|Home | About | Journals | Submit | Contact Us | Français|
Rationale: Increased production of mucus is a prominent feature of asthma. IL-13–driven mucous cell metaplasia is associated with decreased expression of the transcription factor FOXA2 and increased expression of the related transcription factor FOXA3 in animal and cell culture models.
Objectives: Establish how changes in FOXA2 and FOXA3 expression contribute to mucous metaplasia and determine whether FOXA2 and FOXA3 expression is altered in asthma.
Methods: Mice expressing a Foxa2 transgene in airway epithelial cells and mice deficient in Foxa3 were analyzed after allergen sensitization and challenge. Expression of FOXA2, FOXA3, MUC5AC, and the highly IL-13–inducible gene CLCA1 was analyzed in airway biopsies from subjects with asthma and control subjects.
Measurements and Main Results: Expression of a Foxa2 transgene reduced allergen-induced mucous metaplasia by 45% compared with control transgenic mice (P < 0.05) whereas inactivation of Foxa3 had no detectable effects on mucous metaplasia. Expression of FOXA2 was reduced in subjects with asthma and was negatively correlated with MUC5AC and CLCA1 levels in subjects with asthma. In contrast, FOXA3 expression was not significantly correlated with MUC5AC and was positively correlated with CLCA1.
Conclusions: Increasing Foxa2 expression reduced mucous metaplasia in an allergic mouse model. Subjects with asthma had decreased FOXA2 expression, suggesting that therapeutic approaches that increase FOXA2 expression or function could be beneficial for reducing mucus production in asthma. Unlike FOXA2, FOXA3 did not regulate mucous metaplasia.
In mouse asthma models, mucus production is associated with decreased levels of the FOXA2 transcription factor and increased levels of the related protein FOXA3 in airway epithelial cells.
FOXA2 expression is reduced in humans with asthma and increasing FOXA2 expression reduces mucus in a mouse asthma model. FOXA3 did not regulate mucus levels in the same model.
Excessive mucus production is a common feature of asthma and contributes to morbidity and mortality (1–6). Studies using mouse asthma models (7–12) and cultured human bronchial epithelial cells (13, 14) established that the helper T type 2 cytokines IL-4 and IL-13 act directly on epithelial cells to produce mucous metaplasia. In this article, we focus on the functional importance of two related genes, Foxa2 and Foxa3, in this process.
The “Forkhead box a” transcription factors FOXA1, FOXA2, and FOXA3 have overlapping patterns of expression in organs derived from embryonic endoderm such as liver, stomach, and intestine (15). The three FOXA proteins are 95% identical within the DNA-binding domains but less similar in other domains (16, 17). Loss of either FOXA1 or FOXA2 alone does not prevent liver development but hepatic specification was completely abrogated in mice lacking both FOXA1 and FOXA2 in the foregut endoderm (16), indicating that one FOXA family member can sometimes compensate for the loss of another. However, functional relationships between FOXA family members remain incompletely understood.
Foxa2 is expressed at the onset of lung bud formation and continues to be expressed in the pulmonary epithelium in adulthood (15). Disruption of Foxa2 in respiratory epithelial cells caused airspace enlargement, neutrophilic pulmonary infiltrates, and mucous metaplasia (18). Airway epithelial cell FOXA2 expression was decreased by allergen challenge and by IL-4 and IL-13 overexpression in mouse airways (18) and by IL-13 stimulation of human bronchial epithelial cells (14), suggesting that loss of FOXA2 may contribute to mucous metaplasia in these systems. Unlike Foxa2, we found that Foxa3 mRNA was increased in lungs of allergen-challenged mice and FOXA3 mRNA was increased during IL-13–induced mucous metaplasia of cultured human bronchial epithelial cells (14). On the basis of these observations, we hypothesized that increased FOXA3 partially compensates for decreased FOXA2 in allergic airways, thereby limiting mucous metaplasia. An alternative hypothesis was that FOXA3 competes with FOXA2 for DNA binding, and therefore amplifies mucous metaplasia by reducing FOXA2 activity.
In this study, we analyzed how changes in FOXA2 and FOXA3 expression contribute to mucous metaplasia in allergic airway disease and asthma. We used transgenic mice that express a Foxa2 transgene in airway epithelial cells (to counteract allergen-induced loss of FOXA2 expression) and mice deficient in Foxa3 (to prevent allergen-induced FOXA3 expression) (19). We also analyzed airway epithelial FOXA2 and FOXA3 expression in asthma. Our results provide new information about the roles of FOXA2 and FOXA3 in allergic airway disease and asthma. Some of the results of these studies have been previously reported in the form of an abstract (20).
Inducible Foxa2 and enhanced green fluorescent protein (EGFP) transgenic mice were produced by coinjection of CCSP-rtTA-hGH (21) and pTRE-Tight-Foxa2 or pTRE-Tight-EGFP (Clontech, Mountain View, CA) into FVB blastocysts. Antisera recognizing FOXA2 (Upstate Biotechnology, Santa Cruz, CA) and MUC5B (kindly provided by C. W. Davis, University of North Carolina, Chapel Hill, NC ) were used to detect these proteins in lung sections. Foxa3−/− mice on a C57BL/6 genetic background were generously provided by K. Kaestner (19). To detect FOXA3 protein, lung homogenates were analyzed by immunoblotting, using an antiserum against FOXA3 (Abcam, Cambridge, MA). FVB/N and BALB/c mice were purchased from Jackson Laboratory (Bar Harbor, ME). The University of California San Francisco (San Francisco, CA) Committee on Animal Research approved the use of mice for these experiments. Mice were housed in a specific pathogen–free facility.
Six- to 8-week-old transgenic and strain-, age-, and sex-matched control mice were sensitized and challenged with ovalbumin as reported previously (23). In experiments involving Foxa2 transgenic mice, all mice were provided with food containing doxycycline (2 g/kg) to induce transgene expression beginning after the final sensitization and continuing until the mice were killed. Design-based stereology was applied to measure mouse and human epithelial cell mucin stores using point and intercept counting (24), and mucin granule volume using the point-sampled intersect technique (25). Analyses of serum ovalbumin–specific IgE and bronchoalveolar lavage fluid leukocytes (23) and Flexivent measurements of airway reactivity (26) were performed as described previously.
RNA from mouse lungs was reversed transcribed to cDNA and analyzed by SYBR green real-time polymerase chain reaction (PCR). The normalized copy number was determined by comparing the threshold cycle (Ct) of each transcript with the mean Ct for Tubb, Actb, and Gapd. For RNA from human epithelial brushings, cDNA synthesis and two-step quantitative PCR (qPCR) was performed as described previously (27).
NCI-H292 cells were transfected with a human MUC5AC promoter-luciferase plasmid (28) together with pcDNA3.1-FOXA2 or pcDNA3.1-FOXA3 expression plasmids, using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Luciferase activity was measured 48 hours after transfection. The pSV-βGalactosidase plasmid was included in each transfection and we normalized for transfection efficient by determining the ratio of luciferase activity to β-galactosidase activity. All transfections were performed in triplicate.
Information about the subjects and FOXA2 staining is available in the online supplement.
Data are reported as means ± SEM. For analyses of FOXA2, FOXA3, MUC5AC, and CLCA1 expression in bronchial epithelial cells from subjects with asthma and control subjects, microarray data generated in our previous study (27) were downloaded from the Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo; accession number GSE4302). Significance testing was performed by Student t test or by analysis of variance and Tukey-Kramer posttest for multiple groups unless otherwise indicated. Correlations were analyzed by linear regression and analysis of variance.
Confirming previous work (18), we found that nuclear FOXA2 protein staining in airway epithelial cells was reduced during allergen-induced mucous metaplasia (Figures 1A and 1B). Quantitative reverse transcription-PCR analysis indicated that allergen challenge resulted in a decrease of Foxa2 mRNA levels in the lungs (44% decrease; P < 0.05). To determine whether persistent Foxa2 expression would inhibit mucous metaplasia, we produced mice with a Foxa2 transgene driven by a doxycycline-regulated protein (reverse tetracycline transregulator, rtTA) expressed under the control of the Clara cell secretory protein (CCSP) promoter. After doxycycline treatment, FOXA2 immunoreactivity was more intense in the Foxa2 transgenic mice than in control mice (Figure 1C). To determine whether transgene expression was affected by allergen challenge, we measured Foxa2 transgene mRNA in the lungs of saline- and ovalbumin-challenged mice. Allergen challenge had no detectable effect on Foxa2 transgene expression (Figure 1D). These results indicate that the Foxa2 transgene was expressed in airway epithelial cells and that transgene expression persisted after allergen challenge.
We previously reported that Foxa3 mRNA was increased after allergen challenge of FVB/N mice (14). To determine whether increases were seen in other strains more commonly used for asthma models, we compared the responses of BALB/c and C57BL/6 mice with the response of FVB/N mice (Figure 2A). Foxa3 mRNA expression increased by 5.5-fold in BALB/c mice, by 3.0-fold in C57BL/6 mice, and by 2.5-fold in FVB/N mice after allergen challenge (Figure 2A). For unclear reasons, these fold increases were smaller than the increase we previously reported in FVB/N mice using the same model system. We analyzed lungs from saline- and ovalbumin-treated mice by immunoblotting to determine whether the increase in Foxa3 mRNA was accompanied by an increase in FOXA3 protein. FOXA3 protein expression was detectable in wild-type C57BL/6 mice challenged with ovalbumin but not in saline-challenged mice (Figure 2B). As expected, no FOXA3 protein expression was detectable in Foxa3−/− mice even after allergen challenge. These results show that FOXA3 protein is induced during allergic airway disease and confirm the absence of FOXA3 protein in Foxa3−/− mice. We previously showed that IL-13, a critical helper T cell type 2 cytokine produced during allergic inflammation, induces Foxa3 expression in purified airway epithelial cells. We attempted to use the FOXA3 antiserum to localize FOXA3 protein in the allergic lung but obtained similar staining in wild-type and Foxa3−/− mice, indicating that the staining was not specific for FOXA3.
We used Foxa2 transgenic mice and strain-matched wild-type control mice to determine whether transgenic expression of Foxa2 had an effect on mucous metaplasia. Mice were sensitized and challenged with saline or ovalbumin, and received doxycycline throughout the period of allergen challenge to promote transgene expression. Saline-challenged mice had little or no mucin visible in the airway epithelium (data not shown), consistent with previous reports (12, 29). We used computer-assisted stereology to quantify airway epithelial mucin stores in ovalbumin-challenged mice. Airway epithelial mucin stores in ovalbumin-challenged Foxa2 transgenic mice were reduced by 37% compared with ovalbumin-challenged wild-type mice (Figure 3A). Despite the decrease in overall volume of stored mucin within the epithelium, there was no decrease in the volume of mucin within individual cells (Figure 3B). This indicates that the overall decrease in mucin stores reflects a decrease in the proportion of mucin-containing cells within the epithelium and not a decrease in the amount of mucin stored per cell. The Foxa2 transgenic mouse line that we produced includes two transgenes: a Foxa2 transgene and an rtTA transgene that regulates Foxa2 transgene expression. We considered the possibility that expression of the rtTA transgene rather than of Foxa2, in Foxa2 transgenic mice might be responsible for the effect on mucus production. To address this, we measured airway epithelial mucin stores in a control transgenic line. The control transgenic line carried the rtTA transgene and an irrelevant transgene (EGFP) in place of the Foxa2 transgene. After allergen challenge, airway epithelial mucin stores in EGFP transgenic mice were similar to mucin stores in nontransgenic control mice. Mucin stores in Foxa2 transgenic mice were 45% less than in EGFP transgenic mice (Figure 3A). These results indicate that transgenic expression of FOXA2 specifically reduced allergen-induced mucous metaplasia.
We analyzed the role of FOXA3 in allergen-induced mucous metaplasia, using Foxa3−/− mice. Foxa3−/− mice and strain-matched wild-type control mice had similar airway epithelial cell mucin stores after ovalbumin challenge (Figure 3C). This result demonstrates that allergen-induced increases in FOXA3 did not affect mucous metaplasia.
FOXA2 has been shown to reduce MUC5AC transcriptional activity in NCI-H292 human lung mucoepidermoid cells (18), which may explain the ability of FOXA2 to inhibit mucous metaplasia in the airways of allergen-challenged mice. We used a similar approach to compare the effects of FOXA2 and FOXA3 on MUC5AC transcription. A luciferase reporter construct containing 3.8 kb of the human MUC5AC promoter was transfected into NCI-H292 cells together with human FOXA2 or FOXA3 expression vectors. FOXA2 decreased reporter expression by 57% (Figure 3D), consistent with the previous report (18). In contrast, expression of FOXA3 had no effect on MUC5AC transcription reporter expression.
In addition to MUC5AC, allergen challenge also induces expression of MUC5B (12). To determine whether FOXA2 overexpression affected MUC5B, we quantified MUC5B staining in lung sections from ovalbumin-challenged nontransgenic, Foxa2 transgenic, and EGFP transgenic mice, using stereology. MUC5B stores in Foxa2 transgenic mice were 36% less than in nontransgenic control mice (P < 0.05), whereas EGFP transgenic mice were similar to nontransgenic control mice (99% relative to the nontransgenic control mice). We did not detect staining for MUC2, a mucin that is found primarily in the intestine, in airways from control or transgenic mice. Taken together, our results indicate that FOXA2 overexpression can reduce production of both major airway mucins, MUC5AC and MUC5B.
We considered the possibility that the effects of Foxa2 transgene expression might be explained by a general effect on the allergic response rather than a specific effect on epithelial mucus production. To address this possibility, we analyzed other aspects of allergic airway disease in Foxa2 mice. We found that Foxa2 transgenic and strain-matched control mice had similar elevations in serum ovalbumin–specific IgE after allergen challenge (Figure 4A). Bronchoalveolar lavage was performed to assess the effect of allergen challenge on inflammatory cell recruitment (Figure 4B). Allergen challenge induced similar increases in macrophages, eosinophils, and neutrophils in Foxa2 transgenic and control mice, whereas Foxa2 transgenic mice had a significantly greater increase in lymphocytes. Airway reactivity was determined by measuring pulmonary resistance in sedated and mechanically ventilated mice after intravenous administration of increasing doses of acetylcholine (Figure 4C). Foxa2 transgenic mice developed a similar degree of allergen-induced airway hyperreactivity as wild-type control mice. These results indicate that the reduction in mucous metaplasia seen in allergen-challenged Foxa2 transgenic mice was not accompanied by reductions in IgE production, airway inflammation, or airway reactivity.
Although FOXA3 deficiency did not affect mucus production, we explored the possibility that FOXA3 might be involved in other aspects of allergic airway disease. We found that allergen-challenged Foxa3−/− mice had increased IgE production and eosinophilic inflammation but reduced airway reactivity compared with nontransgenic control mice (see Figure E1 in the online supplement).
Humans with mild and moderate asthma have increased airway epithelial mucin stores and an increase in the number of goblet cells compared with healthy control subjects (2). To determine whether FOXA2 expression is also altered in asthma, we quantified airway epithelial cell FOXA2 expression in airway biopsies obtained from five subjects with mild or moderate asthma and five healthy control subjects. The ratio of FOXA2-unstained nuclei to FOXA2-stained nuclei was significantly reduced in asthma (Figure 5A). We found a significant negative correlation between MUC5AC and the ratio of FOXA2-stained nuclei to FOXA2-unstained nuclei (Figure 5B). This indicates that asthma and increased mucin stores are associated with reduced FOXA2 expression in airway epithelial cells.
To further analyze FOXA2 and FOXA3 expression in human subjects, we analyzed data from our genome-wide study of mRNA transcript levels in bronchial epithelial cells from 42 subjects with stable mild or moderate asthma and 28 healthy control subjects (27). FOXA2 expression was significantly lower in subjects with asthma compared with control subjects (Figure 6A). Within the asthma group, FOXA2 expression was negatively correlated with MUC5AC expression (Figure 6B). FOXA2 expression was also negatively correlated with the expression of CLCA1, a gene that is highly induced by IL-13 (14, 27) and is expressed at high levels in epithelial cells from many subjects with asthma (Figure 6C). FOXA3 expression tended to be modestly higher in subjects with asthma compared with control subjects, but this did not reach statistical significance (Figure 6D). Within the asthma group, there was no significant correlation between FOXA3 and MUC5AC expression (Figure 6E). In contrast, there was a highly significant positive correlation between FOXA3 and CLCA1 (Figure 6F). These results suggest that IL-13 (and/or other stimuli that increase CLCA1 expression) decrease FOXA2 expression and increase FOXA3 expression.
Our goal was to evaluate the functional importance of changes in FOXA2 and FOXA3 expression that occur during allergic airway disease and asthma (14). FOXA2 expression is decreased during allergic airway disease, and we investigated the importance of this decrease by producing and characterizing mice with a transgene that increased FOXA2 expression in airway epithelial cells. These studies demonstrated that persistence of FOXA2 expression during allergic responses reduces airway mucus without detectable effects on other aspects of the allergic response that we analyzed. FOXA3 expression is increased during allergic airway disease, and we investigated the importance of this increase by studying FOXA3-deficient mice. Lack of FOXA3 had no detectable effect on allergen-induced increases in airway mucus. In human subjects, we showed that FOXA2 expression was decreased in asthma and that the decrease in FOXA2 expression is associated with increased mucin levels. In contrast, FOXA3 expression was not associated with mucin expression. Taken together, these studies suggest that loss of FOXA2 expression contributes to mucous metaplasia in asthma, whereas FOXA3 is not involved in this aspect of the disease.
Our studies of FOXA2 complement a previous groundbreaking study (18) that demonstrated a major role for this transcription factor in regulating mucus production. In the previous work, conditional deletion of FOXA2 in respiratory epithelial cells induced mucus production in the absence of allergen challenge or other exogenous stimuli. In addition, allergen challenge of wild-type mice was shown to decrease FOXA2 expression, suggesting that allergen-induced increases in mucus might depend on loss of FOXA2. We directly tested this possibility with Foxa2 transgenic mice, and found a significant reduction of allergen-induced airway mucus in these mice. This finding demonstrates that allergen-induced mucus production is at least partially dependent on loss of FOXA2 in airway epithelial cells. The effect of the Foxa2 transgene is likely explained by a cell autonomous effect of FOXA2 on mucus production, because FOXA2 overexpression reduced MUC5AC transcription in cultured lung mucoepidermoid cells ( and Figure 3D) and because other aspects of allergic airway disease, including IgE production, airway inflammation, and airway hyperreactivity, were not reduced in Foxa2 transgenic mice. Although we have previously shown that hyperreactivity persists even when mucus production is dramatically reduced in this allergic model, other studies in mouse models (30) and humans (31) indicate that mucus hypersecretion is often a major contributor to airway obstruction in allergic models and in humans with asthma. Foxa2 transgenic mice were not completely protected against allergen-induced mucus production. This might reflect the existence of FOXA2-independent pathways for mucus production. However, it is also possible that the failure to completely repress mucus production is due to incomplete restoration of FOXA2 expression in airway epithelial cells of allergen-challenged transgenic mice. The promoter we used to drive transgene expression is active in Clara cells, which are believed to be the major precursors of mucus-producing cells in allergic airway disease (32), but other cell types may also be capable of transdifferentiation to mucus-producing cells (33).
We found that Foxa3 mRNA was increased during allergic airway disease (14), but the contributions of FOXA3 in this disease have not been explored previously. Here we extended our previous work by showing that the allergen-induced increase in Foxa3 mRNA is accompanied by a substantial increase in FOXA3 protein in the lung. We attempted to identify FOXA3-expressing cells in lung sections, but the antibodies we used were not suitable for immunohistochemistry because they produced similar staining patterns in wild-type and Foxa3−/− mice (data not shown). However, in previous work we showed that the cytokine IL-13, which plays a key role in the airway epithelial response to allergy, substantially increased FOXA3 expression in cultured normal human bronchial epithelial cells (14). This suggests that the FOXA3 protein detected in allergic mouse lungs is at least partially derived from airway epithelial cells. We hypothesized that the increase in FOXA3 expression might partially compensate for loss of FOXA2 in allergic airways, and thereby limit the extent of mucous metaplasia. Alternatively, we considered the possibility that FOXA3 might competitively inhibit FOXA2 and thereby promote mucous metaplasia. To investigate this, we examined airway mucins in allergen-challenged Foxa3−/− mice. We found similar amounts of stored mucins in airways from Foxa3−/− and matched Foxa3+/+ control mice. We also compared the effects of FOXA2 and FOXA3 on MUC5AC promoter activity in cultured cells. FOXA2 inhibited transcription but FOXA3 had no effect. Analysis of mRNA expression data from our study of freshly isolated bronchial epithelial cells (Figure 6) showed a highly significant positive correlation between FOXA3 expression and expression of CLCA1, a highly IL-13–inducible gene that is greatly increased in asthma (14, 27). This suggests that the IL-13–induced increases in FOXA3 expression that have been reported in mice and in cultured human cells (14) also occur in vivo in people with asthma. However, expression of FOXA3 was not significantly correlated with expression of MUC5AC. Taken together, these findings indicate that changes in FOXA3 expression, unlike changes in FOXA2 expression, do not affect airway mucus accumulation in response to allergen. FOXA2 and FOXA3 have substantially different sequences (16, 34) and it seems likely that structural differences between these two family members lead to differences in interaction with DNA or with other proteins involved in the regulation of gene expression. Allergen-challenged Foxa3−/− mice, unlike Foxa2 transgenic mice, had differences in serum OVA-specific IgE, bronchoalveolar lavage eosinophil counts, and airway reactivity compared with allergen-challenged control mice (Figure E1). Further work will be required to determine whether these differences indicate a direct role for airway epithelial cell FOXA3 in these aspects of the allergic response.
FOXA2 is a promising therapeutic target in asthma and other diseases characterized by excessive mucus production. We found that maintaining FOXA2 expression in airway epithelium is sufficient to cause a substantial decrease in airway mucus in vivo in a mouse model of asthma. We also showed that FOXA2 expression is reduced in human subjects with mild to moderate stable asthma and is negatively correlated with the volume of stored mucins and the level of MUC5AC mRNA. Loss of FOXA2 staining has also been reported in areas of mucous metaplasia within airways from humans with cystic fibrosis, chronic obstructive pulmonary disease, and bronchopulmonary dysplasia (18). This suggests that studies of the function of FOXA2 in mice are relevant for common human diseases. Therapies that increase FOXA2 expression or function in airway epithelial cells might inhibit mucus overproduction, which is an important cause of morbidity and mortality in asthma and other airway diseases (35).
The authors thank A. Barczak and N. Killeen for advice and K. Huang, Y. Wang, X. Bernstein, X. Ren, and S. Kim for technical assistance. C.V. was supported by a fellowship from the Belgian American Educational Foundation and by the Leon Fredericq Fund.
Supported by funding from the NIH and the UCSF Strategic Asthma Basic Research (SABRE) Program.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
Originally Published in Press as DOI: 10.1164/rccm.200811-1768OC on July 23, 2009
Conflict of Interest Statement: S.-W.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; C.V. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; L.T.N. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; R.B. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; C.J.E. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; Y.N. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; X.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; P.G.W. receives $200,000 per year as a research grant from Genentech Inc and is a coinventor on a patent related to asthma biomarkers; J.V.F. received $1,001–$5,000 from Amira, up to $1,000 from Gilead, $1,001–$5,000 from Merck, $1,001–$5,000 from Roche, up to $1,000 from Aerovance for consulting, $5,001–$10,000 from Cytokinetics as a member of their scientific advisory board, more than $100,001 in grants for research in asthma and cystic fibrosis from Genentech, a $50,001–$100,000 grant for research in asthma from Roche, and a grant of more than $100,001 for a clinical trial in COPD from Boehringer Ingelheim, and is the coinventor on a patent for the development of biomarkers of asthma; D.J.E. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript.