Endothelial cell FGF signaling is required for injury response but not for vascular homeostasis Sunday S. Oladipupoa,1, Craig Smitha, Andrea Santefordb, Changwon Parkc, Abdoulaye Seneb, Luke A. Wileyb, Patrick Osei-Owusud,2, Joann Hsua, Nicole Zapatab, Fang Liuc, Rei Nakamurab, Kory J. Lavinea,e, Kendall J. Blumerd, Kyunghee Choic, Rajendra S. Aptea,b,3, and David M. Ornitza,3 Departments of aDevelopmental Biology, bOphthalmology and Visual Sciences, cPathology and Immunology, dCell Biology and Physiology, and eMedicine, Washington University School of Medicine, St. Louis, MO 63110 Edited* by Kari Alitalo, University of Helsinki, Helsinki, Finland, and approved July 30, 2014 (received for review December 30, 2013)
|
choroidal neovascularization oxygen-induced retinopathy retinopathy of prematurity neoangiogenesis
|
that vascular FGF signaling was required to maintain vascular integrity. Although important insights were gained, whether FGF signaling was required specifically in the EC vs. in another vascular cell type was not determined. These studies also did not address whether EC FGF signaling is required during embryonic development, postnatal vascular homeostasis, or in various injury settings. It is becoming increasingly clear that the FGFR pathway interacts with other receptor tyrosine kinases, such as VEGFR2 (Flk1), to regulate normal physiological and pathological processes (8, 15, 17). Recently, Murakami et al. (15) showed that ECs lacking FGF signaling become nonresponsive to VEGF–VEGFR2 signaling, suggesting that the FGF pathway is upstream of VEGF signaling. However, because other studies imply otherwise (10, 18), and both pathways are often being targeted simultaneously in diseases with deregulated angiogenesis (8), additional analyses are needed to clarify these conflicting conclusions. By using engineered mice deficient in Fgfr1 and Fgfr2 in cells of both endothelial and hematopoietic lineages we demonstrate a functional in vivo requirement for cell-autonomous FGFR1/2 signaling in ECs during injury response and pathologic neovascularization. Surprisingly, our data also suggests that EC (and
|
Significance FGF receptor (FGFR) signaling is thought to be essential for vascular development, homeostasis, and pathological angiogenesis. However, the in vivo requirements and the cellular targets of FGF in the vasculature are not known. Here, we show that endothelial FGFR1 and FGFR2 are not required for vascular homeostasis or physiological functions and are likely not required for embryonic development. However, endothelial FGFR1 and FGFR2 are essential for neovascularization after skin or eye injury or following retinal ischemia. These findings reveal a key requirement for cell-autonomous endothelial FGFR signaling in tissue repair and neovascularization following injury and validate the endothelial cell FGFR as a target for diseases associated with aberrant vascular proliferation such as age-related macular degeneration, diabetic retinopathy, and wound healing.
N
eovascularization is critical for tissue repair and pathological conditions, including aberrant ocular angiogenesis and cancer (1–4). Although FGF signaling has been prominently implicated in these processes based on genetic inactivation experiments in mice and in vitro studies, the functional in vivo requirement of this pathway in the endothelial cell (EC) vs. other vascular cell types is not known (5–8). The FGF family is composed of 18 signaling ligands, which interact with four cell surface tyrosine kinase receptors. FGF receptor (FGFR) signaling regulates many biological processes, including survival, differentiation, proliferation, and angiogenesis through the activation of RAS-RAF-MAPK, PI3K, STAT, and PLC gamma pathways (6, 9). The EC response to FGF signals is well described in in vitro models of angiogenesis (10, 11). Moreover, previous gene expression analysis showed that Fgfr1 and Fgfr2 were the predominant Fgfrs in ECs (5), whereas Fgfr3 was sparsely detected (12, 13) and Fgfr4 expression was not reported (8). To this end, and given the critical role of FGFRs 1 and 2 during embryonic development, we tested the hypothesis that EC FGFR1/2 may play a key role during vascular development, homeostasis, and response to injury. Studies aimed at understanding the functional requirement of vascular FGF signaling have demonstrated a critical role in homeostasis and angiogenesis (14–16). In these studies, in vivo expression of an adenoviral-based soluble FGF trap (sFGFR) or a dominant inhibitor of all FGFRs (FGFR1DN) was used to disrupt FGF signaling in the vasculature. These studies showed www.pnas.org/cgi/doi/10.1073/pnas.1324235111
Author contributions: S.S.O., K.C., R.S.A., and D.M.O. designed research; S.S.O., C.S., A. Santeford, C.P., A. Sene, L.A.W., P.O.-O., J.H., N.Z., F.L., and R.N. performed research; K.J.L. and K.J.B. contributed new reagents/analytic tools; S.S.O., C.S., R.N., K.J.B., K.C., R.S.A., and D.M.O. analyzed data; and S.S.O., K.J.L., K.C., R.S.A., and D.M.O. wrote the paper. The authors declare no conflict of interest. *This Direct Submission article had a prearranged editor. 1
Present address: Eli Lilly and Company, Indianapolis, IN 46285.
2
Present address: Department of Pharmacology and Physiology, Drexel University College of Medicine, Philadelphia, PA 19102.
3
To whom correspondence may be addressed. Email:
[email protected] or dornitz@ wustl.edu.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1324235111/-/DCSupplemental.
PNAS | September 16, 2014 | vol. 111 | no. 37 | 13379–13384
DEVELOPMENTAL BIOLOGY
Endothelial cells (ECs) express fibroblast growth factor receptors (FGFRs) and are exquisitely sensitive to FGF signals. However, whether the EC or another vascular cell type requires FGF signaling during development, homeostasis, and response to injury is not known. Here, we show that Flk1-Cre or Tie2-Cre mediated deletion of FGFR1 and FGFR2 (Fgfr1/2Flk1-Cre or Fgfr1/ 2Tie2-Cre mice), which results in deletion in endothelial and hematopoietic cells, is compatible with normal embryonic development. As adults, Fgfr1/2Flk1-Cre mice maintain normal blood pressure and vascular reactivity and integrity under homeostatic conditions. However, neovascularization after skin or eye injury was significantly impaired in both Fgfr1/2Flk1-Cre and Fgfr1/2Tie2-Cre mice, independent of either hematopoietic cell loss of FGFR1/2 or vascular endothelial growth factor receptor 2 (Vegfr2) haploinsufficiency. Also, impaired neovascularization was associated with delayed cutaneous wound healing. These findings reveal a key requirement for cell-autonomous EC FGFR signaling in injuryinduced angiogenesis, but not for vascular homeostasis, identifying the EC FGFR signaling pathway as a target for diseases associated with aberrant vascular proliferation, such as age-related macular degeneration, and for modulating wound healing without the potential toxicity associated with direct manipulation of systemic FGF or VEGF activity.
Fig. 1. Endothelial/hematopoietic FGFR1/2 is dispensable for vascular development and homeostasis in vivo. (A) Limb buds (embryonic day 11.5) from a Flk1-Cre, mT/mG embryo (Left) and a Flk1-Cre, mT/mG, DCKO embryo (Right) showing normal vascular patterns (green). (B) Limb bud (embryonic day 12.0) from a DFF embryo (Left) and a DCKO embryo (Right) immunostained for CD31, showing normal vascular plexus formation. (C) Representative MECA32 (+) immunofluorescence micrographs showing normal microvascular density and morphology in DFF and DCKO ear skin, lung, and kidney, and a normal retina vascular plexus formation visualized in mice perfused with FITC-dextran. (D) Meca32 (+) vessel count (external ear skin), quantitative immunofluorescence normalized to DAPI (lung and kidney), and FITC-dextran fluorescent area (retina), showing no difference in the microvascular network between DFF and DCKO mice (n = 3). All values are mean ± SD. Lung, kidney, and retina were imaged with a 10× objective and the external ear skin was imaged with a 20× objective. Data were analyzed using the unpaired Student t test.
in Flk1-Cre DCKO limb buds compared with DFF and Cre control limb buds. Furthermore, DCKO mice were viable, phenotypically normal, and present in normal Mendelian ratios. Additionally, vascular patterns and density, examined in adult ear skin, lung, kidney, and retina, seemed normal (Fig. 1C) and had similar vascular densities as determined by immunostaining ECs or labeling the vasculature with FITC dextran (Fig. 1D). Western blot analysis of other tissues probed with the vascular marker VE-cadherin also showed no difference between DCKO and DFF littermates (Fig. S1A). Vascular beds in Tie2-cre DCKO adult tissues were also normal (Fig. S1 B–D). Because we did not observe a phenotype in Flk1-Cre or Tie2-Cre or DFF mice, the remaining studies were performed with DFF mice as controls, except where otherwise noted. To determine whether EC and hematopoietic Cre-mediated inactivation of Fgfr1/2 was maintained postnatally, FACS adult lung EC (VE-cadherin/CD31–positive) and bone marrow (CD45positive) cells were analyzed for target gene deletion. Adult lung ECs showed 84% and 87% reduction in Fgfr1 and Fgfr2, respectively (Fig. 2A and Table S1). CD45-positive bone marrow cells showed an 88% reduction of Fgfr1 and almost undetectable Fgfr2, even in control cells (Fig. 2B). Consistent with previous results (22, 23), hematopoietic cell populations, including B cells, macrophages, granulocytes, erythrocytes, T cells, and hematopoietic progenitor cells were unaffected by loss of Fgfr1 (and Fgfr2) postnatally under homeostatic conditions (Fig. 2 C and D). Taken together, these data suggest that EC and hematopoietic FGFR1/2 signaling is not required for normal vascular and hematopoietic development or homeostasis. Regulation of Vegfr2 by FGF signaling has been demonstrated in vitro, suggesting interaction between FGF and VEGF signaling pathways in vascular endothelium (15). However, when Vegfr2 was measured in Flk1-Cre, DFF, and DCKO adult lung ECs no significant changes were observed (Fig. S2A). To rule out potential effects of Flk1 haploinsufficiency, we also probed lysates from Tie2-Cre targeted mouse tissues (DFF and DCKO) for VEGFR2 expression. Analysis of lung, liver, and back skin
hematopoietic) FGFR1/2 signaling is not required for embryonic development or for maintaining vascular integrity and function under homeostatic physiological conditions, despite the wellestablished role for FGF signaling in ECs for vascular development in vitro. Results Endothelial/Hematopoietic FGFR1/2 Is Dispensable for Developmental and Postnatal Angiogenesis. To examine the in vivo cell-autonomous
function of EC/hematopoietic FGFR1/2, we conditionally inactivated floxed alleles of Fgfr1 and Fgfr2 using a Flk1-Cre knockin allele (Flk1-Cre) that is haploinsufficient for Vegfr2 (19) and a Tie2-Cre transgenic allele (Tie2-Cre) (20) to generate Fgfr1/2Flk1-Cre and Fgfr1/2Tie2-Cre mice, respectively (referred to as double-conditional knockout or DCKO mice). Mice homozygous for floxed Fgfr1/2 alleles (DFF) without Cre and mice heterozygous for Cre (Flk1-Cre+/− or Tie2-Cre+/−) served as controls. To confirm EC Cre activation during development, we combined Flk1-Cre with the dual-fluorescent reporter allele mT/mG (21) (Cre-mediated replacement of membrane-targeted tomato with membrane-targeted GFP) to generate mice with the genotype Flk1-Cre, mT/mG. Immunofluorescence showed vascular patterns of mT/mG activation in Flk1-Cre limb buds (Fig. 1A, Left). To determine whether there were any phenotypic consequences of loss of EC FGFR1/2, we compared this pattern with DCKO mice also containing the mT/mG reporter allele (Fig. 1A, Right) and by immunostaining with an anti CD31 antibody (Fig. 1B). Surprisingly, embryonic development and the embryonic vascular network seemed normal 13380 | www.pnas.org/cgi/doi/10.1073/pnas.1324235111
Fig. 2. Flk1-Cre activation is maintained in adulthood and hematopoiesis is normal in mice lacking endothelial/hematopoietic FGFR1/2. (A and B) Quantitative RT-PCR analysis of FACS CD31- and VE-cadherin–positive adult lung ECs (A) and CD45-positive adult bone marrow hematopoietic cells (B), showing efficient depletion of Fgfr1 and Fgfr2 mRNA expression in DCKO mice. Data are represented as relative expression normalized to Hprt. (C and D) FACS analysis of bone marrow hematopoietic cells revealing normal myeloid (C) or lymphoid/stem cell populations (D) in DCKO compared with DFF mice. All values are mean ± SD. Data were analyzed using the unpaired Student t test; *P < 0.05; **P < 0.01; ns, not significant. n = 2–4.
Oladipupo et al.
showed similar levels of VEGFR2 (Fig. S2B). Overall, these data suggest that FGF signaling in ECs does not affect VEGFR2 levels in vivo. Endothelial/Hematopoietic FGFR1/2 Is Dispensable for Adult Basal Vascular Permeability and Functions. To determine whether cell-
autonomous loss of EC FGFR1/2 affected vascular permeability, we examined vascular integrity in the retina using FITC-dextran– based angiography and in skin using the Evans Blue (EB)-based Miles assay (24, 25). Under homeostatic conditions, there was no disruption in vascular permeability in the retina or skin of DCKO and control mice (Fig. 3 A–C). However, when challenged acutely with mustard oil, a proinflammatory agent, vascular permeability increased three- and fivefold in control and DCKO mouse ear skin, respectively, as measured by EB extravasation (Fig. 3 B and C). Notably, mustard oil-stimulated DCKO mouse ear skin showed a modest, but statistically significant (P < 0.04), increase in cutaneous deposition of EB compared with DFF mice. These data suggest that EC FGFR1/2 deficiency/VEGFR2 haploinsufficiency has no effect on normal physiological vascular
permeability but may be important for maintaining vascular integrity during acute inflammation. Hypertension is a common side effect of antiangiogenic therapy and can be attributed in part to diminution of the microvasculature (26, 27). Although we did not observe any identifiable vascular density or structural abnormalities in several DCKO tissues examined (Fig. 1 B and C), we asked whether blood pressure (BP) was altered in DCKO mice as a systemic functional consequence of EC-specific deletion of FGFR1/2. No difference in arterial BP was observed between DCKO and littermate controls (Fig. 3D), consistent with the observed normal steady-state vascular patterns. Similarly, we found that EC FGFR1/2 inactivation did not alter vascular reactivity, as shown by normal vasoconstriction induced by phenylephrine (PE), endothelium-dependent vasodilation stimulated by increasing concentrations of acetylcholine (ACh) after PE preconstriction, or endothelium-independent vasodilation stimulated by increasing concentrations of the nitric oxide donor sodium nitroprusside (SNP) after PE preconstriction (Fig. 3 E–G). Because the EC is often the cellular target in many ocular diseases, we also assessed visual acuity in mice lacking EC Fgfr1/2. Biomicroscopic examination of eyes in vivo showed normal retinal vascular morphology (Fig. S3), and photopic visual acuity testing, as assessed by opto-kinetic tracking analysis (28), showed normal values of 0.86 ± 0.02 and 0.84 ± 0.02 cycles per degree (n = 5, P > 0.1) for DFF and DCKO mice, respectively. Taken together, these results suggest that BP, vascular reactivity, and visual function as measured by visual acuity testing are not perturbed under homeostatic conditions owing to Vegfr2 haploinsufficiency and loss of EC FGFR1/2.
Fig. 3. FGFR1/2 is dispensable for basal vascular permeability, BP, and vascular reactivity in vivo. (A) FITC-dextran (2 MDa) whole-mount stained DFF and DCKO adult mouse retinas showing absence of vascular leak at basal levels (representative of 8–10 retinas per group). (B) Vascular leakage assessed by EB dye permeability showing similar ear skin coloration in vehicle-treated DFF and DCKO mice but increased leakage into the skin in DCKO mice compared with DFF mice following treatment with mustard oil. (C ) Quantitation of EB in ear tissue (mean ± SEM, n = 5–7, *P < 0.04). (D) Mean arterial BP. (E) Mesenteric artery (MA) vasoconstriction in response to PE (n = 4–6 animals, two vessels per animal per group). The data shown are the mean percentages of change in diameter of arteries perfused with Mops containing PE at indicated concentrations, expressed as the percentages of change in vessel diameter relative to baseline. (F) Vasodilation in response to Ach expressed as percentages of increase in arterial diameter after constriction with PE (100 μM). (G) Vasodilation in response to SNP expressed as percentages of increase in arterial diameter after constriction with PE (100 μM).
Oladipupo et al.
wound healing and for limiting scar size following myocardial ischemia (29–32). Despite these well-established functions for FGF signaling, its cellular sites of action in complex tissues are poorly defined. To determine whether cell-autonomous loss of EC FGFR1/2 signaling contributes to the neovascularization associated with wound healing, expression of vascular markers was examined 2–3 μm away from the punch biopsy epithelial margin 6 d after wounding. Compared with control DFF and Flk1-Cre or Tie2-Cre mouse skin, DCKO mice showed a significant decrease (∼1.5- to twofold) in microvascular density as demonstrated by reduced pan-EC markers, MECA32 or CD31, or smooth muscle actin-positive vessels at the wound margin (Fig. 4 A–C and Fig. S4 A–C). We also observed a reduction in total VEGFR2 and CD31 protein levels in DCKO mouse epithelial wound margin (Fig. 4 D–F, compare Flk1-Cre vs. DCKO). In addition to decreased neovascularization at the wound site, Flk1-Cre DCKO mice also showed delayed wound closure at a time point coincidental with decreased neovascular growth (Fig. 4G). We did not detect a significant difference in the percentage of proliferating endothelial cells between Tie2-Cre, DFF, or DCKO mice at day six following skin wounding (Fig. S4D). To identify signaling pathways that are active in the wound healing-associated angiogenic response and potentially impaired consequent to EC/hematopoietic FGFR1/2 disruption, we screened nonwounded and 6-d-postwounded skin with an angiogenesis pathway-focused protein array. Overall, there was an increase in the proangiogenic factors FGF2, VEGFA, and placental growth factor, but not FGF1, at the 6-d time point in DCKO and DFF mice (Fig. S5A). The similar levels of induction of these proangiogenic factors in DCKO and DFF mice are consistent with an absence of feedback regulation of ligand production by EC and hematopoietic FGF signaling. Because inflammation is a key component of the wound-healing response, and because Fgfr1 and Fgfr2 were inactivated in both the EC and hematopoietic lineages, we investigated whether loss of PNAS | September 16, 2014 | vol. 111 | no. 37 | 13381
DEVELOPMENTAL BIOLOGY
Impaired Neovascularization and Delayed Wound Healing in Mice Lacking Endothelial FGFR1/2. FGF signaling is important for
Fig. 4. Impaired endothelial neovascularization and delayed wound healing in mice lacking endothelial FGFR1/2. (A) Representative MECA32 (+) vessels (red) showing increased neovascular growth in Flk1-Cre and DFF control mouse wound edge (2 mm) 6 d after excision wound injury compared with normal unwounded skin. Neovascularization was significantly reduced in wounded DCKO mice (Lower Right). Epidermis is marked in green (Keratin 14) and nuclei in blue (DAPI). (B) Quantitative immunofluorescence of A, MECA32-positive vessels normalized to DAPI (n = 4–10). (C) Smooth muscle actin (SMA)-positive vessel count (three to five 20× fields per mouse) (n = 3–4 mice; see Fig. S2A for corresponding representative images). (D) Western blot of wound margin tissue from Flk1-Cre, DFF, and DCKO mice showing that Vegfr2 heterozygosity (Flk1-Cre mice) has no effect on microvascular density (CD31 levels) in response to skin injury compared with DFF mice. CD31 levels are reduced in DCKO, and not in Flk1-Cre and DFF, mice following injury. VEGFR2 levels are reduced by 50% in Flk1Cre mice compared with DFF controls and further reduced in DCKO mice. (E and F) Quantification of D using imageJ image analysis software. CD31 and VEGFR2 arbitrary density units (ImageJ) normalized to tubulin. (G) Wound healing is significantly delayed at 4 and 6 d after wounding in DCKO compared with Flk1-Cre and DFF control mice (n = 4–10). (H and I) F4/80 immunostaining (quantified in I) showing similar levels in wound-margin tissue from DFF and DCKO mice. A and H were imaged with a 10× objective. *P < 0.05, **P < 0.01, and ***P < 0.001.
FGFR1/2 signaling could affect hematopoietic cell infiltration into wounded tissue. There was no difference in the expression of the macrophage marker F4/80 and pan-hematopoietic cell marker CD45 in control vs. DCKO skin at the 2- and 6-d-postwound time points (Fig. 4 H and I and Fig. S5B), suggesting that there was no impairment of hematopoietic cell recruitment to the injured tissues. Collectively, these data indicate that the abnormal vascular response and delayed wound closure in DCKO mice is consequent to loss of FGFR1/2 signaling in ECs and not hematopoieticderived cells. Endothelial FGFR1/2 Is Required for Neovascular Response in Ocular Injury and Hypoxic Stress-Induced Pathogenic Angiogenesis. To fur-
ther investigate the role of EC FGFR1/2 in vascular response to injury, we examined vascular response to choroidal injury and retina hypoxia. Laser-induced choroidal neovascularization (CNV) is well described as a proliferative neovascular response to injury to the retinal pigment epithelium and Bruch’s membrane underneath the retina and a murine surrogate for the wet form of age-related macular degeneration (33, 34). Following choroidal injury, DCKO mice showed a marked reduction in CNV compared with control Flk1-Cre or Tie2-Cre and DFF mice, as determined by quantifying FITC-dextran–perfused neovascular complexes overlying the injury site (Fig. 5). To identify potential ligands critical to the formation of CNV and 13382 | www.pnas.org/cgi/doi/10.1073/pnas.1324235111
affected by EC FGFR1/2 disruption, we examined noninjured and injured choroidal tissues (24 h or 7 d after injury) by quantitative RT-PCR. Fgf9 and VegfA were similarly induced in DFF and DCKO mice 24 h after laser injury and Fgf2 and Fgf8 were induced, but to a lesser extent, in DCKO mice (Fig. S6 A–D). Seven days after laser injury, the levels of Fgf2 and Fgf8 were decreased in both DFF and DCKO mice (Fig. S6 E and F). These data suggest that induction of some FGF ligands and consequent FGFR1/2 signaling in the EC may play a role in the vascular response to ocular injury. In Flk1-Cre mice, Vegfr2 levels were reduced in choroidal tissue by about 50%, similar to what was observed in skin. This is attributed to heterozygosity of Vegfr2 in Flk1-Cre knockin mice. Twenty-four hours after laser injury, Vegfr2 levels were significantly (P < 0.05) increased in DCKO mice, but not in DFF mice (Fig. S7A). Seven days after laser injury, Vegfr2 levels were proportionally reduced in both DFF and DCKO choroid (Fig. S7B). Note that even though Vegfr2 levels were reduced by 50% in Flk1-Cre choroid compared with DFF choroid (owing to heterozygosity of Vegfr2), the angiogenic response to injury was unaffected (Fig. 5B). Collectively, these data show that the CNV response is critically dependent on EC FGFR1/2 and not on the level of Vegfr2. Next, we examined the vascular response to injury in a mouse model of human retinopathy of prematurity, called oxygeninduced retinopathy, by exposing day-7 neonates to high oxygen Oladipupo et al.
levels (75% O2) for 5 d followed by normoxia (25% O2) for an additional 5 d. Retinas from hyperoxia-exposed control Flk1-Cre and DFF mice showed avascular zones, vascular reperfusion, and neovascular tufts similar to what has been reported previously (Fig. S7 C and D) (35). In hyperoxia-exposed DCKO mice, the central avascular zones were significantly increased and the neovascular response was significantly blunted (Fig. S7 C and D, asterisks in C). Taken together, these results suggest that EC or hematopoietic FGFR1/2 signaling is required for the ocular response to hypoxic stress-induced neovascularization. Discussion Determination of the functional and in vivo requirement of FGF signaling in the vasculature may provide insight into opportunities for targeted cell-specific therapeutic interventions, given the systemic toxicity associated with global FGFR inhibition (36–39) and potential risks associated with global FGFR activation. Because the EC is often a therapeutic target, it is essential to know whether EC FGF signaling is directly required during development, tissue homeostasis, and response to injury. Using a loss-of-function genetic model, we discovered that endothelial and hematopoietic FGFRs 1 and 2 (in addition to Vegfr2 happloinsufficiency) are dispensable for homeostasis and basal vessel permeability and likely are dispensable for embryonic development. However, when subjected to either skin or eye injury, or ischemia-induced pathologic angiogenesis, mice lacking EC FGFR1/2 display a significant reduction in neovascular growth and tissue repair. It is well appreciated that serious vascular alterations associated with many antiangiogenic therapies limit their optimal benefits, including those that target the FGF pathway (8, 37, 40–42). Many of these adverse, unintended, or “off-target” effects on normal vascular physiology include increased permeability, hypertension, thrombosis, proteinuria, and ocular toxicity (retinal detachment and retina vein occlusion). Recent studies suggest that FGF signaling is a key positive regulator of basal EC barrier functions (43, 44) but did not determine whether FGF signaling is required in ECs vs. another vascular cell type. Additionally, it cannot be ruled out that the adenoviral-based soluble FGFR trap might have affected non-FGF pathways. To this end, we created a genetic mouse model in which FGFR1/2 were specifically deleted in ECs. In contrast to the results published by Murakami et al. (43) and Murakami and Sakurai (44), mice lacking EC FGFR1/2 actually maintained normal basal barrier functions in the skin and retinal vascular beds, even on a sensitized (Vegfr2 haploinsufficient) Oladipupo et al.
Materials and Methods Animals. Detailed information about animal generation is provided in SI Materials and Methods. The Washington University in St. Louis Animal Studies Committee approved all animal care and experimental procedures.
PNAS | September 16, 2014 | vol. 111 | no. 37 | 13383
DEVELOPMENTAL BIOLOGY
Fig. 5. Endothelial FGFR1/2 is required for neovascular response in eye injury-induced pathologic angiogenesis. (A) Representative FITC-dextran staining and (B and C) quantification shows a marked suppression of CNV in DCKO mice compared with control Flk1-Cre or Tie2-Cre and DFF mice (n = 8– 10). *P < 0.05. (Scale bars: A, 100 μm.)
genetic background. Explanations for these differences include effects of acute (43, 44) vs. chronic (this study) pathway inhibition, and the targeting of multiple vascular cell types (43, 44) vs. only EC and hematopoietic cells (this study). Additionally, Fgfr1/2 gene inactivation could be incomplete or could be compensated for by increased expression of FGFR3 or FGFR4, which could explain in part the lack of a developmental phenotype. Quantitative RT-PCR demonstrated expression of Fgfr3 (but not Fgfr4) in lung ECs but did not detect significant changes in the expression of these genes in ECs isolated from DFF and DCKO adult lung. Notably, and in agreement with previous findings (43, 44), we observed a modest, but statistically significant, increase in vascular permeability in response to an acute inflammatory stimulus, suggesting a nonhomeostatic role for EC FGF signaling in the maintenance of vascular barrier function. However, other vascular functions, including BP and vascular reactivity, were not dependent on EC and hematopoietic cell FGFR1/2 signaling. The contribution of FGF signaling in response to tissue injury is well established; however, most studies have focused on the role of EC and stromal FGF signaling in tissue repair (31, 45–47). Whether the vascular response to injury in vivo requires a direct FGF signal remains poorly understood. Using different mouse models of cutaneous and ocular injury we showed that EC FGFR1/2 signaling is indeed required for neovascular growth in tissue injury and repair. These observations are consistent with studies by Murakami et al. (15), in which they showed that induction of a dominant negative FGFR1 in ECs resulted in an impaired arteriogenic response and an associated reduction in Vegfr2 expression in a hind-limb ischemia model. In contrast, we find that the impaired angiogenic response to wounding is directly dependent on endothelial FGFR1/2. Recent studies by Zhao et al. (22) have shown that FGFR1 signaling is required for hematopoietic stem and progenitor cell (HSPC) engraftment following secondary adoptive transfer, but not for developmental or homeostatic hematopoiesis. Furthermore, Zhao et al. (22) showed that HSPC mobilization was impaired in mice lacking FGFR1 in the hematopoietic lineage suggesting impaired migration. The Flk1-Cre and Tie2-Cre deleter alleles used in the current study, which are known to target the hematopoietic lineage (19, 48), allowed us to test the requirement for hematopoietic cell FGFR1/2 signaling during development, homeostasis, and response to injury. Consistent with Zhao et al. (22), we found that inactivation of FGFR1/2 signaling in the hematopoietic lineage does not adversely affect hematopoietic development or homeostasis under normal physiological conditions. We also observed that leukocyte infiltration in DCKO and DFF mice was similar following cutaneous wounding. Although we did not examine specifically whether HSPCs display mobilization defects, we suggest that FGF signaling is not required for hematopoietic cell recruitment to injured tissue. In summary, we show that targeted inhibition of EC and hematopoietic cell FGFR1/2 signaling is compatible with normal embryonic development and normal vascular homeostasis and function. We also established an essential role for cell-autonomous EC FGFR signaling in injury response, tissue repair, and pathologic neovascular growth. This role for EC FGFR signaling in vivo highlights it as a viable therapeutic target for treating diseases associated with wound healing, tissue repair, and aberrant angiogenesis. The observation that EC and hematopoietic lineage are not affected by loss of FGFR1/2 signaling suggests that antiangiogenic therapy could be designed to target FGFR signaling in the ECs in tissues undergoing pathological angiogenesis without significant off-target effects on nondiseased tissue.
Whole-Mount Imaging of Embryo Vasculature. Time-mated embryos were fixed in 4% (wt/vol) paraformaldehyde overnight at 4 °C, then tissues were dehydrated in a methanol series, incubated in methanol/hydrogen peroxide, rehydrated, and blocked in PBST [5% (vol/vol) goat serum/PBS and 0.1% Triton X-100]. Additional information about whole-mount immunostaining and imaging can be found in SI Materials and Methods. Detailed methods on endothelial and hematopoietic cell sorting, Evans Blue assay, mean arterial pressure measurement, photopic visual acuity assessment, vascular reactivity assay, visualization of retinal vasculature, and wound-healing assays are provided in SI Materials and Methods. In addition, detailed information on tissue preparation, immunostaining, fluorescence microscopy methods, microvascular density quantification, immunoblotting, angiogenesis antibody array analysis, laser-induced CNV assay, oxygeninduced retinopathy, and quantification of vascular area with gene expression analysis can also be found in SI Materials and Methods.
considered to be statistically significantly. Data were analyzed by using the unpaired Student t test (GraphPad Prism 5). Numbers of mice used per group per experiment are stated in the figure legends.
Statistical Analysis. The data are reported as the mean ± SD and changes with P values less than 0.05 (*), 0.01 (**), 0.001 (***), or 0.0001 (****) were
ACKNOWLEDGMENTS. We thank B. Coleman and M. Scott for help with tissue processing and the Mouse Genetics Core for technical help. This work was supported by National Institutes of Health (NIH) Grants HL105732 (to D.M.O.), T32-HL07275 (to S.S.O.), HL63736, HL55337 (to K.C.), and EY019287 (to R.S.A.), as well as NIH Vision Core Grant P30EY02687 and a Carl Marshall Reeves and Mildred Almen Reeves Foundation Inc. Award (to R.S.A.). This work was also supported by a Research to Prevent Blindness Inc. Career Development Award (to R.S.A.), International Retina Research Foundation (R.S.A.), American Health Assistance Foundation (R.S.A.), Thome Foundation (R.S.A.), a Lacy Foundation Research Award (to A. Santeford), a Knights Templar Eye Foundation Grant (to L.A.W.), and a Research to Prevent Blindness Inc. Unrestricted Grant to Washington University in St. Louis. Transgenic mouse production was made possible through the Washington University Musculoskeletal Research Center (NIH Grant P30 AR057235) and the Digestive Disease Research Core Center (NIH Grant P30 DK052574).
1. Ferrara N, Kerbel RS (2005) Angiogenesis as a therapeutic target. Nature 438(7070): 967–974. 2. Chung AS, Ferrara N (2011) Developmental and pathological angiogenesis. Annu Rev Cell Dev Biol 27:563–584. 3. Murata K, et al. (2014) Inhibition of miR-92a enhances fracture healing via promoting angiogenesis in a model of stabilized fracture in young mice. J Bone Miner Res 29(2): 316–326. 4. Potente M, Gerhardt H, Carmeliet P (2011) Basic and therapeutic aspects of angiogenesis. Cell 146(6):873–887. 5. Presta M, et al. (2005) Fibroblast growth factor/fibroblast growth factor receptor system in angiogenesis. Cytokine Growth Factor Rev 16(2):159–178. 6. Beenken A, Mohammadi M (2009) The FGF family: Biology, pathophysiology and therapy. Nat Rev Drug Discov 8(3):235–253. 7. Itoh N, Ornitz DM (2011) Fibroblast growth factors: From molecular evolution to roles in development, metabolism and disease. J Biochem 149(2):121–130. 8. Lieu C, Heymach J, Overman M, Tran H, Kopetz S (2011) Beyond VEGF: Inhibition of the fibroblast growth factor pathway and antiangiogenesis. Clin Cancer Res 17(19): 6130–6139. 9. Turner N, Grose R (2010) Fibroblast growth factor signalling: From development to cancer. Nat Rev Cancer 10(2):116–129. 10. Seghezzi G, et al. (1998) Fibroblast growth factor-2 (FGF-2) induces vascular endothelial growth factor (VEGF) expression in the endothelial cells of forming capillaries: An autocrine mechanism contributing to angiogenesis. J Cell Biol 141(7):1659–1673. 11. Gospodarowicz D, Brown KD, Birdwell CR, Zetter BR (1978) Control of proliferation of human vascular endothelial cells. Characterization of the response of human umbilical vein endothelial cells to fibroblast growth factor, epidermal growth factor, and thrombin. J Cell Biol 77(3):774–788. 12. Claus P, Grothe C (2001) Molecular cloning and developmental expression of rat fibroblast growth factor receptor 3. Histochem Cell Biol 115(2):147–155. 13. Shin JW, et al. (2006) Prox1 promotes lineage-specific expression of fibroblast growth factor (FGF) receptor-3 in lymphatic endothelium: A role for FGF signaling in lymphangiogenesis. Mol Biol Cell 17(2):576–584. 14. Murakami M, Elfenbein A, Simons M (2008) Non-canonical fibroblast growth factor signalling in angiogenesis. Cardiovasc Res 78(2):223–231. 15. Murakami M, et al. (2011) FGF-dependent regulation of VEGF receptor 2 expression in mice. J Clin Invest 121(7):2668–2678. 16. Hatanaka K, Lanahan AA, Murakami M, Simons M (2012) Fibroblast growth factor signaling potentiates VE-cadherin stability at adherens junctions by regulating SHP2. PLoS ONE 7(5):e37600. 17. Nissen LJ, et al. (2007) Angiogenic factors FGF2 and PDGF-BB synergistically promote murine tumor neovascularization and metastasis. J Clin Invest 117(10):2766–2777. 18. Masaki I, et al. (2002) Angiogenic gene therapy for experimental critical limb ischemia: Acceleration of limb loss by overexpression of vascular endothelial growth factor 165 but not of fibroblast growth factor-2. Circ Res 90(9):966–973. 19. Motoike T, Markham DW, Rossant J, Sato TN (2003) Evidence for novel fate of Flk1+ progenitor: Contribution to muscle lineage. Genesis 35(3):153–159. 20. Kisanuki YY, et al. (2001) Tie2-Cre transgenic mice: A new model for endothelial celllineage analysis in vivo. Dev Biol 230(2):230–242. 21. Muzumdar MD, Tasic B, Miyamichi K, Li L, Luo L (2007) A global double-fluorescent Cre reporter mouse. Genesis 45(9):593–605. 22. Zhao M, et al. (2012) FGF signaling facilitates postinjury recovery of mouse hematopoietic system. Blood 120(9):1831–1842. 23. Park C, et al. (2006) Bone morphogenetic protein receptor 1A signaling is dispensable for hematopoietic development but essential for vessel and atrioventricular endocardial cushion formation. Development 133(17):3473–3484. 24. Apte RS, Richter J, Herndon J, Ferguson TA (2006) Macrophages inhibit neovascularization in a murine model of age-related macular degeneration. PLoS Med 3(8):e310. 25. Miles AA, Miles EM (1952) Vascular reactions to histamine, histamine-liberator and leukotaxine in the skin of guinea-pigs. J Physiol 118(2):228–257.
26. Feihl F, Liaudet L, Waeber B, Levy BI (2006) Hypertension: A disease of the microcirculation? Hypertension 48(6):1012–1017. 27. Mourad JJ, Levy BI (2011) Mechanisms of antiangiogenic-induced arterial hypertension. Curr Hypertens Rep 13(4):289–293. 28. Kolesnikov AV, Tang PH, Parker RO, Crouch RK, Kefalov VJ (2011) The mammalian cone visual cycle promotes rapid M/L-cone pigment regeneration independently of the interphotoreceptor retinoid-binding protein. J Neurosci 31(21):7900–7909. 29. Virag JA, et al. (2007) Fibroblast growth factor-2 regulates myocardial infarct repair: Effects on cell proliferation, scar contraction, and ventricular function. Am J Pathol 171(5):1431–1440. 30. House SL, et al. (2003) Cardiac-specific overexpression of fibroblast growth factor-2 protects against myocardial dysfunction and infarction in a murine model of low-flow ischemia. Circulation 108(25):3140–3148. 31. Braun S, auf dem Keller U, Steiling H, Werner S (2004) Fibroblast growth factors in epithelial repair and cytoprotection. Philos Trans R Soc Lond B Biol Sci 359(1445): 753–757. 32. Broadley KN, et al. (1989) Monospecific antibodies implicate basic fibroblast growth factor in normal wound repair. Lab Invest 61(5):571–575. 33. Sene A, et al. (2013) Impaired cholesterol efflux in senescent macrophages promotes age-related macular degeneration. Cell Metab 17(4):549–561. 34. Kelly J, Ali Khan A, Yin J, Ferguson TA, Apte RS (2007) Senescence regulates macrophage activation and angiogenic fate at sites of tissue injury in mice. J Clin Invest 117(11):3421–3426. 35. Smith LE, et al. (1994) Oxygen-induced retinopathy in the mouse. Invest Ophthalmol Vis Sci 35(1):101–111. 36. Brown AP, Courtney CL, King LM, Groom SC, Graziano MJ (2005) Cartilage dysplasia and tissue mineralization in the rat following administration of a FGF receptor tyrosine kinase inhibitor. Toxicol Pathol 33(4):449–455. 37. van der Noll R, Leijen S, Neuteboom GH, Beijnen JH, Schellens JH (2013) Effect of inhibition of the FGFR-MAPK signaling pathway on the development of ocular toxicities. Cancer Treat Rev 39(6):664–672. 38. Sun HD, et al. (2007) Monoclonal antibody antagonists of hypothalamic FGFR1 cause potent but reversible hypophagia and weight loss in rodents and monkeys. Am J Physiol Endocrinol Metab 292(3):E964–E976. 39. Guagnano V, et al. (2011) Discovery of 3-(2,6-dichloro-3,5-dimethoxy-phenyl)-1-6[4-(4-ethyl-piperazin-1-yl)-phenylamino]-pyrimidin-4-yl-1-methyl-urea (NVP-BGJ398), a potent and selective inhibitor of the fibroblast growth factor receptor family of receptor tyrosine kinase. J Med Chem 54(20):7066–7083. 40. Liang G, Chen G, Wei X, Zhao Y, Li X (2013) Small molecule inhibition of fibroblast growth factor receptors in cancer. Cytokine Growth Factor Rev 24(5):467–475. 41. Armstrong TS, Wen PY, Gilbert MR, Schiff D (2012) Management of treatmentassociated toxicites of anti-angiogenic therapy in patients with brain tumors. Neurooncol 14(10):1203–1214. 42. Dieci MV, Arnedos M, Andre F, Soria JC (2013) Fibroblast growth factor receptor inhibitors as a cancer treatment: From a biologic rationale to medical perspectives. Cancer Discov 3(3):264–279. 43. Murakami M, et al. (2008) The FGF system has a key role in regulating vascular integrity. J Clin Invest 118(10):3355–3366. 44. Murakami M, Sakurai T (2012) Role of fibroblast growth factor signaling in vascular formation and maintenance: Orchestrating signaling networks as an integrated system. Wiley Interdiscip Rev Syst Biol Med 4(6):615–629. 45. Barrientos S, Stojadinovic O, Golinko MS, Brem H, Tomic-Canic M (2008) Growth factors and cytokines in wound healing. Wound Repair Regen 16(5):585–601. 46. Meyer M, et al. (2012) FGF receptors 1 and 2 are key regulators of keratinocyte migration in vitro and in wounded skin. J Cell Sci 125(Pt 23):5690–5701. 47. Gurtner GC, Werner S, Barrandon Y, Longaker MT (2008) Wound repair and regeneration. Nature 453(7193):314–321. 48. Lugus JJ, Park C, Ma YD, Choi K (2009) Both primitive and definitive blood cells are derived from Flk-1+ mesoderm. Blood 113(3):563–566.
13384 | www.pnas.org/cgi/doi/10.1073/pnas.1324235111
Oladipupo et al.
