Isolation and Characterization of Antistasin

c 1987 by The American Society of Biological Chemists, Inc. THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 262, No. 20, Issue of July 15, pp. 9718-9723, 1987

Printed in U.S.A.

Isolation and Characterization of Antistasin AN INHIBITOR OF METASTASIS AND COAGULATION*

(Received for publication, February 19,1987)

George P. TuszynskiSB, Tatiana B. Gasicll, and Gabriel J. Gasicll From the SLankenau Medical Research Center, Philadelphia, Pennsylvania 19151 and the (Laboratory of Experimental Oncology, Pennsylvania Hospital, Philadelphia, Pennsylvania 19107

The purpose of this study was to purify and characterize the agent responsible for the antimetastatic activity of an extract of the salivary glands (SGE) of the Mexican leech Haementeria officinalis. When admin- istered intravenously in mice on the same day as the intravenous inoculation of T241 sarcoma cells, SGE markedly reduces the number and size of lung tumor colonies. In designing a purification protocol for the antimetastatic agent, we postulated that the antimetastatic agent would also display anticoagulant activity. Thus, we discovered that heparin affinity chromatography followed by anion-exchange chromatography results in a fraction highly enriched in both potent anticoagulant activity and potent antimetastatic activity. Approximately, 200-300 pg of purified protein is isolated from 150 mg of SGE. As little as 15 pg of this material inhibits tumor cell metastasis to the same extent as 1.0 mg of the unfractionated SGE. When analyzed on sodium dodecyl sulfate gels the active fraction consists mainly of one polypeptide band having an apparent molecular weight of approximately 17,000 under either reducing or nonreducing conditions. The protein has a PI of approximately 9.5 and a molecular weight of approximately 17,000 under non- denaturing conditions. A specific antiserum prepared against the 17,000-dalton protein indicated that this protein is the major anticoagulant and antimetastatic agent of leech salivary gland extract. We have termed this anticoagulant, antimetastatic agent “antistasin.” We hypothesize that antistasin inhibits coagulation via factor Xa, and not thrombin, since factor Xa, but not thrombin, is rapidly inactivated upon addition of antistasin. The mechanism of antistasin’s antimetastatic activity is currently under investigation.

The therapeutic value of leeches has been recognized since antiquity. In Europe bloodletting with Hirudo medicinalis, the well known European leech, was prescribed as treatment for such diverse afflictions as bruises and arthritis (1). The me- dicinal value of leeches has never been investigated in depth. It has, however, been recognized that leeches possess powerful anticoagulants and antiproteases since ingested blood remains liquid in their gut for weeks (2). Some of the proteins responsible for this anticoagulant activity have been isolated from

* This research was supported by Grant CA33209 from the De- partment of Health Services and a grant from the w. w. Smith Charitable Trust. The information contained in this study is pending patent approval, serial number 908,581. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

5 To whom correspondence should be addressed.

the European leech, such as the thrombin-specific inhibitor (2) hirudin, and low molecular weight antiproteases (3), the eglins. Other leech anticoagulant proteins have been identi- fied as the fibrinolytic agents hementin (4) and hementerin (51, isolated from the South American leeches Haementaria ghilianii and Haementaria lutzi Pinto 1920, respectively. Re- cently, Gasic and co-workers (6) demonstrated that extracts from the salivary glands of H. ghilianii and the Mexican leech Haementaria officinalis not only possess strong antiprotease and anticoagulant activity (6) but also contain substances that inhibit experimentally induced tumor cell metastasis in mice (7,8).

The purpose of this study was to isolate and characterize the active antimetastatic agent from one of these leech spe- cies, the Mexican leech. We found that a novel anticoagulant we have termed antistasin inhibits lung tumor colonization of mice by T241 sarcoma cells.

EXPERIMENTAL PROCEDURES

Materials-All reagents, unless specified otherwise, were purchased from Sigma. Purified factor Xa was of bovine origin. Purified human thrombin was a gift from Dr. J. W. Fenton 11, New York State Department of Health. Reagents for dodecyl sulfate-polyacrylamide gel electrophoresis were obtained from Bio-Rad. Reagents for fast protein liquid chromatography were purchased from Pharmacia P-L Biochemicals. The chromogenic substrates N-benzoyl-L-isoleucylglu- tamylglycyl-L-arginine-p-nitroanilide hydrochloride (S-2222) and H- D-phenylalanyl-L-pipecolyl-L-arginine-p-nitroanilide dihydrochlo- ride (S-2238) were purchased from Ortho Diagnostics. Syngeneic C57BL/6 or C57BL/6 X AF, mice were obtained from Jackson Laboratory. Sarcoma T241 (9), a dimethylbenzathracene-induced tumor from C57BL/6 mice was converted into an ascitic tumor and maintained by weekly transplantations into C57BL/6 mice as described previously for the ascitic form of TA, tumor cells (10).

Preparation of Salivary Gland Extract-The salivary gland extract (SGE)’ was obtained from the anterior and posterior salivary glands of the Mexican leech H. officinalis, and was prepared according to a modified procedure of Budzynski et al. (11) as described by Gasic et al. (8). Briefly, glands were extracted twice at 4 “C by sonication in a buffer of 20 mM HEPES, pH 7.8, containing 10 mM CaC12. The sonicated material is first centrifuged at 250 X g for 20 min and the supernatants pooled and centrifuged at 100,000 X g for 1 h. The resulting supernatant is frozen at -70 “C until needed.

Purification of the Antimetastatic Actiuity-The antimetastatic activity was purified by a two-step chromatographic procedure using prolongation of the prothrombin time as an indicator of the presence of antimetastatic activity. A volume of 25 ml of SGE was pumped at a flow rate of 0.5 ml/min through a 1 X 4-cm column of heparin- agarose equilibrated with 20 mM Tris-HC1, pH 8.7 (Tris buffer). Loading, washing, and elution of the heparin-agarose column were accomplished by use of the gradient maker and liquid chromatography controller (LCC-500) and UV monitoring system which are component parts of the Pharmacia fast protein liquid chromatography

’The abbreviations used are: SGE, salivary gland extracts; HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; FPLC, fast protein liquid chromatography; SDS, sodium dodecyl sulfate.

9718

Isolation and Characterization of Antistasin 9719

system (FPLC). The proteins that did not adhere to the column (breakthrough fraction) were collected and analyzed for anticoagulant and antimetastatic activity. The column was then washed with Tris buffer until no more adsorbance at 280 nm was detected in the wash buffer. The washed column was eluted by a combination of linear and step gradients produced by the gradient marker using Tris buffer and Tris buffer containing 1 M NaC1. First, the gradient marker was programmed to elute the column with a 20-ml gradient prepared from Tris buffer containing no salt and Tris buffer containing 1 M NaCl. When protein peaks began eluting the gradient was held manually and proteins allowed to elute under isocratic conditions. After each protein peak eluted the gradient was restarted. Using this procedure maximum column resolution was obtained. The column fractions containing anticoagulant activity, which in most experiments eluted at 0.55 M NaC1, were pooled, concentrated by centrifugation through Amicon ultrafiltration filters having molecular weight exclusion lim- its of 12,000, desaltedon small columns of Sephadex G-25, and applied to a Mono Q column equilibrated in Tris buffer. The Mono Q column was washed and eluted in a manner similar to that of the heparin- agarose column. The anticoagulant and antimetastatic activities usu- ally eluted at salt concentrations of 0.25 M. The purified protein was concentrated to 1 mg/ml and stored frozen at -70 "C in Tris buffer containing the NaCl concentration present in the elution buffer.

Measurement of Molecular Weight-The molecular weight of the active material was estimated by gel filtration on the FPLC and by SDS-gel electrophoresis. An aliquot of 1 ml containing 200 pg of protein was applied at a flow rate of 1 ml/min to a Superose 6B column equilibrated in Tris buffer, containing 0.55 M NaCl and calibrated with the following molecular weight standards: 45,000 (ovalbumin), 29,000 (carbonic anhydrase), 24,000 (chymotrypsinogen A), and 12,000 (cytochrome c). The molecular weight was determined from a plot of elution volume uersus log of the molecular weight as previously described (12) for gel filtration and from a plot of log molecular weight uersus distance migrated as described for SDS-gel electrophoresis (13).

Measurement of Isoelectric Point-The isoelectric point of the active material was measured by chromatofocusing on a Mono P column using the FPLC according to the instructions provided by Pharmacia for a pH 10 to 7.0 gradient.

Amino Acid and End Group Analysis-The amino acid composition of the active material was determined by the UM Protein Sequencing Facility, Biological Chemistry, University of Michigan Medical School, Ann Arbor, MI, by methods previously described (14).

Preparation of Antiserum and Antibody Absorption Experiments- Three-month-old New Zealand rabbits were immunized with purified M, = 17,000 protein isolated from SDS gels as described by the method of Knudsen (15). Antisera against the 17,000-dalton protein were tested for immunoreactivity by double immunodiffusion as previously described (16). A 10-ml aliquot of SGE was absorbed with either control serum or anti M, = 17,000 serum by passing these protein solutions over 5-ml columns of protein A-Sepharose that had previously been treated with 10 ml of antiserum and washed exten- sively with phosphate-buffered saline. Column flow rates were maintained at approximately 0.5 ml/min and all procedures performed at room temperature. Absorbed solutions were analyzed for immunoreactivity, anticoagulant activity as measured hy the prothrombin time, and tested for inhibition of experimentally induced metastasis.

Measurement of Experimentally Induced Metastasis-Inhibition of experimental metastasis was determined as previously described (7). Briefly, each mouse was injected intravenously with one-third of the protein sample or buffer 2 h prior to intravenous injection of IO5 T241 tumor cells followed by inoculation of one-third of the sample 2 and 4 h, respectively, after injection of the tumor cells. After 14-19 days, animals were killed and the lung tumor colonies counted.

Protein Assays-Protein concentrations were measured by the method of Bradford (17) using bovine serum albumin as standard.

Coagulation and Factor Xu Chromogenic Assays-One-stage clotting assays for measurement of the effect of antimetastatic activity on the prothrombin time, activated partial thromboplastin time, thrombin time, and factor X assay were performed at 37 "C in a final volume of 0.3 ml with the aid of a fibrometer essentially as previously described (18). Briefly, for the determination of the effect of antistasin on the prothrombin time, 100 p1 of a saline dilution of protein followed by 100 pl of thromboplastin reagent (Sigma) containing 0.025 M CaC12 was added sequentially to 100 p1 of normal human plasma and the clotting time determined. For the measurement of the activated partial thromboplastin time, kaolin/cephalin reagent containing 0.025 M CaCI2 was substituted for the thromboplastin reagent. For

measurement of the effect of the antimetastatic material on the thrombin time, the substrate solution was either 100 pl of normal human plasma or a 100-111 solution of 1 mg/ml of fibrinogen in HEPES-buffered saline and was clotted by the addition of 200 pl of a 0.30 unit/ml HEPES-buffered saline solution of thrombin containing various amounts of purified protein. To maximize the effect on thrombin, 10-pl volumes of protein and thrombin were first incubated together for 5 min at 37 "C before dilution and addition to the normal human plasma or fibrinogen. For determination of the effect on factor Xa activity, 100 pl of factor X-deficient plasma was treated with 100 pl of 0.025 M CaClz followed by 100 pl of purified factor Xa or 100 pl ofpurified factor Xa containing various amounts of the antimetastatic protein.

Inhibition of factor Xa activity by the purified leech protein was determined spectrophotometrically using the factor Xa-specific chromogenic substrate $3-2222 by methods similar to those previously described (19). Briefly, various molar ratios of factor Xa (1 pg/ml) and the purified leech protein (2.9-84 pg/ml) were incubated for 15 min at room temperature in 50 mM Tris-C1, pH 8.3, containing 0.1% bovine serum albumin and 2.5 mM CaC12. After the incubation period, S-2222 was added to a final concentration of 0.063 mg/ml and the reaction mixture monitored at 405 nm for production ofpara-nitroan- iline. Inhibition of thrombin activity was measured in a similar manner except that the thrombin specific-substrate S-2238 was substituted for S-2222.

Gel Electrophoresis-SDS-polyacrylamide gel electrophoresis was performed as previously described (13).

RESULTS

Purification of the Antimetastatic Actiuity-We hypothe- sized that protein fractions possessing anticoagulant activity would also possess antimetastatic activity. This hypothesis was based on preliminary observations from early attempts at purification of the antimetastatic agent which showed that fractions possessing antimetastatic activity also displayed anticoagulant activity (data not shown). Antimetastatic activity was purified by a combination of heparin-agarose chromatography followed by anion-exchange chromatography using FPLC. Column fractions were rapidly assayed for anticoagulant activity as determined by prolongation of the prothrombin time. All the anticoagulant as well as all the antimetastatic activity of SGE was quantitatively adsorbed on heparin-agarose. The adsorbed anticoagulant activity was eluted in one peak at approximately 0.55 M NaCl and did not co-elute with the major protein peaks (Fig. 1). Further frac- tionation of the active material on a Mono Q column resulted in the elution of the majority of both the anticoagulant and antimetastatic activity at approximately 0.25 M NaCl (Fig. 2). In a typical purification protocol, we are able to purify 200- 400 pg of protein from 25 ml of SGE. As little as 15 pg of purified protein quantitatively inhibits tumor cell metastasis whereas 1.0 mg of unfractionated SGE is required to achieve the same effect (Table I).

To rule out the possibility that small amounts of heparin that may have artifactually co-eluted with the M , = 17,000 protein in the first step of the purification procedure contrib- uted significantly to the antimetastatic activity of the final purified protein preparation, purified M , = 17,000 protein, a buffer control obtained from a heparin-agarose column that was developed in the absence of leech protein, and the former two samples containing 2 units of heparin were tested in the metastasis assay (Table I). The number of lung tumor colonies obtained with either the buffer control, saline control, or buffer control plus heparin were not statistically different from one another. In addition, heparin had no effect on the antimetastatic activity of the M , = 17,000 protein. These results indicate that heparin does not contribute to the antimetastatic activity of the M , = 17,000 protein.

Characterization of the Antimetastatic Activity-When the active material was analyzed on SDS gels, a major protein

9720 Isolation and Characi

” 2.0 11

0’3 1 0.2

O ’ l I 0

t I

Fraction number

FIG. 1. Heparin-agarose column elution profile. The elution profile was obtained after adsorption and NaCl elution of 25 ml of SGE from a heparin-agarose column as described under “Experimental Procedures.” One-ml fractions were collected. The clotting time measurements were performed on 5 pl of the corresponding column fractions according to the procedure described for the prothrombin time. The clotting times are prolongation of the prothrombin time over the buffer blank.

0.05 1

Fraction number

FIG. 2. Mono Q column elution profile. The elution profile was obtained after adsorption and NaCl elution of the anticoagulant peak obtained in the experiment described in the legend to Fig. 1 and under “Experimental Procedures.” One-ml fractions were collected. The clotting time measurements were performed on 5 pl of the corresponding column fractions according to the procedure described for the prothrombin time. The clotting times are prolongation of the prothrombin time over the buffer blank.

band migrating with apparent M , of approximately 17,000 under both reducing and nonreducing conditions was obtained (Fig. 3, lane 2 ) . This band was selectively adsorbed on heparin-agarose since little or no protein corresponding to the

:erization of Antistasin TABLE I

Effect of purified M, = 17,000 protein and heparin on lung tumor colony (LTC) formation in mice

The data denoted as Experiment 1 are combined from five separate experiments. Buffer denotes the saline control in the absence of the M, = 17,000 protein. The data from all groups were analyzed using a median test (Chi square equals 57.36 and t less than 0.001). Each group was compared with the other. The p values of either SGE or the M, = 17,000 protein-treated groups compared to the control group are both < 0.0001. The p value of the SGE group compared to the M, = 17,000 protein-treated group is > 0.5. In Experiment 2, buffer controls are 0.35 M NaCI, Tris buffer heparin-agarose eluants. Two units of heparin were added to the samples containing heparin. Experimentally induced lung tumor colony formation was measured as described under “Experimental Procedures.” -

Sample Amount Tumor bearing

injected mice/total LTC/mouse

number of mice median (range)

(MI m i s e )

Experiment 1 Buffer 0 24/24 18 (5-162) SGE 1000 15/25 1 (0-9) M, = 17,000 protein 15-75 17/30 1 (0-9)

Experiment 2 Buffer from heparin- 4/4 13 (10-28)

agarose

agarose + heparin Buffer from heparin- 4/4 13 (5-21)

M, = 17,000 protein 15 2/4 2 (0-3) M, = 17,000 + heparin 15 2/4 -

4 (2-6)

MW X

200 - 1 1 6 2 5 96

- - .. -

- 45 - 31 - a

r , * -

1 2 3 4 5 FIG. 3. SDS gel of the M. = 17,000 protein, SGE, and pro-

tein fractions from the heparin-agarose column. Ten pg of purified protein and 50 pg of protein mixtures were analyzed on 12% polyacrylamide slab gels under nonreducing conditions. Lune I , molecular weight standards; lane 2, purified M, = 17,000 protein; lane 3, proteins not adhering to heparin-agarose; lane 4, proteins eluted in fractions 10-20 of heparin-agarose; lane 5, SGE.

17,000-dalton protein was observed in the non-adsorbed heparin-agarose fraction (Fig. 3, lane 3 ) . In addition, proteins eluting from the heparin-agarose column that possessed no anticoagulant activity also were devoid of the M , = 17,000 protein (Fig. 3, lane 4 ) . Analysis by gel filtration on Superose 6 FPLC column resulted in one peak of anticoagulant activity at an M , of 17,000 (Fig. 4). Analysis by column chromatofocusing revealed a single peak of anticoagulant activity at an isoelectric point of 9.5 (Fig. 5). Amino acid analysis shows that the purified protein is rich in arginine and lysine residues, which is consistent with a basic PI (Table 11). NH2-terminal analysis revealed that the terminal residue is blocked. These results indicate that the antimetastatic agent is an M , = 17,000 single-chain protein with a PI of 9.5.

Anticoagulant Activity-The antimetastatic protein is a potent anticoagulant of human and mouse plasma. Our results show that both the extrinsic and intrinsic pathways are inhibited as measured by the prolongation of the prothrombin

Isolation and Characterization of Antistasin 9721

TABLE I1 Amino acid composition of the antimetastatic and anticoagulant agent

Fraction number FIG. 4. Gel filtration profile. The purified protein (200 pg) was

gel-filtered on a preparative Superose 6B column as described under "Experimental Procedures." Ten-pl fractions were analyzed for clotting activity as described for the prothrombin time. The inset shows an SDS gel of the peak fraction. Samples (50 pl) were analyzed under nonreducing conditions on 15% polyacrylamide slab gel.

r0.15

0 0 , 4 2 . l , !,L"" , e"- , , , , , , , 1 0 10 20 30

Fraction number

FIG. 5. Chromatofocusing column elution profile. One hundred-fifty pg of protein obtained from the Mono Q column were chromatographed on a Mono P column after dialysis against 25 mM ethanolamine-HCI buffer, pH 9.7. Ten-pl fractions were analyzed for clotting activity as described for the prothrombin time.

and activated partial thromboplastin times, respectively. For example, 2 pg of protein prolongs the prothrombin time by 70 s and the activated partial thromboplastin time by 56.5 s. This effect is not due to the inhibition of thrombin directly since the purified protein has no effect on the clotting time of thrombin whether or not the substrate is fibrinogen or plasma. For example, the clotting time of 0.10 units of thrombin in the presence of 2.42 pg of protein is 31.4 s, whereas in the absence of protein a clotting time of 33.9 s was obtained. In addition, no inhibition of thrombin activity was observed when thrombin activity was measured using the chromogenic substrate S-2238. I t seems likely, however, that the anticoagulant effect is due to inhibition of factor Xa since the purified protein is a potent inhibitor of purified factor Xa as

464 ne of Drotein were analmed. Amino acid Residues per M, = 17,000

Asx Thr Ser Glx Pro GlY Ala CYS Val Met Ile Leu TYr Phe His LYS Arz

11.1 8.9 8.7 13.5 12.2 17.4

ND" 6.5

4.3 2.2 4.3 6.3 29.3 4.5 3.3 3.9 12.6

' ND, not determined.

I 0.24 1

8 0.18 t /

L / /

OO j l l l l l l l l l l

2 4 6 8 10 Minutes

FIG. 6. Inactivation of factor Xa activity. The inhibition of factor Xa activity was measured spectrophotometrically at 405 nm by recording the rate of the factor Xa-catalyzed cleavage of S-2222 in the presence and absence of the purified leech protein designated as A. The assay is described under "Experimental Procedures."

indicated by a factor Xa clotting assay and by a spectropho- tometric assay using the factor Xa-specific chromogenic substrate s-2222 (Fig. 6). For example, 2.42 pg of protein com- pletely inhibits the clotting activity of 0.30 units of factor Xa in factor X-deficient plasma and incubation of factor Xa with stoichiometric amounts of protein results in more than a 50% inhibition of the rate of the factor Xa-catalyzed cleavage of s-2222.

Anticoagulant and Antimetastatic Activities of the M, = 17,000 Protein-To determine whether the 17,000-dalton protein is the major antimetastatic and anticoagulant protein of SGE, antibodies were raised against the M , = 17,000 polypeptide and antibody-adsorbed SGE was tested for anticoagulant and antimetastatic activity. Antibody-adsorbed SGE possessed no 17,OO-dalton antigen (Fig. 7) nor did it contain anticoagulant and antimetastatic activity (Fig. 8). In addition, fractions from the heparin-agarose (Fig. 1) column that contained no M, = 17,000 antigen or anticoagulant activity (non- adsorbed fraction, fractions 3-10, and fractions 10-18) possessed no antimetastatic activity (Figs. 7 and 8). These results indicate that the M , = 17,000 protein is the major component of SGE and possesses both anticoagulant and antimetastatic activity.

DISCUSSION

Since it has been known for at least 100 years that cancer patients have abnormalities of blood coagulation (20) and

9722 Isolation and Characterization of Antistasin

I

FIG. 7. Immunoreactivity of ant i M , = 17.000 protein antiserum. Double immunodiffusion was performed as previously described (16). Samples wells were loaded with 10 pg of protein and the central well contained 8 pl of antiserum. After development of pre- cipitant lines, the plate was washed in phosphate-buffered saline and photographed. Well la, anti M , = 17,000 protein antiserum; Wells I - 6 are salivary gland extract, purified antimetastatic protein, antibody- absorbed salivary gland extract, protein that did not absorb to heparin-agarose, heparin-agarose peak I (fractions 3-10, Fig. l), and heparin-agarose peak I1 (fractions 10-18; Fig. l), respectively. Anti- body absorption of salivary gland extract is described under “Exper- imental Procedures.”

Control A Adsorbed aalivary

Non-SIICk gland extract

heparin-agarose Anlistasln Salivary gland extract

> m

FIG. 8. Anticoagulant and antimetastatic activity. Both purified fractions and SGE were tested for anticoagulant activity (5 pg each) as measured by the prothrombin time and antimetastatic activity as described in the legend of Table I. Three groups of six mice/ group were each injected with 335 pg of protein obtained from SGE adsorbed with control serum (designated control), with SGE absorbed with anti M , = 17,000 protein antibody (designated adsorbed salivary gland extract), and salivary gland extract. Seven mice were each injected with 17 pg each of purified antimetastatic protein (designated antistasin). Three mice were each injected with 335 pg of protein that did not adsorb to heparin-agarose (designated non-adsorbed heparin- agarose). The results for the metastasis assay are presented as the median and the range shown by the error bars. The ordinate scale for lung tumor colonies/mouse is the same as the scale for clotting time. The procedure for antibody absorption of SGE is described under “Experimental Procedures.”

more recently it has been recognized that the activation of the hemostatic system by tumors may contribute to their invasive nature (reviewed in Refs. 21 and 22), we decided to use inhibition of the plasma prothrombin time as the functional assay for inhibition of tumor cell metastasis in our purification scheme. In early attempts to purify the antimetastatic agent by alkaline gel electrophoresis we consistently found that antimetastasis activity co-isolated with anticoagulant activity. The material migrated on alkaline gels toward the top of the gels suggesting that its PI is high. Our initial

purification procedure lacked a step that would concentrate the material sufficiently to allow for further purification. We investigated the possibility that lectin affinity chromatography might be useful in purification but found that the antimetastatic activity did not bind to wheat germ agglutinin, Lens culinaris lectin, or concanavalin A. The only chromatographic matrix that quantitatively adsorbed both the anticoagulant and antimetastatic activities present in SGE was discovered to be heparin-agarose. This activity could be eluted with high salt and further purified by anion exchange. Hepa- rin-agarose chromatography enabled us not only to concentrate the active material in one step but also purify the anticoagulant activity by a t least an order of magnitude.

The antimetastatic agent has been purified and termed antistasin. Antistasin analyzes on SDS gels and by gel filtration as a major protein band having M , = 17,000 under both reducing and nonreducing conditions. Analysis by chromatofocusing reveals one major band having PI = 9.5. Our inter- pretation of these results is that antistasin most likely exists as a non-disulfide-linked M, single-chain protein. In addition, the subunits of antistasin may contain intrachain disulfide bonding since upon reduction the mobility of the subunits decreases slightly on SDS gels.

Antistasin is a protein that possess both anticoagulant and antimetastatic activity as confirmed by antibody absorption experiments using an antiserum raised against the M , = 17,000 polypeptide. The anticoagulant activity is at least in part due to its inhibitory effect on factor Xa. Surprisingly, antistasin had no effect on thrombin activity. In this respect antistasin differs from other naturally occurring inhibitors such as antithrombin I11 and protease nexin (23). I t is tempt- ing to speculate on the mechanism of antistasin’s antimetastatic activity. Since our data show that coagulation of human and mouse plasma is potently inhibited by low levels of antistasin, it seems unlikely that the antimetastatic effects of antistasin can be achieved in vivo without some concomitant anticoagulant effects. Further experiments are required to determine if antistasin inhibits metastasis formation through its effect on coagulation. It is well known that tumor cells contain a procoagulant that can directly activate factor Xa (24) and it has been argued that fibrin deposition promotes formation of tumor fibrin emboli that lodge in the vascular bed and promotes tumor cell sequestration important in the metastatic spread of tumor cells (25). In addition, fibrin is thought to protect tumor cells against destruction by natural killer cells by providing the tumor cells with a protective covering impermeable to the immune system (26). Therefore, antistasin might inhibit fibrin formation by its inhibition of factor Xa and prevent metastasis formation. The action of antistasin may be similar to the effect of other anticoagulants (27, 28) and protease inhibitors (29, 30) that have been reported to decrease tumor dissemination and metastasis. However, in contrast to the other agents that have been described, antistasin is more specific and may act directly on the tumor-catalyzed pathway of fibrin deposition.

Further studies are in progress to ascertain whether the anticoagulant properties of antistasin are necessary for its antimetastatic activity and to assess the efficacy of antistasin as a therapeutic agent for the treatment of spontaneous metastasis in an animal model. We are currently also analyz- ing the functional and biochemical properties of this interest- ing protein with the long range goal of defining the role of the hemostatic system in the blood-born dissemination of cancer.

Acknowledgments-We are grateful for the excellent technical as- sistance of Marion Berenfeld, Vicki Rothman, and Andew Murphy. We thank Daniel Zelac and James Bradley for their help in the

Isolation and Charaderization of Antistasin 9723

purification procedure and Dr. Karen Knudsen for her excellent Cierniewski, C. S., and Niewiarowski, S. (1985) J. Biol. Chem. review of the manuscript. We also wish to thank Edward G r a d y for 260,12240-12245 performing the statistical analysis. 17. Bradford, M. M. (1976) Anal. Biochem. 72 , 248-254

18. Biggs, R. (1972) Human Blood Coagulation, Haemostasis, and REFERENCES Thrombosis, pp. 594-595, Blackwell Scientific Publications,

1. Haycraft, J. B. (1984) ProC. R. SOC. L ~ n d . BBiOl. SCi. 3 6 9 478 19. Tuszynski, G. P., Mauco, G. P., Koshy, A., Schick, P. K., and Oxford

2. Bagdy, D., Barabas, E., Graf, L., Petersen, T. E., and Magnusson, S. (1976) Methods Enzymol. 45,669-678

Walsh, P. N. (1984) J. Biol. Chem. 259,6947-6951

3. Seemuller, u., Meier, M., Ohlsson, K., Muller, H.-P., and F,ritz, 21. Gasic, G. J., Tuszynski, G. P., and Gorelik, E. (1986) The Znter- H. (1977) Hoppe-Seyler's 2. Physiol. Chem. 358, 1105-1117 action of the Hemostatic and Immune Systems in the Metastatic

4. Malinconico, S. M., Katz, J. B., and Budzynski, A. Z. ( 1 W ) J. Spread of Tumor Cells in the International Review of Experi- Lab. Clin. Med. 103 , 44-58 mental Pathology, Vol. 29, pp. 173-208, Academic Press, Or-

5. Kelen, E. M. A., and Rosenfeld, G. (1975) Haemostasis 4,51-64 lando, FL 6. Murer, E. H., James, H. L., Budzynski, A. Z., Malinconico, 6. M., 22. Gasic, G. J. (1984) Cancer Metastasis R ~ ~ , 3,99-114

and Gasic, G. J. (1984) Thromb. Haemostasis 51, 24-26 23. Scott, R. W., Bergman, B. L., Bajpai, A., Hersh, R. T., Rodriguez, 7. Gasic, G. J., Viner, E. D., Budzynski, A. Z., and Gasic, G. P. H., Jones, B. N., Barreda, C., Watts, S., and Baker, J . B. (1985)

(1983) Cancer Res. 43,1633-1636 J. Biol. Chem. 260,7029-7034 8. Gasic, G. J., Iwakawa, A., Gasic, T. B., Viner, E. D., and Milas, 24. Falanga, A., and Gordon, S. G. (1985) Biochemistry 24, 5558-

L. (1984) Cancer Res. 44, 5670-5676 5567

and Shafie, S. (1980) Nature 284,67-68 10. Gasic, G. J., Gasic, T. B., and Stewart, C. C. (1968) Proc. Natl.

26. Gorelik, E., Wiltrout, R. H., Okumura, K., Habu, S., and Herber-

Acud. Sci. U. S. A . 61,46-52 man, R. (1982) Int. J . Cancer 30,107-112

27. Hoover, H. C., and Ketcham, A. S. (1975) Cancer (Phila.) 3 5 , 4 - 11. Budzynski, A. Z., Olexa, S. A., Brizuela, B. S., Sawyer, R. T., and 14

Stent, G. s. (1981) Proc. Soc. Exp. Biol. Med. 168 ,266275 28. Zacharski, L. R. (1981) in Malignancy and the Hemostatic system 12. Andrews, P. (1964) Biochem. J. 9 1 , 222-233 (Donati, M. B., Davidson, J. F., and Garattini, S., eds) pp. 113- 13. Tuszynski, G. P., Buck, C. A., and Warren, L. (1979) Anal. Bioch. 127, Raven Press, New York

14. Tarr, G. E. (1986) in Microcharacterization of Polypeptides: A Eur. J . Cancer 16,449-454

15. Knudsen, K. A. (1985) Anal. Bioch. 147,285-288 16. Tuszynski, G. P., Srivastava, S., Switalska, H. I., Holt, J. C., 2539-2542

20. Trosseau, A. (1865) New Sydenhum. SOC. 3, 94

9. Liotta, L. A., Trygpason, K., Garbisa, S., Hart, I., Folk, c. M., 25. Wood, S., Jr. (1974) J. Med, &en. 5 , 7-22

93,329-338 29. Giraldi, T., Sava, G., Kopitar, M., Brzin, J., and Turk, V. (1980)

Practical Manual, pp 155-194, Humana Press Inc., Clifton, NJ 30. Saito, D., Sawamura, M., Umezawa, K., Kanai, Y., Furihata, C., Matsushima, T., and Sugimura, T. (1980) Cancer Res. 40,

Documents

Isolation and Characterization of Antistasin