Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Secretory processes in the airways of mammalia are important for maintenance of physiologic conditions. Different secretory cells contribute to the tracheobronchial fluid (1). In the human large airways, the goblet cells of the surface epithelium and the serous and mucous cells of the glands are the principal secretory cells. The latter are thought to produce most of the secretions. Furthermore, several pathologic situations are characterized by altered tracheobronchial secretions. Chronic obstructive lung disease, for example, is characterized by increased mucous secretory components (2). Similarly, cystic fibrosis is characterized by an altered airway secretion caused by the inborn malfunction of the cystic fibrosis membrane conductance regulator. This protein has been localized in secretory cells of the airway, especially in the serous gland cells (3). Gland cells have been isolated and cultured previously by several authors (6) and secretory products have been localized in gland cells by histochemical methods (e.g., 10), lectin- and immunohistochemistry (11), and in situ hybridization (3, 16).

Any molecular understanding of the mechanisms involved in proliferation and differentiation of serous and mucous cells requires an efficient method for both isolation of secretory cell types and isolation of cell type-specific genes. Therefore, the goal of the present study was to (1) develop a procedure to obtain fractions enriched in serous and mucous gland cells of the large airways, and (2) to isolate cell type-specific transcripts from these cells. The pig (Sus scrofa) was used as model because material is easily available and the morphology of the porcine airways is similar to that of the human airways (19; own unpublished observations).

Our data show that secretory gland cells can easily be separated by density gradient centrifugation. We also show that differential display of mRNA (20, 21) is the method of choice for identifying secretory cell-specific transcripts.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Separation of Secretory Gland Cells

Porcine tracheae were obtained from the Munich slaughterhouse. Shortly after the animals were killed by electroshock, tracheae were explanted and transported to the laboratory immersed in 0.1 M phosphate-buffered saline (PBS; 4°C). Secretory gland cell recovery was by the method of Tournier and colleagues (9). Adhering tissue was removed and the airways were washed in PBS. Following removal of the surface epithelium, mucosal tissue of the posterior part of the tracheae was dissected. The whole separation procedure is shown schematically in Figure 1. For each experiment, mucosal tissue of five tracheae was pooled and transferred into 15 ml of the cell buffer described by Culp and associates (22) (110 mM NaCl, 5.4 mM KCl, 26.2 mM NaHCO₃, 1 mM Na₂HP₄ · H₂O, 0.8 mM MgSO₄ · 7 H₂O, 1.8 mM CaSO₄ · 2 H₂O, 20 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid, 2 mg/ml bovine serum albumin, and 1 mg/ml glucose, pH 7.4) together with a mixture of enzymes and antibiotics (200 U/ml deoxyribonuclease I [type IV], 200 U/ml hyaluronidase [type IV], 0.1 mg/ml elastase, 200 U/ml collagenase [type I], 0.05% pronase, 200 U/ml penicillin, 200 µg/ml streptomycin, and 2.5 µg/ml amphotericin B). Enzymes and antibiotics were purchased from Sigma (Deisenhofen, Germany). The tissue digestion was for 5 h at 37°C. After digestion, the cell suspension was filtered to remove multicellular aggregations. Monodispersed cells were separated by discontinuous density gradient centrifugation using two Percoll gradients in incubation buffer of 1.02 and 1.04 g/ml density. The cell suspension was layered above the two solutions of Percoll and the centrifugation was performed at 500 × g for 30 min at room temperature. Three different cell suspensions were collected: fraction 1 was collected from the border of the buffer to the Percoll gradient (1.02 g/ml), an intermediate fraction 2 from the border of the two Percoll gradients (1.02 and 1.04 g/ml), and a fraction 3 from the bottom of the tube. The cell suspensions were washed in incubation buffer and centrifuged at 500 × g for 10 min. At the beginning of this study, different media, enzymes, and density gradients were used for cell separation. The procedure described here, which resulted in a maximum of cell yield and purity of cell fractions, was independently performed 15 times. The cell pellets were used for mRNA extraction or morphologic analysis.

View larger version (41K): [in this window] [in a new window]   Figure 1.   Flowchart showing the procedures used for isolation and characterization of secretory gland cells of the airways.

Morphologic Methods, Cell Numbers, and Cell Viability

Cell fractions of 10 independent experiments were used for morphologic analysis. Cells were fixed in 3.5% glutaraldehyde in PBS (3 h, 4°C) and embedded in Araldite using standard procedures. Ultrathin sections were contrasted with uranyl acetate and lead citrate and viewed under a Zeiss EM 900 electron microscope. Cells containing granules filled with electron-dense material were recognized as serous cells, whereas cells containing electron-lucent material in their granules were recognized as mucous cells. Cells without secretory granules were grouped as nonsecretory cells. The cell yields in the different cell suspensions were determined using a counting chamber. Cell viability was assessed by staining the cells with erythrocin B.

Isolation of mRNA and Differential Display Polymerase Chain Reaction

Cell fractions of seven experiments were used for RNA isolation; cells of five of these experiments were also used for morphologic analysis. The mRNA of serous and mucous cell fractions was isolated using Quick-Prep mRNA Isolation Kit (Pharmacia Biotech, Freiburg, Germany). Reverse transcription was performed using a cDNA synthesis kit (Pharmacia Biotech), 250 ng of mRNA, and a 5'-T₁₂GC or 5'-T₁₂AG oligonucleotide as primer. Differential display polymerase chain reaction (DD-PCR) was performed according to Liang and Pardee (20) in a nonradioactive modification of the method of Lohmann and coworkers (23) using 23 different arbitrary 5' primers (Operon Technologies, Alameda, CA; Pharmacia Biotech) and either the 5'-T₁₂GC or 5'-T₁₂AG oligonucleotide as tailing primer. A total of 43 cycles of PCR were performed with cycles times of 30 s at 94°C, 1 min at 42°C and 1 min at 72°C in an Omn-E thermo-cycler (MWG, Ebersberg, Germany).

The PCR products were separated on horizontal polyacrylamide gels (15% CleanGel-48S; Pharmacia Biotech) containing 7 M urea. The cDNA pattern was visualized by silver staining as described elsewhere (23). PCR products that were present in only one of the two cell types were directly transferred from the gel to a PCR tube for reamplification using the initial set of primers and PCR conditions but with dNTP concentration increased to 200 µM. The products of the second reamplification were purified, separated in agarose gels as described elsewhere (23), and cloned into PBS-plasmid (Stratagene, Heidelburg, Germany).

Dot-blot and Northern Blot Analyses and Determination of Nucleotide Sequence

Dot blots were used to confirm the data obtained by DD-PCR. A total of 400 ng of poly(A)⁺ RNA from the serous, intermediate, and mucous cell fractions were dot-blotted onto nylon membranes (Biodyne B; Pall, Portsmouth, MA) and hybridized with ³²P-dATP-labeled PCR fragments of interest. Labeling, hybridization, and washing were performed following standard procedures (24). Northern blots of 20 µg total RNA from digested mucosal tissue were prepared (24) and probed with ³²P-dATP-labeled PCR fragments. After washing, autoradiography was performed. The nucleotide sequences of the cloned PCR fragments were determined using the T7 Sequencing Kit (Pharmacia Biotech). The nucleotide sequences were used to perform BLAST homology searches at the website of the National Center of Biotechnology Information (NCBI; http://www.ncbi.nlm.nih.gov).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Separation of Secretory Gland Cells by Density Centrifugation

The procedure to obtain fractions enriched in serous and mucous gland cells is shown schematically in Figure 1. For each experiment, mucosal tissue was prepared from five tracheae and enzymatically digested. A total of 1.5 ± 0.4 g of tissue was used for each experiment. Care was taken to remove all of the surface epithelium. The cell suspension obtained was filtered and applied to density gradient centrifugation. Centrifugation resulted in two sharp areas of cellular material at the density steps and some material at the bottom of the tube. Standard transmission electron microscopy (TEM) of the cellular material revealed that the majority of the cells were secretory cells with only few white and red blood cells, fibroblasts, smooth muscle cells, intercellular fibers, and cell debris. Mucous cells were detected in fractions 1 and 2 (see Figure 1), whereas serous cells were present in fraction 3 (collected from the bottom of the centrifugation tube). The ultrastructure of cells from the different fractions is shown in Figure 2. Typically cells detected in the mucous and intermediate fractions (fractions 1 and 2) contained electron-lucent material in their granules (Figures 2a and 2b). Such cells were recognized as mucous cells. In fraction 3, the majority of cells were recognized as serous cells due to their granules filled with electron-dense material (Figure 3, panel 3). The cell yields, cell viability, and percentages of secretory cells before and after the centrifugation are shown in Table 1. The results demonstrate that by using enzymatic digestion and discontinuous density gradient centrifugation, fractions enriched in serous and mucous gland cells can be obtained.

View larger version (73K): [in this window] [in a new window]   Figure 2.   Ultrastructure of cells in the three fractions obtained after density gradient centrifugation. Mucous cells showing large electron-lucent granules (a and b) were isolated from fractions 1 and 2. Serous cells (c) containing small, electron-dense granules were isolated from fraction 3. Bar in (a) and (c) = 7 µm; bar in (b) = 15 µm.

View larger version (65K): [in this window] [in a new window]   Figure 3.   DD-PCR analysis of cells of the mucous (m) and serous (s) cell fractions (fractions 1 and 3). DD-PCR amplifications were carried out in parallel using primer combinations T12AG/OPA13 ( panel 1), T12AG/OPA14 ( panel 2), and T12AG/OPA4 ( panel 3). Whereas in panels 1 and 2 no differences are detectable in the transcript pattern in mucous and serous cells, in panel 3 five PCR products (arrowheads) can be obtained only in RNA from serous cells. The PCR product indicated by a large arrowhead was eluted, cloned, and proven to correspond to a serous cell-specific gene by dot-blot analysis (see Figure 4).

View larger version (69K): [in this window] [in a new window]   Figure 4.   Confirmation of differential expression by dot-blot analysis. (Panel A) The serous cell-specific PCR product shown in Figure 3 (thick arrowhead, panel 3) was radioactively labeled and hybridized to dot blots containing mRNA from serous (1), intermediate (2), and mucous (3) cell fractions. (Panel B) Control experiment showing equal loading of RNA by hybridization of a probe expressed in both serous and mucous cells to the filter shown in panel A. (Panel C) Northern blot of total RNA from unseparated cells using the same probe as in panel A. A single band of about 2.6 kb can be detected. Small arrows indicate positions of Sus scrofa 18S and 28S rRNA.

At the beginning of the present study we also used histochemical (periodic acid Schiff reaction, Alcian blue) and lectin- and immunohistochemical methods to determine the cell purity (data not shown). Because TEM was as fast as the other methods and allowed a clear discrimination between the cell types, we used this method for all 15 cell separation procedures analyzed in this study.

Isolation and Characterization of Cell Type-specific Genes by DD-PCR

The PCR-based differential screening method of Liang and Pardee (20) provides a powerful and efficient tool for detecting variations in gene expression and was used to isolate genes expressed in either the mucous or the serous cell fraction. Poly(A)⁺ RNA was reverse transcribed with either tailing primer T₁₂AG or T₁₂GC. PCR amplification of cDNA was performed using one of these tailing primers and 23 arbitrary decamers. For differential display of mRNA we used the nonradioactive modification of the original procedure (21, 23). A total of 990 PCR products could be obtained when using 23 arbitrary decamers and tailing primer T₁₂AG; 690 PCR products could be obtained with 23 arbitrary decamers and tailing primer T₁₂GC. A typical example of a DD-PCR experiment is presented in Figure 3. PCR amplifications were carried out in duplicate and the products obtained were analyzed in parallel. The three panels shown correspond to three different primer combinations: T₁₂AG/OPA4, T₁₂AG/ OPA13, and T₁₂AG/OPA14. Figure 3 shows that using primer combinations T₁₂AG/OPA13 and T₁₂AG/OPA14, no differences can be observed between mucous (m) and serous (s) cells. When using primer combination T₁₂AG/ OPA4, however, five major bands (marked by arrowheads) can be seen only in lanes from the serous cell fraction; they are absent in the mucous cell fraction. This observation suggests that expression of the genes represented by the five bands is specifically activated in serous cells. Similar experiments were carried out using 46 different primer combinations for reverse transcriptase-PCR amplification. Most primer combinations did not produce any detectable differences between the two cell populations. Out of 1,680 bands analyzed, eight transcripts could be observed to be activated specifically in cells of the mucous cell fraction and 10 transcripts were present only in the serous cell fraction.

Isolation of RNA from cell fractions was repeated seven times in independent experiments and the material was used for DD-PCR. The pattern of PCR products was highly reproducible for all primer combinations.

All bands corresponding to differentially expressed genes were isolated and cloned. To confirm the differential expression, conventional dot-blot analysis was performed. Poly(A)⁺ RNA of cells of the serous, intermediate, and mucous fractions was blotted onto nylon membrane and hybridized to radiolabeled inserts. Figure 4, panel A shows, for example, hybridization of a serous cell-specific PCR fragment to RNA from the three different cell fractions. Out of 18 PCR fragments analyzed, nine revealed a positive cell type-specific signal. Four fragments hybridized much stronger with RNA isolated from the mucous and intermediate cell fractions while five hybridized with RNA from the serous cell fraction.

To obtain information about the transcript size and to determine whether a single transcript is recognized or whether different cloned sequences bind to transcripts of the same size (indicating that these PCR products are parts of one mRNA), Northern blot analysis using total RNA from unseparated cells was performed. PCR fragments that showed cell type-specific expression in the dot-blot analysis were used as probes. Figure 4, panel C shows the results obtained when using clone 4M1(4) as probe and demonstrates that the labeled cDNA binds to a single transcript of about 2.6 kb. Hybridization using the other two probes also revealed single transcripts (data not shown). All nine cell type-specific cDNA fragments have been sequenced and compared with nucleotide and protein databases. The sequence of clone 4M1(4) shown in Figure 5 is 80% identical to a Sus scrofa expressed sequence tag of unknown function (data not shown). The remaining clones displayed no significant sequence similarity and most likely represent 3' untranslated regions of various genes.

View larger version (14K): [in this window] [in a new window]   Figure 5.   Nucleotide sequence of the serous cell-specific PCR fragment used as probe in Figure 4, panels A and C. Positions of the 3' random primer (Operon A4) and the 5' T12AG tailing primer used for DD-PCR are underlined. A BLAST database search revealed 80% identity of this sequence to a Sus scrofa expressed sequence tag (EST) of unknown function and location.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

This study describes the isolation and characterization of cell type-specific genes of secretory cells of the porcine airways glands. So far, secretory gland cells have been isolated by enzymatical digestion (7, 17). No further studies on cell type-specific properties were possible because isolated cell types have not been available so far. The results described here demonstrate that by using enzymatic digestion and discontinuous density gradient centrifugation, fractions highly enriched in serous and mucous gland cells can be obtained reproducibly. Because the cells were viable and could be obtained in large quantity (see Table 1), it should now be possible to use the separation procedure for a variety of cell and molecular biological experiments including in vitro cultivation. TEM was used as method of choice to determine the purity of the obtained cells fractions because this method (1) allows a clear discrimination of secretory and nonsecretory as well as mucous and serous cells, and (2) is faster and perhaps easier to perform than other methods, such as immunohistochemistry.

During recent years DD-PCR has evolved to an important technique in molecular biology and has been reviewed elsewhere (25). Differential display of mRNA has proven powerful in the identification of specific genes (26, 27) and detection of even minor variations in gene expression (28). One problem associated with the DD-PCR technology is the fact that a significant number of DD-PCR clones may contain nondifferential inserts (27). Therefore, differential expression of cloned DD-PCR fragments must be confirmed independently. In the present study, we used simple dot-blot analysis of mRNA of cells of the serous, intermediate, and mucous fractions. We found about 50% of the cloned PCR fragments to represent differentially expressed genes.

The nucleotide sequences of cell type-specific PCR fragments were determined and used for homology search. However, all fragments seem to originate from untranslated parts of mRNAs. This may be due to the bias of the DD-PCR method to amplify 3' regions of transcripts. One clone was found to be nearly identical to a pork-expressed sequence tag of unknown function.

In the present article over 95% of the PCR fragments produced by 46 different primer combinations were found to be expressed in both cell fractions. Only a few PCR products appear to correspond to cell type-specific genes, suggesting that the two secretory cell types are closely related. This similarity in the transcript pattern is consistent with the previous observation of secretory gland cells exhibiting morphologic and biochemical characteristics of both serous and mucous cells in vivo and in vitro (17).

Our study has demonstrated that cell type-specific genes of secretory airway glands can be isolated simply by means of DD-PCR following density gradient cell separation. Secretory cell type-specific molecular markers may now be used for analysis of cell differentiation in cell culture experiments and for in situ characterization of cell types in developmental studies or pathologic conditions. The global approach described here could facilitate a molecular understanding of the pathogenesis of many respiratory diseases.