Macrophages are an abundant inflammatory cell type in the tumor microenvironment that can contribute to tumor growth and metastasis. Macrophage recruitment into tumors is mediated by multiple cytokines, including vascular endothelial growth factor (VEGF), which is thought to function primarily through VEGF receptor (VEGFR) 1 expressed on macrophages. Macrophage infiltration is affected by VEGF inhibition. We show that selective inhibition of VEGFR2 reduced macrophage infiltration into orthotopic pancreatic tumors. Our studies show that tumor-associated macrophages express VEGFR2. Furthermore, peritoneal macrophages from tumor-bearing animals express VEGFR2, whereas peritoneal macrophages from non–tumor-bearing animals do not. To our knowledge, this is the first time that tumor-associated macrophages have been shown to express VEGFR2. Additionally, we found that the cytokine pleiotrophin is sufficient to induce VEGFR2 expression on macrophages. Pleiotrophin has previously been shown to induce expression of endothelial cell markers on macrophages and was present in the microenvironment of orthotopic pancreatic tumors. Finally, we show that VEGFR2, when expressed by macrophages, is essential for VEGF-stimulated migration of tumor-associated macrophages. In summary, tumor-associated macrophages express VEGFR2, and selective inhibition of VEGFR2 reduces recruitment of macrophages into orthotopic pancreatic tumors. Our results show an underappreciated mechanism of action that may directly contribute to the antitumor activity of angiogenesis inhibitors that block the VEGFR2 pathway. [Cancer Res 2008;68(11):4340–6]

Angiogenesis is required for solid tumor growth beyond 1 to 2 mm3 (1). Therefore, inhibitors of this process are attractive as molecularly targeted agents for the treatment of cancer (2). Bevacizumab (Avastin; Genentech) is a well-characterized example of an antibody against human vascular endothelial growth factor (VEGF) that prevents the subsequent binding of VEGF to VEGF receptor (VEGFR)1 and VEGFR2 (3). Recently, bevacizumab has shown clinical effectiveness in the treatment of metastatic colorectal cancer (2, 4, 5). In our laboratory, we have characterized 2C3, a monoclonal antibody to human VEGF that, in contrast to bevacizumab, blocks VEGF interaction with VEGFR2 only (68). We have previously shown 2C3 to be effective in controlling the growth of breast and pancreatic cancer in animal models (810). The antitumor effect of 2C3 treatment is due in part to the reduction in microvessel density seen in tumors after such treatment. However, as many nonendothelial cells express VEGFRs, we are interested in the effect of VEGF inhibitors on the immune cell infiltrate of tumors, particularly macrophages.

Inflammatory cells make up a large component of the overall tumor mass, and of these, macrophages represent an abundant cell type (1113). At first glance, this could represent an attempt by the host to combat the tumor. However, it has been increasingly recognized that the majority of tumors are not recognized as foreign and that the inflammatory/immune infiltrate promotes tumor growth and metastasis (11, 13). One of the mechanisms for this is the elaboration of factors such as matrix metalloproteinases and cytokines by infiltrating cells. Macrophages, for example, contribute to tumor angiogenesis by the secretion of VEGF. Pollard and colleagues (12, 14, 15) have outlined six mechanisms through which macrophages can promote tumor growth and metastasis. Additionally, most clinical studies indicate increased macrophage infiltration into tumors confers a negative prognosis (11, 16). In animal studies, depletion of macrophages has led to decreased tumor growth in breast (15) and Ewing's sarcoma models (17). However, the effect of macrophages in pancreatic cancer progression has not been well-characterized.

The cytokine network involved in tumor progression is elaborate and extensive. In this study, we investigated a potential contribution to macrophage recruitment by the cytokine pleiotrophin (PTN). PTN is involved in growth and differentiation during development, but in the adult, PTN expression is generally restricted to the central nervous system (1821). PTN can act to promote cancer progression by stimulating epithelial to mesenchymal transition (22) and tumor angiogenesis (23). PTN expression is increased in many cancers, including breast and pancreatic cancer (18, 24). Our interest in PTN is derived not only from its involvement in cancer progression but also from the finding that PTN can induce a vascular phenotype on macrophages in vitro (21).

Pancreatic cancer is one of the deadliest gastrointestinal malignancies, and despite preclinical success, new chemotherapy is rarely effective in this disease (25). One aspect contributing to the poor clinical success of effective preclinical therapeutics is model inadequacies (16, 26). To better identify the biological processes contributing to pancreatic cancer growth, we have implemented an orthotopic model in which human tumors are grown in the pancreas of nude mice. This model provides a more accurate description of the pathologic processes occurring in the tumor microenvironment, including angiogenesis and stromal cell infiltration. Additionally, the nude mouse model allows for an assessment of the innate immune system and its response to tumor growth.

In the present study, we used 2C3 to investigate the effect of VEGF inhibition on the immune cell infiltrate in orthotopic pancreatic tumors. Specifically, based on a prior study in a breast cancer model where treatment with 2C3 resulted in a reduction in tumor-associated macrophages (10), we hypothesized that 2C3 would decrease macrophage infiltration into pancreatic tumors. Furthermore, we investigated VEGF-mediated macrophage recruitment and VEGFR expression on macrophages. Finally, we examined PTN as a potential mediator of VEGFR2 expression on macrophages. Overall, our results show that inhibition of VEGF can affect both tumor angiogenesis and the nonendothelial cell component of a tumor, and that these effects may be significant in terms of tumor growth inhibition.

Cell Culture

The human pancreatic carcinoma cell lines MIA PaCa-2, Capan-1, and Panc-1; the human breast cancer cell lines MDA-MB-231 and MCF-7; and the murine monocyte cell line RAW 264.7 were obtained from American Type Culture Collection and maintained accordingly.

Animals

Athymic 6 to 8-wk-old female nude mice (nu/nu) were purchased from Charles River. Animals were housed in a pathogen-free facility and all procedures were performed in accordance with a protocol approved by the Institutional Animal Care and Use Committee of University of Southwestern.

Tumor Model and Treatment

Orthotopic model. Orthotopic pancreatic tumors were established by injection of cells directly into the pancreas as previously described (27). Treatment was initiated 7 d after tumor cell injection with twice weekly i.p. injections of 100 μg of 2C3 or control (C44, an isotype-matched antibody of irrelevant specificity in one experiment, saline in a second experiment). Treatment was continued in all mice until the experiments were terminated. Mammary fat pad tumors were established with 5 × 106 MCF-7 cells.

Isolation of peritoneal macrophages. Peritoneal macrophages were collected by washing the peritoneal cavity with ice-cold RPMI with pen/strep (10 mL × 2 washes) and collecting the lavage fluid. The lavage fluid from tumor bearing (n = 3) and nontumor bearing (n = 3) was pooled and centrifuged at 1,000 rpm for 5 min. The cells were plated overnight and washed with PBS the following morning. This procedure yielded cells that were >95% positive for F4/80. These cells were used for ICC (Fig. 3A) and transwell assays (Fig. 5C and D).

Histology and Immunohistochemical Studies

Methyl carnoys- or formalin-fixed, paraffin-embedded tissues were sectioned by molecular pathology core laboratory at the University of Texas Southwestern Medical Center. Paraffin-embedded sections were deparaffinized by immersion in xylenes and rehydrated in sequential ethanol. Sections were blocked with 20% Aquablock (East Coast Biologics). Primary antibodies used include MECA-32 (DSHB; University of Iowa), F4/80 (Serotec), CD11b (M1/70; BD Bioscience), T014 (rabbit anti-VEGFR2 purified in our laboratory; ref. 28), 7D10 and 10B11 (rat anti-VEGFR2, purified in our laboratory; ref. 29), anti-CD14 and anti-CD86 (both from eBioscience), anti–Mac-3 (PharMingen), anti-VEGFR1 (Santa Cruz Biotechnology), and anti-PTN (Abcam). The appropriate horseradish peroxidase– or fluorophore- conjugated secondary antibody was used (Jackson Immunoresearch). Negative controls were performed by omitting the primary antibody. Fluorescent slides were coverslipped using Prolong with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen). Sections were examined on a Nikon E600 microscope and images were captured with Photometrics coolsnap HQ camera using Metamorph Software.

Reverse Transcription-PCR

RNA was prepared using TRIzol (Invitrogen) according to the manufacturer's instructions. The expression of VEGFR2 was analyzed by quantitative real-time reverse transcription-PCR (RT-PCR) using an assay on demand (Mm00440111_m1) from (Applied Biosystems). We used TBP (Applied Biosystems assay-on-demand) as an internal reference gene to normalize input cDNA. Quantitative real-time RT-PCR was performed in a reaction volume of 20 μL including 1 μL of cDNA, and each reaction was performed in triplicate. We used the comparative cycle threshold method to compute relative expression values.

Migration Assays

Migration assays were done using 24-well plates with 8-μm transwell inserts (Becton Dickinson Labware). Cells (40,000; RAW or peritoneal macrophages) were loaded onto the top of a gelatin-coated filter. The cells were allowed to migrate overnight. Human VEGF (40 ng; R&D) added to the lower chamber was used to stimulate chemotaxis. 2C3 or an isotype-matched control antibody (C44) was used at 40 μg/mL. RAW cells were incubated in serum-free medium ± 100 ng/mL PTN (Sigma) for 6 h before use in the assay. Peritoneal macrophages were plated in SFM with pen/strep. The inserts were fixed and stained using Diff-Quick method (Dade Behring). Cells were counted in 5 to 10 high power fields per insert.

Transforming Growth Factor β ELISA

Tumor lysates were made from orthotopic tumors by mincing the tumor in lysis buffer. Protein content was assayed using BCA assay (Pierce). Transforming growth factor β ELISA was performed according to manufacturer's instructions (Promega).

Flow Cytometry

Tumor lysates were prepared by mincing the tumor in RPMI (Sigma) and filtering through sequentially smaller filters (BD Falcon). The single-cell suspension was labeled with primary antibody for 30 min at 4°C. Antibodies used were FITC-CD11b and phycoerythrin-VEGFR2 (both from Biolegend). Flow cytometry was done on FACScaliber (BD). Propidium iodide–positive cells (Sigma) were excluded and gates were adjusted on the negative control. These gates were then applied to the normal pancreas and tumor lysates. Data analysis was performed using FloJo software (Tree Star, Inc.).

Statistics

Data were analyzed using GraphPad software (GraphPad Prism version 4.00 for Windows; GraphPad Software). Results are expressed as mean ± SE. Differences are analyzed by t test or ANOVA, and results are considered significant at a P value of <0.05.

2C3 decreases macrophage infiltration into pancreatic tumors in vivo. To address the question of whether 2C3 treatment affects macrophage infiltration into orthotopic pancreatic tumors, MIA PaCa-2 tumors were established in nude mice by directly injecting the tail of the pancreas with 106 cells. Animals were treated with saline or 2C3 twice weekly. Pancreatic tumors were excised at sacrifice, and tumor volumes were measured. We show that 2C3 is effective at inhibiting pancreatic tumor growth (Fig. 1A). This confirms previous results in other models demonstrating that inhibition of the VEGFR2 pathway is an effective treatment strategy (8, 10, 29, 30).

Figure 1.

2C3 inhibits pancreatic tumor growth and reduces tumor-associated macrophage infiltration. One million MIA PaCa-2 cells were injected into the tail of the pancreas in nude mice. Treatment with a control antibody or 2C3 was initiated 1 wk after tumor cell injection and continued for 5 wk. Tumor volumes were measured at sacrifice. A, 2C3 inhibited tumor growth compared with control (468.2 mm3 ± 254 mm3 versus 1,599 mm3 ± 254.7 mm3, respectively; P < 0.01). B, sections of tumors from control and 2C3-treated animals were evaluated by immunofluorescence for macrophage markers, and signal intensity was quantified by MetaMorph software. Tumors from 2C3-treated animals showed a significant reduction (P < 0.01) in F4/80 and CD11b-positive cells (red) compared with control-treated animals. Blue, DAPI-stained nuclei. Total magnification, ×200. Images are representative from at least three tumors per group. Quantification of immunofluorescence is based on 3 to 5 high power field per tumor (P < 0.05). C, tumor lysates from control or 2C3-treated animals (n = 3 per treatment group) were evaluated by ELISA for the level of active TGFβ, showing a decrease in active TGFβ after 2C3 treatment (P = 0.056).

Figure 1.

2C3 inhibits pancreatic tumor growth and reduces tumor-associated macrophage infiltration. One million MIA PaCa-2 cells were injected into the tail of the pancreas in nude mice. Treatment with a control antibody or 2C3 was initiated 1 wk after tumor cell injection and continued for 5 wk. Tumor volumes were measured at sacrifice. A, 2C3 inhibited tumor growth compared with control (468.2 mm3 ± 254 mm3 versus 1,599 mm3 ± 254.7 mm3, respectively; P < 0.01). B, sections of tumors from control and 2C3-treated animals were evaluated by immunofluorescence for macrophage markers, and signal intensity was quantified by MetaMorph software. Tumors from 2C3-treated animals showed a significant reduction (P < 0.01) in F4/80 and CD11b-positive cells (red) compared with control-treated animals. Blue, DAPI-stained nuclei. Total magnification, ×200. Images are representative from at least three tumors per group. Quantification of immunofluorescence is based on 3 to 5 high power field per tumor (P < 0.05). C, tumor lysates from control or 2C3-treated animals (n = 3 per treatment group) were evaluated by ELISA for the level of active TGFβ, showing a decrease in active TGFβ after 2C3 treatment (P = 0.056).

Close modal

Next, to investigate immune infiltration into the tumor microenvironment, we performed immunohistochemistry using markers F4/80 and M1/70 to address whether macrophages were reduced in this setting. Our results showed that macrophages were reduced in tumors from 2C3-treated animals compared with control (Fig. 1B). Additionally, tumor lysates from 2C3-treated animals have less TGFβ, consistent with a reduction of macrophages (Fig. 1C). These results show for the first time that VEGF inhibition reduces macrophage infiltration into pancreatic tumors.

Furthermore, we show that inhibiting tumor cell–derived VEGF with 2C3 reduces tumor angiogenesis. Tumors from control or 2C3-treated animals were analyzed by immunohistochemistry for microvessel density. Treatment with 2C3 reduced the number of blood vessels within the tumor (Supplementary Fig. S1) Additionally, there was a trend toward increased pericyte associated blood vessels in the 2C3-treated group, suggesting that VEGF inhibition with 2C3 preferentially targets nonpericyte associated vessels (Supplementary Fig. S1).

Tumor-associated macrophages express VEGFR2. 2C3 has previously been shown to selectively inhibit human VEGF from interacting with both murine (Flk-1) and human VEGFR2 (6, 7). Therefore, for the decrease in macrophages to occur in this model as a result of treatment with 2C3, we hypothesized that tumor-associated macrophages express VEGFR2. We performed immunohistochemistry to colocalize VEGFR2 on macrophages using four different macrophage markers. Our results showed that macrophages express VEGFR2 in pancreatic tumors (Fig. 2A). We also found VEGFR2+ macrophages in tumor sections from two other human pancreatic cell lines, Capan-1 and Panc1 (data not shown). Furthermore, 2C3 treatment reduced the number of VEGFR2+ macrophages in tumor sections (Fig. 2B). Finally, we performed three-color flow cytometry showing VEGFR2+ CD11b+ cells from orthotopic tumors but not in normal, non–tumor-bearing pancreas (Fig. 2C).

Figure 2.

Tumor-associated macrophages express VEGFR2. Tumor sections from control or 2C3-treated animals were evaluated by immunofluorescence for VEGFR2 and various macrophage markers. A, representative images of tumor sections from control-treated animals show colocalization (yellow) of VEGFR2 and macrophage markers (CD14, CD86, F4/80, and Mac-3). Total magnification, ×200. B, representative images of tumors from control and 2C3-treated animals and quantification of the number of VEGFR2+ macrophages. (Total magnification, ×200; *, P < 0.05). C, flow cytometry performed on single-cell suspensions made from a normal, non–tumor-bearing pancreas or from an orthotopic tumor showed a small population of VEGFR2+CD11b+ cells in the tumor but not in normal pancreas. Data shown are representative of two separate experiments. hpf, high power field.

Figure 2.

Tumor-associated macrophages express VEGFR2. Tumor sections from control or 2C3-treated animals were evaluated by immunofluorescence for VEGFR2 and various macrophage markers. A, representative images of tumor sections from control-treated animals show colocalization (yellow) of VEGFR2 and macrophage markers (CD14, CD86, F4/80, and Mac-3). Total magnification, ×200. B, representative images of tumors from control and 2C3-treated animals and quantification of the number of VEGFR2+ macrophages. (Total magnification, ×200; *, P < 0.05). C, flow cytometry performed on single-cell suspensions made from a normal, non–tumor-bearing pancreas or from an orthotopic tumor showed a small population of VEGFR2+CD11b+ cells in the tumor but not in normal pancreas. Data shown are representative of two separate experiments. hpf, high power field.

Close modal

Tumor-bearing animals express VEGFR2 on macrophages distant from the tumor. In non–tumor-bearing animals, VEGFR1 has been shown to mediate monocyte chemotaxis (31). Thus, the finding of VEGFR2+ macrophages in tumor tissue is novel and raised the question of whether this was a systemic effect or if VEGFR2 expression was isolated to tumor-associated macrophages. To address this question, we harvested peritoneal macrophages from tumor-bearing or non–tumor-bearing animals and performed immunocytochemistry for VEGFR2. These studies showed that peritoneal macrophages from tumor-bearing animals expressed VEGFR2, whereas those from non–tumor-bearing animals did not (Fig. 3A). Peritoneal macrophages from both non–tumor-bearing and tumor-bearing animals also express VEGFR1 as determined by immunocytochemistry (Fig. 3A) and PCR analysis (data not shown). We further examined VEGFR2 expression at the RNA level by qPCR. This analysis revealed that peritoneal macrophages from tumor-bearing animals express VEGFR2 at a 2.5-fold higher level than peritoneal macrophages from non–tumor-bearing animals (Fig. 3B).

Figure 3.

Macrophage expression of VEGFR2 is tumor dependent. Macrophages from tumor-bearing (TB) and non–tumor-bearing (NTB) animals were isolated by peritoneal lavage and evaluated for VEGFR2 expression by immunocytochemistry (A) or qPCR (B). A, peritoneal macrophages were plated onto chamber slides and stained with a control antibody (staining control), anti-VEGFR1 (green), or anti-VEGFR2 (red). Peritoneal macrophages from non–tumor-bearing animals are VEGFR1 positive but not VEGFR2. In contrast, peritoneal macrophages from tumor-bearing animals are VEGFR1- and VEGFR2-positive. Images are representative from three experiments. Total magnification, ×200. B, RNA isolated from peritoneal macrophages at the time of cell harvest from tumor-bearing or non–tumor-bearing animals was used for qPCR. Shown is ratio of VEGFR2 transcript/TBP, a housekeeping gene. Ct, cycle threshold.

Figure 3.

Macrophage expression of VEGFR2 is tumor dependent. Macrophages from tumor-bearing (TB) and non–tumor-bearing (NTB) animals were isolated by peritoneal lavage and evaluated for VEGFR2 expression by immunocytochemistry (A) or qPCR (B). A, peritoneal macrophages were plated onto chamber slides and stained with a control antibody (staining control), anti-VEGFR1 (green), or anti-VEGFR2 (red). Peritoneal macrophages from non–tumor-bearing animals are VEGFR1 positive but not VEGFR2. In contrast, peritoneal macrophages from tumor-bearing animals are VEGFR1- and VEGFR2-positive. Images are representative from three experiments. Total magnification, ×200. B, RNA isolated from peritoneal macrophages at the time of cell harvest from tumor-bearing or non–tumor-bearing animals was used for qPCR. Shown is ratio of VEGFR2 transcript/TBP, a housekeeping gene. Ct, cycle threshold.

Close modal

PTN induces VEGFR2 expression on macrophages. The presence of VEGFR2 on macrophages distant from the tumor suggested to us that a soluble factor produced by or in response to the tumor induced VEGFR2 expression on macrophages. Based on prior work by Sharifi et al. (21) demonstrating that PTN can induce a vascular phenotype on monocytes (i.e., expression of endothelial markers such as VEGFR2), we proposed that PTN could mediate the induction of VEGFR2 on tumor-associated macrophages. We found that PTN is expressed in MIA PaCa-2 tumors by RT-PCR (data not shown) and by immunohistochemistry (Fig. 4A). Additionally, we also identified that PTN protein is present in orthotopic Capan-1 tumors (Fig. 4B). Next, we exposed proliferating macrophages (RAW cells on chamber slides) to recombinant PTN and probed for VEGFR2 expression. These studies show that recombinant PTN protein can induce VEGFR2 on macrophages in vitro (Fig. 4C).

Figure 4.

PTN induces VEGFR2 expression. Tumors were harvested and evaluated for the level of PTN by immunohistochemistry. A, frozen sections of orthotopic MIA PaCa-2 tumors were stained with antibodies specific for PTN (red). Images show the presence of PTN. B, tumor sections from CAPAN-1 tumors illustrate a similar pattern of PTN (total magnification, ×200). The images shown in A and B are representative from histology performed on at least three separate tumors. C, RAW 264.7 (RAW) cells were incubated for 12 h in chamber slides in the presence or absence of PTN (10 or 100 ng/mL). Immunocytochemistry showed an increase in VEGFR2 expression after PTN stimulation. Representative images from three independent experiments are shown (total magnification, ×200). D, immunocytochemistry shows expression of VEGFR2 on RAW cells after incubation with conditioned medium from MDA-MB-231 cells but not after exposure to conditioned medium from MCF-7 cells (total magnification, ×200).

Figure 4.

PTN induces VEGFR2 expression. Tumors were harvested and evaluated for the level of PTN by immunohistochemistry. A, frozen sections of orthotopic MIA PaCa-2 tumors were stained with antibodies specific for PTN (red). Images show the presence of PTN. B, tumor sections from CAPAN-1 tumors illustrate a similar pattern of PTN (total magnification, ×200). The images shown in A and B are representative from histology performed on at least three separate tumors. C, RAW 264.7 (RAW) cells were incubated for 12 h in chamber slides in the presence or absence of PTN (10 or 100 ng/mL). Immunocytochemistry showed an increase in VEGFR2 expression after PTN stimulation. Representative images from three independent experiments are shown (total magnification, ×200). D, immunocytochemistry shows expression of VEGFR2 on RAW cells after incubation with conditioned medium from MDA-MB-231 cells but not after exposure to conditioned medium from MCF-7 cells (total magnification, ×200).

Close modal

To further characterize this phenomenon, we analyzed the effect of conditioned medium from MDA-MB-231 and MCF-7 cell lines on VEGFR2 expression by macrophages in vitro. We chose these cell lines based on microarrary data demonstrating that MDA-MB-231 cells express PTN at higher levels than MCF-7 cells,5

5

Drs John Minna and Luc Girard, personal communication.

and other published results indicating the same (32, 33). Exposure of RAW cells to conditioned medium from MDA-MB-231 cells for 12 hours induced VEGFR2 expression, whereas MCF-7 conditioned medium did not induce VEGFR2 expression (Fig. 4D). Furthermore, we analyzed frozen sections of MCF-7 tumors and showed that despite the lack of tumor cell production of PTN, there was PTN in the tumor microenvironment, as well as VEGFR2+ macrophages (Supplementary Fig. S2). These data suggests that the source of PTN in the tumor microenvironment is not entirely tumor-cell derived.

VEGFR2 dominates VEGF-mediated macrophage chemotaxis. To assess the functional significance of VEGFR2 expression by macrophages, we analyzed VEGF-induced migration of RAW cells in vitro and peritoneal macrophages ex vivo. RAW cells (VEGFR1+ VEGFR2−) show a slight increase in migration toward VEGF (Fig. 5A). This stimulation is not affected by the addition of 2C3 to the bottom of the transwell chamber. However, when VEGFR2 is induced by stimulation with PTN (see above; Fig. 4C), RAW cells show a significant migration toward VEGF (Fig. 5B). Importantly, this effect is blocked by the addition of 2C3, demonstrating VEGFR2 dependence. Consistent with these finding, peritoneal macrophages from non–tumor-bearing animals (VEGFR2−) show only minimal migration toward VEGF, which is unaffected by the presence of 2C3 (Fig. 5C) or control antibody (data not shown). Peritoneal macrophages harvested from tumor-bearing animals (VEGFR2+) migrate strongly toward VEGF, and this is abrogated by the addition of 2C3 (Fig. 5D).

Figure 5.

VEGFR2 is the dominant receptor mediating VEGF-induced macrophage chemotaxis. The effect of 2C3 on VEGF-induced macrophage migration was assessed by the use of modified Boyden chamber assays. A, unstimulated RAW cells migrate only minimally to VEGF (40 ng/mL) compared with serum free-medium (SFM), and this is unaffected by the presence of 2C3. B, RAW cells stimulated with PTN (100 ng/mL) migrate significantly more toward VEGF, and this process is blocked by 2C3. C, peritoneal macrophages from non–tumor-bearing animals do not migrate toward VEGF. D, macrophages harvested from tumor-bearing animals do migrate significantly toward VEGF, and this is blocked in the presence of 2C3. The mean number of cells per high-power field in each condition is shown. Four to five high power fields were counted per insert. Assays were done in duplicate or triplicate. Data shown is representative of at least three independent experiments and was analyzed with Kruskal-Wallis test.

Figure 5.

VEGFR2 is the dominant receptor mediating VEGF-induced macrophage chemotaxis. The effect of 2C3 on VEGF-induced macrophage migration was assessed by the use of modified Boyden chamber assays. A, unstimulated RAW cells migrate only minimally to VEGF (40 ng/mL) compared with serum free-medium (SFM), and this is unaffected by the presence of 2C3. B, RAW cells stimulated with PTN (100 ng/mL) migrate significantly more toward VEGF, and this process is blocked by 2C3. C, peritoneal macrophages from non–tumor-bearing animals do not migrate toward VEGF. D, macrophages harvested from tumor-bearing animals do migrate significantly toward VEGF, and this is blocked in the presence of 2C3. The mean number of cells per high-power field in each condition is shown. Four to five high power fields were counted per insert. Assays were done in duplicate or triplicate. Data shown is representative of at least three independent experiments and was analyzed with Kruskal-Wallis test.

Close modal

The major findings in this study are that tumor-associated macrophages express VEGFR2, and this expression is critical for VEGF-induced recruitment of macrophages into tumors. Additionally, we show that systemic macrophages from mice with orthotopic pancreatic tumors express VEGFR2. Finally, we show that the cytokine PTN is sufficient to induce VEGFR2 expression on macrophages.

An increased level of macrophage infiltration into tumors correlates with increased angiogenesis and poor prognosis. It is proposed that macrophages recruited into tumors facilitate a microenvironment that promotes tumor development (14). VEGF is an abundant cytokine in the tumor microenvironment and is known to stimulate macrophage chemotaxis. However, to our knowledge, few studies have looked directly at the effect of inhibitors of VEGF on macrophage infiltration into tumors. Whitehurst et al. (10) investigated the effect of 2C3 treatment in an orthotopic breast cancer model. They show that tumors contain less CD11b+ macrophages after treatment with 2C3, a reduction similar to that found in the present study. In a mouse model of thyroid cancer, Salnikov et al. (34) show that treatment with bevacizumab reduced the number of macrophages in the tumor. Thus, including our study, at least three independent studies have shown VEGF blockade reduces tumor-associated macrophage density. In this set of experiments, we show that the mechanism for reduced macrophage infiltration is dependent on blocking VEGFR2 activity. Although our study focused primarily on pancreatic cancer, our data on MCF-7 tumors is consistent with the results of Whitehurst et al. (10) and suggests that VEGFR2 expression on tumor-associated macrophages is present in epithelial tumors of multiple origins.

The origin of VEGFR2+ macrophages is undetermined and currently being investigated. One possibility is that VEGFR2+F4/80+ cells are a manifestation of fusion of macrophages and endothelial cells. However, we interpret the finding of VEGFR2+ peritoneal macrophages to indicate that induction of VEGFR2 on macrophages is a systemic phenomenon, suggesting a soluble factor produced by or in response to the tumor is mediating VEGFR2 expression by macrophages. An additional source for VEGFR2+ macrophages would include macrophages that are already present in the tumor that become activated by PTN. Finally, the hypothesis that tumor-derived PTN recruits bone marrow–derived cells has not been excluded. This is supported by a recent report that PTN can induce nitric oxide–dependent migration of endothelial progenitor cells in vitro (35).

Our study shows the potential involvement of the cytokine PTN in altering the immune infiltrate in pancreatic cancer. This cytokine has multiple functions affecting tumor growth, including promoting angiogenesis, cancer cell proliferation, and epithelial-mesenchymal transformation (21, 32). In breast cancer, PTN has been shown to alter the tumor microenvironment, including the extracellular matrix in a transgenic model (32). Souttou et al. (18) have investigated the serum levels of PTN in mouse models and indicate that serum levels increase as tumor size increases. In patients with gastrointestinal malignancies, there was an elevation of PTN in some but not all patients (18). We hypothesize that PTN in the tumor microenvironment is sufficient for induction of VEGFR2 expression on macrophages. In our in vitro studies, we have used a concentration that is higher than physiologic but may reflect the level of PTN in the tumor microenvironment.

In this study, we show the effectiveness of 2C3 therapy in a preclinical model of pancreatic cancer. However, we also provide a mechanism by which VEGF participates in the recruitment of macrophages. These data suggest a possible marker (e.g., VEGFR2+ macrophages or circulating monocytes) for identifying pancreatic cancer and a possible marker for the efficacy of antiangiogenic treatment. It is worth noting that higher levels of circulating VEGFR2+ monocytes/macrophages have recently been identified in patients with various advanced solid tumors compared with healthy volunteers (36). The inductive factor in human patients has not been elucidated; however, we show that the cytokine PTN is sufficient to induce VEGFR2 expression in macrophages, a function of PTN that is relevant to the progression of pancreatic cancer. As PTN is not normally expressed in adult tissue, this also suggests a potential new target for therapy.

R.A. Brekken: Peregrine Pharmaceuticals employee, commercial research grant from Peregrine Pharmaceuticals, and ownership interest in Peregrine Pharmaceuticals.

Peregrine Pharmaceuticals has licensed the antibody 2C3 from the University of Texas. R.A. Brekken is a consultant for, has equity interest in, and receives research support from Peregrine Pharmaceuticals, Inc.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

Grant support: NIH R21 CA10669 (J.B. Fleming); the National Pancreas Foundation and the Effie Marie Cain Scholarship in Angiogenesis Research (R.A. Brekken.); and a postdoctoral fellowship from the Susan G. Komen Foundation (S.P. Dineen.) and the Leader Koo Fund for Pancreatic Cancer Research (M. D. Anderson Cancer Center). Additional support was given through a sponsored research agreement with Peregrine Pharmaceuticals, Inc. (R.A. Brekken).

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 all members of the Brekken laboratory for support Michael Dodge for help with quantitative PCR, Dr. Robert Bachoo for providing reagents, and Dr. Philip Thorpe for suggestions, rat anti-VEGFR2 antibodies, and u87 cells; Dr. John Minna for helpful advice and access to and use of quantitative PCR; and Dr. Donald McDonald for critical review of the manuscript.

1
Folkman J. Angiogenesis in cancer, vascular, rheumatoid and other disease.
Nat Med
1995
;
1
:
27
–31.
2
Duda DG, Jain RK, Willett CG. Antiangiogenics: the potential role of integrating this novel treatment modality with chemoradiation for solid cancers.
J Clin Oncol
2007
;
25
:
4033
–42.
3
Shojaei F, Ferrara N. Antiangiogenic therapy for cancer: an update.
Cancer J
2007
;
13
:
345
–8.
4
Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer.
N Engl J Med
2004
;
350
:
2335
–42.
5
Jain RK, Duda DG, Clark JW, Loeffler JS. Lessons from phase III clinical trials on anti-VEGF therapy for cancer.
Nat Clin Pract Oncol
2006
;
3
:
24
–40.
6
Brekken RA, Overholser JP, Stastny VA, Waltenberger J, Minna JD, Thorpe PE. Selective inhibition of vascular endothelial growth factor (VEGF) receptor 2 (KDR/Flk-1) activity by a monoclonal anti-VEGF antibody blocks tumor growth in mice.
Cancer Res
2000
;
60
:
5117
–24.
7
Brekken RA, Huang X, King SW, Thorpe PE. Vascular endothelial growth factor as a marker of tumor endothelium.
Cancer Res
1998
;
58
:
1952
–9.
8
Holloway SE, Beck AW, Shivakumar L, Shih J, Fleming JB, Brekken RA. Selective blockade of vascular endothelial growth factor receptor 2 with an antibody against tumor-derived vascular endothelial growth factor controls the growth of human pancreatic adenocarcinoma xenografts.
Ann Surg Oncol
2006
;
13
:
1145
–55.
9
Zhang W, Ran S, Sambade M, Huang X, Thorpe PE. A monoclonal antibody that blocks VEGF binding to VEGFR2 (KDR/Flk-1) inhibits vascular expression of Flk-1 and tumor growth in an orthotopic human breast cancer model.
Angiogenesis
2002
;
5
:
35
–44.
10
Whitehurst B, Flister MJ, Bagaitkar J, et al. Anti-VEGF-A therapy reduces lymphatic vessel density and expression of VEGFR-3 in an orthotopic breast tumor model.
Int J Cancer
2007
;
10
:
2181
–91.
11
de Visser KE, Eichten A, Coussens LM. Paradoxical roles of the immune system during cancer development.
Nat Rev Cancer
2006
;
6
:
24
–37.
12
Pollard JW. Tumour-educated macrophages promote tumour progression and metastasis.
Nat Rev Cancer
2004
;
4
:
71
–8.
13
Coussens LM, Werb Z. Inflammation and cancer.
Nature
2002
;
420
:
860
–7.
14
Condeelis J, Pollard JW. Macrophages: obligate partners for tumor cell migration, invasion, and metastasis.
Cell
2006
;
124
:
263
–6.
15
Lin EY, Pollard JW. Tumor-associated macrophages press the angiogenic switch in breast cancer.
Cancer Res
2007
;
67
:
5064
–6.
16
Talmadge JE, Singh RK, Fidler IJ, Raz A. Murine models to evaluate novel and conventional therapeutic strategies for cancer.
Am J Pathol
2007
;
170
:
793
–804.
17
Zeisberger SM, Odermatt B, Marty C, Zehnder-Fjallman AH, Ballmer-Hofer K, Schwendener RA. Clodronate-liposome-mediated depletion of tumour-associated macrophages: a new and highly effective antiangiogenic therapy approach.
Br J Cancer
2006
;
95
:
272
–81.
18
Souttou B, Juhl H, Hackenbruck J, et al. Relationship between serum concentrations of the growth factor pleiotrophin and pleiotrophin-positive tumors.
J Natl Cancer Inst
1998
;
90
:
1468
–73.
19
Souttou B, Raulais D, Vigny M. Pleiotrophin induces angiogenesis: involvement of the phosphoinositide-3 kinase but not the nitric oxide synthase pathways.
J Cell Physiol
2001
;
187
:
59
–64.
20
Deuel TF, Zhang N, Yeh HJ, Silos-Santiago I, Wang ZY. Pleiotrophin: a cytokine with diverse functions and a novel signaling pathway.
Arch Biochem Biophys
2002
;
397
:
162
–71.
21
Sharifi BG, Zeng Z, Wang L, et al. Pleiotrophin induces transdifferentiation of monocytes into functional endothelial cells.
Arterioscler Thromb Vasc Biol
2006
;
26
:
1273
–80.
22
Perez-Pinera P, Alcantara S, Dimitrov T, Vega JA, Deuel TF. Pleiotrophin disrupts calcium-dependent homophilic cell-cell adhesion and initiates an epithelial-mesenchymal transition.
Proc Natl Acad Sci U S A
2006
;
103
:
17795
–800.
23
Zhang N, Zhong R, Perez-Pinera P, et al. Identification of the angiogenesis signaling domain in pleiotrophin defines a mechanism of the angiogenic switch.
Biochem Biophys Res Commun
2006
;
343
:
653
–8.
24
Weber D, Klomp HJ, Czubayko F, Wellstein A, Juhl H. Pleiotrophin can be rate-limiting for pancreatic cancer cell growth.
Cancer Res
2000
;
60
:
5284
–8.
25
Korc M. Pathways for aberrant angiogenesis in pancreatic cancer.
Mol Cancer
2003
;
2
:
8
.
26
Man S, Munoz R, Kerbel RS. On the development of models in mice of advanced visceral metastatic disease for anti-cancer drug testing.
Cancer Metastasis Rev
2007
;
26
:
737
–47.
27
Beck AW, Luster TA, Miller AF, et al. Combination of a monoclonal anti-phosphatidylserine antibody with gemcitabine strongly inhibits the growth and metastasis of orthotopic pancreatic tumors in mice.
Int J Cancer
2006
;
10
:
2639
–43.
28
Feng D, Nagy JA, Brekken RA, et al. Ultrastructural localization of the vascular permeability factor/vascular endothelial growth factor (VPF/VEGF) receptor-2 (FLK-1, KDR) in normal mouse kidney and in the hyperpermeable vessels induced by VPF/VEGF-expressing tumors and adenoviral vectors.
J Histochem Cytochem
2000
;
48
:
545
–56.
29
Ran S, Huang X, Downes A, Thorpe PE. Evaluation of novel antimouse VEGFR2 antibodies as potential antiangiogenic or vascular targeting agents for tumor therapy.
Neoplasia
2003
;
5
:
297
–307.
30
Korpanty G, Carbon JG, Grayburn PA, Fleming JB, Brekken RA. Monitoring response to anticancer therapy by targeting microbubbles to tumor vasculature.
Clin Cancer Res
2007
;
13
:
323
–30.
31
Sawano A, Iwai S, Sakurai Y, et al. Flt-1, vascular endothelial growth factor receptor 1, is a novel cell surface marker for the lineage of monocyte-macrophages in humans.
Blood
2001
;
97
:
785
–91.
32
Chang Y, Zuka M, Perez-Pinera P, et al. Secretion of pleiotrophin stimulates breast cancer progression through remodeling of the tumor microenvironment.
Proc Natl Acad Sci U S A
2007
;
104
:
10888
–93.
33
Choudhuri R, Zhang HT, Donnini S, Ziche M, Bicknell R. An angiogenic role for the neurokines midkine and pleiotrophin in tumorigenesis.
Cancer Res
1997
;
57
:
1814
–9.
34
Salnikov AV, Heldin NE, Stuhr LB, et al. Inhibition of carcinoma cell-derived VEGF reduces inflammatory characteristics in xenograft carcinoma.
Int J Cancer
2006
;
119
:
2795
–802.
35
Heiss C, Wong ML, Block VI, et al. Pleiotrophin induces nitric oxide dependent migration of endothelial progenitor cells. J Cell Physiol 2007;2:366–73.
36
Vroling L, Yuana Y, Schuurhuis GJ, et al. VEGFR2 expressing circulating (progenitor) cell populations in volunteers and cancer patients.
Thromb Haemost
2007
;
98
:
440
–50.

Supplementary data