Phospholipid regulation of innate immunity and respiratory viral infection

Abstract

Toll-like receptors (TLRs) coupled to intracellular signaling cascades function as central elements of innate immunity that control transcription of numerous pro-inflammatory genes. Two minor anionic phospholipids present in the pulmonary surfactant complex, palmitoyl-oleoyl-phosphatidylglycerol (POPG) and phosphatidylinositol (PI), antagonize the cognate ligand activation of TLRs 2 and 4. The lipids block recognition of activating ligands by the TLRs, either directly or via the TLR4 coreceptors CD14 and MD2. Antagonism of TLR activation results in inhibition of the initiating step of the pro-inflammatory signaling pathways. Evidence for this mechanism of action comes from direct binding studies between CD14 and MD2 with POPG and PI. Additional evidence for this mechanism of antagonism also comes from monitoring the reduction of protein phosphorylation events that characterize the intracellular signaling by activated TLRs. The pathogenesis of respiratory syncytial virus (RSV) and influenza A virus (IAV) have been linked to TLR4 activation, and we examined the action of POPG and PI as potential antagonists of the pathology of these viruses. Surprisingly, POPG and PI dramatically curtail infection, in addition to inhibiting inflammatory sequelae associated with RSV and IAV infections. The mechanism of action by the lipids is disruption of virus particle binding to host cell plasma membrane receptors, required for viral uptake. The antagonism of activation of TLRs and virus binding to the alveolar epithelium by resident constituents of the pulmonary surfactant system suggests that POPG and PI function in homeostasis, to prevent inflammatory processes that result in reductions in gas exchange within the alveolar compartment.

Keywords: Toll-like receptors; antiviral agent; antivirals; immunology; inflammation; innate immunity; phospholipids; pulmonary surfactant; respiratory system; signaling; toll receptor; viral entry; virology.

Figure 1.

Overview of the actions of anionic phospholipids as antagonists of TLR4 and TLR2 activation and intracellular signaling. The pathways shown are for antagonism by POPG, and the data are nearly identical for PI. TLR4 activation requires recognition of LPS by the receptor accessory proteins, CD14, and MD2. POPG can disrupt LPS-dependent activation of the CD14-MD2-TLR4 complex in the plasma membrane, and the mechanism of action appears to be inhibition of MD2 and CD14 recognition of the LPS ligand. This inhibition prevents the conformational change in intracellular domains of TLR4 dimers that are required for engagement of intracellular signaling processes. The signaling cascades include activation of IRAKs 1, 2, and 4, phosphorylation-dependent inactivation of IκBα, release of NF-κB from the complex with IκBα, nuclear import of NF-κB, initiation of transcription of pro-inflammatory genes (e.g. TNFα, IL-6, and IL-8, among others), and new transcription of the gene for MKP1. A parallel arm of signaling cascades uses phosphorylation of MAP kinases (ERKs, JNK, and p38) and subsequent activation of transcription factor AP-1, which results in new transcription of cyclooxygenase 2 (COX2). TLR2 activation follows the same general scheme with Pam3Cys activating TLR2/1 heterodimers and MALP2 activating TLR2/6 heterodimers. For the TLR2 pathways, POPG and PI appear to block ligand recognition by the TLR heterodimers.

Figure 2.

POPG and PI disrupt intracellular signaling from TLR4 and TLR2 and the downstream production and secretion of TNFα. A, immunoblot data for activation of p38, ERK, and JNK by phosphorylation and inactivation by phosphorylation of IκBα, and new synthesis of MKP-1. LPS (10 ng/ml) and POPG (200 μg/ml) were added to cultured U937 macrophages for 15–60 min, as indicated. Cell extracts were prepared at the indicated time points and subjected to solubilization with SDS-PAGE buffer and subsequently electrophoresed, transferred to nitrocellulose, and immunoblotted with antibodies recognizing the proteins, p38, ERK, and IκBα, and their phosphorylated variants, or MKP-1. B, the effects of LPS (10 ng/ml) and 200 μg/ml lipids (DPPC, dipalmitoylphosphatidylglycerol (DPPG), dimyristoyl phosphatidylglycerol (DMPG), POPG, and PI) upon expression and secretion of TNFα in response to LPS challenge were measured by ELISA. A, Data are from Ref. and are from 1 of 3 experiments. B, data are from three independent experiments, with duplicate samples in each experiment. Values shown are mean ± S.D. Significance: * corresponds to p <0.05. C, direct concentration-dependent binding reactions between soluble human CD14 and solid phase POPG and POPC. D, direct concentration-dependent binding between human MD2 and solid phase POPG and POPC. Error bars in panels C and D correspond to mean ± S.E. for 3 experiments.

Figure 3.

POPG and PI prevent RSV infection in mice. Mice were either uninfected (CONL) or inoculated intranasally with 10⁷ pfu of RSV (in 50 μl PBS) in either the absence (RSV) or presence of 150 μg of POPG (RSV + POPG) or 600 μg of PI (RSV + PI). On day 5 after the infection, mice were euthanized and the lungs lavaged and harvested. A, the left lung was fixed, stained, and processed for microscopy and scored for histopathology and the results are shown. B, representative micrographs of tissue sections obtained from the different experimental groups. C and D, quantification of viral plaque numbers in lung extracts prepared from different experimental groups (either RSV or RSV+ POPG or RSV + PI). Viral burdens in the lungs were determined after homogenizing the right lung and performing quantitative plaque assays with HEp2 cell monolayers as the target for infection. Staining of monolayers was performed overnight using neutral red. Plaques typically appear as zones of unambiguous clearing over a background of the stained cell layer. Data shown are mean ± S.D., and the asterisk indicates p <10⁻⁹ for RSV + PG or p <10⁻⁶ for RSV + PI, when compared with RSV alone.

Figure 4.

POPG and PI prevent influenza A infection in mice. Mice were inoculated with 100 pfu of H1N1-IAV in either the absence or presence of 3 mg of POPG or 600 μg of PI. At 6 days after infection, the mice were euthanized and the lungs harvested. Viral burden was determined from homogenates of the lungs using quantitative plaque assays with monolayers of MDCK cells as the targets. A, the histology of lung sections for uninfected animals (CONL) or those receiving virus alone (IAV day 3 or IAV day 6) or virus plus lipids (IAV + PG, day 3 and day 6). Routinely, these experiments also include an inactive lipid (IAV + POPC), which consistently is not different from treatment with IAV alone. Cultures were harvested at 3 days after infection. B, the results of the plaque assays in scatter plots. Values shown are mean ± S.D. from three independent experiments with duplicate determinations in each experiment. Significance: * corresponds to p <10⁻¹⁰ for PG and p <10⁻⁶ for PI, when compared with H1N1-IAV alone.