Circulating Heparan Sulfate Proteoglycan Anticoagulant from a Patient with a Plasma Cell Disorder

Circulating heparan sulfate proteoglycan anticoagulant from a patient with a plasma cell disorder.

M S Khoory, … , E J Bowie, K G Mann

J Clin Invest. 1980;65(3):666-674. https://doi.org/10.1172/JCI109712.

Research Article

A woman, aged 68, with multiple myeloma (immunoglobulin[Ig]A kappa type) developed an anticoagulant with properties suggestive of heparin. The anticoagulant prolonged the thrombin time but not the reptilase time and was resistant to boiling, proteolytic enzyme digestion, and trichloracetic acid precipitation. The thrombin time was corrected by the addition (in vitro) of protamine sulfate or the addition of purified platelet Factor 4 (PF4) to the plasma. The anticoagulant was isolated by PF4-Sepharose affinity chromatography and analyzed in terms of its molecular weight, uronic acid, and amino acid composition. The proteoglycan isolated had a mol wt of 116,000 and appears to consist of two 38,000 dalton polysaccharide units interconnected by peptide material totaling 39,000 daltons. Electrophoretic analysis of the pronase digested peptidoglycan using the lithium acetate-agarose technique suggested the material was of the heparan sulfate type. The peptidoglycan had about one-tenth the specific activity of commercially available heparin on a weight basis. The isolated proteoglycan was indistinguishable from commercial heparin when analyzed in terms of its ability to act as a cofactor in the antithrombin III inhibition of thrombin.

Find the latest version: https://jci.me/109712/pdf Circulating Heparan Sulfate Proteoglycan Anticoagulant from a Patient with a Plasma Cell Disorder

\MAGED S. KHOORY, \MICHAEL E. NESHEINI, E. J. WALTER BOWIE, and KENNETH G. MANN, Hematology Researchi Sectioni, AMailo Cliniic, Roch5ester, MlinnICesota 55901

A B S T R A C T A womiiani, aged 68, with miiultiple polymerization (3-5). The patienit described in this imiyelomiia (im munoglobulin [IgJA kappa type) developed study appears to be unii(que in that she developed a an anticoagulanit with properties suggestive of heparin. coagulopathy not directly related to the miyeloma The anticoagulanit proloniged the thromiibini tiunle but not protein, but clharacterized by the presence of a circulat- the reptilase timiie and(l was resistanit to boilinig, proteo- ing proteoglycan with properties similar to heparini in lytic enizymie dligestioni, anid tricliloracetic acid precipi- that it functionied as cofiictor to antithrombin III. tation. The thrombin timiie was corrected by the additioni The acid mnucopolysaceharides are presenit as stirLctual (in vitro) of protaminie sulfhate or the additioni of purified miiembers of conniective aiud other supportinig tissues platelet Factor 4 (PF4) to the plasmiia. The aniticoagulanit of mammiiiials, anid their preseniee in a variety of cell linles was isolated by PF4-Sepharose affinity chromnatography has been described (6, 7). The presenice of heparini in and analyzed in terms of' its miiolecular weight, uroniic blood has beeni reportedl (8) altlhouigh this observation acid, anid aminio acid coimlpositionl. The proteoglycan was bx'iindirect proceduires anid lhas nlot beeni conifirmiied. isolated had a miiol wt of 116,000 anid appears to consist The presenlt stu(ly clearly demiionistrattes the presence of of two 38,000 daltoln polysacchari(le unlits initerconi- ani antitlhromiibini III cofactor active miiaterial in humilan nected by peptide miiaterial totalinig 39,000 daltonis. plasmiia and shovs that it was nlot associated withl the Electrophoretic analysis of the proinase digested pep- plasmacytes or the immunoglobtlini fractioni of the tidoglycan using the lithiumii acetate-agarose techniique plasmiia. The isolationi anid partial characterizationi of tlhe suggested the miiaterial was of the heparani sulfate type. material is also presen-itedl. The peptidoglycan ha(l about onie-ten-thl the specific activity of commercially available heparini onl a weight METHODS basis. The isolated was proteoglycani inidistinguishable Aminohexvl Sepharose was obtained fromil Phariiaciaii Finie from commercial heparin wlheni analyzed in termiis of its Chemilicals, Inic., Piscataway, N. J. 10% Agarose and(l P-150 ability to act as a cof'actor in the anitithrolmpbini III were obtainied fromil Bio-Rad Laboratories, Richilloild, Calif. inhibitioni of thrombini. The comimillercial heparini used for routinle assays was froml1 Upjohn Co., Kalamilazoo, Mich. (heef lunlg), wlhereas reference stanidards of the acid mucopolysaccharides were provided by INTRODUCTION Dr. Martiii NMatthewvs,' Universitv of Chicago. Agarose for electrophoretic analysis was SeaKemil (ME), ob)tained firolim Mlultiple myeloma may interfere with hemiiostasis in MIarine Colloids, Inc.' Rockland, Maine. Azoruhiin S dye, several ways (1). These incltilde the dlirect comiiplexinig D-glucuromo-3,6-lactone, and( carbazolq wvere fi'oiln Aldrich of the paraprotein with a coagtulationi f;actor (2), or in- Chemiiical Co., Milwaukee, Wise. The water soluble r-e- direct interaction by interferenice with fibrin miioniomiier agenit I-ethl\1-3-(3-d imllethvlai\tinopropvl-)carl)odiiidi(le was fromil Sigmila Chemiiical Co., St. Louiis, Mo. Crtulde thromhin foi routinie aniticoaggulanit analysis was oltained fromii Parke, D)avis A preliminary report of this studv was presenited ata imieetinlg & Co., Detroit, Mich. Proniase was obtaine(d fromim Sigmiia of the Federation for Amiierican Societies for Experimiienital Chemical Co., (protease IV). Platelet Factor 4 was purifie(d bh Biologv, April 1978, Atlaniitic City, N. J. 1978. Fed. Proc. 37: the methlo(d of Ruiz et al. (9). Purifie(d bovine a-thrombinm use(d 3. (Abstr. 618) for examiiniationi of initeractioni of the inhihitor of antithromhlln Dr. Khoorv wvas supporte(d bh Bloo(o Banikinig anid lemiiostasis III was prepared 1y the miiethlodl of Lunidblad et al. (10). Bovine traininlg granit HL-07069 fromii the National Heart, Ltiung, and(l antithrombin III was prepared uisinig the miietlho(d of Datmns atn(d Blood Institute. Dr. Mann is ani Established Investigator of the Am-ierican Heart Associationi. Address reprinit re(luests to Dr. Manni. I Acidl mucopolysaccharide referenice stanl(lar-ds, April Receive(ifor puiblicatiom 9 Mlarch 1979 ani(I ini revtise(lform 1977, MI. B. Mlatthews, J. A. Ciff)oielli, D)epartmnent of Pediatrics, 9 Novemnber 1979. University of Chicago, Chicago, Ill.

666 J. Clitn. Inoest. © The Americani Society foirtCClinical Intvestig(ation), Inc. 0021-9738I80I0'3I066609) $1.(00 Voltumiie 65 March 1980 6661-674 Rosenberg (11). Analyses of thrombin concentrations were of acid mucopolysaccharide were performed according to the performed by the modified National Institutes of Health (NIH) procedure of Homer (18) in lithium acetate buffer and 0.6 or procedure (12) described by Mann et al. (13). 1.2% agarose gels. Staining was accomplished using toluidine Coagulation tests were performed by published methods blue 0. (14). The appearance of the anticoagulant material throughout Antithrombin III cofactor activitv of the isolated anticoagulant the different steps of the isolation and purification procedures was evaluated with purified coomponents. Solutions of the were monitored by the thrombin time using crude bovine isolated inhibitor at concentrations equivalent to 0.66 U thrombin reconstituted to yield a normal thrombin time of heparin/ml (deten-nined from thrombin-time assays onl nlorimnal 16-18 s in the absence of added inhibitor. As a substrate, human plasma) were incubated at 220C in 0.02 M Tris-HCl, 50 ,ul of a 0.15 M NaCl solution in which the unknown material 0.15 NI NaCl, pH 7.4 with bovine antithrombin III (2.44 ,utM) was dissolved was added to 150 ,ul of normal pool of fresh and human a-thrombin (0.856 KLM). At regular intervals after plasma collected in oxalate; as a control, 50 ,ul of 0. 15 M NaCl the addition of thrombin aliquots of the reaction mixture were was added to 150 ,ul of normal plasma. The anticoagulant withdrawn and assayed for residual thrombin-catalyzed clotting activity of the material was measured against a standard curve activitv according to the NIH procedure (12) as modified by prepared with increasing dilutions of beef lung heparin with Mann et al. (13). Control experiments were also performedl 0.15 M NaCl as a diluent. Because of the variation of the using an eqtuivalent amount of commercial heparin in place thrombin solution from day to day, such a standard curve was of the isolated anticoagulant. In some instances human PF4 freshly plotted every time an assay ofthe anticoagulant activity at a final concentration of 0.14 mg/!lll was included to compare was to be done. the neutralization of the heparin-like cofactor activitv of the Proteolytic digestion was accomplished by incubating the patient-derived inhibitor to the neutralizationi of commercial thawed patient plasma, freshly collected in acid citrate dextrose heparin. and stored at -700C, or the isolated proteoglycan from platelet The patient's leukocvtes, obtained by leukophoresis, were Factor 4 (PF4)2-Sepharose column, with pronase (3 mg/100 ml) pelleted by centrifugation at 800 g for 8 min. Cells were then for 48 h at 450C. At the end of this period, solid trichloroacetic prepared for either short-tenrm culture or homogenization. For acid was added to 5% (wt/vol) and incubation continued for 2 h short-term cultures, the cells were resuspended in RPMI at 40C. The supernate from this as well as two washes of the 1640 medium and cultured in Marbrook vessels (19) as precipitate with deionized water were exhaustively dialyzed described by Katzmann (20). Cell pellets were homogenized against the water and lyophilized. in a pestle tissue grinder maintained at 0°C. Nuclei and PF4 was coupled to aminohexyl Sepharose 4B in the follow- cellular debris were removed by centrifugation at 1,000 g for ing manner: to 1 ml of packed resin was added 2.5 ml human 10 min. The supernate of the short-term leukocyte culture did PF4 (2.6 mg/ml) in 0.5 M NaCl, pH 5.5, 22°C. 20 mg of solid not possess anticoagulant activity, nor did the extracts of water-soluble carbodiimide was then added with stirring and homogenized cells. the pH maintained at 4.5-5.5 by the addition of dilute HCl Case report. The patient was a 68-yr-old female with a over a period of 1 h. Stirring was then continued overnight 10-mo history of multiple myeloma (imunuoglobulin[Ig]A at 22°C. The resin was then filtered, washed with 100 ml of kappa) who had been treated with L-phenvlalanine mustar(d and 0.02 M Tris-HCl, 1.5 M sodium chloride, pH 7.4. Resin was prednisone. She presented with generalized petechiae antd then washed in 0.02 M Tris-HCl, 0.15 M NaCl, pH 7.4 for stor- large hematomata on the upper extremities. age or further use. Laboratorv studies: hemoglobini 10.2 gIlO0 ml, leukocvtes Molecular weight estimations were performed with the 29,500/mm3 with 63% plasma cells; platelet count 12,000/ high speed sedimentation equilibrium technique (15) using a mnm3; calcium 15 rng/100 ml, uric acid 6.6 mg/100 ml, and Beckman model E analytical centrifuge (Beckman Instruments, creatinine 7.2 mg/100 ml. Total serum protein 7.2 g/100 ml Inc., Fullerton, Calif.). No gross curvature was observed in with 4.5 g/100 ml albumin and 0.28 g/100 ml alpha 1, 0.92 g/ high speed sedimentation equilibrium experiments. Therefore, 100 ml alpha 2, 1.4 g/100 ml beta, and 0.82 g/100 ml gammlla weight average molecular weights (NMw) were calculated from globulin. The IgA was 4.9 mg/ml, IgM 0.15 mg/mIl, and IgG direct analysis of the slope of plots of log fringe displacement 4.0 mg/mIl. The urine was positive for Bence Jones protein (lnc) vs. radial distance squared (r2) using the equation: and showed a monoclonal kappa spike on immunoelectro- phoresis. Bone marrow aspiration showed over 90% abnormal 2RT dlnc plasma cells. No mast cells were seen. Coagulation studies Mw = coNl - vp) dr2 are presented in Table I. The Z average molecular weights were calculated by extrap- olation of Mwr, the weight average molecular weight at any RESULTS point r to the cell bottom. A value of 0.53 cm3/g (16) was usedl for v, for the peptidoglycan, since this value (determined for The chromatographic separation of the anticoagulant chondroitin 4-sulfate) is appropriate for the degree of sulfation material present in the patient's plasma on PF4- reported for heparan sulfate. A value of 0.61 cm3/g was used Sepharose is presented in Fig. IA, whereas the chro- for the proteoglycan based upon composition (38-41% protein) matographic behavior of beef lung heparinized-humani and a value of 0.73 cm3/g was used for protein. acid citrate dextrose plasma used as a control is presented Speeds were chosen such that dlnc/dr2 was between 1.5 and 2.5. For all experiments, meniscus depletion and the attain- in Fig. lB. In each case, 5 ml of the anticoagulated ment of equilibrium were verified experimentally. plasma was allowed to percolate through -3 ml of PF4- Amino acid analyses were performed by single column tech- Sepharose. The material not bound in each case repre- niques on a Beckman model 119 amino acid analyzer. Before sents most of the plasma protein present in the original analysis samples were hydrolyzed in 6 N HCI for24 h at 1 10°C. Uronic acid content was measured by the carbazole reaction samples. The plasma used in the heparin control was according to Bitter and Muir (17). Gel electrophoretic analyses anticoagulated by the addition of 250 U of beef lung heparin, and before application to the column had a 2Abbreviationt used itl this paper: PF4, platelet Factor 4. clot time of > 100 s. The initial fractions from this column

Circuilatitng Proteogly catn Atnticoagtulant 667 TABLE I The amount of' anticoagulant material present in the Patient's Coagtulationi Studie,s* patient's plasma was about 30 gsg/ml, based upoIn re- coveries from the PF4-Sepharose columlln. Prothrombin time 12 (11) s For the columnn to which the patient's plasimia was Partial thromboplastin time 50 (45-60) s Activated partial thromboplastin applied, the comiiponents eluted at 0.75 -M salt coin- time 58 (25-40) s tained, in addition to anticoagulant activity and tironic Stypven time 14 (15) s acid, sonme proteinaceous comiiponients. In addition, Reptilase time 21 (14) s immunodiffusioni experiments revealed the presence Thrombin time 600 (16-18) s of trace amnounts of human antithrombin III in the comil- Thrombin time with equal ponents eluted in this fraction, whereas sodiu-m dodecyl volume of normal plasma 600 s suilfate gel electrophoretic analysis showed fainit an-iotunts Thrombin time with 10 ,ug/ml of of a variety of proteins associated with this peak. The protamine sulfate 22 s components eluted in the 0.75 M fraction fromn the pa- Thrombin time after barium tient's plasma were pooled and suibjecte(d to chromatog- sulfate adsorption 24 s on Thrombin time with 50,g/ml of raphy Bio-Rad P-150. Anticoagulanit activity was purified PF4 19 s eluted in the void volume of this column. The miiaterial Antithrombin III 21 (18-24) ,ug/ml retained its absorbance at 280 nm, suiggestinig the con- Fibrinogen 250 (190-300) mg/100 ml tinued presence of protein. Fibrin split products 5 (<5) t.g/ila To coiipare the patient-derived ainticoagulanit with Protamine gel positive (negative) commercial hepariin, the 0.75 M fraction was subjected to proniase digestion and trichloroacetic acid precipita- * Normal values in parentheses. tion to remove the protein comlponenit(s). These steps were taken because commlllercial heparin is obtaine(d had a thrombin clot time of 22.6 s. Similarly, patient onily after rigorous proteolytic and chemical digestioni plasma, which before application to the column had a of the animiial tissue fronm which the polysaccharide is plasma thrombin clot time of > 100 s, exhibited a thrombin derived. After this treatinent all absorbance at 280 nIn clot time for the initial fractions of 25.5 s. Thus, in both was lost from the anticoagulanit preparation, whereas the experimental and control PF4-Sepharose columlln, the anticoagulant activity was completely recoveredl. A the anticoagulant activity was removed on the column, comparison of the gel filtration behavior of this material and the clottability of the plasma coming through the with comnmercial heparin, usinig 10% agarose, is pre- colunmn was restored. Subsequent elution ofthe column sented in Fig. 2. In contrast with commnercial heparin, after washing with 0.5 M sodium chloride, was ac- the patient-derived aniticoagulanit peptidoglycani is complished with 0.75 M salt. The anticoagulant activity excluded from a 10% agarose column and appears to eluted from each column is expressed in terms of the be of substantially larger size than commiiiiercial beef fraction of the quantity of anticoagulant applied to the lulng heparin. column based upon the heparin assay described in The isolated peptidoglyean was subjected to agarose Methods. Under this set of conditions, anticoagulanit gel electrophoretic analysis in lithiumii acetate, pH 3, activity, some protein (as exhibited by absorbanice at to compare its relative electrophoretic mobility to that 280 nm), and uronic acid were elaborated from both of other known sulfate(d polysacchari(le standards. The the chromatography of the patient plasma and chromatog- electrophoretic miobility of the acid mulcopolysaccharides raphy of the heparinized normal plasma. Subsequent are functions of their charge density, and mnost of these elution of both columns with 1.5 M sodium chloride can be distinguished electrophoretically. Represented resulted in the elution of a small amount of additional in Fig. 3 is a photograph of agarose gel electroplhoreto- anticoagulant activity from the patient's plasma aind the grams of reference stanidard acid mucopolysaccharides larger fraction of anticoagulant activity from the column and the anticoaguilant peptidoglycan derived from the to which the heparinized normal plasma was applied. patient's plasma. Reference standard acid miiucopoly- The appearance of two peaks from each column is a saccharides are present in gels B through H. The dye- reproducible occurrence that does not appear to be front marker, amaranth, is in gel K. The patient-derived related to individual heparin components in each sample, anticoagulant is present in gels A, I, and J; as the iso- but rather to anomalous columiin behavior, in that reap- latedi material (gel A), as the isolated material mixed plication of either of the components eluted at 0.75 M with heparin (gel I), and the isolated material mixed or 1.5 M results in a similar chromatographic pattern; with heparan sulfate (gel J). It canl be seen that the that is, two peaks on elution. Most likely this anomalous patient-derived peptidoglycan ainticoagulant is electro- behavior is a reflection of heterogenous bindinig of PF4 phoretically homogeneous in this system and is eintirely to the agarose support rather than heterogeneity in the distinguished from heparini and five of the other sulfated anticoagulant activity applied to the column. polysaccharide reference standards used for comparison.

668 M. S. Khoory, Al. E. Nesheinmi, E. J. IV. Bowie, anid K. G. Mlannu A l.5M

0.75 M 0.5 125 0 1.5M ' 0.4 120 co 0Nco .JO 0.3- >.

-0 4- a "I 0.2 30 *- 0.11 5 20 c

10 20 30 40 Fractions- 1.5 ml each B

1.5M t f a 0.5 25

0.4 0.75 M 4) 10 0i 0 (4 0.3 [

0.2 .m 0.1

0 30 40 Fractions- 1.5 ml each FiGuRE 1 (A) Chromatography of patienit ACD plasma oIn humiiian PF4-aminohexyl Sepharose and (B) of heparin anticoagulated niormal humani acid citrate dextrose plasma on human PF4- amninohexyl Sepharose. Absorbance at 280 nm (0), aniticoagulation activity (0), and uronic acid contenit (x) are plotted vs. fraction member.

The material undergoes coelectrophoresis with heparan tienit is a proteoglycan for which the acid polysaccharide sulfate. Subsequenit to this experimient, reference portion is heparan sulfate, one can estimate a weight- standard heparani sulfate was subjected to PF4-Sepharose fraction proteini of from 38 to 41% based upon the amino affiniity chromatography, anid was showni to behave in acid to hexuronic acid ratio. After pronase treatment, similar fashion to the anticoagu-laint derived fromn the 85% of the amino acids are lost from the total composite patienit's plasma. to provide a ratio of 0.317 amino acids per uronic acid The amino acidl composition of the proteoglycan residue for the pepticloglycan. Again assuming that the relative to uronic acid comnpositioni is presente(d in acid polysaccharidle is heparan sulfate for which the Table 11. The suin for amino acid composition indicates polvsacchari(le chaiin is 44-49% hexuronic acid, the about 2.23 aminiio acids per hexuironiic acid residute. For peptidoglycan resultinig from pronase digestion cor- heparan sulfate, the hexuronic acid content correspondls respondls to -91-92% acid polysaccharide with 8-9% to 44 (21)-49%' of the total polysaccharide Unlit mass. of residuial (not digestible by pronase) peptide material. If one presumnles that the material isolated fromii the pa- Frequently, gel filtration has been used to estimate

(Circulatinig Proteoglycan Aniticoagulant 669 TABLE II Amiino Acid Composition of Anticoagulantl laterial A N. 1.0 S 0.8 Proteoglvca,i Peptidoglyean ,i 0.6 n1/oll/niol uronic (aci(d t 0.4 *g 0.2 Aspartic acid 0.18 0.035 n Threonine 0.14 0.032 5 10 15 20 25 30 35 40 45 50 Serine 0.26 0.030 Elution volume,ml Glutamic acidl 0.32 0.049 Proline 0.16 0.031 FIGURE 2 Elution profiles of pronase-digested, TCA-treated Glycine 0.32 0.037 anticoagulant (0) and Upjohn heparin (0) on a gel filtration Alanine 0.18 0.023 column prepared with 10% agarose. Relative anticoagulant Valine 0.11 activity is plotted vs. elution volume. The void volume for the column is -18 ml. Methionine 0.02 Isoleucin1e 0.05 Leucine 0.12 0.009 the physical size of acid mucopolysaccharides. Gel Tyrosine 0.05 filtration chromatography as a technique for molecular Phenylalanine 0.08 0.014 weight estimation, is somewhat complicated by the need Histidine 0.08 0.009 Lysine 0.09 0.009 to calibrate the gel filtration column in (luestion with Arginine 0.07 0.009 an appropriate set of markers that represent model com- Hexuronic acidl 1.0 1.0 pounds of known chemical structure and molecular Amino acid 2.23 0.317 weight, which are the hydrodynamic equivalents of the molecule whose molecular size is to be estimated. When used for the analysis of complex proteoglycans that of the patient-derived anticoagulant by analytical ultra- may have peptide backbone as well as multichain or centrifugation. A most precise analysis of sedimentation branch chain structures, gel filtration analysis can lead equilibrium data requires knowledge of the total com- to gross error. For these reasons and the fact that we do position of the material for the calculation of the partial not have detailed knowledge of the nature of the struc- specific volume terms. At present, detailed total ture of the anticoagulant, we chose to evaluate the mass compositions for the patient-derived anticoagulant is not available. However, sufficient partial composition data is available to permit estimation of the appropriate A B C D E F G H K partial specific volume term for the peptidoglycan and .If proteoglycan (see Methods). We have performed sedimentation equilibrium experiments with the isolated patient anticoagulant under three sets of conditions. The data are summarized in Table III. Two sets of experiments were performed with the intact, protease nondigested anticoagulant. In dilute aqueous buffers, the molecular weight of the anticoagulant and any noncovalently associated materials are measured. Under this set of conditions, an average molecular weight of 111,000 was obtained, presuming a partial specific volume of 0.61 cm3/g. The sedimentation equilibrium data obtained under these sets of conditions showed some evidence of nonideal behavior near the cell bottom. However, this problem FIGuRE 3 Lithium acetate-agarose electrophoretic analysis was not sufficiently severe so as to preclude an accurate of reference standard acid mucopolysaccharides and patient- interpretation of the data. Sedimentation equilibrium derived peptidoglycan. Gel A contains patient-derived pep- were also with the isolated tidoglycan anticoagulant; gel B, reference standard heparin; experiments performed gel C, heparin obtained from Upjohn; gel D, keratan sulfate; proteoglycan in the presence of 6 M guanidinium chlo- gel E, dermatan sulfate; gel F, heparan sulfate; gel C, hyal- ride. Under this set of conditions, noncovalent inter- anti- uronic acid; gel H, chondroitin-6-sulfate; gel I, patient actions between the covalent proteoglycan unit and coagulant peptidoglycan mixed with reference standard any other noncovalent associated entities would be heparin; gel J, patient peptidoglycan anticoagulant mixed with reference standard heparan sulfate; and gel K, amaranth destroyed and the molecular weight obtained should dye marker. reflect the mass of the covalent anticoagulant unit.

670 M. S. Khoory, M. E. Nesheinm, E. J. W. Bowie, and K. G. Mann TABLE III Molecular WVeight of Antticoagulatit

Preparation Solvent MolecularM weight Amiiino acid

C1113/g % PF4 chromatography 0.15 M NaCl 0.61 111,000 38-41 P-150 chromatography 0.02 M Tris pH 7.4 PF4 chromatography 6 M Guanidinium 0.61 116,000 38-41 P-150 chromatography Chloride PF4 chromatography 0.15 MI NaCl 0.53 42,000 8-9 P-150 chromiatography 0.02 NI Tris Pronase digestion pH 7.4 TCA precipitationi

Under these sets of conditions, an average mol wt of from the patient's plasmna would suggest the structure 116,000 was obtained, again presum-ing that the partial containinlg two polysaccharide chains of 38,000 mol wvt specific volume of the proteoglycan is 0.61 cm113/g. liniked to a peptide chaiin (or chains) wlhose total mass Pronase digestion of the proteoglycan anticoagulanit would be -39,000. Such a structure wouild be -66% removed -86% of the amino acid imiaterial iniitially carbohydrate and 34% peptide. These valtues are within associated with the anticoagulant and the resulting 4-7% of the estimated composition values based uponl polymer gave a mlol wt of 42,000 when examinied in the partial coimipositioin data available. The coinsisteincy the analytical ultracentrifuge, presumiinig a partial specific betweeni the values obtained from the proteoglycan in volume of 0.53 Cm113/g. The relative homogeneity of the the native state in dilute aqueouis buffer ( 111,000) and polysaccharide unit is illustrated in the sedimenitationi that obtaiinedl in 6 NI guaniidiniiumii chloride (116,000), equilibrium data presented in Fig. 4. These data clearly ani( the consistency between the values obtained for show that the polysaccharide is not highly polydisperse. relative fractioin of pepticle to carbohydrate based upoIn In addition the NIz value determiined for this analysis both chemiiical analysis and physical analysis strongly was 48,000, inidicating a relatively narrow ranige of suggest that the appropriate assumilptionis have been molecular weights for the polysaccharide mass distribu- made with respect to partial specific volume for the tion. Approximately 9% of this imiaterial is aminio aci(d intact proteoglycan based Uponl the partially assumptive by compositioin data, and thus the polysacchari(le miiol analysis of comlipositioni of the patient-derived aniti- wt is 38,000. coagulanit. A composite of analysis of the molecular weight anid Functioinal assays of the relative anticoagulaint potency comipositioin data obtained for the proteoglycan isolated of the patient-derived peptidoglycani were coinducted by comiipariing its effects to that of commiiiiercial heparini oIn a plasmiia-thromiibini clot timiie. The patienit-derivedl 1500 r2> anticoagulanit had a specific activity of -8 U/miig, which is substantially less thani the specific activity of the 1000 commlllercial beef luIng heparin, which has a specific 750 - activity of -120 U/miig. The niatture of the initeraction of the patient-derived 500-3 ainticoagtulanit peptidoglycan with antithrombin III, thrombin, and PF4 vere compared to that of beef lunig 350 heparin. The results of these experimeints are presented 250/ in Fig. 5. The thrombin inihibition data for each of the experiments presenite(d in Fig. 5 are expresse(d as pseuldo first-or(ler rectetionis. The aiimounlt of patient-dierived 150 / anticoagulanit or beef lutnig heparini adIdledI to aniti- 50.4 thrombin III (closedl an(l openi circles, respectivelv), r2 (cm2) are eqdiivalenit basedl uponi the plasmiia thromiibini clot timne assay (lescribed in Niethods. Upon a(diniistration FIGURE 4 Sedimentationi equilibrium clistributioni for peptidoglycan, at 18,000 rpm (22.3°C). Concentration (c) is plotted of PF4 to the mixture of heparin, antithrombin III, vs. radial distance sqtuared. The cell bottom is signifiedl by r2,. aind thromnbini or that conitaininlg patient-derived prote-

Circulating Proteogligc.an Anticoa(ulant 671 100 TABLE IV Rate of Thromnbin Inihibition

Component added k* Halt:life

liter* wini -' ntiol-V x 10(-5 "?li Antithrombin III + PF4 1.22 2.33 Antithrombin III + heparint 6.51 0.44 .t 50 Antithrombin III + anticoagulant§ 6.91 0.41 s Antithrombin III + heparin + PF4 1.03 2.76 o Antithrombin III + anticoagulant + PF4 1.10 2.57 20 - I \\ * The second-order rate constant (at 22°C) determined from pseudo first-order plots wvith AT-III in 2.85-fold mol excess over Ila. I Heparin (intestinal mucosa) obtained from Eli Lilly & Co., 1 2 3 4 5 6 Indianapolis, Ind. § Patient-derived proteoglycan anticoagulanit. Time (min) FIGuRE 5 Thrombin inhibition by antithrombin III in the DISCUSSION presence of commercial beef lung heparin or patient-derived anticoagulant. The reaction is analyzed as a pseudo first-order The anticoagulant activity present in this patient's reaction in terms of the logarithm of the percent thrombin plasma could be quantitatively bound to and subse- remaining as a function oftime (minutes). The time-dependent quently eluted from PF4 immobilized on agarose. The decrease of thrombin activity with antithrombin III alone is shown (A). The mixture of antithrbinbin III, thrombin, and chemical and physical data derived from studies of the patient-derived anticoagulant (0.6B heparin equivalent U/ml) anticoagulant suggest that it is likely composed of two is shown (0) and a mixture of antithrombin III, thrombin, polysaccharide chains of nearly identical size intercon- and heparin (0.66 units/ml) is given (0). Results obtained by nected by peptide material. Whereas current data adding PF4 (0.14 mg/ml) to these two samples are designated indicates that the polysaccharide chains present in the by A and L, respectively. anticoagulant proteoglycan are of nearly identical size, no knowledge is presently available as to whether a oglycan anticoagulant, antithrombin III, and thrombin, single or multiple polypeptides are responsible for the equivalent rates of inactivation for thrombin were protein constituent of the proteoglyean. In any event, observed. These rates obtained after neutralization of the protein constituents do not appear to have an influ- heparin and heparin-like anticoagulant are experi- ence on the anticoagulant activity of the proteoglycan, mentally equivalent to the rate of neutralization of the in that pronase digestion leaves this activity intact with equivalent amount of thrombin by the equivalent the peptidoglycan. Further, the anticoagulant activity amount of antithrombin III alone (Fig. 5, open triangles). associated with the peptidoglycan is consistent with It should be noted here that the rate observed and its behavior as an antithrombin III cofactor in a mannier presented for Fig. 5 inhibition of thrombin in the completely analogous to heparin. Electrophoretic analy- presence of heparin is not the ultimate rate that could sis of the peptidoglycan anticoagulant indicates that the be observed upon saturation of the reaction with heparin. isolated material behaves more like heparan sulfate The concentration of heparin used in this experiment than heparin. was chosen such that the rate of decrease of thrombin The molecular weight of commercially available activity could be measured. If saturating amounts of heparin is reported to be in the range of 6,000-16,000 patient proteoglycan or commercial heparin are added (21-23). The molecular weight of the reference stanclardl to the amd'unt of antithrombin III present in the reac- heparan sulfate used for electrophoretic and chromato- tion mixture, the experimental data would show inhibi- graphic comparisons can be estimated to be in the range tion of all thrombin present at the earliest time of of 30,000-40,000 based upon intrinsic viscosity data' analysis. Under conditions of the saturation of the anti- supplied with the material. Both of these materials thrombin III present either by patient-derived proteo- were isolated with harsh chemical extraction procedures glycan or by heparin, complete return to the anti- and proteolytic digestion of the animal tissue. In con- thrombin III alone inhibition rate is Observed upon trast, macromolecular heparin has been observed on administration of PF4. The rate constants derived from more gentle extraction of cultures of mast cells (24, 25), these experiments are presented in Table IV. and from rat skin (26, 27). In these cases, the major

672 M. S. Khoory, M. E. Nesheim, E. J. W. Bowie, acnd K. G. Mann17 heparin fraction obtained was resistent to protease anticoagulant activity was not elaborated in short-term digestion but could be reduced in size by alkaline cultures of her peripheral leukocytes, nor was it present hydrolysis, oxidative-reductive depolyimerization with in the extracts of homogenates of the cultured cells. ascorbic acid, or by cleavage by heparinases. The Thus, although this anticoagulant molecule was generated products obtained by the chemical treatments had mol in the blood of the individual suffering from plasma cell wt near 40,000, whereas the enzymatic digestion products leukemia, no direct connections between the disease had a mol wt of about 14,500. state and the anticoagulant molecule can be identified. Heparan sulfates are found in a variety of tissues Two potential routes by which the anticoagulant including the aorta (28, 29), and apparently are sus- molecule may have appeared in the plasma appear ceptible to change with both age (30) and disease (31), worthy of some speculation. These are: (a) that the including cellular transformation (32-34). In contrast anticoagulant molecule is a normal constituent of to heparin, which appears to be a complex macromo- plasma, ordinarily present only at very low concentra- lecular polysaccharide structure with additional protein tions, which was released 4t a greater rate as a conse- attached, heparan sulfate appears to be a proteoglycan quenice of the disease state; and (b) that tissue damage with a peptide backbone to which a series of polysac- secondary to the primary disease state has led to the charide chains is connected. Jansson and Lindahl (35) release of the proteoglycan. Heparan sulfate and re- reported the isolation of a heparan sulfate proteoglycan lated substances have been reported to be produced fromn aorta that had an apparent mol wt of 70,000 and by endothelial cells in culture (6); in addition, Muller could be cleavedl to produce polysaccharide chains with et al. (41) have reported a heparin-like inhibitor in the a mol wt of 10,000. Chiarugi and(l Urbano (36) reported blood of the newborn, and it is conceivable that the the isolation of a heparan sulfate proteoglycani for anticoagulant described in the present work is related normiial and Rotis sarcoma virus transformed baby to these materials. hampster kidney cells which was -40% protein and had a mol wt of 130,000, somewhat similar to the value ACKNOWLEDGMENTS obtained for our material. We would like to thank Dr. Jerry A. Katzmann for performing Circulating anticoagulants have been known to occur the short-term cell culture experiments described in this in a broad variety of conditions. Some of these antico- paper. We would also like to thank Dr. Matthews and Dr. agulants, such as the lupus anticoagulant, cause an Ciffonelli for providing the mucopolysaccharide reference immediate inhibition of the coagulation mechanism standards used for comparison. Finally, we would like to acknowledge Dr. Robert Kyle and Dr. Philip Greipp for whereas others are directed against specific coagula- making the plasma of this patient available for our studies. tion factors, particularly factor VIII (37). In multiple This research by grant HL-07049 from the National Heart, myeloma, however, specific abnormalities in the clotting Lung, and Blood Institute. mechanisms have not frequently been characterized. Prolongation of prothrombin and partial thromboplastin REFERENCES times have generally been attributed to nonspecific 1. Perkins, H. A., M. R. MacKenzie, and H. Fudenberg. interference of the monoclonal proteins, although in 1970. Hemostatic defects in dysproteinemias. Bloocl. 35: certain instances specific coagulation factor antibodies 5, 695-707. " have been elaborated in myelomna (1). In several studies 2. Glueck, H. E., and R. Hong. 1965. A circulating antico- (4, 5, 38), an antithrombin role has been assigned to a agulant in yl-A multiple myeloma: Its modification by in which have been penicillin.J. ClitG. Invest. 44: 1866-1881. protein the plasma may, or may not, 3. Vigliano, E. M., and H. Horowitz. 1967. Bleeding syndrome the monoclonial immunoglobulini molecule. In two in a patient with IgA myeloma: interaction of protein and studies (39, 40) heparin-like activity for an anticoagulant connective tissue. Blood. 29: 6, 823-836. was suggested but fu-rther analyses were not conducted 4. Cohen, I., J. Amir, Y. Ben-Shaul, A. Pick, and A. De Vries. to specifically identify the nature of the anticoagulant 1970. Plasma cell myeloma associated with an unusual myeloma protein causing impairmnent of fibrin aggregation activity. In the patient whose anticoagulant is the sub- and platelet function in a patient with multiple malig- ject of this investigation, the anticoagulant material nancy. Am. J. Med. 48: 766-776. appears to have no specific relationship to the immuno- 5. Lackner, H., V. Hunt, M. B. Zucker, and J. Pearson. 1970. globulin elaborated by the myeloma cell line, and, in Abnormal fibrin ultrastructure, polymerization and clot re- the disease was in remission and at traction in multiple myeloma. Br. J. Hemizatol. 18: 625- fact, patient's IgA 636. a normal level when the spontaneously occurring anti- 6. Buonassisi, V., and M. Root. 1975. Enzymatic degradation coagulant activity developed. The observation of a of heparin-related mucopolysaccharides from the surface circulating heparani sulfate proteoglyean in this patient of endothelial cell culture. Biochim. Biophyjs. Acta. 385: is provocative in view of reports suggesting differences 1-10. 7. Kraemer, P. M. 1971. Heparan-sulfates of cultured cells with respect to heparan sulf:ate in mucopolysaccharide to acid soluble and precipitable species of different cell components after cell transformation (32-34). Although lines. Biochemistri. 10: 1445-1451. this particular patient had a plasma cell leukemia, the 8. Hormer, A. A. 1975. Heparin Structure Function and

Circulating Proteoglycan Anticoagulant 673 Clinical Implications. R. A. Bradshaw and S. Wessler, 26. Homer, A. A. 1971. Macromolecular heparin from rat skin. editors. Plenum Press, New York. 85-93. Isolation, characterization and depolymerization with 9. Ruiz, C. E., D. N. Fass, V. Fuster, and K. G. Mann. 1976. ascorbate.J. Biol. Chew. 246: 231-239. Isolation and partial characterization of porcine platelet 27. Horner, A. A. 1972. Enzymic depolymerization of macro- antiheparin factor. Circulatiotn. 54(Suppl. II): 198. (Abstr. molecular heparin as a factor in control of lipoprotein 775). lipase activity. Proc. Natl. Acad. Sci. U. S. A. 69: 10. Lundblad, R., H. J. Kingdon, and K. G. Mann. 1976. 3469-3473. Proteolytic enzymies (Part B). Thrombin. Methods 28. Radhakrishnamurty, B., H. A. Ruiz, Jr., and G. S. Beren- Eniziltmol. 45: 156. son. 1977. Isolation and characterization of proteoglycans 11. Damus, P. S., and R. D. Roseniberg. 1976. Antithrombin- from bovine aorta.J. Biol. Chem. 252: 4831-4841. heparin cofactor. M\lethods Enizymt1ol. 45: 653-669. 29. Ogema, T. R., Jr., V. C. Hascall, and R. Eisenstein. 1979. 12. Minimum Requirements for Dried Thrombin. 1946. Na- Characterization of bovine aorta proteoglyean extracted tional Institutes of Health, Bethesda, Md. with guanidine hydrochloride in the presence of protease 13. Mann, K. G., C. M. Heldebrant, and D. N. Fass. 1971. inhibitors.J. Biol. Chem. 254: 1312-1318. Multiple active forms of thrombin. 1. Partial resolution 30. Kumar, V., G. S. Berenson, H. Ruiz, E. R. Dalferes, Jr., differential activities and seqtuential formations. J. Biol. and J. P. Strong. 1967. Acid mucopolysaccharides of Che1 246: 5994-6001. human aorta. 1. Variations with maturation.J. Atheroscler. 14. Bowie, E. J. W., J. H. Thompsoni, Jr., P. Didisheim, and Res. 7: 573-581. C. A. Owenl, Jr. 1971. Laboratory Manual of Hemostasis. 31. Berenson, G. S., S. R. Sriniveasan, P. J. Dolan, and B. W. B. Saunders, Philadelphia. 186. Radhakrishnamurty. 1971. Lipoprotein acid mucopolysac- 15. Yphantis, D. A. 1964. Equilibrium ultracentrifugation of charide complexes from fatty streaks of human aorta. Cir- dilute solutions. Biochem istryl. 3: 297-317. culation. 44(Suppl. II): 6. (Abstr. 20) 16. Muir, H., and T. E. Hardingham. 1975. Structure ofproteo- 32. Johnston, L. S., K. L. Keller, and J. M. Keller, 1979. glycans. In MTP International Review of Sciences, Heparin sulfates of Swiss mouse 3T3 cells. The effect of Biochemistry Series One. 5: 153-222. University Park transformation. Biochim. Biophys. Acta. 583: 81-94. Press, Baltimore, Md. 33. Smith, H. S., A. J. Hiller, E. W. Kingsbury, and C. Roberts- 17. Bitter, T., and H. M. Muir. 1962. A modified uronic acid Dory. 1973. Cell surface properties and the expression carbazole reaction. Atnal. Biochemii. 4: 330-334. of SV40 induced transformation. Nat. New Biol. 245: 18. Horner, A. A. 1967. Electrophoresis of acidic mucopoly- 6769. saccharides in agarose gel. Cant. J. Biochem. 45: 7, 34. Minnikin, S. M., and A. Allen. 1973. Cell surface mucosub- 1009-1013. stances from trypsin disaggregation of normal and virus- 19. Marbrook, J. 1967. Primary immunlliie responise in ctultures transformed lines of baby-hamster kidney cells. Biochen. of spleen cells. Latncet. II: 1279. J. 134: 1123. 20. Katzmann, J. A. 1978. Myeloma-induiced immuntiosuippres- 35. Jansson, L., and U. Lindahl. 1970. Evidence for the ex- sion: a multistep mechaniismii.J. Immunotiol. 121: 4, 1405- istence of a multichain proteoglycan of heparin sulfate. 1409. Biochen. J. 117: 699-702. 21. Roden, L., J. R. Baker, J. A. Cifonelli, and M. B. Matthews, 36. Chiarugi, V. P., and P. Urbano. 1973. Studies on cell cur- 1972. Isolation and characterization of connective tissue rent macromolecules in normal and virus transformed polvsaccharides. Methods Eniziym1ol. 28: 73-140. EHK/21C13 cells. Biochim. Biophys. Acta. 298: 195-208. 37. Feinstein, D. I., and S. I. Rappaport. 1970. Progress in 22. Ciffonelli, J. A. 1975. Relationi of chemical structure of Hemostasis and Thrombosis. T. H. Spaet, editor. Gruen heparin to its anticoagulant activity. In Heparin Structure & Stratton, Inc., New York. 75-95. Function and Clinical Implications. R. A. Bradshaw and 38. Verstraete, Par M., and C. Vermylen. 1959. Recherches S. Wessler, editors. Plenum Press, New York. pp 95-103. sur L'Antithrombine V dans la Maladie de Kahler. Acta 23. Lasker, S. E. 1975. Low miolecular weight derivatives of Haematol. 22: 240-254. heparin that is orally active inl mice. hI Heparin Structure 39. Vermnylen, C., and M. Verstraete. 1961. Antithrombin V: Function and Clinical Implications. R. A. Bradshaw and Critical evaluation of its assessment and properties. Thromb. S. Wessler, editors. Plenumii Press, New York. pp 119-130. Diath. Haemorrh. 5: 267-284. 24. Yurt, R. W., R. W. Leid, Jr., K. F. Austen, and J. E. Silbert. 40. Harbaugh, M. E., E. M. Hill, and R. B. Conn. 1975. Anti- 1977. Native heparin from rat peritoneal mast cells. J. thrombin and antithromboplastin activity accompanying Biol. Chewn. 252: 518-521. IgG myeloma. Am. J. Clin. Pathol. 63: 57-67. 25. Ogren, S., and U. Lindahl. 1975. Cleavage of macro- 41. Muller, A. D., J. M. Van Doorm, and H. C. Hemker. 1977. molecular heparin by an enzyme from mouse mastocvtoma. Heparin-like inhibitor of blood coagulation in normal J. Biol. Chewl. 250: 2690-2697. newborn. Nature (Lond.). 267: 616-617.

674 .M. S. Khooril, Ml. E. Nesheiin, E. J. WV. Bowie, anid K. G. Mannii