Inhibition of Binding of β2-Glycoprotein 1 to Phosphatidylserine by Polymyxin B,a Lupus-Like Anticoagulant

Abstract

Polymyxin B is a cationic peptide that inhibits phospholipid-dependent coagulation tests including activated partial thromboplastin time and to a lesser degree prothrombin time. Thrombin clotting time is insensitive to polymyxin B. β2-glycoprotein 1 (β2GP1) is a cofactor of antiphospholipid antibodies. Antiphospholipid autoantibodies also poses lupus anticoagulant activity through interactions with β2GP1. Using affinity chromatography, polymyxin B can effectively decrease the binding of β2GP1 to immobilize phosphatidylserine. Since then, anticoagulant effect of polymyxin B is most likely due to the binding to negatively charged phospholipids, preventing formation of coagulation complexes.

Lupus anticoagulants are a heterogeneous group of in vitro antibodies inhibiting phospholipid-dependent coagulation assays. ^1,2 There is an increased interest in experimental models of lupus anticoagulant. ³ Substances simulating lupus anticoagulant include annexin V, arachnase, protamine sulfate, and polymyxin B. Annexin V (placental anticoagulant I; lipocortin V) binds to anionic phospholipid vesicles and acts as an anticoagulant in vitro. ⁴ Arachnase contains an extract of venom from brown recluse spider, which prolongs clotting times probably due to its sphingomyelinase D activity. ⁵ Protamine sulfate has an anticoagulant effect. In the presence of heparin, a stable salt is formed, and the anticoagulant activity of both substances is lost. ⁶ Polymyxin B, a cationic peptide currently used mainly as topical antibiotic, is also an in vitro anticoagulant. ⁷ Inhibition of coagulation assays is most likely secondary to binding of polymyxin B to anionic phospholipids, thus impeding the formation of coagulation complexes. Inhibition of phospholipid-dependent coagulation tests by Televancin (Vibativ, Astellas Pharma, US) was also recently described but the possible lupus-like anticoagulant effect was not evaluated. ⁸

Materials and Methods

Platele-poor plasma from 3 healthy donors was pooled and diluted 5 times with Tris 0.05 mol/L, NaCl 0.05 mol/L, and pH 7.3 buffer. The phospholipid affinity column not involving covalent linking was prepared using modified method of McNeil et al. ⁹ Ethanol solution and 54.8 µmol of phosphatidylserine (Sigma) was mixed with 15% acrylamide/5% bisacrylamide and rapidly polymerized by tetramethylethylenediamine in the presence of ammonium persulfate. The rigid gel was then homogenized and suspended in Tris/NaCl buffer. Phosphatidylserine polyacrylamide support was incubated with Tris/NaCl buffer containing 100 000 U or 500 000 U of polymyxin B for 30 minutes at room temperature. The gels were then loaded onto 15 × 20 glass columns and washed with Tris/NaCl. Affinity chromatography was performed using 35 mL of diluted plasma. The flow rate was 0.5 mL/min. The level of β2GP1 was measured by the Laurell electroimmunodiffusion assay using 1.5% rabbit antihuman β2GP1. The results are expressed in percentage derived from the standard dilution curve of the pooled plasma used in the experiment. The level of β2GP1 in 4 times diluted plasma is arbitrarily selected as 100%.

Results

Affinity chromatography on noncovalent-linked phosphatidylserine decreased β2GP1 in plasma by 64.3% (from 70% to 25%). Affinity chromatography on polymyxin B pretreated gels decreased β2GP1 in plasma only by 7.1% (from 70% to 65%). This effect was not dose dependent in the dose range used in this experiment.

Discussion

β2-glycoprotein 1 is an important cofactor of antiphospholipid antibodies and is related to its physiological properties of binding to the negatively charged surfaces. ¹⁰ Lupus anticoagulant is a type of antiphospholipid antibody. Autoimmune antiphospholipid antibodies with lupus anticoagulant activity are β2GP1 dependent. ¹¹ Our results show that pretreatment of phosphatidylserine polyacrylamide gel with polymyxin B effectively decreases the binding of the β2GP1. This most likely reflects the masking of the binding site of phosphatidylserine by polymyxin B. Since then, in vitro anticoagulant effect of the polymyxin B is probably through impeding the formation of coagulation complexes on the surface of negatively charged phospholipids.

In 1992, American Pathologists Coagulation Resource Committee sent a sample to various laboratories in the United States in its routine survey proficiency program. The polymyxin B sample performed very nicely as a lupus anticoagulant for all participants using soluble (ellagic acid) activators. For those laboratories that utilized particulate activators (micronized silica and kaolin), it was found that polymyxin B was not a good agent to simulate the lupus anticoagulant effect. Other methods seeking to utilize spider venom and annexin have not proved to be fruitful (1992 CAP Surveys Set CG2-B).

The presence of lupus anticoagulant especially those depending on the presence of an anti-β2GP1 antibodies is a significant risk factor for developing thrombosis. ¹² Accurate assessment of lupus anticoagulant is important since positive results may change the length of anticoagulation for patients with thrombosis. Many laboratory assays for testing of lupus anticoagulant have been described. Unfortunately there is no gold standard diagnostic assay in view of heterogeneity of lupus anticoagulant. ¹³ Most laboratories around the world use dilute Russell viper venom time and activated partial thromboplastin time–based assays. ¹⁴ It is unclear whether higher titer lupus anticoagulants are stronger risk factors for thrombosis than the lower titer. Moreover, current tests used in the diagnosis of lupus anticoagulant are not quantitative, and there are no criteria to define weak positives and strong positives. ¹⁵

Whether polymyxin B or other lupus-like anticoagulant can be used as a positive control in testing for lupus anticoagulant or used as a reference system in reporting of lupus anticoagulant is still an open question. The most advanced research in this area is conducted by Thomas Exner, PhD (personal communication, August 2010).