Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Lung mucus plays an important role in host defense by trapping inhaled airborne pathogens and facilitating their removal from the airways via a mucociliary transport mechanism. When overproduced, however, mucus contributes significantly to the morbidity and mortality of patients with respiratory diseases, including asthma, chronic bronchitis, and cystic fibrosis (1, 2). Mucus glycoproteins (mucins) are largely responsible for the maintenance of the viscoelasticity of lung mucus and its adhesive properties for airborne pathogens (3). In the airways secretion, MUC5AC is a major mucin secreted from the goblet cells of the surface epithelium (4, 5). Mucins are large molecular weight glycoproteins of which 10 to 20% are protein and 80 to 90% are complex carbohydrates. Mucin polypeptide typically contains tandem repeat regions that are rich in serine and threonine where most mucin-type glycans are localized. Mucin carbohydrate is made of five sugars: D-N-acetylgalactosamine (GalNAc), D-N-acetylglucosamine (GlcNAc), D-galactose (Gal), L-fucose, and sialic acid. Mucin oligosaccharides are very heterogeneous in structure and sizes, which can range from one to more than 20 residues per chain linked through GalNAc to serine or threonine on the protein backbone (3). Mucin polypeptide is synthesized at the endoplasmic reticulum while glycosylation takes place in the Golgi apparatus. Mucin oligosaccharides are synthesized by adding one sugar at a time, and the product of an enzymatic step serves as the acceptor for a subsequent step. This template-independent synthetic process is responsible for the heterogeneity of mucin carbohydrate structures. Furthermore, changes in the expression of mucin glycosyltransferases as a result of changes in physiologic conditions (6) or during the progression of diseases (7) can lead to changes in mucin carbohydrate structures.

Mucus hypersecretion in the respiratory tract and mucus cell metaplasia (MCM) in the airway epithelium are found in many human obstructive pulmonary diseases associated with inflammation in the lung (1, 2). Several MCM animal models (8), including a murine model generated by ovalbumin treatment (8), have been reported. This mouse model has been employed for the study of atopic asthma but it may not be suitable for the study of bacteria-induced MCM. Bacterial infection of the lung is known to induce inflammatory response, which leads to overproduction of mucus. For example, Gram-positive and Gram-negative bacteria can upregulate the expression of MUC2 and MUC5AC genes in human airway cell lines and tissues (13). Up-regulation of mucin gene expression by bacterial lipopolysaccharide (LPS) also has been reported in human airway cells (13, 14). However, these studies did not examine the effects of bacteria or bacterial LPS treatment on mucin glycosylation, the steps conferring mucins their major biologic functions.

In this report, we show that a single instillation of Pseudomonas aeruginosa LPS into murine lungs induces MCM and hyperplasia. We also show that the complementary DNAs (cDNAs) encoding key glycosyltransferases may be used as molecular probes to monitor the sequence of events involved in the induction of MCM. The first probe is mucin core 2 6 N-acetylglucosaminyltransferase (C2GnT)-L cDNA, which encodes a glycosyltransferase that catalyzes the transfer of GlcNAc from uridine diphosphate-GlcNAc to the Carbon-6 of GalNAc in Gal1-3GalNAcSer/Thr acceptor, forming core 2 structure, Gal1-3(GlcNAc1-6)GalNAc. Core 2-associated sialyl Lewis x structure is an essential part of the ligands recognized by P- and L-selectins (15) and possibly E-selectin as well (15, 18). Sialyl Lewis x can be generated by introduction of C2GnT-L cDNA into cell lines that do not express C2GnT-L (15, 19), underscoring the importance of the upregulation of C2GnT-L in the formation of selectin ligands during inflammation. In addition, ablation of C2GnT-L gene leads to a defect in neutrophilic inflammation (20), further confirming the essential role played by this enzyme in inflammation. C2GnT-L also was detected in the mucus cells of the surface epithelium and the submucosal gland of bovine airway epithelium (21). Therefore, mucin C2GnT-L cDNA may be useful for monitoring the influx of neutrophils during the initial phase and the generation of mucus cells during the late phase of inflammation. The second probe is C2GnT-M cDNA, which encodes C2GnT-M, a glycosyltransferase responsible for the synthesis of all branched 1-6 GlcNAc structures found in mucin-type glycans (22). These structures include core 2 (Gal1-3[GlcNAc1-6]GalNAcSer/Thr); core 4 (GlcNAc1-3[GlcNAc1-6]GalNAcSer/ Thr); and blood group I (GlcNAc1-3[GlcNAc1-6]Gal). The human C2GnT-M cDNA was cloned recently (22, 23), and the mRNA is expressed primarily in the mucus secretory tissues, including gastrointestinal tract and trachea (23). Hence, the enzyme may be useful for monitoring the appearance of mucus cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals and LPS Treatment

Female C57BL6 mice approximately 5 wk old and 20 to 25 g body weight were used for the current study. Three animals per group were used for bronchoalveolar lavage (BAL) experiments, three animals per group for Northern blot analysis, and two animals per group for Western blot analysis. After being fully anesthetized by an intraperitonial injection with avertin (5 mg/20 g body weight), each mouse received an intratracheal injection of 100 µg LPS (P. aeruginosa serotype 10; Sigma, St. Louis, MO) dissolved in 50 µl saline. Mice that were treated with 50 µl saline served as the control. The reagent was delivered orally to the trachea by a bolus injection using a 24-gauge, 0.75-in angiocath (Becton Dickinson Infusion Therapy Systems Inc., Sandy, UT). The protocol was approved by the Institutional Animal Care and Use Committee at the University of Nebraska Medical Center.

Lung Lavage, Tissue Fixation, and Histologic Staining of Tissue Sections and Leukocytes

Lung lavage, tissue fixation, and staining of cells in lung lavage fluid were performed as described previously (25). Briefly, after the animal had been killed and its chest opened to expose the lung, a blunt-ended 22-gauge needle was inserted into the trachea. BAL was performed in situ three times sequentially with 1 ml saline each time and the recovered fluids were pooled for each animal. Leukocytes obtained from each mouse in the group were washed and counted with a hemacytometer. For differential cell counts, cells were centrifuged onto a slide in a tabletop centrifuge at 1,000 × g for 1 min and the slides were stained with May- Giemsa stain, and differential cell counts were performed by counting 300 cells. To prepare lungs for tissue sectioning, the lungs were excised completely from the chest, inflated with 1 ml of 10% formalin, and then immersed in 10% formalin. Paraffin embedding and tissue staining with Alcian blue (AB; pH 2.5) and periodic acid-Schiff (PAS) were performed using standard methodologies.

RNA Isolation and Northern Hybridization

For RNA isolation, the whole lung was homogenized in TRI reagent (GIBCO-BRL, Gaithersburg, MD) immediately after removal from the animal. PolyA RNA was isolated from total lung RNA prepared from three animals using a PolyATract messenger RNA (mRNA) isolation system (Promega, Madison, WI). For Northern blot analysis, equal amounts of polyA RNA (1.0 µg/lane) were subjected to electrophoresis in a 1% agarose gel containing 0.66 M formaldehyde. PolyA RNA was transferred to a Nytran membrane by capillary action using a Turboblotter system (Schleicher & Schuell, Keene, NH) and ultraviolet-crosslinked. Prehybridization (1% sodium dodecyl sulfate [SDS] and 1 M NaCl at room temperature for 30 min) and hybridization (HS-114F) were carried out according to the instructions of the High Efficiency Hybridization System (Molecular Research Center, Inc., Cincinnati, OH). ³²P-labeling of the DNA probes prepared by reverse transcription/polymerase chain reaction (RT-PCR) described in the next section was performed by the random-primer labeling method using Prime-It II (Stratagene, La Jolla, CA) with [³²P]deoxycytidine triphosphate (3,000 Ci/mmol; Amersham, Arlington Heights, IL). The probes were purified on G-50 Sephadex columns (Boehringer Mannheim, Indianapolis, IN), denatured by heating at 100°C for 5 min, and added to the hybridization solution at a concentration of 1 × 10⁶ cpm/ml. Hybridization of all Northern blots was performed at 55°C for 24 h and washing was carried out in a solution containing 1% SDS and 1× saline sodium citrate (SSC) (SSC: 0.15 M NaCl and 0.015 M sodium citrate) three times (10 min each time) at room temperature and four times (20 min each time) at 55°C. For autoradiography, membranes were exposed to Kodak X-OMAT AR film (Eastman Kodak Co., Rochester, NY) with the use of intensifying screens at 70°C. Membranes were stripped with 1% SDS and 40 mM Tris-HCl, pH 7.5, at 100°C for 5 min and reprobed. One membrane was reprobed with C2GnT-L cDNA and then glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA after probing with Muc5ac cDNA. Another membrane was probed with C2GnT-M cDNA first and then with GAPDH cDNA after stripping.

cDNA Probes

Mouse C2GnT-M cDNA probe was prepared from the mouse full-length C2GnT-M cDNA cloned by cross-species RT-PCR based on human C2GnT-M sequence (22, 23). The primers for amplifying a 341-bp RT-PCR product corresponding to residues 43 to 383 of the open reading frame (ORF) of mouse C2GnT-M cDNA were: 5'-TTTGCGTACTCTTGGTGGTG-3' (5'-primer) and 5'-CATACTGTTCTGCCCTTTCA-3' (3'-primer). All other cDNA probes were generated from mouse stomach polyA RNA by RT-PCR using sequence-specific primers. The primers for amplifying a 294-bp RT-PCR product corresponding to residues 282 to 575 of the ORF of mouse Muc5ac cDNA (26) were: 5'-CATCTCTACAACCCAAACTA-3' (5'-primer) and 5'-GAGGAGGGTTTGATCTGTTT-3' (3'-primer). The Muc5ac cDNA probe corresponded to residues 2502 to 2317 upstream from the stop codon of human MUC5AC cDNA (27). The primers for amplifying a 248-bp PCR product corresponding to residues 414 to 889 of the ORF of mouse C2GnT-L cDNA (28) were: 5'-GAGATGATCCTTACAGCAAT-3' (5'-primer) and 5'-TCTTATGATGAACCACAATG-3' (3'-primer). The primers for amplifying a 214-bp RT-PCR product corresponding to residues 8 to 221 of the ORF of mouse GAPDH cDNA (29) were: 5'-ATCTTCTTGTGCAGTGCCAG-3' (5'-primer) and 5'-GTAGTTGAGGTCAATGAAGG-3' (3'-primer). The PCR products were cloned into a pCRII vector using the TA Cloning Kit (Invitrogen, San Diego, CA) and then confirmed by nucleotide sequencing.

Western Blot Analysis

Lung homogenate was prepared by homogenizing a freshly excised lung with a Tekmizer (TR-5; Tekmar Company, Franklin Lakes, NJ) in 1 ml of phosphate-buffered saline containing a mixture of protease inhibitors (1:1,000) (p8340; Sigma) plus 1 mM ethylenediaminetetraacetic acid and 0.2 mM phenylmethylsulfonyl fluoride added immediately before use. For Western blot analysis, the proteins in the supernatant of the lung homogenate (1,000 × g, 5 min) were separated by polyacrylamide gel electrophoresis on 4% stacking/6% separating polyacrylamide gels and transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA). Then, the membranes were blocked with 3% bovine serum albumin in Tris-buffered saline (TBS) and incubated with HO1 primary antibody (1:1,000), kindly provided by Dr. David Ho at the University of Minnesota at Minneapolis (26). HO1 polyclonal antibody was generated in chickens using mouse gastric Muc5ac glycoprotein as the immunogen. The membranes were then washed with 0.05% Tween 20 in TBS and incubated with horseradish peroxidase-conjugated rabbit anti-chicken immunoglobulin (Ig) Y (1:4,000) (Sigma) for 1 h. After two additional washes, the signal was developed with ECL-Plus (Amersham). The polyclonal antibody HO8, generated in chickens using mouse Muc5ac tandem repeat peptide conjugated with keyhole limpet hemacyanin as the immunogen (26), failed to detect mouse lung Muc5ac glycoprotein and was not used further.

Statistical Analysis

Statistical analysis was performed using an unpaired t test program (StatView; Abacus Concepts, Berkeley, CA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Histologic Examination of Lung Tissue Sections Stained with Hemotoxylin and Eosin and AB-PAS

Histopathologic examination of hemotoxylin and eosin (H&E)-stained lung tissue sections of mice treated with LPS showed massive bronchopneumonia with accumulation of neutrophils inside the bronchial lumen on Day 2 post-treatment (Figure 1B), indicating infiltration of acute inflammatory cells. The inflammation subsided on Day 4 (Figure 1C) and essentially disappeared on Day 7 (Figure 1D). No apparent increase in inflammatory cells was detected in the lungs of saline-treated mice (Figure 1A). In LPS-treated mice, the AB-PAS-positive cells were not detected in the tissue sections on Day 2 post-treatment (Figure 2B), but occasional AB-PAS-positive cells were detected on Day 4 (Figure 2C). On Day 7, the number of AB-PAS-positive cells and the intensity of AB-PAS staining in these cells reached their highest level in the LPS-treated mice (Figure 2D). MCM and hypertrophy were observed in the epithelial cells on the conducting airways from trachea to bronchioles. However, no AB-PAS-positive airway epithelial cells were detected in saline-treated mice (Figure 2A).

View larger version (118K): [in this window] [in a new window]   Figure 1.   H&E staining of paraffin sections of lung tissues from mice treated with saline (A) or LPS (B, C, and D) (original magnification: ×50). The histopathologic examination of the lung specimens from mice treated with saline showed no inflammatory cells (A). Lung specimens from mice killed on Day 2 (B) after LPS instillation showed bronchopneumonia with accumulation of neutrophils in the lung (arrows) and inside the bronchial lumen, indicating infiltration of acute inflammatory cells. The inflammatory cells were hardly detectable on Days 4 (C) and 7 (D).

View larger version (104K): [in this window] [in a new window]   Figure 2.   AB-PAS staining of paraffin sections of lung tissues from mice treated with saline (A) or LPS (B, C, and D) (original magnification ×125). No AB-PAS-positive cells were detected in the lung tissue sections of saline-treated mice (A). In LPS-treated mice, mucus cells were not detected on Day 2 but began to be detected on Day 4 and reached a maximum on Day 7. Mucus cells are identified with an arrow.

Analysis of the Inflammatory Cells in BAL Fluid

The type and time-course changes of the inflammatory cells in BAL fluid (BALF) after intratracheal instillation with LPS and saline is shown in Table 1. The total number of inflammatory cells, including neutrophils, macrophages, and lymphocytes in the BALF of LPS-treated mice, was significantly higher than those of saline-treated mice on Days 2, 4, and 7 post-treatment. Epithelial cells and eosinophils constituted less than 1% of the total number of cells in BALF. The number of inflammatory cells in BALF on Day 2 post-treatment reached a peak for both saline and LPS-treated mice, which were 2.6 and 226 times that of the untreated mice, respectively. In both cases, neutrophils made up the majority of the inflammatory cells in BALF, i.e., 77 and 94%, respectively. In saline-treated mice, the total number of inflammatory cells and neutrophils returned to baseline on Day 4 and thereafter. In LPS-treated mice, the total number of inflammatory cells, neutrophils, macrophages, and lymphocytes in BALF peaked on Day 2 and decreased drastically on Day 4 and further on Day 7 post-treatment. In all cases, neutrophils were the major inflammatory cell type.

Expression of Muc5ac, C2GnT-M, and C2GnT-L mRNAs in the Lungs of Mice Treated with LPS and Saline

Northern blot analysis of polyA RNA isolated from LPS-treated mice revealed a time-dependent change in the expression of Muc5ac, C2GnT-M, and C2GnT-L mRNAs. The intensity of one major mRNA species for Muc5ac (13.5 kb), C2GnT-M (2.4 kb), and C2GnT-L (2.4 kb) (Figures 3A, 4A, and 5A) was normalized to that of GAPDH mRNA (Figures 3B, 4B, and 5B) to assess the change in the expression of these mRNAs as the result of LPS treatment. The distinctive GAPDH mRNA band shown in Figures 3 and 5 indicated that the polyA RNA was of high quality.

View larger version (29K): [in this window] [in a new window]   Figure 3.   Northern blot analysis of polyA RNA isolated from the lungs of saline- and LPS-treated mice using mouse Muc5ac and GAPDH cDNAs as the probes. Each group used three animals, and total RNA was combined for isolation of polyA RNA. The 13.5-kb band, which corresponds to Muc5ac mRNA (26), was barely detected in saline-treated mice but was greatly enhanced in LPS-treated mice from Days 2 to 7 post-treatment (A). The Muc5ac mRNA expression was highest on Day 4. The distinctive GAPDH mRNA band indicates that the RNA sample is of high quality. The stomach (S) epithelium expresses very high concentrations of Muc5ac mRNA. Normalization of Muc5ac mRNA to that of GAPDH is shown in (B).

View larger version (28K): [in this window] [in a new window]   Figure 5.   Northern blot analysis of PolyA RNA isolated from saline- and LPS-treated mice using mouse C2GnT-L cDNA and GAPDH cDNA as the probes. (A) C2GnT-L mRNA was essentially not detectable in saline-treated mice but was expressed in LPS-treated mice. The expression of C2GnT-L mRNA peaked on Day 2 and remained high on Day 4 after LPS treatment. (B) Normalization of C2GnT-L mRNA (2.4 kb band) to GAPDH mRNA is shown. Stomach epithelium (S) expresses very low levels of C2GnT-L mRNA.

The expression of Muc5ac message (13.5 kb) (26) in LPS-treated mice on Day 2 post-treatment was 4.6 times that of the saline-treated mice (Figures 3A and 3B). The abundance of Muc5ac message in LPS-treated mice reached a maximum, which was 9.2 times the control, on Day 4 post-treatment. The expression of Muc5ac mRNA in LPS-treated mice remained high on Day 7, but only 4.8 times that of the saline-treated control. The Muc5ac message in the saline-treated mice was low and did not vary appreciably between Days 2 and 7. The concentration of Muc5ac mRNA in the stomach epithelium was higher than that in LPS-treated airway epithelium.

The expression pattern of C2GnT-M mRNA after LPS treatment was different from that of Muc5ac mRNA. The C2GnT-M mRNA expression was not enhanced appreciably until Day 4 after LPS treatment and continued to increase on Day 7 (Figure 4A). LPS treatment increased C2GnT-M mRNA expression 1.6-fold on Day 4 and 3.0-fold on Day 7 as compared with the saline-treated mice (Figure 4B). The C2GnT-M mRNA level in the stomach epithelium was substantially higher than that in the saline-treated airway epithelium, but lower than that in the airway epithelium on Day 7 post-treatment.

View larger version (29K): [in this window] [in a new window]   Figure 4.   Northern blot analysis of polyA RNA isolated from saline- and LPS-treated mice using mouse C2GnT-M cDNA and GAPDH cDNA as the probes. (A) C2GnT-M mRNA was barely detectable in saline-treated mice, but the expression was enhanced in LPS-treated mice. The expression of C2GnT-M mRNA reached its highest level on Day 7 after LPS treatment. (B) Normalization of the C2GnT-M mRNA (2.4 kb band) to GAPDH mRNA is shown. The level of C2GnT-M mRNA expressed in the stomach epithelium (S) is higher than the saline-treated but lower than the LPS-treated airway epithelium on Day 7.

The expression pattern of C2GnT-L mRNA in LPS-treated mice was different from those of Muc5ac and C2GnT-M mRNA. In saline-treated mice, C2GnT-L mRNA expression did not vary appreciably between Days 2 and 7 post-treatment (Figure 5A). However, in LPS-treated mice, C2GnT-L mRNA expression peaked on Day 2 and decreased somewhat on Day 4, and substantially on Day 7, although still higher than that of the control (Figure 5A). The expression of C2GnT-L mRNA on Days 2, 4, and 7 were 3.3, 2.7, and 1.6 times those of the comparable controls, respectively (Figure 5B). The C2GnT-L mRNA in the stomach epithelium was lower than that in saline-treated airway epithelium.

Western Blotting Analysis of Muc5ac Glycoprotein

The Western blot of lung Muc5ac glycoprotein prepared from saline-treated and LPS-treated mice as detected with HO1 polyclonal antibody (26) is shown in Figure 6. Muc5ac glycoprotein was barely detectable in saline-treated mice, whereas significant amounts of Muc5ac glycoprotein were found in LPS-treated mice from Days 2 to 7. Specifically, an appreciable amount of Muc5ac glycoprotein was observed on Day 2, which reached maximum on Day 4 and then decreased on Day 7 to a level slightly lower than that on Day 2. 

View larger version (83K): [in this window] [in a new window]   Figure 6.   Western blot analysis of proteins separated on a 4% stacking/6% separating polyacrylamide gel and transferred onto a PVDF membrane. One hundred micrograms protein was used for each sample except the stomach sample, which contained 20 µg protein. The membrane was treated with chicken antibody HO1 against mouse Muc5ac (26), followed by horseradish peroxidase- conjugated rabbit anti-chicken IgY (1:4,000) for 1 h and then developed with ECL-Plus. Muc5ac glycoprotein was detected in LPS-treated lungs but not in saline-treated lungs. The stomach Muc5ac glycoprotein (S), which was used as a positive control, showed a very strong signal.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

MCM and mucus hypersecretion are the hallmarks of many lung diseases, including asthma, chronic bronchitis, allergy, and cystic fibrosis (1). Many stimuli, such as bacterial LPS (14), cytokines (30, 31), and elastase (32, 33), are known to induce mucus secretion. For studying mucus hypersecretion, several MCM animal models, including rats (9), hamsters (10), guinea pigs (11), rabbits (12), and mice (8, 34), have been developed. However, for studying the mechanism of the genesis of MCM, a mouse model is the most attractive because many genetically engineered mice with ablation of the genes, such as cytokines (35) and elastase (36), that are potentially involved in the development of MCM are available. The mouse MCM model produced by ovalbumin challenge (8, 34) has been widely used for the study of atopic asthma, but this model is probably more suitable for studying allergen-induced MCM. Our laboratory has been interested in understanding the role of P. aeruginosa in the development of MCM in the lungs of patients with cystic fibrosis. We chose to develop a mouse MCM model induced with P. aeruginosa LPS because it closely resembles the conditions in the lungs of patients with cystic fibrosis.

We were successful in generating a mouse MCM model by a single intratracheal instillation of P. aeruginosa LPS. In response to LPS challenge, the circulating neutrophils, macrophages, and lymphocytes migrated in large numbers to the lung within 2 d, which drastically decreased over the next few days. The occurrence of the peak migration of neutrophils to the lung on Day 2 after LPS treatment agreed with a previous report (37) in which neutrophilic infiltration was examined at 24, 48, and 72 h after a single intranasal instillation of LPS in mice. In another report, Blyth and colleagues (34) found that 24 h after a single intratracheal instillation of 100 µg of Escherichia coli LPS in Balb-C mice, there was a significant increase of neutrophil influx into lung lumen. The reported number of neutrophils (9 × 10⁶/mouse) (34) in BALF at the 24-h time point after treatment was slightly lower than the number of neutrophils (1.22 × 10⁷/mouse) we recovered in BALF 48 h after P. aeruginosa LPS treatment (Table 1). At the 24-h time point, Blyth and coworkers (34) did not detect any eosinophils in the BALF and goblet cells in the epithelium, which agrees with our observation at 48 h post-treatment. However, in their study they did not observe a significant increase of macrophages and lymphocytes, which is at variance with our observation (Table 1). We do not know if the different observation made in our study and that of Blyth and colleagues is due to the difference in mouse species used, i.e., C57BL6 versus Balb-C. We also found that mucus cells did not begin to appear until Day 4 after LPS treatment and reached a maximum on Day 7. This result could explain why Blyth and associates (34) did not detect goblet cells one day after E. coli LPS treatment but did after three LPS treatments at 3-d intervals. We also noted that the influx of inflammatory cells to the lung coincided with the increase of mucin C2GnT-L mRNA in the lung. However, the concentration of C2GnT-L mRNA remained high at the time, i.e., Day 4, when the number of inflammatory cells drastically decreased (Figure 5 and Table 1). C2GnT-L has been identified in leukocytes (6, 7) as well as in the mucus cells of the surface epithelium and the submucosal gland of bovine airway epithelium (21). It is also known that activation of leukocytes is accompanied by an increase of C2GnT enzyme activity at least 10-fold (6). The dramatic increase in C2GnT-L mRNA on Day 2 (Figure 5) reflects the massive influx of inflammatory cells to the lungs (Figure 1B and Table 1) because mucus cells were not detected until Day 4 (Figure 2). Therefore, the high level of C2GnT-L mRNA in the lungs on Day 4 is probably contributed by inflammatory cells and mucus cells. Furthermore, the higher than normal amount of C2GnT-L mRNA detected on Day 7 most likely comes from mucus cells.

The LPS induction of Muc5ac gene expression we observed in mice agrees with a previously published report in which the expression of both MUC5AC and MUC2 genes was upregulated by Gram-positive and Gram-negative bacteria as well as LPS in human airway cell lines and tissues (13). Our study shows that the peak expression of Muc5ac mRNA occurs after the majority of the inflammatory cells have been cleared from the lungs. However, the peak expression time of C2GnT-M mRNA lagged behind the peak expression time for Muc5ac mRNA and Muc5ac glycoprotein as detected with HO1 antibody. The results are not totally unexpected because synthesis of mucin polypeptide precedes the action of glycosyltransferases, the enzymes responsible for mucin glycosylation. However, it is intriguing that the maximal amount of Muc5ac glycoprotein was detected on Day 4 by Western blotting, whereas maximal expression of C2GnT-M and mucus cell hypertrophy occurred on Day 7. There are several possible explanations for the apparent incongruence among these three parameters. First of all, the HO1 antibody was generated using mouse gastric Muc5ac glycoprotein as immunogen. The polyclonal antibody strongly recognizes native gastric Muc5ac glycoprotein but reacts poorly with nonglycosylated peptide tandem repeat (26). Recently, we found that stable transfection of a C2GnT-L cDNA into a human pancreatic cancer cell line, which does not express C2GnT but expresses MUC1, resulted in masking the MUC1 peptide tandem repeat region (19). The same phenomenon may occur in the LPS-treated mice because the peak expression of C2GnT-M coincides with the hypertrophy of mucus cells. This is consistent with our observation that the mucus cells on Day 7 contained large amounts of mucins as revealed by AB-PAS staining, but the Muc5ac glycoprotein on Day 7 reacted poorly with the HO1 antibody. It is possible that Muc5ac glycoprotein detected on Day 2 post-treatment represents primarily sparsely glycosylated species, whereas the Muc5ac glycoprotein detected on Day 4 consists of a mixture of sparsely and moderately glycosylated species, which contained fewer branched glycans. This explanation is supported by the low abundance of C2GnT-L mRNA (Figure 5) and moderate expression of C2GnT-M mRNA (Figure 4) in gastric tissue. These results suggest that gastric Muc5ac glycoprotein contained fewer branched glycans, and the HO1 antibody raised against gastric Muc5ac glycoprotein may recognize poorly the extensively branched glycans that are abundant in lung Muc5ac glycoprotein produced in LPS-treated mice on Day 7 post- LPS treatment. Alternatively, differences in the expression of other glycosyltransferases in gastric epithelium and LPS-treated tracheal epithelium may contribute to different Muc5ac carbohydrate structures in these tissues. Characterization of the mucin glycan structures, including the epitopes recognized by the HO1 antibody, and the expression pattern of gastric and tracheal glycosyltransferases should provide direct answers to these questions.

The decrease of Muc5ac mRNA on Day 7 after its peak expression on Day 4 is of great interest. This result suggests that Muc5ac mRNA may not be around after Muc5ac polypeptide is made. The implication is that maximal expression of Muc5ac mRNA occurs only in mucus cells that are undergoing metaplastic process and "mature" mucus cells may have low or no expression of Muc5ac mRNA. If this is the case, a Muc5ac mRNA may not be an accurate indicator of "mature" mucus cells. In situ hybridization using Muc5ac-specific cDNA probe coupled with immunohistochemical staining using Muc5ac-specific antibody should provide the answer to this important question.

Our proposed sequence of events beginning from the instillation of LPS in the mouse lung to the induction of MCM is summarized in Figure 7. Instillation of LPS in the lung induces the release of inflammatory mediators (38), which leads to the activation of neutrophils and endothelial cells and upregulates the expression of C2GnT-L mRNA in these cells. Upregulation of C2GnT-L increases the synthesis of the mucin-type core 2 glycans terminated with sialyl Lewis x, which is an essential part of the ligands for P- and L-selectins (16). Adhesion of neutrophils to endothelial cells is initiated by the interaction of P-selectin found on activated endothelial cells and circulating leukocytes and platelets (15, 16) with P-selectin ligand PSGL-1 found on the surface of leukocytes and platelets. Interaction of L-selectin on leukocytes with core 2-associated sulfated sialyl Lewis x on the surface of endothelial cells also plays crucial roles in the early phase of neutrophil migration. Firm binding of activated neutrophils to the endothelial cells and subsequent extravasation across the endothelium is mediated in part by E-selectin. Upon reaching the lung, the activated neutrophils release their contents, which include elastase, to induce MCM/hyperplasia (10). This process starts with the synthesis of Muc5ac mRNA (Figure 3) (39), which is followed by the production of mucin glycan-specific glycosyltransferases, including C2GnT-L (Figure 5) and C2GnT-M (Figure 4) in the mucus cells, to produce fully glycosylated Muc5ac (Figure 6) detected by AB-PAS stain (Figure 2).

View larger version (21K): [in this window] [in a new window]   Figure 7.   Proposed sequence of events between LPS treatment and the induction of MCM. LPS treatment induces the secretion of cytokines/chemokines from the epithelium (38), which trigger the activation of neutrophils in which the C2GnT-L mRNA level was increased. The activated neutrophils adhered to, rolled on, and extravasated across the endothelium to reach the lung to release their contents. These released products initiate in the lung a series of events, including enhancement of the expression of Muc5ac, C2GnT-M, and C2GnT-L mRNAs, which result in the production of fully glycosylated Muc5ac detected by AB-PAS.