Acute respiratory infection with mouse adenovirus type 1

Abstract

Studies of the pathogenesis of adenovirus respiratory disease are limited by the strict species-specificity of the adenoviruses. Following intranasal inoculation of adult C57BL/6 mice with mouse adenovirus type 1 (MAV-1), we detected MAV-1 early region 3 (E3) and hexon gene expression in the lungs at 7 days post-infection (dpi). We detected MAV-1 E3 protein in the respiratory epithelium 7 dpi. We did not detect viral mRNA or protein at 14 dpi, but MAV-1 DNA was detected by PCR at 21 dpi. Chemokine transcript levels increased between 7 and 14 dpi in the lungs of infected mice. MAV-1 infection induced a patchy cellular infiltrate in lungs at 7 and 14 dpi. This is the first report demonstrating the presence of MAV-1 in the respiratory epithelium of infected mice and describing chemokine responses in the lung induced by MAV-1 respiratory infection. MAV-1 infection of mice has the potential to serve as a model for inflammatory changes seen in human adenovirus respiratory disease.

Introduction

Human adenoviruses cause a wide range of upper and lower respiratory tract infections in children (Carballal et al., 2002; Edwards et al., 1985; Larranaga et al., 2000; Pacini et al., 1987). Adenoviruses are also a common cause of upper respiratory tract infection and pneumonia in military recruits (Huebner et al., 1958; Kolavic-Gray et al., 2002; Mogabgab, 1968). Immunocompromised patients experience greater problems due to adenovirus infections, including more aggressive respiratory infections, transplant loss and death (Kojaoghlanian et al., 2003). A potential long-term consequence of persistent adenovirus infection is an increased risk for the development of asthma and chronic obstructive pulmonary disease (COPD) (Hogg, 1999, 2001). Human adenoviruses are also significant as vectors for in vitro and in vivo delivery of genes into mammalian cells (reviewed in Kay et al., 2001).

Chemokines are chemotactic cytokines that serve as chemoattractants for cells involved in the inflammatory response to stimuli such as infection (Luster, 1998; Rollins, 1997; Taub, 1996). Combined in vitro, ex vivo and in vivo data suggest that both intact (Alcorn et al., 2001; Booth et al., 2004; Harrod et al., 1998; Kajon et al., 2003; Leland Booth and Metcalf, 1999; Otake et al., 1998) and replication-deficient (Amin et al., 1995; Kodama et al., 2001; Noah et al., 1996; Schwarz et al., 1999) human adenoviruses are capable of inducing chemokine responses in human, murine and bovine lung. Chemokines potentially play a significant role in the pathophysiology of asthma (Blease et al., 2000; Lukacs, 2001) and COPD (Beeh et al., 2003; Qiu et al., 2003; Saetta et al., 2002; Traves et al., 2002). Chemokine responses to adenovirus vectors contribute to difficulties encountered with limited persistence of the vector in its host, low levels of expression of desired genes and potential morbidity associated with use of the vector (reviewed in Liu and Muruve, 2003).

A variety of models have been used to study chemokine responses to adenovirus infection. In vitro systems have used infection of cell lines with adenovirus or transduction with adenovirus vectors (Alcorn et al., 2001; Kodama et al., 2001; Leland Booth and Metcalf, 1999; Schwarz et al., 1999). An ex vivo lung slice model has been used to examine chemokine responses to adenovirus infection in the context of a relevant tissue (Booth et al., 2004). The majority of these studies have focused on upregulation of interleukin 8 (IL-8), a CXC chemokine that serves as a major neutrophil chemoattractant and activator (Baggiolini et al., 1994). Relatively few studies have used in vivo animal models to examine chemokine responses to respiratory infection with human adenovirus (Harrod et al., 1998; Kajon et al., 2003). In vivo studies of human adenovirus pathogenesis are limited by the strict species-specificities of the adenoviruses. Because exposure of mice to human adenovirus does not result in a fully permissive infection (Ginsberg et al., 1991; Kajon et al., 2003), this system does not allow for a complete assessment of host inflammatory responses to persistent adenovirus infection. Human adenovirus infection of cotton rats (Sigmodon hispidus) has shown the most promise as an animal model of human adenovirus respiratory disease (Pacini et al., 1984; Prince et al., 1993), but it depends on infection with a high dose of a virus that is not fully adapted to the host. This drawback is particularly relevant in the study of links between adenovirus infection and chronic lung disease such as asthma or COPD, which are more likely to be associated with effects of persistent rather than acute infection.

Mouse adenovirus type 1 (MAV-1) serves as an excellent animal model system for studying adenovirus pathogenesis, providing the means to define viral and host factors involved in both acute and persistent adenovirus infections (reviewed in Smith and Spindler, 1999). The MAV-1 virion is structurally similar to that of human adenoviruses (Wigand et al., 1977), and its genomic organization is also similar to that of human adenovirus (Meissner et al., 1997). With the exception of the early region 3 (E3) gene (Beard et al., 1990), MAV-1 gene products have sequence similarity to those of human adenoviruses (Ball et al., 1988, 1989, 1991; Beard et al., 1990; Cauthen and Spindler, 1996; Kring et al., 1992; Kring and Spindler, 1990). Of particular interest, given the potential association between the human adenovirus early region 1A (E1A) gene and long-term sequelae of infection such as COPD (Elliott et al., 1995; Matsuse et al., 1992), MAV-1 E1A has approximately 40% similarity to human adenovirus E1A, with the strongest similarity in conserved region 2 (CR2) (Ball et al., 1988). MAV-1 E1A shares functional roles of human adenovirus E1A, binding to the mouse cellular proteins, retinoblastoma protein and pRb-related protein (Fang et al., 2004; Smith et al., 1996) and transactivating the human adenovirus type 5 E3 promoter (Ball et al., 1988).

Few studies directly address respiratory infection with MAV-1 or the role of chemokine responses to MAV-1. Following intraperitoneal (i.p.) MAV-1 inoculation of adult C57BL/6 mice, increased mRNA levels for the chemokines IP-10, MCP-1 and TCA-3 were detected in the brain between 3 and 4 days post-infection (dpi) (Charles et al., 1999). MAV-1 organ tropism following intranasal (i.n.) inoculation of outbred mice is similar to that observed following i.p. inoculation (Kajon et al., 1998). In each case, virus was identified in multiple organs including lung, spleen and brain (Kajon et al., 1998). Following i.n. inoculation of newborn BALB/c mice less than 24 h old, peribronchiolar infiltrates comprised of macrophages, and lymphocytes were noted 3 dpi (Gottlieb and Villarreal, 2000).

We and others have used mouse models to study the chemokine responses to respiratory infections with a variety of viruses, including acute viral infections with respiratory syncytial virus (Domachowske et al., 2000; Haeberle et al., 2001; John et al., 2003; Mejias et al., 2004; Miller et al., 2004; Power et al., 2001; Tekkanat et al., 2002) and influenza virus (Dawson et al., 2000; Tumpey et al., 2000) as well as persistent viral infections such as those caused by murine gammaherpesvirus 68 (Sarawar et al., 2002; Weinberg et al., 2002, 2004). In this study, we demonstrate that respiratory infection is established following i.n. inoculation of C57BL/6 mice with MAV-1. Respiratory infection with MAV-1 results in chemokine upregulation and cellular inflammation in the lung. These data indicate that respiratory infection with MAV-1 may serve as a useful model to study adenovirus respiratory disease.

Results

Clinical signs of MAV-1 infection

In order to study MAV-1 respiratory infection, we inoculated 4- to 6-week-old C57BL/6 mice i.n. with 10⁵ plaque-forming units (PFU) of MAV-1 and carefully monitored them until 14 dpi. At no time point did mice infected with MAV-1 exhibit clinical signs suggesting respiratory distress such as tachypnea or labored respirations. CNS symptoms such as ataxia, hyperreflexia, hyper-esthesia, paresis and flaccid paralysis were absent in all infected mice.

Detection of MAV-1 in the lung

Using in situ hybridization, Kajon et al. (1998) demonstrated the presence of MAV-1 in the lungs of adult outbred mice following both i.n. and i.p. inoculation. We sought to confirm and expand on these findings in the current study using inbred mice. In three separate experiments (n = 2 to 6 mice per condition at each time point), we harvested total RNA from lungs of mice at 1, 4, 7 and 14 dpi. We used ribonuclease protection assay (RPA) to detect the expression of MAV-1 E3 and hexon genes in the lung at each time point. We did not detect E3 gene expression at 1 or 4 dpi (data not shown). We detected E3 gene expression in 4 of 6 mice at 7 dpi (Fig. 1A) but did not detect E3 gene expression in any mice at 14 dpi. Expression of the late gene encoding the structural protein, hexon, was also detected at 7 dpi but not at 14 dpi in the lungs of infected mice (Fig. 1A). Paralleling these data, we found evidence of infectious virus in lung homogenates at 7 dpi but not at 14 dpi by plaque assay titration on mouse 3T6 fibroblast monolayers (data not shown). As expected, we did not detect E3 or hexon gene expression in the lungs of mock-infected animals at any time point (Fig. 1A).

Fig. 1.

Detection of MAV-1 gene expression and DNA in lung. (A) RNA was obtained from the lungs of mice infected i.n. with 10⁵ PFU of MAV-1-or mock-infected with conditioned media. RNA was analyzed using RPA to detect expression of the MAV-1 E3 and hexon genes and the GAPDH and L32 mouse housekeeping genes. Protected viral probe fragments are shown from RNA isolated from mouse lungs at 7 and 14 dpi. Lanes 1 to 3 present data for individual mice from one experiment at 7 dpi, and lanes 4 to 9 present data for individual mice from a separate experiment at 7 and 14 dpi. (B) DNA was obtained from lungs harvested from infected and mock-infected mice at 21 dpi. PCR was used to amplify a segment of the MAV-1 E1A gene. β-actin was successfully amplified from every sample as a control (data not shown).

MAV-1 has a tropism for endothelial cells in the CNS and other organs (Charles et al., 1998; Guida et al., 1995; Kajon et al., 1998; Moore et al., 2003). To determine whether MAV-1 is capable of infecting respiratory epithelial cells following i.n. inoculation, we performed immunohistochemistry on paraffin-embedded sections of lungs harvested at various time points following infection to localize MAV-1 E3 protein. At 7 dpi, virus-specific staining was evident in respiratory epithelial cells lining the medium and large airways of the lung (Fig. 2). At 14 dpi, MAV-1-specific staining was not present in respiratory epithelial cells of infected lungs, although there was occasional staining of vascular endothelial cells in the lung (arrow, Fig. 2). No staining was noted using preimmune rabbit serum as a primary antibody (data not shown).

Fig. 2.

Localization of MAV-1 in lung. Immunohistochemistry was performed on paraffin-embedded sections from the lungs of mice infected i.n. with 10⁵ PFU of MAV-1- or mock-infected with conditioned media. Polyclonal rabbit antiserum was used to detect MAV-1 E3 protein at 7 and 14 dpi. Staining of mock-infected lungs is shown for comparison. Dark brown staining indicates virus-specific staining, which is present in respiratory epithelial cells at 7 dpi. Open arrow indicates virus-specific staining of vascular endothelial cells at 14 dpi. Scale bar, 100 μm.

Expression of hexon and E3 genes was not detected in lungs harvested from a separate group of mice at 21 dpi (data not shown). However, we detected the presence of MAV-1 DNA in the lungs of 2 of 3 mice at 21 dpi by PCR (Fig. 1B). Together, these data suggest that i.n. inoculation with MAV-1 resulted in a productive pulmonary infection in C57BL/6 mice involving respiratory epithelial cells. Following acute infection, MAV-1 genome persisted in the lung in the absence of detectable viral gene expression.

Chemokine responses to MAV-1 infection

Chemokine upregulation induced by MAV-1 has been described in the CNS but not the lung following i.p. inoculation (Charles et al., 1999). We used RPA to quantify chemokine gene expression in the lungs of mice infected i.n. with MAV-1. Data from a representative experiment are presented in Fig. 3. Prior to infection or at 1 and 4 dpi, we noted little chemokine expression in the lungs of mice, and levels of chemokine expression did not differ between MAV-1-infected and mock-infected mice (data not shown). In contrast, MAV-1 infection upregulated the expression of nearly all chemokine genes measured at later time points relative to mock infection (Fig. 3A). Levels of expression of IP-10 and Ltn peaked at 7 dpi, while levels of expression of MIP-1α, MIP-1β and RANTES peaked at 14 dpi. MCP-1 and TCA-3 expression was increased in MAV-1-infected animals to similar levels at 7 and 14 dpi. No significant differences in MIP-2 expression between MAV-1-infected and mock-infected mice were detected at any time point. Expression of the eotaxin gene was not reliably detected by RPA in lungs of MAV-1-infected or mock-infected mice at any time point (data not shown). Similar RPA results were obtained in separate experiments from infected mice at identical time points (data not shown).

Fig. 3.

Chemokine responses to MAV-1 in the lung. (A) RPA measurement of chemokine gene expression at the indicated time points in the lungs of mice infected i.n. with 10⁵ PFU of MAV-1 or mock-infected with conditioned media. Quantified data are presented as percentages of the expression of the L32 housekeeping gene. (B) ELISA measurement of protein levels of the representative CC chemokine RANTES in the lungs of MAV-1-infected and mock-infected mice. For both RPA and ELISA, data are presented as means ± standard deviation of values obtained from three mice at each condition (except mock infection at 7 dpi, where n = 2) from one representative experiment corresponding to lanes 4 to 9 in Fig. 1. Statistical significance was determined using Student’s t test (*P < 0.05). Similar results were obtained from a separate independent experiment.

To verify that altered levels of chemokine gene expression after infection were reflected in chemokine protein levels, we measured RANTES protein levels in lung homogenates at 7 and 14 dpi (Fig. 3B). At 7 dpi, levels of RANTES protein were not increased in the lungs of infected compared to mock-infected mice. At 14 dpi, when RANTES gene expression was dramatically upregulated in the lungs of infected mice (Fig. 3A), levels of RANTES protein were also significantly elevated compared to mock-infected mice (Fig. 3B).

Cellular inflammatory responses to MAV-1 infection in the lung

Chemokines are chemotactic cytokines that form a concentration gradient to attract cells involved in the inflammatory response to pathogens and other stimuli (Luster, 1998). Based on the chemokine upregulation we detected in the lungs and brains of infected mice, we predicted that a cellular inflammatory response would accompany MAV-1 respiratory infection. To assess cellular inflammation, we examined hematoxylin and eosin-stained sections of lungs and brain from MAV-1-infected and mock-infected mice. No cellular infiltrate was present in the lungs of mock-infected mice at any time point (Fig. 4A) or in the lungs of infected mice at 1 and 4 dpi (data not shown). A mild patchy interstitial pneumonitis was present in lungs of infected mice at both 7 and 14 dpi (Fig. 4A) that was characterized by a predominantly mononuclear infiltrate and thickened alveolar walls. In addition, scattered areas of hypercellularity were focused around medium and large airways in the lungs of infected mice at 7 dpi; these areas consisted predominantly of mononuclear cell infiltrates. These changes were also seen, though to a lesser extent, at 14 dpi. Thus, maximum cellular inflammation in the lung was present at 7 dpi when virus was detected in the lung by RPA, immunohistochemistry and plaque assay and was reduced when virus had been cleared from the lung at 14 dpi.

Fig. 4.

Cellular inflammatory responses to MAV-1 respiratory infection. (A) Hematoxylin and eosin-stained sections of lungs of MAV-1-infected and mock-infected mice demonstrate a mononuclear cellular infiltrate surrounding airways that is most prominent at 7 dpi and a mild interstitial pneumonitis that is present at both 7 and 14 dpi. Scale bar, 100 μm. (B) Hematoxylin and eosin-stained cytospins of BAL fluid demonstrate an increased number of neutrophils (open arrows) in the airways of MAV-1-infected mice at 7 dpi. (C) Neutrophils as a percentage of all cells in bronchoalveolar lavage fluid collected from mock-infected (n = 3) and MAV-1-infected (n = 5) mice at 7 dpi. Each symbol represents an individual mouse.

In a separate experiment, we used bronchoalveolar lavage (BAL) to characterize the airway cellular inflammatory response to MAV-1 respiratory infection at 7 dpi, when the greatest amount of cellular inflammation was noted in the lungs (Fig. 4B). BAL fluid from mock-infected animals had a monotonous population of alveolar macrophages. These cells were also evident in BAL fluid from MAV-1-infected animals. In addition, BAL fluid from infected animals contained a variable degree of neutrophils (open arrow, Fig. 4B), ranging from 1.5% to 28.3% of the total population of cells (Fig. 4C). In contrast, BAL fluid from mock-infected animals contained only between 0.3% and 0.5% neutrophils. Only rare leukocytes of other types were noted in either mock-infected or MAV-1-infected animals, and their percentages did not differ among groups (data not shown).

Discussion

We demonstrate here that MAV-1 is capable of establishing a respiratory infection in the lungs of C57BL/6 mice and describe the novel finding that MAV-1 protein is present in respiratory epithelial cells following i.n. inoculation. We observed a parallel between viral gene expression and chemokine upregulation in the lungs. As expected from the chemokine upregulation we detected, MAV-1 infection in the lung elicited a cellular inflammatory response that resulted in an interstitial pneumonitis and foci of inflammation around larger airways.

Previous work from our laboratory showed that MAV-1 DNA could be detected using in situ hybridization in the lung following both i.n. and i.p. inoculation of NIH Swiss outbred mice with MAV-1, but its distribution was limited to the vascular endothelium (Kajon et al., 1998). Here, we found MAV-1 protein in both vascular endothelium and respiratory epithelium following i.n. inoculation of C57BL/6 mice. Several possibilities may explain the apparent difference in cell tropism. The lack of detection of MAV-1 nucleic acid in the earlier report is possibly due to a lower viral dose, 10³ PFU, versus 10⁵ PFU used here. The higher dose may increase the efficiency of entry into a cell. Following virus entry, the higher dose may result in increased viral replication to levels whereby viral protein is detectable by immunohistochemistry. We believe that a less likely explanation is that detection of protein in respiratory epithelial cells using immunohistochemistry was more sensitive than detection of MAV-1 DNA by in situ hybridization.

Another possibility explaining the apparent cell tropism difference between this study and the previous study (Kajon et al., 1998) is that different mouse strains were used, C57BL/6 and NIH Swiss outbred mice, respectively. Differential host susceptibility to MAV-1 infection is seen among various inbred and outbred mouse strains (Guida et al., 1995; Kring et al., 1995; Spindler et al., 2001). The 50% lethal dose (LD₅₀) for outbred NIH Swiss outbred mice is 17 PFU (Kring et al., 1995), much lower than the LD₅₀ of >10^4.4 PFU for C57BL/6 mice (Spindler et al., 2001). The underlying basis for these differences in susceptibility is unknown but is likely to involve mouse strain-specific differences in immune responses to MAV-1. For instance, chemokine responses to MAV-1 infection in the brain differ between C57BL/6 and BALB/c mice (Charles et al., 1999). It is therefore possible that our ability to detect MAV-1 in the respiratory epithelium in the present study and not in our previous study (Kajon et al., 1998) is based on differing abilities of the two mouse strains to efficiently prevent or control viral infection in respiratory epithelial cells.

Histopathologic descriptions of adenovirus respiratory disease in humans are quite limited and largely include descriptions of fatal disease and disease in immunocompromised individuals (Becroft, 1967, 1971; Chany et al., 1958; Zahradnik et al., 1980). MAV-1-induced respiratory disease in C57BL/6 mice appears to share some characteristics with human disease. The respiratory epithelium appears to be involved in both cases, although the significant epithelial damage seem in some human cases (Becroft, 1967, 1971; Chany et al., 1958) was not prominent in this study. Histopathologic features of respiratory disease induced by MAV-1 are also similar in some ways to those present in the cotton rat model of human adenovirus type 5 (Prince et al., 1993), including the presence of viral antigen in respiratory epithelial cells and a cellular infiltrate of peribronchiolar regions that was most pronounced at 7 dpi in each case. MAV-1 infection of C57BL/6 mice resulted in a patchy interstitial pneumonitis (Fig. 4A); an early intra-alveolar cellular exudate, which was comprised predominantly of macrophages and polymorphonuclear leukocytes in the cotton rat model (Prince et al., 1993), was present following MAV-1 infection, although to a lesser extent than in human adenovirus infection of cotton rats. Although the riboprobe template used to measure chemokine gene expression in this report includes fewer CXC than CC chemokines, it is perhaps notable that MIP-2, a CXC chemokine, was the only chemokine detected that was not upregulated following MAV-1 infection. MIP-2 serves as a potent neutrophil chemoattractant in lung injury (reviewed in Guo and Ward, 2002), and the absence of MIP-2 upregulation in the setting of MAV-1 respiratory infection in part may explain the less prominent neutrophil influx into the airways in the present report.

Adenoviruses are capable of establishing persistent infections in their hosts (reviewed in Horwitz, 2001). Persistent adenovirus infection may contribute to chronic lung disease. The human adenovirus E1A protein is expressed in airway epithelial cells and type 2 alveolar cells in human lung tissue (Elliott et al., 1995). Human adenovirus E1A DNA was detected more often in lung tissue from patients with COPD than in tissue from control subjects (Matsuse et al., 1992). Human adenovirus DNA was isolated more frequently in the lungs of children with asthma than in healthy children (Marin et al., 2000). Adenovirus capsid protein was detected in BAL fluid of children with steroid-resistant asthma but not in children without persistent asthma (Macek et al., 1994). In a guinea pig model, persistent human adenovirus infection was associated with chronic lung inflammation (Vitalis et al., 1996), enhanced inflammatory response to cigarette smoke exposure (Vitalis et al., 1998) and lung destruction following chronic cigarette smoke exposure (Meshi et al., 2002). In addition, persistent adenovirus infection of guinea pigs modifies steroid responsiveness to ovalbumin-induced allergic lung disease (Yamada et al., 2000), potentially by inhibiting AP-1-mediated effects of steroid on the expression of the chemokines eotaxin and MCP-1 (Yamada et al., 2002).

Persistent MAV-1 infections have also been documented (reviewed in Smith and Spindler, 1999). In mice infected with MAV-1, high titers of virus were found in the urine for up to 24 months (van der Veen and Mes, 1973) and in the liver up to 52 dpi (Wigand, 1980). Using PCR amplification and in situ hybridization, Smith et al. demonstrated persistence of viral DNA in the brain, spleen, kidneys and lymph nodes of NIH Swiss outbred mice for up to 55 weeks following i.p. inoculation of MAV-1 (Smith et al., 1998). We detected MAV-1 DNA in the lungs of infected mice at 21 dpi using PCR (Fig. 1B). Using nested PCR to amplify a portion of the MAV-1 E1A coding sequence, we have also detected MAV-1 genome in the lungs of infected BALB/c mice at 44 dpi (data not shown). MAV-1 therefore has the potential to serve as a powerful tool to study the role of persistent adenovirus infection in chronic lung disease.

It is clear from the data presented in this report that i.n. inoculation of C57BL/6 mice results in an acute respiratory infection. Given the availability of multiple transgenic and knockout mouse strains together with mutant MAV-1 deficient in the production of E1A (Smith et al., 1996) or E3 proteins (Beard and Spindler, 1996; Cauthen et al., 1999), MAV-1 infection will allow the study of specific host inflammatory responses and specific viral genes involved in the pathogenesis of adenovirus-induced respiratory infection.

Materials and methods

Viruses, mice and experimental infections

Wild-type MAV-1 was grown and passaged in NIH 3T6 fibroblasts, and titers of viral stocks were determined by plaque assay on 3T6 cells as previously described (Cauthen and Spindler, 1999). Four- to six-week-old male C57BL/6J mice were purchased from Harlan and maintained in microisolator cages. Under light isoflurane anesthesia, mice were inoculated i.n. with 10⁵ PFU in 20 μl sterile PBS. Control mice were inoculated i.n. with conditioned media at an equivalent dilution in PBS. Mice were euthanized at the indicated time points. Organs were harvested, snap frozen in a dry ice and ethanol bath and stored at −20 °C until processed further. All animal work complied with all relevant federal and institutional policies.

Histology

In a subset of mice, lungs and brain were harvested and fixed in 10% formalin. Prior to fixation, lungs were gently inflated with PBS via the trachea to maintain lung architecture. After fixation, organs were embedded in paraffin, and 5 μm sections were cut for histopathology and immunohistochemistry. Sections were stained with hematoxylin and eosin to evaluate cellular infiltrates. All sectioning and staining was performed by the University of Michigan Department of Pathology Histology Research Laboratory.

Immunohistochemistry

Sections were deparaffinized in Citri-solv (Fisher) and rehydrated in sequential ethanol baths (100%, 95% and 75% ethanol for 5 min each) followed by a PBS wash (10 min). Antigen retrieval was accomplished using sequential washes in 100 mM Tris–HCl, pH 10, boiling in a microwave oven (4 washes, 5 min each). Slides were allowed to cool gradually to room temperature and were then washed twice in phosphate-buffered saline (PBS). Endogenous peroxidase activity was quenched with hydrogen peroxide, and slides were blocked by incubating with horse serum (1:667 in PBS). MAV-1 protein was detected using αE3-1 antiserum (Beard and Spindler, 1995) diluted 1:2000 in blocking buffer. This polyclonal rabbit antiserum was raised against a GST fusion protein containing the unique amino acid sequence from the E3 gp11K protein. Rabbit preimmune serum was used as a negative control. The secondary antibody was biotinylated anti-rabbit antibody (Vectastain ABC Kit, Vector Laboratories). Detection was accomplished using the Peroxidase Substrate Kit (Vector Laboratories) according to the manufacturer’s instructions. Slides were counterstained for 5 min with hematoxylin.

Bronchoalveolar lavage

Airspace inflammatory cells were collected by BAL. Lungs of euthanized animals were lavaged three times with 0.8 ml phosphate-buffered saline. Cells in the BAL fluid were pelleted by centrifugation, cytospun onto glass slides and stained with hematoxylin and eosin. Cells from ten randomly selected high-power fields were counted on each slide, and the proportion of different leukocyte subtypes calculated.

Isolation of RNA and DNA

The left lung was homogenized with sterile glass beads and 1 ml TRIzol Reagent (Gibco-BRL) using a Mini Beadbeater (Biospec Products). The homogenates were incubated at room temperature for 10 min, and then 200 μl of chloroform was added to each sample. Following a 3-min incubation at room temperature, the tubes were centrifuged at 12,000 ×g for 15 min at 4 °C. The aqueous phase was transferred to a new microcentrifuge tube, and total RNA was precipitated with isopropanol. DNA was extracted from approximately one-third of the right lung using the DNeasy Tissue Kit (Quiagen Inc.) according to the manufacturer’s protocol.

Ribonuclease protection assays

Total RNA was analyzed by RPA to measure viral, chemokine and cytokine gene expression as previously described (Rochford et al., 2001; Weinberg et al., 2002). The RPAs are performed in probe excess, so they give a quantitative measure of the specific RNAs present. Expression of murine chemokines and cytokines was measured using the mCK-5c riboprobe template (PharMingen). An MAV-1 E3 probe was prepared from plasmid pZU14 as described (Fang et al., 2004). The full-length E3 template measured 714 nucleotides (nt), and the size of the protected fragment was 645 nt. A genomic MAV-1 hexon probe was prepared from plasmid pHEX as described (Fang et al., 2004). The full-length hexon template measured 395 nt, and the size of the protected fragment was 337 nt. The templates for internal β-actin (Fang et al., 2004) and L32 (Hobbs et al., 1993) controls were previously described. Linearized E3 and β-actin templates were combined in equimolar concentrations for use in riboprobe syntheses driven by T3 bacteriophage RNA polymerase. Linearized MAV-1 hexon and L32 templates combined in equimolar concentrations for use in riboprobe syntheses were driven by T7 bacteriophage RNA polymerase. All syntheses used [α-³²P]UTP (Amersham) as the labeling nucleotide. Probe bands were visualized by autoradiography and were quantified where indicated using the Storm PhosphorImager and ImageQuant software (Molecular Dynamics).

PCR amplification of MAV-1 E1A

DNA extracted from the lungs of mice was amplified using PCR to detect the presence of MAV-1 E1A. 80 ng of DNA was amplified by PCR in a reaction volume of 20 μl containing 4 μl of 5× GoTaq Reaction Buffer (Promega, final concentration 1.5 mM MgCl₂), 0.5 U GoTaq DNA polymerase (Promega), 0.2 mM dNTPs and 150 ng each of E1A-FO1 (5′ ATG TCG CGG CTC CTA CG 3′) and E1A-RE1 (5′ CAA CGA ACC ATA AAA AGA CAT CAT 3′). The PCR amplification was carried out for 35 cycles at an annealing temperature of 50 °C. Amplified products were separated on a 1.8% agarose gel, stained with ethidium bromide and photographed using an Electrophoresis Documentation and Analysis System (Kodak).

Measurement of chemokine protein

Protein levels of the murine chemokine RANTES were measured in supernatants from organ homogenates obtained as described above. Specific chemokines were quantified using a standard method of sandwich ELISA as previously described (Hogaboam et al., 1998; Walley et al., 1997). In brief, microtiter plates were coated with the appropriate chemokine capture antibody. Samples were added to wells for 1 h at 37 °C. Recombinant murine chemokine standard curves were used to calculate the chemokine concentrations. Detection and processing were performed with the appropriate biotinylated polyclonal rabbit anti-cytokine antibody (3.5 μg/ml), streptavidin–peroxidase conjugate (Bio-Rad) and chromogen substrate (Bio-Rad). Plates were read on an ELISA plate scanner at 492 nm.