To study the localisation of human WFDC2 protein we used a monoclonal antibody raised against recombinant human WFDC2 [19
]. This antibody has previously been used for both immunohistochemical analysis of WFDC2 in ovarian cancers as well as in ELISAs [19
]. The antibody was also shown to specifically identify human WFDC2 in western blots of in vitro
translation reactions (results not shown).
In the light of our previous observation that significant expression of WFDC2 RNA was seen in tissue from the nasal passages [21
], we initially stained sections from the maxillary sinus (antral) mucosa and nasal polyps. In these sections there was significant staining in the ductular epithelium of the minor glands, with less pronounced staining in the mucous cells of the glands themselves (Figure ). The surface epithelium of the nasal passages was also weakly positive. We stained nasal polyps from >15 individuals all of which demonstrated WFDC2 in ductular cells of the minor mucous glands (Figure ), although the intensity of staining varied and some regions of individual glands were found to be negative. The respiratory epithelium overlying the polyps was found to stain in a similar manner to the maxillary sinus. We then went on to examine pulmonary tissue and in sections from central bronchi the most prominent staining was again seen in sub-mucosal glands (Figure ) and was restricted to what phenotypically appear to be serous cells within the demilunes (Figure ). Cells within the surface bronchial epithelium from both major and minor airways were also stained (Figure ), although in the 10 samples studied both the intensity of staining and the number of cells stained varied (results not shown). In contrast to the positive staining seen in sections of the bronchi, none of the >20 samples of peripheral lung studied revealed clear WFDC2 staining in alveolar epithelium (Figure ).
Figure 1 Distribution of WFDC2 in the respiratory tract and salivary glands. Immunohistochemistry for WFDC2 was performed as described in materials and methods. Sections show staining in nasal antral mucosa (A, B), nasal polyps (C), airway sub-mucosal glands (more ...)
In the light of our finding that WFDC2 staining was prominent in the minor glands of the upper airways, and our previous observation that WFDC2 was highly expressed in salivary gland RNA, we extended our studies to look at WFDC2 staining in both major and minor salivary glands. For all of these samples we stained sections from at least 3 individual patients. The three major salivary glands showed a striking distribution of WFDC2, predominantly in the ductular epithelium of the glands. In the parotid gland this intense staining of the excretory ducts was also accompanied by some staining in a few of the serous cells of the acini (Figure ). Similar staining in the excretory ducts was seen in the sub-mandibular gland along with staining in the striated and intercalated ducts, although in this tissue staining was not seen in serous cells within the acini themselves (Figure ). In the sublingual gland the strong ductular staining was also accompanied by staining within the serous demilunes of the predominantly mucinous acini, although the intensity of this staining was less marked than in the ducts. (Figure ). In samples of the posterior portion of tongue again the minor glands demonstrated strong WFDC2 staining (Figure ) and minor glands of the vallecular region of the tongue, associated with tonsil, also stained with WFDC2 (Figure ). In both of these glands the majority of the cells of the ducts expressed the protein very intensely. Most of the tonsil tissue, including tonsillar crypts or germinal centres, did not stain (results not shown).
These results clearly show that WFDC2 is present in a variety of glands associated with the upper respiratory tract, nasal passages and oral cavity and also suggest that the protein is present in a subset of upper airway epithelial cells. In addition, a significant site of WFDC2 protein localisation appears to be ductular cells of the major salivary glands. WFDC2 is not found at significant levels in the peripheral lung.
To complement our studies of the distribution of WFDC2, and to gain further insights into the cells that express the protein, we carried out more immunohistochemical studies, using additional antibodies on serial sections of multiple tissues. In the submandibular gland it is clear that expression of WFDC2 and the related 2-WAP domain containing protein, SLPI are almost mutually exclusive. As previously shown, (Figure ) WFDC2 was present in the ducts of the gland whereas SLPI was found in both the serous and (rarer) mucous cells of the gland (Figure ). SLPI staining was absent from the ducts. Staining for SPLUNC1 (a mucous cell marker [27
]) was found to be strongest in the mucous cells of the gland with weak staining in the ductular cells (Figure ). A similar distribution pattern was seen in the submucosal glands of the bronchiolar epithelium where WFDC2 was found in the ductular cells and a subset of cells in the serous demilunes of the predominantly mucinous acini (Figure ). The majority of mucous cells did not stain with WFDC2. In this tissue, however, SLPI staining was found in some of the ductular tissues that express WFDC2 as well as the mucous cells of the gland (Figure ). SPLUNC1 was clearly localised in the mucous cells of the gland (Figure ). These results show that WFDC2 is distributed in a distinct subset of cells within both major and minor salivary glands but it is not completely co-expressed with the related protein SLPI.
Figure 2 WFDC2 and SLPI do not co-localise in major and minor glands. The expression of WFDC2 (A, D) was directly compared with that of the related 2-WAP domain containing protein SLPI (B, E) and the mucous cell marker SPLUNC1 (C, F) in multiple glandular tissues (more ...)
As chronic inflammation within the lung has been associated with alterations in the levels of SLPI and elafin [23
], we studied WFDC2 expression in chronically inflamed lung using airway and peripheral lung sections from 10 patients who had undergone transplantation for CF. In sections of large airways from these samples there was no clear alteration in either the site or level of staining within the bronchiolar epithelial cells or the submucosal glands (results not shown). However, the situation in the smaller airways within the peripheral lung sections was markedly different. WFDC2 staining in the abnormal (hyperplastic) epithelium was found to be more generally distributed in the epithelium (Figure ) compared to the situation seen in similar sized airways from CF disease-free lung (Figure and Figure ). In addition to the greatly increased staining seen within the epithelium, the inflammatory mass within the airway lumen was also found to stain strongly for WFDC2 (Figure and ). Sections stained with SLPI clearly did not display the same increased expression as was seen with WFDC2. In non-diseased small airways, SLPI was limited to cells with the phenotypic characteristics of goblet cells (Figure ). In CF sections stained with the same antibody, SLPI immuno-reactivity was markedly reduced compared to normal lung (Figure ) and was not detectable in the inflammatory exudates within the lumen of the small airways (Figure ). In contrast staining of sections with the goblet cell marker Muc5AC clearly showed that there was a significant increase in the number of mucous secreting cells in the CF airway (Figure ). The luminal content of these small airways also stained strongly for Muc5AC (Figure ). These results indicate that WFDC2 immunoreactivity is increased in the lungs of patients with CF and furthermore suggest the protein is secreted into the luminal contents of the diseased lung. This suggests that although expression of the WFDC2 gene may be mediated by inflammatory signals this may be different to that previously described for the related proteins SLPI and elafin.
Figure 3 WFDC2 is abnormally expressed in the Cystic Fibrosis lung. Sections of normal (A, B) and Cystic Fibrosis lung (C – I) were stained with WFDC2 (A, C, D, G), SLPI (B, E, H) and mucin 5AC (F, I) as described in materials and methods. The original (more ...)
To directly address the question of potential regulation of WFDC2 by inflammatory mediators we performed a series of RT-PCR studies using primary human lung derived cells. In our previous study we had shown that WFDC2 was expressed in a number of lung cancer cell lines, some of which also expressed SLPI and elafin [21
]. Given the localisation of WFDC2 staining in cells of the bronchiolar epithelium we initially chose to study primary cells from this region. We confirmed that WFDC2 mRNA was readily detectable in submandibular and parotid gland and in NCI-H358 cells (results not shown). When RT-PCR reactions were performed on bronchial epithelial cells directly harvested from healthy donor lungs WFDC2 mRNA was readily detected (Figure , lane H). Reduced expression of WFDC2 was seen as the cells underwent the process of de-differentiation when plated in collagen-coated dishes for 3 days, allowed to expand (P1), and plated again for a second passage (P2). During this period of submerged culture the cells lose features of differentiation, confirmed by the complete loss of expression of the Clara cell secretory protein (CCSP) gene, a marker of differentiated bronchiolar cells (Figure ). When the cells were placed at an air-liquid interface they re-differentiated as evidenced by the visual appearance of cilia (results not shown) and re-expression of CCSP. At this stage, expression of WFDC2 was increased to levels similar to those observed in directly harvested cells (Figure , lane d14). Similar results were found in two other sets of cells (results not shown).
Figure 4 WFDC2 is expressed in primary tracheal cells, differentiated tracheobronchial epithelial cells and alveolar Type II epithelial cells. A. Expression of WFDC2 was studied by PCR using freshly harvested human tracheal cells (H) and samples taken from the (more ...)
To further investigate the role of differentiation of the ALI cells in influencing the expression of WFDC2 we modulated the differentiation status of the cells by removing retinoic acid from the culture medium. 50 nm RA is normally present in the culture medium of these cells and has been shown to be required for full differentiation of the cultures [39
]. Removal of RA from these cultures led to a progressive loss of WFDC2 expression such that following 18 days of RA withdrawal, expression was essentially lost (Figure ). Levels of expression of SLPI in the same samples were clearly not influenced by the removal of RA (Figure ) whereas, in marked contrast, expression of elafin almost completely mirrored that of WFDC2 with the RA depleted cells having significantly more expression of the elafin gene. The requirement for RA to maintain the mucous secretory phenotype of the cells was shown by the loss of expression of Muc5AC that paralleled the loss of WFDC2 (Figure ). Again we found similar results in two further sets of experiments with cells from different donors (results not shown).
We also examined the role of pro-inflammatory stimuli on the expression of WFDC2 in the ALI cells. For these studies differentiated cells were exposed to levels of IL-Iβ, TNFα and E. coli LPS known to induce expression of a variety of responsive genes. Neither TNFα nor LPS was found to induce expression of WFDC2 under any conditions (results not shown). Similar negative findings were found with IL-1β (Fig ) but this cytokine does upregulate elafin gene expression. The expression of SLPI mRNA was not significantly altered by any of these treatments (results not shown). These results appear to suggest that although expression of WFDC2 in ALI TBE cells is influenced by differentiation status, classical pro-inflammatory mediators do not appear to regulate the gene.
We have previously shown that SLPI and elafin are constitutively expressed in primary type II alveolar epithelial cell cultures and that elafin expression is induced by IL-1β in these cells [26
], thus we performed similar induction studies with WFDC2. WFDC2 gene expression was readily detected in control Type II cells (Figure ). Stimulation with the same pro-inflammatory mediators as was used for the ALI cells suggested that WFDC2 was slightly induced by IL-1β, but was not responsive to either TNFα or LPS (Figure ). Again, and consistent with our previous studies [26
], elafin expression was induced by IL-1β but was not altered by either TNFα or LPS treatment in the Type II cells. SLPI was not altered by any of these stimulations. We have also found a similar lack of induction of WFDC2 gene expression in lung cancer cell lines treated with pro-inflammatory mediators (results not shown). These results suggest that WFDC2 expression is not responsive to pro-inflammatory mediators in primary human lung cells and that the increased staining seen in the CF lungs may be due to a phenotypic alteration of the cell population rather than a direct transcriptional effect on the WFDC2 gene.
In light of the multiple reports of disregulated WFDC2 expression in ovarian cancers [13
] and the smaller number of reports that the gene may also be differentially expressed in sub-groups of pulmonary cancers [17
] we stained sections of a variety of lung carcinomas with the WFDC2 antibody. To do this we used commercial tissue arrays that contained 150 individual cases from a range of different lung cancers. Representative results of this study are shown in Figure and the data is summarised in Table . For representation of the data, each section was scored as being either negative, exhibiting focal positivity or being positive throughout the tumour. In focal positive cases the number and intensity of cells staining varied widely as did the intensity of staining in the positive cases. Due to the subjective nature of this assessment we have not attempted to sub-divide either of these two groups. It is clear from this data that both adenocarcinomas (Figure ) and bronchioloalveolar carcinomas (BAC, Figure ) exhibit the greatest percentage of positive staining tumours (>80%). In the majority of these strong, positive staining was identified throughout the tumour (Figure and ), although focal positive staining ranging from a few cells within the tumour (Figure ) to focal staining of all of the abnormal ductular tissue (Figure ) was also seen. 12 of 63 cases of adenocarcinomas/BACs were negative. Focal staining was seen in 18 of 60 cases of squamous carcinoma (Figure ), the majority of these tumours were negative (Figure ). Focal staining was also seen in a single case of adenoid cystic carcinoma (Figure ) whilst 11 of 14 cases of small cell carcinoma (Figure ) and a single case of mesothelioma were negative (Figure ). The majority of cases of large cell carcinoma (8 of 11) were negative with the remainder exhibiting weak focal positive staining (results not shown). These results suggest that WFDC2 staining is predominantly associated with adenocarcinomas although a percentage of squamous, small cell and large cell carcinomas also exhibit focal positive staining.
Figure 5 Distribution of WFDC2 in Pulmonary Neoplasms. Immunohistochemistry for WFDC2 was performed on three commercial lung cancer tissue arrays as described in materials and methods. Representative samples were chosen for imaging. Examples of focal (A, B) and (more ...)
Summary of WFDC2 staining in lung tumours. Cores were designated as exhibiting no staining (Negative), focal positive staining (Focal) or as being positively stained (Positive).