Topical administration of the kappa opioid receptor agonist nalfurafine suppresses corneal neovascularization and inflammation

Corneal neovascularization (CNV) causes higher-order aberrations, corneal edema, ocular inflammation, and corneal transplant rejection, thereby decreasing visual acuity. In this study, we investigated the effects of topical administration of the kappa opioid receptor agonist nalfurafine (TRK-820) on CNV. To induce CNV, intrastromal corneal sutures were placed on the corneal stroma of BALB/c mice for 2 weeks. Nalfurafine (0.1 µg/2 μL/eye) was topically administered to the cornea once or twice daily after CNV induction. The CNV score, immune cell infiltration, and mRNA levels of angiogenic and pro-inflammatory factors in neovascularized corneas were evaluated using slit-lamp microscopy, immunohistochemistry, flow cytometry, and polymerase chain reaction. The mRNA expression of the kappa opioid receptor gene Oprk1 was significantly upregulated following CNV induction. Topical administration of nalfurafine twice daily significantly suppressed CNV and lymphangiogenesis, as well as reduced the mRNA levels of angiogenic and pro-inflammatory factors in the neovascularized corneas. Moreover, nalfurafine administration twice daily reduced the numbers of infiltrating leukocytes, neutrophils, macrophages, and interferon-γ-producing CD4+ T cells in the neovascularized corneas. In this study, we demonstrated that topical administration of nalfurafine suppressed local CNV in a mouse model along with the activation of KOR, suggesting that nalfurafine may prevent and control CNV in humans.


Results
Increased mRNA expression levels of KOR and angiogenic, lymphangiogenic, and inflammation factors in the neovascularized cornea. To investigate the effects of CNV induced by suture knots in the central cornea, mRNA levels of angiogenic, lymphangiogenic, and inflammatory factors in the neovascularized tissue were analyzed using reverse transcription-quantitative polymerase chain reaction (RT-qPCR) on days 7 and 14 post neovascularization induction. Figure 1a shows a representative slit photograph of the neovascularized cornea. mRNA levels of Pecam1 (encoding CD31) and Lyve1 (encoding lymphatic vessel endothelial hyaluronan) were significantly higher in the cornea affected by neovascularization than in the control group (Fig. 1b, n = 3, P < 0.001, one-way analysis of variance [ANOVA] for both days 7 and 14 post-induction; Fig. 1c, n = 3, P < 0.001, for both days 7 and 14 post induction). Interestingly, mRNA levels of Oprk1 that encodes kappa opioid receptors were also significantly increased at days 7 and 14 post neovascularization induction compared to that in the control tissue (Fig. 1d, n = 3, P = 0.022 and P = 0.002, respectively). Vegfa and Vegfc mRNA levels were significantly increased at days 7 and 14 post neovascularization induction compared to the level in control (Fig. 1e,f; n = 3, P < 0.001). mRNA levels of Flt1 that encodes VEGFR-1 were significantly increased only at day 14 post neovascularization induction compared to the control level (Fig. 1g, n = 3, P = 0.049). In contrast, mRNA levels of Kdr and Flt4 that encode VEGFR-2 and VEGFR-3, respectively, were significantly elevated at days 7 and 14 post neovascularization induction compared to their levels in the control tissue (Fig. 1h,i; n = 3, Kdr, P = 0.001, P < 0.001, respectively; Flt4, P < 0.001). In addition, significant elevations of mRNA expression in the cornea affected by neovascularization were noted for Nrpl, encoding the VEGF co-receptor neuropilin-1 18 , at day 14 post-induction (Fig. 1j, n = 3, P = 0.049) and Ifng, encoding the pro-inflammatory cytokine interferon-γ (IFN-γ) 19 , at days 7 and 14 post neovascularization induction (Fig. 1k, n = 3, P < 0.001). The relative degrees of increases in mRNA expression of several angiogenic factors in the neovascularized tissue are illustrated by a heatmap in Fig. 1l. Particularly high increases in expression levels at days 7 and 14 post-induction compared to the levels in control were noted for the genes encoding angiogenic factors VEGF-A, VEGF-C, and VEGFR-3 ( Fig. 1l).
Downregulation of the mRNA expression of angiogenic factors in the naïve cornea following topical administration of nalfurafine. To investigate dose-dependent effects of the topically administered KOR agonist nalfurafine on angiogenic signals, we administered it once or twice a day to the naïve cornea. The dosage used (0.1 μg/2 μL/eye) was based on the dosage used in a previous study 16 . The mRNA levels of the genes encoding angiogenic and lymphangiogenic factors were analyzed in the samples of the naïve cornea using RT-qPCR (Fig. 2). We found that Vegfa, Vegfc, Flt1, Kdr, Flt4, and Nlp1 mRNA levels were significantly lower in the "nalfurafine twice a day" group than in the control group ( Fig. 2a-f, n = 3; Vegfa, P = 0.001; Vegfc, P = 0.002; Flt1, P = 0.004; Kdr, P = 0.004; Flt4, P = 0.001; and Nrp1, P < 0.001). Smaller, once a day dosage led to a significant decrease in expression of only Vegfc mRNA (Fig. 2b, n = 3, P = 0.039). Figure 2g shows a heatmap of relative changes in mRNA expression of the angiogenic factors depending on the dosage.
Topical administration of nalfurafine upregulated and downregulated the mRNA expression levels of KOR and angiogenic factors, respectively, in neovascularized cornea. Next, we investigated the effects of topical administration of nalfurafine on mRNA expression of KOR and angiogenic and lymphangiogenic factors in the neovascularized cornea (Fig. 3). RT-qPCR results showed that the mRNA levels of Oprk1 encoding the kappa opioid receptors were significantly upregulated in the nalfurafine twice a day group than in the control group (Fig. 3a, n = 3, P = 0.002). We also observed that mRNA levels of Vegfa, Vegfc, Flt1, Kdr, Flt4, and Nrp1 were all significantly lower in the nalfurafine twice a day group than in the control group ( Fig. 3b-g, n = 3, Vegfa, P < 0.001; Vegfc, P < 0.001; Flt1, P = 0.040; Kdr, P < 0.001; Flt4, P = 0.002; and Nrp1, P < 0.001). Notably, lower dosage (once a day treatment) caused numerical decreases in the expression levels of Vegfa, Kdr, Flt4, and Nrp1, but the effect did not reach statistical significance in any of these cases ( Fig. 3b-g). Figure 3h shows a heatmap of relative changes in mRNA expression of angiogenic factors in neovascularized www.nature.com/scientificreports/ Vegfa encoding vascular endothelial growth factor (VEGF) A, (e) Vegfc (f), Flt1 encoding VEGF receptor 1 (g), Kdr encoding VEGFR-2 (h), Flt4 encoding VEGFR-3 (i), Nrp1 encoding neuropilin-1 (j), and Ifng encoding interferon-γ (k). Data are presented as the mean ± standard error of the mean (n = 3 in all cases). Statistical significance of differences is illustrated as follows: *P < 0.05; **P < 0.01; ***P < 0.001 (one-way analysis of variance followed by the Bonferroni test). (l) Heatmap shows the distribution of fold changes in mRNA levels of angiogenetic and inflammation factors in the cornea after neovascularization induction. ns, no significant difference. www.nature.com/scientificreports/ (a-f) Significant decreases in mRNA levels were noted at 14 days post treatment in samples that were exposed to nalfurafine once and/or twice a day for the following VEGF signaling components: Vegfa (a), Vegfc (b), Flt1 (c), Kdr (d), Flt4 (e), and Nrp1 (f). Data are presented as the mean ± standard error of the mean (n = 3 in all cases). Statistical significance of differences is illustrated as follows: *P < 0.05; **P < 0.01; ***P < 0.001 (one-way analysis of variance followed by the Bonferroni test). (g) Heatmap shows relative decreases in expression levels of angiogenic factors in the naïve cornea depending on the dosage of nalfurafine. ns, no significant difference. www.nature.com/scientificreports/ cornea depending on nalfurafine dosage. Overall, the expression of angiogenic factors was lower in the nalfurafine twice a day group compared to that in control and nalfurafine once a day groups.
Topical administration of nalfurafine decreased inflammatory cell infiltration and lowered mRNA expression of inflammation-related molecules in the neovascularized cornea. Infiltrating inflammatory cells secrete VEGF and inflammatory cytokines [20][21][22][23][24] . To determine the effect of nalfurafine on inflammatory cell infiltration in the neovascularized cornea, the numbers of CD45 + leukocytes, CD4 + T cells, CD45 + CD11b + Ly6G + neutrophils, and CD45 + CD11b + F4/80 + macrophages infiltrating the neovas- Relative mRNA expression levels of angiogenic factors in the control and treated neovascularized corneal samples were analyzed using RT-qPCR (b-g). Significant decreases in mRNA levels were observed,14 days post treatment in samples that were exposed to nalfurafine once and/or twice a day for the following VEGF signaling components: , Flt4 (f), and Nrp1 (g). Data are presented as the mean ± standard error of the mean (n = 3 in all cases). Statistical significance of differences is illustrated as follows: *P < 0.05; **P < 0.01; ***P < 0.001 (one-way analysis of variance followed by the Bonferroni test). (h) Heatmap shows relative decreases in expression levels of angiogenic factors in the neovascularized cornea depending on the dosage of nalfurafine. ns, no significant difference.   Figure 5 shows the expressions of CD4 + IFN-γ + T cells and inflammation cytokines in the neovascularized cornea following topical administration of nalfurafine. IFN-γ-producing activated T cells are recognized as the principal mediators of corneal inflammatory responses 25,26 . The numbers of CD4 + IFN-γ + T cells in the neovascularized cornea were analyzed at day 14 post neovascularization induction (Fig. 5a). Flow cytometry analysis showed a marked reduction in the quantity of CD4 + IFN-γ + T cells in the neovascularized corneal samples treated with topically administered nalfurafine once or twice a day compared to their number in the control group (Fig. 5b, n = 3, P = 0.002, P < 0.001, respectively). Additionally, the Ifng mRNA level was significantly decreased in the neovascularized cornea after the treatment with topically administered nalfurafine once or twice a day compared to that in control samples (Fig. 5c, n = 3, P = 0.017, P < 0.001, respectively). Among the mRNA expressions of inflammatory cytokines, Tnf-α was decreased following nalfurafine once or twice a day treatment (Fig. 5d, n = 3, P < 0.001, P < 0.001, respectively), while Il1β was suppressed only in the twice a day group (Fig. 5e, n = 3, P < 0.001). Moreover, mRNA expression levels of Krt12 that encodes the corneal epithelial differentiation marker cytokeratin 12 27 did not differ significantly between the groups (Fig. 5f, n = 3, P > 0.05).
The protein levels of VEGF, IFN-γ, TNF-α, and IL1β were measured using enzyme-linked immunosorbent assay (ELISA). VEGF (Fig. 5g, P = 0.018, P = 0.006, respectively ), IFN-γ (Fig. 5h, P = 0.032, P < 0.001, respectively), TNF-α (Fig. 5i, P = 0.040), and IL1β ( Fig. 5j, P = 0.020) expression levels were significantly reduced in the nalfurafine once or twice a day treatment compared with the control group. Figure 6a shows representative images of neovascularized cornea at days 3 and 14 after neovascularization induction following topical administration of nalfurafine. Topical administration of nalfurafine twice a day significantly reduced the neovascularization score compared to that in the control group at days 10 and 14 post neovascularization induction (Fig. 6b, n = 15 per each group; P = 0.034, P = 0.007, respectively, two-way ANOVA followed by the post hoc Bonferroni test). Figure 6c shows representative images of CD31 staining of the whole cornea and the processed images on day 14 post neovascularization induction. Fractions of the corneal area covered with blood vessels (CD31 + ) in the samples treated with topically administered nalfurafine once or twice a day were significantly smaller than those in the control group (Fig. 6d, n = 5, P < 0.001). Furthermore, the mRNA level of Pecam1 was significantly reduced by the treatment with nalfurafine twice a day compared to the level in the control group on day 14 post neovascularization induction (Fig. 6e, n = 3, P < 0.001). Figure 6f shows representative images of LYVE1 staining of the whole cornea and its processed images on day 14 post neovascularization induction. The fractions of the corneal area covered with lymphatic vessels (LYVE1 + ) in the samples exposed to topically administered nalfurafine once or twice a day were significantly reduced compared  (Fig. 4 g) and the histogram (Fig. 4 h) showing the frequency of CD45 + CD11b + Ly6G + neutrophils in the neovascularized corneas 14 days post neovascularization induction in control conditions and following treatment with nalfurafine (0.1 µg/2 μL/eye) once or twice a day. The number of CD45 + CD11b + Ly6G + neutrophils was significantly lower in tissues treated with topical administration of nalfurafine once or twice a day compared with the control group (Fig. 4i, n = 3, P = 0.004, P < 0.001, vs. control, respectively). Representative flow cytometry plots (Fig. 4j) and the histogram (Fig. 4 k) showing the frequency of CD45 + CD11b + F4/80 + macrophage. The abundance of CD45 + CD11b + F4/80 + macrophage infiltrating the corneas was significantly reduced in the corneal samples treated with nalfurafine once or twice a day (Fig. 4 l, n = 3, P < 0.001, P < 0.001, vs. control, respectively). mRNA levels of Ptprc (Fig. 4 m), CD11b (Fig. 4n), Gr-1 (Fig. 4o), and Adgre1 (Fig. 4p) were significantly reduced following topical administration of nalfurafine twice a day, 14 days post neovascularization induction. www.nature.com/scientificreports/ , and IL1β (j) measured using ELISA were significantly reduced following topical administration of nalfurafine twice a day, 14 days post neovascularization induction. Data are presented as the mean ± standard error of the mean (n = 3 in all cases). Statistical significance of differences is illustrated as follows: ns, no significant difference; *P < 0.05; **P < 0.01; ***P < 0.001 (one-way analysis of variance followed by the post hoc Bonferroni test). IFN, interferon; Tnf-α, Tumor necrosis factor-α; Il1β, interleukinβ; ELISA, enzyme-linked immunosorbent assay. www.nature.com/scientificreports/ The fractions of the total cornea area covered with blood vessels (CD31 + ) were significantly reduced in corneas that were treated with nalfurafine once or twice a day compared to that in control, untreated corneas on day 14 post neovascularization induction (mean ± standard error of the mean, n = 5, statistical significance is indicated as follows: *P = 0.033 and ***P < 0.001 vs. control). (e) Relative mRNA expression levels of Pecam1, encoding CD31, were significantly decreased in corneas that were treated with nalfurafine twice a day compared to that in control, untreated corneas on day 14 post neovascularization induction (mean ± standard error of the mean, n = 3, statistical significance is indicated as follows: ***P ≤ 0.001 vs. control). (f) Images showing lymphatic vessel endothelial hyaluronan receptor1 (LYVE1) staining of the whole cornea and the processed images. (g) The fractions of the cornea area covered with lymphatic vessels were significantly reduced in corneas that were treated with nalfurafine once or twice a day compared with the control, untreated corneas on day 14 post neovascularization induction (mean ± standard error of the mean, n = 3, **P < 0.01, ***P < 0.001 vs. control, respectively). (h) Relative Lyve1 mRNA expression was significantly decreased in corneas that were treated with nalfurafine twice a day compared to that in control, untreated corneas on day 14 post neovascularization induction (mean ± standard error of the mean, n = 3, statistical significance is indicated as follows: *P = 0.016, vs. control). ns no significant difference. www.nature.com/scientificreports/ to that in the control group (Fig. 6g, n = 5, P < 0.01, P < 0.001). In addition, Lyve1 mRNA level was significantly decreased in corneal samples after topical administration of nalfurafine twice a day on day 14 post neovascularization induction (Fig. 6h, n = 3, P = 0.016).

Discussion
Effective treatment of CNV represents an unmet medical need and is critical for the retention of visual function.
In this study, we investigated the efficacy of the topical administration of the KOR agonist nalfurafine in the treatment of CNV in mice. We demonstrated that topical administration of nalfurafine attenuated angiogenesis and immunological responses by upregulating KOR expression levels and downregulating expression levels of several angiogenic and inflammatory factors in the cornea. These results suggest that topical administration of nalfurafine and other KOR agonists might be an effective treatment option for the management of CNV. We showed that CNV induction in the cornea led to a significant increase in the expression of Oprk1 that encodes KOR (Fig. 1d), confirming the notion about the stimulatory effect of peripheral inflammation on the expression of KOR in the corneal microenvironment in accordance with a previous study 28 . Another previous study showed the increased expression of KOR in human umbilical vein endothelial cells in vitro 16 . These results indicate that the increased expression of KOR may be induced by the increased vascular endothelial cells itself in the local cornea during the angiogenesis process. As shown by the schematic in Fig. 7, inflammatory mediators activate G proteins that upregulate cyclic adenosine monophosphate (cAMP) generation, which in turn activates the protein kinase A pathway for VEGF production, whereas KOR activation suppresses VEGF production by promoting Gαi/o inhibition of adenylyl cyclase, which decreases cAMP levels and attenuates protein kinase A activation 29,30 It has been reported that the downregulation of KOR expression in hepatocellular carcinoma had a strong association with poor prognosis 31 . We showed here that topically administered nalfurafine, which activates KOR, inhibited the expression of genes encoding VEGF-A, VEGF-C, as well as VEGFR-1-3 (Figs. 2, 3), resulting in decreased CNV and lymphangiogenesis in neovascularized corneas (Fig. 6). These findings suggest that the KOR system controls the development of blood and lymphatic vessels in the corneal microenvironment 32 . Therefore, ligands that activate the KOR system, such as nalfurafine, may be useful as anti-angiogenic agents in the cornea.
CNV occurs as a result of an imbalance between angiogenic and anti-angiogenic factors in favor of angiogenic molecules [32][33][34][35] . This imbalance causes pathogenic angiogenesis in CNV, which in turn activates pro-angiogenic factors including VEGF, VEGFR, IFN-γ, TNF-α, and IL1β in the neovascularized microenvironment and strongly potentiates further development of CNV 21,36,37 . Therefore, controlling the expression levels of VEGF, VEGFR, and inflammation cytokines is an effective strategy to inhibit angiogenesis in the local corneal tissue 21,38,39 . In this study, topical administration of nalfurafine twice a day significantly inhibited CNV compared to the extent of the pathology in the untreated control or nalfurafine once a day groups (Fig. 5). In addition, as shown in Fig. 3, Vegfa mRNA expression was significantly reduced in the treatment group administered nalfurafine twice www.nature.com/scientificreports/ a day compared to that in the treatment group administered nalfurafine once a day or in the untreated group. This result suggests that sustained KOR activation achieved by topical administration of nalfurafine inhibits CNV. Furthermore, we showed that treatment with nalfurafine downregulated the expression of genes that encode the VEGF co-receptors VEGFR-2 and neuropilin-1 18 in the neovascularized cornea. A previous study has reported that KOR activation suppresses the expression of VEGFR-2 and neuropilin-1 by inhibiting cAMP/PKA signalling 40 . Furthermore, because lymphatic vessels multiply during angiogenesis 41 , the decreased expression of VEGFR-3 is secondary to the suppression of increased angiogenesis by KOR activation, which also suppresses lymphangiogenesis. Here, we found that topical administration of nalfurafine to the neovascularized cornea reduced the expression of genes encoding VEGF and VEGF-R, which is in agreement with previous studies 16,42 , further confirming that nalfurafine may be effective as a topical agent in CNV.
In CNV, angiogenesis of the corneal local tissue is accompanied by the induction of local inflammation, which damages the corneal epithelium and stroma, leading to the infiltration of inflammatory cells, primarily neutrophils and macrophages 6,8,[43][44][45] . Infiltrating neutrophils secrete VEGF 20,21 . Abnormally growing blood and lymphatic vessels trigger local infiltration of inflammatory cells to the cornea and further promote angiogenesis 38,46 . Topical administration of nalfurafine reduced the number of infiltrating CD45 + CD4 + and CD4 + IFN-γ + cells, CD45 + CD11b + Ly6G + neutrophils, CD45 + CD11b + F4/80 + macrophage, and inflammatory cytokine expression in the neovascularized cornea in our experiments (Figs. 4, 5). Because nalfurafine inhibited the infiltration of these cells, the secretion of VEGF-A and VEGF-C may have been reduced. This indicates that activating the KOR system suppressed the infiltration of inflammatory cells to the damaged sites in the neovascularized cornea likely by reducing the levels of associated inflammatory cytokines.
In infectious keratitis and corneal transplantation concurrent with CNV, both the suppression of angiogenesis and pain control are important. Opioid systems are mainly responsible for the control of neurophysiological functions, including pain, mood changes, physical dependence, reward, and addictive behaviours 47 . The opioid system comprises MOR, DOR, and KORs and their endogenous ligands 48 . The KOR agonists are considered as promising alternatives to the MOR agonists: the former also have strong analgesic properties but are devoid of the side effects caused by MOR activation. Given that we showed the suppressive effects of the KOR agonist nalfurafine on CNV in mice, it is possible that eye drops containing nalfurafine may be able to control pain as well as suppress angiogenesis in local corneal tissue.
This study has several limitations. We identified the inhibitory effects of nalfurafine on CNV in a murine model. However, because CNV may be triggered by various pathological conditions, further studies using alternative CNV models, such as the alkali burn model and corneal micropocket assay should be considered. Although sustained activation of KOR with a selective agonist was thought to be important for the inhibition of CNV, the optimal concentration and doses of nalfurafine for topical administration to the cornea were not determined in this study because we used the previously published dosing regimen for this drug 16 . In addition, the treatment with much higher doses of KOR agonists for longer periods might lead to the development of tolerance to kappa opioids, and this possibility should be investigated 16,49 . Furthermore, although a potential advantage of using nalfurafine and other KOR agonists is the combination of attenuated angiogenesis and pain relief due to their analgesic effects, we did not directly investigate pain in this study. Moreover, we did not compare the antiangiogenic effects of nalfurafine with those of other KOR agonists 50 . Therefore, further studies are required to assess the possible anti-angiogenic effects of a broader spectrum of KOR agonists in CNV. Nonetheless, our study is an important step forward, as there is currently no ophthalmic medication that both provides pain relief and has anti-angiogenic effects in clinical ophthalmic practice. The results of our study warrant further experiments to clarify the effects of nalfurafine and other KOR agonists and pave way to the potential clinical application of these drugs for CNV treatment.
In conclusion, we have shown that that topical administration of nalfurafine, a clinically used KOR agonist, suppressed local CNV in a mouse model. Therefore, KOR agonists might be useful for preventing and controlling CNV. Topical administration of eye drops containing nalfurafine. The KOR agonist (-)-17-cyclopropylmethyl-3,14b-dihydroxy-4,5a-epoxy-6b-[N-methyl-trans-3-(3-furyl) acrylamido] morphinan hydrochloride (nalfurafine, TRK-820, TORAY, Tokyo, Japan) 13,14 has been approved in Japan as a treatment for hemodialysisrelated uremic pruritus 17 . Nalfurafine was topically administered to the naïve or neovascularized cornea once or twice a day (0.1 µg/2 μL/eye). The dose-dependent effects of topically administered nalfurafine on angiogenic signals were investigated by once a day or twice a day application of the drug to the naïve cornea.

Methods
Suture-induced CNV model. A murine CNV model was created as previously described 41,51 . First, 1.5mm trephine was used to mark the central cornea of BALB/c mice. Three intrastromal figure-of-eight suture knots were placed into the central cornea using 11-0 nylon sutures (AB-0550S; MANI, Tochigi, Japan) for 14 days. www.nature.com/scientificreports/ Assessment of cornea neovascularization. CNV was evaluated once a week using a slit-lamp biomicroscope for 8 weeks. We used the standard neovascularization grading scheme (range from 0 to 8) 52 . Neovascularization grading were assigned as follows: 0, no vessels; 1, vessels peripheral cornea only (1-2 quadrants); 2, vessels peripheral cornea only (3-4  RNA isolation and RT-qPCR. The excised corneas were immediately submerged in the RNAlater solution (Ambion, Austin, TX, USA). Total RNA was isolated from five corneas per group by using a NucleoSpin RNA isolation kit (Macherey-Nagal GmbH, Duren, Germany). The cDNA was reverse-transcribed from total RNA by using random primers and a ReverTra Ace qPCR RT kit (Toyobo, Osaka, Japan) following the manufacturer's guidelines. The qPCR primers specific for mouse mRNA are shown in Table 1. qPCR was performed with the Applied Biosystems 7500 Fast Real-Time PCR Systems using a FAST-SYBR Green master mix (Thermo Fisher Scientific K.K., Tokyo, Japan). Results were analyzed using the comparative cycle threshold method, and Gapdh mRNA expression in the same cDNA sample was used as the internal control.
Corneal whole mount and immunofluorescence staining. Corneas were excised on day 14 after suture-induced CNV and washed with phosphate-buffered saline. The corneal epithelium was removed after incubation with 20 mM EDTA for 60 min at 37 °C, fixed in acetone for 15 min at 20-22 °C, and blocked in 2% bovine serum albumin for 60 min. Corneas were then double-stained for CD31 and lymphatic vessel endothe- Table 1. Primer sets for real-time PCR.

Adgre1
TCT GGG GAG CTT ACG ATG GA  TGG TTC TGA ACA GCA CGA CA   Cd11b  CGG AAG GAT TCA GCA AGC CAG AAC  AGC TGG ACT CAG CAG GCT TTAC   Flt1  GAG GAG GAT GAG GGT GTC TAT AGG T GTG ATC AGC TCC AGG TTT GACTT   Flt4  GTG CTC AAA GAG GTG ACC GA  TGA GGA GGC ACA TTC ACC AC   Gapdh  AAG GGC TCA TGA CCA CAG TC  GGA TGA CCT TGC CCA