Triamcinolone Acetonide Inhibits IL-6– and VEGF-Induced Angiogenesis Downstream of the IL-6 and VEGF Receptors

Abstract

purpose. To test whether triamcinolone acetonide (TA) inhibits angiogenesis induced by IL-6 or VEGF and whether this inhibition is through antagonism of the IL-6 or the VEGF receptor 2.

methods. A rat cornea micropocket assay was used to initiate IL-6– and VEGF-mediated angiogenesis. The ability of TA or neutralizing VEGF antibody to inhibit IL-6– or VEGF-mediated neovascularization was analyzed by measuring vessel length, vessel extension, and vessel area. The phosphorylation of signal transduction activator 3 (STAT3), VEGF receptor, and extracellular signal-regulated kinase 1/2 (ERK1/2) was determined by Western blot in human umbilical vein endothelial cell (HUVEC) lysates after stimulus with IL-6 or VEGF, with and without TA pretreatment. The effect of IL-6 or TA on STAT3 expression in cornea was determined by Western blot.

results. IL-6 induced corneal angiogenesis in a dose-dependent manner, with 350 ng producing a peak at day 6. VEGF antibodies and TA blocked IL-6-mediated limbal neovascularization. TA also directly inhibited angiogenesis stimulated by a VEGF pellet; the glucocorticoid receptor antagonist mifepristone neutralized TA inhibition of angiogenesis. TA did not inhibit IL-6-induced STAT3 phosphorylation and did not inhibit VEGF-induced phosphorylation of the VEGF receptor 2 or of ERK1/2 in endothelial cells, but TA decreased IL-6-induced STAT3 expression in cornea.

conclusions. IL-6– and VEGF-mediated corneal neovascularization are blocked by TA through the mifepristone-sensitive steroid receptor. TA inhibits IL-6-induced STAT3 expression in cornea, but it does not inhibit activation of the IL-6 or the VEGF receptor in cultured human endothelial cells. This finding has two implications. The fact that TA directly inhibits VEGF action implies that other factors may be critical to angiogenesis and sensitive to glucocorticoids.

Diabetic retinopathy and retinal vein occlusion have an inflammatory and immune response similar to that of ischemia/reperfusion syndromes in other tissues. Both retinovascular disorders, like stroke or neoplasm, induce the secretion of IL-6, a multifunctional glycoprotein with diverse signaling functions. Elevated levels of IL-6 have been reported in serum and vitreous of patients with diabetic macular edema, central retinal vein occlusion, and iris neovascularization. ^{1 2}  

With the realization that IL-6 secretion may underlie some instances of pathologic ocular vasopermeability and angiogenesis, we were interested in determining whether the beneficial effect of existing therapies, among them intravitreal glucocorticoids such as triamcinolone acetonide (TA), ameliorate conditions associated with hypersecretion of IL-6. Although anecdotal evidence indicates that glucocorticoids are useful in treating these conditions, ^{3 4 5 6 7 8 9 10 11 12} the side effects of steroids can be catastrophic and hence beg the question of whether a better understanding of how glucocorticoids act could produce novel therapies that lack the complications of steroid treatment. ¹³ In this article, we demonstrate that IL-6 promotes canonical angiogenesis in a cornea micropocket assay that is blocked by VEGF-specific antibodies and by TA. TA downregulates the target of the IL-6 receptor, signal transducer activator 3 (STAT3), in cornea. Curiously, TA also directly blocks the effect of VEGF on neovascularization. We have previously reported that TA induces rapid nongenomic destabilization of VEGF mRNA. ¹⁴ The novel finding that TA directly inhibits VEGF implies that TA acts at another unknown control point in the angiogenesis pathway that is separate from decreased VEGF secretion. This effect appears to be downstream of the IL-6 and VEGF receptors in cultured human endothelial cells. 

Methods

Rat Corneal Micropocket Assay

Animal experiments were performed in accordance with guidelines established by the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Review Committee for Animal Care at the Cleveland Clinic Foundation. The rat corneal micropocket assay was performed as described. ¹⁵ Briefly, sustained-release pellets were prepared by combining TA, IL-6, VEGF, or antibody with an equal volume of 12% (wt/vol) poly2-hydroxyethylmethacrylate polymer (Hydron; Sigma, St Louis, MO) such that each pellet contained 150 or 350 ng rat IL-6 (Sigma), 50 or 100 ng human VEGF 165 (Sigma), or 3 μg TA (Bristol-Myers Squibb, New York, NY). Female Fischer 344 rats (8–10 weeks of age) were anesthetized with xylazine and ketamine, and corneal micropockets were surgically created in both eyes. The left eye was used as a control in which pellets (Hydron; Sigma) with control buffer were inserted. On the seventh postoperative day, animals were deeply anesthetized and perfused with India ink by intracardiac injection to label the vessels. Six eyes were treated in each group, and the total vascular area surrounding the pellet was measured in corneas by a computer algorithm (Image-Pro Plus 5.0; Media Cybernetics, Silver Spring, MD) and presented as mean ± SE. Levels of endotoxin in recombinant IL-6 were measured to be less than 0.1 ng/μg. For the mifepristone group, 4 μg mifepristone was injected into the subconjunctival space adjacent to the cornea micropocket at the time of placement of the pellets. 

Chick Chorioallantoic Membrane Assay

Fertilized 3-day-old white Leghorn eggs (Case Western Reserve University Farms) were cracked, and embryos with the yolk intact were placed in glass-bottomed Petri dishes. Methylcellulose disks of 3-mm diameter containing IL- 6 (100 ng) or an equal amount of carrier buffer were implanted on the chorioallantoic membrane assay (CAM) on day 3, injected with India ink, and photographed after 3 days. Angiogenesis was assessed by comparing the branchpoint dilatation and tortuosity of the vessels under the transparent disks. 

HUVEC Cell Culture and IL-6/VEGF Treatment

Human umbilical vein endothelial cells (HUVECs) were isolated and grown as described previously. ¹⁵ Passage 2 and passage 7 HUVECs were grown until confluent and serum starved for 24 hours. IL-6 (100 μg/mL) was diluted 1/1000 into HUVEC cultures to give a final concentration of 100 ng/mL. Cells were pretreated with 1 μg/mL TA for 12 hours and then treated with IL-6 for 6 hours before cell lysates were prepared. For VEGF treatment, cells were prepared identically and treated with 50 ng/mL VEGF-A for 10 minutes, 30 minutes, and 6 hours. 

Western Blot

Cells were harvested in each particular experiment by scraping them into 2 mL chilled PBS and pelleting them in a 15-mL conical tube at 200g for 5 minutes. Cell lysates were made by directly resuspending cells in 2× sample buffer with 1 mM dithiothreitol (DTT). Lysates were subjected to 4% to 20% PAGE and were electrotransferred to polyvinylidene difluoride (PVDF) membrane for immunoblotting. Membranes were blocked with 5% nonfat dried milk in TBS-T and probed with anti-STAT3 (1:100; Santa Cruz Biotechnology, Santa Cruz, CA), anti-pSTAT3 (1:500; Cell Signaling, Danvers, MA), anti-ERK1/2 (1:500; Calbiochem, San Diego, CA), anti-pERK1/2 (1:500; Calbiochem), anti-VEGF-R2 (1:1000; Santa Cruz Biotechnology), and anti-pVEGF-R2 (1:500; Santa Cruz) antibody, washed 3× for 10 minutes with TBS-T and secondary antibody hybridization, and exposed by chemiluminescence (Western Lightning, Perkin Elmer, CA). 

Results

IL-6-Induced Corneal Neovascularization In Vivo

Newly formed vessels extended toward and then reached the IL-6 pellet implant within the cornea micropocket by day 6. The formation of new capillaries was evident in all animals implanted with 150 ng or more (n = 6) IL-6. A robust neovascularization response was produced by 350 ng IL-6 (n = 6; Fig. 1c ), whereas 150 ng IL-6 induced a more modest effect (n = 6; Fig. 1b ). Control pellets of PBS did not induce any neovascularization response (Fig. 1a) . Control pellets containing human serum albumin also showed no effect on neovascularization (n = 10; data not shown). 

VEGF-Dependent IL-6-Induced Angiogenesis

To determine whether IL-6-mediated angiogenesis is dependent on VEGF, we tested whether a neutralizing antibody to VEGF could inhibit the neovascularization mediated by IL-6. Coinsertion of a pellet containing neutralizing VEGF antibody along with IL-6 pellets resulted in an inhibition of IL-6-mediated neovascularization (n = 6; Fig. 2c ), whereas control nonspecific antibody had no effect (n = 6; Fig. 2b ). The neutralizing anti-VEGF antibody was efficient in its ability to inhibit neovascularization mediated by VEGF (Figs. 2g 2h 2i) , suggesting that IL-6 mediates angiogenesis through a VEGF-dependent pathway. Control nonspecific antibody and anti-VEGF antibody did not induce neovascularization (Figs. 2d 2e 2f) . 

TA Inhibition of IL-6– and VEGF-Induced Angiogenesis

To determine whether TA could block IL-6-induced angiogenesis, pellets containing TA and IL-6 were coinserted in the corneas of rats. The presence of the TA pellet caused a dramatic inhibition of 350 ng IL-6-induced angiogenic response in all the corneas tested (n = 6; Fig. 3b ). Inasmuch as TA is proven to decrease VEGF secretion and IL-6 angiogenesis is VEGF dependent, we next tested whether TA could block VEGF directly and expected that the VEGF pellet would overwhelm the TA pellet. However, we observed (Figs. 4a 4b)the complete inhibition of VEGF-induced angiogenesis by coinsertion of a TA pellet alongside a VEGF pellet (Fig. 4b)compared with a VEGF pellet at day 4. Total skeletal vessel length for each experiment was quantified to demonstrate the VEGF-dependent stimulatory effect of IL-6 and the inhibitory action of TA (Fig. 5) . 

TA Inhibition of VEGF-Induced Angiogenesis by a Mifepristone-Sensitive Steroid Receptor

To determine whether the effect of TA is specific and receptor mediated, we used the antagonist of the steroid receptor, mifepristone, to negate the inhibitory action of TA. A single subconjunctival injection of 4 μg mifepristone at the time of construction of the cornea micropocket blocked the action of TA (Fig. 6e) , demonstrating that the inhibition of angiogenesis by TA is receptor mediated. Although this was strong evidence of a nongenomic effect of TA through the cytosolic steroid receptor, we were unsure whether mifepristone could inhibit other receptor targets of glucocorticoids. A VEGF pellet induced robust angiogenesis (Fig. 6b) , but TA inhibited this effect (Fig. 6c) . A PBS control pellet had no effect (Fig. 6a) , and mifepristone alone injected into the subconjunctival space had no effect (Fig. 6d) . 

IL-6 and VEGF-R2 Response to TA

To determine whether TA blocks activation of the IL-6 receptor, HUVECs were first pretreated with 1 μg/mL TA for 12 hours after serum starvation for 24 hours and then were stimulated with IL-6 for 6 hours, which correlates with the second of a biphasic peak of STAT3 phosphorylation (pSTAT3). We noted a robust pSTAT3 species in both the IL-6 and the IL-6/TA cell lysates prepared from HUVECs (Fig. 7c) . Similarly, we tested whether VEGF-R2 was phosphorylated at 10 minutes, 30 minutes, and 6 hours after a protocol that was identical to the one described except that VEGF-A, not IL-6, was used as a stimulus. The VEGF receptor was phosphorylated at each of these time points (30 minutes and 6 hours; data not shown) in response to VEGF (Fig. 7a) . We next confirmed that VEGF-R2 was functional by testing whether extracellular signal-regulated kinase 1/2 (ERK1/2) was phosphorylated. Again we noted that ERK1/2 became phosphorylated at 10 minutes after VEGF stimulus despite pretreatment with 1 μg/mL TA (Fig. 7b) . Phosphorylation of the VEGF receptor and its immediate target of phosphorylation ERK1/2 implied that the receptor was stimulated and functional despite TA treatment. 

TA Inhibition of IL-6-Induced STAT3 Expression in Cornea

To extend our findings in cultured human endothelial cells, we next attempted to duplicate the phosphorylation experiments. However, we were unable to demonstrate robust pSTAT3 in the cornea, perhaps because of low levels of signal. This might have been because the pSTAT3 antibody was an anti-human antibody. We next checked the levels of STAT3 in the cornea and found that IL-6 increased the amount of STAT3 in cornea and that TA blocked this upregulation (Fig. 8) . 

Discussion

We have used the rat cornea micropocket to demonstrate that IL-6-induced angiogenesis is VEGF dependent and thereby links the molecular mechanism of immune-mediated angiogenesis to a common proangiogenic signal. The mechanism that directs induction of VEGF by IL-6 has recently begun to be defined ^{16 17 18} but may be specific to cell type and stimulus. IL-6 binds the IL-6R α-chain (gp80), which subsequently recruits gp130 to trigger signaling through Jak/STAT, Ras/mitogen-activated protein kinase (MAPK), and PI3-K/Akt. ^{17 19} Although the PI3-K/Akt pathway controls the level of active hypoxia-inducible factor-1 (HIF-1) by regulating the nuclear translocation of the α-subunit during hypoxia, VEGF synthesis has also been induced through the ERK1/2 MAPK pathway in cells maintained in low pH and by the STAT3-dependent pathway in IL-6-stimulated cervical cancer cells. ^{20 21} Specific pharmacologic inhibitors of Akt and MAPK failed to block IL-6 induction of VEGF, whereas the proangiogenic IL-6 effect was nullified in cells expressing a dominant-negative STAT3 mutation. ¹⁹ We demonstrate that in vivo, TA decreases the expression of STAT3 in cornea. Although limbal blood vessels were included in the cornea preparations used for Western blot, we assume that the predominant STAT3 signal comes from the great excess of extravascular tissues in cornea explants. On the contrary, TA did not decrease the expression or phosphorylation of STAT3 after IL-6 treatment in cultured human umbilical vein endothelial cells. Although these results appeared contradictory, it may be that these different cell types responded differently and that cornea specimens were predominantly avascular. 

Evidence clearly indicates that TA decreases the secretion of VEGF into human vitreous and decreases the level of VEGF mRNA in cultured human retinal pigment epithelial cells challenged with hypoxia or oxidation. ^{22 23} Dexamethasone also decreases VEGF mRNA in hypoxic rat glioma cells. ²⁴ These data demonstrate the broad action of glucocorticoids, which can decrease VEGF levels in oxidative, hypoxic, or, as described in this report, immune-mediated experimental systems. The fact that TA is reported to decrease VEGF secretion suggests that if VEGF itself, not inducers of VEGF secretion such as IL-6, is used as a stimulus, TA would be ineffective at inhibiting its action. On the contrary, we found that TA also directly blocks VEGF in the cornea micropocket. Both the IL-6 receptor and the VEGF receptor are shown here to be functional during TA treatment, proving that in addition to decreasing the RNA stability of VEGF, TA inhibits angiogenesis downstream of the VEGF receptor. Others have reported the resistance of retinal blood vessels pretreated with glucocorticoid to leakage despite the presence of VEGF in vivo. ²⁵ Our data agree with this model. This finding is significant because it implies that a second cofactor, induced by VEGF and inhibited by TA, modulates angiogenesis. The inhibition of total skeletal length of new vessels in our experiment by TA is often close to 50%, suggesting the involvement of angiogenic factors other than VEGF. Specific identification of the molecular mechanism of TA global inhibition of VEGF may yield potent new strategies to inhibit angiogenesis once these factors are known. 

 Supported by the National Institutes of Health/National Institute of Child Health and Human Development Grant K12HD049091 (JES).

 Submitted for publication December 27, 2005; revised April 22 and May 23, 2006; accepted September 25, 2006.

 Disclosure: Q. Ebrahem, None; A. Minamoto, None; G. Hoppe, None; B. Anand-Apte, None; J.E. Sears, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Angiogenic dose response of IL-6. Representative images of a rat cornea micropocket. (a–c) Dose response of 150 and 350 ng IL-6 pellets in a rat cornea micropocket. The PBS control pellet has no effect. (d, e) An IL-6 filter placed on a chick chorioallantoic membrane provokes vasopermeability, seen here as leakage of India ink.

Figure 2.

 

IL-6-induced angiogenesis is VEGF dependent. Representative images of a rat cornea micropocket. (a–c) Coinsertion of anti-VEGF antibody and IL-6 pellets demonstrates blockade of IL-6-induced angiogenesis by anti-VEGF antibody. Nonspecific antibody has no effect on angiogenesis induced by IL-6. (d–f) Coinsertion of the anti-VEGF antibody or the nonspecific antibody pellet provokes no vascular inflammation or angiogenesis in the company of a sham PBS pellet. (g–i) Coinsertion of VEGF and anti-VEGF antibody pellet shows specificity of the anti-VEGF antibody and nonspecificity of the control antibody.

Figure 3.

 

TA inhibits IL-6-induced angiogenesis. Representative images of a rat cornea micropocket. (a–c) Coinsertion of a TA pellet with an IL-6 pellet inhibits corneal angiogenesis. Although this may interrupt the VEGF-dependent angiogenic stimulus, it has not been conclusively proven here that TA inhibits VEGF secretion or acts through other multiple pathways to block angiogenesis.

Figure 4.

 

TA inhibits VEGF-induced angiogenesis. Although TA is known to inhibit VEGF secretion, these images demonstrate that TA also blocks VEGF directly, suggesting multiple points of inhibition by TA in the angiogenesis cascade, multiple cell targets, or both.

Figure 5.

 

Quantification of corneal angiogenesis. A computer algorithm was used to quantify the length of new vessels from the limbus and was presented with SD based on n = 6 experiments for each error bar. IL-6 induced angiogenesis. This effect was blocked by TA and by anti-VEGF antibodies.

Figure 6.

 

Reversal of TA inhibition by mifepristone. Cornea micropockets were prepared as before with a negative control of PBS (a) and a positive control of 50 ng VEGF (b). TA blocks VEGF angiogenesis (c). Mifepristone (4 μg) was injected subconjunctivally at the time of surgery for pellet implantation adjacent to micropockets containing either PBS (d) or TA/VEGF pellets (e). Mifepristone reversed the inhibitory effect of TA, indicating that the action of TA was receptor mediated. Each experiment was repeated six times. Results of a representative experiment are shown.

Figure 7.

 

TA does not block the IL-6 or the VEGF-2 receptor in cultured HUVECs. Immunoblot using anti-VEGF receptor 2 antibody and anti-phospho-VEGFR-2 (top), anti-ERK1/2 antibody and anti-phospho-ERK1/2 (middle), and anti-STAT3 and anti-phospho-STAT3 (bottom). Western blot demonstrates that VEGF-R and ERK1/2 are functional and activated in the presence of TA and VEGF after 10 minutes of VEGF stimulation in HUVECs, demonstrating that TA does not inhibit activation of the VEGF receptor 2 or the early targets of signal transduction. STAT3, the transcription factor stimulated by the IL-6 receptor, is phosphorylated at 6 hours despite TA treatment. These data suggest that TA does not block signal transduction through the VEGF-R2 or the IL-6R in HUVECs.

Figure 8.

 

TA decreases IL-6-induced expression of STAT3. Cornea micropockets were created holding PBS control pellet, 350 ng IL-6, 1 μg TA, or both. Whole cornea, including limbal vessels, was excised and lysed in sample loading buffer by sonication. Protein levels were quantified, and 15 μg total protein was separated by PAGE and analyzed by immunoblot. TA decreases STAT3 expression in IL-6-stimulated cornea specimens.