Themis2/ICB1 is a signaling scaffold that selectively regulates macrophage Toll-like receptor signaling and cytokine production

Abstract

Background: Thymocyte expressed molecule involved in selection 1 (Themis1, SwissProt accession number Q8BGW0) is the recently characterised founder member of a novel family of proteins. A second member of this family, Themis2 (Q91YX0), also known as ICB1 (Induced on contact with basement membrane 1), remains unreported at the protein level despite microarray and EST databases reporting Themis2 mRNA expression in B cells and macrophages.

Methodology/principal findings: Here we characterise Themis2 protein for the first time and show that it acts as a macrophage signalling scaffold, exerting a receptor-, mediator- and signalling pathway-specific effect on TLR responses in RAW 264.7 macrophages. Themis2 over-expression enhanced the LPS-induced production of TNF but not IL-6 or Cox-2, nor TNF production induced by ligands for TLR2 (PAM3) or TLR3 (poly IratioC). Moreover, LPS-induced activation of the MAP kinases ERK and p38 was enhanced in cells over-expressing Themis2 whereas the activation of JNK, IRF3 or NF-kappaB p65, was unaffected. Depletion of Themis2 protein by RNA inteference inhibited LPS-induced TNF production in primary human macrophages demonstrating a requirement for Themis2 in this event. Themis2 was inducibly tyrosine phosphorylated upon LPS challenge and interacted with Lyn kinase (P25911), the Rho guanine nucleotide exchange factor, Vav (P27870), and the adaptor protein Grb2 (Q60631). Mutation of either tyrosine 660 or a proline-rich sequence (PPPRPPK) simultaneously interrupted this complex and reduced by approximately 50% the capacity of Themis2 to promote LPS-induced TNF production. Finally, Themis2 protein expression was induced during macrophage development from murine bone marrow precursors and was regulated by inflammatory stimuli both in vitro and in vivo.

Conclusions/significance: We hypothesise that Themis2 may constitute a novel, physiological control point in macrophage inflammatory responses.

Figure 1. Themis2 is inducibly tyrosine phosphorylated and interacts with Lyn kinase.

Panel a. RAW cells were challenged with LPS (10 ng/ml) for the indicated period. Phosphotyrosine-containing proteins were recovered with 4G10-agarose, and western blotted with anti-Themis2 antibodies. Panels b,c. Anti-Flag immunoprecipitates of LPS-stimulated RAW cells stably expressing Flag-tagged Themis2 were western blotted with anti-Flag-HRP and anti-Lyn kinase. Data shown is representative of three similar experiments.

Figure 2. A Y660F mutation of Themis2 inhibits its LPS-induced tyrosine phosphorylation and Lyn interaction.

Panel a. Flag-tagged Themis2, or a Y660F mutant, was immunoprecipitated from LPS-stimulated (10 ng/ml, 0–120 min) stable RAW cell transfectants. For each time point the relative number of cell equivalents of immunoprecipitated material analysed is indicated. Flag immunoprecipitates were western blotted for phosphotyrosine with 4G10-HRP, stripped and reprobed with anti-Flag-HRP. Panel b. Flag-tagged proteins were immunoprecipitated from RAW cells stably expressing Themis2, BAP or the Y660F mutant point mutant of Themis2. For Themis2 and BAP, two independently generated pools of stable transfectants (denoted a or b) were compared. Immunoprecipitates were western blotted for Flag and Lyn kinase. In each panel a representative of three similar experiments is shown.

Figure 3. Vav and Grb2 associate with Themis2.

Panel a. Flag-tagged Themis2 or BAP was immunoprecipitated from matched numbers of stable RAW cell transfectants which had been activated or not with LPS (10 ng/ml, 60 min). Flag-tagged Themis2 and associated Vav or Grb2 were detected by western blot. Vav and Grb2 present in whole cell lysates from Themis2 expressing RAW cells is shown for reference (input). Panel b. Flag-tagged proteins were immunoprecipitated from matched numbers of stable RAW cell transfectants over-expressing wild type Themis2 (Thms2) or its proline (2PA) or tyrosine (Y660F) mutants. Flag-tagged protein and associated Vav and Grb2 were detected by Western blot. In each case a representative of three similar experiments is shown.

Figure 4. Themis2 over-expression modulates LPS-induced p38 and ERK activation.

Parental RAW cells or stable Themis2 transfectants were challenged or not with LPS (10 ng/ml) as indicated. Whole cell extracts were resolved by SDS PAGE and western blotted with antibodies for the phosphorylated MAP kinase indicated then stripped and re-probed with the relevant total MAPK antibody. Panel a depicts a representative of four similar experiments which were analysed by scanning densitometry, the means ± sem of which are shown in panel b. Values were normalised using total antibody data for each protein and expressed relative to the 15 min time point in parental RAW cells.

Figure 5. Themis2 over expression has no effect on LPS-induced nuclear localisation of NF-κB p65 or IRF3.

Nuclear extracts of parental RAW cells or stable Themis2 transfectants were prepared following LPS challenge for the periods indicated (0–60 min, panel a; 0–4 hr, panel b), western blotted as shown then stripped and re-probed for actin. A representative of three similar experiments is shown.

Figure 6. Themis2 regulates TLR-induced cytokine expression.

Panel a. Parental RAW cells or cells stably over-expressing Themis2 or BAP were challenged (o/n) as indicated. TNF production was measured by ELISA. Data represent the mean ± sem of six experiments. * denotes p<0.05. Panel b. Parental or stably transfected RAW cells were challenged (o/n) with LPS and the presence of TNF and IL-6 in supernatants was analysed by ELISA. Data were normalised against values obtained in parental RAW cells and represent the mean ± sem of nine experiments. * denotes p<0.05. Panel c. RAW cells stably over-expressing wild-type Themis2 or the 2PA or Y660F mutants were stimulated (o/n) with LPS (10 ng/ml), PAM3 (10 ng/ml) or poly I∶C (20 µg/ml). TNF in cell supernatants was measured by ELISA and expressed relative to the levels generated by wild type Themis2 transfectants. Data represent the mean ± sem of 6 experiments. * denotes p<0.05, ** denotes p<0.005. Panels d and e. Primary human macrophages were transfected, or not, on days 3 and 5 with the siRNA indicated (100 nM). On day 7, levels of Themis2 protein and actin were measured by western blot (panel d) and LPS-induced generation of TNF was measured by ELISA (panel e). Data represent the mean ± sem of four experiments, * denotes p<0.05. Panel f. Parental RAW cells were nucleofected with constructs (2 µg each) encoding renilla luciferase, TNF-firefly luciferase and either Flag-tagged Themis2 or the empty CMV Flag vector. LPS-induced (10 ng/ml, 4 hrs) levels of firefly luciferase activity were normalised against levels of renilla luciferase in the same triplicate wells then expressed as a percentage of the level detected in empty vector-transfected cells. Data represent the mean ± sem of nine experiments. ** denotes p<0.001.

Figure 7. Regulation of Themis2 expression.

In panels a–d, detergent extracts of matched numbers of cells were western blotted for Themis2 and actin. Data are representative of three similar experiments. Panel a. Whole splenocytes were separated into non-adherent (lymphocyte-enriched, L) and adherent (macrophage-enriched, M) populations; bone marrow cells undifferentiated (BM) or M-CSF-differentiated macrophages (MΦ). Panel b. RAW cells were treated (48 hrs) with or without IFNγ (10 ng/ml), TGFβ (20 ng/ml) or dexamethasone (0.1 µM). The relative levels of expression in this experiment, corrected for actin expression levels, are presented below each lane. A representative of three similar experiments is presented. Panel c. Lymphocyte-enriched (L) or macrophage-enriched (M) splenocyte populations were recovered as above from control mice (con), immunised but disease-free mice (imm) or immunised arthritic mice (arth) 10 days after immunisation with type II collagen. Panels d,e. Mice (5 per time point) were inoculated (50HAU of X31 influenza, nasally) or not, weighed and splenocytes recovered at the indicated time points. Weight (mean±sem) is expressed as a percentage of the value measured at time zero for each animal.