Pulmonary surfactant phosphatidylglycerol inhibits respiratory syncytial virus–induced inflammation and infection

POPG Inhibits RSV-Elicited Cytokine Production in Human Epithelial Cells.

We first examined the effects of POPG on cytokine production (IL-6 and IL-8) induced by RSV in bronchial epithelial cells. Normal human bronchial epithelial cells (NHBE) were pretreated with POPG for 1 h, followed by infection with RSV at a multiplicity of 2/cell. The untreated and RSV infected cultures were examined at 48 h and 72 h after infection. As shown in Fig. 1, control cultures secreted low levels of IL-6 and IL-8 into the culture medium at 48 h after sham infection, whereas the RSV infected cells increased production of these cytokines by 5- to10-fold. Cultures of NHBE that were treated with RSV and POPG showed dramatically suppressed levels of both IL-6 and IL-8 that were not different from those in uninfected cells. The full range of variability among different preparations of NHBE cells is shown in Fig. S1. Nearly identical findings were also obtained with the BEAS2B cell line (Fig. S2). Control experiments monitoring epithelial cell protein synthesis and retention of responsiveness to the TLR5 ligand, flagellin, revealed that POPG treatment was not pleiotropically inhibitory to metabolism or signaling processes (Fig. S3). These findings clearly demonstrate that POPG is a potent suppressor of the inflammatory cytokine production, which accompanies viral challenge.

POPG Prevents Cell Death Caused by RSV Infection in Bronchial Epithelium.

RSV infection induces syncytia formation and cell death in epithelial cells, and we next examined the ability of POPG to interfere with the cytopathic effects (CPE) of the virus. At 48 h after infection, RSV produced prominent contraction of the cells within monolayers of NHBE cells, and numerous synctytia became apparent. None of this CPE was evident in either control cultures or infected cultures treated with POPG. At 72 h after viral infection, the cultures treated with RSV contained predominantly lysed cells, large syncytia, and scattered islands of cells adherent to the substratum as shown in the micrographs in Fig. 2. In contrast, the RSV infected cultures treated with POPG showed a healthy cell monolayer that was indistinguishable from cultures of uninfected cells. POPG treatment alone produced no significant alteration in the appearance of the cells. From these findings, we conclude that POPG can play a major role in protecting the primary NHBE cells from infection by RSV and subsequent cell death. POPG also protected BEAS2B cells from RSV infection (Fig. S4).

POPG Specifically Attenuates Proinflammatory Cytokine Production Induced by RSV.

Previous studies have provided evidence that the antagonism of LPS activation through LBP/CD14/MD2/TLR4 is dependent upon the molecular class and species of phospholipid used as an antagonist (17–20). We examined the molecular specificity of the POPG effect on RSV-elicited cytokine production by comparing the effects of POPG with a phospholipid counterpart containing a choline polar moiety, palmitoyl-oleoyl-phosphatidylcholine (POPC). These data are presented in Fig. 1 and show that POPC fails to attenuate IL-6 and IL-8 production by NHBE. In addition, POPC is ineffective at suppressing the CPE induced by RSV in NHBE cells as shown in Fig. 2. Thus the actions we observe for POPG are specific to this class of phospholipid, as compared with the most abundant class of phospholipid present in pulmonary surfactant. This result parallels the actions of POPG and POPC upon the CD14/MD2/TLR4 complex (17).

RSV Binds CD14 AND POPG.

CD14 and TLR4 have been implicated in regulation of the innate immune response to RSV, but the mechanism of this process is not clearly understood (16). We undertook experiments to examine the binding interactions between the virus, CD14, and POPG. The data presented in Fig. 3A demonstrate direct physical interactions between RSV and CD14. This binding interaction is concentration dependent and saturable. The addition of POPG inhibits the RSV-CD14 interaction in a concentration dependent reaction as shown in Fig. 3B. The maximum inhibition of binding occurs at ≈200 μg/mL POPG. An examination of the direct physical interaction between RSV and POPG is shown in Fig. 3C. The virus binds to POPG in a concentration-dependent and saturable reaction. In contrast to POPG, the lipid POPC is a weak-binding ligand for the virus that interacts with RSV nonspecifically. The consequences of the interactions of POPG with RSV upon epithelial cell responses are shown in Fig. 3D, and demonstrate the lipid concentration dependent inhibition of IL-8 production from BEAS2B cells elicited by the virus. The apparent IC50 of the POPG upon IL-8 production is ≈50 μg/mL, and this value correlates well with the apparent IC50 of the lipid for disrupting viral binding to CD14 (Fig. 3B). From these data, we conclude that one important mode of action of POPG is its direct interaction with the virus, which can interfere with viral recognition of specific cell surface ligands required for viral infection and induction of inflammation.

We next examined whether POPG could effectively inhibit the direct binding of RSV to cell surfaces. For these experiments, HEp2 cells in suspension were used as targets for viral binding. HEp2 cells are routinely used in quantitative plaque assays for RSV (23). After the binding reaction, the cells were analyzed for attached virus by FACScan, using fluorescent antibody. The results presented in Fig. 3E demonstrate significant binding of the RSV to the cell surface in the absence of lipid as detected using FACScan. The addition of POPG markedly inhibits the association of the RSV with the cell surface. In contrast, POPC does not significantly alter the viral attachment to the cell surface. The data in Fig. 3F summarizes the findings from three independent experiments, with cell surface binding expressed as the mean fluorescence intensity. The data clearly demonstrate that 200 μg/mL POPG directly inhibits the binding of the virus to the cell surface by more than 80%, even at a viral multiplicity of infection of 50.

POPG Prevents Plaque Progression and Viral Spreading After an Infection Is Established.

The preceding findings provide strong evidence that POPG can effectively inhibit RSV infection and the subsequent inflammation. We next examined the efficacy of POPG as an antiviral agent after the establishment of a viral infection. In these experiments, monolayers of HEp2 cells were infected with RSV for 2 h to enable viral attachment and internalization. Subsequently, unattached viruses were removed, and the cells were overlayed with agar under standard conditions for a plaque assay (23), except that the agar medium composition was modified to contain either no lipid supplements, or 200–500 μg/mL POPG, or 500 μg/mL POPC. After 6 days of culture, viral plaques were quantified using Neutral Red staining as shown in Fig. 4A. Initially, cursory inspection of the Neutral Red stained plates showed the expected plaque number on RSV-infected plates, but no plaques on plates treated with RSV plus POPG. The monolayers exposed to RSV and POPC produced the same number of plaques as those treated with virus alone. However, careful inspection of the Neutral Red–stained plates from cultures treated with RSV plus POPG, revealed the presence of unusual appearing regions of the monolayer, which we refer to as indefinite plaques. In subsequent experiments, staining with anti-RSV antibody revealed that the usual complete plaques had an expected bullseye appearance, created by lysed cells surrounded by a perimeter of cells darkly staining for viral antigen, whereas the indefinite plaques appeared as minute foci of viral infection as shown in Fig. 4B. Higher magnification views of the different plaque types are shown in Fig. 4C. Quantification of plaques demonstrated that POPG treatment eliminated nearly all of the complete plaques, and produced indefinite plaques at 25–50% of the frequency of normal plaques as shown in Fig. 4D. These findings clearly show that POPG can act to suppress viral spreading in epithelial monolayers after an RSV infection has been established.

In additional experiments, we also quantified the suppression of plaque formation in vitro by treating HEp2 cells with virus in the presence or absence of 200 μg/mL POPG. The infection of HEp2 cells with RSV alone produced 3.2 ± 0.6 × 10⁷ plaques. When the same viral inoculum was administered in the presence of POPG, only 3.1 ± 0.1 × 10³ plaques were obtained. Thus, POPG effectively prevents the in vitro infection of epithelial cells by 4 log units.

Intranasal Administration of POPG Inhibits RSV Infection in Mice.

The potency of POPG as an anti-RSV agent was next examined using a mouse model of viral infection. BALB/c mice were inoculated with RSV intranasally, either in the absence, or presence of 75–150 μg of POPG. At 3 and 5 days after the infection, the animals were euthanized and the lungs were harvested and processed for RSV burden, quantified by plaque assay. Comparisons were made between groups that were uinfected, RSV infected, RSV infected plus POPG treated, inactivated RSV infected, and POPG treated alone, groups. The data obtained after 5 days of viral infection are presented in Fig. 5A. The results provide clear evidence that POPG suppresses in vivo infection in the mouse by a factor of 1,700, when compared with animals treated with RSV alone (RSV infection = 5355 ± 693 plaques; RSV infection + POPG = 3.1 ± 1.2 plaques). No plaques were obtained from uninfected animals, or animals treated with UV-inactivated RSV or POPG alone. Similar results were obtained after 3 days of infection, except the number of viral plaques produced with RSV treatment (822 ± 115 plaques per left lung; n = 3) were significantly lower than those obtained after 5 days of infection. Figure 5B shows the effects of viral infection upon the influx of inflammatory cells into the lung, recovered in the BALF. Control mice contain low levels of lymphocytes and neutrophils in lavage but their percentage increases 3- to 20-fold after viral infection. In addition, the total cells recovered from the lavage of infected animals is increased an average of 2.1-fold. The treatment of RSV-infected animals with POPG completely attenuates the increase in lymphocytes and neutrophils recovered in lavage, elicited by virus. With RSV infection, IFN-γ levels increase from undetectable levels to nearly 8 ng/mL of lavage, at 5 days after infection; and this antiviral response is completely abrogated by treatment with POPG (Fig. 5C). SP-D levels also increase 3.5-fold in response to RSV infection, and this response is also completely inhibited by POPG treatment as shown in Fig. 5D. The effects of RSV and POPG upon lung histopathology were also evaluated and are presented in Fig. 6. The histopathology scores are shown in Fig. 6A. Control mice had a pathology score of 2.7 ± 0.4, and RSV-infected mice had a score of 8.6 ± 0.6. Treatment of mice with RSV and POPG resulted in a pathology score of 3.4 ± 0.5, which is significantly different from RSV infection, but not different from control mice or animals treated with POPG alone. Representative tissue sections are shown in Fig. 6B and provide evidence that POPG also suppresses the appearance of inflammatory cellular infiltrates and pneumonia, which accompanies the viral infection. Microscopy also demonstrates that POPG installation into the lungs does not significantly alter the appearance or inflammatory status of the tissue.

Collectively, the data presented in Figs. 5 and 6 demonstrate that POPG can block RSV infection and replication in vivo, and protect the lungs from inflammatory cell infiltration and inflammatory cytokine production. These findings suggest that POPG may have significant potential for preventing RSV infections in vulnerable human populations, and treating infections after they become established.

Virus Preparation and Titration.

The human RSV A2 strain VR-1540 was obtained from American Tissue Culture Collection. Virus stocks were prepared under endotoxin-free conditions. The virus was grown in HEp2 cell monolayers using Dulbecco’s Modified Eagle’s Medium/Ham’s F-12 Medium (DMEM/F12: GIBCO) plus 5% growth bovine serum (GBS, HyClone) and purified using methods previously described (28, 29). Viral titers and growth were determined by quantitative plaque assays (23). RSV replication and plaque formation were also determined by immunostaining for viral antigens using HRP-conjugated polyclonal goat anti-human RSV antibody (AbD Serotec).

Bronchial Epithelial Cell Tissue Culture, Infection, and Surfactant Phospholipid Treatments.

Beas2B cells were obtained from ATCC and cultured using LHC9 medium (Lonza). Normal human primary bronchial epithelial cells (NHBE) were obtained by bronchoscopy and airway brushing of normal, nonsmoking volunteers with no history of lung disease or current medications (30, 31), or from surgical specimens of tracheas and mainstem bronchi, as previously described (31). All human tissue was acquired under protocols approved by the National Jewish Health Institutional Review Board. Cell culture and characterization was performed as described previously (31). NHBE cells were seeded into 24-well plates and grown until 80% confluent, before exposure to RSV. Cells were infected at a viral multiplicity of 1.5–2.0, and incubated for 72 h. Phospholipids were obtained from AVANTI and liposomes were prepared as previously described (17). For examination of phospholipid inhibition of RSV infection, the cells were preincubated with POPG and POPC liposomes (200 μg/mL) for 1 h and subsequently infected with virus in the presence of liposomes for up to 72 h. IL-6 and IL-8 production in culture supernatants were assayed by ELISA (eBioscience, Biosource). Previous studies have demonstrated that NHBE cells and BEAS2B cell are positive for expression of TLR4, CD14, and MD2 by either Western blotting or PCR (32–34); and HEp2 cells are positive for CD14 expression (35, 36).

Binding of RSV to CD14 and Phospholipid.

RSV viral stocks were diluted into 50 μl PBS (PBS, pH 7.4) and coated overnight in 96-well, half-area plates (Corning) at 4°C. The wells were washed with 20 mM Tris buffer (pH 7.4) containing 150 mM NaCl, 5 mM CaCl₂ (buffer A) and blocked with buffer A + 5% BSA (Sigma) for 2 h, at 37°C. Recombinant human CD14 protein (R&D Systems, 1 μg/mL) was added and incubated at 37°Cfor 1 h. Subsequently the wells were washed with buffer A and labeled with monoclonal anti-human CD14 antibody (R&D Systems) diluted 1:1,500 in buffer A + 5% BSA at 37°C for 1 h, followed by the incubation with HRP-labeled anti-mouse IgG (1:2,000) for 1 h. After washing the wells with buffer A, the bound antibody was detected with o-phenylenediamine as substrate and measuring absorbance at 490 nm. In competitive binding reactions, varying concentrations of POPG in buffer A + 5% BSA were added to wells immediately before the addition of CD14 (1 μg/mL).

For measurement of direct interactions between RSV and lipids, 1.25 nmol of the specified phospholipids in 20-μl aliquots of ethanol were added to 96-well half-area plates, and the solvent was evaporated using a warm-air blower. Nonspecific binding was blocked with bufferA + 5% BSA, and varying concentrations of RSV were added to the wells and incubated at 37°C for 1 h. After the wells were washed with buffer A, HRP-conjugated anti-human RSV antibody diluted in bufferA + 5% BSA (1:500) was added and incubated for 1 h at 37°C. The amount of bound RSV was detected as described above.

Binding of RSV to HEp2 Cells and Inhibition by Phospholipid.

Adherent HEp2 cells were detached from tissue culture plates by 10-min exposure to PBS containing 10 mM EDTA, at 0°C. HEp2 cells (3 × 10⁵) in suspension were incubated with RSV at a multiplicity of 50/cell for 10 min at 37°C in either the presence or absence of 200 μg/mL phospholipid liposomes. Subsequently, the cells were diluted and washed three times with PBS at 4°C, by centrifugation. The cells were fixed overnight in 1% buffered paraformaldehyde, followed by washing three times with PBS. The fixed cells were labeled for 1 h at 4°C, with monoclonal mouse anti-human RSV antibody (AbD Serotec) diluted 1:100 with PBS + 5% BSA. Unbound primary antibody was removed by washing the wells three times with cold PBS. Next, the cells were incubated with phycoerythrin (PE) conjugated anti-mouse IgG (e-Bioscience) for 1 h at 4°C. The unbound fluorescent antibody was removed by washing the cells using centrifugation at 4°C in PBS. The cell associated fluorescence was determined using FACScan. Typically, 20,000 cells were used to quantify the fluorescence of the HEp2 cells. Graphic analysis of the fluorescence and calculation of mean fluorescence intensity was performed using Cell Quest software.

In Vivo Suppression of RSV Infection by POPG.

Female BALB/c mice, 6 weeks old, were obtained from Jackson Laboratory. The animals were anesthetized by i.p.injection of Avertin at a dose of 0.25 g/kg. The viral inoculation volume for each group was adjusted to 50 μL. Sham infections were performed using the same amount of medium plus PBS. We diluted virus with PBS in a total volume of 50 μl that contained 1 × 10⁷ plaque forming units. Liposome POPG was likewise diluted in PBS, and each mouse received 75–150 μg of the lipid. RSV and POPG mixtures were prepared in PBS and inoculated intranasally into mice. Aliquots of RSV stocks were inactivated by exposure to 1,800 mJ of UV radiation in a Stratalinker UV cross-linker (UVP), for the purpose of eliminating viral infectivity without significantly altering the conformation of viral proteins, or inactivating endogenous mediators, as previously described (37). UV inactivation reduced RSV infectivity by 8 log units. At the end of the experiment, mice were euthanized by i.p.injection of 0.25 mL of Nembutal (10 mg/mL). Animal experiments followed all prescribed guidelines and were approved by the Institutional Animal Care and Use committee.

Measurement of Viral Infection and Inflammation.

To quantify viral titers in mice, we homogenized the left lungs in DMEM/F12 at 0°C, shortly after removal from the animals. Freshly prepared lung extracts were centrifuged at 2,000 × g for 15 min at 4°C, and serial dilutions of the supernatants were assayed for viral plaque formation.

For histopathology, lungs were fixed overnight in 10% formalin, washed with PBS, and subsequently dehydrated in 70% ethanol for 48 h. The lungs were paraffin embedded, and 4-μm sections were prepared and stained with H&E. The tissue was evaluated in a double-blinded fashion using light microscopy and a histopathologic inflammatory scoring system, as described previously (38). We collected bronchoalveolar lavage fluid (BALF) by intratracheal instillation of the lungs with 1.0 ml of saline, followed by recovery of the solution. Total recovery was typically 0.7–0.9 mL, and the cells were sedimented at 5,000 × g for 15 min at 4°C. The resultant cell pellet was resuspended in 150 μL of PBS, stained with Trypan Blue, and the total cell number was quantified using light microscopy. Differential cell counts were performed on cytospin preparations stained with Trypan Blue. The cell-free supernatants resulting from the centrifugation of BALF were stored at –80°C and used for cytokine measurements. IFN-γ was measured by ELISA using primary and HRP-conjugated secondary antibodies from BD Biosc0iences. SP-D was measured by ELISA using a polyclonal antibody for solid phase capture and an HRP-conjugated polyclonal antibody for detection. Rabbit polyclonal antibodies against mouse SP-D were generated from purified recombinant protein.

Statistical Analysis.

All results are expressed as mean ± SE. ANOVA was used to determine the level of difference between all groups. Groups were compared by unpaired Student’s t test. Differences are considered to be significant at P < 0.05.