Enveloped viruses disable innate immune responses in dendritic cells by direct activation of TAM receptors

Abstract

Upon activation by the ligands Gas6 and Protein S, Tyro3/Axl/Mer (TAM) receptor tyrosine kinases promote phagocytic clearance of apoptotic cells and downregulate immune responses initiated by Toll-like receptors and type I interferons (IFNs). Many enveloped viruses display the phospholipid phosphatidylserine on their membranes, through which they bind Gas6 and Protein S and engage TAM receptors. We find that ligand-coated viruses activate TAM receptors on dendritic cells (DCs), dampen type I IFN signaling, and thereby evade host immunity and promote infection. Upon virus challenge, TAM-deficient DCs display type I IFN responses that are elevated in comparison to wild-type cells. As a consequence, TAM-deficient DCs are relatively resistant to infection by flaviviruses and pseudotyped retroviruses, but infection can be restored with neutralizing type I IFN antibodies. Correspondingly, a TAM kinase inhibitor antagonizes the infection of wild-type DCs. Thus, TAM receptors are engaged by viruses in order to attenuate type I IFN signaling and represent potential therapeutic targets.

Figure 1. Gain- and Loss-of-Function Studies for Evaluating the Impact of TAM Receptors on Virus Infection

(A) Existing model of TAM receptor-ligand facilitation of enveloped virus entry. PtdSer, phosphatidylserine. (B) Immunoblots showing expression of individual TAM receptorsin HEK 293T cells either untransfected (293) or stably transfected with individual TAM receptors. (C) Levels of HIV-1 viral DNA produced 24 hr postinfection of the indicated HEK 293T transformants (from B) by pseudotyped viruses bearing envelope glycoproteins from VSVg, Ebola, Marburg, or MLV-A viruses. Levels shown were normalized to those obtained with untransfected HEK 293T cells for each pseudotype (defined as 100%, dashed blue line). (D) Immunoblots showing constitutive activation of all three types of TAM receptor in HEK 293T cells that ectopically express Axl. First, TAM receptors were immunoprecipitated with the indicated antibodies and subjected to SDS-PAGE before being immunobloted with receptor-specific antibodies (IB: Tyro3, Axl, or Mer) or with a phosphotyrosine-specific antibody (IB: pY). (E) BMDCs generated from WT or specific TAM receptor knockout mice were infected with HIV-1-derived virus pseudotyped with the indicated viral glycoproteins, and the levels of reverse transcribedHIV-1 DNA were measured at 24 hr postinfection. Levels shown were normalized to those seen with WT BMDCs for each pseudotype (100%, dashed blue line). A, Axl; M, Mer; T, Tyro3; KO, single knockout; DKO, double knockout; TKO triple TAM receptor knockout. *p < 0.05; **p < 0.01; ***p < 0.001. Error bars are SEM of samples from three independent experiments. VSVg, EbGP, MARVGP, and MLV-A denote the glycoproteins of vesicular stomatitis, Ebola, Marburg, and amphotropic murine leukemia virus, respectively. See also Figure S1.

Figure 2. TAM Activity Stimulates Postentry Events during Virus Infection

(A) BMDCs prepared from wild-type (WT, blue) or Axl^−/−Mer^−/− (AM DKO, red) mice were infected with VSVg-pseudotyped HIV-1-derived viruses carrying a Vpr-BLAM fusion protein (see the Experimental Procedures) and maintained at either 4°C or 37°C, as indicated, for 4 hr. Then, cells were loaded with the fluorescent β-lactamase (BLAM) substrate CCF2, and BLAM activity was assayed. Error bars are SD of six samples. Results are representative of three independent experiments. (B) Axl-transfected (293 Axl) and kinase-dead Axl-transfected (293 Axl^KD) HEK 293T cells were treated with (+) or without (−) 10 nM recombinant mouse Gas6, and lysates from these cells were immunoprecipitated with anti-HA. (The cDNA expression constructs for both Axl proteins were HA-tagged.) Protein equivalents of the IPs were immunoblotted with anti-pY (top blot) or anti-HA (bottom blot). (C) Wild-type (293 WT, panels 1–3), Axl-transfected (293 Axl, panel 4), and kinase-dead Axl-transfected (293 Axl^KD, panel 5) stable HEK 293T cell lines were infected with WNV at moi = 10, and the presence of internalized WNV was scored by confocal immunofluorescence microscopy for WNV E protein (E16, red, panels 1 and 3–5) at 4 hr postinfection. Panel 2 is a control staining with a monoclonal antibody (CHKV-152) against chikungunya virus. Nuclei of HEK 293T cells are visualized with DAPI (blue). One representative of three independent experiments is shown. (D) Axl-transfected (Axl, blue) and kinase-dead Axl-transfected (293 Axl^KD, red) stable HEK 293T cell lines were infected with WNV at moi = 0.01 and scored for virus replication as plaque-forming units (PFU; see the Experimental Procedures) at 4 and 24 hr postinfection. Error bars are SD of three samples. Results are representative of three independent experiments.

Figure 3. Enveloped Virus Potentiates Ligand-Dependent TAM Receptor Activation

(A) Left, WT BMDCs were incubated with increasing concentrations (0.5–2 nM) of purified human Protein S (Pros1), in either Gla-depleted cell culture medium alone (medium) or along with enveloped VSVg-pseudotyped HIV-1-derived virus that was produced in Gla-depleted medium (VSVg). Cell lysates were prepared for immunoblot 5 min after virus or Pros1 challenge. Mer was specifically immunoprecipitated from these samples, subjected to SDS-PAGE, and immunoblotted with a phosphotyrosine-specific antibody (IB: pY). Immunoblot analysis was used to confirm equal amounts of Mer in each sample (IB: Mer), and the amount of protein in each lysate was assessed by immunoblot for Gapdh (IB: Gapdh). Right, virus potentiation of Mer autophosphorylation in the presence of 1 nM Pros1 for 5 min is seen for the enveloped VSVg-pseudotyped virus, but not with WT MAV-1, a nonenveloped virus. (B) Left, the same experiments as in (A, left) except that WT BMDCs were incubated with 0.25–1 nM of recombinant mouse Gas6 (Gas6), and the tyrosine autophosphorylation of Axl was monitored (IB: pY). Right, Virus potentiation of Axl autophosphorylation in the presence of 1 nM Gas6 was seen only with the enveloped VSVg-pseudotyped virus and not with the nonenveloped MAV-1. (C) Microtiter wells, precoated with purified BSA, Gas6, or Pros1 (10 μg/ml ON for each protein) and incubated with equivalent amounts of VSVg-HIV (50 ng/ml p24 concentration) at 37°C for 2 hr with or without 5 mM EDTA, and the amount of bound virus was measured by ELISA for p24 protein (see Experimental Procedures). Error bars are SD of three samples. Results are representative of three independent experiments. See also Figure S2.

Figure 4. Enveloped Virus Infection Abrogates the Cellular Antiviral Response in a TAM-Dependent Manner

(A–C) WT BMDCs (blue bars) or TAM TKO BMDCs (red bars) were challenged with the pseudotyped viruses of Figure 1B. Expression levels of IFNα4 (A), IFNβ (B), and SOCS1 (C) mRNAs were measured by qRT-PCR at the indicated time points after virus addition. The levels of the indicated mRNAs were normalized to control β-actin mRNA. *p < 0.05; **p < 0.01; ***p < 0.001. Results comparable to those for IFNα4 and IFNβ mRNAs were seen for measurements of TNFα, IRF-5, and IRF-7 mRNAs (see Figure S3). Error bars are SEM of samples from three independent experiments. See also Figure S3.

Figure 5. The Primary Effect of TAM Receptor-Ligand Interactions on Virus Infection in BMDCs Is to Inhibit the Cellular Antiviral Response

(A) WT BMDCs and TAM TKO BMDCs were incubated with VSVg-pseudotyped virus in either the absence (−) or presence (+) of 100 μg each of neutralizing anti-mouse IFNα and IFNβ antibodies, and the levels of HIV-1 DNA were measured at 24 hr postinfection. Results were normalized to the level of viral DNA seen with WT BMDCs incubated with no antibody (100%, dashed blue line). ***p < 0.001. Error bars are SEM of samples from three independent experiments. (B) WT BMDCs were preincubated for 30 min with or without BMS-777607 and stimulated with Gas6 for 10 min. Receptor activation was monitored with the immunoprecipitation and immunoblotting protocol shown in Figure 3; this protocol employed a phosphotyrosine-specific antibody (pY). Immunoblot was used to confirm equal amounts of Mer and Axl in each sample. (C) WT BMDCs and TAM TKO BMDCs were incubated with EbGP-pseudotyped virus in the absence (−) or presence (+) of 300 nM BMS-777607, and the levels of HIV-1 viral DNA were measured at 24 hr postinfection. Results were normalized to those seen with WT BMDCs (100%, dashed blue line). **p < 0.01. Error bars are SEM of samples from three independent experiments. (D) WT BMDCs (blue bars), AM DKO BMDCs (red bars), and IFNAR KO BMDCs (green bars) were incubated with WNV in the absence (−) or presence (+) of 1 μM BMS-777607, and the levels of infectious virus (PFU per ml) were measured at 24 hr postinfection. ***p < 0.001. Error bars are SEM of three independent samples. Results are representative of three independent experiments. BMS-777607 effects are not due to cytotoxicity (see Figure S4).

Figure 6. Virus-Ligand Activation TAM of TAM Receptors Inhibits the Antiviral Immune Response

A model showing that the major effect of TAM receptor-ligand interactions on enveloped virus infection of dendritic cells is at the level of inhibiting the cellular innate immune response, including TAM inhibition of type I IFN signaling.