Since its discovery in 1917, heparin has been a fascinating, and in a way elusive, molecule. Heparin is a glycosoaminoglycan (GAG) formed by repeated sulphated oligosaccharide units. Natural preparations of heparin (usually obtained from bovine lung or porcine intestinal mucosa) can vary in the length of the polymeric unit and therefore have different molecular weights. As such, the biological actions of this GAG can vary quantitatively between different batches of the molecule. The initial activity ascribed to heparin was its capacity to prolong the process of blood clotting, an effect due to its potentiating interaction with the natural inhibitor of thrombin, antithrombin III. These properties have led to widespread use of heparin as an anticoagulant although scant attention has been paid to other biological activities of this GAG.

This situation has changed in recent years. In fact is has long been recognised that heparin has a wide range of biological effects in addition to its well characterised anticoagulant property, including the ability to display anti-inflammatory efficacy (reviewed by Jacques1 and Tyrrel and colleagues2). The anti-inflammatory activity of heparin has been reinforced by positive, although small, clinical trials in patients suffering from a range of inflammatory diseases, including rheumatoid arthritis3 and bronchial asthma.4 In addition, a number of clinical studies have recently demonstrated the anti-inflammatory activity of heparin in the treatment of inflammatory bowel disease at doses that do not produce antihaemorrhagic complications (for example, see Dwarakanath and colleagues5). Given that it is now well recognised that different portions of the heparin molecule exhibit anti-inflammatory activity, and that a pentasaccharide sequence retains the ability to inhibit antithrombin III,6 it is possible that the anti-inflammatory actions of heparin are distinct from its anticoagulant activity.2

Under various experimental and clinical conditions, heparin was found to actively reduce the process of leucocyte recruitment into the site of injury or of application of inflammatory stimuli. In this issue (see page 88), Salas et al provide evidence for the first in vivo mechanism responsible for the antimigratory action of heparin.7 In fact, intravital microscopy techniques have allowed direct observation of inflamed microvascular beds with definition of the paradigm of white blood cell extravasation. Leucocyte interaction with the endothelium of an inflamed post-capillary venule is initially intermittent and dynamic (cell rolling), it then becomes static (firm adhesion), and is finally followed by diapedesis.8 Using the multipotent cytokine, tumour necrosis factor-α (TNF-α) to promote this cascade of events in vivo, Salas and colleagues7 reported that heparin downregulated TNF-α induced leucocyte rolling, adhesion, and migration into gut tissue without affecting changes in vascular permeability. These data extend and confirm previous studies in which heparin reduced leucocyte adhesion to vascular endothelial cells in vitro9 and recruitment of inflammatory cells into other tissues during an experimental inflammatory reaction.2

What is the molecular mechanism(s) responsible for the antimigratory action of heparin? The study by Salas and colleagues7begins to clarify this important issue in an in vivo context. Heparin is acidic in nature and can therefore bind many cationic proteins, including leucocyte and endothelial adhesion molecules (reviewed by Tyrrel and colleagues2). In this new study, heparin at a dose of 500 μg/kg, which was unable to alter the partial thromboplastin time, affected TNF-α induced leucocyte extravasation without modifying TNF-α induced endothelial adhesion molecule (CD62P, CD54, and CD106) expression. This is consistent with in vitro observations that heparin inhibits adhesion of leucocytes to umbilical endothelium cells without altering expression of adhesion molecules.9 Treatment of rats with heparin did not modify the increase in adhesion molecules (CD11b/CD18 and CD62L) measured in polymorphonuclear leucocytes two or five hours post-TNF-α injection7 although increased ex vivo binding capacity of circulating neutrophils to heparin was noted. Of the three adhesion molecules tested, heparin binding occurred at least in part to the β₂ integrin CD11b/CD18, as demonstrated by displacement by a selective monoclonal antibody. These in vivo observations partially confirm the data generated in vitro, emphasising the selective interaction between heparin and CD11b/CD18, but not with CD62L or CD11a/CD18. The authors concluded that heparin blocks CD11b mediated cellular events such as firm adhesion and in this manner this GAG can effectively interfere with the process of leucocyte extravasation that is central to the host inflammatory response.7

As is often the case with ground breaking studies, several other questions can be formulated as a result of this work. In particular, heparin was effective in inhibiting not only cell adhesion to and migration through the mesenteric post-capillary venule endothelium (effects easily ascribed to interference with CD11b mediated events) but it also attenuated TNF-α induced cell rolling, a phenomenon clearly independent of this β₂ integrin.8Therefore, an interaction(s) other than with CD11b must be occurring between heparin and an unknown endogenous protein(s) which sustains white blood cell rolling. Future studies will address this aspect although binding to CD62L (L-selectin) can be excluded.7 Potential interference with CD62P (P-selectin), as demonstrated in vitro, or with selectin counterligands, may be proposed. This possibility is reinforced by the time dependency of heparin inhibition of cell rolling on the inflamed post-capillary venule endothelium (at two but not five hours post-TNF-α).

Will new anti-inflammatory drugs capable of controlling diseases such as colitis be developed out of this research on heparin? This is a pertinent question that is now being addressed in several laboratories. A number of chemically modified fractions of heparin that retain an anti-inflammatory effect have been identified, yet are lacking in anticoagulant activity.2 ,9 Recently, a pentasaccharide sequence containing the antithrombin binding site was described6 that interfered selectively with the coagulation cascade but did not produce the haemorrhagic side effects of heparin (due to the ability of the long polymer to bind platelet factor 4).2 ,6 In addition, pentosanpolysulphate has been introduced into clinical practice in the USA as a treatment for interstitial cystitis based on the antiadhesive effects of heparin.10 There is no reason why more successful molecules cannot be identified based on a better understanding of the antiadhesive action of heparin, and recent studies are addressing this aspect.9 The observation of Salas and colleagues7 will undoubtedly give impulse to this line of research.

In conclusion, the novel study published in this issue ofGut has shed further light on the biological activities of heparin. Fragments of this natural product may in the future lead to the development of novel drugs with a wide range of clinical uses in the treatment of inflammatory diseases, including those of the gastrointestinal system.