Experimental Anti-Inflammatory Drug Semapimod Inhibits TLR Signaling by Targeting the TLR Chaperone gp96

Abstract

Semapimod, a tetravalent guanylhydrazone, suppresses inflammatory cytokine production and has potential in a variety of inflammatory and autoimmune disorders. The mechanism of action of Semapimod is not well understood. In this study, we demonstrate that in rat IEC-6 intestinal epithelioid cells, Semapimod inhibits activation of p38 MAPK and NF-κB and induction of cyclooxygenase-2 by TLR ligands, but not by IL-1β or stresses. Semapimod inhibits TLR4 signaling (IC50 ≈0.3 μmol) and acts by desensitizing cells to LPS; it fails to block responses to LPS concentrations of ≥5 μg/ml. Inhibition of TLR signaling by Semapimod is almost instantaneous: the drug is effective when applied simultaneously with LPS. Semapimod blocks cell-surface recruitment of the MyD88 adapter, one of the earliest events in TLR signaling. gp96, the endoplasmic reticulum-localized chaperone of the HSP90 family critically involved in the biogenesis of TLRs, was identified as a target of Semapimod using ATP-desthiobiotin pulldown and mass spectroscopy. Semapimod inhibits ATP-binding and ATPase activities of gp96 in vitro (IC50 ≈0.2-0.4 μmol). On prolonged exposure, Semapimod causes accumulation of TLR4 and TLR9 in perinuclear space, consistent with endoplasmic reticulum retention, an anticipated consequence of impaired gp96 chaperone function. Our data indicate that Semapimod desensitizes TLR signaling via its effect on the TLR chaperone gp96. Fast inhibition by Semapimod is consistent with gp96 participating in high-affinity sensing of TLR ligands in addition to its role as a TLR chaperone.

FIGURE 1

Effects of Semapimod on pro-inflammatory responses in IEC-6 cells. (A) Activating phosphorylation of p38 MAPK and MKK3/6 MAPKK following 15 min pre-treatment with 10 μM Semapimod and 15 min treatment with 100 ng/ml LPS or 1 ng/ml IL-1β, as indicated. (B) Levels of iNOS, COX-2 and β-actin proteins following 15 min pre-treatment with 10 μM Semapimod and 12 h treatment with 100 ng/ml LPS or 1 ng/ml IL-1β, as indicated. (C) Levels of COX-2 protein following 15 min pre-treatment with (solid boxes) or without (open boxes) 10 μM Semapimod and 12 h treatment with: 1 ng/ml IL-1β; 100 ng/ml LPS; osmotic stress (Osm) with 0.5 M glycerol; oxidative stress with 200 μM H₂O₂ or 20 μM peroxynitrite (PN); protein misfolding stress with 3 μM proteasome inhibitor MG132. Ctrl, control, untreated cells. *p<0.01. COX-2 levels are relative to β-actin levels. (D) Activating phosphorylation of p38 MAPK after pre-treatment with or without 10 μM Semapimod for 15 min and treatment for 15 min with 100 ng/ml LPS, or 1 μg/ml Pam₃CSK₄, or 1 μg/ml CpG DNA, as indicated. (E) Levels of cell death following 24 h incubation of IEC-6 cells with or without 10 mM Semapimod as indicated. (F) Activating phosphorylation of p38 MAPK after pre-treatment with 10 μM Semapimod and treatment with LPS or indicated concentrations of flagellin. All data are representative of at least three independent experiments.

FIGURE 2

Semapimod blocks activation of NF-kB by a group of TLR ligands. (A, B) Levels of IkBα in cells pre-treated with or without 10 μM Semapimod and treated with LPS or IL-1β as indicated. (C) Levels of IkBα after pre-treatment with 10 μM Semapimod and treatment with 100 ng/ml LPS, or 1 μg/ml CpG DNA, or 1 μg/ml Pam₃CSK₄, as indicated. (D) Levels of IkBα following pre-treatment with or without Semapimod and treatment with 100 ng/ml LPS or indicated concentrations of flagellin. Results are representative of at least three independent experiments.

FIGURE 3

Semapimod desensitizes responses to LPS. (A) Phospho-p38/p38 ratio as function of LPS concentration in the presence of 0, 0.2, 0.5, and 10 μM Semapimod. (B) Phospho-p38/p38 ratio following treatment with indicated concentrations of LPS in μg/ml as function of Semapimod concentration. Values are percentages of full activation (100 ng/ml LPS for 15 min) in the absence of Semapimod. SD of the individual points are 20% or less. (C) Time course of inhibition by 2 μM Semapimod of IkBα degradation induced by 15 min treatment with 100 ng/ml LPS. Time of addition of Semapimod is relative to the beginning of LPS treatment.

FIGURE 4

Semapimod blocks LPS-induced recruitment of MyD88 to cell surface. Top: Anti-MyD88 immunostaining following 10 min treatment with 100 ng/ml LPS, with or without 15 min pre-treatment with 2 μM Semapimod as indicated. Arrowheads indicate localization of MyD88 to the cell surface. Bar=5 μm. NS, normal rabbit serum. Similar results were obtained in 4 independent experiments. Bottom: Ratios of surface to sub-surface MyD88 signal in cells treated with or without LPS and Semapimod, as indicated. n=40 in each treatment group. *, significant difference from other groupss, p<0.0001.

FIGURE 5

Semapimod abrogates modification of gp96 with ATP-desthiobiotin. (A) Silver-stained gel of IEC-6 proteins modified by ATP-desthiobiotin and collected on streptavidin-agarose. Cells were pre-treated with 10 μM Semapimod for 15 min and then treated with LPS for 15 min as indicated. The prominent protein band that is present or absent depending on Semapimod treatment is indicated by arrowhead. (B) Immunoblot analysis of IEC-6 proteins modified by ATP-desthiobiotin. Cells were treated with or without Semapimod, and ATP-desthiobiotin-modified proteins were analyzed by Western blotting with anti-gp96, anti-HSP90, and anti-β-actin Abs. β-actin is pulled down by ATP-desthiobiotin because it is an ATP-binding protein. M, marker lanes; positions of protein size markers are indicated on the left. Data are representative of at least three independent experiments.

FIGURE 6

Semapimod inhibits ATP-binding and ATPase activities of gp96. (A) Effect of Semapimod on ATP-binding activity of gp96. (B) Effect of Semapimod on ATPase activity of gp96. Inset, effect of Semapimod on ATPase activity of HSP90. Values are percentages of activities in the absence of Semapimod. Error bars show SD. The data are from the four independent assays.

FIGURE 7

Comparisons between Semapimod and other gp96 inhibitors. (A) IkBα and phospho-p38 levels after pre-treatment with 8 μM geldanamycin, 8 μM radicicol, or 20 μM NECA and treatment with 100 ng/ml LPS as indicated. (B) TLR4 and TLR9 immunofluorescence in SW480 cells before (left) and after (right) 3 h treatment with 2 μM Semapimod. Bar=5 μM. NS, normal rabbit serum. (C) Co-immunoprecipitation of FLAG-tagged TLR4 and gp96 from lysates of HEK293 cells transiently expressing FLAG-TLR4, with or without 10 μM Semapimod. Cells were treated as indicated with Semapimod for 2 h prior to lysis, and the drug was also added to cell lysates. All data are representative of at least three independent experiments.

FIGURE 8

Model of Semapimod effects on TLR receptor complexes. Semapimod rapidly shifts cell surface TLR complexes into the low affinity state by inhibiting cell surface gp96. In addition, by inhibiting intracellular gp96 (dashed arrow), Semapimod blocks TLR trafficking from the ER to the cell surface.