Suppression of Herpes Simplex Virus Type 1 (HSV-1)–induced Pneumonia in Mice by Inhibition of Inducible Nitric Oxide Synthase (iNOS, NOS2)

Female CBA/J (H-2^k) mice, purchased from The Jackson Laboratory (Bar Harbor, ME), were used at 12–26 wk of age.

Mice were infected intranasally with 1 × 10⁶ PFU HSV-1, strain KOS 1.1 (25), kindly provided by Drs. R. Finberg and J. Brubaker (Dana-Farber Cancer Institute, Boston, MA), given in a final volume of 0.03 ml sterile 0.9% sodium chloride (Fujisawa USA, Inc., Deerfield, IL) per mouse. For survival experiments, mice were infected with 2 LD₅₀ (5 × 10⁷ PFU per mouse in this system). The virus stock used for this study was clarified supernatant derived from infected Vero cell cultures (Advanced Biotechnologies, Inc., Columbia, MD). The Vero cells and culture media were screened and shown to be free of mycoplasma and endotoxin contamination.

l-NMMA (Alexis Corp., San Diego, CA) dissolved in PBS (pH 7.4), was given at a dose of 2 mg/day i.p. in a volume of 0.2 ml from the day of virus infection until analysis of mice as described (12). Infected control mice were treated with the same amount of N^G-monomethyl-d-arginine (d-NMMA) (Sigma Chem. Co., St. Louis, MO) or with PBS. Treatment of uninfected mice with l-NMMA did not affect any of the tested parameters, when compared to uninfected, untreated mice (data not shown).

Mice were euthanized by CO₂ asphyxiation 7 d after infection, blood was collected by cardiac puncture, and the trachea was cannulated using a 19-gauge tubing adapter (Becton Dickinson, Rutherford, NJ) inserted through an incision immediately distal to the larynx. Bronchoalveolar lavage (BAL) was performed with sterile 0.9% sodium chloride, using three separate 1.0-ml aliquots, each of which was infused and withdrawn five times. The cells were centrifuged and resuspended in RPMI medium containing 10% FCS for further analysis. When indicated, the left lung was processed for virus titration, and the right lower lobe of the lung was used for histological examination. Previous studies demonstrated that the virus titration and histology were not affected by the lung lavage procedure (Adler, H., L. Kobzik, and I.J. Rimm, unpublished results).

BAL lymphocytes were stained for flow cytometry and analyzed using CellQuest on a FACScan^® (Becton Dickinson). The following monoclonal antibodies, conjugated either with fluorescein isothiocyanate or phycoerythrin were used for staining: anti-B220, anti-CD4, anti-CD8, anti-CD3, anti-α/β T cell receptor, and anti-γ/δ T cell receptor (PharMingen, San Diego, CA). Fluorescein isothiocyanate- or phycoerythrin-conjugated isotype-matched antibodies were used as negative controls.

The left lung was homogenized by grinding in 1 ml of sterile 0.9% sodium chloride and stored frozen at −80°C. Viral titers were determined from thawed aliquots by plaque assay on Vero cells, and the number of plaques was determined after staining the cells with crystal violet.

For histology, the lower lobe of the right lung was inflated with O.C.T. compound (Miles Inc., Elkhart, IN) and fixed in 10% formalin in PBS (Fisher Scientific, Fair Lawn, NJ). After fixation, the samples were dehydrated and embedded in paraffin. 5-μm-thick sections were cut and stained with hematoxylin/ eosin (H/E). The samples were analyzed and scored microscopically in a blind study on coded slides as described (26). Two types of pathological changes were scored separately. Histopathologic changes around airways and vessels were termed periluminal, and histopathologic changes within the alveolar parenchyma were scored as parenchymal abnormalities. A final score of pneumonitis from 0 to 9 was assigned for each pathologic process (periluminal and parenchymal) as the product of severity grade by the extent. The score that is presented reflects the sum of both scores.

For immunohistochemistry, mouse lungs were inflated by intratracheal instillation of O.C.T. compound and immediately frozen in a dry ice/2-methylbutane bath. Immunostaining was performed on 8-μM cryostat sections using a standard peroxidase anti-peroxidase protocol (27). Primary antibodies included rabbit anti-NOS2 or anti-nitrotyrosine (Upstate Biotechnology Inc., Lake Placid, NY) or normal rabbit IgG (Sigma) used at 1 μg/ml. Diaminobenzidine was used as a chromogen, and slides were counterstained with hematoxylin before microscopic evaluation.

As parameter of lung function, dynamic compliance (C_dyn, defined as the increase in lung volume obtained from a given increase in distending pressure at points of zero flow) was determined in vivo 7 d after infection. C_dyn is a parameter that generally reflects the condition of the lung parenchyma, and is decreased by pulmonary interstitial edema, or inflammation, or by any condition that reduces alveolar patency. We used a previously described plethysmographic technique for making measurements of lung function (28). Anesthetized mice (70–90 mg/kg Na pentobarbital i.p.) underwent placement of a tracheostomy tube for mechanical ventilation (7 ml/kg/breath, 150 breaths/min) and thoracotomy to eliminate any effect that the chest wall may have on pulmonary function.

HSV-1–specific IgG2a antibodies in serum samples of mice 7 d after infection were determined using a specific ELISA (29). In brief, microtiter plates coated with HSV-1 Macintyre Strain Viral Lysate (Advanced Biotechnologies Inc., Columbia, MD) were incubated with serial dilutions of a standard serum or test samples, and after extensive washing, alkaline phosphatase conjugated goat anti–mouse IgG2a (Southern Biotechnology, Birmingham, AL) was added. The plates were developed using p-nitrophenyl phosphate (Sigma) in DEA buffer. The reaction was stopped with 3 N NaOH, and absorbance was read at 405 nm using an ELISA reader. In all assays, samples and standards were run at least in duplicates. The calculation of the HSV-1–specific antibodies was performed by the method of Zollinger and Boslego (30). As expected, no HSV-1–specific IgG2a antibodies were detected in sera of uninfected mice. In contrast, in all sera of infected mice, regardless of their treatment, HSV-1–specific IgG2a antibodies were readily detected (data not shown). Thus, all infected animals seroconverted in response to infection.

The Student's t test (unpaired) and the Wilcoxon rank sum test were used for data analysis where appropriate. Treatment-group differences in the viral titer per lung were analyzed by one-way analysis of variance (ANOVA).

Intranasal infection of mice with HSV-1 led to the development of pneumonia. Histopathological changes in infected animals were most evident at day 7 after infection. The pneumonia was characterized histologically by mononuclear infiltrates (Fig. 1, compare A with D), an increase in the histology score (Fig. 2, compare uninfected/ l-NMMA with infected/PBS), an increase in the number of cells recovered by BAL (Fig. 3, A and B, compare uninfected/l-NMMA with infected/PBS), a change in the CD4/ CD8 ratio (Table 1, compare uninfected/untreated with infected/PBS), and by a decrease in lung compliance (see below).

To demonstrate the induction of iNOS and the production of NO during HSV-1–induced pneumonia, we performed immunohistochemistry on lung tissue of uninfected and infected mice at days 3 and 7 after infection, using antibodies against murine iNOS and nitrotyrosine. iNOS was detected within inflammatory cells in infected mice, but not in uninfected mice. Similarly, nitrotyrosine staining was observed in infected mice, but not in uninfected mice (data not shown). Thus, these results demonstrated that NO is indeed produced during the course of HSV-1 infection, confirming previous data (24). NO reacts with O₂⁻ to generate peroxynitrite, which is involved in microbial killing as well as toxicity (9, 31, 32). Nitrotyrosine is a marker of peroxynitrite generation, since nitrotyrosine is a product resulting from peroxynitrite action (12).

Since iNOS and nitrotyrosine staining demonstrated the involvement of NO in HSV-1 pneumonia, experiments were designed to test the effect of l-NMMA treatment on HSV-1 pneumonia. Mice were infected intranasally with 1 × 10⁶ PFU of HSV-1 per mouse, and histological examination of lungs was performed at day 7 after infection. In HSV-1–infected, PBS-treated control mice, pneumonitis was characterized by mononuclear infiltrates around vessels and bronchioles (periluminal), and a diffuse intraalveolar infiltrate consisting mainly of mononuclear cells (parenchymal) (Fig. 1,A). In contrast, the development of pneumonia was strongly suppressed in mice treated with l-NMMA (Fig. 1,B). The suppression of pneumonia was specific, since the biologically inactive stereoisomer d-NMMA did not show this effect (Fig. 1,C). Normal lungs were observed in uninfected mice either treated with l-NMMA (Fig. 1,D) or left untreated (data not shown). Histological scores reflecting the periluminal and parenchymal changes in the lungs of infected mice were significantly lower in infected animals treated with l-NMMA, compared to infected mice treated with d-NMMA or PBS. In fact, treatment with l-NMMA suppressed the pneumonitis scores nearly to the level observed in uninfected mice (Fig. 2).

Intranasal infection with HSV-1 led to a pronounced increase in the total cell number as well as in the lymphocyte number recovered by BAL at day 7 after infection, when compared to uninfected animals (Fig. 3, A and B). l-NMMA treatment, however, markedly inhibited the influx of inflammatory cells. In contrast, treatment with d-NMMA showed no effect. Thus, the BAL data were consistent with the results obtained by histology.

Since lymphocytes were the predominant cells recruited to the lung after HSV-1 infection, we analyzed whether treatment with l-NMMA changed the composition of the lymphocyte population. Neither treatment with l-NMMA nor d-NMMA altered the lymphocyte subpopulations in infected mice, when compared with infected, PBS-treated control mice (Table 1). All groups of infected mice showed a dominance of CD3⁺ T cells. The CD4/CD8 ratio changed from 2:1 in uninfected mice to ∼1:4 in infected mice. The vast majority of the cells were α/β TCR⁺; very low numbers of γ/δ TCR⁺ cells were detected (data not shown).

To assess the effect of l-NMMA treatment on the development of pneumonia, we also measured lung compliance as a parameter of in vivo global lung function. Pulmonary compliance is a quantitative measure of bronchopulmonary injury; additionally these results estimate the magnitude of the physiological derangement resulting from the infection. A reduction in lung compliance, or the most closely related human pulmonary function test parameter, total lung capacity, to 30% of a previously normal level would be considered a life-threatening change of pulmonary status. Lung compliance was 0.032 ± 0.002 (mean ± SE; n = 3) in uninfected mice treated with l-NMMA, which was similar to values measured in uninfected, untreated mice (data not shown). In infected, d-NMMA– treated mice, compliance was reduced to 0.010 ± 0.003 (n = 3), i.e., to 30% of the level in uninfected mice. In contrast, treatment of infected mice with l-NMMA significantly improved lung compliance to 0.029 ± 0.001 (n = 4), when compared to infected, but d-NMMA–treated mice (P <0.004; Student's t test). Thus, l-NMMA treatment improved lung compliance of infected animals nearly to the levels measured in uninfected mice.

Mice were infected intranasally with 5 × 10⁷ PFU of HSV-1 per mouse (∼2 LD₅₀), and treated with l-NMMA or PBS daily for 7 d after infection. Survival of mice was monitored over a period of 14 d after infection. l-NMMA treatment led to a significant improvement in the survival rate (Fig. 4).

To assess the role of NO in the antiviral defense, pulmonary viral titers were determined at days 2, 3, 4, and 7 after infection. Similar viral titers were observed at day 2 after infection between mice treated either with l-NMMA or PBS (Fig. 5). However, a 17-fold higher pulmonary viral titer was observed at day 3 in the l-NMMA–treated mice, when compared to mice treated with PBS. At day 4 after infection, a 2.5-fold higher viral titer was measured in the l-NMMA–treated mice. Virus was cleared at day 7 after infection in all groups of mice. The difference in pulmonary viral titer between the l-NMMA and PBS group was significant.

The significantly higher pulmonary viral titer observed at day 3 after infection in the l-NMMA–treated mice was not accompanied by an increase in the extent of pneumonitis, as demonstrated by a comparable histopathology score and a similar number of BAL inflammatory cells for both the l-NMMA– and PBS-treated mice (Table 2).

Intranasal infection with HSV-1 led to the development of pneumonia. We demonstrated that inhibition of NO production led to a significantly improved survival rate as well as to a strong suppression of the pneumonia despite the presence of a 17-fold higher viral titer in the lung at 3 d after infection. These results suggested that the proinflammatory effects of NO are more important than its antiviral effector functions for the development of HSV-1–induced pneumonia. Furthermore, these results also support the notion that the inflammatory response to HSV-1 infection, rather than direct viral cytopathic effects, is critical for the development of pneumonia. Immunopathology is involved in other HSV-1 infections, for example, an immune-mediated pathogenesis that involves T cells is the basis for herpetic stromal keratitis (33, 34). A T cell–mediated immunopathology is also believed to be involved in the pathogenesis of CMV-induced pneumonia (35). Consequently, reduction of virus growth did not prevent, but did moderate, murine CMV pneumonitis (36, 37). Therefore, treatment of HSV-1–induced pneumonia with antiviral drugs alone might not be sufficient, and a modulation of components of the immune response may also be required. Administration of selective inhibitors of NOS activity in the treatment of inflammatory airway diseases has been suggested (38, 39).

In our study, inhibition of NO production led to a 17-fold increase in pulmonary viral titer at 3 d after infection, confirming earlier reports on an antiviral activity of NO against HSV (18, 19). Therefore, it was remarkable that inhibition of NO production strongly suppressed the development of pneumonia. A recent study of the pathogenesis of influenza virus-induced pneumonia in mice demonstrated that inhibition of NO production by l-NMMA administration resulted in a significant improvement of the survival rate of virus-infected mice (12). However, in this case, l-NMMA treatment did not significantly affect the propagation of influenza virus in vivo.

Inhibition of NO production might lead to suppression of pneumonia by preventing the influx of inflammatory cells to the lung. NO can cause vasodilation, leading to alterations in blood flow, which may promote leukocyte infiltration. NO can also modulate leukocyte adhesion to endothelial cells and alter microvascular permeability during inflammation (9, 31). In vivo neutralization of IFN-γ reduced the magnitude of the pulmonary cellular infiltrate after pulmonary influenza virus infection (40). Thus, IFN-γ may recruit leukocytes into inflamed lung parenchyma. IFN-γ is a strong iNOS inducer, able to induce NO production as a single stimulus (41). Thus, the mechanism of IFN-γ action in this system may be to increase NO production. IFN-γ is also induced during HSV-1 infection of the lung (Adler, H., and I.J. Rimm, unpublished results).

NO or its reaction products might directly cause tissue damage and the release of cytokines and chemokines, which then attract inflammatory cells to the site of virus infection. Influenza A virus infection selectively induced chemokines responsible for the predominant mononuclear leukocyte infiltration in virally infected tissue (42). NO induces NF-kB binding activity and may lead to secretion of proinflammatory lymphokines and cytokines from activated lymphocytes (8, 43).

Treatment with l-NMMA clearly did not prevent the development of pneumonia by interfering with the viral infection of the lung for the following reasons: (a) higher viral titers were detected in l-NMMA treated animals (Fig. 5). (b) The change in the CD4/CD8 ratio of the BAL lymphocytes, which is characteristic for a viral infection of the lung (44), was not affected by the treatment (Table 1). (c) All infected animals seroconverted in response to the intranasal infection (data not shown).

Although there is currently controversy regarding the production of NO by human macrophages (45), increasing numbers of reports demonstrate iNOS induction or NO production in human diseases (46–50). Moreover, NO can be produced by iNOS of a number of other cells types including for example pulmonary epithelial cells (31, 39, 51). In conclusion, we demonstrated an important role of NO in HSV-1–induced pneumonia in mice that may have substantial implications for the therapy of this disease.

The authors thank Dr. Steven Burakoff for helpful discussions and Dr. Paul Martin for critical reading of the manuscript.

This study was supported by grants from the National Institutes of Health, the Mallinckrodt Foundation and the Aid for Cancer Research. I.J. Rimm is a Claudia Adams Barr Investigator. H. Adler is supported by a grant from the Deutsche Forschungsgemeinschaft.