Antithrombin (AT) exhibits anti-inflammatory properties that reduce mortality in sepsis models. We examined the effects of AT on leukocyte-endothelial interactions in sepsis by using intravital microscopy to measure leukocyte rolling and firm adhesion in post-capillary venules of the cremaster muscle in live mice. Human AT (0.25U g−1 i.v. of Thrombate III, Bayer Corporation) or vehicle was infused prior to proinflammatory stimuli. We evaluated venules with diameters between 20 and 50 μm and shear rates between 200 and 800 s−1. The average diameter and shear rate was not different between treatment groups. The number of leukocytes rolling past a defined vessel point was expressed as leukocyte rolling flux, which normalizes for leukocyte count and flow rate. The number of firmly adherent leukocytes (stationary for 30 s) was normalized for the area of vessel wall analyzed. We evaluated C57BL/6 mice in three inflammatory models. When inflammation was caused by surgical isolation of the cremaster, AT decreased leukocyte rolling flux (28±2 vs. 21±2%, P=0.02) but not the number of firmly adherent leukocytes (41±5 vs. 51±7 per μm2, NS). In contrast, AT did not effect leukocyte rolling flux when inflammation was induced by tumor necrosis factor alpha (0.5 μg intrascrotally). Thus in this model, either pro- and anti-inflammatory effects are balanced or leukocyte rolling involves other interactions that are AT insensitive. Similarly, AT did not affect leukocyte rolling flux induced by an injection of lipopolysaccharide (LPS at 1 μg i.p.) 8 h before observation. However, in this model, AT did inhibit the number of firmly adherent leukocytes (72±6 vs. 39±6 per μm2, P<0.01). Thus, the anti-inflammatory effects of AT can also inhibit leukocyte-endothelial interactions that mediate firm adhesion.

We then tested if these anti-inflammatory effects require a heparan sulfate proteoglycan (HSPG) based AT receptor. We used Hs3st1−/− mice, which lack the isoform of heparan sulfate 3-O-sulfotransferase that is predominantly responsible for the biosynthesis of anticoagulant HSPGs, which bind AT. Although healthy without challenge, 2 of 2 Hs3st1−/− mice died 6 h after LPS administration (1 μg i.p.), whereas none of 18 wild-type (WT) mice (P<0.01) of various strains died from this regimen. Consistent with this lethality stemming from enhanced inflammation, 2 h after LPS, Hs3st1−/− mice had more firmly adherent leukocytes than WT siblings (48±5 vs. 29±7 per μm2, P=0.02). AT decreased leukocyte rolling flux in WT (22±3 vs. 15±1%, P=0.01) and Hs3st1−/− (21±2 vs. 13±1%, P<0.01) mice. AT had no effect on the number of firmly adherent leukocytes observed in WT mice; however AT increased the number of firmly adherent leukocytes in Hs3st1−/− mice (from 48±5 to 80±8 per μm2, P<0.01) - consistent with a loss of AT anti-inflammatory activity. Flow cytometry revealed that Alexa-647-labeled AT bound equally to peripheral blood leukocytes isolated from Hs3st1−/− mice or their WT siblings. Conversely, endothelial HSPGs are very reduced in Hs3st1−/−, compared to wild-type mice.

Thus AT can decrease specific leukocyte-endothelial interactions that occur in sepsis, but the effects are highly model dependent, possibly involving a balance of AT’s pro- and anti-inflammatory activities. HSPGs, on cells other than leukocytes, mediate AT anti-inflammatory activity, as LPS- treated mice lacking this AT receptor have enhanced lethality and exhibit enhanced leukocyte firm adhesion that is augmented by AT treatment.

Author notes

Corresponding author