The anticoagulant activity of heparin: measurement and relationship to chemical structure

For many years the anticoagulant activity of heparin has been estimated by coagulation assays, in which the prolongation of clotting times by heparin is measured under various conditions. More recently, assays have been developed which measure the inhibitory action of heparin on isolated coagulation enzymes, notably Factor Xa and thrombin, using specific amidolytic peptide substrates. The anticoagulant activity of heparin arises primarily from its ability to bind to antithrombin III (AT III), altering the conformation and enhancing the activity of this major protease inhibitor. Passage of heparin through an immobilised AT III column yields two fractions: a high affinity fraction with 300-350 iu mg-1 anticoagulant activity, comprising one-third of the total, and a low affinity fraction with an activity of less than 10 iu mg-1, comprising the remaining two-thirds. Studies in several laboratories have demonstrated that a specific pentasaccharide sequence is required for AT III binding. The authors have shown that the presence or absence of this sequence can be detected by high-field proton NMR, thus providing a semi-quantitative method for a functionally important group. A second major influence on anticoagulant activity is molecular weight distribution. Studies in the authors' laboratory on a series of fractions of 5000-35,000 showed that whereas anticoagulant activity in APTT clotting assays decreased with decreasing molecular weight (Mr), activity in anti-Xa assays was maintained or increased in the low Mr fractions. However, in vivo studies showed that high affinity fragments with anti-Xa activity only were poor antithrombotic agents. It appears that the presence of the AT III binding site alone is not sufficient for full antithrombotic activity; an extra length of polysaccharide chain of at least 15 residues is required. Molecular weight distribution is readily assessed by HPLC, although the lack of suitable reference materials hampers assignment of absolute molecular weights. Important determinants of anticoagulant activity can now be assessed by physicochemical techniques but, at present, these techniques are not precise enough to replace anticoagulant assays as predictors of in vivo behaviour.