures and cultures treated with inactivated HNE served as controls. HNE treatment caused a 40 to 50% decrease in the silver grain count on the external apical surfaces of secretory cells (p less than 0.05) and ciliated cells (p less than 0.01). Silver grain counts in secretory and ciliated cell cytoplasm, submucosa, and nontissue background were not significantly different from controls. The percentage of nongranulated secretory cells and the number of secretory granules in granulated secretory cells were similar in the HNE-treated and untreated controls. There was no evidence of constitutive release of radiolabeled glycoproteins, or of discharge of secretory granules from the secretory cells. We conclude that HNE releases mucins and other glycoconjugates from the external surfaces of both secretory and ciliated cells in tracheal organ culture.

A study was made of the evolution of emphysema and airway injury induced in the lungs of male golden Syrian hamsters by a single intratracheal injection of 350 micrograms human neutrophil elastase (HNE). Saline control and HNE-treated groups of 8 animals were studied 1, 3, 6, 12, and 18 months posttreatment. HNE treatment caused a significant increase in all lung volumes and a significant decrease in maximum expiratory flows at all study times. The mean linear intercept (MLI) values of the left lung were significantly increased over control values. There was no progression with time in MLI values, lung volumes, or lung compliance. Secretory-cell metaplasia was present at 1 month and persisted throughout the study. The HNE-treated lungs showed clusters of ferric iron-containing macrophages in the terminal airspaces. The amount of iron in the lungs, determined morphometrically, was greatest at 1 month, was decreased by 6 months, and then did not change further to 18 months. At 18 months the amount of iron was still significantly above control amounts. We conclude that the airway and parenchymal lesions induced by HNE persist without progression for 18 months. Clearance of ferric iron, which was probably a result of the hemorrhage induced by HNE treatment, continued for 6 months with no evident subsequent clearance.

An ultrastructural morphometric analysis of bronchial secretory cells was carried out on hamsters treated intratracheally with 300 micrograms of human neutrophil elastase (HNE) in 0.5 ml saline, saline alone, or left untreated. Five to 6 animals were killed at 2 h and at 3, 8, and 16 days after treatment. Electron micrographs were prepared from the hilar region of the left main intrapulmonary airway; 125 +/- 16 (mean +/- SE) granulated secretory cells extending from basement membrane to lumen were analyzed for each group. The number (Ng) and area (Ag) of granular profiles per cell, the area of cell profiles (Ac), and the volume density of secretory granules per cell (Vv) were determined using an electronic image analyzer. There were significant decreases in Ng, mean Ag, and Vv in the 2-h HNE group when compared with the saline group. Values of Ng, mean Ag, and Vv were similar for HNE and saline groups at 3 days, but were significantly increased at 8 and 16 days. The Ac of HNE-treated groups was similar to their saline control groups at all time points except at 16 days when the HNE-treated group enlarged to double that of its saline control group. An ultrastructural differential cell count showed a decrease in frequency of granulated secretory cells at 2 h and an increase at 8 and 16 days; there was an inverse change in the frequency of nongranulated secretory cells at these times. The proportion of ciliated, preciliated, and indeterminate cells remained constant over time in all treatment groups.(ABSTRACT TRUNCATED AT 250 WORDS)