Type-Specific Induction of Interleukin-8 by Adenovirus - ATS Journals

Type-Specific Induction of Interleukin-8 by Adenovirus J. Leland Booth and Jordan P. Metcalf Pulmonary and Critical Care Division, Department of Internal Medicine, University of Oklahoma Health Sciences Center and the Program in Molecular and Cellular Biology, Oklahoma Medical Research Foundation, Oklahoma City, Oklahoma

Infection with adenovirus (Ad) causes acute pneumonia in a type-specific fashion because type 7 but not type 5 Ad has been isolated as a causative agent. We postulated that the type specificity of induction of pneumonia may be related to type-specific cytokine induction in lung cells. To test this hypothesis, we infected human fetal lung fibroblasts and the lung epithelial cell line A549 with live type 5 and type 7 Ad. Virus inactivated by irradiation was used as a control. Type 7 but not type 5 Ad induced interleukin (IL)-8 protein production in both cell types in a dose- and time-dependent manner. Inactivated virus had no effect on the production of IL-8 protein. Type 7 but not type 5 virus also stimulated IL-8–specific messenger RNA (mRNA) production in these cells. Because half-life of IL-8 mRNA was prolonged in both type 5– and type 7–infected A549 cells, induction likely involves enhancement of message stability as well as other effects. Virus early gene expression did not consistently correlate with IL-8 message induction and followed induction in fibroblasts. These results suggest that there is type-specific induction of IL-8 production during infection of lung cells with Ad. Induction involves message stabilization and may not require viral gene expression. Because IL-8 is one of the important mediators of lung inflammation, typespecific induction of this and other cytokines may account for the different consequences of lung infection with different types of Ad. Booth, J. L., and J. P. Metcalf. 1999. Type-specific induction of interleukin-8 by adenovirus. Am. J. Respir. Cell Mol. Biol. 21:521–527.

Although adenovirus (Ad) exposure can cause a variety of localized infections as well as disseminated disease in immunocompromised and in nonimmunocompromised hosts, there appear to be type-specific consequences of infection. For example, type 7 Ad has been isolated during episodes of pneumonia in military recruits and children (1, 2), whereas type 5 is primarily implicated as a cause of upper respiratory tract infection (3). On the basis of these observations, we postulate that the type-specific nature of the consequences of lung infection might be due to type-specific differences in the elaboration of cytokines in response to infection of lung cells with these viruses. It is likely that alveolar epithelial cells are important in the inflammatory response because Ad has been shown to infect alveolar epithelial cells during acute infection (4). Pulmonary fibroblasts occupy a key position in the lung and may also participate in lung in(Received in original form January 29, 1999 and in revised form May 7, 1999) Address correspondence to: Jordan P. Metcalf, M.D., 118 Massman, Oklahoma Medical Research Foundation, 825 NE 13th St., Oklahoma City, OK 73104. E-mail: [email protected] Abbreviations: adenovirus, Ad; enzyme-linked immunosorbent assay, ELISA; fetal bovine serum, FBS; glyceraldehyde-3-phosphate dehydrogenase, GAPDH; interleukin, IL; lactate dehydrogenase, LDH; lipopolysaccharide, LPS; mitogen-activated protein kinase, MAPK; multiplicity of infection, MOI; messenger RNA, mRNA; phosphate-buffered saline, PBS; standard error of the mean, SEM. Am. J. Respir. Cell Mol. Biol. Vol. 21 pp. 521–527, 1999 Internet address: www.atsjournals.org

flammation via release of interleukin (IL)-8 (5). IL-8 is an important mediator of the response to many stimuli, including viruses (6). IL-8 may also be important in lung inflammation, because IL-8 levels are increased in patients with asthma and obstructive lung disease (7, 8). To investigate the possibility that Ad stimulates IL-8 in a type-specific manner, we infected A549 cells (a human lung alveolar epithelial cell line) and primary human fetal lung fibroblasts (GM5387 cells) with type 5 and type 7 Ad. Exposure of cells to Ad inactivated by radiation was used to determine whether viral gene expression is necessary for IL-8 gene stimulation. Type 7 but not type 5 Ad induced IL-8 protein in a dose- and time-dependent manner. Inactivated virus of either type did not stimulate IL-8 gene expression, suggesting that viral gene expression is necessary for IL-8 gene activation by Ad to occur. Induction of IL-8 by Ad is likely regulated at the transcriptional level and at the level of message stability because type 7 but not type 5 Ad increased endogenous IL-8–specific messenger RNA (mRNA) whereas both serotypes enhanced IL-8 mRNA stability. The data suggest that type-specific Admediated inflammation in the lung may be due to typespecific induction of cytokine genes.

Materials and Methods Cell Culture A549 cells were purchased from the American Type Culture Collection (ATCC CCL-185) (Rockville, MD). This line

522

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 21 1999

was cultured from lung carcinoma tissue and is considered to be of alveolar type II origin (9). Cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bovine serum (FBS), 2 mM L-glutamine, and 80 mg/ml gentamicin. Normal human fetal lung fibroblast GM5387 cells were obtained from the NIGMS Mutant Cell Repository (Camden, NJ) and grown in Eagle’s minimal essential media with Earles’ Salts, 20% FBS, 2 mM L-glutamine, and 80 mg/ml gentamicin. Virus Preparation Ad type 5 and type 7 (Ad5 and Ad7, respectively) were purchased from ATCC (VR-5, VR-7) and propagated on A549 cells grown in DMEM supplemented with 10% FBS. Virus was harvested using freeze–thaw lysis, purified using CsCl gradients, and dialyzed extensively against phosphate-buffered saline (PBS)/10% glycerol. Lipopolysaccharide (LPS) levels by limulus amebocyte lysate assay (BioWhittaker, Walkersville, MD) in final virus preparations were similar to those seen in cell culture media and virus-free buffer used as a control and were , 1 EU/ml. Purified virus was titered using a plaque assay on A549 cells in soft agar (10). Viral Infection A549 or GM5387 cultures were exposed to various multiplicities of infection (MOIs) of live Ad5 or Ad7. Evidence of infection was confirmed both by cytopathic effect and by fluorescent staining for an Ad-specific antigen (Microscan; Baxter, West Sacramento, CA). Additional cultures were exposed to Ad that had been inactivated (plaqueforming units [PFU] 5 0) by g irradiation (4 3 106 rads). Cells exposed to an equal volume of virus-free buffer and cultured under identical conditions served as a negative control and tumor necrosis factor (TNF) (1 U/ml) was used as a positive control. Cytotoxicity Assay Cytotoxicity was assessed by determining the percentage of cellular lactate dehydrogenase (LDH) released into the culture medium. Briefly, cell supernatants were collected and adherent cells lysed with an equal volume of PBS/1% Triton X-100. To 2.5 ml of these samples was added 25 ml 0.75 mM pyruvate/1.3 mM nicotinamide adenine dinucleotide, reduced (Sigma, St. Louis, MO), and after incubation for 30 min at 308C, 25 ml of color reagent (Sigma) was added. After further incubation for 20 min at room temperature, the reaction was terminated by 250 ml 0.4 N NaOH and absorbance was determined at 450 nm and plotted against absorbance using known amounts of pyruvate substrate. Cytotoxicity is expressed as the percent of LDH release and equals LDH supernatant activity/(supernatant 1 cell-extract LDH activity) 3 100. IL-8 Protein Determination At varying time periods after exposure to live or inactivated virus, cell media were collected and stored at 2708C before IL-8 protein determination. The adherent cells were then washed twice with PBS, lysed and collected in 0.1 N NaOH, and stored at 2708C before determination of total cellular protein using the Bradford method (11). IL-8

protein was determined using a commercially available IL-8 protein enzyme-linked immunosorbent assay (ELISA) (R&D Systems, Minneapolis, MN). Isolation of RNA Cells were infected as previously described before harvest. All solutions used for RNA preparation were treated with diethylpyrocarbonate at 1:1,000. At various time periods after infection, cells were scraped off the plates, pelleted by centrifugation, and washed with PBS. Whole cellular RNA was prepared by the guanidinium–isothiocyanate– phenol–chloroform method of Chomczynski and Sacchi (12). The RNA was measured by ultraviolet absorbance at 260 nm and stored at 2708C before Northern blot analysis. Northern Blot Analysis RNA was dissolved in a 3-(N-Morpholino)propanesulfonic acid (MOPS)/formaldehyde/formamide solution and heated for 5 min at 608C. Samples (7.5 mg/lane) were run on an agarose/formaldehyde (1.5% agarose/2.2 M formaldehyde/MOPS) gel. The gel was stained with ethidium bromide. After overnight capillary transfer to a Gene Screen (Dupont/NEN, Boston, MA) membrane, the RNA was probed with a 32P end-labeled 30–base pair oligomer complementary to the IL-8 protein coding region (13). Fresh buffer was added to the blot along with the amount of probe to give 1 3 106 counts/ml. Salmon sperm DNA was added to block nonspecific binding. After incubation, the blot was removed, washed, and exposed. The process was repeated using a 32P random primer–labeled complementary DNA (cDNA) probe for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (14) and a similarly labeled cDNA probe for Ad5 and Ad7 E1A. To estimate the stability of IL-8 mRNA transcripts, A549 cells were exposed to actinomycin D (10 mg/ml; Sigma) 3 h after exposure to Ad5 or Ad7. In two separate experiments, total cellular RNA was extracted before and 30, 60, 120, 240, and 480 min after actinomycin D addition for IL-8 mRNA determination and quantitation. Viral Infectivity Assay Ad infection was quantified for both viruses in both cell types by a modification of the method described by Wickham and colleagues (15). Cells were exposed to virus at 50 MOI for 0, 5, 10, 20, 30, and 60 min. They were then removed from the plates using 0.25% trypsin/0.1% ethylenediaminetetraacetic acid/50 mM azide to remove externally bound virus and inhibit further internalization. After further washing, the cells were transferred to slides, fixed and permeabilized with acetone, and stained for Ad using a fluorescein isothiocyanate–conjugated antihexon-specific antibody (Chemicon, Temecula, CA). Viral Replication Assay Replication of both viruses in the A549 and GM5387 fibroblasts was determined by exposure of both cell types to viruses in a manner identical to that used for IL-8 induction. At 3, 6, 12, and 24 h after exposure, cells were harvested in 1 ml of medium. The amount of infectious particles present in the cell extract was determined by plaque assay and expressed as PFU per milliliter of extract.

Booth and Metcalf: Interleukin-8 and Adenovirus

Statistical Analysis Results are reported as means 6 standard error of the mean (SEM). Statistical differences between groups were determined using multiple comparisons among treatment means using the Bonferroni t test. The effect of increasing doses of different treatments was determined using the repeated-measures analysis of variance (16). IL-8 mRNA halflife was determined by linear regression analysis (16).

Results Type-Specific Ad Induction of IL-8 Protein Production in A549 Cells A549 cells were exposed to either type 5 or type 7 Ad (MOI 5 50) for various time periods. This dose results in infection of a majority of both cell types within 1 h of exposure (see VIRAL INFECTIVITY section, later). Ad7 induced IL-8 production in A549 cells beginning as soon as 3 h after infection and continuing for the 24 h of the experiment (Figure 1A; P , 0.05, all comparisons). Infection of the same cells with Ad5 (MOI 5 50) did not increase IL-8 production. Virus of either type 5 or type 7 inactivated by g irradiation also failed to increase IL-8 production (Figure 1A). Heat-inactivated virus (568C, 4 h) of either type also failed to induce IL-8 (not shown). LDH release in Ad5-

Figure 1. Type-specific induction of IL-8 protein. (A) A549 cells were exposed to live or irradiated inactivated Ad5 or Ad7 for various times. Cells exposed to buffer served as a negative control and TNF (1 U/ml) was used as a positive control. IL-8 was determined at times indicated using ELISA and normalized for total cellular protein as described in M ATERIALS AND METHODS. All measurements were performed in triplicate. (B) GM5387 cells exposed to the same virus preparations as in A. Data are expressed as means 6 SEM of three separate experiments in both panels.

523

and Ad7-infected cells at 24 h was similar (24 and 26%) though higher than that in control cells (17%). Type-Specific Ad Induction of IL-8 Protein Production in GM-5387 Fibroblast Cells GM5387 cells were exposed to either Ad5 or Ad7 (MOI 5 50) for various time periods. Ad7 induced IL-8 production in these cells by 3 h after infection and was elevated at subsequent time points (Figure 1B; P , 0.05, all comparisons). It is notable that although irradiated Ad7 failed to induce IL-8 at any time point, there was modest induction of IL-8 over control cell levels by irradiated Ad5 at 6 h after exposure (6.9 versus 1.8 ng/mg cell protein, P , 0.05). Heat-inactivated virus (568C, 4 h) of either type failed to induce IL-8 (not shown). LDH release in control and Ad5and Ad7-infected cells was similar at 24 h (29, 28, and 30%, respectively). Increasing MOI of Ad7 but Not Ad5 Increases IL-8 Production in A549 Cells A549 cells were infected with increasing doses of human type 5 and type 7 Ad from 0.001 to 50 MOI. The cells were cultured in the presence of virus for 24 h before collection of the supernatants for measurement of IL-8 protein. Unexposed cells cultured under identical conditions served as

Figure 2. Effect of increasing amounts of Ad5 or Ad7 on induction of IL-8 protein. (A) A549 cells were exposed to increasing MOI of Ad5 or Ad7 for 24 h. Buffer-exposed cells cultured under identical conditions served as a negative control. IL-8 was determined at times indicated using ELISA and normalized for total cellular protein as described in M ATERIALS AND METHODS. All measurements were performed in triplicate. (B) GM5387 cells exposed to increasing MOI of Ad5 and Ad7 as in A. Data are expressed as means 6 SEM of three separate experiments for both panels.

524


a negative control. There was a type-specific difference in the responses of these cells to increasing doses of virus. The results demonstrated a significant increase in IL-8 with increasing MOI of Ad7 (P , 0.01) that was not seen with exposure of the cells to increasing doses of Ad5 (Figure 2A). Increasing MOI of Ad7 but Not Ad5 Increases IL-8 Production in GM-5387 Fibroblast Cells GM5387 cells were infected with increasing doses of human type 5 and type 7 Ad from 0.001 to 50 MOI. Cells were cultured in the presence of virus for 24 h before collection of the supernatants for measurement of IL-8 protein. Unexposed cells cultured under identical conditions served as a negative control. As in the A549 cells, there was a type-specific difference in the response of these cells to increasing doses of virus. The results again showed a significant increase in IL-8 with increasing MOI of Ad7 (P , 0.01) that was not seen with exposure of the cells to increasing doses of Ad5 (Figure 2B). Type-Specific Induction of Endogenous IL-8 RNA Production in A549 Cells A549 cells were exposed to Ad5 or Ad7. At various times after infection, whole-cell RNA was extracted and IL-8– specific RNA was quantified by Northern analysis using an IL-8 probe. Ethidium bromide staining of the RNA was equivalent (Figure 3A). The blots were probed with a GAPDH probe to assess levels of ongoing synthesis of cellular RNA (Figure 3B). When these levels are constant the signal can be used to confirm equal loading of RNA (14). Infection with Ad7 increased IL-8 mRNA levels by 6- and 11-fold over uninfected control cells at 3 and 6 h after in-

Figure 3. Type-specific induction of endogenous IL-8 mRNA by Ad in A549 cells. A549 cells were exposed to live Ad5 or Ad7 for various times before RNA extraction. Buffer-exposed cells cultured under identical conditions served as a negative control. (A) RNA gel stained with ethidium bromide. (B) The same blot as in A, reprobed for GAPDH. (C) The same blot as in A, reprobed for IL-8 mRNA. (D) The same blot as in A, reprobed for Ad E1A mRNA.

fection, respectively, and then fell to control levels (Figure 3C). IL-8 mRNA levels were highest in the infected cells 3 h after infection and declined at each subsequent point measured. GAPDH mRNA levels decreased in Ad7infected cells during the experiment (Figure 3B). When corrected for GAPDH mRNA levels, there was an increase in relative IL-8 mRNA of 5-, 18-, 27-, and 9-fold over control levels at 3, 6, 12, and 24 h after infection, respectively, over uninfected control cells. There was no increase in IL-8 mRNA over uninfected control cells at any time point in cells infected with Ad5 (Figure 3C). Expression of Ad7 E1A mRNA was detected at 3 h after infection and decreased thereafter (Figure 3D). Ad5 E1A mRNA was detected 12 h after infection and increased at 24 h (Figure 3D). Type-Specific Induction of Endogenous IL-8 RNA in GM5387 Cells GM5387 cells were exposed to Ad5 or Ad7. At various times after infection, whole-cell RNA was extracted and IL-8–specific mRNA was quantified by Northern analysis using an IL-8 probe. Equal loading of RNA was confirmed by ethidium bromide staining (Figure 4A). The blots were probed for GAPDH (Figure 4B) and, as in the A549 cells, there was a decrease in the mRNA levels of this housekeeping gene with increasing time after infection in the Ad7-exposed cells. IL-8 mRNA levels were increased in the Ad7-infected cells 3.5-, 2.5-, 1- (equivalent), and 2-fold over control cells at 3, 6, 12, and 24 h, respectively, after infection (Figure 4C). When corrected for GAPDH mRNA levels, there was an increase in relative IL-8 mRNA of 3-, 3.2-, 3.4-, and 3.6-fold over control levels at 3, 6, 12, and 24 h, respectively, after infection. IL-8 mRNA levels were highest at the initial time point and declined at all subse-

Figure 4. Type-specific induction of endogenous IL-8 mRNA by Ad in GM5387 cells. GM5387 cells were exposed to live Ad5 or Ad7 adenovirus for various times before RNA extraction. Buffer-exposed cells cultured under identical conditions served as a negative control. (A) RNA gel stained with ethidium bromide. (B) The same blot as in A, reprobed for GAPDH. (C) The same blot as in A, reprobed for IL-8 mRNA. (D) The same blot as in A, reprobed for Ad E1A mRNA.


quent points measured (Figure 4C). As in the A549 cells, there was no increase in IL-8 mRNA at any time point in Ad5-infected cells (Figure 4C). In contrast to the data from A549 cells, expression of Ad7 E1A mRNA was not detected until 6 h after infection and increased thereafter (Figure 4D). Ad5 E1A mRNA was first detected 12 h after infection and increased at 24 h (Figure 4D). IL-8 mRNA Stability in A549 Cells A549 cells were exposed to both viruses and IL-8 message stability in these cells was determined by inhibition of mRNA synthesis with actinomycin D. In control cells, IL-8 mRNA levels fell rapidly with a half-life of 38 min (Figure 5, R2 = 0.998), similar to results reported by others (17). IL-8 stability was prolonged to approximately 70 min in both the Ad5- and Ad7-infected cells (Figure 5, R2 = 0.939 and 0.908, respectively). Viral Infectivity in A549 and GM5387 Cells Both cell types were exposed to virus for various times and the percentages of cells infected were determined by immunostaining for Ad hexon antigen. In A549 cells, infection occurred rapidly by both virus strains with over 80% infected 20 min after exposure (Figure 6A). Essentially all cells (98%) were infected by both viruses by 60 min after exposure (Figure 6A). Infection of GM5387 cells was slower, as , 70% of cells were infected by Ad7 and , 40% were infected by Ad5 20 min after exposure (Figure 6B). At 60 min after exposure, 93 and 60% of these cells were infected by Ad5 and Ad7, respectively (Figure 6B). These results demonstrate that both viruses are infectious in each cell type and that a difference in infectivity is unlikely to account for the difference in induction of IL-8 by Ad7, at least in the A549 cells.

525

IL-8 and protein induction, cells were collected and the amounts of infectious virus produced determined by plaque assay. In both cell types, both viruses began to produce significant quantities of infectious virus after an initial delay. Interestingly, there was initially less infectious Ad5 than Ad7 virus produced in both cell types (Figure 7). By 24 h after infection, the total amounts of Ad5 and Ad7 virus present in the cell extract were similar (Figure 7). The results again demonstrate that both viruses replicate well in both cell types, although Ad5 replication is initially lower than Ad7 replication. It is possible that this difference in replication is related to type-specific induction of IL-8, but this is as yet unknown.

Discussion These studies show that Ad stimulates expression of the IL-8 gene in human lung cells in a type-specific manner. Specifically, we have shown that infection of these cells with Ad7 but not Ad5 increases IL-8 protein production, and this is associated with an increase in IL-8–specific mRNA, despite concurrent suppression of GAPDH mRNA. Because we demonstrated this effect in pulmo-

Viral Replication in A549 and GM5387 Cells A549 and GM5387 cells were exposed to Ad5 and Ad7 virus. At the same times after infection used for determining

Figure 5. Effect of Ad5 and Ad7 infection on IL-8 mRNA stability in A549 cells. Three hours after infection with Ad5 or Ad7 virus, actinonmycin D (10 mg/ml) was added to A549 cells and RNA was extracted at various times for evaluation by Northern blot analysis using an IL-8 probe (see MATERIALS AND METHODS). Cells exposed to an equal quantity of virus buffer were treated in the same manner for use as a negative control. Data are expressed as a percentage of the IL-8 signal present in RNA samples from cells collected just before the addition of actinomycin D and are representative of two individual experiments.

Figure 6. Infectivity of Ad5 and Ad7. (A) A549 cells were exposed to Ad5 and Ad7 for various times and infectivity was determined by detection of intracellular Ad antigen (see MATERIALS AND METHODS). (B) Infectivity determined in GM5387 cells using the same methods as in A. Data are expressed as a percentage of infected cells as a percentage of the total and are the means 6 SEM of three separate experiments for both panels.

526


Figure 7. Replication of Ad5 and Ad7. (A) A549 cells were exposed to Ad5 and Ad7 virus. At various times after infection cells were collected in 1 ml of media, and virus replication was determined by plaque assay (see MATERIALS AND METHODS). (B) Replication of Ad5 and Ad7 in GM5387 cells using the same methods as in A. Data are expressed as PFU per milliliter and are the means 6 SEM of three separate experiments for both panels.

nary fibroblast cells as well as the pulmonary epithelial A549 cell line, fibroblasts may contribute to the inflammatory response to Ad during acute infection via the elaboration of IL-8. Ad-induced cytotoxicity does not explain IL-8 induction because there was no type-specific increase in LDH release in either cell type. The virus was propagated in cells that likely release cytokines during propagation. Several factors decrease the possibility that cytokines or other contaminants in the viral stocks could be responsible for our results. Viruses were highly purified by gradient centrifugation and there were no cytokines detectable in the purified virus stocks. We did not see any induction in IL-8 levels when cells were exposed to irradiated virus, which suggests that the effect is due to infection of the cells by virus and not to contaminating cytokines. Finally, there was a clear difference in the IL-8 response with increasing doses of the two strains of Ad, despite similar methods of propagation and purification. Therefore, we conclude that the induction of IL-8 by Ad7 is due to the virus. Both viruses infect and replicate in both cell types (Figures 6 and 7). There are differences in the initial rates of in-

fection and replication which suggest that differences in the early steps in the infectious cycle may be related to differences in IL-8 induction. It is unlikely that delay in viral replication is entirely responsible for the lack of IL-8 induction by type 5 virus because exposure to this virus for up to 96 h failed to stimulate IL-8 production (data not shown). The results also give some insight into the mechanism of IL-8 induction in these studies. Northern analysis suggests that strain-specific induction of IL-8 by Ad is due to a difference in steady-state IL-8 mRNA levels. This appears to be partially due to increased IL-8 mRNA half-life in Ad7-infected cells. Because prolonged mRNA half-life is also demonstrated in Ad5-infected cells that do not have increased steady-state IL-8 mRNA levels, there must be an additional factor involved in Ad7-mediated IL-8 mRNA induction, likely stimulation of transcription. Our observations also complement studies by others relating to IL-8 induction by the replication-deficient Ad5 “gene therapy” vectors and are comparable inasmuch as we used similar virus doses and exposure times (18, 19). This vector, which lacks the Ad E1A region, has been shown to induce IL-8 in A549 cells (18) or in human bronchial epithelial cells at high MOI (19). In contrast, wild-type (WT) Ad5 (containing E1A) did not induce IL-8 in any of the cells tested (Figures 1 and 2 [19]). Our studies comparing the timing of Ad 7 E1A gene expression to IL-8 induction show no consistent relationship in the cell lines tested. Specifically, IL-8 mRNA induction in A549 cells parallels Ad7 E1A expression, whereas induction in fibroblasts precedes detectable Ad7 E1A mRNA. These results suggest either that E1A is not responsible for IL-8 induction, that the effect of E1A is type-specific or cell-specific, or that minimal amounts of E1A are needed for induction. If E1A does not act alone to induce IL-8, it is still possible that it could act with a cofactor to induce IL-8, and Keicho and associates demonstrated that stably transfected E1A can cooperate with LPS to induce this cytokine (20). We do not feel that LPS is important in our results for several reasons: (1) the amount of LPS in our system is an order of magnitude less than the lowest dose shown to augment IL-8 release in E1A-transfected cells (20); (2) the amount of LPS present was the same in the virus preparations, virus-free control buffer, and media; and (3) we do not see IL-8 induction in Ad5-infected cells producing the same type of E1A as those used by Keicho (20). An alternate possibility that does not involve viral DNA expression and is suggested by the data in fibroblasts is that IL-8 induction occurs due to a step in the Ad infectious pathway before viral gene expression. These findings may also explain the modest induction of IL-8 by irradiated type 5 virus (Figure 2B) because irradiation may have inactivated the virus in a manner similar to that of E1A-deficient virus, resulting in IL-8 induction. Additional data regarding the mechanism of induction of IL-8 by the E1A-deficient vectors have suggested that the mechanism is due to stimulation of mitogen-activated protein kinase (MAPK) (21). This study also demonstrated inhibition of Ad5 vector–mediated IL-8 induction and MAPK activation by forskolin. Unfortunately, because both WT- and E1A-deficient Ad5 activate MAPK whereas only E1A-deficient Ad5 induces IL-8, the role of MAPK activation in this process is unclear.


Our findings demonstrating that this induction is typespecific are potentially important in understanding the different consequences of infection with different strains of the virus. Some Ad types cause primarily upper respiratory tract or gastrointestinal symptomatology, whereas other types, including type 7, can cause pneumonia (1, 2). Acute Ad pneumonia is accompanied by tremendous inflammation with alveolitis, bronchiolitis, and alveolar edema with primarily mononuclear cell and lymphocytic cell infiltration. Polymorphonuclear cells are also present in limited numbers (4). The induction of IL-8 by Ad infection of lung cells may be important in this process inasmuch as it is chemotactic for neutrophils and possibly for T lymphocytes under certain conditions (22–24). In addition, IL-8 stimulates many proinflammatory actions of neutrophils, including formation of 5-hydroxyeicosatetraenoic acid and platelet-activating factor (25, 26). Our data showing induction of IL-8 by Ad7 but not Ad5 in lung cells supports the hypothesis that the ability of particular strains of Ad to cause serious infection in certain organ systems may be related to the ability of that strain to induce cytokines in cells of that system. We found induction of IL-8 in both a pulmonary fibroblast as well as a pulmonary epithelial cell line. Because alveolar inflammation, cellular infection, and alveolar cell destruction are prominent in Ad pneumonia, the results in the alveolar epithelial-like A549 cells probably reflect the cellular response to acute infection. Because interstitial inflammation is a frequent finding in human Ad pneumonia, it is also likely that induction of cytokine release by infected interstitial cells participates in the inflammatory response in vivo. There is some evidence that chronic infection with Ad may be a risk factor for the development of chronic obstructive pulmonary disease (27). Persistence of the virus has also been detected in high frequency in children with steroid-resistant asthma (28). IL-8 is a likely mediator of the inflammation seen in these patients because levels of this cytokine are elevated in lavage fluid and are expressed in the epithelial cells of patients with these conditions (7, 8). Our results demonstrating stimulation of IL-8 release from lung cells infected with Ad suggest a possible mechanism for the pulmonary consequences of infection with this virus. The type-specific nature of this response may also explain the type specificity of the consequences of infection. Further, induction of IL-8–specific mRNA and prolongation of IL-8 half-life defines a site of stimulation by Ad at the level of message stability. The finding that IL-8 induction does not correlate with and may precede virus gene expression suggests that IL-8 induction is not related to viral DNA gene expression or is cell-specific. Acknowledgments: This study was supported by the following: a Presbyterian Health Foundation Grant, National Institutes of Health grant NIH K08 HL03106, and an American Lung Association Research Grant. The authors acknowledge the assistance of Dr. Robert Petrone with the statistical analysis and Drs. Phillip Silverman and J. Donald Capra for review of the manuscript.

References 1. Dudding, B. A., S. C. Wagner, J. A. Zeller, J. T. Gmelich, G. R. French, and F. H. Top, Jr. 1972. Fatal pneumonia associated with adenovirus pneumonia in three military trainees. N. Engl. J. Med. 286:1289–1292. 2. Sly, P. D., M. E. Soto-Quiros, L. I. Landau, I. Hudson, and H. NewtonJohn. 1984. Factors predisposing to abnormal pulmonary function after

527

adenovirus type 7 pneumonia. Arch. Dis. Child. 59:935–939. 3. Brandt, C., H. Kim, and A. Vargosdo. 1969. Infections in 18,000 infants and children in a controlled study of respiratory tract disease: I. Adenovirus pathogenicity in relation to serologic type and illness syndrome. Am. J. Epidemiol. 90:484–500. 4. Abbandanzo, S. L., C. K. English, E. Kagan, and R. A. McPherson. 1989. Fatal adenovirus pneumonia in a newborn identified by electron microscopy and in situ hybridization. Arch. Pathol. Lab. Med. 113:1349–1353. 5. Rolfe, M. W., S. L. Kunkel, T. J. Standiford, S. W. Chensue, R. M. Allen, H. L. Evanoff, S. H. Phan, and R. M. Strieter. 1991. Pulmonary fibroblast expression of interleukin-8: a model for alveolar macrophage-derived cytokine networking. Am. J. Respir. Cell Mol. Biol. 5:493–501. 6. Baggiolini, M., B. Dewal, and B. Moser. 1994. Interleukin-8 and related chemotactic cytokines—CXC and CC chemokines. Adv. Immunol. 33:97–179. 7. Nocker, R. E., D. F. Schoonbrood, E. A. van de Graaf, C. E. Hack, R. Lutter, H. M. Jansen, and T. A. Out. 1996. Interleukin-8 in airway inflammation in patients with asthma and chronic obstructive pulmonary disease. Int. Arch. Allergy Immunol. 109:183–191. 8. Marini, M., E. Vittori, J. Hollemborg, and S. Mattoli. 1992. Expression of the potent inflammatory cytokines, granulocyte-macrophage-colony-stimulating factor and interleukin-6 and interleukin-8, in bronchial epithelial cells of patients with asthma. J. Allergy Clin. Immunol. 89:1001–1009. 9. Lieber, M., B. Smith, A. Szakal, W. Nelson-Rees, and G. Todaro. 1976. A continuous tumor cell line from a human lung carcinoma with properties of type II alveolar epithelal cells. Int. J. Cancer 17:62–70. 10. Lawrence, W., and H. Ginsberg. 1967. Intracellular uncoating of type 5 adenovirus deoxyribonucleic acid. J. Virol. 1:851–867. 11. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248–254. 12. Chomczynski, P., and N. Sacchi. 1987. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162:156–159. 13. Goodman, R. B., R. G. Wood, T. R. Martin, O. Hanson-Painton, and G. T. Kinasewitz. 1992. Cytokine-stimulated human mesothelial cells produce chemotactic activity for neutrophils including NAP-1/IL-8. J. Immunol. 148: 457–465. 14. Hla, T., and T. Maciag. 1991. Cyclooxygenase gene expression is down-regulated by heparin-binding (acidic fibroblast) growth factor-1 in human endothelial cells. J. Biol. Chem. 266:24059–24063. 15. Wickham, T. J., P. Mathias, D. A. Cheresh, and G. R. Nemerow. 1993. Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 73:309–319. 16. Zar, J. H. 1996. Biostatistical Analysis. Prentice-Hall, Englewood Cliffs, NJ. 185–325. 17. Nakamura, H., K. Yoshimura, H. A. Jaffe, and R. G. Crystal. 1991. Interleukin-8 gene expression in human bronchial epithelial cells. J. Biol. Chem. 266:19611–19617. 18. Amin, R., R. Wilmott, Y. Schwarz, B. Trapnell, and J. Stark. 1995. Replication-deficient adenovirus induces expression of interleukin-8 by airway epithelial cells in vitro. Hum. Gene Ther. 6:145–153. 19. Noah, T. L., I. A. Wortman, P. C. Hu, M. W. Leigh, and R. C. Boucher. 1996. Cytokine production by cultured human bronchial epithelial cells infected with a replication-deficient adenoviral gene transfer vector or wildtype adenovirus type 5. Am. J. Respir. Cell Mol. Biol. 14:417–424. 20. Keicho, N., W. M. Elliott, J. C. Hogg, and S. Hayashi. 1997. Adenovirus E1A upregulates interleukin-8 expression induced by endotoxin in pulmonary epithelial cells. Am. J. Physiol. 272:L1046–L1052. 21. Bruder, J. T., and I. Kovesdi. 1997. Adenovirus infection stimulates the Raf/ MAPK signaling pathway and induces interleukin-8 expression. J. Virol. 71:398–404. 22. Leonard, E. J., A. Skeel, T. Yoshimura, K. Noer, S. Kutvirt, and D. Van Epps. 1990. Leukocyte specificity and binding of human neutrophil attractant/activation protein-1. J. Immunol. 144:1323–1330. 23. Bacon, K. B., and R. D. Camp. 1990. Interleukin (IL)-8-induced in vitro human lymphocyte migration is inhibited by cholera and pertussis toxins and inhibitors of protein kinase C. Biochem. Biophys. Res. Commun. 169:1099–1104. 24. Wilkinson, P. C., and I. Newman. 1992. Identification of IL-8 as a locomotor attractant for activated human lymphocytes in mononuclear cell cultures with anti-CD3 or purified protein derivative of Mycobacterium tuberculosis. J. Immunol. 149:2689–2694. 25. Schroder, J. M. 1989. The monocyte-derived neutrophil activating peptide (NAP/interleukin 8) stimulates human neutrophil arachidonate-5-lipoxygenase, but not the release of cellular arachidonate. J. Exp. Med. 170:847–863. 26. Bussolino, F., M. Sironi, E. Bocchietto, and A. Mantovani. 1992. Synthesis of platelet-activating factor by polymorphonuclear neutrophils stimulated with interleukin-8. J. Biol. Chem. 267:14598–14603. 27. Matsuse, T., S. Hayashi, K. Kuwano, H. Keunecke, W. A. Jefferies, and J. C. Hogg. 1992. Latent adenoviral infection in the pathogenesis of chronic airways obstruction. Am. Rev. Respir. Dis. 148:177–184. 28. Macek, V., J. Sorli, S. Kopriva, and J. Marin. 1994. Persistent adenoviral infection and chronic airway obstruction in children. Am. J. Respir. Crit. Care Med. 150:7–10.