In this report we demonstrate that Eyes Absent promotes angiogenesis specifically through its tyrosine phosphatase activity. While a role for the EYAs in angiogenesis has not been previously reported, it is not an unexpected observation in light of the association between the EYAs and cell migration, and the cardiovascular defects associated with Eya mutations in humans and in mice. Effects of EYA inhibition were apparent in the processes of both vasculogenesis (HUVEC capillary tube formation, formation of the zebrafish aorta and posterior cardinal vein) and sprouting angiogenesis (aortic ring assay, formation on intesegmental vessels in zebrafish). Upon bathing zebrafish embryos in the EYA inhibitors we see a clear effect on vascular development at doses that do not affect the general morphology, suggesting a degree of specificity. On the other hand when the inhibitors were administered relatively late no significant change in the development of intersegmental vessels was seen. Hence from the currently available data it is not possible to specify whether the inhibitors act exclusively on either vasculogenesis or angiogenesis. In vitro EYA inhibition does not appear to attenuate endothelial cell proliferation. Hence the pro-angiogenic function of EYA could be a direct result of its ability to activate a cell migration-promoting pathway. The EYA substrate involved in this process is yet to be identified.
However another possibility is that dephosphorylation of H2AX at Tyr142 
underlies the pro-angiogenic role played by EYA. Angiogenesis can be promoted by hypoxia during normal development as well as in pathological conditions such as retinopathy of prematurity (ROP), diabetic retinopathy and tumor angiogenesis (reviewed in 
). Hypoxia leads to DNA damage and the phosphorylation of H2AX at Ser139 (to yield γ-H2AX) by ATR kinases. Re-oxygenation and the production of reactive oxygen species lead to double-strand breaks and the formation of γ-H2AX though the action of ATM kinase. Recent evidence suggests that such activation of H2AX contributes to angiogenesis and proliferative retinopathies
. The constitutively Tyr142-phosphorylated form of H2AX 
promotes recruitment of the p53 apoptosis machinery and disfavors DNA damage repair 
. Activation of ATM/ATR also leads to EYA activation and dephosphorylation of H2AX-pY142
, thus permitting DNA repair 
. Hence hypoxia-induced angiogenesis could be promoted by the tyrosine phosphatase activity of the EYA proteins via dephosphorylation of H2AX-pY142
. In this manner EYA is likely to play a role in both developmental and pathological retinal angiogenesis.
In this study small molecule inhibitors of the Eyes Absent tyrosine phosphatase activity were used as chemical probes of EYA's cellular function because they permit separation of EYA's signal transduction and transcriptional activities. However the inhibitors also have other potential uses. The growing evidence that EYAs are over-expressed in malignancies raises the possibility that inhibitors could be a targeted therapy for appropriate subsets of breast and ovarian cancers, as well as malignant peripheral nerve sheath tumors 
. Furthermore, since the EYA PTP activity promotes DNA damage repair 
, it is likely that EYA inhibitors could potentiate commonly used genotoxic therapeutic regimens such as chemotherapy and ionizing radiation.
Interestingly Benzbromarone was also identified in an early screen for PTP1B inhibitors 
, and more detailed analyses and crystallographic studies of Benzbromarone-derivatives were later reported 
. These compounds bind to an allosteric site on PTP1B far from the active site and are noncompetitive inhibitors that prevent formation of the active “closed” conformation of the enzyme. In contrast our studies suggest that Benzbromarone binds in a hydrophobic pocket adjacent to the active site of EYA, and it is not competitive with small substrate mimics such as pNPP and phosphotyrosine. Furthermore both the docking studies () and kinetic analyses () demonstrate that, despite the presence of a phenol group in all of the EYA inhibitory compounds in , they are unlikely to be acting as simple tyrosine-mimics. Inhibition of both EYA3 and EYA2 were tested in this analysis and the compounds are similarly effective. This result is not surprising given the high degree of sequence conservation in the predicted inhibitor binding sites. Analyses of benzbromarone related compounds in has also permitted a separation of the chemical scaffolds necessary for EYA inhibition versus PTP1B inhibition; compound 1a is preferentially an EYA inhibitor while compound 1c is a PTP1B inhibitor. These data thus provide a basis for the design of more potent and selective EYA inhibitors.
EYA is a metallo-enzyme 
. However the inhibitors described here are not metal-chelators, unlike compounds identified in a recently reported in silico
screen performed using the isolated catalytic domain of EYA2 
. While metal chelators can be high affinity inhibitors of metalloenzymes, as therapeutics they present challenges such as depletion of physiologically needed metal ions and inhibition of non-targeted metalloenzymes. Hence the identification of EYA inhibitors that do not appear to coordinate the active site divalent metal in the EYAs opens up new possibilities.
Compound 1 (Benzbromarone) and compound 1a (Benzarone) are both known uricosuric agents. They suppresses uric acid reabsorption via inhibition of the urate transporter SLC22A12 (URAT1) 
. The compounds have a short half-life (about 3 hours 
), but their major metabolite 6-hydroxybenzbromarone has a longer half-life of 30 hours and retains the uricosuric activity. These compounds are no longer marketed for use in the treatment of gout in most countries because of reported hepatotoxicity and their ability to inhibit CYP2C9 which causes undesired interactions with drugs that use this enzyme for clearance, such as warfarin (reviewed in 
). However there remains controversy about whether the benefit-risk ratio warrants a complete withdrawal 
. There have been no previous reports of either compound affecting cell motility or angiogenesis. Hence the data presented here represents an opportunity for re-purposing a well-characterized drug as an anti-angiogenic and anti-metastatic agent.
In summary we have identified a class of compounds that strongly inhibit the EYA tyrosine phosphatases. These compounds display selectivity towards EYA relative to a representative classical PTP, PTP1B. The inhibitors effectively attenuate the known ability of the EYAs to promote cell motility and have been used as a tool to reveal a pro-angiogenic function for the EYA tyrosine phosphatase. These inhibitors could thus be useful in identifying other biological role(s) of this multi-tasking protein, as well as in probing the molecular mechanisms by which it functions. The compounds also represent viable leads for the development of EYA-targeted therapeutics with potential uses in the treatment of cancer metastasis, potentiating DNA damaging chemotherapy, tumor angiogenesis, as well as other vasculopathies.