Capillary Morphogenesis gene 2 mediates multiple pathways of growth factor-induced angiogenesis by regulating endothelial cell chemotaxis

Abstract

Pathological angiogenesis contributes to diseases as varied as cancer and corneal neovascularization. The vascular endothelial growth factor (VEGF) - VEGF receptor 2 (KDR/VEGFR2) axis has been the major target for treating pathological angiogenesis. However, VEGF-targeted therapies exhibit modest efficacy over time indicating that new therapeutic strategies are needed. Therefore, identifying new targets that mediate angiogenesis is of great importance. Here, we report that one of the anthrax toxin receptors, capillary morphogenesis gene 2 (ANTXR2/CMG2), plays an important role in mediating angiogenesis. Inhibiting physiological ligand binding to CMG2 results in significant reduction of corneal neovascularization, endothelial tube formation and cell migration. We also report the novel finding that CMG2 mediates angiogenesis by regulating the direction of endothelial chemotactic migration without affecting overall cell motility.

Introduction

Pathological angiogenesis is closely related to many diseases including cancers and various ocular diseases that lead to blindness. Most of the current treatment strategies are targeting the VEGF - VEGFR2 axis. However, such therapies only provide modest efficacy overtime and accompany with unpleasant side effects. Thus, identify new targets are needed and is critical for alternative therapeutic strategies development. It was previously demonstrated that one of the anthrax toxin receptors, CMG2, has a role in mediating angiogenesis^{1, 2}. Our previous work demonstrated that inhibition of CMG2 potently reduces angiogenesis in the cornea¹ and inhibit endothelial cell migration^{3, 4}. Thus, we are interested to investigate the role(s) of CMG2 in mediating angiogenesis.

While the knowledge of CMG2’s role in angiogenesis remain limited, its function as one of anthrax toxin receptors, is very well established. CMG2 and its homolog, tumor endothelial marker 8 (ANTXR1 / TEM8), were consider the primary sites of anthrax toxins entry. Both CMG2 and TEM8 share 40% amino acid homology including sharing homology in an intracellular domain of unknown function that is shared with no other protein in the mammalian genome⁵. Within their von Willebrand Factor A (VWA) ligand binding domains, CMG2-TEM8 homology rises to 60%⁶. During anthrax intoxication, the 83kDa protective antigen (PA), one of the three subunits from Bacillus anthracis, binds to the VWA ligand binding domain of the receptor. It is then cleaved by a furin-like protease that cut and releases a 20kDa fragment, leaving a 63kDa PA at the receptor surface. The cleaved PA oligormerize to form and PA-receptor heptamer, which acts as a binding platform for the other two toxin subunits, lethal factor (LF) and edema factor (EF), and traffic the toxins into the cell via endocytosis⁷.

The endogenous functions of the anthrax toxin receptors are still poorly understood, although, it was suggest that these receptors interacts with extra-cellular matrix (ECM) proteins^{2, 8}. Studies observed that mutations on CMG2 or TEM8 would lead to build up of hyaline materials that result in implication of skeletal growth and alternation of vascular patterns. Lost of function mutations in TEM8 would cause GAPO syndrome, a disease characterized by vascular anomalies, skeletal defects, growth retardation and progressive fibrosis of various organs^{9, 10}. Mutations in CMG2 in humans leads to Hyaline Fibromatosis Syndrome (HFS), a rare but serious autosomal recessive disorder, characterized by accumulation of hyaline material in connective tissue in the skin and other organs and by the presence of non-cancerous nodules^11–13 containing excess collagen I, collagen VI, and glycosaminoglycans. To explain the symptoms, it has been hypothesized that patients with a CMG2 mutation have a defect in either the synthesis or degradation of ECM^{12, 14}, presumably related to CMG2 dysfunction.

In addition to ECM binding, it is suggested that these receptors have angiogenic related activity^{1, 15}. Our previous work demonstrated that a furin-cleavage-site mutant of protective antigen (PA^SSSR) inhibits basic fibroblast growth factor (bFGF) induced corneal neovascularization, VEGF-induced corneal neovascularization, and tumor growth^{1, 16}. Because PA^SSSR is known to bind both anthrax toxin receptors (CMG2 and TEM8) as well as integrin β1^{17, 18}, it has remained unclear which of these receptors was responsible for mediating the observed anti-angiogenic effect of PA^SSSR in vivo. While CMG2 has a much tighter affinity for PA than the other two receptors^18–20, it is considered to be the major receptor for PA. Thus, we hypothesized that CMG2 is the relevant receptor that are responsible to PA’s anti-angiogenic affect. Additionally, our previous data showed that CMG2 inhibition reduces angiogenesis in both in-vivo and ex-vivo assays, however, the biology behind such observations remain unclear. Therefore, the focuses of this paper are to understand the anti-angiogenic relationship between CMG2 and PA^SSSR, and further investigate the role(s) of CMG2 in mediating angiogenesis in endothelial cells.

Result and Discussion

As mentioned previously, PA^SSSR interacts both anthrax toxin receptors and integrin β1. While CMG2 and tumor endothelial marker 8 (TEM8) bind PA^SSSR much more tightly than does integrin β1 (Kd for CMG2, 200pM; TEM8, 100nM^{18, 21}; integrin β1, 1μM), suggesting that CMG2 and TEM8 are more likely to mediate the antiangiogenic effects of PA^SSSR. We performed a series of experiments to compare the relative contributions of CMG2 and TEM8 to corneal neovascularization in mice, and to determine whether blockade of either of these receptors was sufficient to explain the observed effect. First, we used either a CMG2 or TEM8 extracellular domain fused to an antibody Fc-domain to disrupt the ligand-receptor interaction by competing for endogenous ligand binding. In the corneal micropocket assay, administration of CMG2-Fc significantly inhibited bFGF-induced vessel growth when compared to the untreated control, but the TEM8-Fc fusion did not (Fig. 1A). We confirmed this observation by administering antibodies specific to either the CMG2 or TEM8 extracellular domains. The anti-CMG2 antibody significantly reduced bFGF-induced corneal neovascularization in a concentration dependent manner (Fig. 1B). In contrast, treatment with the anti-TEM8 antibody L2 at, a dose and schedule previously shown to inhibit tumor growth in mice (20 mg/kg/week) ²² resulted in no significant decrease in corneal neovascularization when compared to the vehicle control (Fig. 1C). Together, these results show that inhibiting the interaction of CMG2 with its physiological ligand significantly inhibits corneal angiogenesis, while inhibition of TEM8-ligand interaction does not.

Figure 1. Blocking the interactions of CMG2 with its natural ligand inhibits both FGF and VEGF induced corneal neovascularization in vivo.

(A) Corneal micropocket assay on mice treated with soluble ECD of either CMG2 or TEM8, via intraperitoneal injection. (B-C) Corneal micro-pocket assay on mice treated with either anti-CMG2 antibody (B) or anti-TEM8 L2 anti-body (C). Corneal neovascularization in these assays was induced by bFGF before treatment; vessel area on both the left and right corneas were measured from each mouse. (D) Representative image of corneal neovascularization in both wild type (WT) and CMG2 −/− male mice. (E-F) Comparison of corneal neovascularization between WT and CMG2−/− mice induced by either bFGF (E) or VEGF (F). CMG2−/− mice showed significant reductions in neovascularization for both bFGF- and VEGF-induced angiogenesis. (G-H) A similar experiment was performed with WT and TEM8−/− mice with either bFGF (G) or VEGF (H). Data presented are pooled from both genders. Error bars are standard error of mean, * p<0.05; ** p<0.01; *** p<0.001.

To confirm the importance of CMG2 rather than TEM8 as a mediator of angiogenesis in the cornea, we next sought to identify phenotypic changes that result from CMG2 or TEM8 knockout. We performed the corneal micropocket assay, using either bFGF or VEGF to induce vessel growth in CMG2 and TEM8 knockout mice. CMG2 knockout (CMG2−/−) mice exhibit a striking reduction in both bFGF (Fig 1E) and VEGF-induced (Fig 1F) corneal vascularization compared to WT mice, confirming the importance of CMG2 in corneal angiogenesis. Interestingly, female mice were particularly susceptible to this effect (Fig S1A-B), exhibiting a >85% reduction in their response to VEGF (Fig S1B). In contrast, TEM8 knockout (TEM8−/−) mice exhibited no significant reduction when stimulated with VEGF (Fig 1H) and only modest reductions in bFGF-induced neovascularization (15%; p<0.05), (Fig 1G, S1C-D). These data confirm that CMG2 plays a quantitatively substantial role in corneal neovascularization.

Supplementary Figure S1. Gender-specific quantification of bFGF- and VEGF-induced corneal vascularization in WT, CMG2 KO, and TEM8 KO mice.

(A-B) Quantification of (A) bFGF-induced and (B) VEGF-induced neovascularization in male and female CMG2−/− mice as compared to WT. (C-D) Quantification of (C) bFGF-induced and (D) VEGF-induced neovascularization in male and female TEM8−/− mice as compared to WT. For both genders, greater decreases in corneal vascularization were observed in CMG2 KO than for TEM8 KO.

It remained possible that both anthrax toxin receptors might still be responsible for mediating the anti-angiogenic effect of PA^SSSR. To assess this possibility, we evaluated the efficacy of PA^SSSR in reducing corneal vascularization in CMG2−/− or TEM8−/− mice. Importantly, CMG2−/− mice did not show reduction in bFGF-induced corneal angiogenesis (Fig2A). In contrast, PA^SSSR treatment significantly reduced corneal neovascularization in both wild-type control animals and TEM8−/− mice (Fig 2B). We conclude that CMG2 blockade is responsible for the anti-angiogenic effects of PA^SSSR.

Figure 2. CMG2 is the main mediating receptor of PA^SSSR-induced angiogenic inhibition.

(A) Comparative levels of bFGF-induced corneal neovascularization between CMG2−/− and WT mice treated with or without PA^SSSR, and (B) TEM8−/− and WT mice treated with or without PA^SSSR. Results showed that PA^SSSR reduced vessel formation on TEM8−/− mice but not CMG2−/− mice. . Error bars are standard error of mean. (C) Western blot of HMVEC lysates with either CMG2 knockdown (CMG2si) or TEM8 knockdown (TEM8si). Band intensity was normalized to control (untreated cells). (D) EGF- and VEGF-induced tubule formation assays with both CMG2si- and TEM8si-HMVEC, treated with or without different concentrations of PA^SSSR. (E) Quantification of tube formation assays by counting the number of networks on each field. Error bars are standard deviation.

To establish whether the effect of CMG2 knockout or blockade is endothelial cell intrinsic, we performed tube formation assays in Matrigel, using HMVEC cells, which express both CMG2 and TEM8. We note that due to culture limitations, CMG2 ablation in HMVEC using CRISPR is not possible; however, knockdown using siRNA was successful toknockdown (KD) CMG2 or TEM8 (Fig 2C) and compared the anti-angiogenic effects of PA^SSSR on cells expressing only one of the two receptors. In TEM8 KD cells (which primarily express CMG2), PA^SSSR administration resulted in a concentration-dependent reduction in the extent of tube network formation (Fig 2D-E). In contrast, tube formation in CMG2 KD cells was not altered by PA treatment (Fig 2D-E). Together with knockout data described above, these results demonstrate that PA^SSSR retains its anti-angiogenic effect in both TEM8−/− mice and TEM8 KD HMVECs but not in CMG2−/− mice or CMG2 KD cells. Hence, we conclude that PA^SSSR exerts its anti-angiogenic effects via CMG2 on endothelial cells.

We next worked to determine the CMG2-mediated endothelial process that PA^SSSR disrupts to decrease angiogenesis. During angiogenesis, endothelial tip cells receive migratory signals and orient themselves to migrate up a chemotactic gradient, while the stalk cells follow and proliferate to form new vessel(s)²³ Thus, CMG2 could regulate angiogenesis by impacting cell proliferation and/or cell migration. We have previously demonstrated that PA^SSSR affects cell migration, but not proliferation in HMVEC¹, a primary endothelial cell type. Consistent with this result, knockdown of neither CMG2 nor TEM8 knockdown significantly reduce HMVEC proliferation (Fig S2A-D). We also observe no reduction in cell proliferation in EA.hy926 endothelial cells (a fusion of human umbilical vein cells with lung carcinoma cells) treated with PA^SSSR under conditions that selectively target CMG2 (Fig S3A). Together, these data suggest that CMG2 inhibition affect angiogenesis by modulating endothelial cell migration, rather than cell proliferation.

Supplementary Figure S2. Effects of CMG2 and TEM8 siRNA-mediated knockdown on Ea.hy926 cell proliferation and migration.

(A-B) Western blot analysis of Ea.hy926 lysates for (A) CMG2 and (B) TEM8 before and after introduction of siRNA. (C-D) Proliferation of both WT and siRNA-treated Ea.hy926 cells with both (C) CMG2-specific and (D) TEM8-specific siRNA. Proliferation was not significantly altered from WT in either CMG2-targeted or TEM8-targeted cells. (E) Migration of WT, CMG2 siRNA and TEM8 siRNA-treated Ea.hy926 cells. No significant difference in migration was observed from WT. (F) Western blot analysis of Ea.hy926 lysates for CMG2 expression 3 days and 6 days after addition of siRNA.

Supplementary Figure S3.

Binding of CMG2 to extracellular matrix proteins, as determined by ELISA. Proteins were coated on wells, after which CMG2-GST-biotin was added and signal read out with streptavidin-HRP/TMB. KD values were calculated by fitting each dataset to a 4-parameter logistic curve. A, positive control (PA); B, fibronectin; C, collagen I; D, collagen VI. PA x axis scale is logarithmic with pM concentration units; all others are logarithmic with nM concentration values. Y axis is normalized absorbance at 450 nm. Collagen IV and laminin-111 were also assayed for binding, but curves are not displayed due to poor fit. In each case, similar binding affinites (500 nM-1 uM) were observed (see Supplementary Table S1).

Our data support the idea that PA^SSSR exerts its anti-angiogenic effects by competing with interaction(s) between the CMG2 von Willebrand Factor A (vWA)^{24, 25} domain (including the metal ion-dependent adhesion site (MIDAS)^{5, 26} and endogenous ligand(s). However, the specific ligands that regulate angiogenesis by binding CMG2 remain unidentified. ECM proteins are likely candidates because of CMG2’s homology with integrins²⁵, modest available data showing CMG2 interaction with ECM proteins², and the observation that in individuals with hyaline fibromatosis syndrome (HFS), CMG2 mutations in the VWA domain result in widespread accumulation and of extracellular matrix (ECM) proteins⁸. To verify interaction of CMG2 with ECM proteins, we used ELISA to measure CMG2 K_d for a series of different ECM proteins (Table S1; Fig S3 A-D). Interestingly, each of the ECM proteins tested (Collagens I, VI, laminin and fibronectin) interacted with CMG2 with near-indistinguishable near-micromolar K_d values (Table S1). Similar K_d values for different ECM proteins in these assays did not reflect artefacts in the binding assay, since both positive and negative control assays (CMG2 + PA^SSSR and CMG2 + PA^SSSR in the presence of EDTA, respectively) mirrored previously published K_d values¹⁸. Previously published data on CMG2–matrix interactions, which⁸ have indicated that CMG2 may bind preferentially to collagens IV² or VI⁸, are not necessarily at odds with this result. Differences in protein preparation, matrix adhesion to wells, and number of binding sites per molecule are all factors which influence relative signal observed in blot- and ELISA-based assays. However, the similar affinity of CMG2 for the matrix proteins tested indicates that CMG2, like integrins, may play a significant role in cell adhesion and motility through direct interactions with the extracellular matrix.

Supplementary Table S1. CMG2 binds several broadly expressed matrix proteins with high affinity.

Results of an ELISA-based quantification of CMG2 binding to several broadly-expressed matrix proteins, including collagens, laminin-111, and fibronectin. For positive control, PA was assayed for binding to CMG2. Such similar affinities between matrix proteins indicate that CMG2 shows no preference for binding to any one of the matrix proteins assayed.

To provide evidence for the assertion that PA^SSSR inhibits interaction between ECM and CMG2, we performed ex vivo cell adhesion assays with EA.hy926 cells on plates coated with different ECMs (Collagens I, IV, and VI, Human Fibronectin, and Laminin-111). Treatment with PA^SSSR at concentrations identical to the CMG2 K_d value (200pM) significantly reduced EA.hy926 cell adhesion on plates coated with each ECM, but not on uncoated plates (Fig S4B). We note that at this concentration CMG2, but not TEM8, is blocked by PA^SSSR. Together, these data suggest that CMG2 plays a role in mediating cell adhesion to ECM proteins²⁵, and that PA^SSSR binding to CMG2 inhibits that interaction.

Supplementary Figure S4. Matrix binding to CMG2 is an important component of cell adhesion and migration.

(A) EA.hy926 endothelial cell proliferation in different concentrations of PA^SSSR. Ethanol (EtOH) treated cells was used as the negative control. (B) EA.hy926 endothelial cell adhesion to plates coated with various matrix proteins, or no ECM proteins (“NT”), both with and without 200 pM PA. Addition of PA significantly inhibits adhesion to matrix-coated plates. Adhesion to uncoated plates was not affected by PA^SSSR treatment. (C-D) EA.hy926 migration on different ECM coated surface and compare to the PA^SSSR treatment. Chemotaxis are the vertical displacement measured towards a serum gradient form across the migration chamber (C). While chemokinesis are measured by total distance cell migrated in random direction (D). statistic significance between untreated and PA treated cell was calculated by student T.Test. Error bars are standard deviation of mean, * p<0.05; ** p<0.01; n.s not significant

In angiogenesis a major function of cell adhesion is to enable cell migration. We hypothesize that CMG2 may play a role in enabling cell movement. In that case, PA^SSSR would inhibit cell migration as well as adhesion. Indeed, we find that PA^SSSR significantly reduces cell migration on multiple different ECM substrates (Fig S4C-D). Specifically, we used fetal bovine serum as chemoattractant in a microfluidic perfusion device which contains a perfusion chamber that allows formation of a small molecule and macromolecule concentration gradient (data not shown). In this setup, we can observe migration of individual cells in real time. As expected, PA^SSSR treatment resulted in decreased migration over nearly all ECM surfaces (Fig S4C-D), even at the 200pM concentration expected to result in only 50% bound CMG2. Combining these adhesion and migration data, we conclude that ECM proteins mediate endothelial cell adhesion and migration through interaction with CMG2, and that PA^SSSR can disrupt this interaction (Fig S4B-D).

Notably, the nature of this migration assay allows us to compare two different aspects of migration: initiation of cell movement (chemokinesis) and directional motion towards growth factor-containing serum (chemotaxis). Surprisingly, targeting CMG2 in EA.hy926 cells with PA^SSSR at the CMG2 Kd (200pM) showed a 70% decrease in observed chemotaxis on non-coated surface (70%, P<0.01; Fig S4C. Given that we expect only 50% CMG2 occupancy at these PA^SSSR concentrations, this chemotactic inhibition is dramatic. In contrast, the total distance traveled by PA^SSSR treated cells was only modestly (∼20%) reduced, compared to untreated cells (Fig S4D). Thus, CMG2 targeting has a dramatic effect on chemotaxis but little effect on chemokinesis.

Futher support for involvement of CMG2 in chemotaxis rather than chemokinesis is provided by transwell migration experiments performed on CMG2 knockdown HMVEC cells. Like many other traditional migration assays, the transwell assay does not provide a stable chemoattractant gradient to allow chemotaxis measurement²⁷, and thus only measures chemokinesis. Evaluation of these CMG2 knockdown cells via transwell migration assays showed no significant difference between wild-type and CMG2 or TEM8 knockdown HMVEC (Fig S4E), consistent with our previous observation that CMG2 inhibition impacts chemotaxis but not chemokinesis.

We next examined the effect of CMG2 depletion on chemotactic migration in EA.hy926 cells, using CRISPR to target Exon1. Loss of CMG2 function in knockout cells was validated with a flow cytometry assay in which uptake of a fluorescent PA-AF568 conjugate was evaluated at CMG2 K_d 200pM (Fig 3A). CMG2 knockout cells showed substantial inhibition of intracellular fluorescence relative to WT cells, and knockout cell fluorescence was indistinguishable from that of negative control WT cells (Fig 3A). The ability of transfected CMG2-mClover to rescue the PA uptake defect was confirmed, demonstrating that off-target effects cannot explain loss of PA uptake in these cells (Fig 3A). Importantly, evaluation of the migration of these EA.hy926 knockout cells in our microfluidic device demonstrated that while WT cells migrated toward serum (Fig 3B), CMG2−/− cells migrated only in a random manner (Fig 3C). Significantly, knockout cells showed complete abolition of chemotactic migration (Fig 3E) but only slight alteration of chemokinetic migration (P<0.001) (Fig 3F). To confirm that the loss of chemotaxis was not a consequence of off target CRISPR effects, we complemented the CMG2 mutation with a CMG2-clover expression vector and found that CMG2 add-back restored EA.hy926 cell chemotaxis (Fig 3D-E). Hence, CMG2 is a critical component of chemotactic sensing or response in these cells, and may play a role in regulating endothelial cell chemotaxis in vivo. Of interest, CMG2 is required for reorientation of the mitotic spindle in zebrafish embryos²⁸ suggesting that it might also play a role in orienting the microtubule cytoskeleton in a chemotactic context.

Figure 3. CMG2 KO endothelial cells lose the ability to migrate toward chemoattractants.

(A) Differential PA-AF586 conjugate uptake by wild type (WT), CMG2 add back (AB) CMG2 knockout (CMG2 −/−) EA.hy926 cells via flow cytometry (10,000 cells per condition). Median fluorescence intensiy of each condition was normalized against WT signal after subtracting from the unstain control. Error bars are normalized standard deviation from three individual replicates.. (B) Aggregated track plots of individual EA.hy926 cell migration in the CellASIC migration chamber. Both WT cells (B), CMG2−/− cells (C), and CMG2 AB cells were subjected to an FBS gradient. (E-F) Quantification of both Ea.hy926 chemotaxis (displacement toward gradient, E) and Chemokenesis (total migration, F) in the CellASIC migration assay. * p<0.05; ** p<0.01; *** p<0.001; n.s not significant

We next investigated the effect of CMG2 deletion on developmental angiogenesis. To do so, we compared retinal vessel development between WT and CMG2−/− mouse. We found that the total vascular area was similar between the two genotypes (Fig. 4). However the vascular growth pattern observed was different between knockout and WT mice, with knockout mice having a more bush-like pattern versus the tree-like pattern observed in WT animals (Fig. 4A). Careful quantitation showed that the total number of veins in the retina of WT and CMG2−/− mice is similar (Fig 4A-B), but CMG2−/− mice have slightly fewer primary arteries (Fig 4A, C). Importantly, CMG2−/− mice showed 2.5-fold more primary branches per artery than the WT cells (Fig 4A, D). This difference in vessel patterning could be consistent with the loss of chemotactic migration observed in knockout cells. We speculate that in the absence of CMG2, endothelial cells effectively lose chemotactic signals (though haptotactic and other signals may be retained), reducing effective directional movement. This hypothesized weaker response in CMG2−/− mice to the developmental angiogenic gradient could leave some areas of the retina with poor perfusion. Angiogenic signals from these areas could then lead to additional branching to alleviate hypoxia, resulting in a vasculature that remains functional, but is patterned differently.

Figure 4. CMG2 KO increases vessel branching in the mouse retina.

(A) Representative mages of vessel formation in the retina of both WT (left) and CMG2−/− (right) mice. (B-D) Quantified vessel formation from retinal assays. (B) Comparison of vascular radius, normalized to the retinal radius, between WT and CMG2−/− mice. (C) Quantification of arterial branching in both WT and CMG2−/− mouse retinas. (D) Artery and vein counts per retina as in WT and CMG2−/− mice. CMG2−/− mouse retinas exhibit fewer veins and arteries than WT, but only artery count is significantly lower than WT mice.

CMG2 mediates chemotaxis that is induced by multiple different growth factors, indicating that the process it regulates is core to that process. We have shown here that CMG2-deficient mice showed potent reduction in corneal neovascularization induced by both bFG and VEGF. In addition, Vink et al. has reported that CMG2 mediates migration of smooth muscle cells towards PDGF²⁹. Consistent with this observation, Reeves et al. have suggested that CMG2 works closely with MT1-MMP, an important cofactor in PDGF signaling pathway³⁰. Therefore, we have decided to measure endothelial cell migration under a gradient of individual growth factor. Interestingly, EA.hy926 cells display a chemotactic migration towards bFGF and PDGF, and can be significantly inhibited with 200pM of PA^SSSR treatment (Fig S5A). While cells in insulin treatment didn’t show much chemotactic migration (Fig S5A). We also measure migration under the gradient of boiled FBS (Growth factors are denatured by boiling and removed by centrifugation). Result show that EA.hy926 cells remain a high chemotaxis phenotype, but it is not affected by PA treatment (Fig S5C). Neither the chemokinesis in all these conditions were not affected by PA^SSSR, further suggest that CMG2 plays a critical role in mediating chemotaxis towards some molecules that can be denatured at high temperature, like bFGF, VEGF, and PDGF. Together, these data suggest that CMG2 may act as a central regulator for chemotaxis that is initiated by a series of different growth factors. The mechanism of CMG2’s impact on chemotaxis will be important for developing new efficacious therapies to treat pathological angiogenesis. It remains to be seen whether interaction of CMG2 with growth factors receptors at the cell surface can be detected.

Supplementary Figure S5. CMG2 directs cell chemotaxis towards various growth factors in serum.

(A-B) EA.hy926 endothelial cell migration under a gradient of various growth factors (20ng/mL bFGF and PDGF, 300ng/mL VEGF) and fetal bovine serum (FBS; 10%). Displacement on Y direction (direction towards chemoattractant gradient) was measured and displayed as chemotaxis (A). Total migration distance was measured and displayed as chemokinesis (B). (C-D) EA.hy926 endothelial cell migration under a gradient of boiled FBS (10%). FBS was boiled and centrifuged to remove denatured and precipitated proteins. Supernatant (10%) was used in the migration to measure chemotaxis (C) and chemokinesis (D). 200pM of PA^SSSR was used to treat the cells in each condition and compared with non-treated cells (Ctrl). 40 cells were tracked from each conditions and error bar is standard deviation of three individual replicates. * p<0.05; ** p<0.01; n.s not significant

Methods

Protein preparation

Cell culture

EA.hy926 (CRL-2922) cells are the result of a fusion of human umbilical vein cells with lung carcinoma cells. Cells were cultured in 10% FBS + DMEM and incubated at 37°C in a humidified environment with 5% CO₂ until ready for passaging.

EA.hy cell proliferation assay

15,000 EA.hy926 cells were seeded into each well in a 96 well plate and incubated for 1 h to attach. After cell attachment, media with treatments were added. Ethanol fixed cells were used as negative control. After a 24 h incubation, 20 μL CellTiter-Blue Reagent (Promega) was added to each well for 4 h. Fluorescence signal (Ex: 560nm / Em: 590nm) was measured using a BioTek Synergy H2 plate reader. All readings were normalized to the non-treated control.

Cell ASIC migration assay

The assay protocol followed the CellASIC ONIX M04G-02 Microfluidic Gradient Plate User Guide. All media put into the plate (excepting the cell suspension) was filtered through a 0.2 μm syringe filter. EA.hy cells (3 × 10⁶ cells/mL) were loaded in and incubated overnight with DMEM + 10% FBS under flow at 37°C. Assays were then performed with a stable gradient of DMEM + 0 – 10% FBS with or without peptide treatment. Brightfield images at 10x magnification were taken every 10 min over 12 h on an Olympus IX73 microscope and the ORCA-Flash4.0 camera (Hamamatsu). Individual cells were tracked with Image J manual tracking plugin. Data was transferred into the ibidi Chemotaxis and Migration Tools 2.0. to export endpoint y displacement and accumulated displacement. P-values were calculated using Student’s t-test and error bars are standard error of the mean (n=50).

CMG2 ECM ELISA

All steps were performed at room temperature unless otherwise indicated. For binding assays, matrix protein (Rockland Immunochemicals, Corning, EMD Millipore) was adsorbed onto 96-well polyethylene plates (Greiner) by incubating 2 ug/mL matrix protein in HBS with 2 mM Mg²⁺ and Ca²⁺ (Buffer A) in individual wells at 4° C overnight. Bacillus anthracis protective antigen (PA) was treated similarly but was incubated at 1 μM in the same buffer. After incubation, wells were washed 3x with Buffer A and blocked with 5% BSA (GoldBio) in HBS with 2 mM Mg²⁺, 2 mM Ca²⁺, and 0.1% Tween-20 (Buffer B) for 1 hour. Blocking was followed by 3 washes with Buffer B, after which varying concentrations of a biotinylated CMG2-AviTag construct (for matrix, 2138 to 10 μM; for PA, 4 μM to 1 pM) were dissolved in Buffer B and incubated in wells for 4 hours. After incubation with CMG2-AviTag, wells were again washed 3 times, after which streptavidin-HRP (Thermo Scientific) diluted in Buffer B with 5% BSA was incubated in wells for 1 hour. Wells were then washed 6 times with Buffer B, after which 1x TMB solution (Thermo Scientific) was added to wells. Once color was visible in wells, reaction was quenched with 0.2 M H₂SO₄. Wells were read out in a BioTek H4 Hybrid plate reader (BioTek) by quantifying well absorbance at 450 nm. Data were analyzed in Microsoft Excel, and Kd values were calculated in MATLAB using the sigm_fit function.

Adhesion assay

EA.hy 926 cells were grown in DMEM supplemented with 10% FBS. 20μg/mL collagen I, IV, VI, laminin, fibronectin, PA^SSSR, and BSA were preared in PBS and coated in 96well plate (100◻g/well) overnight at 4°C. After coating, collagen I, IV, VI, laminin, fibronectin, PA^SSSR, and BSA were removed and 5% BSA were coated for 1 hour. Cells were deprived with serum by scraped and cells were at 100g for 10 minutes and the supernatant was discarded. Cells then resuspended at 2-3×10⁵ cells/mL in DMEM. 100μL of cells were added to each of the coated wells. The plate was incubated at 37°C for 1 hour to allow the cells to adhere on the surface. Each well was washed gently for two times using warm serum-free DMEM by multi-channel pipettor. 100μL of serum free DMEM and 20μL of cell-titler blue (Promega Cat#G8080) were added into each well, and incubated for 4 hours at 37°C. After 4 hours incubation, fluorescent signal of 96 well plate was measured at 560_Ex/590_Em by a plate reader.

EAhy 926 CMG2 KO development

HEK 293T cells (3.8 × 10⁶ cells) were seeded in 10cm tissue culture dish for 18 hours. 12μg of pCMV, 5μg of pVSVG, 12.5μg of Lenti-CRISPR (targeting CMG2 exon 1, sgRNA sequence: GCACCAACAGCCACAGCCCG), 90μL of 1mg/mL PEI and 600μL of serum-free DMEM were mixed into a tube and incubated for 15 minutes before adding to the 10cm culture dish. The dish was replaced with 10mL of 10% FBS DMEM in 4 hours and followed by 48 hours incubation. Media was removed from the transfected cells into a conical tube, and was spin at 2500g for 3 minutes to remove debris. The supernatant (lentivirus) and 10μg/mL polybrene were added to 40% confluent EA.hy926 cells and incubated for 24 hours. The next day, media was removed, fresh 10% FBS DMEM and 1μg/mL puromycin were added to the plate for 3-5 days. After 3-5 days, the cells that survived in the selection were diluted into single colony in 96 well plate.

PA uptake flow cytometry

CMG2 KO EA.hy 926 cells and WT EA.hy926 cells were split in 12 well plate at 50% confluent. 200pM PA-Alexa Fluor 546 was added into each well and incubated overnight. Then cells were trypsinized and resuspended into a microcentrifuge tube. Cells were washed once with complete media. Then 50,000 cells of each condition were sent to Cytoflex (RIC facility at BYU). Data was collected and analyzed in our lab by using FlowJo software.

Corneal micropocket assay

The corneal micropocket assay was performed as described¹⁶ using pellets containing 80 ng of basic fibroblast growth factor (bFGF) or 180 ng of human carrier-free recombinant vascular endothelial growth factor (VEGF; R&D Systems) in C57BL/6J mice. The treated groups received daily or twice daily i.p injections for 5 (bFGF) or 6 (VEGF) consecutive days of protein in PBS. Treatment was started on the day of pellet implantation; control mice received only vehicle i.p. The area of vascular response was assessed on the 5th (bFGF) or 6th (VEGF) postoperative day using a slit lamp. Typically, at least 10 eyes per group were measured.

HMVEC proliferation assay

Human microvascular endothelial cells (Cambrex) were maintained in EGM-2 (Cambrex) according to the vendor’s instructions and used before passage 7. On day 0, proliferating cultures of BCE or HMVEC-d cells were seeded at ∼10% confluence into 96-well plates. After attachment, medium was exchanged for medium containing 1 pmol/L to 1 μmol/L of the indicated protein. Cells were allowed to grow for 7 days and then quantified using CyQUANT (Invitrogen) according to manufacturer’s directions. The degree of proliferation in culture was measured by comparing wells in each plate fixed in absolute ethanol on day 0 with experimental wells, with fold proliferation calculated by dividing CyQUANT fluorescence in experimental wells by that in day 0 wells. Groups were compared using Student’s t test, with Bonferroni correction where appropriate.

HMVEC Trans-well migration assay

Human microvascular endothelial cells were maintained as above. Polycarbonate Transwell inserts, 6.5 mm diameter with 8.0 μm pores, were coated with fibronectin (BD Biosciences). Cells were harvested and resuspended in EBM (Cambrex) containing 0.1% bovine serum albumin (Fisher Chemical). Cells (10,000 per well) were plated onto wells containing medium alone or medium containing the molecule to be tested. These wells were suspended above wells containing 5 to 10 ng/mL recombinant human VEGF (R&D Systems) or full serum medium. Cells were allowed to migrate for 4 h. Membranes were rinsed once in PBS and then fixed and processed using Diff-Quick (Dade Diagnostics). Cells on the top of the membrane were removed using cotton-tipped applicators, and the membrane was removed from the insert using a scalpel. Membranes were then mounted on slides, and the number of cells in a microscopic field was counted manually.

HMVEC CMG2 / TEM8 KD

HMVEC tube formation assay

Human microvascular endothelial cells were maintained as above. Before the assay, a 1- to 2-mm layer of Matrigel was plated into the wells of a 12-well cluster. Approximately 10⁵ cells were plated on this layer in EGM-2. Plates were examined at 12, 14, 16, 18, and 24 h for differences in network formation. In each experiment, good network formation was observed in untreated control wells.

CMG2 and TEM8 KO mice

Western Blotting

Supplementary note

1. bFGF was more used in corneal assays because we saw a better angiogenic effects

2. 200pM PA was targeting CMG2