Cancer Research 2010 AACR Elections Sign up for Cancer Research eTOC's
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by El-Sheikh, A.
Right arrow Articles by Edgington, T. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by El-Sheikh, A.
Right arrow Articles by Edgington, T. S.
[Cancer Research 65, 11109-11117, December 1, 2005]
© 2005 American Association for Cancer Research

Experimental Therapeutics, Molecular Targets, and Chemical Biology

A Selective Tumor Microvasculature Thrombogen that Targets a Novel Receptor Complex in the Tumor Angiogenic Microenvironment

Amr El-Sheikh1, Per Borgstrom2, Gourab Bhattacharjee1, Mattias Belting1,3 and Thomas S. Edgington1

1 Department of Immunology, The Scripps Research Institute; 2 Vascular Biology and Angiogenesis Program, The Sidney Kimmel Cancer Center, La Jolla, California and 3 Department of Cell and Molecular Biology, Lund University, Lund, Sweden

Requests for reprints: Thomas S. Edgington, Department of Immunology (SP258), The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037. Phone: 858-784-8225; Fax: 858-784-8480; E-mail: tse{at}scripps.edu.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
We have previously shown that part of the heparin-binding domain of the vascular endothelial growth factor (VEGF), designated HBDt, localizes very selectively to surfaces of the endothelial cells of i.t blood vessels. Here, we have coupled the HBDt to the extracellular domain of tissue factor (TFt), to locally initiate the thrombogenic cascade. In tumor-bearing mice, infusion of this HBDt.TFt results in rapid occlusive thrombosis selective only for tumor microvasculature with resultant infarctive destruction of tumors. We now show that infusion of an optimal combination of this HBDt.TFt and its requisite cofactor (factor VIIa) in tumor models results in significant tumor eradication. Binding studies and confocal microscopy indicate that the target for the HBDt.TFt seems to be a trimolecular complex of chondroitin C sulfate proteoglycan, neuropilin-1, and VEGF receptor-2, overexpressed together only in highly angiogenic sites of the tumor microenvironment. The HBDt.TFt was also colocalized with the trimolecular receptor complex in endothelial sprouts from tumor tissues, and its binding inhibited the growth of such sprouts. In vitro, we show that the HBDt structure has its highest affinity for chondroitin 6 sulfate. We show the potential of this HBDt.TFt as a candidate therapeutic and elucidate its target in vivo.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
To search and discover novel targets specific for tumor vascular endothelium is a challenging task but a critical aspect of developing more effective tumor therapeutics. The existence of such targets expands on our perception of the uniqueness of the tumor microenvironment. To be an optimally effective target, it should not only be expressed selectively by tumor microvasculature but should also be a common feature found in a broad spectrum of solid tumors.

We previously showed that the rather novel vascular endothelial growth factor (VEGF) heparin-binding domain substructure (HBDt), when expressed on the surface of a modified M13 phage, selectively enabled accumulation in vivo on the vascular surface of the CD31-positive endothelium of tumor vasculature (1). The VEGF165 isoform is known to bind to VEGF receptor-2 (VEGFR-2, KDR), neuropilin-1 (Npn-1), and an unknown heparan sulfate (HS) proteoglycan (2). In addition to VEGFR-2, Npn-1 is also known to be highly expressed by endothelial and tumor cells (3) and is known to play important roles in endothelial mitogenesis, tumor cell migration, and invasion (4–7). The role of Npn-1 expression in tumor angiogenesis has been observed in several types of tumors (4, 8). Furthermore, it was reported that expression of Npn-1 in tumors seems to correlate with a shift from benign stromal tissue to that associated with malignancies (9).

The VEGF165 isoform, which contains HBDt, is known to participate in the formation of molecular complexes, containing VEGFR-2 and Npn-1. The HBDt-encoded VEGF/exon 7 is thought to interact with the Npn-1 b1b2 domain (10–12). The b1b2 region of Npn-1 was also analyzed and shown to be a heparin-binding domain, which enhanced the binding of VEGF165 to Npn-1 (13). The size of the oligosaccharide fragment needed for enhancement of binding was at least 20-monosaccharide unit in length, suggesting the potential involvement of a large polysaccharide chain, possibly a chondroitin sulfate proteoglycan (CSPG).

We fused the HBDt structure to the extracellular domain of human tissue factor (TFt) through a linker to explore the feasibility of targeting the truncated extracellular domain of TFt to the tumor microvessel plasmalemma in a mode to locally initiate the thrombogenic cascade. Local infarctive necrosis of tissue follows loss of local vascular perfusion; a concept conserved in evolution from marine invertebrates to mammals.

We have now analyzed the ability of the HBDt.TFt fusion protein to bind heparin and other carbohydrate moieties in vitro and to endothelial cells in culture. We also investigated the effect of HBDt.TFt addition on angiogenic sprouting ex vivo. We also analyzed the complex dynamics of targeting of the HBDt.TFt protein to tumor intravascular microenvironment when infused in vivo in various murine tumor models, monitoring thrombus formation by real-time intravital microscopy. We investigated the effect of HBDt.TFt administration and the resultant thrombosis on the destruction of tumors. Finally, we confirmed docking of HBDt.TFt in the i.t. vasculature and colocalization with its target receptor molecules by confocal microscopy in vivo and ex vivo. Our data show that HBDt.TFt accumulates in i.t. vasculature preferentially, attributable to recognition of a unique target, and docks on tumor endothelium plasmalemma, in a manner to permit functional initiation of the coagulation cascade. This results in vascular occlusion of tumors, resulting in local infarctive necrosis and significantly reducing the mass of solid tumors. We support the potential of this strategy for therapeutic purposes and illustrate the unique biology of the tumor microenvironment.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
HBDt.TFt Cloning, Expression, and Purification
The VEGF165 HBDt wild-type and mutant sequences (1) were synthesized and fused to the extracellular domain of tissue factor (amino acids 1-211). The two sequences were separated by a 15-amino-acid inert sequence [3(4G-S)] and cloned into the BamHI and HindIII sites of pTrcHis/CAT (Invitrogen Corp., Carlsbad, CA). This was expressed in BL-21 Star (DE3) pLysS One Shot (Invitrogen). The protein was isolated from the inclusion bodies of Escherichia coli, analyzed by gel electrophoresis, and folded in folding buffer (14, 15). Fast protein liquid chromatography was used to purify the HBDt.TFt proteins, using MonoQ and HIC hydrophobicity columns (Amersham Biosciences, Inc., Piscataway, NJ). The purified protein was analyzed for tissue factor function in a Xa generation assay and by binding activity to heparin matrix. Heparin matrix beads (Sigma, St. Louis, MO), 10% (v/v) in 0.2 mol/L NaCl, and 20 mN Na2HPO4 (pH 7.2) were mixed with 10 µg of protein, with agitation at room temperature for 1 hour; centrifuged briefly to precipitate the beads with bound protein; and washed with saline. The amount of endotoxin in the protein preparation was analyzed using the Pyrochrome chromogenic test (Cape Cod, Inc., East Falmouth, MA) and determined to be 30.2, 42.1, and 70.23 pg/mL in the preparations of HBDt.TFt, X.TFt, and TFt, respectively.

Heparin and Oligosaccharide Binding
Heparin matrix beads (Sigma), 10% (v/v) in 0.2 mol/L NaCl, and 20 mN Na2HPO4 (pH 7.2) were mixed with purified HBDt.TFt (10 µg) with agitation at 4°C overnight and centrifuged briefly to precipitate the beads with bound protein. After several washes with PBS, 100 µg of competitor oligosaccharides were added and incubated at 4°C overnight. An aliquot (1:20) of the beads was used for Western blot analysis and detected with monoclonal mouse anti-TFt antibody 10H10, which is dependent on the conformation integrity of the protein (16). The bands were quantified using Scion software (Scion Corp., Frederick, MD).

Microscopy
Tissues were embedded in Tissue-Tek ornithine carbamyl transferase (Sakura Finetechnical Ltd., Tokyo, Japan) and frozen, and 5-mm sections were mounted on positively charged slides (Fisher Scientific, Tustin, CA). Sections were fixed in cold acetone for 30 seconds, blocked with 1:10 porcine serum (30 minutes), and washed at room temperature in saline. Primary antibodies used included rat biotin monoclonal antimouse CD31 (BD Biosciences PharMingen, La Jolla, CA), mouse monoclonal anti-fibroblast growth factor receptor (FGFR; Chemicon, Temecula, CA), mouse monoclonal anti-FLK/KDR and VEGFR-2 (Chemicon, Temecula, CA), mouse monoclonal anti-HS and anti-CS (Seikagaku America Associate of Cape Cod, Falmouth, MA), rabbit polyclonal anti-neuropilin-1 (Zymed Laboratories, Inc., South Francisco, CA, and Sigma-Aldrich, Inc., St. Louis, MO), and rabbit polyclonal anti-neuropilin 2a (Zymed Laboratories). Secondary antibodies used included Texas red streptavidin, horse anti-mouse IgG, goat anti-rabbit, goat anti-mouse IgM, (Vector Labs, Burlingame, CA), and Alexa Blue-647 (Molecular Probes, Eugene, OR). Antibodies (1:100-300) were incubated with sections for 30 minutes, washed with saline, and mounted in Vectashield (Vector Labs). A Zeiss confocal microscope (objective lens power x40) and Software Bio-Rad 1024 (Hercules, CA) were used to produce images.

Intravital Microscopy
Fluorescence microscopy was done using a microscope (Mikron Instrument, San Diego, CA) equipped with epi-illuminator and video-triggered stroboscopic illumination from a xenon arc lamp (MV-7600, EG&G, Salem, MA) with attached silicon-intensified target camera (SIT68, Dage-MTI, Michigan City, IN). A Hamamatsu image processor (Argus 20) with firmware version 2.50 (Hamamatsu Photonic System, Bridgewater, NJ) was used for image enhancement and to capture image data to a computer database. A Leitz PL1/0.04 objective was used to obtain an overview of the chamber and for determination of tumor size. A Zeiss long-distance objective 10/0.22 was used to capture images for calculation of vascular variables. A Zeiss Achroplan x20/0.5 W objective was used to capture images for calculation of mitotic and apoptotic indices. To quantify tumor area (AT), it is defined as number of pixels with photo density above 75 (256 Gy levels):

Cells
Generation of H2B-eGFPN1 cell lines. Human histone H2B-eGFP, a kind gift from Dr. Kanda (Salk Institute, La Jolla, CA), was excised from the SalI/NotI sites in the BOS H2BGFP-N1 vector. The H2B-eGFP retroviral vector was generated by cutting first with NotI followed by Mung-Bean Nuclease digestion to remove the overhanging end. The fragment was gel purified followed by SalI digestion. The resultant SalI/blunt fragment was cloned into the SalI/HpaI sites in the LXRN retroviral vector (Clontech, Palo Alto CA). VSVG-pseudotyped retrovirus was generated by transfecting GP293 GagPol-expressing 293 cells (Clontech) in T75 flasks at 75% confluence with 5 µg of this resultant L-H2BGFP-RN-L vector with 5 µg VSV-G vector (Clontech) and harvesting viral supernatants 48 hours after transfection. Retroviral supernatants were used immediately or concentrated by centrifugation and stored at –80°C until use. Tumor cells were transduced with VSVG-pseudotyped H2B-GFP LXRN viral supernatant for 48 hours in 8 µg/mL polybrene followed by selection in 300 µg/mL G418 for 2 weeks. Pooled H2B-GFP-expressing tumor cells were usually heterogeneous as determined by fluorescent microscopy but could be sorted via fluorescence-activated cell sorting if necessary as for resistant cell lines, such as Lewis lung carcinoma. Most cells used for chamber studies were passaged in DMEM, 4.5 g/L glucose supplemented with pyruvate, glutamine, nonessential amino acids, and gentamicin (50 µg/mL) and maintained in a humidified 5% CO2 atmosphere at 37°C. Cells are routinely tested for Mycoplasma contamination with the Genprobe Mycoplasma detection kit.

Preparation of tumor spheroids. Liquid overlay plates were generated using 1% molecular biology grade agarose melted in DMEM in a microwave. The melted agarose solution was plated into round-bottomed 96-well plates at 50 µL/well and allowed to cool at room temperature. Tumor cells grown as preconfluent monolayers were trypsinized and diluted to a final volume of 250,000 tumor cells/mL. Viability was determined using trypan blue. Tumor suspensions were dispersed at 100 µL/well into the 96-well agarose-coated round-bottomed plates. Spheroids were allowed to compact for 48 hours and picked from the plates followed by transfer into serum-free media immediately before implantation into chambers.

Animals
The dorsal skin fold chamber in nude mice was prepared as previously described (17), and all surgical procedures were done under strict sterile conditions. Male mice (25-35 g body weight) were anesthetized (7.3 mg ketamine hydrochloride and 2.3 mg xylazine/100 g body weight, i.p.) and placed on a heating pad. Two symmetrical titanium frames were implanted into a dorsal skin fold, so as to sandwich the extended double layer of skin. A 15-mm full thickness layer was excised. The underlying muscle (M. cutaneous max.) and s.c. tissues were covered with a glass coverslip incorporated in one of the frames. After a recovery period of 2 to 7 days, tumor spheroids were carefully placed in the chamber. Syngeneic CT26 colon carcinoma in BALB/c mice and tissues were prepared as described before (1).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
In vitro investigation of the HBDt.TFt protein. We evaluated the HBDt.TFt construct in vitro by analyzing the presence and integrity of the HBDt domain at the NH2 terminus and the TFt domain at the COOH terminus of the fusion protein. Figure 1A depicts Western blot analysis of the binding of HBDt.TFt (lane 3) to heparin matrix compared with that of TFt alone (lane 1). In addition, Fig. 2B shows the relative amounts of protein bound to heparin beads. No binding of TFt to the heparin matrix was observed. The HBDt.TFt bound well to the heparin matrix and significantly more than the mutant HBDtMut.TFt (lane 2) carrying the substitution of Cys24 by glycine as described previously (1). We examined the ability of various oligosaccharides to competitively displace the bound HBDt.TFt from the heparin beads. Various oligosaccharides (lanes 4-10) were added to the bound HBDt.TFt on the heparin matrix and incubated overnight, then analysis of the amount of protein remaining bound to the washed heparin matrix was done by Western blot. We observed that chondroitin C sulfate (i.e., chondroitin 6 sulfate or C6S) was the most potent competitor for binding to the HBDt followed by structurally similar chondroitins A and D. Taken together, the data indicates that the HBDt binding function is dependent on the three-dimensional structure of the HBDt domain and that C6S has the greatest affinity for this domain. We also examined the ability of the TFt domain to initiate coagulation by the Xa generation assay (data not shown).



View larger version (42K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. A, Western blot analysis showing the amount of TFt alone (lane 1), mutated HBDt.TFt (lane 2), and HBDt.TFt alone (lane 3) or after competition with different oligosaccharides (lanes 4-10) bound to heparin matrix. B, proteins (10 µg) were added to heparin beads and allowed to bind overnight at 4°C, then 100 µg of an oligosaccharide was added and incubated overnight at 4°C. Beads were washed and an aliquot (1:20) of the beads was taken for analysis, and protein was detected with mouse anti-TFt antibody.

 


View larger version (59K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. HBDt.TFt inhibits ex vivo angiogenesis by tumor endothelial cells. HBDt.TFt or TFt was added at 20 µg/mL to aortic rings of mice (A-C) from BALB/c, or to tumor (CT26) pieces (F-G), grown ex vivo in Matrigel. C, effect of HBDt.TFt addition to aortic endothelial cells during angiogenesis ex vivo at day 4. A, sprouting pattern obtained normally. B, effect of added TFt alone. D, summary of number of sprouts from aortic rings obtained on day 4. Columns, average of three independent experiments. G, effect and binding of HBDt.TFt to CT26 tumor sprouting using confocal microscopy. E, sprouting pattern obtained normally. F, effect of TFt alone. K, summary of inhibitory effect of HBDt.TFt addition on the number of sprouts obtained during ex vivo angiogenesis by CT26 tumor fragments from BALB/c mice. Columns, average of three independent experiments. HBDt.TFt and TF were visualized using FITC-labeled anti-TFt antibody (green). CD31 was visualized using Texas red–labeled anti-mouse CD31 antibody. Nuclei were visualized using Topro-3 dye (blue). Arrows, triple colocalization (white). H-I, sprouting of CT26 pieces. White, triple colocalization of HBDt.TFT (FITC green) with Texas red–labeled receptor complex molecules: VEGFR-2 (G), Npn-1 (H), and C6S (I), along with the Topro-3 nuclear stain (blue).

 
The effect of the HBDt domain on ex vivo angiogenic sprouting by tumor endothelium. The effect of HBDt.TFt was examined on vascular sprouting using mouse aorta and murine mammary carcinoma (CT26) and subsequently quantitating angiogenic sprouts in culture. A previously established protocol (18) for aortic ring sprouting was used. Figure 2 depicts the normal mouse aortic ring sprouting pattern in Matrigel and media containing syngeneic mouse serum and exogenous growth factors (A) VEGF and basic FGF (bFGF) compared with that with the addition of (B) TFt or (C) HBDt.TFt. Addition of excess HBDt.TFt reduced the number of aortic sprouts by ~14.6% by day 4 compared with the extracellular domain of tissue factor (TFt) alone (Fig. 2D). However, the difference is not statistically significant. On the other hand, Fig. 2E-G shows the endothelial sprouting patterns from tumor fragments under comparable conditions. The inhibitory effect of HBDt.TFt addition to sprouting using CT26 tumor fragments was greater (65.7%) and more statistically significant (P < 0.05; Fig. 2K). We also examined by confocal microscopy the colocalization of HBDt.TFt with tumor endothelial cells during sprouting. Figure 2G shows the colocalization (arrows) of FITC-labeled HBDt.TFt (green) with Texas red–labeled CD31 (red). The colocalization color is white due to the effect of the blue nuclear stain. The binding of HBDt.TFt is observed in sites where sprouting is inhibited (Fig. 2G, arrows). Colocalization of HBDt.TFt to its target in tumor fragment sprouts was analyzed. Figure 2H-J shows the colocalization of FITC-labeled HBDt.TFt (green) with Texas red–labeled VEGFR-2 (H), Npn-1 (I), and C6S (J) in sprouting tumor fragments. These data show that the target (VEGFR-2 + Npn-1) of the HBDt domain exists in active angiogenic areas of tumors, which are rich in C6S. It also confirms that the binding of the HBDt to its target can inhibit a biological function needed for endothelial cell proliferation albeit ex vivo. Similar results were also obtained using the Lewis lung murine tumor model (data not shown).

In vivo infusion of HBDt.TFt, biodistribution, thrombosis, and tumor necrosis. To examine the immediate effect of HBDt.TFt infusion with respect to the ability to occlusively thrombose i.t. vessels of tumors, we used intravital microscopy. The HBDt.TFt protein was infused in live unanesthetized nude mice carrying micro green fluorescent protein (GFP)–positive murine tumors implanted in dorsal skin fold chambers. The microcirculation of tumors and surrounding areas was monitored using light and fluorescence microscopy, and the degree of cell death was based on disappearance of viable GFP-positive cells. Optimization of the amount of HBDt.TFt and the rate of infusion were critical factors to achieve desirable results. Therefore, we did a complex set of experiments to determine how each variable contributed to the efficacy of tumor-selective thrombosis. We addressed one variable constant at a time and examined the effect on thrombus formation where the result was described by an arbitrary thrombosis index based on speed and stability of thrombus and lack of hemorrhage in tumor or other organs. Figure 3A shows an example of one such relationship between the amount of HBDt.TFt and factor VIIa (FVIIa) infused and the degree and quality of thrombosis produced. Thrombosis was achieved with HBDt.TFt alone infused at 2.5 µg/g body weight infused at a constant rate over 120 seconds. Lower doses did not thrombose, and higher doses were associated with local hemorrhage. Thrombosis was observed within 3 to 4 minutes following infusion. Extension of the infusion time beyond 120 seconds, even with higher doses or higher rates, was associated with failure of generation of stable thrombosis in tumors. Addition of recombinant human FVIIa reduced the dosage of the HBDt.TFt by ~80%. The optimum (immediate and stable, with no peripheral hemorrhage) thrombus was achieved with 0.5 µg HBDt.TFt + 0.25 µg of FVIIa/g body weight infusion at a constant rate at 120 seconds. This ratio reflects a 4:1 molar ratio of TFt to FVIIa, contrary to what was the predicted ratio (i.e., 1:1). With this combination, thrombosis was achieved at the end of infusion, and the thrombi were stable for at least 30 minutes of observation. Infusion of equimolar amounts of HBDt.TFt and FVIIa resulted in thrombolysis after 30 minutes, and infusion of a 2:1 molar ratio of HBDt.TFt to FVIIa resulted in thrombolysis in 15 minutes. Therefore, there seems to be an inverse relationship between the amount of FVIIa infused if >0.25 µg of FVIIa/g body weight and the stability of thrombi. Interestingly, decreasing the amount of FVIIa below 0.25 µg of FVIIa/g body weight (but not below 0.05 µg of FVIIa/g body weight) resulted in slower formation but more stable thrombosis.



View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. A, relationship among the dose of infused HBDt.TFT, its requisite ligand FVIIa, and the quality of the resultant thrombus in the tumor microvasculature. HBDt.TFt was infused in live unanesthetized nude mice carrying micro GFP-positive murine tumors implanted in dorsal skin fold chambers. The blood microcirculation of tumors and surrounding areas was monitored through a light and fluorescent microscope, and optimization of the amount of HBDt.TFt and rate of infusion were done. An example of the relationship between the amounts of HBDt.TFt (open) and FVIIa (gray) infused and the quality of resulting thrombosis (black). X-axis, experiment number on top of the table containing the amounts of proteins, HBDt.TFt and VIIa, infused [µg/min/g body weight (BW)]. Thrombosis index (bottom row). Y-axis, amounts of protein infused or thrombosis index. The arbitrary thrombosis index is based on (i) the speed and stability of thrombus and (ii) lack of hemorrhage in tumor or other organs. Optimum, fast, and stable thrombosis was achieved with 2.5 µg/g body weight infused at a rate of 1.25 µg/min/g body weight. Thrombosis was observed within 3 to 4 minutes after infusion. With respect to FVIIa, optimal tumor-specific thrombosis was achieved with FVIIa at 0.05 µg/min/g body weight. A combination of 0.5 µg HBDt.TFt + 0.05 µg FVIIa/g body weight infused by tail vein over a 120-second interval was used in analysis of breast carcinoma N202-bearing animals for evaluation of arrest and/or reduction of tumor volume. B, effect of HBDt.TFt treatment on murine mammary carcinoma N202. Nude mice carrying N202 tumors in the skin window were infused with 0.5 µg HBDt.TFt + 0.05 µg of FVIIa/g body weight on days 0 and 3. Healthy tumor cells were visualized by fluorescent emission from the GFP, with 488-nm excitation. Progressive cell death (dark) as a result of the HBDt.TFt infusion during the course of 6 days. Healthy tumor cells were visualized by fluorescent microscopy (light), and the live tumor cell mass was represented as a percentage of the original starting signal on day 0 (or day 13 after tumor implantation). Points, time course of a group of tumors (n = 4).

 
To examine the effect of HBDt administration on tumors, we infused a combination of 0.5 µg HBDt.TFt TFt + 0.05 µg FVIIa/g body weight by tail vein over a 120-second interval in breast carcinoma N202-bearing animals for evaluation of arrest of growth and/or reduction of tumor volume. Figure 3B shows the progressive disappearance of healthy live tumor cells following treatment. Healthy tumor cells are visualized by fluorescence due to their stable transformation with GFP-labeled H2B DNA. Visualization of tumors in this case was facilitated by their discus shape. The tumors shown in Fig. 3B have a diameter of ~2 mm and a height of ~1 mm.

Injections were administered twice on days 0 and 3, because it was observed that the residual tumor started to regrow on day 3. Figure 3B shows the change in N202 mammary tumor size of a group (n = 4) of tumors treated with the combination of 0.5 µg HBDt.TFt + 0.05 µg FVIIa/g body weight infused by tail vein over a 120-second interval. Tumor size was estimated by quantification of fluorescent emission of viable tumor cells, in contrast to the dark (necrotic or apoptotic cells). Injections were done on day 0 (day 13 after implantation) and day 4 (day 17 after implantation). Tumor size was reduced from 2 mm3 (100%) on day 0 by ~90% in 1 week with two administrations.

To examine whether the HBDt domain can target the thrombogenic TFt domain in a functional assembly to tumor vasculature, we analyzed several organs for the presence of the infused protein. We injected BALB/c mice carrying colon tumor (CT26) with the HBDt.TFt construct, and several tissues were harvested after certain time points. HBDt.TFt was detected in those organs by Western blot using anti-human TFt antibody, specific for the infused human truncated TFt and not the endogenous mouse TF at 5 minutes (Fig. 4A) and at 60 minutes (Fig. 4B). The graph (Fig. 4) shows the amount of HBDt.TFt detected in tumor, brain, liver, spleen, kidney, heart, and muscle 5 and 60 minutes after infusion. The results show the immediate and consistent accumulation of the infused construct in the tumor. Initially, there was a transient accumulation of HBDt.TFt in the liver and kidney at 5 minutes. However, by 60 minutes, the protein was detected only in tumor tissues. The protein was still detected in blood at approximately the same level at the latter time point (data not shown). This shows that the targeting HBDt domain enabled selective, immediate, and stable accumulation of the construct in the tumor, whereas the transient accumulation of the construct in the liver and kidney was subject to clearance mechanisms. Infusion of TFt at similar levels, as a control, instead of the HBDt.TFt was only detected in blood and not detected in tumor or other organs at 60 minutes after infusion.



View larger version (46K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. Biodistribution of infused HBDt.TFt in the syngeneic murine colon tumor model. BALB/c mice carrying the CT26 tumor were infused i.v. with HBDt.TFt, and tissues were harvested and analyzed by Western blot. The protein was detected using anti-human tissue factor antibody, specific for the infused human truncated tissue factor and not the endogenous murine tissue factor, after 5 minutes (A) and 60 minutes (B). C, bands were quantified by video densitometry and the results summarized.

 
We also analyzed the tissues harvested from the same animals histologically for thrombosis. Figure 5A shows the amount of thrombus formed after 5 minutes of infusion of the HBDt.TFt construct. At this time point, thrombus formation is evident in the vasculature associated with the periphery of the tumor, whereas the smaller vasculature associated with the inner core of the tumor mass appears thrombus-free. Peripheral tumor thrombosis was stable at 60 minutes after infusion (Fig. 5B) and extended inwardly to more central vasculature. In comparison, infusion of TFt instead of HBDt.TFt was associated with no thrombus formation in tumors or elsewhere even after 60 minutes (Fig. 5C). Vascular thrombosis was absent in all other tissues examined at both 5 and 60 time points after infusion, as shown in Fig. 5D-J of sections from the brain, heart, liver, spleen, kidney, lung, and muscle at 5 minutes. The absence of thrombus in the kidney and liver at the 5-minute mark (Fig. 5F and G) confirms that the small amounts of tissue factor detected by Western blot analysis were due to clearance and not associated with local thrombogenesis.



View larger version (83K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5. Histologic analysis of thrombus formation resulting from HBDt.TFT infusion in the CT26 syngeneic colon tumor model. A, amount of thrombus formed after 5 minutes of infusion of HBDt.TFt; thrombus formation is evident in the vasculature associated with the more peripheral aspect of the tumor mass, whereas smaller vessels in the more central vasculature of the tumor seem to be thrombus-free. Thrombus formation in the periphery remained stable at 60 minutes (B) and seemed to extend to the central vasculature. By comparison, infusion of TFt, rather than HBDt.TFt, had no ability to thrombose tumor vasculature even after 60 minutes (C). Thrombi were absent from all other tissues examined at both 5- and 60-minute time points following infusion. D-J, sections from the brain, heart, liver, kidney, spleen, lung, and muscle, respectively, at 5 minutes.

 
Colocalization of infused HBDt.TFt with endothelial surface molecules. We observed the binding of HBDt.TFt to VEGFR-2 (KDR), and Npn- 1, and C6S on CD31-positive endothelial cells from the sprouts of CT26 tumor fragments (Fig. 2E-G). We also did confocal microscopy studies on tissues from the dorsal skin fold microtumors and from tumor-bearing mice to address the colocalization of HBDt.TFt with possible targets, such as CD31, bFGFR, VEGFR-2 (KDR), Npn- 1, and Npn-2, in tumor-associated vascular channels and elsewhere in other tissues and organs. We also examined the colocalization of the thrombogen with oligosaccharides, HS, and chondroitin C sulfate (i.e., C6S), which may be interacting with the HBDt in vivo. In Fig. 6, the colocalization, or lack thereof, of the infused HBDt.TFt with different cell surface markers is depicted. We observed the colocalization of the infused labeled HBDt.TFt (blue) with labeled VEGFR-2 (red; Fig. 6D), Npn-1 (red; Fig. 6E), and C6S (red; Fig. 6H), whereas no colocalization was observed for bFGFR (Fig. 6C), Npn-2 (Fig. 6F), or HS (Fig. 6G). In the case of the endothelial cell marker CD31, the localization of HBDt.TFt in this (Fig. 6B) and other models (data not shown) was observed in close proximity within the CD31-positive vascular channels but not directly with the endothelial receptor. Negative controls for fluorescent secondary antibodies were also included in the study. Figure 6I depicts a positive stain for HBDt.TFt with negative IgG control for VEGFR-2, whereas Fig. 6J depicts the opposite.



View larger version (70K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6. Colocalization of HBDt.TFt with relevant cell surface molecules in the N202 murine mammary tumor model. Nude mice carrying the murine N202 tumor in their skin window were infused with 0.5 µg HBDt.TFT + 0.05 µg FVIIa/g body weight at a constant rate for 120 seconds. One hour later, the animals were sacrificed, and tissues were harvested. A, H&E stain of the N202 tumor, including the vascular thrombosis in the center and periphery (angiogenic corona) of part of the N202 tumor disc. B-H, confocal microscopic images, where the HBDt.TFT was visualized using Alexa 647 (blue) and Texas red–labeled different surface markers, as well as the GFP-positive tumor cells (green). The different cell surface markers examined are the endothelial marker CD31 (B), bFGFR (C), VEGFR-2 (D), Npn-1 (E), Npn-2 (F), HP (G), and C6S (H), each in red. Colocalization between the blue-labeled HBDt.TFT and the red-labeled cell surface markers renders as purple. Nonbinding controls for Texas red (I) and Alexa 647 IgGs (J).

 
Similar results were obtained for the CT26 and Lewis lung tumor models infused with the HBDt.TFt. Only results from the N202 model are shown because the architectural organization of the N202 tumor spheroidal discs allows better assessment of colocalization with respect to the degree of vessel maturity and tumor topography. The results confirm the expression of the HBDt target in the highly angiogenic area surrounding the tumor mass (periphery) known to be rich with newly formed vessels. The center of tumors with its mature vessels does not appear to express the target or only weakly.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
We previously reported the accumulation of a modified M13 phage in CD31-positive channels within tumors, which was directed by expression of HBDt on the surface of the phage. We also reported that the novelty of this domain, unlike most known heparin-binding domains, is a result of its three-dimensional structure-function relationship rather than simply positively charged amino acids. Here, (a) we show the high degree of selectivity for this HBDt domain for tumor vasculature; (b) provide evidence for the novelty of its receptor; and (c) illustrate the potential use of this HBDt domain for targeting tumor therapeutics.

We cloned a fusion protein with the HBDt as the NH2-terminal domain and the extracellular domain of TFt as the COOH-terminal domain separated by a short spacer sequence) in the pTricHis vector and expressed it in E. coli. We have analyzed the ability of the FLPC-purified protein to bind heparin and other oligosaccharide moieties and functionally to initiate the coagulation protease cascade in vitro. We also determined the amount of endotoxin in the preparations of the protein and its proper controls to confirm its safety and lack of involvement in the observations.

Fusion of the TFt to HBDt allowed the detection, visually and functionally, of the protein in vitro and in vivo. Furthermore, it endowed the targeting HBDt domain to localize the thrombogenic TFt domain in a functional local context supportive for the active thrombogenic cascade. We have found that the basis of novelty of HBDt and its selectivity for tumor microenvironment lies in its high affinity for C6S (Fig. 1). For example, we found that HBDt binds to C6S in the sprouting endothelium from tumor fragments ex vivo (Fig. 2) and inhibits the growth of the sprouts and proliferation of the CD31-positive endothelium in the tumor environment. The biological significance of the proposed trimolecular target complex in tumor neoangiogenesis was shown in this sprouting assay. Colocalization between HBDt.TFt with VEGFR-2 (Fig. 2H), Npn-1 (Fig. 2I), and C6S (Fig. 2J) shows that HBDt recognizes its trimolecular target in tumor neoangiogenesis, and we propose that by competing with the natural ligand (VEGF165), it inhibits endothelial cell proliferation.

The selectivity of HBDt for tumors neoangiogenic vessels in vivo is observed in experiments (Figs. 4 and 5), where the infused HBDt.TFt is exclusively detected in the syngeneic murine colon tumor (CT26) after 60 minutes. Histologic examination of thrombus formation of all organs analyzed shows that thrombi are formed exclusively in tumor vessels and nowhere else, including the kidney and liver, which showed a brief accumulation at 5 minutes. We can only infer that the protein detected in these organs is present due to clearance mechanisms (e.g., Kupffer cells in the liver phagocytose protein aggregates). Infusion of the control protein TFt alone resulted in no thrombus formation (Fig. 5B), further confirming the requirement for the HBDt domain for proper docking on the local endothelial anionic plasmalemma and initiation of the coagulation cascade, and further indication that the thrombus formation is not attributable to the trace endotoxin content. We show by confocal microscopy that the infused HBDt.TFt colocalized with VEGFR-2, Npn-1, and C6S, which are expressed at high density in the neoangiogenic corona of the tumor (Fig. 6).

It has been reported that HBDt binds Npn-1, the VEGFR-2 coreceptor (3, 19, 20), and the binding and proliferation effects were enhanced in the presence of heparin. It was later reported that the b1b2 region of Npn-1 appears to be a heparin-binding domain, which enhance the binding of VEGF165 to Npn-1 (13). The role of VEGFR-2 (21, 22) and Npn-1 (8, 23–27) and their interaction in angiogenesis and neoangiogenesis are documented. The binding of VEGFR-2 to Npn-1 is mediated through the heparin-binding domain of VEGF165 (12), particularly through the exon 7 encoded region we term HBDt. Computational models have been proposed to explain the different receptor binding and phosphorylation events as resulting from a bridge formed by the HBDt domain of VEGF165 compared with the lack of it as in the case of VEGF121 (28). In addition, there are studies identifying aberrant forms of Npn-1 (and Npn-2) in tumor cells (26). However, current studies do not provide a factual basis why neoangiogenesis and the coexpression of the VEGFR-2 and Npn-1 in the tumor environment may differ from that of angiogenesis occurring in nonneoplastic tissues. Here, we propose that the addition of the C6S-expressing entity modifies the VEGFR-2 + Npn-1 receptor bimolecular complex and redefines it as a unique trimolecular complex within the tumor environment. Although the identity of the core protein carrying the C6S remains to be resolved, there are several possibilities. One is that glycosylation of VEGFR-2 (29) or Npn-1 (26) may differ in the tumor microenvironment and be the C6S carrier. A likely possibility is the expression of a CSPG by the local endothelial cells. The possible involvement of a long oligosaccharide on the b1b2 domain of Npn-1 implies a CS modification that provides the C6S (13). In fact, there is a shift in CS expression associated with tumorigenesis and malignancy (30–35). Several CSPGs are known to date, and some can be candidates for those expressed in tumors (e.g., syndecan-1, betaglycan, and decorin; ref. 33), as well as playing a role in forms of development and neoplasia (36). The role of expression of CSPGs in central nervous system development is well documented (37). Npn-1 is a receptor for both VEGF and semaphorins, the latter being a neuronal mediator (10), which acts as a pivotal mediator in neuronal and angiogenic development (38). Therefore, it is our hypothesis that a CSPG, carrying C6S, interacts with Npn-1 in the tumor microenvironment and plays an important role in tumor neoangiogenesis. The presence of C6S specifically in tumor microenvironment may facilitate the association of VEGFR-2 and Npn-1 in a manner more conducive to endothelial cell signaling and proliferation. This C6S carrying PG is the target of the HBDt as shown in the model illustrated in Fig. 7.



View larger version (29K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 7. Model for the mechanisms of HBDt.TFt action as a selective tumor vascular thrombogen. The HBDt domain localizes the construct to the surface of tumor vascular endothelium through the recognition of a trimolecular target composed of the VEGFR-2, Npn-1, and a C6S oligosaccharide, carried by a PG. The localization of the construct facilitates the docking of the extracellular tissue factor (TF) domain, with its corequisite FVIIa, on an appropriate anionic surface, which promotes initiation of the coagulation cascade as shown above leading to the local generation of thrombus.

 
To advance HBDt as a targeting entity with therapeutic potential, we created a fusion protein of HBDt with TFt, which successfully thrombosed tumors in vivo and reduced their size drastically nearly eradicating them (Fig. 3). This evidence for successful targeting suggests the possibility of using HBDt in other molecules delivering radioactive isotopes or cytotoxic agents.

It should be noted that addition of FVIIa reduced the amount of HBDt.TFt needed to achieve more effective tumor selective thrombosis. This might well be expected considering the very limited amount of murine FVIIa available (39), which would be inadequate to form HBDt.TFt/FVIIa functional complexes. We are currently investigating why the molar ratio of the HBDt.TFt/FVIIa is not the expected 1:1 molar ratio.

In conclusion, we describe a potential tumor selective vascular thrombogen, which is able to target a variety of in vivo tumors, such as colon, mammary, prostate, and lung tumors and produce local occlusive thrombosis and local infarctive necrosis. The endothelial target of the HBDt is novel and appears to result from the assembly of three distinct molecules (VEGFR-2, neuropilin1-1, and C6S) uniquely coexpressed by tumor neoangiogenic vessels.


   Acknowledgments
 
Grant support: NIH grant P01 HL016411.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Malcolm Wood, Ph.D., for his help with fluorescent microscopy and Barbara Parker for her help with the article preparation.

Received 8/ 2/05. Revised 9/ 2/05. Accepted 9/21/05.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. El-Sheikh A, Liu C, Huang H, Edgington TS. A novel vascular endothelial growth factor heparin-binding domain substructure binds to glycosaminoglycans in vivo and localizes to tumor microvascular endothelium. Cancer Res 2002;62:7118–23.[Abstract/Free Full Text]
  2. Robinson CJ, Stringer SE. The splice variants of vascular endothelial growth factor (VEGF) and their receptors. J Cell Sci 2001;114:853–65.[Abstract]
  3. Soker S, Takashima S, Miao HQ, Neufeld G, Klagsbrun M. Neuropilin-1 is expressed by endothelial and tumor cells as an isoform-specific receptor for vascular endothelial growth factor. Cell 1998;92:735–45.[CrossRef][Medline]
  4. Soker S, Kaefer M, Johnson M, Klagsbrun M, Atala A, Freeman MR. Vascular endothelial growth factor-mediated autocrine stimulation of prostate tumor cells coincides with progression to a malignant phenotype. Am J Pathol 2001;159:651–9.[Abstract/Free Full Text]
  5. Miao HQ, Lee P, Lin H, Soker S, Klagsbrun M. Neuropilin-1 expression by tumor cells promotes tumor angiogenesis and progression. FASEB J 2000;14:2532–9.[Abstract/Free Full Text]
  6. Halder JB, Zhao X, Soker S, et al. Differential expression of VEGF isoforms and VEGF(164)-specific receptor neuropilin-1 in the mouse uterus suggests a role for VEGF(164) in vascular permeability and angiogenesis during implantation. Genesis 2000;26:213–24.[CrossRef][Medline]
  7. Miao HQ, Klagsbrun M. Neuropilin is a mediator of angiogenesis. Cancer Metastasis Rev 2000;19:29–37.[CrossRef][Medline]
  8. Klagsbrun M, Takashima S, Mamluk R. The role of neuropilin in vascular and tumor biology. Adv Exp Med Biol 2002;515:33–48.[Medline]
  9. Vanveldhuizen PJ, Zulfiqar M, Banerjee S, et al. Differential expression of neuropilin-1 in malignant and benign prostatic stromal tissue. Oncol Rep 2003;10:1067–71.[Medline]
  10. Gu C, Limberg BJ, Whitaker GB, et al. Characterization of neuropilin-1 structural features that confer binding to semaphorin 3A and vascular endothelial growth factor 165. J Biol Chem 2002;277:18069–76.[Abstract/Free Full Text]
  11. Whitaker GB, Limberg BJ, Rosenbaum JS. Vascular endothelial growth factor receptor-2 and neuropilin-1 form a receptor complex that is responsible for the differential signaling potency of VEGF(165) and VEGF(121). J Biol Chem 2001;276:25520–31.[Abstract/Free Full Text]
  12. Soker S, Miao HQ, Nomi M, Takashima S, Klagsbrun M. VEGF165 mediates formation of complexes containing VEGFR-2 and neuropilin-1 that enhance VEGF165-receptor binding. J Cell Biochem 2002;85:357–68.[CrossRef][Medline]
  13. Mamluk R, Gechtman Z, Kutcher ME, Gasiunas N, Gallagher J, Klagsbrun M. Neuropilin-1 binds vascular endothelial growth factor 165, placenta growth factor-2, and heparin via its b1b2 domain. J Biol Chem 2002;277:24818–25.[Abstract/Free Full Text]
  14. Stone MJ, Ruf W, Miles DJ, Edgington TS, Wright PE. Recombinant soluble human tissue factor secreted by Saccharomyces cerevisiae and refolded from Escherichia coli inclusion bodies: glycosylation of mutants, activity and physical characterization. Biochem J 1995;310:605–14.
  15. Donate F, Kelly CR, Ruf W, Edgington TS. Dimerization of tissue factor supports solution-phase autoactivation of factor VII without influencing proteolytic activation of factor X. Biochemistry 2000;39:11467–76.[CrossRef][Medline]
  16. Ruf W, Edgington TS. An anti-tissue factor monoclonal antibody which inhibits TF VIIa complex is a potent anticoagulant in plasma. Thromb Haemost 1991;66:529–33.[Medline]
  17. Lehr HA, Leunig M, Menger MD, Nolte D, Messmer K. Dorsal skinfold chamber technique for intravital microscopy in nude mice. Am J Pathol 1993;143:1055–62.[Abstract]
  18. Belting M, Dorrell MI, Sandgren S, et al. Regulation of angiogenesis by tissue factor cytoplasmic domain signaling. Nat Med 2004;10:502–9.[CrossRef][Medline]
  19. Soker S, Fidder H, Neufeld G, Klagsbrun M. Characterization of novel vascular endothelial growth factor (VEGF) receptors on tumor cells that bind VEGF165 via its exon 7-encoded domain. J Biol Chem 1996;271:5761–7.[Abstract/Free Full Text]
  20. Soker S, Gollamudi-Payne S, Fidder H, Charmahelli H, Klagsbrun M. Inhibition of vascular endothelial growth factor (VEGF)-induced endothelial cell proliferation by a peptide corresponding to the exon 7-encoded domain of VEGF165. J Biol Chem 1997;272:31582–8.[Abstract/Free Full Text]
  21. Millauer B, Wizigmann-Voos S, Schnurch H, et al. High affinity VEGF binding and developmental expression suggest Flk-1 as a major regulator of vasculogenesis and angiogenesis. Cell 1993;72:835–46.[CrossRef][Medline]
  22. Costa C, Soares R, Schmitt F. Angiogenesis: now and then. APMIS 2004;112:402–12.[CrossRef][Medline]
  23. Parikh AA, Fan F, Liu WB, et al. Neuropilin-1 in human colon cancer: expression, regulation, and role in induction of angiogenesis. Am J Pathol 2004;164:2139–51.[Abstract/Free Full Text]
  24. Parikh AA, Liu WB, Fan F, et al. Expression and regulation of the novel vascular endothelial growth factor receptor neuropilin-1 by epidermal growth factor in human pancreatic carcinoma. Cancer 2003;98:720–9.[CrossRef][Medline]
  25. Akagi M, Kawaguchi M, Liu W, et al. Induction of neuropilin-1 and vascular endothelial growth factor by epidermal growth factor in human gastric cancer cells. Br J Cancer 2003;88:796–802.[CrossRef][Medline]
  26. Fukahi K, Fukasawa M, Neufeld G, Itakura J, Korc M. Aberrant expression of neuropilin-1 and -2 in human pancreatic cancer cells. Clin Cancer Res 2004;10:581–90.[Abstract/Free Full Text]
  27. Chen C, Li M, Chai H, Yang H, Fisher WE, Yao Q. Roles of neuropilins in neuronal development, angiogenesis, and cancers. World J Surg 2005;29:271–5.[CrossRef][Medline]
  28. Mac Gabhann F, Popel AS. Differential binding of VEGF isoforms to VEGF receptor 2 in the presence of neuropilin-1: a computational model. Am J Physiol Heart Circ Physiol 2005;288:H2851–60.[Abstract/Free Full Text]
  29. Takahashi T, Shibuya M. The 230 kDa mature form of KDR/Flk-1 (VEGF receptor-2) activates the PLC-{gamma} pathway and partially induces mitotic signals in NIH3T3 fibroblasts. Oncogene 1997;14:2079–89.[CrossRef][Medline]
  30. Goulas A, Papakonstantinou E, Karakiulakis G, Mirtsou-Fidani V, Kalinderis A, Hatzichristou DG. Tissue structure-specific distribution of glycosaminoglycans in the human penis. Int J Biochem Cell Biol 2000;32:975–82.[CrossRef][Medline]
  31. Goulas A, Hatzichristou DG, Karakiulakis G, Mirtsou-Fidani V, Kalinderis A, Papakonstantinou E. Benign hyperplasia of the human prostate is associated with tissue enrichment in chondroitin sulphate of wide size distribution. Prostate 2000;44:104–10.[CrossRef][Medline]
  32. Tsara ME, Papageorgacopoulou N, Karavias DD, Theocharis DA. Distribution and changes of glycosaminoglycans in neoplasias of rectum. Anticancer Res 1995;15:2107–12.[Medline]
  33. Tsara ME, Theocharis AD, Theocharis DA. Compositional and structural alterations of proteoglycans in human rectum carcinoma with special reference to versican and decorin. Anticancer Res 2002;22:2893–8.[Medline]
  34. Theocharis AD, Tsara ME, Papageorgakopoulou N, Vynios DH, Theocharis DA. Characterization of glycosaminoglycans from human normal and scoliotic nasal cartilage with particular reference to dermatan sulfate. Biochim Biophys Acta 2001;1528:81–8.[Medline]
  35. Theocharis AD, Tsara ME, Papageorgacopoulou N, Karavias DD, Theocharis DA. Pancreatic carcinoma is characterized by elevated content of hyaluronan and chondroitin sulfate with altered disaccharide composition. Biochim Biophys Acta 2000;1502:201–6.[Medline]
  36. Schwartz NB, Domowicz M. Chondrodysplasias due to proteoglycan defects. Glycobiology 2002;12:57–68R.
  37. Domowicz M, Mangoura D, Schwartz NB. Cell specific-chondroitin sulfate proteoglycan expression during CNS morphogenesis in the chick embryo. Int J Dev Neurosci 2000;18:629–41.[CrossRef][Medline]
  38. Bates D, Taylor GI, Minichiello J, et al. Neurovascular congruence results from a shared patterning mechanism that utilizes semaphorin3A and neuropilin-1. Dev Biol 2003;255:77–98.[CrossRef][Medline]
  39. Fair DS. Quantitation of factor VII in the plasma of normal and warfarin-treated individuals by radioimmunoassay. Blood 1983;62:784–91.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Proc. Natl. Acad. Sci. USAHome page
D. Simberg, T. Duza, J. H. Park, M. Essler, J. Pilch, L. Zhang, A. M. Derfus, M. Yang, R. M. Hoffman, S. Bhatia, et al.
Biomimetic amplification of nanoparticle homing to tumors
PNAS, January 16, 2007; 104(3): 932 - 936.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by El-Sheikh, A.
Right arrow Articles by Edgington, T. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by El-Sheikh, A.
Right arrow Articles by Edgington, T. S.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online