CA2064585C - Inhibition of tumor growth by blockade of the protein c system - Google Patents

Abstract

Composition, and methods of use thereof, for the inhibition of tumor growth and killing of tumors having extensive microcirculation wherein the active agent is a compound blocking the protein C s ystem, preferably anti-protein C antibody, anti-protein S antibody and C4b binding protein. In the most preferred embodiment, the protei n C blocking compound is provided in combination with a cytokine such as tumor necrosis factor (TNF), gamma interfero n, interleukin-1, interleukin-2 and granulo-cytemacrophage colony stimulating factor. Example are provided demonstra ting the administration of the protein C blocking compound, alone or in combination with TNF, to dogs having canine venereal tumor s, a fibrosarcoma, and an adenocarcinoma, to pigs with melanoma, and a primate with a malignant lymphoma, followed by sign ificant tumor reduction.

Description

20615~
W O 91/01753 _ 1 - PC~r/US90/04337 I~n~IBIl~O N O F llU~IO R G R O~VllI BY
BLOCKADE OF THE PR O TEIN C SYSTE M
Background of the Invention ~ This is generally in the area of compositions for treatment of cancer, in particular, compositions cont~ining blockers of the protein C system in combination with a lymphokine.
A variety of mech~nicmc in tumors capable of promoting 5 clot formation have been described (Dvorak, H.F. Human Path.
18,275-284 (1987); Rickles, F.R., Hancock, W.W., Edwards, R.L., et al. Sem.Thromb.Hemost. 14,88-94 (1988)). Initially, the discovery of extravascular fibrin deposits in a variety of animal and human tumors ~lol,lpled the search for these tumor-associated clotting mesh~nicmc.
10 This extravascular fibrin disposition has been found in association with prothrombin, factor VII and factor X in certain tumor cells in situ by immllnohistochemical techniques. A heat- and acid-stable glycoprotein present in mucin produced by certain adenocarcinomas is capable of catalyzing the collvel~ion of factor X to factor Xa. A
68,000 dalton cysteine protease has been identified in a number of tumor lines which activates factor X, independent of the actions of factor VII, and tissue factor. Some tumor homogenates display tissue factor activity, while other tumor cell lines have cell membranes with receptors for factor Va and are capable of catalyzing the conversion of 20 prothrombin to thrombin. Abnormalities in co~ tion parameters observed in certain cancer patients has plolllpled the hypothesis that tumor-associated clotting may be of sufficient magnitude to cause systemic activation of the clotting system.
Alterations in the fibrinolytic system are observed in 25 tumors and Llansrolllled cells. Pl~cminogen activator activity has been found to be higher in extracts of surgically excised human cancer tissues than in surrounding benign tissue. In addition to other possible roles for plasminogen activator, such as participation in tumor invasion of normal tissue, this suggests that tumors have the capacity 30 to promote fibrin degradation. Urokinase-type plasminogen activator appears to be produced by tumors with greater frequency than tissue-wo 9l/01753 Pcr/US9O/o4337 206~ 2-type pl~cminogen activator. It is not known if the production of pl~cminogen activators by tumors is involved in preventing thrombus ~c~lmlll~tion within the bed~of the tumor.
Protein C is a vitamin K-dependent plasma protein.
5 Activated protein C serves as a natural antiCQ~ nt by inhibiting the clotting c~cc~de at the levels of factor V~ and factor VIIId (Walker, FJ., Sexton, P.W. and Esmon, C.T. Biochim.Biophys.Acta 571,333-342 (1979); Fulcher, C.A., Gardiner, J.E., Griffin, J.H., et al. Blood 63,86-49 (1984)). Protein C is rapidly convel~ed to activated protein C by a 10 complex of thrombin and the endothelial cell surface protein, thrombomodulin. Thrombomodulin forms a l:l stoichiometric complex with thrombin and increases the rate at which thrombin activates protein C by apl)~o~;...~tely 20,000-fold. This activation occurs primarily in the capillaries, where the availability of a large endothelial surface area per unit of plasma volume favors complex formation between thrombin and thrombomodulin. The activation of protein C in the microvasc~ tllre forms a potential feedback loop which, when thrombin is formed in the circulation, generates activated protein C. The activated protein C in turn inhibits further thrombin formation. This has been demonstrated directly in dogs, where low level inll~vel,ous thrombin infilcjon results in the generation of activated protein C and antico~ll~tio~ of the znim~l A more complete review of the roles of thrombomodulin and protein C in regulation of blood co~ tion is by C. T. Esmon, in J. Biol. Chem.
264(9), 4743-4746 (1989).
Thrombomodulin has been id~ntified on a variety of cultured endothelial cell lines and on the l~lmin~l surface of blood vessels (F,smon, N.L. Semin.Thromb.Hemost. 13,454463 (1987)).
Thrombomodulin has also been identified by functional and immlm( chemical means on tumor cells, inchltling human lung carcinoma -line CL-185 and Bowes melanoma cells (Marks, C.A., Bank, N.U., Mattler, L.E., et al. Thromb.Hemost. 54,119 (1985)), 2 0 6 4 ~ 8 ~ PCI /US90/W337 -3- ~ ~

A549 human lung cancer cells (Maluyallla, I. and Majerus, P.W.
Blood 69,1481-144 (1987)), and angiosarcomas (Yonezawa, S., Maluyallla, I., Sakae, K, et al. Am.J.Clin.Pathol. 88,405-11 (1987)).
The functional role of thrombomodulin on the tumor cells is subject 5 to speclll~tion Tumor associated thrombomodulin may help to protect the tumor from excess fibrin formation. If the protein C-thrombomodulin system is involved in detellllirullg the hemostatic balance in the tumor vasc~ lre, blocking protein C activation may shift the hemost~tic balance and result in thrombosis of tumor vessels.
In addition to functioning as an anticoagulant, activated protein C promotes fibrinolysis. This action involves complex formation between activated protein C and two inhibitors of pl~cminogen activator, pl~cminogen activator inhibitor 1 (PAI-1) and pl~cminogen activator inhibitor 3 (PAI-3). Complex formation between activated protein C and PAI-1 or PAI-3 may serve to protect pl~cminogen acliv~tol~ from inhibition and can thus potentially result in an increase in fibrinolytic activity. The extent to which in vivo protein C influences fibrinolytic activity in the body or in tumor beds in particular is unknown.
Protein S, another vital~lin K-dependent placma protein, serves as a cofactor for the antico~ nt and fibrinolytic effects of activated protein C (Walker, FJ. Semin. Thromb. Hemost. 10,131-138 (1984); de Fouw, NJ., Haverkate, F., Bertina, R.M., et al. Blood 67, 1189-1192 (1986). Protein S exists in two forms in plasma. Forty percent of the protein S is free and serves as a cofactor for activated protein C, and 605~o is in complex with C4b binding protein and is functionally inactive. C4b binding protein is an acute phase protein (Boerger, L.M., Morris, P.C., Thurnau, G.R., et al. Blood 69,692-694 (1987); Dahlblaclc, B. J.Biol.Chem. 261,12022-12027 (1986)) and thus infl~mm~tion, by elevating the levels of C4 binding protein, may shift protein S to the inactive form by the law of mass action. This shift in protein S status may predispose to thrombosis. Hereditary protein S

W O 91/01753 ~ PCT/US90/04337 ~5 -4-deficiency, at least in some kindreds, is linked to an increased risk of venous thrombosis. See, for example, P. C. Comp, et al., J. Clin.
Invest. 74, 2082-2088 (1984).
While hetero~ygous protein C deficiency in certain kindreds 5 is also associated with an increased risk of venous thrombosis, two other protein C deficiency states are characterized by tissue necrosis:
homozygous protein C deficiency and the coumarin-in~luced skin necrosis obsened in hetero~ygous protein C deficient individuals after the initiation of coumarin therapy. In homozygous deficient 10 individuals, skin necrosis occurs on the first or second day of life, reslllting in a clinical condition termed pUl~JUI~ lmin~n~ neonatalis.
Thrombosis of the small vessels of the skin is characteristic, leading to loss of large areas of skin, which is often fatal. Major vessel thrombosis is also possible.
When oral antico~ nt therapy is initiated in hetero~ygous protein C deficient patients, extensive microvascular thrombosis may occur. The skin is again the primary target and extensive loss of skin and underlying tissue can occur. The post~ ted mech~nicm is a rapid fall in protein C levels in the heterozygous deficient individuals before the levels of clotting factors with a long half-life, such as ~rolhrolllbin and factor IX decrease to levels adequate for systemic antico~ tion. A transient hypercoagulable state may exist in these individuals and it is during this period that the tissue necrosis occurs.
Hetero~y~ous protein C deficiency is relatively common in the population and may occur as frequently as 1 in 300 individuals.
Coumarin necrosis is rare and this suggests that factors other than protein C deficiency alone must be present. A review of the literature indicates that most patients developing necrosis have some infl~mm~tory condition as well such as an infection, recent surgery or extensive venous thrombi. These accompanying infl~mm~tory changes could decrease thrombomodulin ~-ession, increase tissue factor ~o 9l/01753 2 0 6 ~ 5 ~/US90/04337 e~l~ression and favor a shift of free protein S to the inactive C4b binding protein-protein S complex. These events would further down re~ se the protein C system and promote microvascular clot - formation, suggesting that protein C deficiency accompanied by 5 infl~ tion can result in tissue necrosis.
Although spontaneous regression of solid tumors can occur on rare occasions following febrile illnesses, Dr. William B. Coley demonstrated in Annals of Surgery 14,199-220 (1891) that regression of certain solid tumors in hllm~n~ could follow the ~flmini~tration of 10 heat killed bacteria. The response of the tumors was highly variable and the di~gnosic of the tumor type was not always made by histologic eY~min~tiQn However, some of the patients treated with Coley's toxins had long term tumor regression and possible cure. Coley's work and that of other investi~tors testing the effects of bacterial 15 products on animal tumors, led to the discovery of tumor necrosis factor, a 157 amino acid molecule capable of r~lsin~ tumor necrosis in a number of murine tumors (Old, LJ. Nature 330,602-603 (1987);
Old, LJ. Scientific American 258,59-75 (1988); Gifford, G.E. and Flick, D.A. Tumor Necrosis Factor and Related Cytotoxins. edited by 20 Bock, G. and Marsh, J. Chilchester, p. 3-20 (John Wiley and Sons, 1987). Tumor necrosis factor is produced by macrophages in response to endotoxin. Tumor ncwosis factor triggers a number of physiologic responses on skeletal muscle, adipose tissue, endothelium, cartilage, leukocytes and the hypoth~l~mu~, and has a direct cytotoxic effect on 25 some tumor cell lines (Tracey, K J., Lowry, S.F. and Cerami, ~
Tumour Necrosis Factor and Related Cytokines, p. 88-108 (1987).
Tumor necrosis factor causes infl~mm~tory changes at the endothelial level by increasing the adhesion of PMNs, blood monocytes and related leukocyte cell lines. This may result from the 30 tumor necrosis factor induced production of endothelial-leukocyte adhesion molecules (E-LAMs) by the endothelium (Pober, J.S., Bevilacqua, M.P., Mendrick, D.L., et al. J.Immunol. 136,1680-1687 ~, ~

wo 91/017S3 Pcr/US9o/o4337 (1986); Pober, J.S., Lapierrc, L~, Stolpen, A.H., et al. J.lmn~lln-)l.
138,3319-3324 (1987); Bevilacqua, M.P. and Gimbrone Jr., M~
Sem Th-ornh~Iemoct 13,425433 (1987). Tumor necrosis factor also stimul~tcs en~olhclial cell production of platelet activating factor S which may also ~lol~ote microvascular throl,.bosis by platelet activation and activation of adherent polyrnorphonuclear leukocytes (Pobcr, J.S. Tllmf~ur Necrr~cic Factor ~ntl Rel~ted Cytokines p. 170-184 (1987).
The effects of tumor necrosis factor on the endothe~ lm in~lude increased tissue factor activity and decreased thrombomodulin eA~)ression (Moore, K, Esmon, C.T., and Esmon, N.L, et al. Blood 73, 159-165 (1989). Tumor .lecro~c factor and interleukin-l can dec~ease the prodllction by human cultured umbilical vein endotbelial celLc of tiscue-type pl~cminogen activator and increace the production of p1~cmino~cn activator inhibitor type 1 (schleef, R.R., Bevilacqua, M P., Sawdcy, Mr., et al. J Riol.Chem 263,5797-5803 (1988);
N;~ m~n R.L, HaDar, KA., Silverstein, R.L, et al. J.Exp.Med.
163,1595-1600 (1986); Emeis, JJ. and Kooistra, T. J.Exp.Med.
163,1260-1266 (1980); Bevilacqua, M.P., Scbleef, R.R., Gimbrone, M~, Jr., et al. J.Clin.Invest. 78,587-591 (1986). These fin~lingc, coupled witb the rnicrovascular thrombosis observed in certain protein C ~efi~ency states, su~est that ~C~ t~neolts ~dminictration of tNmor ncc.osis factor and inhibition of the protein C system u~ vivo may rcsult in i..~,.~scular thrombosis.
There is evidence that activated protein C influences the production of ~NF in the intact anim~l Tbe ~dn.;ni~l.alion of activated protein C protects ~ainst shock induced by the infusion of E. coli, as described in U.S. Patent No. 5,009,889 entitled "Treatment of Dysfunctienal VascNlar Endothelium Using Activated Protein C' issued April 23, 1991, by Fletcher B. Taylor ~r and Charles T.
Esmon. In thc ~nim~lc receiving activated protein C, tbe production of tumor ncciosis factor in resyol~se to the E. coli is markedly W O 91/01753 PC~r/US90/04337 2064;)8~

reduced, raising the possibility that blockade of protein C activation in vivo could result in enhanced production of TNF by macrophages or natural killer cells in the tumor bed in the presence of low levels of endotoxin. This could in turn contribute to further toxic effects on the 5 tumor.
p~ ino, et al., in J.~mm~mol. 138,4023-4032 (1987);
Tumour Necrosis Factor and Related Cytokines. p. 21-38 (1987); has proposed that injection of tumor necrosis factor into Meth A sarcoma bearing mice actually causes a series of events which result in tumor 10 rejection: 1) hemorrhagic tumor necrosis, initiated in the first one to four hours involving PMN activation and their localization to the tumor; 2) direct cytostatic/cytotoxic effects on the tumor as growth stops, at 24 to 72 hours; and 3) a specific T-cell mediated imm~-ne response to the tumor, at two to four weeks. Blockade of the protein 15 C system may potentiate certain aspects of the TNF mefli~ted necrosis in addition to promoting microvascular thrombosis. Blockade of the protein C system should result in increased thrombin generation by tissue factor at the endothelial surface. The thrombin which is produced can increase endothelial cell production of platelet 20 activating factor, which would prime margin~ted granulocytes and thus enh~nce the PMN medi~ted endothelial cell injury.
Antico~ ntc block the Sl-w~ llan reaction, presumably by ~revell~ing fibrin deposition and microvascular thrombosis (Edwards, R.L and Rickles, F.R. Science 200,541-543 (1978). If 25 tumor killing by tumor necrosis factor does have characteristics in common with the Shwal ~lllan reaction, blocking protein C activation could allow intravascular thrombosis to proceed unimpeded and increase the extent and severity of damage to the microv~cc~ tnre, and thus increase tumor killing. Since the use of tumor necrosis 30 factor in the treatment of human m~lign~ncies is accompanied by serious side effects, protein C blockade, which may potentiate the WO 91/01753 Pcr/usso/o4337 ~,~6~' ~ -8- ~

tumor-directed effects of a given dose of tumor necrosis factor, should be worth investigation.
It is therefore an object of the present invention to provide a method and compositions to block the natural anticoagulant 5 pathways, to thereby promote microvascular co~ ion in the new capillaries growing into tumors and by so doing block the process of angiogenesis.
It is a further object of the present invention to provide a method and compositions to promote an immune response against 10 tumors by blocking the natural antico~ nt pathways.

S~ y of the Invention Compositions, and methods of use thereof, for the inhibition of tumor growth and killing of tumors having extensive microcirculation wherein the active agent is a compound blocking the 15 protein C system, preferably anti-protein C or anti-activated protein C
antibody, anti-protein S antibody, inactivated activated protein C
(inactivated APC), or C4b binding protein, ~Aminictered systemically or at the site of the tumor. In the most preferred embodiment, the protein C blocking co~ )oulld is provided in combination with a 20 cytokine that stimlll~tes natural killer and Iymphokine-activated killer cell-meAi~ted cytotoxicity, activates macrophages, stim~ tes Fc receptor e~ression on mQnon~ ear cells and antibody-dependent cellular cytotoxicity, enh~nces HLA class II antigen expression, and/or stimlll~tes proco~ nt activity, such as tumor necrosis factor (~NF), 25 g~mm~ interferon, granulocyte-macrophage colony stim~ ting factor, interleukin-1 or interleukin-2. It is possible to stop or reverse the hyperco~ rathy following ~Aminictration of the protein C blocking compound by ~dmini~tration of an agent inhibiting the blocking compound, such as by ~lminictration of protein C, or more preferably 30 activated protein C, when the blocking compound is anti-protein C
antibody. Although initial studies show that a single treatment is wo 91/017~3 Pcr/usso/o4337 ~206 1~85 effective in significantly redllring tumor size, different protein C
blocking compounds, alone or in combination with different cytokines and/or other anti-tumor agents, such as chemotherapeutic agents or anti-tumor monoclonal antibodies, can be ~rlministered simultaneously 5 or subsequently to the initial treatment to reduce tumor size.
Radiation and hyperthermia can also be used to sensitize the tumor to the protein C blocking agent.
Examples are provided demonstrating the systemic and local ~ isl~ation of the protein C blocking compound, alone or in 10 combination with ~NF, to dogs having canine venereal tumors (characterized as a sarcoma), a fibrosarcoma, and an adenocarcinoma, to pigs having melanomas, and a baboon with a m~lign~nt lymphoma, followed by significant tumor reduction. .A~lministration of the composition to a dog with a lymphoma was partially effective.

Brief Description of the Drawings Figure 1 is a graph of tumor size (% volume) over time (days) following ~.l...;..i!~l.a~ion of TNF alone (a); anti-protein C
antibody in combination with TNF (b); and anti-protein C antibody alone (c) to dogs ino~ ted with the canine venereal tumor.
Figure 2 is a graph of tumor size (% volume) over time (days) following ~dmini~tration of TNF alone (a); anti-protein C
antibody in combination with lNF (b); and anti-protein C antibody alone (c) to dogs inoclll~te~l with the canine venereal tumor.
Figures 3A and 3B are perspective views of the fibrosarcoma in a dog prior to treatment (Figure 3A) and eleven days post treatment (Figure 3B).
Figures 4A and 4B are views of a laryngeal adenocarcinoma in a dog prior to treatment, Figure 4A in cross-section with the laryngeal cartilages, and Figure 4B in perspective.

wo 91/017S3 Pcr/us9o/w337 ~ ~4~
~etailed Description Or the In~ention Blockage of the natural anticoa~ nt pathways, in particular the protein C pathway, uses the natural proco~1~nt l~rolx, Lies of the tumor to target the tumor capillaries for microvascular thrombosis, le~ling to hemorrhagic necrosis of the tumor. This method provides a new approach to the treatment of solid tumors either alone or in conjl~n~ion with biological rcs~ ce modifiers, chemotherapy or radiation treatments.
Tumors contain proteins which predispose to tbe formation of blood clots in the vessels in the tumor bed. Tumors aLso cont~in other proteins and cellular ele.l~nt~ which pl~el1t tbrombosis of tumor blood vessels. Tumor r.ec.osis results from altering the hC.~os~ ;c b~l~nce between procQ~l~nt and anticoa~ nt mech~nicmc to favor thrombosis of the tumor microv~sc~ tl1te. The heTnQst~tic b~1ance of the tumor can be altercd by blocking the con~e,sion of protein C to its active forrn (activated protein C). Tbe procoa~)4nt mer~l~nisms present in the tumor bed will then function without opposition and cause thrombosis of the tumor vesseLs.
Tbe protein C c~ccade can be inhibited directly and specifically by several means, inc1u-1ing antibodies to protein C or activated protein C, antibodies to protein S, inactivated protein C and C4b binding proteirL The l,refel.ed embo(liment of the method uses a monoclonal antibody to buman protein C. In the most ~,refe-.ed embodiment this is a mQnoc1on~1 ~ntibo~y known as HPC4, described in U.S. Patent No. 5,202,253 entitled "Monoclonal Antibody Against Protein ~ by Charles T. Esmon and Naorni L Fcn~on The epitope for this antibody spans the acthation site in protein C and as a result blocks protein C activation. As an experimental tool it is important to note that the antibody clossreacts with protein C from canine, porcine and at least two primate plasmas, baboon and marmoset. It does not cross react with bovine or mouse protein C. The inhibitory effect can be reversed in~t~ntly by ~ inisl.ation of acthated protein C to which WO 91/01753 PCI'/US90/04337 ' -11- 206~58i5 the antibody does not bind. The antibody therefore provides a means to selectively inhibit the protein C pathway in vivo and to reverse the process if thrombotic complications ensue at sites other than the tumor. Other antibodies to protein C and activated protein C are S known, for example, the antibodies described by Ohlin and Stenflo in J. Biol. Chem. 262(28), 13798-13804 (1987); Laurell, et al., FEBS Lett 191(1), 75-81 (1985); and Suzuki, et al., J. Biochem. (Tokyo) 97(1), 127-138 (1985).
Antibodies to protein S can be obtained using methodology 10 knov~n to those skilled in the art for the preparation of hybridomas from ~nim~l.c i...,.~ e-l against protein S purified using the method of Comp, et al., in J. Clin. Invest. 74, 2082-2088 (1984).
Blocking protein C activation in vtVo in the dog for 18-30 hours does not result in infarction of body organs nor does it result in detectable skin necrosis, which is observed in homozygous protein C
deficient newborn inf~rltc. However, the potential risk of thrombotic complications outside the tumor v~Cc~l~tllre is a consideration with any blockade of the protein C system, whether congenital or induced by a specific monoclonal antibody. An alternative method for blocking the protein C system, possibly with less risk of systemic thrombosis, involves blocking protein S function. Blocking protein S
function may be tolerated without the development of major thrombotic side effects, such as thrombophlebitis. During pregnancy an acquired protein S de~lciency occurs and protein S activity drops to 38% +/- 17% (mean +/- 1 S.D.). Thrombotic complications occur rarely during pregnancy, suggesting that protein S deficiency is tolerated in many individuals without severe clotting.
Protein S activity is blocked in vivo with specific monoclonal antibodies against human protein S. The monoclonals which inhibit protein S function block protein S in vivo. The monoclonals which bind protein S, but do not block function, should not block the system. The ~(lminictration of the blocking monoclonals W O 91/01753 PC~r/US90/04337 ~,6~ -12-is anticipated to decrease tumor growth and/or cause regression, whereas the non-inhibitory antibodies should not.
In the most plefelled embodiment, monoclonal antibodies of the same species as the patient to be treated are used. Since a 5 single dose may be effective, however, it is possible to ~mini~ter cross-species antibodies and still see tumor reduction. The antibody binding region can also be cloned and recombinantly e~ressed, using methods known to those skilled in the art, for use in blocking the protein C pathway.
A second approach to blockade of the protein C system does not involve the use of monoclonal antibodies. C4b binding protein can be purified and infilced at high levels. C4b binding protein in vivo should result in a shift in protèin S from the free, and functionally active, form to the inactive C4b binding protein-protein S
complex, rendering the protein C system inactive.
Human plasma contaills free and bound protein S. C4b binding protein can be purified using affinity chromatography with a monoclonal antibody directed against the human C4b binding protein.
C4b binding protein (C4bBP) can be obtained using the method of Dahlback, Biochem. J. 209, 847-856 (1983) or Nussenzweig, et al., Methods En_ymol. 80, 124-133 (1981). Methods yielding high levels of C4b binding protein are as follows.
A method for purification of C4b binding protein is as follows: 30 liters of plasma is thawed at 37~C and 10,000 units porcine heparin added with stirring. The plasma is diluted with an equal volume of 10 mM ben_amidine HCl, 0.025'o NaN3, 2 mM
ethylene~ minetetraacetic acid (EDTA), and 0.8 units heparin/ml in 20 mM Tris buffer, pH 7.5. The diluted plasma is then batch adsorbed with gentle stirring for 1 hr with 30 gm QAE-SephadexTM
(Pharmacia Fine Chemicals, Piscataway, NJ) which has been rehydrated in 20 mM Tris buffer, pH 7.5 cont~ining 0.1 M NaCl, 5 mM berl7~mi(line HCl, and 0.02% sodium a_ide. The QAE-Sephadex wo 91/01753 ~ ~ ~ Pcr/US9O/04337 . ~, , ~
206~585 is allowed to settle for 45 min and the supernatant siphoned of~ The resin is then packed in a 10 x 30 cm cohlmn and washed with 1 L 150 mM NaCl, 10 mM benzamidine HCl, in 20 mM Tris buffer, pH 7.5.
The column is step eluted with 1 L 20 mM Tris, pH 7.5 buffer cont~ining 0.5 M NaCl, 10 mM benzarnidine HCl, and 0.02~ sodium azide. The eluted material is made 1 mM in diisopropylfluorophosphate and 10,000 units heparin added.
Ammonium sulfate is added to the supernatant to 30~o saturation at 4~C and held at that temperature with gentle stirring for 1 hour.
Following precipitation, the pellet is resuspended and desalted into 20 mM Tris buffer, pH 7.5, cont~ining 0.1 M NaCl and S mM
benzamidine HCl. The desalted material is then loaded onto a 2.5 x 5 cm column of a monoclonal antibody directed against C4b-binding protein linked to Affigel lOTM (Pharmacia) (S mg antibody/ ml gel).
The column is eluted with 80% ethylene glycol in 1 mM MOPS
buffer, pH 7.5, at room temperature at 10 ml/hr or with distilled water CO~t~illil-g 1 mM MOPS, pH 7.5, and applied to a heparin agarose column as described below.
Protein S can simultaneously be isolated from the plasma and the C4bBP-protein S complex by QAE adsorption and elution of the plasma before adsorption with the monoclonal antibodies. This allows isolation of both the free protein S and the C4bBP protein S
complex, as well as protein C and some other plasma factors. The QAE eluate is prepared by rlilllting 30 L of human plasma 1:1 with 0.02 Tris HCl, pH 7.5, 10 mM benzamidine HCl, 1 unit/ml heparin, and adsorbing for one hr with 30 g of preswollen QAE SephadexTM.
The QAE is allowed to settle for 30 min, the supernatant is decanted or siphoned off and the QAE packed into a 10 cm diameter column, washed with approxi,.,~tely two liters of 0.15 M NaCl in the Tris buffer and eluted with 0.5 M NaCl. This eluate, which is a bright green color, is adsorbed for one hr with anti-C4bBP antibody BP45 tn remove the C4bBP and C4bBP-protein S complex and then for one hr wO 9l/01753 6~ Pcr/usso/o4337 with the Ca2~ dependent anti-protein S antibody S163 to remove free protein S. The BP45 co!umn is then processed and protein S and ,.
C4bBP recovered~ as described below. The yield of protein S from the S163 is about 10 mg, the yield of C46BP is about 50-100 mg and yield 5 of protein S from the complex is about 10 mg.
A variation of the method for isolating C4bBP in high yield is described below. Three liters of frozen plasma are thawed at 37O
C, made 10 mM in benzamidine HCl (BHCl), and cooled in an ice bath to 4~ C. PEG 8000 is added with stirring to 6% wttvol and stirred for one hr. This is cellllifuged at 4000 rpm in the PR6000 for 30 min at 4~ C and the supernatant discarded. The pellet is resuspended in 1 L of 0.3 M NaCl, 0.02 Tris 10 mM BHCL, pH 7.5, for 1 hr at 25~ C on a magnetic stirrer. The suspension is then celllliruged for 30 min at 4000 RPM in the PR 6000 at 25O C. The supernatant con~aills the C4bBP. The pellet is resuspended again, as above for 1 hr, and recenlliruged. This supernatant also contains C4bBP. Small samples of each fraction are assayed for C4bBP. The two fractions are pooled and adsorbed to a BP 45 column for 1 hr at 25 ~C. The col~lmn is washed overnight with 4 L of 0.25 NaCl in 0.02 Tris 10 mM benzamidine HCl (BHCL), then washed with one column volume of 0.1 M NaCl in 0.02 M Tris to remove the BHCL. The C4bBP is eluted with 80% ethylene glycol, 1 mM MOPS.
The eluate is applied to a heparin agarose column (5 X 8 cm) medium mesh. Application can be rapid, a~,l)roY;",~tely 20 cm pressure head. The C4bBP binds to the column while the protein S
washes through. The column is then sequentially washed with (1) 100 ml of 80% ethylene glycol, (2) 100 ml of 0.1 M NaCl, 0.01 M MOPS
pH 7.5. The C4bBP is then eluted with 1 M NaCl, 0.01 M MOPS, pH
7.5, using about 1 col volume elution buffer/hr during the elution to keep the C4bBP very concentrated. The yield is about 100 mg from 3 liters plasma with no detectable cont~min~nt~ by gel electrophoresis.

W O 91/01753 A P(~r/US90/04337 -15- 2 0 6 1 ~ 8 ~

For purification of protein S, the C4bBP breakthrough from the heparin-agarose column is applied to an S163 column. Since this is a Ca2' dependent monoclonal antibody the sample is made 2 mM in Ca2~. The S163 column is washed with 0.l M NaCl, 0.02 M
Tris HCl pH 7.5 in 2 mM Ca2~, and eluted in the above buffer with 2 mM EDTA replacing the Ca2~. Yields from this approach are a~prc.-;,.,~tely 10 mg protein S that is about 80~o uncleaved from 6 liters of plasma.
Pyrogen-depleted C4b binding protein is ~tlminictered to 10 raise plasma levels 5-fold, which should result in a greater than 90%
reduction of free protein S and a co,les~onding reduction of protein S
activity.
The protein C blocking agent is preferably ~lmini~tered in combination with a cytokine that stim~ tes natural killer and 15 lymphokine-activated killer cell-me~ te~ cytotoxicity, activates macrophages, stiml~l~te~ Fc receptor e~,ression on mononuclear cells and antibody-dependent cellular cytotoxicity, enhances HLA class II
antigen e~ression, and/or stim~ te~ proco~ nt activity, such as tumor necrosis factor (l~NF), interleukin-1 (IL-1), interleukin-2 (IL-2), 20 g~mm~ interferon (g~mm~-IFN), or granulocyte-macrophage colony stimlll~ting factor (GMCSF). Recombinant TNF, IL-2 and GMCSF
can be obtained from Cetus Corporation, 1400 53rd Street, Emeryville, CA, or Biogen Corp., Cambridge, M~ IL-1 can be obtained can be obtained from Genentech, South San Francisco, CA, 25 or ~offm~n-LaRoche, Nutley, NJ. Gamma interferon can be obtained from Genentech, Biogen or Amgen Biologicals, Thousand Oaks, CA.
Tumor necrosis factor is capable of producing significant systemic toxicity in humans including fever. chills, anorexia and n~ e~, hypotension and abnormalities of liver function (Spriggs, D.R., 30 Sherman, M.~, Frei III, E. and Kufe, D.W. Tumour Necrosis Factor and Related Cytokines. p. 124-139 (1987). Other monokines may be more suitable for enhancing the tumor directed effects of protein C

wo 91/01753 Pcr/US90/04337 'l.~6~: -16- .~

blockade. For example, interferon-g~mm~ displays a number of properties which suggest that the combination of protein C blockade and interferon-g~mm~ might be succes~ in tumor killing. These characteristics inc1ude stimlll~tion of natural killer and lymphokine-5 activated killer cell-me~i~ted ~ytotoxicity, activation of macrophages with Ul vivo and in vitro stim~ tion of peroxide generation, stimn1~tion of Fc receptor e~yression on mononuclear cells and antibody-dependent cellular cytotoxicity, and enhancement of HLA class II
antigen expression. Interferon-g~mm~ is relatively well tolerated by 10 human subjects at doses which provide significant immune stimnl~tion.
Another alternative to the use of TNF is g~mm~ irradiation in conjunction with blockade of protein C activation. Canine tumors are relatively sensilive to X-irradiation and radiation doses of 1000-3000 rads cause complete tumor regression. However, a dose of 100-15 300 rads delivered to the tumor bed is expected not to kill the tumorbut should be sufficient to cause mild local infl~mm~tion in the tumor.
This infl~mm~tion may be sufficient to cause expression of tissue factor and a loss of thrombomodulin e~yres~ion on endothelial cells in the tumor, which in turn could render the tumors more sensitive to 20 the effects of protein C blockade.
Hyperthermia can also be used to sensitize cutaneous or subc~lt~neous tumors to the effects of the protein C blocking agent, alone or in combination with the cytokine. Methods for inducing localized or systemic hypothermia are known to those skilled in the 25 art.
Optionally, other compositions and methods for treatment can be ~lmini~tered simultaneously or subsequently to treatment with the protein C blocking compound and the cytokine, such as radiation which induces infl~mm~tion in the tumor or chemotherapy with an 30 agent such as cis-p1~tin~1m or 5-fluorouracil (5-FU), which selectively inhibits the most rapidly replicating cells.

W O 91/017~3 i'~ PY~r/US90/04337 -17- 206~ j8~

The therapeutic regime actually used is dependent on the patient and the type of tumor to be treated. In the preferred embodiment, the protein C blocking compound and/or cytokine is ~lminictered in a single dosage, either systemically or at the site of 5 the tumor. In general, an amount of protein C blocking compound in excess of the amount of material in the body to be inhibited will be .1,,,i,,i~lered with each dose. For example, anti-protein C antibody is ~dminictered in an amount equal to the molar concentration of protein C in the plasma, or about 10 I-g antibody per milliliter of 10 blood (i.e., at apl,lox;",~tely a two fold molar excess of protein C
binding sites). C4b binding protein is ~dmini~tered in a dosage equal to or greater than five times the plasma level of C4b binding protein.
TNF is ~-lminictered systemically in a dosage of apprc,l;.,.~tely ten percent of the LDso of 100 ~g TNF/kg body weight, 15 or about 10 /~g TNF/kg body weight. TNF can be ~tlmini~tered into the tumor at a dosage between a~rox;..~tely one microgram and 200 microgram/m2, preferably less than 25 micrograms/m2. TNF is FDA
approved for use in cancer patients, however, the toxicity is such that the beneffts generally do not justify usage in hllm~n~. As discussed by Bartsch, et al., in Eur. J. Cancer Clin. Oncol. 25(2), 287-291 (1989), the side effects of inllalulllor ~timini~tration of rTNF include chills, fever, anorexia, and f~ti~-e, similarly to systemic r~NF treatment. As, reported by Kahn, et al., in J. Acquir. Immune Defic. Syndr. 2(3), 217-223 (1989), the m~xi,-,ll,,, intralesional dose tolerated in treatment of AIDS-associated Kaposi's sarcoma was less than 100 micrograms/m2, preferably 25 micrograms/m2. Neither study showed significant clinical usage for rTNF alone in the treatment of tumors.
The monoclonal antibody against protein C is the most preferred protein C blocking agent because the protein C blockade can be reversed immediately by the ~tlmini~tra~ion of activated protein C. If an immlme response against the antibody arises, secondary treatment can utilize the C4bBP since this would not elicit an immune wO 91/01753 ,l,~6~ Pcr/us9o/o4337 ~ i -18-response. The most preferred cytokine at this time is TNF. Tumors not responding to the combination of the protein C blocking agent and TNF can be challenged with the combination of the protein C
blocking agent and interferon, Il,1 or 2, and/or GMCSF. If the 5 tumor is still not completely~ resolved, the patient can be further treated with monoclonal antibodies against specific tumor antigens, chemotherapeutic agents, and/or radiation.
As ~ csed above, the hypercoagulable state can be reversed by ~lmini~tration of a compound binding to and inhibiting 10 the protein C blocking compound, such as the HPC4 anti-protein C
epitope peptide or by ~-lminictration of activated protein C which does not bind to this antibody. A conlilluous infusion of 30 ~Lg/kg/min recombinant tissue pl~cminogen activator can also be used.
Based on studies of coronary artery thrombosis in dogs, this dose is 15 anticipated to be sufficient to cause lysis of microvascular clots and decrease the euglobulin clot lysis time by at least 50% during the TPA
infusion. The tissue pl~cminngen activator infusion will be continued over a six-hour period. Heparin may also be useful, although clinical experience with homozygous protein C deficient infants indicates that 20 protein C concentrates, but not heparin, are effective therapeutically.
Example 1: Tre~tTnent of a transpl~rt~ble canine venereal tumor using anti~ t~ C and TNF.
The anti-protein C antibody, either with or without added TNF, has been used to block the growth of a transplantable canine 25 venereal tumor. The tumor is an undifferentiated round-cell neoplasm, characterized as a sarcoma, described by R B. Epstein, et al., in Neoplasm Immunity: Experimental and Clinical 299-, Crispen, ed. (Elsevier North Holland, Inc 1980) TNF causes hemorrhagic necrosis in susceptible tumors in mice This tumor killing and 30 necrosis was postulated to be secondary to microvascular thrombosis since it is known that TNF down regulates natural antico~ nt activities and elicits the formation of coagulant complexes However, wo 91/Ol7S3 Pcr/US9o/o4337 ....

in most ~nim~lc, hemorrhagic nc~,o~is is not observed. A compound which blocks the protein C pathway, such as thc anti-protein C
antibody HPC-4, described in U.S. Patent No. 5,202,253, is used to render an animal hypercoa~l~ble since it blocks protein C activation.
S The tumor can be easily tr~ncmitte~ to normal dogs by harvesting fresh tumor from a tumor-b~a"ng dog and, following trypsin tre~tmP-nt of the excised tumor, injecting ~,vashed tumor cell s~spe~cions subcutaneously into the recipients. The tumor displays a logarithmic growth pattern. Since the tumor grows directly beneath the skin, tumor volume can easily be measured on a daily basis, using a formula for estim~ting the volume of a sphere based on the height, width and depth of the tumor. The sub~Jt~neous location facilitates needle biopsy and excision of the tumors. Up to six tumors (three on each side of the back) can be placed on each ~nim~l Al,~ro~ tely one-third of tumor-bearing dogs eventually spontaneously reject the tumor. Ho~.e~er, this occurs at 4-6 month.c of growth, which is beyond the 21 days of growth to 50 days of growth used in the study.
The monoclonal antibody, desi~n~te~l HPC4, used to block protein C activation has been previously used succe-ssfully to block protein C activation during E. coli infusion induced shock in baboons, as described in U.S. Patent No. 5,009,889 entitled '~reatrnent of Dysfunctional Vascular Endothelium Using Activated Protein C" issued April 23, 1991 by Fletcher B. Taylor Jr., and Charles T. Esmon.
The antibody is Ca~ dependent and blocks the activation of protein C
by the thrombin-thrombomodulin complex both in vi~ro and in ViYo by binding the heavy chain of protein C at the activation site. A
synthetic peptide which encompasses the activation region (residues 6-17 of the heavy chain) is bound by the antibody in the presence of calcium ions, as further described by Stearns, et al., J. Biol. Chem.
263(2), 826-832 (1988).
The binding of the HPC4 antibody to the synthetic peptide bas f~cilit~ted the purification of the monoclonal antibody from mouse '1 .,?~
~'t~ ~

W O 91/01753 ~ ~ P(~r/US90/04337 ~6 ~ -20-ascites fluid and thus has facilitated studies involving in vivo ~lminictration of this antibody to large experimental ~nim~lc such as baboons and dogs. The half-life of the HPC4 anti-protein C
monoclonal antibody has been determined to be a~prnxi,.,~tely 18 hours. At present, 1.1 grams of the HPC4 monoclonal can be obtained from 200 ml of mouse ascites fluid by affinity chromatography using the immobilized peptide. Affinity chromatography permits extensive washing of the bound antibody and signiffcant reduction in endotoxin levels in the antibody preparations and no detectable cont~min~ting mouse proteins.
The combination of the HPC4 with TNF was tested in three sets of dogs to see if it would f~cilitate tumor killing or reduce the rate of growth using a dog model. Transplanted venereal tumors were placed subcutaneously into six dogs, with each dog in the first set receiving four tumors and each in the second group receiving six tumors. Growth was monitored by measuring the size of the tumors as described by F~pstein) et al. Two dogs were given TNF alone (10 ~lg/Kg body weight), two dogs were given HPC4 alone (20 ~lgtrnl plasma volume), and two were given a combination of both.
All animal work was conducted in a blinded manner to decrease the possibility of observer bias. The tumor-bearing dogs were assigned to the various treatments in a randornized manner. The individual measuring tumor growth was unaware of the particular treatment the various ~nim~lc were given. The veterinary pathologist pelro,llling the autopsies did not know the treatment schedule nor did the pathologist ev~ ting the light and electron microscopic data.
The results of the first animal study are shown in figure 1.
Fresh tumor was prepared and injected subcutaneously at six sites on each of three dogs. Twenty-one days later the average volume of the tumors on the ~nim~lc was apl)roxi.~.ately 2 ml. An intravenous dose of tumor necrosis factor of 10 ~g/kg was selected because although a transient decrease in blood ples~lre and a transient 40~o drop in the white blood cell count occur, the dose is not lethal and is well below W O 91/01753 ~ PC~r/US90/04337 -21- 206~85 the 100 I g/ml dose of TNF found by others to be uniformly toxic in dogs. The anti-protein C monoclonal antibody (HPC4) was ~fl.ni~ ered at a dose of 1 mg antibody/kg, calculated to give plasma levels of a~lo~i...~tely 20 ~lg antibody/ml. In the case of the dog given both the antibody and TNF, the antibody was ~-lmini~tered first.
The volumes of the tumors (siY. tumors per dog) were determined by measuring the dimen~ions of the tumors with calipers. The average tumor volume over time for each of the three dogs is shown in figure 1.
By the fourth day, the tumors on the antibody only treated dog were no longer palpable and could not be measured. The tumors on the TNF treated dog grew at a rate co~ al~ble to that expected from previous studies on untreated, tumor-bearing ~nim~lc This expected growth rate is shown as a dashed line. The tumors in the dog treated with the combination of antibody and TNF demonstrated a reduced rate of growth and had only increased in average volume by 60~o when the study was termin~ted. At autopsy tumors were not found on the dog receiving antibody only. The tumors on the antibody-TNF treated dog were necrotic throughout with few normal tumor cells. Tumors on the TNF treated dogs showed some mild central necrosis only and had large regions of viable tumor cells.
Other organs in the ~nim~ls, inçllltling the kidneys, adrenal glands and heart, were eY~mined grossly and by light microscopy and did not show evidence of microvascular thrombosis or vascular damage.
These early tumors usually just cease growing and when ~lmini~tration is at an early stage when the tumor is small, no dramatic effects are noted, although the histology clearly shows hemorrhagic necrosis in the tumor bed within two to siY. hours of the ~d...i..i~lfation of the antibody with or without the TNF. In one dog 30 with a set of large tumors, about the size of tennis balls, ~tlmini~tration of HPC4 resulted in the tumors becorning inflamed, soft and ultimately lu~tuling. All of these tumors ultimately necrosed.

WO 91/017~3 ~6~Q~ Pcr/usgo/0433~

The dog showed few side effects of the treatment and did not seem to be in any obvious discolllrol 1.
A second set of dog experiments was then performed and the growth of the tumors followed for an extended period of time.
S The results are shown in Figlire 2. The conditions and doses of TNF
and anti-protein C moi~oclonal antibody were the same as in the first study. The tumors on dogs treated with the anti-protein C
monoclonal antibody alone and the anti-protein C monoclonal antibody-TNF combination both demonstrated a marked decrease in 10 the rate of growth. To monitor histologic cll~n~es in the tumors, true cut needle biopsies were performed on day 2 and day 4. Total excision was carried out on day 36. Tissue was eY~mined and compared in regard to Iymphocytic infiltrate, spindle cell changes, necrosis and polymorphonuclear leukocytic infiltration. The results 15 are shown in Table I.
TABLE I: TUMOR HISTOLOGY
Lymphocytes PMN Necrosis Dav Dav 2 4 36 _ 4 36 2 4 36 TNF 1+ 1+ 2+ 1+ 2+ 1+ - - 1+
Antibodyalone 4+ 2+ 4+ - 3+ 1+ - 1+ 3+
TNF + Antibody 2+ 1+ 4+ +/- 3+ 3+ - - 2+
The values range from absent (-) to severe (4+ ).
There was no single parameter which correlated well with the degree of the tumor growth rate. There was ~roll~inent lymphocytic in~1ltration in dogs treated with anti-protein C monoclonal antibody 5 only and anti-protein C monoclonal antibody-TNF. Necrosis observed was a co~ tion type, mostly seen in the center of the tumors, and appeared to better correlate with the size of the tumor. There was proliferation of spindle cells and fibrosis which was slightly prorninent in dogs treated with anti-protein C monoclonal antibody or the 10 antibody and TNF. PMN infiltration was seen in all cases. However, Wo 91/01753 Pcr/us9o/o4337 ."~ , -23- ~
2~6~585 in the dog treated with the combination, the infiltration by PMN was rather diffuse and frequently associated with focal necrosis.
One interesting finding was the presence of granulomatous nodules at the periphery of the tumor nodule made up with clusters of degenerating tumor cells at the center ~ù~roùllded by a zone of histiocytes. These were found only in the dog treated with anti-protein C monoclonal antibody and TNF. This finding is more characteristic of the granulonlalous infl~mm~tion which occurs in tumors following BCG aflmini~tration~ but which has not been found following TNF treatment of tumors in mice. This suggests that the combination of TNF and protein C blockade may induce a cell-merli~ted infl~mm~tory reaction in the tumor.
Another finding is a decreased capillary blood supply to the tumors of the dogs which have received protein C blockade, alone or in combination with TNF. Normally, the tumors receive blood via an extensive capillary inflow which origin~tes primarily in the skin and subcutaneous tissues and to a lesser extent from the deeper tissues.
In the dogs which were treated with protein C blockade, the vessels entering from the su--ounding tissue into the tumors are visibly decreased. There is also a marked reduction in the capillary oozing of blood when these tumor are eY~se~l This raises the possibility that angiogenesis has been decreased in the tumors where protein C
blockade occurred.
Since the histomorphological changes observed at two and four days following protein C blockade do not explain the major differences in tumor growth rate observed in the three ~nim~ls, the light and electron microscopic characteristics of the tumors at much earlier times were st~ ied Four dogs with six tumors apiece were treated as follows: 1) placebo, 2) antibody-TNF, 3) TNF, and 4) antibody only. One tumor was removed at -10 min, 10 min, 30 min, 2 hrs, 6 hrs, and 24 hrs from each ~nim~l Representative portions of each tumor were snap frozen in liquid nitrogen for Wo 91/01753 Pcr/US90/o4337 ~,~6~ -24- ~

imm1mQhistochemistry, fixed in 3.0~ buffered glutaraldehyde for electron microscopy and fixed in 10% buffered formalin for routine histochemical studies.
On light microscopy, neutrophil lllargi~-~tion occurred at 30 5 mim1tes following tr~ahnent in the dog receiving HPC4/TNF and in subsequent tissue'samples. Hemorrhagic necrosis was evident in the HPC4/TNF animal at 6 and 24 hours. Hemorrhagic necrosis and neutrophil margination were most marked in the center of the tumors.
In the tumors of hte dog receiving HPC4, neutrophil margination l0 occurred at two hours and isolated areas of hemorrhagic necrosis were seen at 24 hours. Increased margination did not occu in the dog receiving TNF only or in the control dog. Light, sc~nning, and s,-,;ccion electron microscopic studies did not show capillary fibrin deposition. Failure to detect microvascular fibrin may be due to ex 15 vivo fibrinolysis following biopsy. No areas of gross vascular infarction were observed in the brain, heart, kidneys, liver, adrenal glands, or lungs, and no hemorrhagic necrosis was seen on histologic ey~min~tion of re~rcsell~alive samples of these organs. Complete autopsies, inr1~1tling gross and histologic ~ on of the liver, spleen, kidneys, 20 adrenal glands, intestin~s~ brain, heart, and lungs have been conducted on 15 ~nim~lc. No thrombosis of internal organs occurred. There is no evidence of extensive intravascular co~ tion following the ~",;"i~lration of HPC4 (n=2) or HPC4/TNF (N=2), as documented by no significant change in the platelet count over a four day period 25 following treatment.
Example 2: Treatment of an unresectable fibrosarcoma with anti-protein C and TNF.
A dog having a biopsy-proven fibrosarcoma of the soft pallet growing up into the nasal passages was treated with a single 30 inl,a~,ellous injection of 10 l~g TNF/kg body weight followed by a single injection of l mg anti-protein C (HPC4)/kg body weight.

~'VO 91/01753 Pcr/us9o/o4337 ~ -25- 206~58S

Tre~sment resulted in significant necrosis and greatly hlll,loved the ability of the dog to breath through his nose.
The location and the size of the tumor, measuring approxim~tely 3.5 cm x 5.0 cm, at the time of treatment is shown in Figure 3A. The 5 reduction in size after eleven days, with only a small necrotic portion rem~ining, is shown in Figure 3B.
Example 3: Treatment of an unl~ ect~ble adenocarcinoma with anti-protein C and TNF.
A second dog having a biopsy-proven adenocarcinoma in the 10 laryngeal area was also treated with a single injection of l0 ~g TNF/kg body weight followed by one mg anti-protein C (HPC4)/kg body weight. The animal entered the hospital with severe difficulty in bre~shing due to obstruction by the tumor. The tumor measured app-oYi~ tely 3 cm x 2 cm by palpation. Figures 4A and 4B are 15 views of the laryngeal adenocarcinoma prior to treatment, Figure 4A
in cross-section with the laryngeal cartilages, and Figure 4B in perspective.
Twelve days following treatment, the tumor volume was reduced 50%. One week later the tumor was no longer detectable. The dog 20 returned to full functional status, is able to bark, and has been walking 4 to 5 miles per day with her owners. At six weeks some detectable tumor mass had reappeared and additional treatment was w~anted.
Both ~nim~ls in examples 2 and 3 tolerated the therapy very well, 25 despite being geriatric cases, both ~nim~lc being over 10 years of age.
Example 4: Tr~ e~t of a Iymphoma using anti-protein C and TNF.
A German shepherd dog with naturally occurring advanced lymphoma was treated with a single ~tlmini.ctration of TNF and HPC4.
30 There was a 50~o reduction in the average size of his multiple tumors for a period of applo~i.l.~tely one week following therapy with HPC4/TNF. Two other dogs with naturally occurring tumors, one wo 91/01753 Pcr/usso/o4337 with an osteosarcoma and the other with a melanoma, had no response to treatment with HPC4/TNF.
The use of HPC/TNF in these dogs has been followed by vomiting and a loss of appetite for 24 hours following treatment, a 5 response identical to that seen in control ~nim~lc ~-lminictered TNF
alone. - ~
Example 5: Treatment of dogs with tumors using anti-protein C
only.
Five dogs with naturally occuring tumors were treated with HPC4 10 alone. One dog with a hemangiosarcoma showed improvement. The four other ~nim~lc, having a hemangiosarcoma, a histiocytoma, a m~mm~ry carcinoma, and an adenocarcinoma, had no h~l~rovement.
Example 6: Treatment of Sinclair swine ~l~r~ma with TNF, anti~ t~ C monoclonal antibody, or TNF in combination with anti-protein C.
Sincl~ir mini~tnre swine spontaneously develop multiple cutaneous me!~nomas which have the ability to metastasize and regress. These ~nim~ls are an accepted model for human melanomas because of the many features in common with human melanomas: tumors develop 20 spontaneously; swine possess a wide spectrum of benign melanocyte lesions capable of m~lign~nt transformation; melanomas in pigs histopathologically resemble human superficial spreading melanoma;
metastatic disease is correlated with deeply invasive cutaneous tumors;
the pattern of metastatic spread is analogous to the distribution of 25 metastases in human melanoma (as in human melanoma, the highest incidence of metastasis is to lymph nodes, lungs, and liver, with metastases to multiple organ systems in one-third of the ~nim~lc); the histopathology of cutaneous regression is similar; a tumor-related immlme response occurs in the host; and a genetic component is pigs 30 is comparable to a genetic component of some melanomas in humans.
This model is reviewed by Oxenhandler, et al., Amer. J. Pathol. 96(3), 707-714 (1979) and Hook, et al., AJP 108(1), 130-133 (July 1982).

W O 91/01753 ~ P(~r/US90/04337 -27- ~i.
206~58!;
There are also advantages in having an animal model control where the history of the tumors, the treatment of the ~nim~ls, and the response of the ~nim~l~ is known and can be compared to a control.
In contrast to many of the dog studies where the tumor was naturally S OC~;ullillg, the pigs were able to be treated prior to the tumors beco...i.~g life threatening.
10 day old 1 kg pigs were treated with either TNF (10 I g TNF/kg body weight/animal), the anti-protein C monoclonal antibody HPC4 (one mg HPC4/kg body weight/animal), or TNF in combination with HPC4 (10 ~g TNF/kg body weight followed by one mg HPC4/kg body weight/animal). The compositions were injected in the tissues directly under the melanomas.
The tumor on the pig which received the combination of HPC4 and TNF showed extensive hemorrhage and necrosis. Over a 72 hour period the volume of the tumor was reduced by 80%. There was no reduction in size or necrosis in the tumors treated with either TNF or HPC4 alone, nor in the control.
Example 7: Treatment of an adult baboon with a m~ nt lymphoma with HPC4/TNF.
An adult baboon with a histologically documented m~lign7~nt lymphoma was identified. The animal had large axillary and inguinal lymph nodes infiltrated with tumor as well as a large midline tumor mass, ~l~arenlly in the abdominal wall. The animal was treated with a single ~-lmini~tration of HPC4/TNF at the dosages used in the dog studies.
A decrease in tumor size was noted on the third day following treatment. The tumors had a soft rubbery consistency and the abdominal wall mass was no longer easily defined by palpation.
Ul~o-lunately, the animal experienced tumor lysis syndrome (high uric acid, calcium, and phosphorus) with evidence of renal failure, and died two days later. Autopsy showed loss of lymphoid cells in the involved Iymph nodes with evidence of necrosis and degeneration of WO91/~1753 ,~ -28- PCI/US90/04337 the rem~ining lymphoid cells. The pancreas, large intestine, and brain were normal. Within the the liver, lungs, eyelids, heart and kidneys aggregates of lymphoid cells were seen. The nuclei of many of these cells appeared pyknotic, suggesting cellular degeneration and necrosis.
Lymphoid cells within the spleen were also pyknotic. There was no evidence of thrombosis of the great vessels and no evidence of pulmonary emboli. : ~
In sllmm~ry~ the data demonstrates that blocking protein C
activation in some tumor-bearing ~nim~ls, either alone or in combination with cytokine ~tlminictration, has a marked effect on tumor growth and in some cases results in regression of the tumors.
The data consistently shows a large difference between the tumor growth in dogs treated with TNF alone versus antibody alone, and a large difference between TNF only versus antibody and TNF.
However, it is not clear if there is a significant difference between the antibody alone and the antibody and TNF treated dogs. The difference in the swine was significant.
It may be possible to o~ e the effects of protein C blockade in a number of ways. ln the examples, the TNF and antibody were ~dminictered within a single 15 mimltes period. Based on the potential priming effect the TNF may have on tumor necrosis, TNF
could be ~-lminictered 6 hours before the protein C blockade to increase tumor damage. Since TNF has a short plasma half-life, subsequent inll~vellous ~ lion of TNF during the first 18 hours the anti-protein C monoclonal antibody is in the circulation may also be used to increase tumor damage. Prolonged protein C
blockade on the tumors, both with and without TNF, may also increase tumor damage, using sequential doses of the anti-protein C
monoclonal antibody at 12 hour intervals.
The HPC4 anti-protein C monoclonal antibody preparations are relatively endotoxin-free, but still result in the ~ Ll~tion of between 15 ng and 30 ng endotoxin (total dose per dog per wo ~I/ol7s3 Pcr/US90/W337 experiment) as measured by the ~ im~ us amebocyte Iysate assay (Associates of Cape Cod, Inc., Wood Hole). This endotoxin could elicit tumor necrosis &ctor produ~tion in the recipient ~nim~lc, which may amplify the effect of the antibody ~lministered in the absence of S the TNF. Endotoxin can be removed from the anti-protein C
monoclQn~l antibody ~ep~hd~ions using immobilized polymyxin B
(Pierce Biochemical, Illinois). Anti-protein C morlorlon~l antibody preparations can also be filtcred on G-200 SephadexTM to remove any possible microaggregates imme~ tely before infusion, which may be 10 cleared more rapidly than non-aggregated antibody.
A single infusion of anti-protein C monoclonal antibody has a significant effect on tumor gro~,vth in preliminary studies. Repeated infusions of the anti-protein C monoclonal antibody over a period of 5 - days may be more effective in blocl~in~ tumor growth than a single 15 ~lminisl~ration of the antibody. Sufficient anti-protein C monoclonal antibody should be employed to keep plasma levels of the antibody over IO J~g/ml. Dog protein C levels range from 3-5 ~g/ml, as mea~ured by Laurell rocket electrophoresis using goat anti-dog protein C. This level of anti-protein C monoclonal antibody should be 20 adequate to block all protein C activation ~csllmin~ that a 1:1 stoichiometric complex forms between the monoclonal antibody and the canine protein C.
To determine if adequate quantities of the monoclonal antibody have in fact been ~d~..;..islered to complex all circul~ting canine 25 protein C, aliquots of heparinized plasma can be passed over a column of the anti-protein C monoclonal antibody (HPC4) immobilked on Affigel-10~ (Bio-Rad Laboratories, California). The immobilked antibody binds to free protein C, but not to activated protein C or circul~ting protein C - antibody complex. The column 30 bound protein C can be q~ntit~ted as previously described by Vigano D'Angelo, S., Comp, P.C., Esmon, C.T., et al. J.Clin.lnvest. 77,416-425.
(1986). An alternative approach is to remove the antibody-protein C

~a W O 91/01753 ~ ~ ~ P(~r/US90/04337 ~. .
complex from the plasma with an immobilized polyclonal antibody directed against the antibody. The residual free protein C can then be q~l~ntit~ted by Laurell rocket electrophoresis, or if increased sensilivily is required, by a radioil",...~"o~csay.
The natural fibrinolytic system in dogs is very potent and may be capable of lysis of microvascular thrombi, thereby serving to protect the tumors from blockade of the protein C system. This can be blocked by ~lmini~tration of an inhibitor such as epsilon-arnino caproic acid (EACA) before and following the ~lmini~tration of the anti-protein C monoclonal antibody. EACA is rapidly secreted into the urine and therefore will be given inlravellously at one-hour intervals for 12 hours. The effect of EACA can be monitored by the euglobulin clot lysis time method described by Comp, P.C. and F~morl, C.T. The ~2e~ tion of Co~ tion~ edited by Mann, K.G.
and Taylor Jr., F.B., p. 583-588 (Amsterdam: Elsevier North-Holland, 1980). EACA may enh~nce the anti-tumor effects of the anti-protein C monoclonal antibody.
The role of platelets in the tumor directed effects of the anti-protein C monoclonal antibody is difficult to assess. A preliminary survey of the electron microscopic data at two hours after protein C
blockade indicate that platelets are not present in the fibrin deposition in the tumor capillaries of the ~nim~ls receiving the anti-protein C
monoclonal antibody. The fibrin strands which are observed appear to be closely associated with the endothelial surface. This suggests that the endothelial surface may be capable of serving as a site of clot initiation under these conditions. This is consistent with the observation of fibrin directly associated with the vessel wall in rabbits inf~l~ed with interleukin-1, described by Nawroth, P.P., Handley, D.A., Esmon, C.T., et al. Proc.Natl.Acad.Sci.USA 83, 3460-344 (1986). If platelets are necec~ary for the effects of the anti-protein C monoclonal antibody on the tumors, thrombocytopenia may protect the tumor from the protein C blockade.

:

Claims (24)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A composition for inhibition of tumor growth in a patient comprising:
a pharmaceutical carrier containing an effective dosage of a compound to block the Protein C coagulation system selected from the group consisting of anti-Protein C
antibodies, anti-Protein S antibodies, inactivated Protein C
and C4b binding protein in combination with the cytokine having at least one activity selected from the group consisting of stimulation of natural killer and lymphokine-activated killer-cell mediated cytotoxicity, activation of macrophages, stimulation of Fc receptor expression on mononuclear cells and antibody-dependent cellular cytotoxicity, enhancement of HLA class II antigen expression, and stimulation of procoagulant activity, wherein the combination is in an effective dosage facilitating to cause hemorrhagic necrosis of the tumor in combination with the compound blocking the Protein C system and the dosage of cytokine in the combination is not effective in the absence of the Protein C blocking compound.

2. The composition of claim 1 wherein the Protein C blocking compound is an anti-Protein C antibody.

3. The composition of claim 1 wherein the cytokine stimulates natural killer and lymphokine-activated killer cell-mediated cytotoxicity.

4. The composition of claim 1 wherein the cytokine is selected from the group consisting of tumor necrosis factor gamma interferon, granulocyte-macrophage colony stimulating factor, interleukin-1 and interleukin-2.

5. The composition of claim 1 further comprising a chemotherapeutic agent selectively inhibiting growth of rapidly replicating cells.

6. The composition of claim 1 further comprising endotoxin in a dosage stimulating production of tumor necrosis factor.

7. The composition of claim 1 in combination with a source of radiation causing inflammation of the tumor.

8. The composition of claim 1 in combination with a chemotherapeutic agent inhibiting replication of cells.

9. The composition of claim 4 comprising anti-Protein C
antibodies in a dosage equal to or greater than the molar concentration of Protein in the blood of the patient and tumor necrosis factor in a dosage delivering less than 100 µg TNF/kg body weight.

10. The composition of claim 4 comprising anti-Protein S
antibodies in a dosage equal to or greater than the molar concentration of Protein S in the blood of the patient and tumor necrosis factor in a dosage delivering less than 100 µg TNF/kg body weight.

11. The composition of claim 4 comprising C4b binding protein in a dosage equal to or greater than five times the molar concentration of C4b binding protein in the blood of the patient and tumor necrosis factor in a dosage delivering less than 100 µg TNF/kg body weight.

12. The use for inhibition of tumor growth in a patient of a compound capable of blocking protein C, in a dosage blocking the protein C anticoagulation system and facilitating hemorrhagic necrosis of microvasculated solid tumors, wherein the compound is not a cytokine.

13. The use of claim 12 wherein the Protein C blocking compound is selected from the group consisting of anti-Protein C antibodies, anti-Protein S antibodies, inactivated Protein C and C4b binding protein.

14. The use of claim 12, in combination with a cytokine in a dosage not causing hemorrhagic necrosis of the tumor, wherein the cytokine has at least one activity selected from the group consisting of stimulation of natural killer and lymphokine-activated killer cell-mediated cytotoxicity, activation of macrophages, stimulation of Fc receptor expression on mononuclear cells and antibody-dependent cellular cytotoxicity, enhancement of HLA class II antigen expression, and stimulation of procoagulant activity.

15. The use of claim 14 wherein the cytokine is selected from the group consisting of tumor necrosis factor, gamma interferon, granulocyte-macrophage colony stimulating factor, interleukin-1 and interleukin-2.

16. The use of claim 12 with endotoxin in a dosage stimulating production of tumor necrosis factor.

17. The use of claim 13 wherein the Protein C blocking compound is anti-Protein C in a dosage equal to or greater than the level of Protein C in the blood.

18. The use of claim 17 wherein the dosage is equal to or greater than approximately 10 µg antibody/ml of blood.

19. The use of claim 13 wherein the Protein C blocking compound is C4b binding protein in a dosage equal to or greater than five times the plasma levels of C4b binding protein.

20. The use of claim 13 wherein the Protein C blocking compound is anti-Protein S in a dosage equal to or greater than the level of Protein S in the blood.

21. The use of claim 20 wherein the dosage is equal to or greater than approximately 10 µg antibody/ml of blood.

22. The use of claim 14 with a chemotherapeutic agent selectively inhibiting growth of rapidly replicating cells.

23. The use of claim 12 with a compound inhibiting blocking of the Protein C
system by the blocking agent.

24. The use of claim 23 wherein the inhibitory compound is selected from the group consisting of Protein C, activated Protein C, Protein S, and factor Xa.