Inhibition of Tumor Growth by RGD Peptide–Directed Delivery of Truncated Tissue Factor to the Tumor Vasculature Free

An angiogenic phenotype is an absolute requirement for tumor progression in solid tumors as well as in hematologic malignancies (1–4). This neoangiogenesis is necessary for sustaining tumor expansion because it provides sufficient supply of oxygen and nutrients as well as elimination of waste products.

Antiangiogenic therapies interfere with the complex process of growth, migration, and differentiation of blood vessels. In contrast, vascular targeting aims at destruction of tumor blood vessels with the result of tumor infarction. Vascular targeting requires the identification of target molecules that are present at sufficient density on the surface of vascular endothelium in solid tumors but are absent from endothelial cells in normal tissues. Such molecules could be used to target cytotoxic agents to the vascular endothelium of the tumor rather than to the tumor cells themselves. Promising candidate molecules include basic fibroblast growth factor, vascular endothelial growth factor and vascular endothelial growth factor receptor-2, endoglin, endosialin, a fibronectin isoform (ED-B domain), the integrins α_vβ₃, α_vβ₅, α₁β₁, and α₂β₁, aminopeptidase N, NG2 proteoglycan, and the matrix metalloproteinases 2 and 9 (5–16).

Selective activation of blood coagulation in tumor vessels with subsequent tumor necrosis is an alternative approach of vascular targeting. Such a strategy using the soluble extracellular domain of tissue factor [truncated tissue factor (tTF)] has been reported by a few groups only. The extracellular domain of tissue factor is not a coagulant while free in the blood circulation but becomes a powerful and specific coagulant once bound to the cell surface of the tumor vasculature by a targeting ligand. Specific targeting of tTF to tumor vessels has been accomplished with antibodies and peptides directed against a variety of tumor vessel markers, including MHC class II (17), the vascular cell adhesion molecule-1 (18), the ED-B domain of fibronectin (19), and prostate-specific membrane antigen (20). In all of these studies, tTF homed selectively to tumor vessels and rapidly induced thrombosis with subsequent tumor necrosis.

In contrast to these studies, when coupling a small peptide comprising the RGD motif to the NH₂ terminus of tTF, no significant inhibition of tumor growth was observed despite induction of some thrombosis in small and medium-sized tumor vessels (21).

In our study, we generated the fusion protein tTF-RGD consisting of tTF and a small RGD peptide coupled to the COOH-terminal region of tTF, thus selectively targeting α_vβ₃ and α_vβ₅ integrins on tumor endothelial cells. tTF-RGD efficiently inhibited tumor growth in mice by thrombotic occlusion of tumor vessels without any apparent side effects in other organs.

Cell lines, antibodies, and reagents. Tumor cell line CCL185 (human adenocarcinoma) was described earlier (22). Tumor cell line M21 (human melanoma) was described earlier and kindly provided by Dr. Silletti (University of California, San Diego, CA; ref. 23). Tumor cell line HT1080 was kindly provided by Dr. Bremer (University of Muenster, Muenster, Germany). The anti-tTF antibody was obtained by Diagnostic International (Karlsdorf, Germany), integrin α_vβ₃, and GRGDSP peptide were obtained by Chemicon (Temecula, CA). Microvascular endothelial cells were purchased from Technoclone (Heidelberg, Germany).

Expression, purification, and characterization of truncated tissue factor and truncated tissue factor-RGD. The cDNA coding for tTF containing amino acids 1 to 218 and tTF-GRGDSP in which the hexapeptide is linked to the COOH terminus of tTF (tTF-RGD) were amplified by PCR using the primers 5′-CATGCCATGGGATCAGGCACTACAAATACTGTGGCAGCATATAAT-3′ (5′-primer) and 5′-CGGGATCCTATTATCTGAATTCCCCTTTCTCCTGGCCCAT-3′ (3′-primer) for tTF and 5′-CATGCCATGGGATCAGGCACTACAAATACTGTGGCAGCATATAAT-3′ (5′-primer) and 5′-CGGGATCCTATTATGGAGAATCACCTCTTCCTCTGAATTCCCC-3′ (3′-primer) for tTF-RGD. With the DNA Ligation kit (Novagen, Madison, WI), the cDNA was cloned into the expression vector pET-30(+)a (Novagen) using the BamHI and NcoI sites of the vector.

The vectors were introduced in competent Escherichia coli cells (BL21 DE3) according to the manufacturer's protocol (Novagen). After stimulating with isopropyl-l-thio-B-d-galactopyranoside (Novagen), the cells were harvested and 5 to 7 mL lysis buffer [10 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 1 mmol/L MgCl₂, 10 μg/mL aprotinin, 2 mg/mL lysozyme] per gram wet weight and 20 μL benzonase (Novagen) were added to the pellet. Then, the cells were incubated for 90 minutes at room temperature and centrifuged at 12,000 × g for 20 minutes at 4°C. The pellet was resuspended and homogenized by sonicating in washing buffer [10 mmol/L Tris/HCl (pH 7.5), 1 mmol/L EDTA, 3% Triton X-100]. To solubilize the inclusion bodies, 2 to 4 mL guanidinium buffer (6 mol/L GuCl, 0.5 mol/L NaCl, 20 mmol/L NaH₂PO₄, 1 mmol/L DTT) per gram wet weight were added. After incubation overnight at room temperature, the suspension was centrifuged at 5,000 × g for 30 minutes at 4°C. The supernatant was filtered through a 0.22 μm filter and loaded onto a nickel nitrilotriacetic acid column (Novagen).

Purification and refolding was done with the His Bind Buffer kit (Novagen) according to the manufacturer's protocol. To remove the salt, the suspension was dialyzed in a Slide-a-Lyzer 10 K dialysis cassette (Pierce, Rockford, IL) against TBS buffer [20 mmol/L Tris, 150 mmol/L NaCl (pH 7.4)]. Subsequently, tTF and tTF-RGD were analyzed under denaturing conditions on SDS-PAGE and Western blot.

Factor X activation by truncated tissue factor and truncated tissue factor-RGD. The ability of tTF and tTF-RGD to enhance the specific proteolytic activation of factor X by factor VIIa was assessed as described by Ruf et al. (24). Briefly, to each well in a microtiter plate was added 20 μL of (a) 50 nmol/L recombinant factor VIIa (Novo-Nordisc, Bagsværd, Denmark) in TBS containing 0.1% bovine serum albumin, (b) 0.16 nmol/L to 1.6 μmol/L tTF/tTF-RGD in TBS-bovine serum albumin, and (c) 25 nmol/L CaCl₂ and 500 μmol/L phospholipids (phosphatidylcholine/phosphatidylserine, 70:30 MM; Sigma, Munich, Germany). After 10 minutes at room temperature, 20 μL of the substrate factor X (Enzyme Research Laboratories, South Bend, IN) were added in a concentration of 5 μmol/L. Aliquots were removed from the reaction mixture every minute and stopped in 100 nmol/L EDTA. Spectrozyme factor Xa (American Diagnostica, Greenwich, CT) was added and rates of factor Xa generation were monitored by the development of color at 405 nm with a microplate reader (Bio-Rad, Hercules, CA).

Binding of truncated tissue factor-RGD to purified α_vβ₃ and endothelial cells. The binding of tTF-RGD to purified immobilized α_vβ₃ (Chemicon) was analyzed by ELISA techniques as described earlier (25). Briefly, purified α_vβ₃ was adsorbed overnight onto microtiter wells (1-5 μg/mL, 50 μL/well) before blocking with casein blocker (Pierce). Purified biotinylated tTF-RGD in binding buffer [50 mmol/L Tris (pH 8.0), 150 mmol/L NaCl, 1 mmol/L MgCl₂, 1 mmol/L CaCl₂, 0.5 mmol/L MnCl₂] was added to the wells. Control wells received no integrin. Bound protein was detected with a horseradish peroxidase/anti-biotin monoclonal antibody and quantitated at 450 nm with tetramethylbenzidine (Bio-Rad). Binding steps were done in the absence and presence of the synthetic peptide GRGDSP (Chemicon) as competitive ligand to show the specificity of this interaction. Subsequently, the specific binding of tTF-RGD to cell surface integrins on endothelial cells was evaluated.

To this end, the differential binding of biotinylated tTF and tTF-RGD on endothelial cells in suspension was analyzed by fluorescence-activated cell sorting. Briefly, microvascular endothelial cells, which were 90% confluent, were trypsinized with 10% trypsin and washed twice with PBS. Cells were incubated with biotinylated tTF-RGD (0.1 μg/10⁶ cells) in binding buffer (Hanks' buffer containing 1% bovine serum albumin, 10 mmol/L CaCl₂, 10 mmol/L MgCl₂, 10 mmol/L MnCl₂) for 60 minutes at 4°C. After incubation and two more wash steps with binding buffer, streptavidin-phycoerythrin was employed for detection of bound protein. Briefly, cells were incubated with streptavidin-phycoerythrin for 30 minutes at 4°C, washed one more time in binding buffer, and fixed in 1% paraformaldehyde in PBS. Fluorescence-activated cell sorting analyses were done for phycoerythrin (excitation 488 nm, emission 680 nm).

Tumor mouse models. All procedures on animals were done in agreement with German regulations (Tierversuchsgesetz §8 Abs. 2) and specifically approved in the form of a project license. Single-cell suspensions (2 × 10⁶ in 100 μL) of CCL185 (human adenocarcinoma of the lung), M21 (human malignant melanoma), or HT1080 (human fibrosarcoma) were injected s.c. into the right anterior flank of male BALB/c nude mice (9-12 weeks old) obtained from Charles-River (Sulzbach, Germany). Tumor growth was allowed to a volume of ∼30 to 100 mm³ (CCL185) and 300 to 1,500 mm³ (M21 and HT1080), respectively. At this point, mice were randomly assigned to different experimental groups. Group 1 received 0.9% NaCl (200 μL), group 2 received tTF (30 μg/mouse in 200 μL of 0.9% NaCl), and group 3 received tTF-RGD (30 μg/mouse in 200 μL of 0.9% NaCl) via i.v. tail vein injection. Depending on the growth kinetics, injections were repeated twice weekly for five times (CCL185) or daily for five times (M21 and HT1080). Tumor size in vivo was evaluated using a standard caliper measuring tumor length and width in a blinded fashion. Tumor volumes were calculated using the standardized formula: length × width² × π / 6. According to our project license, animals had to be sacrificed when tumors became too large, if mice lost >20% of body weight, or at signs of pain. Representative photographs of animals and tumors were taken at the end of each experiment. Next, mice were sacrificed by cervical dislocation in deep CO₂ anesthesia in agreement with standard regulations and the project license. Immediate debleeding was achieved by injection of 0.9% NaCl directly into the left cardiac ventricle and opening of the femoral vessels. Tumors and organs (heart, lung, liver, and kidney) were excised and fixed in 4% buffered paraformaldehyde solution before embedding in paraffin. For toxicity studies, an extra set of animals was injected with a higher dose of tTF-RGD (50 μg) and analyzed for signs of thrombosis in heart, lung, kidney, or liver 1, 4, or 24 hours after injection in a similar manner. Furthermore, 9- to 12-week-old immunocompetent BALB/c mice (n = 10) were injected with 100 μg (=5 mg/kg body weight) and observed for 2 weeks.

Magnetic resonance imaging of mouse bearing human fibrosarcoma to determine tumor blood flow. Mice bearing tumors (HT1080) of ∼1,000 mm³ were randomly chosen for magnetic resonance imaging (MRI) analysis as described previously (26). In the treatment group (n = 9), mice received a single dose of 30 μg i.v. tTF-RGD in 200 μL of 0.9% NaCl; in the control group (n = 9), 200 μL of 0.9% NaCl were injected. Six hours after injection, the MRI was done on a clinical 1.5 Tesla MR system (Interna, Philips, Best, the Netherlands) using a microscopy surface coil with a diameter of 47 mm (Philips). Animals were anesthetized by an i.p. injection of ketamine (125 mg/kg body weight) and xylazine (12.5 mg/kg body weight) and the jugular vein was cannulated by a 2.5 French catheter (SIMS Portex, Kent, United Kingdom). To depict anatomic details, the tumors were imaged using a T2-weighted turbo spin echo sequence. For measuring the perfused tumor volume [vascular volume fraction (VVF)], a fast spoiled gradient dual echo EPI sequence was acquired before and after i.v. injection of ultrasmall superparamagnetic iron oxide nanoparticles (SHU 555 C, Schering AG, Berlin, Germany). The change in transverse relaxation rate (ΔR2*) was calculated and the VVF was determined by calibration of the ΔR2* of the tumor by the ΔR2* of the muscle. Thereafter, the mice were sacrificed and the tumor was removed for histologic studies. All procedures on animals were done in accordance with the German regulations (Tierversuchsgesetz §8 Abs. 2) and specifically approved in form of a project license.

Histology. Histologic analyses were done as described earlier (3). Briefly, tissues embedded in paraffin were cut to 4 μm sections and transferred onto glass slides. At least, 10 sections per sample were available for evaluation. H&E-stained sections were examined using conventional light microscopy for signs of intratumoral necrosis and thrombosis in a blinded fashion. Thrombosis was defined as complete or incomplete occlusion of intratumoral vessels by closely packed RBC, blurring of the vessel outline, and presence of aggregated platelets and fibrin deposition. Regression analysis was done as described (27). Briefly, the different histologic stages of regression were determined by two independent persons in a blinded fashion on 5 to 10 different sections for each tumor. Grade I regression is defined as no or only minor visible signs of tumor cell regression, Grade II is defined by visible signs of therapy induced tumor cell regression (e.g., necrosis adjacent to thrombosed vessels) with grade IIa defined as >10% vital tumor tissue and grade IIb defined as <10% vital tumor tissue. Grade III is defined as complete tumor cell regression meaning almost complete lack of visible vital tumor tissue.

Statistical analyses. Statistical significance of differences was tested by the Mann-Whitney rank-sum test for independent groups. Two-sided Ps < 0.05 were considered significant.

Cloning, expression, and characterization of truncated tissue factor/truncated tissue factor-RGD. For the selective targeting of tumor vasculature, we used a RGD peptide, which is a ligand for cell surface integrins on tumor endothelium, mainly α_vβ₃ and α_vβ₅ (28–30). The RGD motif GRGDSP was fused to tTF using a PCR assembly cloning strategy for expression in E. coli. The fusion protein tTF-RGD consisted of a (His)₆ sequence to allow ready purification by nickel affinity chromatography, enterokinase, and thrombin cleavage sites linked to the NH₂ terminus of tTF and the RGD peptide. The fusion protein and tTF were expressed at ∼25 mg/L and were localized in inclusion bodies. The inclusion bodies could be solubilized in denaturing buffer and purified to ∼95% purity. After refolding and dialyzing, we confirmed the identity of the protein by mass spectroscopy and Western blot (data not shown).

Functional characterization of truncated tissue factor and truncated tissue factor-RGD. The ability of tTF and the tTF-RGD fusion protein to enhance the specific proteolytic activation of factor X by factor VIIa was shown by Michaelis-Menten analyses. The calculated Michaelis constants (K_m) for tTF and the tTF-RGD was within the range of 0.15 nmol/L as reported (ref. 24; data not shown). This shows that the ligation of the peptide sequence to the COOH terminus of tTF did not affect its functional activity.

Binding of truncated tissue factor-RGD to α_vβ₃ and endothelial cells in vitro. The specific binding of tTF-RGD to immobilized α_vβ₃ was shown in a purified receptor-binding assay (Fig. 1A). Binding of our construct to immobilized α_vβ₃ was dose dependent and saturable. The specificity of this RGD-dependent interaction is underlined by the competition using the synthetic peptide GRGDSP. In the presence of a 10- to 100-fold molar excess of the unlabeled synthetic RGD peptide, tTF-RGD showed merely no significant binding to α_vβ₃.

Subsequently, the specific binding of tTF-RGD to receptors on microvascular endothelial cells was evaluated by fluorescence-activated cell sorting. Differential binding of biotinylated tTF and tTF-RGD on endothelial cells in suspension is shown in Fig. 1B. Indeed, the measured fluorescence intensity was 8-fold higher for tTF-RGD compared with tTF. Furthermore, the binding of 0.1 μmol/L tTF-RGD to endothelial cells was reduced to 25% in presence of 1 μmol/L synthetic peptide GRGDSP. These results underscore again the RGD dependency of tTF-RGD binding to receptors on endothelial cells, such as integrins α_vβ₃ or α_vβ₅.

Effect of truncated tissue factor-RGD on growth of M21 tumors in mice. The antitumor activity of the tTF-RGD fusion protein was determined in BALB/c nude mice bearing 300 to 1,000 mm³ M21 tumors. The controls and tTF-RGD were given i.v. five times at intervals of 24 hours. The pooled results of two independent experiments are presented in Fig. 2A and Table 1. At the doses used, a statistically significant reduction in tumor growth rate was observed after the second injection of tTF-RGD. This effect was sustained until the end of the experiment on day 7 for tTF-RGD in comparison with tTF (P = 0.016) or saline (P = 0.008). In detail, four of seven tTF-RGD–treated tumors showed a decrease in tumor volume of >40%, two tumors showed a decrease of 10%, and one tumor showed a tumor volume increase of 4%. Within the control groups, a tumor volume increase of at least 40% was observed in all tumors. No effect on tumor growth was observed with boiled tTF-RGD, ruling out a possible effect solely due to the RGD peptide (data not shown).

During treatment, tumors showed macroscopic signs of necrosis within 24 hours of treatment start similar to the results published by other groups (Fig. 3; refs. 17–19).

Effect of truncated tissue factor-RGD on growth of HT1080 tumors in mice. Furthermore, the antitumor activity of the tTF-RGD protein in BALB/c/nude mice bearing fibrosarcomas (HT1080) was evaluated. These tumors grow fast and are well vascularized. The pooled results of two independent experiments are represented in Fig. 2B and Table 1. After the second injection of tTF-RGD, a significant growth inhibition of the HT1080 tumors in comparison with control groups was observed. This effect was sustained until the end of the experiment on day 7 (P = 0.021 for tTF-RGD versus saline and P = 0.005 for tTF-RGD versus tTF). As it is shown in Table 1, control tumors showed a tumor volume increase of 30%, whereas tTF-RGD–treated tumors decreased ∼25% to 35% in size.

Effect of truncated tissue factor-RGD on growth of CCL185 tumors in mice. Additionally, the antitumor activity of the tTF-RGD protein in BALB/c/nude mice bearing 30 to 100 mm³ CCL185 tumors was evaluated. Because of the slower tumor growth in the experiment, the drug and the controls were given five times at intervals of 3 to 4 days. The pooled results of two independent experiments are represented in Fig. 2C and Table 1. After the fourth injection of tTF-RGD, a significant growth inhibition of the CCL185 tumors in comparison with control groups was observed. This effect was sustained until the end of the experiment on day 26 (P = 0.033 for tTF-RGD versus saline and P = 0.024 for tTF-RGD versus tTF). Histologic analysis revealed that >90% of tTF-RGD–treated tumors reached a regression grade of IIb and III (<10% vital tumor tissue), whereas all control tumors showed >10% of vital tumor tissue.

In all three different tumor models, no significant treatment-related toxicity in living mice was observed during or after cessation of treatment. Especially, no clinical signs indicating major embolism or stroke (dyspnea, asymmetrical movements, and sudden death) occurred. Furthermore, no deleterious side effects have been observed in immunocompetent mice (BALB/c) after injection of 100 μg tTF-RGD (=5 mg/kg body weight), a dose that is thrice higher than the dose used for tumor studies.

Truncated tissue factor-RGD reduces tumor blood flow as determined by magnetic resonance imaging. Contrast-enhanced MRI was done on mice bearing human fibrosarcoma before and 6 hours after application of tTF-RGD (30 μg in 200 μL NaCl) or saline, respectively. The perfused tumor volume (VVF) was determined by application of ultrasmall superparamagnetic iron oxide nanoparticles and the resulting change in transverse relaxation rate (ΔR2*) in comparison with muscle tissue. Color-coded imaging of mouse bearing human fibrosarcoma revealed significant reduced tumor blood flow after treatment with tTF-RGD in comparison with control animals (Fig. 4). Furthermore, the VVF of tumors treated with tTF-RGD was significantly lower than in the control group (1.22 ± 0.29, n = 9 for tTF-RGD versus 2.57 ± 0.29, n = 9 for NaCl; P < 0.005). Thus, the presumable mode of action of the tTF-RGD fusion protein (i.e., the thrombotic occlusion of tumor vessels) is underlined by these observations.

Besides, histologic studies revealed a substantial amount of tumor vessels, which were thrombosed, and confirmed the macroscopic impression of gross tumor necrosis (Fig. 5A and B). The high selectivity of the tTF-RGD for tumor blood vessels is shown by the fact that no visible thrombosis or necrosis occurred in normal tissues, such as heart, kidney, liver, and lung (Fig. 5E-H). Even repeated high doses of tTF-RGD (4 mg/kg bodyweight) did not cause any thrombosis or visible organ damage.

The initial reports on the selective induction of intratumoral thrombosis using tTF-antibody complexes or tTF fused to antibody fragments directed to artificial or natural markers of angiogenesis showed impressive efficacy in terms of tumor growth inhibition (17–19). However, this approach might have some inherent disadvantages. The bigger size of tTF-antibody complexes or even single-chain antibody fragments, which are fused to tTF, may result in sterical hindrance of tTF in terms of binding to factor VIIa and factor X. Thus, coagulation activation may be less efficient. Another, thus far only theoretical, limitation for the use of tTF fused to antibodies is the nonspecific uptake by the reticuloendothelial system, such as liver, spleen, and bone marrow, resulting in potentially deleterious induction of blood coagulation in these organs.

Cell surface–binding peptides are useful alternatives for targeting cancer and tumor vasculature as has been shown by different groups (10, 13, 16, 29, 31, 32). Promising candidates to target tTF to the tumor vasculature include small peptide fusion proteins selective for natural markers on tumor endothelium (33). Integrins α_vβ₃ and α_vβ₅ as well as other integrins have been identified as markers of activated endothelium and seem to play a crucial role in developmental and tumor angiogenesis (28, 30, 34–36). RGD peptide fusion proteins, which bind to these endothelial ligands, have been identified as promising agents to target the tumor endothelium (10). The smaller size of RGD peptides in comparison with previously used antibodies may yield some advantages.

Based on the known crystal structure of the tTF-factor VIIa complex (37), we hypothesized that the fusion of the small GRGDSP peptide to the COOH terminus of tTF would allow the generated tTF-peptide fusion protein to adopt an orientation perpendicular to the phospholipid membrane of the endothelial cell similar to native tissue factor. On the other hand, ligation of the small peptide to the COOH terminus of tTF should not result in sterical hindrance for the interaction of tTF with factor VIIa and its substrate factor X. Besides, by ligating the RGD peptide to the COOH terminus of tTF, we were able to express tTF-RGD without any further enzymatic steps for coupling. To rule out a possible effect on tumor growth solely due to the RGD sequence itself, experiments were repeated with tTF-RGD, which was boiled at 96°C for at least 10 minutes before administration. No effect on tumor growth was observed using this boiled fusion protein (data not shown), indicating that an intact tissue factor molecule is necessary to induce vessel thrombosis and achieve the demonstrated inhibitory effects on tumor growth.

Our fusion protein showed the expected functionality in vitro in terms of generating factor Xa comparable with previously published data (24, 38). tTF-RGD bound effectively and specifically to immobilized α_vβ₃ in a purified receptor-binding assay as well as on surface receptors of endothelial cells.

This report shows for the first time effective in vivo inhibition of tumor growth using the fusion protein tTF-RGD by selective thrombosis in the tumor vasculature. In vivo application to mice bearing the slow-growing human adenocarcinoma CCL185 resulted in significant tumor growth retardation. The faster-growing tumor models of human melanoma M21 and human fibrosarcoma HT1080 showed even significant regression in tumor size after treatment with tTF-RGD in comparison with controls. Histologic analysis of tTF-RGD–treated tumors revealed a regression score of IIb or III (<10% of vital tumor tissue) for 90% of each cohort. Inducing partial or complete thrombotic occlusion of tumor vessels as presumed mechanism is indicated by the histologic data. The fact that tTF fused to the RGD target moiety but not to the untargeted tTF induced thrombosis with subsequent tumor necrosis supports the hypothesis that, by binding of tTF to the cell surface of the tumor endothelium, tTF recovers in part its native function. Moreover, significant reduced tumor blood flow in tTF-RGD–treated animals as determined by contrast-enhanced MRI underlines the proposed mechanism.

Nevertheless, regrowth of tumors was observed after cessation of treatment. This phenomenon might be overcome by a better formulation or scheduling of the fusion protein or combination with other therapeutic modalities. As most approaches targeting the tumor vasculature with coaguligands, including the one presented here, could not completely inhibit tumor regrowth in the less vascularized superficial rim area, we would propose a combination therapy either with cytotoxic agents (e.g., doxorubicin) or local radiation. Besides these modifications, the use of cyclic RGD peptides might further enhance potency because of the known higher affinity of cyclic RGD peptides for integrins (39).

The specificity of our approach of targeting integrins on tumor endothelium is underlined by the observation of no apparent side effects, such as macroscopically or microscopically visible thrombosis or cell damage in liver, kidney, heart, or lung. Most likely, this is due to the differential expression of the targeted integrins in tumor endothelium versus endothelium of other organs. In addition, the higher phosphatidylserine expression within the tumor vasculature compared with the endothelium of normal organs might have contributed to the selectivity of this approach as suggested by others (18). Phosphatidylserine expression on the cell surface is essential for coagulation because it enables the binding of coagulation factors, such as factor VIIa and factor X, thus coordinating the assembly of the coagulation initiation complexes (40–42). It has recently been shown that normal tissues segregate phosphatidylserine to the inner surface of the plasma membrane phospholipid bilayer where it is unable to participate in thrombotic reactions, whereas tumor endothelial cells translocate phosphatidylserine to the external surface of the plasma membrane, thus supporting coagulation activation by tTF bound to tumor endothelium (18). In fact, antibodies against vascular cell adhesion molecule 1 conjugated to phosphatidylserine containing liposomes have been successfully used to induce selective thrombosis within the vasculature of experimental tumors (43).

Of course, it is a major concern that activation of coagulation occurs when administer the tTF-RGD fusion protein. Although, as mentioned above, we did not observe macroscopically or microscopically visible thrombosis or cell damage in liver, kidney, heart, or lung, a certain degree of coagulation activation might have occurred, thus increasing the risk of thrombosis or cardiovascular complications. However, we do not expect a major effect on clotting variables on application of the tTF-RGD fusion protein for several reasons. First, the truncated form of tissue factor reveals only 1:1,000 of the activity of the full-length form while in solution (44). Secondly, α_IIbβ₃, the receptor for RGD recognition sequences of fibrinogen and von Willebrand factor on platelets, as well as platelet α_vβ₃ have a very low affinity and avidity toward its ligands on resting platelets (45–47). Therefore, significant binding of tTF-RGD on unstimulated platelets with subsequent coagulation activation should not occur. Third, phosphatidylserine expression is low in unstimulated platelets, thus reducing the ability of platelets to promote coagulation activation (48).

The significant antitumor activity of tTF-RGD is in apparent contrast to the data of Hu et al., who were able to induce thrombosis only in smaller-sized vessels of experimental lung and colon carcinomas but could not show any significant effect with regard to tumor growth retardation (21). The group speculated that the lack of efficacy of their RGD-tTF fusion protein might be explained by the distribution of the target receptor within the tumor vasculature. Integrins targeted by RGD are mostly expressed on newly formed smaller-sized vessels in which selective thrombosis causes lesser necrosis in the tumor. Our data might indicate that this limitation can be partly overcome by increasing the dose of tTF-RGD. Using an average of 30 μg per animal and injection, we were able to show significant tumor necrosis and retardation in contrast to the 10 μg used by Hu et al. This suggests that not only the distribution of targeted integrins might be the limiting factor but also the low affinity of the ligand to the receptor causing less accumulation of the therapeutic agent within the tumor. The latter can possibly be partly overcome by increasing the absolute amount of injected tTF-RGD.

However, the tremendous difference in efficacy between RGD-tTF (21) and our generated tTF-RGD fusion protein might be alternatively explained by the fact that the RGD targeting moiety is fused to the COOH terminus of tTF in our study in contrast to the published study in which the RGD sequence is coupled to the NH₂ terminus (21). Thus, binding of RGD-tTF to the luminal surface of the tumor endothelium mediated by its NH₂-terminal region is probably not appropriate for the interaction with factor VIIa and its substrate factor X on the phospholipid membrane. Indeed, when considering the known crystal structure of the tTF-factor VIIa complex, linkage of the RGD peptide to the NH₂ terminus of tTF would cause an orientation that is opposite to the orientation of native tTF. Therefore, the cofactor activity of RGD-tTF (21) is presumably much lower than the cofactor activity of our generated tTF-RGD when bound to integrins on the tumor endothelium. As a consequence, the rate of thrombin generation induced by RGD-tTF bound to integrins may not be sufficient to occlude medium and large-sized vessels given the tremendously lower tumor endothelial cell surface to blood volume ratios based on geometric considerations. The apparent low toxicity of the tTF-RGD fusion protein might be due to its relatively high specificity for tumor endothelium and its lack of nonspecific uptake by the reticuloendothelial system such as liver, spleen, and bone marrow. However, only a head-to-head comparison of different tTF-based vascular targeting approaches would allow the selection of the agent with the best efficacy/toxicity ratio.

In summary, we have shown that tTF-RGD, targeting integrins on tumor endothelium, can induce extensive thrombosis in tumor vessels, thus leading to effective antitumor therapy in three experimental solid tumor models in mice. Additional studies are warranted to integrate this promising approach in current treatment strategies employing chemotherapy, radiotherapy, or other molecular targeted therapy of cancer.