haematologica reports 2005; 1(issue 9):October 2005 59

[haematologica reports]2005;1(9):59-60

STEVENSON JLCHOI SHWAHRENBROCK MVARKI AVARKI NM

Glycobiology Research and Training Center, the MooresCancer Center, BiomedicalSciences and Biological SciencesGraduate ProgramsDepartments of Medicine,Pathology, Biology, and Cellular& Molecular Medicine, University of California, San Diego, La Jolla, CA, USA

Heparin effects in metastasis and Trousseau'ssyndrome: anticoagulation is not the primarymechanism

The existence of recurring and success-ful International Conference on Throm-bosis and Hemostasis Issues in Cancer

arises is due to the well-documented andfrequent clinical association betweenmalignancy and excessive thrombosis inhumans. Unfractionated heparin (UFH) hasbeen the classic anticoagulant of choice totreat cancer patients with thrombosis, andrecent studies have indicated that UFHcould be replaced by various low molecularweight heparins (LMWHs), which have sev-eral advantages for clinical use. Heparinsare also the agent of choice in treatingpatients with classic Trousseau's Syndrome(spontaneous migratory thrombophlebitiswith microangiopathic symptoms, typical-ly associated with mucin-producing carci-nomas). Meanwhile, UFH has been shownto reduce tumor metastasis in many murinemodels. Also, retrospective and post-hocanalyses of clinical studies suggest thatUFH and some LMWHs might improve out-comes in human cancer.

Given all these findings, a naturalassumption has been that heparins mustameliorate cancer metastasis and Trous-seau's syndrome via their well-knownactivity in blocking the fluid phase coagu-lation pathway, by potentiating antithrom-bin inactivation of factor Xa (as well as IIa,in the case of UFH). In support of thisnotion, some studies have shown thathirudin (a natural leech-derived thrombininactivator) can also attenuate the meta-static process. However, this assumptionbelies the frequent failure of anti-coagula-tion by Vitamin K antagonists (which alsoreduce factor X and II levels) to ameliorateeither tumor progression or Trousseau'sSyndrome. Furthermore, the doses ofheparin or hirudin used in mouse studieshave been far in excess of those that mightbe clinically tolerated in humans (surpris-ingly, heparin pharmocokinetics in mice hasnot been a subject of well-documentedstudies). Also, heparins are natural com-pounds that comprise a polydisperse mix-ture of glycosaminoglycan chains of vari-

able length, sulfation and epimerization,only a small fraction of which can actuallypotentiate antithrombin action. Finally,heparins have many other potential biolog-ical actions that are not directly related tofluid phase coagulation.

One such non-anticoagulant action ofheparin is inhibition of P- and L-selectinrecognition of mucin-like glycoprotein lig-ands e.g., appropriately glycosylated andsulfated forms of PSGL-1 on leukocytes andmolecules such CD34 and MAdCAM onendothelium. P- and L-selectin are celladhesion/signaling receptors that normallymediate interactions of leukocytes andactivated platelets with one another, and/orwith the endothelium lining blood vessels(see Figure 1, left panel). In this regard, wehave shown that P- and L-selectin can playcritical roles in mouse models of tumormetastasis, and that heparin has limitedadditional efficacy in mice deficient inthese selectins. Also, while heparin effec-tively inhibits a P- and L-selectin-mediat-ed mouse model of carcinoma mucin-induced Trousseau's syndrome, hirudintreatment had no effect. Meanwhile, wehave noted that heparin inhibition ofhuman P- and L-selectin occurs in vitro atconcentrations that are in the clinicallyacceptable range. Selectin blockade thusposes a new paradigm to explain howheparin might attenuate hematogenoustumor metastasis and Trousseau’s syn-drome. This non-anticoagulant action ofheparin would affect critical selectin-medi-ated interactions that occur very early inthe cascade of pathological events that fol-low entry of carcinoma cells or carcinomamucins into the bloodstream (see Figure 1,right panel).

As for other events in tumor progressionthat might be affected by heparins (e.g.,heparanase action, angiogenesis and inte-grin function), almost all are expected tobe involved only after the initial selectin-mediated interactions have occurred. Thus,effective inhibition of two of the earliestmediators of the metastatic cascade (P-

Session VI • Glycosaminoglycans, cancer and thrombosis

haematologica reports 2005; 1(issue 9):October 200560

and L-selectin) could make inhibition of subsequentevents by heparin less practically important. Even ifthis is not entirely true, the fact remains that the oth-er known effects of heparin are also beneficial for thecontrol of tumor progression. Based on all these con-siderations, we have suggested that the failure of vita-min K antagonists to affect tumor progression shouldbe ignored, and that heparin therapy for metastasisprevention in humans should be revisited, with thesenew paradigms in mind. While anticoagulation mayindeed be a major mechanism in rodent experimentsinvolving high doses heparin or hirudin, this is proba-

bly not the case when using heparin in clinically prac-tical doses in humans.

Our current studies are focused on comparing andcontrasting the in vitro and in vivo selectin-inhibitoryproperties of various heparins and heparinoids in cur-rent clinical use in the USA, and at optimizing theirdosing in mice to better mimic the human clinical sit-uation. Our data suggest marked differences in theability of various clinical approved heparins to inhibitP- and L-selectin-mediated processes in vitro and invivo, and show the lack of significance of theirantithrombotic activity in controlling tumor spread inmice at clinically-relevant doses. This data will proveto be important for designing prospective clinical tri-als of heparin in preventing metastasis.

Selectin blockade can have beneficial effects notonly in carcinoma metastasis and Trousseau’s Syn-drome, but also in a variety of other pathological cir-cumstances, involving acute or chronic inflammationand in reperfusion injury. Caution is therefore raisedabout the current trend towards preferring variousLMWHs in clinical practice, based on other criteria.Despite equivalent anticoagulation, hitherto unsus-pected benefits of selectin inhibition in various clini-cal circumstances may be unwittingly discarded. Final-ly, we are working to define heparin fragments withpotent selectin-inhibitory properties relative to anti-coagulant activity - allowing for clinical inhibition ofselectin-mediated pathologies, with only a limitedeffect on the patient’s coagulation state. These datawill eventually be important in making more specificsuggestions for a neo-adjuvant trial in patients withrecently diagnosed carcinomas.

AcknowledgementsWe gratefully acknowledge many prior members of

the lab as well as expert collaborators for some of thepublished results described here. Because of spaceconsiderations, the literature citations listed are lim-ited to those arising from our own work.

Session VI • Glycosaminoglycans, cancer and thrombosis

Figure 1. The cartoon shows P- and L-selectin mediatedinteractions in normal physiology (left panel), and poten-tial P- and/or L-selectin-mediated interactions involvingtumor cells, their membrane-associated and secretedmucins, blood cells and endothelial cells (right panel). P-and L-selectin ligands on both normal cells and tumorcells are often comprised of mucin-like molecules bearingthe correct glycosylation and sulfation patterns to permitrecognition by these selectins. Such selectin-bindingmucins can also be secreted into the blood circulation bycarcinoma cells. Some clinical formulations of heparinscontain subsets of molecules that can inhibit all theseselectin-mediated interactions at clinically-tolerable lev-els, potentially attenuating tumor spread and/or Trous-seau’s syndrome.

References

Koenig A, Norgard-Sumnicht K, Linhardt Rand Varki A: J Clin Invest 101: 877-889,1998

Kim YJ, Borsig L, Varki NM and Varki A: ProcNat'l Acad Sci U S A 95:9325-93301998

Kim YJ, Borsig L, Han HL, Varki NM, and Var-ki A: Am J Pathol 155: 461-472, 1999

Borsig L, Wong R, Feramisco J, Nadeau DR,Varki NM, and Varki A: Proc Nat'l AcadSci USA 98:3352-3357, 2001.

Borsig L, Wong R, Feramisco J, Nadeau DR,Varki NM, and Varki A. Trends Mol Med7:377, 2001.

Varki N and Varki A. Semin Thromb Hemostas28:53-66, 2002.

Borsig L, Wong R, Hynes RO, Varki NM, and

Varki A: Proc Nat'l Acad Sci USA, 99:2193-2198, 2002.

Wahrenbrock M, Borsig L, Le D, Varki N andVarki A. J Clin Invest, 112:853-862, 2003.

Taverna D, Moher H, Crowley D, Borsig L, Var-ki A, and Hynes RO. Proc Nat'l Acad SciUSA 101:763-768, 2004.

Stevenson J, Choi S and Varki A. Clin CancerRes (in press).