Hypoxia-recruited angiogenic neutrophils

In this issue of Blood, Massena et al identify a novel CD49d⁺CXCR4^highVEGFR1^high population of neutrophils that specifically migrate to sites of hypoxia and enhance angiogenesis.¹ 

Tissue hypoxia results from an imbalance between oxygen supply and demand. Adaptation to this metabolic stress includes a switch to glycolytic metabolism and angiogenesis. Best studied in tumor biology, the hypoxic core of a tumor secretes proangiogenic vascular endothelial growth factor A (VEGF-A) to mitigate the metabolic burden and ultimately gain access to circulation. Inflammation and hypoxia are now appreciated to be inextricably linked, with hypoxia occurring at sites of inflammation and inflammation incited as a result of tissue hypoxia. In particular, neutrophils have long been known to migrate to ischemic tissues and mitigate the deleterious effects of reperfusion through lipid peroxidation.²  Moreover, upon recruitment to inflamed tissues, neutrophils can themselves elicit tissue hypoxia through the action of the respiratory burst.³  Myeloid cells, including neutrophils and monocytes, have been observed at sites of angiogenesis and are known to secrete VEGF-A and matrix metalloproteinase 9 (MMP9), likely contributing to angiogenesis.⁴  However, the mechanisms of their recruitment and subsequent impact on neovascularization are largely unknown.

In the present study, Massena and colleagues used transgenic mice deficient in VEGF receptors 1 (VEGFR1; Flt-1 tk^−/−) and 2 (VEGFR2; tsad^−/−) and generated bone marrow chimeric mice to differentiate between neutrophil-dependent and endothelial-dependent responses to VEGF-A. The authors demonstrate that VEGFR1, expressed on a defined subset of circulating neutrophils, and VEGFR2, expressed on endothelia, are both required for efficient trafficking and diapedesis of neutrophils at sites of hypoxia.

Neutrophils responding to inflammatory stimuli typically are recruited to and migrate across endothelia following engagement of lymphocyte-associated antigen 1 (LFA-1; CD11a/CD18) or macrophage antigen 1 (Mac-1; CD11b/CD18) integrins. Massena and colleagues demonstrate that recruited neutrophils, migrating in response to VEGF-A, preferentially used very late antigen-4 (VLA-4) integrin (CD49d/CD29) over LFA-1 or Mac-1. Interestingly, CD29 (ITGB1) is known to be induced by hypoxia⁵  in fibroblasts, which may partially explain the selective recruitment of this CD49d⁺ population of neutrophils to sites of hypoxia. The authors subsequently identified a subset of circulating neutrophils in both mice and humans that coexpresses CD49d, VEGFR1, and CXC chemokine receptor 4 (CXCR4). Stromal cell–derived factor-1 (SDF-1), the primary ligand for CXCR4, is a known hypoxia-dependent chemokine.⁶  Chemokines, such as SDF-1, are known to bind glycosaminoglycans on the surface of endothelia to promote leukocyte homing and chemotaxis. Taken together, these findings suggest that a specific subpopulation of neutrophils present in circulation are primed to respond to and preferentially track to sites of localized hypoxia.

Under basal conditions, the majority of neutrophils reside in bone marrow. Following stimulation, neutrophils enter circulation and exhibit a short (8-16 hours) half-life. However, following recruitment to tissues, tumors, or sites of infection with subsequent exposure to inflammatory stimuli, such as granulocyte macrophage colony-stimulating factor,⁷  neutrophil longevity is significantly extended compared with their circulating counterparts. To maintain homeostasis, aging neutrophils are thought to be efferocytosed by macrophages in the spleen, liver, or return to bone marrow. Given the relatively small percentage of circulating neutrophils identified in the current study that express this triple-positive phenotype, it is unclear what stage of differentiation such a population represents. The authors acknowledge that CD49d expression is downregulated following maturation, which may indicate an immature neutrophil population. Somewhat paradoxically, CXCR4 expression on neutrophils has been characterized as a retention signal for bone marrow neutrophils,⁸  a regulator of stress granulopoiesis,⁹  and a means to remove senescent neutrophils from circulation by returning to the bone marrow.¹⁰  Whether this newly defined CD49d⁺CXCR4^highVEGFR1^high subpopulation represents a less differentiated, activated, functionally distinct, or longer-lived neutrophil has yet to be established. In terms of functionality, the authors begin to address these questions, demonstrating that CD49d⁺ neutrophils exhibit enhanced chemokinesis in response to VEGF-A. Furthermore, in supplemental Figure 9, the authors demonstrate that human CD49d⁺ neutrophils exhibit a higher ability to degrade basement membrane than CD49d⁻ neutrophils, likely through the action of MMP9 release, supporting the suggestion that this newly identified triple-positive population of neutrophils is proangiogenic.

To summarize, the authors Massena et al describe a novel mechanism to support the role of a specific subpopulation of neutrophils that specifically traffic to sites of hypoxia. This process is dependent on neutrophil VEGFR1 and endothelial VEGFR2 expression. Recruited neutrophils coexpress CD49d, CXCR4, and VEGFR1 and use VLA-4 integrin to facilitate extravasation. This study sheds new light on the immune contribution to angiogenesis and has important ramifications for potential targeted therapies in the future.