Chemokine Signal Behind VEGF-Driven Retinal Neovascularization Identified

Abnormal blood vessel growth in the retina is a leading cause of blindness worldwide, driving vision loss in proliferative diabetic retinopathy, retinopathy of prematurity and age-related macular degeneration. Current treatments block vascular endothelial growth factor, or VEGF, the central molecular driver of this pathological angiogenesis, but they fall short for many patients, producing partial responses, resistance and recurrence while requiring repeated injections into the eye. A new study published in Angiogenesis now reveals a previously underappreciated layer of control in this process: a chemokine signaling cascade that operates downstream of VEGF and appears essential for the abnormal vessel growth that damages the retina.

The research, led by Chandra Sekhar Yadavalli, Raj Kumar, Suresh Govatati and Gadiparthi N. Rao at the University of Tennessee Health Science Center, focuses on interleukin-8, a small secreted signaling protein of the ELR-positive CXC chemokine family. Elevated levels of IL-8 had already been detected in the vitreous humor of patients with proliferative diabetic retinopathy, hinting at a role in retinal angiogenesis, but its actual contribution had never been directly tested. Because mice do not possess an IL-8 gene, the team worked with the murine orthologues CXCL1 and CXCL2, which share functional similarities with the human protein, including signaling through the same receptor, CXCR2.

The investigators began with mouse retinal microvascular endothelial cells, the building blocks of the retinal vasculature. When they treated these cells with VEGFA, the principal angiogenic member of the VEGF family, they observed a time-dependent rise in CXCL1 and CXCL2 at both the messenger RNA and protein levels, with maximal induction at four hours that was sustained thereafter. This established that VEGFA does not merely act on endothelial cells directly but also prompts them to manufacture chemokines that could act on themselves and their neighbors.

To test whether those chemokines matter functionally, the researchers added recombinant CXCL1 or CXCL2 to the mouse endothelial cells and measured the four hallmark events of angiogenesis: migration, proliferation, sprouting and tube formation. Each chemokine alone significantly enhanced all four responses, and combining them produced little additional effect, suggesting overlapping actions. The critical experiment came with neutralizing antibodies. Blocking CXCL1 or CXCL2 individually blunted the cells’ migratory and proliferative responses, while blocking both together nearly abolished proliferation and completely eliminated tube formation. When the team applied these neutralizing antibodies to VEGFA-treated cells, the combinatorial blockade completely suppressed VEGFA-induced migration, proliferation, sprouting and tube formation, demonstrating that VEGFA requires these chemokines to exert its full pro-angiogenic effect.

The study then traced the signaling pathway that connects VEGF receptor activation to chemokine production, and here the work intersects with an unexpected player: the unfolded protein response, a cellular stress program usually associated with protein-folding problems in the endoplasmic reticulum. VEGFA triggered the cleavage and activation of ATF6, a stress-sensing transcription factor, along with upregulation of its downstream chaperone GRP78. When the researchers depleted ATF6 or GRP78 using small interfering RNAs, VEGFA could no longer induce CXCL1 or CXCL2 expression. The team had previously shown that VEGFA induces expression of VEGFR3, traditionally a receptor for VEGFC and VEGFD, through VEGFR2, and that both receptors were needed for ATF6-GRP78 activation. In the new work, silencing either VEGFR2 or VEGFR3 blocked VEGFA-induced ATF6 cleavage, GRP78 induction and chemokine expression, while VEGFR2 depletion abolished VEGFA-induced VEGFR3 expression but not the reverse. The result is a defined hierarchy: VEGFR2 drives VEGFR3 expression, which together enables ATF6-GRP78 signaling and, ultimately, chemokine production.

Crucially, the same cascade appears conserved in human cells. In human retinal microvascular endothelial cells, VEGFA induced IL-8 and CXCR2 expression with kinetics mirroring the mouse chemokine response. Recombinant IL-8 alone promoted human endothelial migration, proliferation, sprouting and tube formation, and neutralizing antibodies against IL-8 suppressed both IL-8-driven and VEGFA-driven angiogenic responses. Depleting ATF6, GRP78, VEGFR2 or VEGFR3 in the human cells likewise blocked VEGFA-induced IL-8 production, confirming that the VEGFR2-VEGFR3-ATF6-GRP78 axis governs chemokine induction across species.

The receptor side of the pathway also came into focus. VEGFA increased CXCR2, the receptor shared by CXCL1, CXCL2 and IL-8, in both mouse and human retinal endothelial cells, without affecting the related receptor CXCR1. Neutralizing CXCR2 inhibited chemokine-induced and VEGFA-induced migration, proliferation and tube formation in both cell types, and its induction likewise depended on VEGFR2, VEGFR3, ATF6 and GRP78. This positions CXCR2 as the functional conduit through which the chemokines exert their pro-angiogenic effects downstream of VEGF signaling.

To determine whether the mechanism operates in living tissue, the team used the oxygen-induced retinopathy model, a standard animal model of pathological retinal neovascularization in which mouse pups are exposed to 75 percent oxygen from postnatal day 7 to day 12 and then returned to room air, creating relative retinal hypoxia. Retinas examined 72 hours after oxygen withdrawal showed ATF6 cleavage and elevated GRP78, CXCL1/2 and CXCR2, and siRNA-mediated depletion of ATF6, GRP78, VEGFR2 or VEGFR3 blunted these inductions in vivo. Immunostaining revealed increased endothelial proliferation in post-OIR retinas, which neutralizing antibodies against CXCL1/2 attenuated, and retinal flat mounts from 120 hours post-OIR showed abundant filopodia-like sprouting protrusions that were completely suppressed by blocking CXCL1/2 or CXCR2.

At the level of whole retinal vasculature, the results were striking. Post-OIR retinas treated with control antibodies displayed dense, disordered vascular networks characteristic of pathological neovascularization, whereas retinas treated with CXCL1/2 or CXCR2 neutralizing antibodies showed significantly reduced neovascular area and vascular density. Importantly, blocking the chemokine axis had no effect on the normal retinal vasculature of normoxic pups, indicating that CXCL1/2-CXCR2 signaling is required for pathological but not physiological retinal angiogenesis. The authors note a caveat: the increased avascular area observed after blockade could partly reflect impaired reparative angiogenesis during recovery from the oxygen insult, so the axis may also contribute to vessel regrowth after injury.

The findings carry translational weight. IL-8 has been implicated in anti-VEGF resistance, since it can induce VEGF expression and sustain angiogenesis even when VEGF signaling is therapeutically suppressed, and it promotes vascular permeability and leukostasis that worsen retinal edema. The authors propose that IL-8 and CXCR2 act both downstream and, based on prior work from other laboratories, upstream of VEGF, potentially forming a feed-forward inflammatory loop that amplifies VEGF bioavailability and endothelial responsiveness. Targeting IL-8 signaling, they suggest, may therefore be a more desirable strategy than anti-VEGF therapy alone. They also caution that whether ATF6-GRP78 plays a direct or permissive role in chemokine induction remains unknown, and that validating IL-8’s role in human disease will require additional studies, possibly with an IL-8 humanized mouse model, since mice lack the human gene. Within those limits, the study delivers a complete signaling chain, from VEGFR2 through VEGFR3, ATF6 and GRP78 to CXCL1/2, IL-8 and CXCR2, and demonstrates in a living animal model that interrupting that chain curbs the abnormal vessel growth that destroys sight.

Subject of Research: The VEGFR2-VEGFR3-ATF6-GRP78-dependent induction of CXCL1/2, IL-8 and CXCR2 signaling in pathological retinal neovascularization

Article Title: VEGFR2-dependent VEGFR3 expression leads to CXCL1/2 and CXCR2 induction in the development of retinal neovascularization

Article References: Yadavalli, C. S., Kumar, R., Govatati, S., & Rao, G. N. (2026). VEGFR2-dependent VEGFR3 expression leads to CXCL1/2 and CXCR2 induction in the development of retinal neovascularization. Angiogenesis, 29(4), 82. https://doi.org/10.1007/s10456-026-10104-1

Image Credits: AI Generated

DOI: 10.1007/s10456-026-10104-1

Keywords: retinal neovascularization, VEGFA, IL-8, CXCL1, CXCL2, CXCR2, ATF6, GRP78, VEGFR2, VEGFR3, oxygen-induced retinopathy, diabetic retinopathy