|Home | About | Journals | Submit | Contact Us | Français|
Tumor endothelial cells express vascular endothelial growth factor receptor 2 (VEGFR-2). VEGF can direct toxins to tumor vessels through VEGFR-2 for antiangiogenic therapy. This study aimed to selectively damage the VEGFR-2-overexpressing vasculature of pancreatic cancer by SLT-VEGF fusion protein comprising VEGF and the A subunit of Shiga-like toxin which inhibits protein synthesis of cells with high VEGFR-2 expression.
Expression of VEGF and VEGF receptors was evaluated in human pancreatic cancer cells (AsPC-1, HPAF-2) and in normal human endothelial cells (HUVEC) by reverse transcription-polymerase chain reaction. Cells were treated with SLT-VEGF (0.1–10 nM), and cell viability, proliferation, and endothelial tube formation were assessed. Orthotopic pancreatic cancer (AsPC-1, HPAF-2) was induced in nude mice. Animals were treated with SLT-VEGF fusion protein alone or in combination with gemcitabine. Treatment began 3 days or 6 weeks after tumor induction. Primary tumor volume and dissemination were determined after 14 weeks. Microvessel density and expression of VEGF and VEGF receptors were analyzed by immunohistochemistry.
SLT-VEGF did not influence proliferation of pancreatic cancer cells; HUVECs (low-level VEGFR-2) reduced their proliferation rate and tube formation but not their viability. SLT-VEGF fusion protein reduced tumor growth and dissemination, increasing 14-week survival (AsPC-1, up to 75%; HPAF-2, up to 83%). Results of gemcitabine were comparable with SLT-VEGF monotherapy. Combination partly increased the therapeutic effects in comparison to the respective monotherapies. Microvessel density was reduced in all groups. Intratumoral VEGFR-2 expression was found in endothelial but not in tumor cells.
SLT-VEGF is toxic for tumor vasculature rather than for normal endothelial or pancreatic cancer cells. SLT-VEGF treatment in combination with gemcitabine may provide a novel approach for pancreatic cancer.
Adenocarcinoma of the exocrine pancreas is the fifth leading cause of cancer-related death in Western countries. The estimated overall 5-year survival rate of less than 5% is due to the tumor's propensity toward aggressive growth, early metastasis, and resistance to cytotoxic agents and radiation. More than 80%of patients are diagnosed with pancreatic cancer at a locally advanced or metastatic stage, which excludes a curative surgical resection . New therapeutic approaches based on the biologic characteristics of this disease may improve response rates and survival. One promising approach is the inhibition of angiogenesis. Like all solid neoplasms, pancreatic cancer depends on the process of angiogenesis, the formation of tumor blood vessels, for both local and metastatic growth beyond the size of a few cubic millimeters . Inhibition of angiogenesis is an attractive target for tumor therapy because it theoretically offers the hope of long-term control of neoplasm progression . Among the identified proangiogenic regulators, vascular endothelial growth factor (VEGF; a.k.a. VEGF-A) and its two tyrosine kinase receptors, the fms-like tyrosine kinase receptor (Flt1, VEGFR-1) and the kinase insert domain-containing receptor (KDR/FLK1, VEGFR-2), have been identified as key mediators of the regulation of pathologic blood vessel growth and maintenance . VEGF induces endothelial cell proliferation and enhances vascular permeability . Previous studies have shown that VEGF is overexpressed in human pancreatic cancer [6–8]. Moreover, a high expression of VEGF was associated with liver metastasis and a poor prognosis for patients with ductal pancreatic adenocarcinoma [9,10]. In most epithelial tumors, including pancreatic cancer, VEGFR-1 and VEGFR-2 are expressed almost exclusively on endothelial cells, and there is evidence that endothelial cells at sites of angiogenesis express significantly higher numbers of VEGFR-2—the key proangiogenic receptor—than quiescent endothelial cells . Blocking the effects of VEGF on endothelial cells by receptor antagonists [12–14] or neutralizing anti-VEGF antibodies [15,16] inhibits the expansion of a variety of neoplasms. However, there is recent evidence that the effects of these anti-VEGF strategies seem to be transient, in particular when the treatment is interrupted .
An alternative approach to destroy tumor endothelium would be a targeted delivery of potent toxins to tumor endothelial cells. We conducted in vitro and in vivo experiments to evaluate the effects of SLT-VEGF fusion protein comprising VEGF121 and catalytically active A subunit of Shiga-like toxin 1 (SLT-1) produced by Escherichia coli O157:H7 [18,19]. SLT-1 is composed of a single copy of a 32-kDa A subunit associated with a ring-shaped pentamer of 7-kDa B subunits that bind to the cellular receptor globotriaosylceramide known as Gb3/CD77 and enters cells through CD77-mediated endocytosis. SLT-1 was chosen as a potential “natural killer” of endothelial cells because it is known that damage to endothelial cells caused by SLT-1 plays a causative role in the pathogenesis of hemorrhagic colitis and hemolytic uremic syndrome induced by E. coli O1 57:H7 [20,21]. SLT-VEGF is internalized through VEGFR-2 mediated endocytosis, and its cytotoxicity correlates with VEGFR-2 expression . In animal tumor models, SLT-VEGF selectively depletes VEGFR-2-overexpressing CD31+ endothelial cells in tumor vasculature as judged by immunohistochemical analysis  and whole body imaging of VEGF receptors . In contrast, CD31+ tumor endothelial cells that express relatively low levels of VEGFR-2 are marginally affected by SLT-VEGF .
The aim of this study was to evaluate a novel antiangiogenic treatment strategy for pancreatic cancer by selectively damaging endothelial cells at sites of angiogenesis with SLT-VEGF in a clinically relevant orthotopic nude mouse model of pancreatic cancer. In addition, we investigated whether a combination of SLT-VEGF with the cytotoxic agent gemcitabine increases the therapeutic potential in this animal model. As a control, we assessed the effects of SLT-VEGF on proliferation and viability of the human pancreatic cancer cell lines AsPC-1 and HPAF-2 and the human endothelial cell line HUVEC in vitro.
The human pancreatic adenocarcinoma cell lines AsPC-1 (ATCC, CRL-1682) and HPAF-2 (ATCC, CRL-1997) of ductal origin and the noncancerous endothelial cell line HUVEC (C-12200) were obtained from the American Type Culture Collection (ATCC, Manassas, VA) and from PromoCell GmbH (Heidelberg, Germany). The cells were cultured in RPMI-1640 medium (AsPC-1; Invitrogen, Karlsruhe, Germany), minimum essential medium (HPAF-2; PAA, Cölbe, Germany), or endothelial cell growth medium (HUVEC; PromoCell), supplemented with 10% heat-inactivated fetal bovine serum (FBS-Gold; PAA), penicillin G (100 U/ml), streptomycin (100 µg/ml), and amphothericin B (0.25 mg/ml). Cells were incubated at 37°C in humidified air with 5% CO2. The medium was replaced twice a week, and cells were maintained by serial passaging after treating with 0.1% trypsin.
Total cellular RNA was extracted from cell cultures using the Nucleo-Spin RNA II Kit (Macherey & Nagel, Düren, Germany) according to the manufacturer's instructions and resuspended in 50 µl of water treated with 0.1% DMPC (Sigma-Aldrich, Seelze, Germany). RNA concentration was determined using a BioPhotometer (Eppendorf Scientific, Hamburg, Germany). Reverse transcription of total RNA (2 µg) primed with an oligo (dT) oligonucleotide and supplemented with dNTPs (Sigma-Aldrich) was done with M-MLV reverse transcriptase (Promega, Mannheim, Germany) according to the instructions of the manufacturer. First-strand complementary DNA was amplified with transcript-specific oligonucleotides using Ready-Mix Taq PCR Reaction Mix (Sigma-Aldrich).
The primers (TIB MOLBIOL, Berlin, Germany) for the respective genes were designed as follows: VEGF, sense 5′-CGA AGTGGTGAA GTT CAT G-3′ and antisense 5′-TTC TGT ATC AGT CTT TCC TGG TGA G-3′; Flt1/VEGFR-1, sense 5′-GAA GGC ATG AGG ATG AGA GC-3′ and antisense 5′-CAG GCT CAT GAA CTT GAA AGC-3′; KDR/FLK1/VEGFR-2, sense 5′-CAT GTA CGG TCTATG CCA TTC-3′ and antisense 5′-CGT TGG CGC ACT CTT CCT-3′; and β-actin, sense 5′-TTC CTG GGC ATG GAG TCC TGT GG-3′ and antisense 5′-CGC CTA GAA GCA TTT GCG GTG G-3′.
Polymerase chain reaction (PCR) products and a 100-bp DNA molecular weight marker were analyzed by electrophoresis on a 1% agarose gel. The gel was then visualized and photographed under ultraviolet light.
SLT-VEGF fusion protein was constructed, expressed in E. coli strain Origami(DE3)pLysS (Novagen, Madison, WI), purified, and extensively characterized in vitro and in vivo as described elsewhere [18,19,22]. SLT-VEGF retained physical integrity and functional activity after 2 hours of incubation inmurine serumat 37°C (Supplementary Material and Figure W1). However, SLT-VEGF was inactivated by conjugation of a single fluorescent dye (Supplementary Material and Figure W2). For in vitro assays and intraperitoneal injection, SLT-VEGF was dissolved in 0.9% NaCl (pH 7.5). Further dilutions for in vitro studies were made with medium and filtered before use.
To examine the effect of SLT-VEGF on in vitro cell proliferation, 2 x 105 cells from the human pancreatic cancer cell lines AsPC-1, HPAF-2, and the noncancerous human endothelial cell line HUVEC were seeded in six-well culture plates in 2 ml of the respective cell culture medium. The medium was changed the next day (day 1) and SLT-VEGF (0.1–10 nM) was added. After 72 hours (day 4), the cells were trypsinized and counted in a standard hemocytometer. Cell viability was assessed by a colorimetric dye reduction assay with monotetrazolium (MTT; Boehringer, Mannheim, Germany) according to the manufacturer's instructions. Briefly, cells were seeded in 96-well plates at a density of 5 x 103 cells per well in 0.2 ml of the respective medium. Medium was changed the next day (day 1), and SLT-VEGF (0.1–10 nM) was added to triplicate wells. After 72 hours (day 4), 10 µl of MTT (5 mg/ml) solution, and after additional 4 hours of incubation at 37°C 10%, SDS was added to the cells, 100 µl per well. The plates were allowed to stand overnight (37°C at 5% CO2). The change in absorbance measured at 550 nm with a plate reader (Biotek Instruments, Inc, Burlington, VT) has been shown to strongly correlate with the number of viable cells. All experiments were generated in triplicates and repeated three times.
The ability of SLT-VEGF fusion protein to inhibit angiogenesis in vitro was evaluated in a capillary tube formation assay using HUVECs cultured on a synthetic basement membrane matrix. Under these conditions, HUVECs are capable of morphologic differentiation into an extensive network of capillary-like structures composed of highly organized three-dimensional cords . The 96-well cell culture plates were coated with a 100-µl layer of the synthetic basement membrane substrate Matrigel (Becton Dickinson, Heidelberg, Germany) at 10 mg/ml concentration and were incubated at 37°C for 30 minutes to promote gelling. HUVECs were seeded in coated plates, 1.8 x 104 cells per well, SLT-VEGF fusion protein was diluted in cell medium and added to HUVECs in triplicate wells to final concentrations of 1 or 10 nM. Stimulation with VEGF (250 pg/ml) served as positive control. After an 18-hour incubation at 37°C and 5% CO2 humidified atmosphere, the complete capillary tube network within a designated area of a low magnification (x10) field was counted under light microscopy, and data were expressed as percentage of complete capillary tube formation relative to untreated HUVEC control cultures incubated under the same conditions. The assays were done in triplicates in three independent experiments.
We used the transplantation technique previously described for an orthotopic nude mouse and rat model of pancreatic cancer [24,25]. Four-week-old male nude mice were obtained from Charles River Laboratories (Charles River, Sulzfeld, Germany). Donor mice were anesthetized with isoflurane (Forene; Abbott, Wiesbaden, Germany) inhalation. Ten million cells of either the AsPC-1 or HPAF-2 cell line were injected subcutaneously into the animals' flanks. The animals were killed by a lethal dose of isoflurane inhalation and opening of the thorax after 4 weeks, when the subcutaneous tumors had reached a size of 1 cm in largest diameter. The donor tumors were harvested and minced by a scalpel (no. 11) into small (1 mm3) fragments. Tumor recipient nude mice were anesthetized with isoflurane, followed by intraperitoneal injection of xylazine hydrochloride (Rompun, 12 mg/kg body weight [BW]; Bayer, Leverkusen, Germany) and Esketaminhydrochloride (Ketanest S, 40 mg/kg BW; Parke-Davis/Pfizer, Karlsruhe, Germany). The animals' abdomens were opened by a midline incision, and the pancreatic tail with the spleen was gently exteriorized. Two small tissue pockets were prepared in the pancreatic parenchyma as an implantation bed with a microscissor (RS-5610 VANNAS; Roboz, Rockville, MD). One donor tumor fragment was placed into each pancreatic tissue pocket in such a way that the neoplastic tissue was completely surrounded by pancreatic parenchyma. The pancreas was relocated into the abdominal cavity, which was then closed in two layers with 4-0 absorbable suture (Vicryl, Ethicon, Germany). For pain relief, a subcutaneous injection of carprofen (Rimadyl, 4 mg/kg BW; Pfizer) was given after surgery.
The animals were allocated randomly into four treatment groups (SLT-VEGF, prophylaxis and therapy; SLT-VEGF and gemcitabine, prophylaxis and therapy) and one control group. Administration of SLT-VEGF (200 µg/kg, intraperitoneally every second day), of gemcitabine (125 mg/kg, intraperitoneally once a week), as well as the combination of both substances began either 3 days (prophylaxis) or 6 weeks (therapy) after tumor induction. Control animals received equivalent amounts of saline. The therapy was continued up to 14 weeks after tumor induction.
All animals underwent autopsy after 14 weeks after the orthotopic tumor implantation by a lethal dose of isoflurane. The perpendicular diameters of the primary orthotopic tumor were measured with calipers, and the volume was calculated using the following formula: volume = length x width x depth / 2. A dissemination score was used to assess local tumor infiltration as well as distant metastasis [24,25]. Local infiltration was determined at the following sites: spleen, stomach, liver (hilus), kidney, retroperitoneum, diaphragm, mesentery, bowel loops, and abdominal wall. Isolated tumor nodules with no anatomic connection to the primary tumor were considered distant metastases. The sites of evaluation included liver, kidney, spleen, lung, diaphragm, mesentery, retroperitoneum, mediastinum, and the suture line. Tumor dissemination was quantified as follows. Each manifestation of tumor infiltration or metastasis was credited with one point. Additional points were awarded for massive local infiltration (e.g., including more than half of the circumference of the spleen), multiple metastatic nodules (>1 in parenchymal organs; >10 in diaphragm, mesentery, and retroperitoneum), and metastatic nodules more than 50 mm3. Clinical consequences of tumor growth were incorporated into the following scoring system: formation of ascites (2 points if volume > 5ml), development of jaundice, ileus, and cachexia. The primary tumor and all sites of potential infiltration or metastasis were harvested, fixed in 4% formaldehyde, and embedded in paraffin. Then 3-µm-thick tissue sections were obtained and stained with hematoxylin and eosin for microscopic examination. The sections were reviewed to confirm the findings of the macroscopic dissemination score.
Immunohistochemical staining was performed on paraffin-embedded tissue of the collected primary tumor tissue. Three-micrometer-thick sections were cut, using a rotation microtome (RM2125RT; Leica, Wetzlar, Germany). The sections were deparaffinized in xylene and rehydrated in graded alcohols and distilled water. After antigen retrieval with 0.01% EDTA pH 8.0, endogenous peroxidase activity was blocked with 1% hydrogen peroxide in distilled water for 25 minutes followed by washing with distilled water and finally PBS + 0.1%Tween for 5 minutes. To bind nonspecific antigens, the sections were incubated with 1x Power Block (BioGenex, San Ramon, CA) for 5 minutes. The primary antibodies were purified antirabbit CD-31 (platelet endothelial cell adhesion molecule), VEGF, VEGFR-1, and VEGFR-2 purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies diluted 1:100 to 1:150 in antibody diluent (DCS LabLine, Hamburg, Germany) were applied to tissue sections and incubated for 30 minutes at 37°C. As a negative control, sections were incubated with antibody diluent instead of the primary antibody. This was followed by incubation with biotinylated antirabbit immunoglobulin G (1:200; Santa Cruz Biotechnology) for 30 minutes at 37°C. After washing with PBS + Tween-20 and incubating with peroxidase-conjugated avidinbiotin complexes (KPL, Gaithersburg, MD), the immune complexes were visualized by 3,3′-diaminobenzidine (Sigma-Aldrich). The sections were then counterstained with Mayer hematoxylin, upgraded alcohols, mounted, and analyzed by standard light microscopy.
Anti-CD31 was used as an endothelial marker to highlight intratumoral microvessels. Microvessel density (MVD) was quantified as described by Weidner . Areas of highest neovascularization were found by scanning the sections at a magnification of x100, and individual microvessel counts were made on 10 fields at x200 magnification (≈0.74 mm2 per field).
All results are presented as mean ± SEM. Continuous normally distributed variables were analyzed by the Student's t test. Discontinuous variables (dissemination score and MVD) were analyzed by the Mann-Whitney rank sum test. Differences in survival were tested by the χ2 test. P < .05 was considered as statistically significant.
Expression of VEGF, Flt1/VEGFR-1 and KDR/FLK1/VEGFR-2 messenger RNA (mRNA) in the human pancreatic cancer cell lines AsPC-1 and HPAF-2 as well as in the human endothelial cell line HUVEC were determined with reverse transcription (RT)-PCR (Figure 1). All three cell lines exhibited both VEGF (400 bp) and Flt1/VEGFR-1 (300 bp) transcripts. Only HUVECs were positive for KDR/FLK1/VEGFR-2 (200 bp).
The effect of SLT-VEGF on the proliferation and viability of the human pancreatic cell lines AsPC-1, HPAF-2, and the noncancerous human endothelial cell line HUVEC was studied during a period of 72 hours. HUVECs express relatively low numbers of VEGFR-2/cell (~10–25,000) and therefore were not expected to be particularly sensitive to toxin-VEGF fusion proteins, as its cytotoxicity of such proteins strongly depends on the VEGFR-2 density in targeted cells [18,27]. The fusion protein was applied to the cells at three different final concentrations: 0.1, 1, and 10 nM. Figure 2 shows proliferation and viability changes during the treatment. Cell count and an MTT-assay indicate that SLT-VEGF did not influence proliferation of pancreatic cancer cells significantly, even at the highest concentrations (10 nM). Proliferation of the human endothelial cell line HUVEC decreased significantly by 37% in cell count (P < .001); however, viability of HUVEC detected by an MTT-based assay was only marginally affected (approximately 10%) and only at the highest SLT-VEGF concentration.
The biologic activity of SLT-VEGF was evaluated in an in vitro bioassay of angiogenesis. The HUVEC capillary tube formation assay on a Matrigel synthetic basement membrane is a widely used in vitro system to model effectively the distinct temporal and spatial events underlying angiogenesis in vivo . SLT-VEGF displayed statistically significant dose-related inhibition of complete HUVEC capillary tube formation. VEGF-induced capillary tube formation was inhibited by 24.3% at 1 nM and by 50.8% at 10 nM SLT-VEGF. Exogenous stimulation with VEGF significantly increased HUVEC tube formation (Figure 3).
All control animals developed extensive tumor growth (AsPC-1: 1550 ± 131 mm3; HPAF-2: 3920 ± 184 mm3), whereas there was significantly reduced tumor growth in the SLT-VEGF and gemcitabine monotherapy groups and in SLT-VEGF plus gemcitabine combination group. For tumors derived from poorly differentiated AsPC-1 cells, treatment with SLT-VEGF resulted in a reduced tumor volume of 487 ± 56 mm3 (-69%) in the prophylaxis group and 492 ± 72 mm3 (-68% reduction) in the therapy group. Treatment with gemcitabine was somewhat less effective and reduced the volume of AsPC-1 tumors to 678 ± 115 mm3 (-56% reduction) in the prophylaxis group and to 928 ± 491 mm3 (-40% reduction) in the therapy group. However, the combination of SLT-VEGF with gemcitabine was significantly most effective in AsPC-1 tumors than either monotherapy and inhibited primary tumor growth to 177 ± 31 mm3 (-89% reduction) in the prophylaxis group and to 360 ± 76 mm3 (-77% reduction) in the therapy group (Figure 4A).
Tumors derived from the moderately differentiated HPAF-2 cells were reduced by SLT-VEGF treatment to 1069 ± 267 mm3 (-73% reduction) in the prophylaxis group and to 1587 ± 167 mm3 (-60% reduction) in the therapy group. Treatment with gemcitabine was significantly more effective and diminished tumor growth of the HPAF-2 tumors to 184 ± 133 mm3 (-95% reduction) in the prophylaxis group and to 361 ± 284 mm3 (-91% reduction) in the therapy group. Interestingly, combination of both substances did not prove to be more effective than gemcitabine alone and reduced tumor growth of HPAF-2 tumors to 178 ± 204 mm3 (-95% reduction) in the prophylaxis group and to 711 ± 270 mm3 (-82% reduction) in the treatment group (Figure 4A).
Local infiltration and distant metastasis were summarized by a dissemination score. Control animals with tumors derived from the poorly differentiated AsPC-1 cell line reached a high score (20.4 ± 2.5 points). Treatment with SLT-VEGF tended to result in a reduction of tumor spread to 13.2 ± 2.4 points in the prophylaxis group and was not effective (18.6 ± 2.0 points) in the therapy group. Animals treated with gemcitabine showed no difference in tumor spread in comparison to control animals (22.6 ± 4.2 points, prophylaxis group; 23.7 ± 9.2 points, therapy group). Combined treatment with SLT-VEGF and gemcitabine resulted in a significant reduction of tumor dissemination in animals with AsPC-1 tumors in the prophylaxis group (8.0 ± 1.1 points) and in the therapy group (9.8 ± 2.4 points; Figure 4B).
Tumors derived from the HPAF-2 cell line also displayed a high dissemination score (17.0 ± 1.2 points). Administration of SLT-VEGF reduced tumor spread to 10.4 ± 2.2 points in the prophylaxis group but not in the therapy group (15.9 ± 2.7 points). Treatment with gemcitabine inhibited tumor dissemination significantly to 1.2 ± 1.0 points in the prophylaxis group and to 3.4 ± 1.5 points in the therapy group. Tumor spread was almost completely suppressed after combined treatment with SLT-VEGF and gemcitabine in the prophylaxis group (0.58 ± 1.2 points) and significantly reduced in the therapy group (4.4 ± 2.3 points; Figure 4B).
The AsPC-1 tumors showed a relative aggressive behavior in vivo, which was reflected in a low 14-week survival of the controls (17%). Monotherapy with SLT-VEGF revealed a trend to increase survival rates in the therapy group (25%) and increased 14-week survival in the prophylaxis group to 75% (P < .05). Treatment with gemcitabine resulted in an increase of survival rates to 58% in both therapy and prophylaxis groups. Combination of SLT-VEGF with gemcitabine revealed a tendency toward increased 14-week survival in the therapy (58%) and the prophylaxis groups (83%; Figure 4C).
Tumors derived from the HPAF-2 tumors grew less aggressively than AsPC-1 tumors and resulted in higher 14-week survival (33% vs 17%). Monotherapy with SLT-VEGF increased survival to 75% in the therapy group and 83% in prophylaxis group. Gemcitabine alone led to an increased survival of 83% in the therapy group and of 92% in the prophylaxis group. The combined treatment of SLT-VEGF and gemcitabine resulted in a 100% survival in both (therapy and prophylaxis) groups (Figure 4C).
As a parameter of angiogenic activity, MVD was determined by immunohistochemistry using anti-CD31 as an endothelial marker. MVD was enhanced 4.6- and 5.3-fold in the untreated AsPC-1 and HPAF-2 primary tumors, respectively, compared with normal exocrine pancreas (AsPC-1, 70.0 ± 4.3/0.74 mm2; HPAF-2, 80.5 ± 4.3/0.74 mm2; native pancreas, 15.6 ± 1.5/0.74 mm2; P < .001).
Treatment with SLT-VEGF significantly reduced MVD of AsPC-1 primary tumors in the prophylaxis group (40.8 ± 4.3/0.74 mm2) and in the therapy group (48.9 ± 6.1/0.74 mm2). Administration of gemcitabine resulted in a reduced MVD in animals of the prophylaxis group (56.1 ± 3.3/0.74 mm2) as well as in the therapy group (54.0 ± 18.7/0.74 mm2). Combined treatment with SLT-VEGF and gemcitabine showed no additional benefit in reducing MVD (46.0 ± 6.1/0.74 mm2 prophylaxis group; 52.2 ± 6.3/0.74 mm2 therapy group; Figure 5, A–C).
In animals with HPAF-2 tumors, treatment with SLT-VEGF resulted in a significant reduction of MVD in both groups (45.9 ± 3.5/0.74 mm2, prophylaxis; 49.3 ± 3.5/0.74 mm2, therapy). Monotherapy with gemcitabine reduced MVD in the prophylaxis group (57.0 ± 2.4/0.74 mm2) and the therapy group (46.3 ± 11.7/0.74 mm2). Combination of SLT-VEGF and gemcitabine resulted in no further reduction of MVD in the prophylaxis group (50.8 ± 14.3/0.74 mm2) and in the therapy group (60.3 ± 4.6/0.74 mm2; Figure 5, A, B, and D).
To determine the immunohistochemical localization of VEGF and its receptors Flt1/VEGFR-1 and KDR/FLK1/VEGFR-2, highly specific polyclonal antibodies were used. Moderate to strong VEGF immunoreactivity was found in the cytoplasm of pancreatic cancer cells of AsPC-1 and HPAF-2 tumors (Figure 6, A and B). Both cell lines were also mildly to moderately immunoreactive for VEGFR-1 but not for VEGFR-2 (Figure 6, C–F). VEGFR-2 was only detected in endothelial cells of the tumors (Figure 6, G and H).
Targeting proangiogenic mediators such as VEGF and VEGFR-2 has emerged as a promising anticancer treatment strategy, especially for devastating malignancies such as exocrine pancreatic cancer, which is virtually noncurable by surgical resection and/or chemoradiation [29,30]. Growth of primary tumor and metastatic lesions beyond the size of a few cubic millimeters requires neovascularization that combines angiogenesis and vasculogenesis. Therefore, inhibition of angiogenesis is an attractive and promising target for tumor therapy because it theoretically offers the hope of long-term control of neoplasm progression . Kuehn et al.  demonstrated that MVD and expression of proangiogenic factors such as VEGF were significantly higher in pancreatic cancer in comparison to the normal pancreas. The VEGF receptors are expressed at high levels in the activated endothelium of tumor vasculature in general  and in pancreatic cancer, specifically , whereas normal mature blood vessels show negligible levels of VEGF receptors . Our previous research demonstrated that systemic treatment with VEGF coupled to diphtheriatoxin (DT-VEGF) led to specific targeting and damaging of tumor blood vessels .
In this study, we investigated a more specific approach: the effect of the immunotoxin SLT-VEGF that selectively targets endothelial cells overexpressing VEGFR-2 [17,18,22]. Using two cancer animal models, syngeneic 4T1 murine breast cancer and PC3 human prostate cancer xenografts, we have previously demonstrated that three to five biweekly injections of SLT-VEGF fusion protein inhibited tumor growth approximately two-fold . Detailed immunohistochemical analysis of 4T1 tumors revealed that SLT-VEGF treatment only mildly (1- to 1.4-fold) decreases the overall prevalence of CD31+ tumor endothelial cells. In contrast, the prevalence of VEGFR-2-overexpressing tumor endothelial cells decreased four- to five-fold . These findings were supported by the results of whole body imaging of VEGFR-2 in tumor vasculature with VEGF-based near-infrared fluorescent tracers . In these experiments, SLT-VEGF treatment dramatically inhibited VEGFR-2-mediated accumulation of tracers in tumor endothelial cells, reflecting the depletion of endothelial cells overexpressing VEGFR-2 . Taken together, in vitro and in vivo results indicate that SLT-VEGF may be less toxic for normal endothelial cells of the host organism but targets predominantly proliferating tumor endothelial cells with reportedly upregulated expression of VEGFR-2 . Using two cancer animal models, syngeneic 4T1 murine breast cancer and PC3 human prostate cancer xenografts, we have previously demonstrated that SLT-VEGF fusion protein inhibited tumor growth by selective elimination of tumor endothelial cells expressing high levels of VEGFR-2 . This indicates that SLT-VEGF may be less toxic for normal endothelial cells of the host organism but targets predominantly proliferating tumor endothelial cells with reportedly upregulated expression of VEGFR-2 . Indeed, as we report here human endothelial cell line HUVEC, a model system for normal endothelial cells with physiological level of VEGFR-2 expression showed a somewhat decreased proliferation (Figure 2A) and tube formation (Figure 3) but not viability (Figure 2B). In contrast, DT-VEGF was reported to be cytotoxic to HUVECs .
The antitumor effects of SLT-VEGF were evaluated in an orthotopic nude mouse model of pancreatic cancer, which was established in our laboratory . Thismodel mirrors the development of this clinical disease and allows assessment not only of primary tumor growth but also of metastasis and survival. Importantly, both pancreatic cancer cell lines AsPC-1 and HPAF-2 used in these models were not sensitive to SLT-VEGF (Figure 2), indicating that its therapeutic effects are due predominantly to the sensitivity of endothelial cells. Furthermore, in both models, selected regimens of SLT-VEGF (200 µg/kg, gemcitabine, or both drugs) did not exhibit any clinical evidence of toxicity, such as a change in food intake or activity in treatment groups, compared with control animals.
Poorly differentiated AsPC-1 tumors displayed an aggressive growth pattern, killing most control animals before the end of the 14-week observation period. SLT-VEGF treatment, started either 3 days (prophylaxis group) or 6 weeks (treatment group) after tumor implantation, reduced microvascular density in primary tumors 1.7- and 1.4-fold, respectively. Remarkably, these rather modest changes in microvascular density were associated with reduced primary tumor volume (up to -69%), increased survival (up to +58%; Figure 4, A and C), and decreased tumor dissemination in the prophylaxis group (Figure 4B).
Moderately differentiated HPAF-2 cells grew to large, partly cystic tumors in control mice and displayed a moderate local and systemic pattern of spread. The response to SLT-VEGF treatment was similar to that observed in poorly differentiated AsPC-1 tumors: moderate decrease in microvascular density (Figure 5, A and B), rather dramatic decrease in tumor volume and increase in survival, and inhibition of tumor dissemination in the prophylaxis group (Figure 4B).
It is interesting to compare the effects of SLT-VEGF with those of gemcitabine, the frontline drug for advanced pancreatic cancer. Gemcitabine monotherapy resulted in an increased survival (AsPC-1, up to 58%; HPAF-2, up to 92%; Figure 4C), and a decreased primary tumor volume (Figure 4A). Tumor dissemination of AsPC-1 tumors was almost unaffected in gemcitabine treatment group (Figure 4B). It seems that SLT-VEGF is somewhat more effective in inhibiting tumor growth and metastatic dissemination, as well as overall survival, in the poorly differentiated AsPC-1 model. In contrast, gemcitabine was more effective in inhibiting tumor growth and metastatic dissemination in the moderately differentiated HPAF-2 model. Importantly, in both models, combination of both drugs provided for a tendency toward better survival, lower primary tumor volume, and reduced tumor dissemination, especially in the prophylaxis groups of both AsPC-1 and HPAF-2 tumors (Figure 4, A–C). However, this tendency did not reach statistical significance for all parameters. It may be due to the significant antitumor effects of the respective monotherapies at tested regimens, especially in the HPAF-2 group, which were difficult to improve by the combination. In turn, it suggests that there might be opportunities for decreasing doses of each component without losing therapeutic benefits of the combination. Furthermore, these results suggest that combination of SLT-VEGF and gemcitabine might require careful optimization to reach the most advantageous synergy. For example, because SLT-VEGF is more cytotoxic to cells overexpressing VEGFR-2, treatment with this cytotoxin might be more effective if timed to the stages when tumor vasculature is enriched with such cells, for example, in response to treatment with other drugs.
Interestingly, SLT-VEGF and gemcitabine induced rather similar decreases in microvascular density in both tumor models. SLT-VEGF targets tumor endothelial cells overexpressing VEGFR-2, selectively depleting tumor vasculature of such cells . However, gemcitabine, a fluorinated analog of deoxycytidine, is toxic to proliferating endothelial cells because it inhibits DNA replication either directly or indirectly through inactivation of the RNA reductase. Thus, if actively growing tumor endothelial cells overexpress VEGFR-2, they would be targets for both SLT-VEGF and gemcitabine, and both drugs might lead to similar decreases in microvascular density.
In summary, this study adds further evidence that angiogenesis is a critical component of pancreatic cancer growth and metastasis. Currently, there are multiple clinical trials in progress for combination of antiangiogenic drugs, such as bevacizumab, sunitinib, or sorafenib, with chemotherapeutic agents, such as gemcitabine (www.cilicaltrials.gov). Although there is no definitivemodel for combination of antiangiogenic and chemotherapeutic drugs, there are suggestions that delivery of chemotherapeutics to tumor cells might be more efficient when provided by tumor vasculature responding to antiangiogenic treatment. It remains to be established whether such improved delivery takes place and whether it is associated with “normalization” of tumor vasculature  or with continuous revascularization of tumors in response to drugs, particularly those delivered through metronomic scheduling . However, regardless of mechanism, it seems that pancreatic cancers are sensitive to combinations of chemotherapeutic and antiangiogenic drugs. In this respect, our finding that targeting VEGFR-2-overexpressing tumor endothelial cells with SLT-VEGF fusion protein resulted in a significant biologic improvement in two clinically relevant orthotopic pancreatic tumor models, comparable to that induced by gemcitabine, seems particularly significant. We believe that optimization of the SLT-VEGF/gemcitabine regimen provides a promising treatment strategy for advanced pancreatic cancer.
SLT-VEGF stability in PBS andmurine serumwas assessed after 2 hours of incubation at 37°C. Western blot analysis was used to assess physical integrity, and cytotoxicity assay was used to assess functional activity.