The lectin-like domain of thrombomodulin confers protection from neutrophil-mediated tissue damage by suppressing adhesion molecule expression via nuclear factor kappaB and mitogen-activated protein kinase pathways

Abstract

Thrombomodulin (TM) is a vascular endothelial cell (EC) receptor that is a cofactor for thrombin-mediated activation of the anticoagulant protein C. The extracellular NH(2)-terminal domain of TM has homology to C-type lectins that are involved in immune regulation. Using transgenic mice that lack this structure (TM(LeD/LeD)), we show that the lectin-like domain of TM interferes with polymorphonuclear leukocyte (PMN) adhesion to ECs by intercellular adhesion molecule 1-dependent and -independent pathways through the suppression of extracellular signal-regulated kinase (ERK)(1/2) activation. TM(LeD/LeD) mice have reduced survival after endotoxin exposure, accumulate more PMNs in their lungs, and develop larger infarcts after myocardial ischemia/reperfusion. The recombinant lectin-like domain of TM suppresses PMN adhesion to ECs, diminishes cytokine-induced increase in nuclear factor kappaB and activation of ERK(1/2), and rescues ECs from serum starvation, findings that may explain why plasma levels of soluble TM are inversely correlated with cardiovascular disease. These data suggest that TM has antiinflammatory properties in addition to its role in coagulation and fibrinolysis.

Figure 1.

Generation of mice lacking the lectin-like domain of TM. (A) Strategy to introduce TM lacking the NH₂-terminal lectin-like domain into ES cells and mice via homologous recombination. The WT allele for the TM gene encodes a lectin-like domain (LLD), six EGF-like repeats (EGF), a serine-threonine rich region (STR), a transmembrane domain (TM), and a cytoplasmic tail (CT). The amino acid sequence of the NH₂-terminal domain that was deleted is shown. (B) Southern blot of EcoR1-digested ES cell gDNA from WT and homologously recombined cells (lanes 1 and 2, respectively) detected with the 3′-external Probe E. The targeted allele is represented with a 12-kb fragment. (C) Southern blot of EcoR1/Xho1-digested tail gDNA from TM^wt/wt, TM^LeDneo/wt, and TM^{LeDneo/LeDneo} mice (lanes 3, 4, and 5) detected with Probe E. WT and targeted alleles are represented by 10.8- and 7.3-kb bands, respectively. (D) Southern blot of EcoR1/Xho1-digested tail gDNA from TM^LeD/wt, TM^wt/wt, and TM^LeD/LeD mice (lanes 6, 7, and 8) using Probe E. (E) Southern blot of EcoR1/Xho1-digested tail gDNA from TM^wt/wt, TM^LeD/LeD, and TM^LeD/wt mice (lanes 9, 10, and 11) using a HindII/Pme1 internal probe within the 5′-UTR. The WT and targeted alleles are represented by 10.8- and 4.7-kb bands, respectively. (F) PCR confirmation of Cre excision of neomycin gene. Primers s2520 and as2700 were used with tail gDNA from TM^LeDneo/wt, TM^LeD/wt, and TM^LeD/LeD mice (lanes 12, 13, and 14). The WT allele is seen as an ∼170-bp amplicon, whereas Cre excision yields an ∼260-bp fragment. Amplification did not occur across the intact neomycin gene, explaining a single band in lane 12. (G) PCR confirmation of the deletion of the lectin-like domain. Primers s99 and as1005 resulted in an ∼930-bp amplicon from the WT allele, and ∼260 bp from the targeted allele. Gel shows PCR results using tail gDNA from TM^wt/wt, TM^LeDneo/wt, TM^{LeDneo/LeDneo}, and TM^LeD/LeD mice (lanes 15, 16, 17, and 18).

Figure 2.

Response of TM^LeD/LeD mice to endotoxin. (A) TM^LeD/LeD mice and TM^wt/wt sibling controls received 40 ug/gm intraperitoneal LPS and survival time was measured. For each group, n = 22. (B) 6 h after 20 ug/g intraperitoneal LPS, serum cytokine levels and peripheral leukocyte counts (WBC) were measured. TNFα and IL-1β levels are significantly higher in TM^LeD/LeD and TM^{LeDneo/LeDneo} mice. For each group, n = 18. (C) Sections of lungs from untreated TM^wt/wt and (D) TM^LeD/LeD mice were immunostained with MPO for monocytes/PMNs. Positively stained cells, >95% PMNs, are readily detectable in the representative area from the TM^LeD/LeD mice.

Figure 3.

MI/R. (A) In representative MI/R experiments, infarct size relative to AAR, delineated by the dashed and solid lines, respectively, is larger in hearts from TM^LeD/LeD mice (LeD) as compared with those from TM^wt/wt mice (wt). (B) Infarct size, after MI/R, as a function of AAR is larger in hearts from TM^LeD/LeD mice. (C and D) Labeled PMNs were injected upon reperfusion during MI/R, and PMN homing to each ventricle was quantified. (C) Representative sections with more PMNs (arrows) homing to the hearts of TM^LeD/LeD mice. Arrowheads show the limit of the AAR, delineated by thioflavin fluorescence. (D) Fold increase, relative to that observed with TM^wt/wt mice, of adherent PMNs to the LV and right ventricle (RV) of hearts from TM^wt/wt and TM^LeD/LeD mice. *, P < 0.05.

Figure 4.

Activation of PC is normal in the absence of lectin domain of TM. (A) Plasma levels of human PC (hPC) and human APC (hAPC) after infusion of hPC as described in Materials and Methods. The results reflect one of two representative experiments with five mice in each group. (B) Northern analysis of 20 μg total lymphatic EC RNA from TM^LeD/LeD (LeD) or TM^wt/wt (wt) mice, detected with a TM cDNA probe. (C) ECs from TM^wt/wt or TM^LeD/LeD mice were cultured and cell surface thrombin–dependent activation of PC was measured. Results are representative of studies on three clones.

Figure 5.

PMN adhesion to TM^LeD/LeD and TM^wt/wt ECs. (A) Adhesion of PMNs from TM^wt/wt or TM^LeD/LeD mice to ECs in flow chamber (dynamic) or static adhesion models was measured. Results reflect the mean of at least three independent experiments performed a minimum of three times on three different clones. PMN adhesion was significantly greater to non-TNF–treated TM^LeD/LeD ECs than to TM^wt/wt ECs (P < 0.005). Anti-TM antisera increased PMN adhesion in TM^wt/wt ECs (P < 0.005), but had no effect on PMN adhesion to TM^LeD/LeD ECs. (B) PMNs were assessed for adhesion to TM^wt/wt or TM^LeD/LeD ECs in a flow chamber in the presence or absence of TNFα and blocking anti–ICAM-1 and/or anti–P-selectin antibodies, as described in Materials and Methods and Results. PMN adhesion to ECs is represented as a fold increase over that observed with quiescent TM^wt/wt ECs.

Figure 6.

Up-regulation of adhesion molecules in TM^LeD/LeD mice. (A) Flow cytometric detection of adhesion molecules. After gating on isolectin positive ECs (20,000 events) of lung suspensions from TM^wt/wt (red) and TM^LeD/LeD mice (green), surface expression of VCAM-1 and ICAM-1 was detected. In this representative experiment, more adhesion molecule expression is observed for TM^LeD/LeD cells. (B) Western immunoblot of heart lysates from TM^wt/wt (wt) or TM^LeD/LeD (LeD) mice 3 h after treatment with intraperitoneal PBS (−) or 20 ug/g LPS (+). In each lane, 200 μg total protein was added. (C) Western immunoblots of heart lysates of mice treated as in B. Total ERK_1/2 and tubulin show equal loading of protein. Generation of activated ERK_1/2 (pERK_1/2) is increased in LPS-exposed hearts of TM^LeD/LeD mice.

Figure 7.

Effects of soluble lectin-like domain of TM on PMN adhesion, NF-κB, and MAP kinase activation. (A) In a static adhesion model, TM_lec155 significantly decreased PMN adhesion to TM^LeD/LeD ECs and fEND.5 cells. (B) Regulation of ERK activation by soluble lectin-like domain of TM. HUVECs were exposed to TNFα (1), TNFα plus GST (2), TNFα plus GST-TM_lec155 (3), or TNFα plus TM_lec155 (4) for 20 min. Western immuno-blots of cell lysates detected TNFα-induced up-regulation of pERK_1/2 and NF-κB, which were suppressed by the addition of GST-TM_lec155 or TM_lec155. No change in total ERK_1/2 was detected. In the absence of TNFα, there was no detectable pERK_1/2 or NF-κB (not depicted).