In this report, we provide data that demonstrate the effectiveness of anti-VEGF therapy as a modulator of immune cell infiltration, and intra-tumoral and serum cytokine levels in multiple preclinical models of breast cancer. It is becoming increasing clear that the effect of anti-VEGF agents extends beyond the inhibition of angiogenesis, as many immune cells express VEGFRs, including macrophages, neutrophils, MDSCs, DCs and T-cells 
. We and others have shown a reduction in macrophages in tumors from animals treated with anti-VEGF therapy 
. In preclinical models of colon cancer, sunitinib treatment reduced the accumulation of MDSCs and plasmacytoid dendritic cells in tumors compared to control treatment 
. In metastatic renal cell carcinoma patients, sunitinib therapy reduced the level of circulating MDSCs and reduction in MDSCs in response to sunitinib therapy correlated with an increase in T-cell IFN-γ production 
. These studies suggest that sunitinib or other anti-VEGF therapies function as modulators of antitumor immunity. Here, we demonstrate that inhibition of VEGF binding to VEGFR2 with r84 is more effective than other anti-VEGF strategies in controlling breast tumor growth and the infiltration of immune suppressor cells.
Using the MDA-MB-231 human breast cancer model, we demonstrate that inhibition of VEGF binding to VEGFRs with bevacizumab or r84 effectively controls tumor growth (). Furthermore, only selective inhibition of VEGF binding to VEGFR2 with r84 is able to prevent an increase in MVD from week 1 to week 4 of therapy (). Interestingly, when these experiments were repeated using the immunocompetent 4T1 inflammatory breast cancer model, mcr84 and the VEGFR1 & VEGFR2 binding peptoid, GU81, were able to control tumor growth (). Though mcr84 was able to reduce MVD as seen in the MDA-MB-231 model, 4T1 tumors from animals treated with chronic mcr84 had an increase in vascular area compared to control (). In an effort to explain this increase in vascular area, we evaluated intra-tumoral cytokine levels in tumors from all treatment groups at the one and three week time points (Table S3
). Though hypoxia and VEGF are the main angiogenic stimuli, other cytokines, including IL-1β, IL-6 and TNF-α can induce angiogenesis 
. For example, in cardiac myocytes IL-1β increases VEGFR2 expression 
. In mcr84-treated tumors, we found increased levels of IL-1β and VEGFR2 compared to control-treated tumors (, Fig. S2C
). By linear regression analysis, increases in intra-tumoral IL-1β following anti-VEGF therapy correlated with increased vascular area (). VEGFR2 is the main receptor responsible for VEGF-induced angiogenesis and is activated by VEGF-A, C, or D 
. It is plausible that VEGF-C/D may bind to increased levels of VEGFR2, resulting in increased VEGFR2 phosphorylation and signaling. Increased VEGFR2 signaling can then promote the increased vascular area observed in mcr84- treated tumors.
The effects of anti-VEGF therapy extend beyond its effects on tumor blood vessels. In both the MDA-MB-231 and 4T1 models, chronic anti-VEGF therapy reduced macrophage infiltration in all treatment groups. In non-tumor bearing animals, monocyte and macrophage migration is driven in part by placental growth factor (PlGF), VEGF and VEGFR1 
. However, we have shown previously that VEGFR2 is expressed on macrophages from tumor-bearing animals, and is the dominant receptor driving VEGF-induced macrophage chemotaxis in tumor-bearing animals 
. Therefore, the reduction in macrophage migration seen following anti-VEGF therapy is likely due to the inhibition of VEGFR2 signaling (; ). Interestingly, the inhibition of both VEGFR1 and VEGFR2 activation with bevacizumab, GU81, or sunitinib did not reduce macrophage infiltration better than agents that inhibited the activation of VEGFR2 alone (r84 or RAFL-2).
Neutrophils are often described as ‘first responders’ and have been shown to be capable of mediating the angiogenic switch in engineered animal models of cancer 
. The mechanism underlying the increase in 7/4+
cells after anti-VEGF therapy is unclear. VEGF can stimulate neutrophil migration in vitro
via VEGFR1 activation 
. Furthermore, these cells were shown to express VEGFR1 and VEGFR2 by RT-PCR, suggesting that even though VEGFR2 is present, VEGFR1 is the primary receptor mediating VEGF-induced migration of these cells. In support of this, we have previously shown an increase in neutrophil infiltration in r84-treated tumors 
. In this study, we have further characterized the effect of anti-VEGF therapy on neutrophil infiltration utilizing an immunocompetent model of breast cancer. In the 4T1 model, tumors from animals treated with agents that block VEGFR1 activation (GU81 and sunitinib) had reduced neutrophil infiltration compared to control-treated tumors. In contrast, tumors from animals treated with mcr84, where VEGFR1 signaling was intact, had an increase in neutrophil infiltration (), suggesting that VEGFR1 is the dominant receptor involved in VEGF-mediated neutrophil migration in tumor-bearing animals.
VEGF is a key mediator in the development and maturation of dendritic cells 
. Activation of VEGFR1 on dendritic cells inhibits hematopoetic stem cell differentiation along the dendritic cell lineage, whereas VEGFR2 is important for dendritic cell maturation 
. Previously, using the MDA-MB-231 model, we found an increase in mature dendritic cells in animals treated with r84, but not bevacizumab 
. In this study, we demonstrate that this effect is seen after only one week of therapy, as inhibition of VEGF binding to VEGFR2 with mcr84 reduced the number of total dendritic cells, while increasing the mature fraction of these cells (). Though the antigen presenting ability of these cells is not known, in human breast cancer specimens, increased CD83+ dendritic cells is associated with an improved prognosis 
The role of VEGF in myeloid-derived suppressor cells (MSDC) differentiation and migration has been characterized in recent years 
. In mouse models and patients with cancer, these immunosuppressive cells are increased in the blood, spleen and tumors 
. Many factors induce MDSC expansion and activation, including VEGF, IL-1β, and IL-6, making these attractive targets for MDSC inhibition. In this study, we reveal an interesting connection between these cytokines in mediating MDSC infiltration into tumors. In MDA-MB-231 tumors, treatment with anti-VEGF agents that block both VEGFR1 and VEGFR2 (bevacizumab and GU81), but not other receptor tyrosine kinases resulted in increased intra-tumoral IL-1β levels and MDSC accumulation (, ). It is interesting to note that sunitinib, which also blocks both VEGFR1 and VEGFR2 signaling, does not result in increased IL-1β and MDSC accumulation. This is likely due to the fact that sunitinib also blocks PDGFRβ, GSF-1R, Flt-3, and cKit, any of which may be important for increased expression of IL-1β. Furthermore, changes in IL-1β levels in response to anti-VEGF therapy were highly correlative with changes in MDSC infiltration at the one and four week time-points (), indicating that IL-1β is a key cytokine mediating the infiltration of MDSCs following anti-VEGF therapy. Interestingly, in the 4T1 immunocompetent model of breast cancer, changes in intra-tumoral IL-1β levels correlate negatively with changes in MDSC infiltration after three weeks of therapy (, ), where increases in IL-1β following treatment with mcr84 were associated with reduced MDSC infiltration. These findings indicate a possible bimodal role of IL-1β in MDSC infiltration, where a low level of IL-1β induces MDSC infiltration and increased levels following anti-VEGF therapy inhibit MDSC infiltration.
MDSCs do not express receptors for IL-1β; however, they do express receptors for IL-6, which is capable of inducing MDSC infiltration in the absence of IL-1β signaling 
. Therefore, we investigated other cytokines that may be involved in regulating immune cell recruitment. Similar to IL-1β, tumors from animals treated chronically with mcr84 had increased IL-6 and CXCL1 levels. However, only changes in CXCL1, not IL-6, correlated negatively with changes in MDSC infiltration, as seen with IL-1β (, ). Therefore, we propose that CXCL1 is an inhibitor of MDSC infiltration subsequent to markedly increased IL-1β levels following anti-VEGF therapy.
Treg are immune suppressor cells that maintain peripheral tolerance 
. Like MDSCs, these cells are increased in the blood and tumors of cancer patients and mouse models of cancer 
. The generation of Treg is a complicated process that involves many cytokines, such as IL-2, TGF-β, IFN-γ and TNF-α 
. Though acute anti-VEGF therapy did not affect Treg infiltration, we found reduced levels of Treg after chronic anti-VEGF therapy in all treatment groups (; ). Furthermore, only changes in intra-tumoral IFN-γ levels correlated with changes in Treg infiltration, further confirming its importance in Treg infiltration/generation ().
Finally, having demonstrated changes in intra-tumoral cytokine levels and immune cell infiltration with anti-VEGF therapy, we looked at serum levels of IL-β and IL-6, as potential biomarkers of response to anti-VEGF therapy. For animals treated with mcr84 or GU81, we found that decreases in serum levels of IL-1β and IL-6 were highly correlative with changes in tumor size in the presence of anti-VEGF therapy (). Interestingly, increases in serum IL-6 levels in sunitinib-treated animals correlated with increases in tumor size, suggesting different mechanisms for the cytokine aberrations seen between selective versus broad spectrum anti-VEGF strategies ().
In conclusion, we have demonstrated differences in the ability of anti-VEGF therapy to affect tumor vasculature and modulate immune cell infiltration, intra-tumoral and serum cytokine levels depending on the mechanism of VEGF inhibition. We have demonstrated that selective inhibition of VEGF binding to VEGFR2 with r84 is effective at controlling tumor growth, inhibiting the infiltration of suppressive immune cells (MDSC, Treg, macrophages) while increasing the mature fraction of dendritic cell infiltrates. Furthermore, we have identified two possible biomarkers (IL-1β and IL-6) for assessing the efficacy of anti-VEGF therapy in breast cancer patients.