Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Chronic obstructive pulmonary disease (COPD) is characterized by an inflammatory process in the membranous and respiratory bronchioles that results in early airway closure, increased peripheral airway resistance, and destruction of the alveolar surface (1). Although cigarette smoking is the major risk factor, only 15-20% of heavy smokers develop COPD (2, 3). Epidemiological studies suggest that childhood infections represent an independent risk factor for the development of COPD in adulthood (4, 5) and studies from our laboratory have focused on the role of adenovirus infection as one of the etiologic agents responsible for the increased risk. This virus was selected because it is a common cause of bronchiolitis and pneumonia and it is known that acute infections are followed by persistence of viral DNA in tonsils and peripheral blood lymphocytes (reviewed in Ref. 6). Studies from our laboratory subsequently demonstrated that more adenoviral E1A DNA can be detected in the lungs of COPD patients than in controls (6) and that the E1A protein is expressed in epithelial cells lining the airways, alveoli, and submucosal glands of human lungs (7). Reports from several other laboratories suggest that the E1A region of the adenoviral genome might play an important role in enhancing the inflammatory process. Duerksen-Hughes and coworkers (8) have reported that epithelial cells transfected with E1A show an increased susceptibility to destruction by cytokines such as tumor necrosis factor. The competitive binding of the E1A proteins to the product of the retinoblastoma (RB) gene results in the release of transcription factors normally bound by RB (9, 10) and a disruption in the inhibitory control of RB product on cell division and tissue growth. Also, Liu and Green (11) have shown that the adenoviral E1A, a transcriptional activator, can interact with the DNA-binding domains on several transcription factors and in this manner can be recruited to diverse promoters of host genes. We speculate that the upregulation by E1A of genes encoding inflammatory cytokines may drive the inflammatory process present in the airways of all smokers to produce airway obstruction.

Glucocorticoids are widely used in the treatment of inflammatory diseases (12) and the therapeutic effects of inhaled steroids are likely due to a combination of actions, culminating in a decrease in airway inflammation (13). The effectiveness of glucocorticoids in the treatment of inflammatory diseases of the lung is probably linked to their ability to alter the expression of genes that mediate the inflammatory process. The possibility that latent adenovirus infection amplifies the inflammatory response present in the airways and the fact that glucocorticoids are effective antiinflammatory drugs led us to postulate that genes relevant to the inflammatory process might be regulated by both adenovirus E1A and glucocorticoids.

Differential display (14) is used to identify and isolate genes that are differentially expressed. It has been successfully used to identify oncogenes (15, 16), tumor suppressor genes (16, 17), and genes involved in development (18). Widespread use of this technique has led to a number of improvements including advances in primer design (19). In this study, we used the differential display assay to identify adenovirus E1A-regulated genes and glucocorticoid-regulated genes in A549 human lung epithelial cells. Our strategy was to clone and sequence the complementary DNAs (cDNAs) identified by the differential display assay. Using this approach we have identified groups of genes that are regulated by adenovirus E1A, others that are regulated by glucocorticoids, and those that are regulated by both E1A and glucocorticoids.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Cell Culture and RNA Preparation

All cells were cultured in Dulbecco's medium containing 10% fetal calf serum. A549 human lung cells and two human bronchial epithelial cell lines, H441 and BEAS-2B, were obtained from the American Type Culture Collection (Rockville, MD). They were grown to confluence, split into two groups, and were treated for 24 h either with or without 1 µM dexamethasone. A549 cells were used to identify and subsequently to isolate glucocorticoid-regulated genes by differential display and these cells along with the other two cell lines were used to verify the glucocorticoid regulation of these genes. Three clones (E4, E11, E20) of A549 cells stably transfected with a plasmid containing the adenovirus E1A gene (pE1Aneo, kindly provided by Dr. Frank Graham, McMaster University, Hamilton, ON) using neomycin selection (20) and two clones (C4, C8) of A549 cells transfected with a control plasmid containing the neomycin resistance gene but lacking the adenovirus E1A gene (20) were also studied. These cells were cultured in the presence of 0.5 mg/ml neomycin. E1A-expressing cells (E20) and control cells (C4) were used to identify adenovirus E1A-regulated genes by differential display; and these cells along with the other three transfected A549 clones (E4, E11, and C8) and the parent A549 cells were used for verification of this regulation. Total RNA was isolated from cells by the acid guanidinium thiocyanate-phenol-chloroform extraction method (21). DNase treatment of total RNA was performed according to Liang and coworkers (22). Fifty micrograms of total RNA were incubated for 30 min at 37°C with 10 units of RQ1 DNase I (Promega, Madison, WI) in 10 mM Tris-HCl (pH 8.3)- 50 mM KCl- 1.5 mM MgCl₂. After phenol extraction, RNA was precipitated with ethanol. The isolated RNA was quantified by monitoring absorbance at 260 nm.

mRNA Differential Display

Three separate RNA preparations were made from the control and glucocorticoid-treated A549 cells and from the E20 and C4 cells and analyzed by the differential display assay. This assay was performed using the supplier protocol (GenHunter, Boston, MA). Total RNA (0.2 µg) as described previously was used as a template to generate a cDNA pool using reverse transcriptase (RT) and one of three anchored oligo(dT) primers (H-T₁₁G, H-T₁₁A, H-T₁₁C; GenHunter). An aliquot (2 µl) of each cDNA pool was then amplified using one of the arbitrary 13-base upstream primers (H-AP1-16; GenHunter) and one of the three anchored oligo(dT) (downstream) primers. Each polymerase chain reaction (PCR) contained 0.2 µM upstream primer, 0.2 µM oligo(dT) downstream primer, 2 µM dNTP, 0.5 µM [-³⁵S]dATP (1,000 Ci/mmol) or 0.1 µM [-³³P]dATP (1,000-3,000 Ci/mmol), 10 mM Tris-HCl (pH 8.4), 1.5 mM MgCl₂, 1 unit of Taq polymerase (AmpliTaq; Perkin-Elmer, Mississauga, ON, Canada), and 2 µl of cDNA in a total volume of 20 µl. In experiments to identify glucocorticoid-regulated genes, amplification was performed at 94°C for 15 s, 40°C for 2 min, and 72°C for 30 s for 40 cycles followed by a 5-min extension reaction at 72°C. The differential display assays used to identify adenovirus E1A-regulated genes were performed using 30, 35, and 40 cycles of amplification. Following amplification, the products of the PCR were separated by electrophoresis on denaturing 6% polyacrylamide gels. The gels were transferred to Whatman (Clifton, NJ) 3MM paper, dried, and autoradiographed.

Reamplification, Cloning, and Sequencing of cDNA Bands

The differentially expressed cDNAs identified by autoradiography were then recovered from the electrophoresis gel. The cDNA band was cut out using a scalpel blade, the gel slice rehydrated in water and boiled, and the DNA recovered by ethanol precipitation. The eluted DNA was reamplified using the same upstream and anchored primers used in the original PCR. An aliquot of the PCR was then analyzed on a 1.5% agarose gel and the PCR products detected by ethidium bromide staining. The reamplified PCR products were then cloned. The ends of the PCR products were polished using T4 DNA polymerase and ligated to pCR-Script SK(+) (Stratagene, La Jolla, CA) by blunt end ligation. Minipreps of each clone were prepared and those containing cDNA inserts were selected. Southern blot analysis of 10 individual clones was used to determine if a reamplified cDNA contained one unique cDNA species or several different species. Double-stranded DNA sequencing of selected cDNA clones was performed using T3 or T7 oligonucleotide primers (Stratagene). DNA sequence identity searches were performed using the basic local alignment search tool (BLAST) search program (23) and the DNA sequence databases, GenBank (gb; managed by the National Centre for Biotechnology Information), the European Molecular Biology Laboratory (emb), and the DNA Data Bank of Japan (dbj) were searched. The best match from the DNA sequence databases, in which the length of the match exceeded 50 nucleotides and the percentage sequence match was > 90%, was accepted for identification. The number of matching nucleotides divided by the nucleotide length of the shorter of either the cloned cDNA or the database match was used to determine the degree of sequence identity.

Northern Blot Analysis

Twenty micrograms of total RNA isolated from control and dexamethasone-treated A549 cells was electrophoresed through a 0.86 M formaldehyde gel and transferred to Hybond-N membranes (Amersham, Oakville, ON, Canada). Membranes were hybridized in 5× standard saline citrate (SSC)-5× Denhardt's -1% sodium dodecyl sulfate (SDS)-100 µg/ml salmon sperm DNA-50% formamide at 42°C overnight with radiolabeled DNA probes that were generated by random-prime labeling (24) using cDNA inserts excised by HindIII digestion and purified by agarose gel electrophoresis or using a cDNA of 1.6 kbp for the 18S rRNA (25). The 18S rRNA was chosen as an internal reference among other constitutive genes because its expression was not affected by dexamethasone treatment (1 µM for 24 h) in A549 cells (Figure 2a). After washing twice with 1× SSC containing 0.1% SDS at room temperature for 15 min, followed by washing with 0.25× SSC containing 0.1% SDS at 55°C for 30 min, the membranes were autoradiographed. Three separate RNA preparations from the control and dexamethasone-treated A549 cells was analyzed in this manner for each of the cloned differentially displayed cDNAs that proved to be unique by DNA sequence analysis.

View larger version (29K): [in this window] [in a new window]   Figure 2.   Confirmation of differential mRNA expression of glucocorticoid-regulated genes identified by the differential display assay. (a) Northern blot analysis. Total RNA (20 µg/lane) obtained from control (C) and dexamethasone-treated (D) A549 cells was hybridized with 32P-labeled cDNA probe of clone GC7 and 18S rRNA. The autoradiographs showing hybridization of GC7 probe to a 4.6-kb mRNA (arrowhead) and 18S hybridizing RNA are representative of three separate experiments that gave similar results. To quantitate the results of the Northern analysis the optical density of the GC7 hybridizing band was first normalized with the corresponding 18S rRNA band. The bar graph below the autoradiograms shows the average scan units ± SD from the three experiments, in which the normalized optical density of the GC7-hybridizing band from the dexamethasone-treated lane (D) is expressed relative to that of the corresponding band in the control lane (C), which was set at an arbitrary value of 1. (b) RNase protection assay. 32P-Labeled antisense RNA probes for cDNA clones of interest were mixed and hybridized with 20 µg of total RNA from either control (C) or dexamethasone-treated (D) A549 cells. An 18S rRNA cRNA probe of 180 nucleotides, which protects a fragment of 95 nucleotides (25), was included in each assay to control for lane loading and RNA degradation. The autoradiograms showing GC2, -7, and -8 protected fragments of approximately 300, 275, and 210 nucleotides, respectively (arrowheads), and the corresponding 18S cRNA protected fragment are representative of three assays that gave similar results. To quantitate the results of the RPA the optical density of the GC2, -7, and -8 protected fragments was first normalized with the corresponding 18S rRNA band. The bar graphs below the autoradiograms show the average scan units ± SD from the three experiments, in which the normalized optical density of the GC hybridizing bands from the dexamethasone-treated lane (D) is expressed relative to that of the corresponding band in the control lane (C), which was set at an arbitrary value of 1.

RNase Protection Assay

The RNase protection assay (RPA) was performed using the HybSpeed RPA protocol and kit (Ambion, Austin, TX). Labeled antisense RNA probes were synthesized by either T3 or T7 RNA polymerase from linearized plasmids containing cDNA inserts of interest. The poly(A) tail in the cDNA sequence was used to establish the sense/antisense orientation. The labeled probe was mixed with 20 to 80 µg of total RNA from either control or dexamethasone-treated A549 cells and incubated under conditions for hybridization of complementary transcripts. An 18S rRNA cRNA probe of 180 nucleotides that protects a fragment of 95 nucleotides (25) was added to each RNA sample to control for lane loading and RNA degradation. Following hybridization, a 30-min digestion with RNase T1 and RNase A at 37°C was performed. The samples were electrophoresed through a 6% polyacrylamide gel and autoradiographed. The RPA was repeated three times for each of the cloned differentially displayed cDNAs that proved to be unique by DNA sequence analysis.

Semiquantitative RT-PCR

A specific primer pair was designed for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA using the Oligo Selection Program (26). GAPDH, a housekeeping gene (27), was used as an internal control in this RT-PCR analysis. Similarly, primer sets were designed for each of the glucocorticoid-regulated cDNA clones, except for GC7 and GC8, whose differential expressions were confirmed by Northern analysis or RPA, and for GC12, which was found by sequence analysis to be identical with GC2. The GAPDH probe was selected to hybridize to a portion of the target segment between the primers, which ensured the identity of the amplified DNA of 446 bp. Sequences of the oligonucleotide primers and GAPDH probe were as follows:

GAPDH sense 5'-CCCATCACCATCTTCCAG-3'

GAPDH antisense 5'-ATGACCTTGCCCACAGCC-3'

GAPDH probe 5'-CTAAGCATGTGGTGGT-

GCA-3'

GC1 sense 5'-GCACATATTAGGACCA-

AAG-3'

GC1 antisense 5'-AGTTAGTCTGTCTCCATC-3'

GC2 sense 5'-GAATATTGTAAGTCAGC-

CAC-3'

GC2 antisense 5'-TAAGAAACTCCTCTGGA-

AAG-3'

GC3 sense 5'-TTTGCTCTGTTTGTTGGG-3'

GC3 antisense 5'-TGGGATCTCACTATGTTG-3'

GC4 sense 5'-CACCTAAGAAAGCAAAGA-3'

GC4 antisense 5'-TGGTCCATGATAGTGAAG-3'

GC5 sense 5'-TGTTCATGTAGATGTCCC-3'

GC5 antisense 5'-GAATGTCAGAAGGTGA-

TAG-3'

GC6 sense 5'-GTAAAACTGCAAGCTGAC-3'

GC6 antisense 5'-TTGGAGCGATCCATTATC-3'

GC9 sense 5'-TTCAACTGTACTGAAAGGG-3'

GC9 antisense 5'-TTGGAAGGGGTATTTGAC-3'

GC10 sense 5'-CCAGGGAAAGGAATATTG-3'

GC10 antisense 5'-ATGCTTTGTTTCCAGGTG-3'

GC11 sense 5'-TGAGCAGGGGGTTCCAGC-3'

GC11 antisense 5'-CCTCTGACCAAAGAAAA-

CACAGC-3'

GC13 sense 5'-CAGTACAGAAAGGAAAGG-

ATA-3'

GC13 antisense 5'-ATGAAGGAGATTATGTT-

GTG-3'

GC14 sense 5'-GAACATTGCTAAAGTGC-

TTC-3'

GC14 antisense 5'-TTTGAGGTCTTATGGTCC-3'

GC15 sense 5'-GAATTGAACCTGAACCTG-3'

GC15 antisense 5'-TAGCCTCCCAAAGTATTG-3'

Two RNA preparations were isolated from control or dexamethasone-treated A549, H441, and BEAS-2B cells, and from A549 cells and A549 cells transfected with the control (C4 and C8) or E1A expression (E4, E11, and E20) plasmid. From the first of the two series of RNA preparations, 15, 3.75, and 5 µg of DNase I-digested total RNA from control and dexamethasone-treated A549 cells, from control and dexamethasone-treated H441 and BEAS-2B cells, and from the A549 cells and A549 cells transfected with control (C4 and C8) or E1A expression (E4, E11, and E20) plasmids, respectively, were reverse transcribed in a volume of 300, 75, and 100 µl, respectively, using random hexamers and Moloney murine leukemia virus (Mo-MuLV) RT.

To determine the optimal cycle number for GAPDH PCR, 20 µl of the heat-treated reverse transcription reaction from the first series of RNA was amplified using the GAPDH primer pair in a 200-µl PCR. The temperature profile for each cycle was 94°C for 15 s, 49°C for 45 s, and 72°C for 45 s and 21 to 39 cycles were tested at 3-cycle increments. The last cycle was followed by incubation at 72°C for 5 min. At the end of 21, 24, 27, 30, 33, 36, and 39 cycles, 20 µl of the PCR was removed and the amplification products were analyzed on a 1.5% agarose gel stained with ethidium bromide and transferred onto Hybond-N membranes (Amersham). These membranes were hybridized with labeled GAPDH probe and autoradiographed.

A parallel series of PCRs was performed using the same conditions and cycle patterns, except that the GAPDH primers were tested this time in the same tube with each of the cDNA primer sets listed previously (except for the GC2 primers) on cDNA generated by the RT reaction from the control and dexamethasone-treated A549 cells, with the primers for GC2 or GC10 on cDNA from the control and dexamethasone-treated H441 and BEAS-2B cells, or with the primers for GC2, GC10, or GC11 on cDNA from A549, C4, C8, E4, E20, and E11 cells. In this second set of PCRs, 20 µl of the reaction was removed after 21, 24, 27, 30, 33, 36, and 39 cycles and the amplification products were analyzed as described previously. In this case the membranes were hybridized with the respective labeled cDNA probe and only the products amplified after 24 cycles were tested with the GAPDH probe.

From the second of the two series of RNA preparations, 7.5, 2, or 2.5 µg of DNase I-digested total RNA from control and dexamethasone-treated A549 cells, from control and dexamethasone-treated H441 and BEAS-2B cells, or from A549 cells and A549 cells transfected with control (C4 and C8) or E1A expression (E4, E11, and E20) plasmid, respectively, were reverse transcribed as described previously in a volume of 150, 40, and 50 µl, respectively. In a third set of PCRs, cDNAs in 10 µl of the heat-treated sample from this second series of RT reactions were amplified using the GAPDH primers in the same 100-µl PCR with the primer sets as described previously for the second set of PCRs on the respective RT-generated cDNAs. The temperature profile used previously from 21 to 30 cycles in 3-cycle increments was tested for all cDNAs except GC4, which was tested from 24 to 33 cycles. At the end of each three-cycle increment, starting from the shortest amplification interval, 20 µl of the PCR was removed and the amplification products analyzed as for the products from the second PCR series. As amplification of GC5 was not achieved in any of the preceding PCRs, this cDNA was tested further at 40 cycles.

Quantitation of Autoradiographic Signals from mRNA Differential Display, Northern Blot Analysis, RPA, and Semiquantitative RT-PCR

The autoradiographic signal was quantified with an Ultroscan laser densitometer at 633 nm in the one-dimensional mode and GSXL analysis software version 2.1 (Pharmacia-LKB, Uppsala, Sweden). Several autoradiographic exposures of each blot were taken to ensure that the signal from each band was in the linear range of the film.

For mRNA differential display, the densitometric signals of the differentially expressed bands in the lanes from the control A549 cells or the C4 cells were given an arbitrary reading of 1.0 and relative values were calculated for the corresponding bands in lanes from the dexamethasone-treated A549 cells or the E20 cells, respectively. To correct for unequal loading in the Northern blot analyses and RPA, the data were expressed as the densitometric reading of the mRNA band relative to that of the corresponding 18S rRNA band. This corrected densitometric reading of the mRNA band in the lane from the control A549 cells was given an arbitrary value of 1.0 and a relative value for the corresponding band in the lane from the dexamethasone-treated A549 cells was determined.

For semiquantitative RT-PCR, the optimum number of cycles was determined first for GAPDH and then for each of the differentially expressed cDNAs of interest in control and dexamethasone-treated A549, H441, and BEAS-2B cells and also in A549 cells and in A549 cells transfected with control (C4 and C8) or E1A expression (E4, E11, and E20) plasmid. The densitometric readings of each of the bands representing the PCR target at a given cycle number was used to determine the number of cycles required to give a detectable autoradiographic signal that was below saturating conditions and yet within the linear range of the assay (28). To ensure that equal amounts of RNA were added to each PCR and to verify a uniform amplification process, densitometric readings of the GAPDH target, amplified as an internal control, at its optimal cycle were used. Differential expression was assessed by comparing the optical density of the band representing the cDNA of interest at its optimal cycle from the dexamethasone-treated or E1A-expressing A549 cells with that of the appropriate control cell after taking into consideration the optical density of the corresponding GAPDH band at its optimal cycle from the two cell lines being compared.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Identification of Glucocorticoid-regulated Genes by Differential Display

RNA isolated from A549 lung epithelial cells cultured in the presence or absence of dexamethasone was analyzed by the differential display assay. In this set of experiments, 3 anchoring primers (H-T₁₁G, H-T₁₁A, and H-T₁₁C) and 8 upstream primers (H-AP1-8), for a total of 24 primer pair combinations, were used. The reproducibility of each of the differentially expressed mRNA bands was confirmed by performing the assay in triplicate using RNA isolated from cells in three separate experiments. Representative autoradiograms of differentially displayed bands along with the results of a quantitation of this differential expression are shown in Figure 1. Fifteen differentially expressed cDNA bands were identified in this manner and cloned. Because the cDNAs isolated from differential display gels frequently contain a mixture of PCR products (29), the homogeneity of the recombinant clones was determined by Southern blot analysis of DNA isolated from 10 recombinant colonies from each ligation reaction. Approximately half (8 of 15) of the cDNA bands contained cDNAs that were of a single species. The remaining cDNA clones (7 of 15) were found to contain a mixture of inserts from which the predominant cDNA was systematically identified by Southern blot analysis. The 15 isolated cDNAs are listed in Table 1 along with the manner in which glucocorticoids regulate their expression in A549 cells.

View larger version (65K): [in this window] [in a new window]   Figure 1.   Differential display of mRNA isolated from control and glucocorticoid-treated A549 cells. Total RNA from control (C) and dexamethasone-treated (1 µM) (D) A549 cells was compared by differential display. Representative autoradiograms are shown from one of the three repeated differential display assays, all three of which gave similar results, and in which differential expression using three different one-base anchored oligo(dT) primers (H-T11G, H-T11A, and H-T11C) and four arbitrary 13-mers (H-AP1, H-AP3, H-AP6, and H-AP8) was found between the untreated and the dexamethasone-treated A549 cells. Four differentially expressed cDNA bands labeled GC2, GC7, GC10, and GC14 and their size in base pairs are indicated (arrowheads). The graphs below each autoradiogram show the average results ± SD of the quantitation of this differential expression from the three assays, in which the optical density of the band of interest in lane D is expressed relative to that of the corresponding band in lane C, which was set at an arbitrary value of 1.

Sequencing of Cloned cDNA Inserts

The glucocorticoid-regulated cDNAs identified by the differential display assay were sequenced using plasmid vector primers. The DNA sequences were used to search DNA sequence databases at the nucleotide level using the BLAST search program (23). The percentage sequence identity between the cDNA sequence and the match found in the database was based on the length of the shorter of the two. Nucleotide homology between the glucocorticoid-regulated cDNAs and DNA sequences in the databases above 90% was considered a match. Table summarizes the results obtained for the 15 glucocorticoid-regulated cDNAs identified by the differential display assay. Two cDNAs were matched with known genes. GC4 matched cardiac ventricular myosin light chain 2 and GC13 matched the RNA polymerase subunit hRPB. Nine of the clones matched expressed sequence tag (EST) sequences, which represent sequences from RNA isolated from a number of different tissues. Two of those matching EST sequences, GC2 and GC12, were found to contain identical inserts. GC2 and GC12 are PCR products that resulted from amplifications using identical upstream primers and downstream primers that differed by one nucleotide (H-T₁₁G versus H-T₁₁C). Four of the clones did not match any sequences in the current DNA sequence databases.

Verification of mRNA Expression of Glucocorticoid-regulated cDNA Clones by Northern Blot Analysis, RPA, and Semiquantitative RT-PCR

Differential mRNA expression of the 14 unique glucocorticoid-regulated cDNA clones was then verified using three different mRNA expression assays. Northern blot analysis of RNA isolated from control and glucocorticoid-treated A549 cells showed differential mRNA expression in only 1 of the 14 cDNAs (GC7) cloned (Table , Figure 2a). The more sensitive RPA revealed differential mRNA expression in 3 of 14 of the cloned cDNAs (GC2, 7, and 8) (Table , Figure 2b), including the cDNA clone that was positive by Northern blot analysis.

Amplification of the GAPDH cDNA from control and dexamethasone-treated A549 (Figure 3a), H441, and BEAS-2B cells, and from A549 cells, A549 cells transfected with control plasmid (C4 and C8), or E1A-transfected A549 cells (E4, E11, and E20) was tested from 21 to 39 PCR cycles, in 3 cycle increments. Densitometric analysis determined that 24 cycles was below saturating conditions and yet in the linear range of the assay in all cases. Hereafter, the autoradiographic signal for GAPDH at 24 cycles of amplification was used as an internal control in each semiquantitative RT-PCR. The expression of GAPDH did not appear to be altered by dexamethasone treatment (Figure 3a).

View larger version (37K): [in this window] [in a new window]   View larger version (31K): [in this window] [in a new window]   View larger version (31K): [in this window] [in a new window]   Figure 3.   Verification of differential mRNA expression by semiquantitative RT-PCR. (a) RT-PCR amplification of GAPDH mRNA from control (C) and dexamethasone (D)-treated cells was from 21 to 39 cycles in 3-cycle increments. The representative autoradiograms shown here from A549 cells are from one of three separate analyses on A549, H441, and BEAS-2B cells. The average densitometric readings in absorbance units (AU) ± SD of the autoradiographic signals from the three analyses are plotted against cycle number and shown above the autoradiograms. (b) Semiquantitative RT-PCR amplification of mRNAs representing the differentially displayed cDNAs of interest from control (C) and dexamethasone (D)-treated A549 cells was performed in 21 or 24 cycles in increments of 3 cycles. Representative autoradiograms from the analysis of GC10 and GC13 are shown, with the results of the cDNA of interest shown in the top panel and that of GAPDH after 24 cycles of amplification in the bottom panel. The autoradiograms represent results from one of two separate analyses for each cDNA. For the differentially displayed cDNA, the average densitometric readings in absorbance units from the two trials were plotted against cycle number and shown above the autoradiogram. For GAPDH, the densitometric reading of the band in lane C was assigned an arbitrary value of 1 and the average relative value of the corresponding band in lane D was determined and shown above the autoradiogram. (c) Semiquantitative RT-PCR amplification of GC10 mRNA from control (C) and dexamethasone (D)-treated H441 and BEAS-2B cells was done from 21 to 30 cycles in 3-cycle increments. Representative autoradiograms from one of two separate analyses are shown, with the results of the GC10 mRNA shown in the top panel and that of GAPDH after 24 cycles of amplification in the bottom panel. The average densitometric readings in absorbance units of the GC10 bands from the two trials were plotted against cycle number and shown above the autoradiogram. For GAPDH, the densitometric reading of the band in lane C was assigned an arbitrary value of 1 and the average relative value of the corresponding band in lane D was determined and shown above the autoradiogram.

All 11 differentially displayed cDNA clones that were undetectable by either Northern blot analysis or RPA were subjected to semiquantitative RT-PCR analysis. The autoradiographic results of this analysis consistently showed a single band of the expected target size. In all cases, the optical density of the GAPDH band at 24 cycles from the control cells was similar to that of the dexamethasone-treated cells. Examples of these findings are illustrated in Figures 3a and 3b. Glucocorticoid regulation of mRNA expression in opposite directions was confirmed for GC10 and GC13 (Table , Figure 3b): after 24 cycles of amplification enhanced expression of GC10 was found in dexamethasone-treated A549 cells whereas reduced expression of GC13 was observed. At the same time the expression of GAPDH was not altered. Similarly, differential regulation of GC1, GC4, GC6, GC14, and GC15 by dexamethasone was verified (Table ). Within the number of cycles tested, GC3, GC9, and GC11 did not show differential expression (Table ). These results were consistent in both series of PCRs tested. GC5, which did not match any sequence in the DNA sequence databases, was undetectable by RT-PCR even after 40 cycles and was considered to be an artifact generated by the differential display assay or subsequent steps such as cloning. Semiquantitative RT-PCR verified differential mRNA expression in 7 of 11 cDNAs that were tested. In summary, 10 of 14 cDNA clones that were identified by differential display were confirmed to have true differential mRNA expression.

Verification of Glucocorticoid Induction of the mRNA Expression of GC2 and GC10 in Other Airway Epithelial Cells

Glucocorticoid induction of the mRNA expression of GC2 and GC10 was tested twice by semiquantitative RT-PCR in control and dexamethasone-treated H441 and BEAS-2B cells. The results confirmed the glucocorticoid induction of GC10 (Figure 3c) and GC2 (data not shown) in both H441 and BEAS-2B cells.

Identification of Adenovirus E1A-regulated Genes by the Differential Display Assay

The strategy of sequencing the cDNA inserts identified by the differential display assay was then used for adenovirus E1A-regulated genes in A549 cells. RNA was isolated from either C4, an A549 cell line transfected with the control plasmid, or E20, an A549 cell line transfected with the E1A gene. Initially, the primer pairs used to identify the glucocorticoid-regulated genes (three different anchored primers and eight upstream primers H-AP1-8) were used. As well, eight additional upstream primers (H-AP9-16) were used for a total of 48 different primer combinations. Nineteen cDNAs (Table 2) were identified using the upstream primers H-AP1 to H-AP8. Three cDNAs (Table 3) were identified using upstream primers H-AP9 to H-AP16. DNA sequence analysis of 19 of these cDNAs listed in Tables and revealed matches with known genes and ESTs. Three cDNA clones did not match sequences in the current DNA sequence databases.

cDNA Clones that Are Regulated by Both Adenovirus E1A and Glucocorticoids

DNA sequence analysis revealed that three cDNA clones were common to both the E1A-regulated and glucocorticoid-regulated groups identified by differential display. Two of these, GC2 and GC10, of the verified glucocorticoid-regulated cDNAs had nucleotide sequences identical to those of EA12 and EA19, respectively, of the E1A-regulated genes. Hereafter, these two common cDNAs will be referred to as GC2 and GC10. The third, GC11, was found to be identical with EA2; however, the former was not confirmed as a glucocorticoid-regulated cDNA (described previously). Differential display analysis comparing C4 cells to E20 cells showed that E1A inhibited the mRNA expression of clones GC2 and GC10 (Table ), whereas glucocorticoids increased the expression of these genes in A549 cells (Table ). Semiquantitative RT-PCR of RNA extracted from A549, C4, and C8 cells compared with those from E4, E20, and E11 cells verified the GC10 inhibition by E1A (Figure 4). This assay also verified E1A inhibition of GC2 when C4 cells were compared to E20 cells; however, inhibition by E1A was not observed consistently when the other four cell lines were considered (data not shown). The results of the semiquantitative RT-PCR were consistent in both trials.

View larger version (40K): [in this window] [in a new window]   Figure 4.   Semiquantitative RT-PCR amplification of GC10 mRNA from E1A cells (A549, C4, and C8) or E1A+ cells (E4, E11, and E20) was from 21 to 30 cycles in 3-cycle increments. The autoradiograms represent results from one of two separate analyses on each of the cell lines, with the results of the GC10 mRNA in the bottom left panel and that of corresponding GAPDH band after 24 cycles of amplification immediately to the right of these. Densitometric readings in absorbance units (AU) of the GC10 bands from the E1A cells and from the E1A+ cells from the two analyses were pooled and the average AU ± SD (n = 6) plotted against cycle number are shown above the autoradiograms. For GAPDH, the densitometric reading of the band in the lane representing the A549 cells was assigned an arbitrary value of 1 and the average relative value of the corresponding band in the lane representing the control and E1A transfectants from the two analyses was determined and shown to the right of the autoradiogram.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

This article is based on a strategy of cloning and sequencing cDNAs identified by the differential display assay to isolate genes from lung epithelial cells that are regulated by adenovirus E1A and/or glucocorticoids. It takes advantage of the sensitivity of the PCR in amplifying mRNAs that are expressed at low levels. Genes that are expressed at low levels in cells that have been isolated using the differential display assay include important signaling molecules such as transcription factors (30, 31). To verify that the glucocorticoid-regulated cDNAs identified by the differential display assay are truly differentially expressed and not an artifact of the assay, we first examined the mRNA expression of these cDNAs in A549 cells using three different mRNA expression assays. Predictably, the mRNA expression detected was related to the sensitivity of the mRNA assay. Northern blot analysis, a relatively insensitive technique, confirmed differential mRNA expression in only 1 of 14 (GC7) of the glucocorticoid-regulated cDNAs. The RPA, which is approximately 10 times more sensitive than Northern blot analysis, confirmed differential expression in 3 of 14 cDNAs (GC2, GC7, and GC8). Semiquantitative RT-PCR, the most sensitive mRNA expression assay available, demonstrated differential expression in 7 of 11 cDNAs examined (GC1, GC4, GC6, GC10, GC13, GC14, and GC15). Four cDNAs (GC3, GC5, GC9, and GC11) were shown not to be differentially expressed by this method. In fact, the expression of one of these, GC5, could not be duplicated, even after 40 cycles of amplification. Besides the verification of differential expression of 10 cDNAs in A549 cells, comparable expression in 2 other pulmonary epithelial cells lines, H441 and BEAS-2B, was demonstrated by semiquantitative RT-PCR for GC2 and GC10. Therefore, we are confident that 10 of 14 cDNAs identified by the differential display assay were differentially expressed in A549 cells in response to treatment by glucocorticoids.

Our results indicate that there are a number of indirect measures of PCR primer specificity. Of the 15 glucocorticoid-regulated cDNAs identified by differential display, 2 cDNAs were found to be the same (GC2 and GC12). These cDNAs were amplified with identical upstream primers but with downstream primers differing by one nucleotide. Our findings confirm the results of others (19), that anchored oligo(dT) downstream primers have good specificity and reproducibility in the differential display assay. Of the E1A-regulated cDNAs, compared with the 19 identified by differential display using the first set of upstream primers (H-AP1 to H-AP8), only 3 additional cDNAs were found using a second set of upstream primers (H-AP9 to H-AP16). The reason why the second set of upstream primers was much less effective in this PCR-based assay is unknown. The upstream primers were originally designed as random nucleotide sequences that will bind 100 to 500 nucleotides upstream from the mRNA poly(A) tail (14). None of the E1A-regulated cDNAs identified by the second set of upstream primers matched with glucocorticoid- or E1A-regulated cDNAs that were isolated using the first set of upstream primers. This indicates that the upstream primers were specific and were not simply amplifying genes randomly. Finally, most of the 14 unique cDNAs identified by the differential display analysis of glucocorticoid-regulated genes are expressed at low levels in lung epithelial cells, as subsequent Northern analysis was able to detect only one of these. It was surprising that genes that are highly expressed in the cell are not overrepresented in the set of genes identified by the differential display assay. This suggests that primer specificity is high and that the electrophoretic separation of the amplified fragments is sufficient to permit the isolation of a high percentage of rare cDNAs.

Identification of the cDNAs detected by the differential display assay takes advantage of the large number of EST sequences that have been generated as part of the human genome project (32). Similar to the differential display assay, most EST sequencing strategies involve the use of anchored oligo(dT) primers that amplify the 3'-untranslated region. The EST sequences in the databases are partial nucleotide sequences from an extensive number of mRNAs expressed in a large number of tissues at varying stages of development or differentiation (33). This information has increased the knowledge concerning tissue-specific expression of genes (33) and has led to the identification of at least one disease-associated gene (34). A most valuable aspect for us is that it greatly facilitates the interpretation of sequence data generated from the cDNAs identified by the differential display assay. Conversely, while the function of many of the genes identified as ESTs is not known, the sequence information of the cDNAs selected by the differential display assay may permit the identification of EST genes that are involved in pathological processes such as inflammation.

The known genes classified as glucocorticoid-regulated and adenovirus E1A-regulated by this study can be grouped into a number of different categories. These categories include transcription/translation genes (GC13, EA1, EA13, EA17, and EA18), cytoskeletal/contractile element genes (GC4), genes encoding metabolic enzymes (EA5), and cell regulation/signal transduction genes (EA6, EA11). Additional sequence information regarding the remaining genes will be needed to classify them further. As demonstrated by the Northern blot, RPA, and semiquantitative RT-PCR mRNA expression assays, the majority of the genes we detected by the differential display assay are expressed at low levels in A549 cells. RNA expression of most of these genes was detectable only with PCR-based assays. The primer pair combinations used in this assay were estimated to amplify about 10% of the mRNAs expressed in the cell (19). It is anticipated that additional primer pair combinations will identify other genes. In addition, the higher number of PCR cycles used in our differential display assays, ranging from 30 to 40 cycles, likely favored the identification of rare cDNAs. Owing to the exponential nature of PCR, abundant cDNAs would amplify a saturating amount of PCR product at these high cycle numbers and, therefore, would not be detected as differentially expressed genes.

Adenovirus E1A has been implicated in the pathogenesis of pulmonary inflammatory diseases and glucocorticoids have been shown to be effective antiinflammatory agents in these diseases. We reasoned that genes that were regulated by both glucocorticoids and adenovirus E1A could be used to select genes for further characterization. A mechanism to link the opposing effects of glucocorticoids and adenovirus E1A comes from studies of the transcription coactivator, CREB-binding protein (CBP)/p300. Adenovirus E1A, glucocorticoids, and other intracellular signaling molecules have been shown to bind to CBP/p300 (35). The competition for limited amounts of CBP/p300 in the cell by glucocorticoids and adenovirus E1A may explain the opposing effects of these antiinflammatory agents and E1A on gene expression. Our preliminary results have permitted the identification of 10 confirmed glucocorticoid-regulated and 1 confirmed and 21 putative adenovirus E1A-regulated genes. Interestingly, two cDNAs, GC2 and GC10, were common to both groups and their mRNA expression in A549 cells was initially found to be increased by glucocorticoids and inhibited by E1A. When additional lung epithelial cells were tested for mRNA expression, while both cDNAs demonstrated glucocorticoid induction, only GC10 was inhibited by E1A. The opposing effect of E1A and glucocorticoids on the expression of GC10 in several pulmonary epithelial cell lines supports the concept that this approach should be helpful in the identification of genes involved in the pathogenesis of lung inflammation that are also targets for antiinflammatory drugs such as glucocorticoids.