Leukocyte engagement of fibrin(ogen) via the integrin receptor alphaMbeta2/Mac-1 is critical for host inflammatory response in vivo

Abstract

The leukocyte integrin alpha(M)beta(2)/Mac-1 appears to support the inflammatory response through multiple ligands, but local engagement of fibrin(ogen) may be particularly important for leukocyte function. To define the biological significance of fibrin(ogen)-alpha(M)beta(2) interaction in vivo, gene-targeted mice were generated in which the alpha(M)beta(2)-binding motif within the fibrinogen gamma chain (N(390)RLSIGE(396)) was converted to a series of alanine residues. Mice carrying the Fibgamma(390-396A) allele maintained normal levels of fibrinogen, retained normal clotting function, supported platelet aggregation, and never developed spontaneous hemorrhagic events. However, the mutant fibrinogen failed to support alpha(M)beta(2)-mediated adhesion of primary neutrophils, macrophages, and alpha(M)beta(2)-expressing cell lines. The elimination of the alpha(M)beta(2)-binding motif on fibrin(ogen) severely compromised the inflammatory response in vivo as evidenced by a dramatic impediment in leukocyte clearance of Staphylococcus aureus inoculated into the peritoneal cavity. This defect in bacterial clearance was due not to diminished leukocyte trafficking but rather to a failure to fully implement antimicrobial functions. These studies definitively demonstrate that fibrin(ogen) is a physiologically relevant ligand for alpha(M)beta(2), integrin engagement of fibrin(ogen) is critical to leukocyte function and innate immunity in vivo, and the biological importance of fibrinogen in regulating the inflammatory response can be appreciated outside of any alteration in clotting function.

Figure 1

Modification of the fibrinogen γ chain gene. (A) Overall structure of the Fibγ^390–396A gene-targeting vector, the WT fibrinogen γ chain gene, and the targeted Fibγ^390–396A allele. Exons are depicted as solid areas, and the introduced nucleotide substitutions and PvuII site are indicated in text boxes as 390–396A and PvuII, respectively. The PvuII fragments that were diagnostic for the WT and targeted alleles by Southern blot assays are indicated by thin lines. Arrowheads indicate the position of diagnostic PCR primers used for detecting homologous recombination events in ES cells. Routine animal genotyping was done using primers 1 and 2 followed by PvuII digestion. (B) Partial nucleotide sequence of exons 9 and exon 10 of the WT and γ^390–396A genes. Asterisks indicate nucleotides that were mutated in the Fibγ^390–396A allele. A vertical arrow indicates the position of the exon 9 to 10 splice junction. Amino acids altered by the nucleotide substitutions are in italics. Underlined amino acids indicate the glutamine and lysine that participate in transglutaminase-mediated cross-linking. Amino acids that are known to be critical for platelet integrin receptor α_IIbβ₃ binding (23) are bracketed. (C) Representative PCR analyses to establish animal genotypes using DNA template from ear biopsies of WT, hemizygous, and homozygous mutant Fibγ^390–396A mice. Primers 1 and 2 were used to amplify a 527-bp fragment, which was subsequently digested with PvuII to yield the diagnostic fragments of 304 bp and 223 bp. ddH₂O, double distilled water.

Figure 2

Characterization of Fibγ^390–396A fibrinogen. (A) Western blot of fibrinogen in plasma from WT and mutant mice. (B) Coomassie blue–stained SDS polyacrylamide gel (reducing conditions) showing affinity-purified fibrinogen preparations from WT and Fibγ^390–396A mice. (C) Comparative analysis of thrombin-induced fibrin polymerization in plasma from WT and homozygous Fibγ^390–396A mice. (D) Analysis of fXIIIa-mediated fibrin cross-linking in reaction mixtures containing either purified WT or γ^390–396A fibrinogen. The electrophoretic positions of Aα, Bβ, γ chains are indicated at left along with γ-γ dimer and α polymer cross-linking products.

Figure 3

Fibrinogen γ^390–396A supports normal platelet aggregation. Comparative analysis of ADP-induced platelet aggregation in vitro using platelet-rich plasma prepared from WT and Fibγ^390–396A mice. The fibrinogen dependence of aggregation in vitro was demonstrated with Fibγ^Ø5 mice that express a mutant form of fibrinogen lacking the α_IIbβ₃-binding motif. The decline in transmitted light that immediately follows ADP addition is a consequence of platelet shape change.

Figure 4

Fibrinogen γ^390–396A does not support α_Mβ₂-dependent adhesion of primary neutrophils. Primary human neutrophils (10⁶ cells/ml) were transferred to either uncoated wells, wells coated with 10 ∝g/ml WT mouse fibrinogen (A), wells coated with 10 ∝g/ml fibrinogen γ^390–396A (B), or wells coated with 10 ∝g/ml fibrinogen γ^Ø5 (C). Shown are representative views of bound cells after 25 minutes. (D) Quantitative comparison of neutrophil adhesion to WT, fibrinogen γ^390–396A, and fibrinogen γ^Ø5. The specificity of neutrophil engagement was established by preincubating neutrophils with 20 ∝g/ml of the rat anti-mouse α_M I-domain Ab, M1/70, or control rat anti-mouse IgG for 15 minutes at room temperature before addition to the fibrinogen-coated wells. The data shown are means ± SD. HPF, high power field; Con, control.

Figure 5

Fibrinogen γ^390–396A fails to support α_Mβ₂-dependent adhesion of monocytes, macrophages, or the monocytoid cell line, THP-1. (A) Comparative analysis of primary human blood monocyte adhesion to immobilized WT and γ^390–396A fibrinogen (coating concentration 2 ∝g/ml). Specificity of integrin-mediated binding was established by preincubation of cells with blocking α_M- or β₂-specific mAb’s (44a and IB4, respectively). The nonblocking α_M mAb, OKM1, was used as a control. (B) Comparative analysis of mouse peritoneal macrophage adhesion to WT and γ^390–396A fibrinogen. The blocking Ab M1/70 was used to demonstrate α_Mβ₂-specificity. (C) THP-1 cell adhesion as a function of fibrinogen coating concentration using either WT or γ^390–396A fibrinogen. (D) Adhesion of THP-1 cells to WT or γ^390–396A fibrinogen in the presence of either EDTA or α_Mβ₂ Ab, M1/70. The data shown are means ± SD. fl, fluorescence.

Figure 6

Cell adhesion to fibrinogen is supported by integrin α_Mβ₂, but not the related integrin, α_Lβ₂. (A) Adhesion of HEK293 transfectants expressing either α_Mβ₂, α_Lβ₂, or no β₂ integrin (mock transfectants) to microtiter wells coated with either 10 ∝g/ml WT or γ^390–396A fibrinogen. (B) Adhesion of α_Mβ₂-expressing HEK293 cells to WT or γ^390–396A fibrinogen in the absence or presence of Ab’s against the α_M I-domain (M1/70), β₂ (IB4), and irrelevant rat anti-mouse IgG. The data shown are means ± SD.

Figure 7

Fib γ^390–396Amice exhibit a severe impediment in innate immunity resulting in inefficient clearance of S. aureus in an acute peritonitis model. (A) WT S. aureus was introduced by intraperitoneal injection into WT or Fibγ^390–396A mice, and peritoneal lavage fluid was collected from each animal 1 hour after inoculation to determine the total number of CFUs. (B) Comparative analysis of bacterial clearance in WT and Fibγ^390–396A mice challenged with a S. aureus strain lacking the bacterial fibrinogen-binding protein ClfA. Note that bacterial clearance is inefficient in Fibγ^390–396A mice relative to WT animals regardless of the presence or absence of ClfA. The data shown are means ± SDs. P < 0.035 and P < 0.021 for WT and ClfA S. aureus, respectively (Mann-Whitney U test).

Figure 8

Time course analysis of S. aureus clearance and leukocyte trafficking within the peritoneal cavity of WT and Fibγ^390–396A mice following intraperitoneal infection. (A) Higher levels of S. aureus in Fibγ^390–396A mice relative to WT animals were appreciable 30 minutes after inoculation, and obvious differential expansion was apparent after 3 hours. P < 0.02, Mann-Whitney U test (n = 5 pairs). (B) Representative microscopic views of peritoneal lavage cytospins prepared at the 3-hour time point from WT (left) and Fibγ^390–396A mice (right). Bacterial phagocytosis (arrows) was found in both genotypes, but few residual bacteria were found within WT mice. In contrast, leukocytes from mutant animals appeared overwhelmed with bacteria, which often led to phagocyte disruption (double arrowhead). (C–F) Total cell counts and leukocyte differentials within peritoneal lavage fluid from WT and Fibγ^390–396A mice. Unchallenged mice (C) displayed no significant genotype-related difference in resident cells (P – 0.5, Mann-Whitney U test using four pairs). Three hours after infection (D) a major increase in neutrophil levels was observed in mice of both genotypes compared with naive mice (compare panels C and D), with a trend toward fewer total cells and fewer neutrophils in mutant mice (P ♠ 0.1, Mann-Whitney U test using five pairs). However, comparison of WT and Fibγ^390–396A mice challenged with heat-killed bacteria for 5 hours (E) or 24 hours (F) indicated that there was, if anything, an increase in total leukocyte and neutrophil accumulation in mutant mice (P > 0.2 for all comparisons between genotypes using a Mann-Whitney U test using five pairs). The data shown are means ± SD. Neut, neutrophil; Mac, macrophage.