Basal-like Cells Constitute the Proliferating Cell Population in Cystic Fibrosis Airways

Cystic fibrosis (CF) airways are chronically and recurrently exposed to inflammatory and infectious mediators that injure the superficial epithelium (1). Proteases and oxidants induce epithelial sloughing and activate epithelial repair responses. Immediately after injury, epithelial migration, proliferation, and then differentiation occur (2,3). Although increased epithelial proliferation has been reported in CF (4, 5), characteristics of the proliferating cell populations have not been well characterized.

Recovery of the injured airway has been studied predominantly in rodent models and in vitro using human airway epithelial cells. After injury, the tracheobronchial airways are maintained by several potential progenitor cells including basal cells (6, 7) and nonciliated secretory cells (8, 9). Proliferation of progenitor cells is associated with activation of at least two growth factor receptors, epidermal growth factor receptor (EGFR) (10-13) and ErbB2 (14, 15). In this article, we sought to determine if airway epithelial cells from CF lung sections had the same characteristics during repair as the well-characterized in vivo and in vitro models of airway repair after injury. We used immunohistochemical analyses of airway sections from CF lung transplant patients and control patients without airway disease to determine whether proliferating cells had basal cell characteristics and whether EGFR or ErbB2 were expressed in these cells.

METHODS

Subjects

Airway sections from six subjects with CF who underwent lung transplantation and from six subjects without chronic airways diseases who underwent lobar resections were included in the analysis. Ages and diagnoses of subjects are included in Table 1. Airway sections were fixed in 3.7% neutral-buffered formalin, paraffin embedded, and 5-µm thick tissue sections were placed on either Superfrost/Plus or ProbcOn Plus microscope slides (Fisher Scientific; Pittsburgh, PA) and dried at room temperature overnight, followed by 20 min in a convection oven (70°C). Slides were deparaffinized in xylene, rehydrated in graded alcohols, and rinsed in distilled water. Slides from each subject were stained with hematoxylin and eosin and Alcian blue/periodic acid-Schiff for histologic evaluation. Sections containing cartilaginous airways were selected for analysis. The protocol was approved by the Institutional Review Board for Clinical Investigations, Duke University Medical Center.

Immunohistochemistry

Methods for antigen retrieval and primary antibody detection varied slightly for each primary antibody. Negative controls employed nonimmune IgG (mouse IgG; Southern Biotechnology Associates, Birmingham, AL, for Ki-67, cytokeratins 5/14, and EGFR immunohistochemistry; and rabbit pre-immune serum for ErbB2 immunohistochemistry). Immunohistochemistry with nonimmune antibodies in place of primary antibodies resulted in no staining (results not shown). Immunohistochemistry was performed for Ki-67, cytokeratins 5/14, EGFR, and ErbB2. Ki-67 nuclear antigen is expressed in proliferating cells and expressed throughout the cell cycle (G^sub 1^, S, G^sub 2^, and M phases) but not in resting cells (G^sub 0^). We evaluated Ki-67 expression in both highly stained areas "proliferative zones" and random fields of airway epithelium. Staining for cytokeratins 5/14 was used to identify basal and parabasal cells of the human airway epithelium (6). Primary antibodies and antigen retrieval procedures for immunohistochemistry are summarized in Table 2. A full description of the immunohistochemistry protocols is provided in an online supplement.

Statistics

Data are presented as mean ± SEM. Comparison of CF and control patients for MIB-I staining, goblet cell number, and epithelial nuclear number and cell height were performed by Wilcoxon rank-sum test (16). Analyses were performed using Statistix 8.0 (Analytical Software, Tallahassee, FL). Differences were considered significant for p < 0.05.

RESULTS

Histology of CF Versus Control Airway Sections

Airway sections for subjects with CF and control subjects were evaluated by histology to determine whether superficial epithelia were intact and to evaluate sections for goblet cell hyperplasia (Figure 1). The CF sections were heterogeneous including regions of epithelial sloughing, hyperplastic cuboidal cells, and goblet cell hyperplasia (CF sections had 0.16 ± 0.01 goblet cells per 100 µm of basement membrane versus control sections that had 0.09 ± 0.03 goblet cells per 100 µm, p = 0.09). Interestingly, CF superficial epithelial cells were hypertrophied compared with control cells (Figure 1) manifest by decreased number of nuclei present per 100 µm of basement membrane due to larger cell size (CF 0.77 ± 0.05; control 1.58 ± 0.14; p = 0.002), and increased height of the pseudostratified columnar epithelia measured from the basement membrane to the tips of columnar cells (CF 63 ± 2.3 µm; control 37 ± 2.2 µm; = 0.002).

Proliferation Index of CF Versus Control Airway Sections

We examined epithelial proliferation by immunostaining for Ki-67 (MIB-1 clone), a nuclear proliferation antigen (Figures 1C and 1D), and by counting the percentage of cells staining positively using image analysis. Immunohistochemistry revealed that in the pseudostratified columnar epithelia, MIB-1 staining was restricted to the nuclei located adjacent to the basement membrane, the basal cell compartment. There was no MIB-1 staining in nuclei of goblet or ciliated cells. Importantly, there was increased MIB-1 staining in the CF sections. Analysis of random fields revealed that CF sections had a significantly higher percentage of total nuclei staining for MIB-1 (25.1 ± 2.1% [mean ± SEM]) than control cells (4.6 ± 0.9%, p = 0.002, Figure 1E). Moreover, only CF airways contained MIB-1 staining clusters of proliferating cells that extended through the full thickness of the epithelium.

The foci of proliferating cells (Figure 2A) were localized adjacent to regions of pseudostratified columnar epithelium and appeared as hyperplastic regions of cuboidal cells with strong MIB-1 nuclear staining (Figure 2B). The MIB-1 staining in the proliferative zones appeared to be contiguous with MIB-1 staining in basal cells in adjacent regions. Each CF airway contained an average of 4.8 ± 0.7 proliferative zones (mean ± SEM; two airways evaluated on average per CF patient). The length of these proliferative zones averaged 540 ± 55 µm (range, 103-2,468 µm); the width of the proliferative zones was similar to that of the rest of the epithelium (i.e., 63 ± 2.3 µm [mean ± SEM]). We speculate that these hyperplastic cuboidal cells may represent proliferating progenitor cells before differentiation during epithelial repair.

Cytokeratins 5/14 Immunostaining

To determine whether MIB-1 proliferating cells had characteristics consistent with basal-like cells, we performed immunostaining for cytokeratins 5/14 (Figure 3), a marker for basal cells (6, 17). In the pseudostratified columnar epithelia for both control and CF sections, cytokeratins 5/14 immunostaining was restricted to the basal cell compartment (Figures 3A and 3B). We found a direct correlation between cytokeratins 5/14 immunostaining and MIB-1 immunostaining in the hyperplastic cuboidal cells (Figure 3C).

EGFR and ErbB2 Immunostaining

Immunostaining was performed for both EGFR and ErbB2 to determine whether the proliferating population of cells appeared to express one or both of these growth factor receptors. In pseudostratified columnar epithelia of both control and CF sections, EGFR localization was restricted to basal cells (Figures 4A and 4B), whereas ErbB2 expression extended throughout the columnar epithelia, but often was most intense at the luminal surface (Figures 5A and 5B). In contrast, in the CF proliferative zone, the cuboidal cells predominantly expressed EGFR (Figure 4C), whereas ErbB2 expression was limited to a few differentiated ciliated cells at the most luminal aspect of the section (Figure 5C). These results suggest that EGFR and ErbB2 may be differentially expressed in proliferating versus terminally differentiated cells.

DISCUSSION

We used a monoclonal antibody to detect Ki-67/MIB-1, a nuclear antigen associated with the G^sub 1^, S, G^sub 2^, and M phases of the cell cycle to determine whether cells were proliferating (18). Our findings confirm that CF bronchial epithelial cells have increased proliferation indices compared with control cells. MIB-1 staining was restricted to basal cells in the pseudostratified columnar epithelia of normal airways. Our results are consistent with a study of cell proliferation in the bronchial mucosa of patients with asthma and patients with chronic bronchitis, which also observed that immunostaining for another proliferation marker, proliferating cell nuclear antigen, was restricted to basal cells and absent from ciliated and secretory epithelia (19). Interestingly, in addition to greater basal cell proliferation in the pseudostratified columnar epithelia, we observed hyperplastic zones of proliferating cuboidal epithelial cells that appeared to be in transition between regions of differentiated pseudostratified columnar epithelia. Hyperplastic domains of cells in the superficial airway epithelia have been observed in a previous study, but were not characterized (4). We sought to further characterize these hyperplastic cells by immunohistochemical analysis.

The first question we addressed was whether or not the hyperplastic cells expressed cytokeratins 5/14, a property characteristic of basal and parabasal cells of human airway epithelium (6, 20). Our results demonstrated that cytokeratins 5/14 expression was restricted to basal cells and to the zones of hyperplastic cells. Interestingly, in a previous study of rat tracheal epithelial cells during engraftment of a denuded trachea, populations of predominantly ciliated/secretory cells lost expression of characteristic proteins, and became ultrastructurally indistinguishable from each other, highly proliferative, and positive for cytokeratin 14 expression (21). Alternatively, the presence of cytokeratins 5/14 suggests that the origin of the "proliferative zone" cells may be a subpopulation of basal cells that comprise a proliferating population, a transit-amplifying clone. Importantly, our results support the hypothesis that, after airway injury in CF, basal-like cells constitute a source of epithelial regeneration. Because of the similar phenotype of proliferating epithelial cells between CF airway sections and injured rat and mouse airway sections, our results support the use of in vivo rodent models of airway injury to study the molecular mechanisms of epithelial repair and regeneration.

Increased epithelial proliferation is likely a response to injury induced by bacterial exposure, neutrophil-dominant inflammation, and other toxic mediators from viral infections, cigarette smoke, and air pollutants. These injurious substances have been reported to activate EGFR through direct ligand-mediated activation (13, 22, 23), or via transactivation (24). Because EGFR and ErbB2 receptor tyrosine kinases are associated with epithelial repair and proliferation, the second question we addressed was whether or not the proliferating cell population in CF airways expressed EGFR or ErbB2. We observed that in the pseudostratified columnar epithelia, EGFR expression was limited to basal cells. However, in the proliferative zones, there was an expanded zone of cells expressing EGFR that correlated with increased MIB-1 staining in these cells. Hardie and colleagues (5) reported that although CF bronchial epithelia did not differ in EGFR expression from control tissues, CF epithelia had more intense staining for transforming growth factor a, an EGFR ligand, than seen in the control airway sections. Their results support a different mechanism of increased EGFR activation in the lungs of patients with CF: upregulation of EGFR ligands. Consistent with our results, increased expression of EGFR has been reported in bronchopulmonary dysplasia (25), injured human bronchial epithelia (26), and asthmatic bronchial epithelia (27, 28). Whether EGFR or EGFR ligands are upregulated, these reports indicate a strong correlation between inflammatory airway diseases, EGFR regulation, and epithelial repair after injury. This suggests that caution should be exercised in planning strategies to reduce EGFR activity as a mechanism to reduce goblet cell metaplasia (29) because EGFR expression and activation appears to be a critical mediator in the early processes of epithelial repair.

Although activation of members of the ErbB receptor tyrosine kinase family is required for epithelial repair after injury, specific roles for this process have not been assigned to specific receptors. We observed increased EGFR glycoprotein and diminished ErbB2 glycoprotein in the zone of hyperplastic cuboidal cells. Our findings are consistent with in vivo and in vitro reports implicating EGFR expression and activation during epithelial proliferation. For example, following naphthalene-induced mouse lung injury, increased EGFR expression is localized to the site of proliferating epithelial cells (11). Normal human bronchial epithelial cells proliferate in response to autocrine ligand activation of EGFR (10) or interleukin-13-induced EGFR activation by transforming growth factor a release (12). However, in addition to these reports of EGFR-mediated epithelial proliferation, ErbB2 activation may also contribute to epithelial proliferation. ErbB2 is required for fetal lung cell proliferation (15) and for human bronchial epithelial cell repair after mechanical injury (14). Therefore, although ErbB2 does not appear to colocalize with EGFR in the foci of proliferating cells, we cannot rule out an important role for ErbB2 in proliferation or another component of the process of epithelial repair following injury.

Interestingly, in our study, ErbB2 was localized throughout the pseudostratified epithelium but the intensity of staining tended to be greatest at the luminal aspect of airway sections, whereas EGFR was localized to the basal cell compartment. Other investigators have localized ErbB2 to the same cellular compartment as EGFR (14,30). The differences in immunohistochemical localization in the pseudostratified columnar epithelia may reflect recognition of different antigens as immunohistochemistry with different anti-ErbB2 antibodies results in different localization of ErbB2 in the rat female reproductive tract (31). At the subcellular level, ErbB2 staining appears to be predominantly basolateral, consistent with previous reports of ErbB2 expression in polarized epithelial cells (14, 32). Localization of ErbB2 may be due to differential expression of other proteins that interact with ErbB2 such as Lin-7 that anchors ErbB2 to the basolateral surface (33) or MUC4 that translocates ErbB2 to the apical surface (32). Differences in ErbB2 localization and expression level may affect its activation and role in epithelial repair (14).

In the neonatal period, CF airways appear normal with distension of submucosal gland ducts with mucus as the only reported abnormality (34). This suggests that CF airway pathology is not the result of absence of cystic fibrosis transmembrane conductance regulator (CFTR) function alone, but likely related to early onset of infection and inflammation. CF airway pathologic changes progress during infancy and early childhood and are characterized by mucus plugging, inflammatory infiltrates, and epithelial metaplasia (1). The airway surface is marked by goblet cell hyperplasia, squamous metaplasia, and cuboidal epithelium (4). Consistent with a previous report (35), we also describe epithelial hypertrophy in the CF airway. Our photomicrographs capture the CF airways in different phases of epithelial response to injury, and highlight the patchy nature of epithelial processes within the same airway. It is not known whether secretory cells originate from other terminally differentiated cells, or derive from proliferation and differentiation of a transit-amplifying cell compartment (36). Our observations favor a mechanism of repair requiring epithelial proliferation to create goblet cell hyperplasia.

Because obstruction of airways from mucus plugging is a significant cause of morbidity and mortality in CF, the pathogenesis of goblet cell hyperplasia is an area of intense investigation. The fate of the proliferating cell population and determinants of ciliary versus secretory differentiation are still not known. The mechanism of activation or regulation of ErbB receptors and their role in airway epithelial remodeling are still not well understood. These questions are critical areas for research so that therapies can be designed to promote normal epithelial repair and retard goblet cell and submucosal gland hyperplasia.

Conflict of Interest Statement: None of the authors have a financial relationship with a commercial entity that has an interest in the subject of this manuscript.

© 2005 American Thoracic Society Provided by ProQuest LLC. All Rights Reserved.