Bifunctional lipocalin ameliorates murine immune complex-induced acute lung injury

Abstract

Molecules that simultaneously inhibit independent or co-dependent proinflammatory pathways may have advantages over conventional monotherapeutics. OmCI is a bifunctional protein derived from blood-feeding ticks that specifically prevents complement (C)-mediated C5 activation and also sequesters leukotriene B4 (LTB4) within an internal binding pocket. Here, we examined the effect of LTB4 binding on OmCI structure and function and investigated the relative importance of C-mediated C5 activation and LTB4 in a mouse model of immune complex-induced acute lung injury (IC-ALI). We describe two crystal structures of bacterially expressed OmCI: one binding a C16 fatty acid and the other binding LTB4 (C20). We show that the C5 and LTB4 binding activities of the molecule are independent of each other and that OmCI is a potent inhibitor of experimental IC-ALI, equally dependent on both C5 inhibition and LTB4 binding for full activity. The data highlight the importance of LTB4 in IC-ALI and activation of C5 by the complement pathway C5 convertase rather than by non-C proteases. The findings suggest that dual inhibition of C5 and LTB4 may be useful for treatment of human immune complex-dependent diseases.

Keywords: Acute Lung Injury; Complement; Immune Complex; Immunotherapy; Inflammation; Leukotriene; Lung Injury; Parasite.

FIGURE 1.

OmCI binds LTB₄. A, bOmCI (5 μg) competes with LTB₄-specific polyclonal Ab for binding to LTB₄ but does not compete for LTC₄ or TXB₂. Negative control was PBS only. Tick lipocalin protein controls were 5 μg of OmCLI (mass, 16.9 kDa) and 5 μg of RaHBP2 (mass, 20.4 kDa). Shown are four replicates per sample from three independent experiments (95% CI). B, detection of LTB₄ binding to OmCI shown by red shift in the characteristic absorbance spectra of the leukotriene. Peak absorbance shifts from 271 nm with shoulders at 261 and 281 nm for LTB₄ only (i.e. molar ratio 0:1) to 277 nm with shoulders at 267 and 287 nm at LTB₄ to bOmCI molar ratios of 1:1 and higher. For clarity, absorbance values measured at 1 nm intervals are shown as smoothed lines. C, LTB₄ binding causes a mobility shift in bOmCI by Coomassie Blue-stained native PAGE. The molar ratio of bOmCI to LTB₄ is shown above each lane.

FIGURE 2.

Structure of bOmCI with and without LTB₄ in the binding pocket. A and B, schematic representations of the bOmCI molecule from the LTB₄ cocrystal. Blue to red, from N to C terminus. The cysteine side chains and LTB₄ ligand are shown in stick representation, with the ligand carbon atoms colored yellow and the oxygen atoms red. Views A and B are rotated forward by 90° around the horizontal axis. βB-βC and βH-α3 loops, which undergo conformational changes, are labeled. For the strand and helix labels see Ref. . C and D, representation of most significant, but still minor, conformational differences between bOmCI-palmitoleic acid and bOmCI-LTB₄. C, the conformation of βB-βC loop-(61–70) changes with ligand size. Green, bOmCI-LTB₄ complex; blue, bOmCI-palmitoleic acid complex; loop-(61–70), which swings out to better accommodate LTB₄, is shown in darker shades in both molecules. The fatty acid ligands and Pro-61, the residue mediating the change in structure and contacting the ligand, are in stick representation. The yOmCI-ricinoleic acid complex is not shown, but its loop aligns with the conformation seen in the bOmCI-palmitoleic acid complex. D, His-117 and the βH-α3 loop (residues 135–141) are coupled. Green, bOmCI-LTB₄ complex; yellow, yOmCI-ricinoleic acid (PDB accession code 1CM4). The βH-α3 loop (residues 135–141) is in darker shades. His-117, in which the side chain conformation accompanies the loop rearrangement, is represented in sticks.

FIGURE 3.

Details of ligand binding. A, bOmCI-LTB₄ 1.9 Å F_o − F_c residual electron density contoured at the +1.0 σ level, is computed without modeling of the ligand. The final model for the LTB₄ ligand is shown with carbon (yellow) and oxygen (red) surrounded by the OmCI pocket residues (carbon, green; oxygen, red; nitrogen, blue; and sulfur, yellow); this picture was produced with the program PyMOL (44). B, schematic representation of the bOmCI residues forming hydrogen bonds (with distances in Å) and non-bonding hydrophobic contacts to LTB₄ (upper panel) and palmitoleic acid (lower panel). These pictures were produced with the program LigPlot (45).

FIGURE 4.

C5 and LTB₄ binding activities of OmCI are independent. A, LTB₄ (2-fold molar excess) has no effect on the classical C pathway inhibitory activity of bOmCI assayed by hemolysis. Bars show 95% CI (n = 4 for each data point). B, spectroscopy showing that LTB₄ remains bound (characteristic red-shifted spectrum) to bOmCI in low pH (5.5 and 4), 10 mm sodium acetate buffers. C and D, SPR measurement of binding between bOmCI and human C5 (hC5) in the absence (C) and presence (D) of LTB₄. Curves were fitted using a 1:1 Langmuir model (shown in black). Mean K_D and chi-squared values are shown with their standard deviation calculated from 15 replicates. Replicates were performed over a minimum of two independent experiments. The RUs in D are lower than in C because less bOmCI-LTB₄ than bOmCI coupled to the sensor surface. E and F, binding of radiolabeled LTB₄ by bOmCI alone and in complex with hC5. E, plot showing that ligand binding is saturable. RaHBP2 and hC5 alone do not show saturable binding, and cpm values are equivalent to PBS only (not shown). Raw data shown (n = 2 for each data point) are representative of three independent experiments. The highest concentration of OmCI-hC5 (i.e. 10000 nm) was not assayed, as the amount of hC5 required was prohibitively expensive. F, plot showing interaction of hC5 with bOmCI has no effect on the binding kinetics to LTB₄. The data are representative of two independent experiments (n = 2 for each data point) and show cpm values after subtraction of average cpm (n = 16) of the negative control (RaHBP2). Logarithmic regression line functions are: bOmCI, y = 134.67 ln(x) + 521.6, R² = 0.98; bOmCI-hC5, y = 132.87 ln(x) + 479.2, R² = 0.97.

FIGURE 5.

OmCI inhibits IC-ALI. A, OmCI (100 μg) reduces neutrophils in BALF and lung (myeloperoxidase activity) and prevents hemorrhage, protein and Evan's Blue (EB) leak, and edema. B, dose-response effect of OmCI (50, 100, and 200 μg/mouse) on neutrophil recruitment. All treatment groups except saline (NaCl) were dosed intranasally with chicken anti-OVA IgG with or without OmCI. All groups then received OVA by injection into the tail vein. Lung inflammation was evaluated 4 h after induction of IC. The mean values + S.D. of a representative study are given (n = 6 mice/group).

FIGURE 6.

OmCI (100 μg) inhibits vascular leak, tissue damage, and inflammation in response to IC. A, OmCI attenuates vascular leak as shown by blue macroscopic discoloration due to Evans blue leakage. OmCI+LTB₄ signifies OmCI saturated with LTB₄ before its addition to lung. B, OmCI inhibits endothelial damage, neutrophil recruitment, and hemorrhage into the alveolar space (H&E staining, magnification ×100). C, semiquantitative score of endothelial damage, neutrophil recruitment, and hemorrhage in the lung. Acute inflammation was evaluated at 4 h. Representative data for one of three independent experiments are shown (n = 6 mice/group). Values are shown as mean + S.D.

FIGURE 7.

OmCI (100 μg) neutralizes LTB₄-induced inflammation. A, inhibition of leukotriene biosynthesis by MK886 (1 mg/kg by gavage) reduces neutrophil recruitment induced by IC-ALI. B, intranasal instillation of LTB₄ (1 μg) induces neutrophil recruitment in the lung, which is inhibited by prior administration of 2-fold molar excess bOmCI but not by a similar amount of control tick lipocalin histamine-binding protein RaHBP2 (100 μg). Representative data for one of three independent experiments are shown (n = 6 mice/group). Values are shown as mean + S.D.

FIGURE 8.

OmCI saturated with LTB₄ (bOmCI-LTB₄) is unable to bind additional LTB₄. A, representative absorption spectrum of bOmCI with and without LTB₄ that was used for experiments in mice, crystallization, and SPR. B, unlike bOmCI, in an enzyme immunoassay bOmCI-LTB₄ is unable to compete with LTB₄-specific polyclonal Ab for binding to LTB₄. Negative control was PBS only. A 6-fold excess (30 μg; see legend for Fig. 1A) of each protein, including the histamine-binding protein 2 (RaHBP2), was used for each replicate. Four replicates/sample (95% CI) are shown.

FIGURE 9.

Saturation of OmCI with LTB₄ (OmCI-LTB₄) reduces the inhibitory effect of the lipocalin on IC-ALI. Representative data for one of three independent experiments are shown (n = 6 mice/group). Values are shown as mean + S.D.

FIGURE 10.

OmCI does not inhibit cleavage of C5a from thrombin but does reduce the concentration of C5a in the alveolar space. A, assay of the inhibitory effect of OmCI on cleavage of C5a from hC5 by human thrombin (hT). The average background value (i.e. PBS only, blank wells) was subtracted from all samples. Four replicates/sample (95% CI) are shown. Combined data from two independent experiments are shown. B, C5a concentration in BALF in each treatment group (n = 8 mice/group) measured at 4 h after induction of IC-ALI. The box plot shows median values, upper and lower quartiles, largest and smallest observations for each group, and the single outlier in the OVA group. Significant differences between groups, determined by analysis of variance using Tukey's multiple comparison of means, are indicated by asterisked bars (*, p < 0.02; **, p < 0.005; ***, p < 0.001).

FIGURE 11.

Interactions and feedback mechanisms between OmCI, complement, Fcγ receptors, and LTB₄ in early phase of mouse IC-ALI. A, simplified view of co-dependent proinflammatory pathways focused on alveolar macrophages (AMΦ) in the alveolar space in the absence of OmCI. B, the same pathways as in A with OmCI present. For clarity, inflammatory mediators stimulated by TCC, C5a, and LTB₄ and cellular responses within the lung are not shown. Note that only the LTB₄ receptor BLT1 is shown, as BLT2 expression in mice is restricted to mast cells and keratinocytes. The C5a receptor C5L2 is omitted from the figure, as its actions are subject to debate. Blue arrow, stimulatory pathway; red line, inhibitory pathway. Solid line, full activation of pathway; dotted line, partial activation of pathway.