Chrysin Alleviates DNA Damage to Improve Disturbed Immuno-Homeostasis and ProAngiogenic Environment in Laser-Induced Choroidal Neovascularization

Background: Choroidal neovascularization (CNV) is a devastating pathology of numerous ocular diseases, such as wet age-related macular degeneration (wAMD), which causes irreversible vision loss. Although anti-VEGF therapy has been widely used, poor response or no response exits in some patients, suggesting that some other important angiogenic components play roles. Therefore, the underlined mechanism need to be claried and new target of anti-angiogenic therapy is urgently needed. Damaged retinal pigment epithelium (RPE) cells have been demonstrated to activate inammasome, drive a degenerative tissue environment and an enhanced pro-angiogenic response, which emphasizes the dysfunction of RPE, may be the hallmark of the pathogenesis. Methods: C57BL/6J male mice aged between 6 and 8 weeks were subjected to laser-induced CNV models. Chrysin was administered intragastrically at 25 mg/kg daily for 3 days or one week after laser-treated. Then to observe the CNV areas and CNV thickness, immunouorescence staining of choroidal atmount, SD-OCT and uorescein angiograghy were performed, respectively. To further conrm the effect of chrysin on stress-induced DNA damage in RPE cells, RPE cells were administered with A2E and western-blot, cell viability assay, immunouorescence chromosome PNA-FISH and SA-β-gal staining were performed. To elucidate the underlying mechanism, we performed RNA-seq and bioinformatics analyses. Results: In this study, we demonstrated that chrysin could successfully alleviated choroidal neovascularization. We show that DNA damage of RPE cells is remarkable in laser-induced choroidal neovascularization, resulting in inammation response, which can be ameliorated by chrysin through inactivation of STAT3. Also, we identify that chrysin can reduce DNA damage, especially telomere erosion, simultaneously compromise the dysfunction of RPE and the secretion of SASP factor in vitro. Mechanistically, KEGG pathway analyzes show that chrsyin improves inammatory imbalance mainly through down-regulation of IL17 pathway In decreased DNA damage, telomere in the of senescence cell reduced and alleviating the parainammation in a stress-induced RPE damage model in vitro. our results suggest that chrysin mainly down regulate the IL17 pathway in this inammatory imbalance environment. our results emphasize the critical role of interplay between DNA damage, RPE cell dysfunction, inammatory imbalance and angiogenesis in CNV development. Besides, chrysin may be of promising therapeutic value for the treatment of neovasvular diseases.


Background
Choroidal neovascularization (CNV) is the hallmark of numerous ocular diseases, such as wet age-related macular degeneration (wAMD), which is the major cause of vision impairment among elderly population in developed countries. The etiology of CNV is believed to be multifactorial and remains unclear.
Intravitreal injection of anti-VEGF agents, which is the rst-line therapy for CNV treatment, has achieved great success in suppressing pathological angiogenesis and improving vision. However, the bene cial effect of anti-VEGF therapy begins to wear off and a substantial number of patient revealed poor response or no response to anti-VEGF therapy [1,2], suggesting other proangiogenic mechanisms may play roles and need to be elucidated urgently.
Intravitreally injected with chrysin successfully reduced the intensity of uorescein leakage in laserinduced CNV lesion [12]. However, the mechanism is undiscovered. In present study, we demonstrated that chrysin could protect RPE cells from DNA damage, improve the disturbed immuno-homeostasis and pro-agiogenic environment, which effectively suppress CNV development. Results indicate that chrysin may be an effective therapeutic supplement for CNV and emphasize the critical role of interplay between DNA damage, RPE cell dysfunction, in ammation and angiogenesis in laser-induced CNV.
Adult C57BL/6 mice (aged between 6-8 weeks old, weighed 20 ± 1 g) were used in this study. All animals were treated according to the guidelines of the ARVO statement for the Use of Animals in Ophthalmic and Vision Research. The experimental procedures were approved by the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University (Shanghai, China). The mice were housed and maintained in the animal care services facility and subjected to a 12-hour light/dark cycle with constant access to nourishments. Chrysin (Sigma-Aldrich, C801052) was dissolved in dimethyl sulfoxide (DMSO); a nal use concentration of DMSO was <0.5%. A2E formation and treatment A2E was synthesized as described previously. [27] RPE cells were incubated with A2E in culture medium for 24h and washed three times to remove extracellular A2E. After A2E loading, RPE cells were exposed to 460 ± 20 nm wavelength light (4000lx; Osram, Augsburg, Germany) for 20 min, as described previously [27] .
Laser-induced Mouse CNV model C57BL/6J male mice aged between 6 and 8 weeks and approximately 20 g were included. After application of tropicamide (Santen, Osaka, Japan) for pupil dilatation, animals were anesthetized with intraperitoneal injection of 1% pentobarbital sodium (0.1 mL/10 g body weight) (Guge Biotech, Wuhan, China). Covered with loxacin eye ointment (Xing Qi Pharmaceutical Companies, Shenyang, China), and four laser spots were distributed around the optic nerve head with an argon laser (110 mW, 100ms, 50μm, OcuLight Infrared Laser System 810 nm, Iridex Corp., Mountain View, CA, USA). Appearance of a gray bubble indicative of the rupture of Bruch`s membrane were included. If retinal bleeding occurred, the animal was eliminated. Eyes were enucleated at different time points.

Perfusion xation
Mice were perfused transcardially with cold 4% paraformaldehyde. In brief, mice were administered with an over dose (0.2mL) of 1% sodium pentobarbital and monitored until the point when the animal fails to respond to pinching of the foot. Incisions in the abdomen and diaphragm were made to expose the heart and perfusion needle was placed into ascending aorta. Cold 4% paraformaldehyde was poured into left ventricle of mouse through a peristaltic perfusion pump (Cole-Parmer Master ex, NewYork, USA).
Twitching of muscles suggests that the perfusion is proceeding properly. After the e uent runs clear, pump was stopped, eyes were harvested, post-xed for two hours and then placed in PBS to make choroidal atmounts or put into 30% sucrose solutions to make frozen sections.

Fluorescence angiography
FA was performed at day 7, after laser to observe the severity of CNV leakage. Firstly, mice were anesthetized with 1% sodium pentobarbital (Guge Biotech, Wuhan, China) i.p. at a volume of 5 μL·g -1 body weight. Secondly, each mouse was injected intraperitoneally with 0.05mL of 10% uorescein sodium (Fluorescite; Alcon, Tokyo, Japan), and fundus angiogram photos were captured at the middle stage (2-3 minutes after dye injection) using a digital fundus camera (Heidelberg Retina Angiograph, Vista, CA).

SD-OCT
The preparation of mice was described in the section of laser-induced Mouse CNV model. The Bioptigen SDOIS (Bioptigen, Inc., Durham, NC) was used in this study, which is a noninvasive imaging Class , Type B, IPXO, continuous operation medical device. The SDOIS apparatus is comprised of a base system as well as an animal imaging mount and rodent alignment stage (AIM-RAS), which contains a SD-OCT hand held probe (HHP). After the HHP lens was situated close to the right eye of the animal, the InVivoVue Clinic application was activated and the scanning began-following setup of subject pro le for image acquisition, we selected the rectangular scanning protocol consisting of a 3 mm by 3 mm perimeter with 1000 A-scans per B-scan with a total B-scan amount of 100. This a modi cation of the recommended parameters of 1.4 mm by 1.4 mm set by the company for performing rectangular scans.

Western blot
To obtain protein in choroids, mice were killed promptly by cervical vertebra dislocation, eyeballs were harvested instantly and put in cold PBS. Corneas, lenses, vitreouse and retina of mouse eyes were removed and the RPE-choroid tissues were put and chopped into homogenate using a tissue chopper (SONICS &MATERIALS INC.; NEWTOWN, USA) in radio-immunoprecipitation assay (RIPA) lysis buffer with Phenylmethanesulfonyl uoride (Beyotime biotechnology, China).

Cell viability assay
The Cell Titer 96 Aqueous One Solution cell proliferation assay (Promega, Madison, WI, USA) was performed. Brie y, RPE cells were seeded in 96-well at-bottomed microliter plates in eight repeat cultures at a concentration of 1 × 10 4 cells/well. After treatment, each well was incubated with 20 μM MTS assay solution for 2 h at 37°C and the absorbance was measured using an enzyme-linked immunosorbent assay plate reader at 490 nm emission wavelength. Cell viability was expressed as the percentage of absorbance in cells with indicated treatments to that in cells with solvent control treatment.
All antibody incubations were performed in a moist chamber. Cells were then washed three times for 10 min in 0.8× PBS, 50 mM NaCl, and 0.5% Triton X 100. Slides were then rinsed in PBS, counterstained with DAPI, mounted in VECTASHIELD, and stored at 4°C in the dark.
Chromosome PNA-FISH Brie y, cells were washed with PBS and 10 mL of fresh culture medium with 60 μL of colcemid (10 ng/mL) added and incubated at 37°C, after which the cells were collected. Cells were then centrifuged at 300 × g. The supernatant was then aspirated. Next, 25 mL of KCl were added and mixed by inverting. A total of 100 μL of fresh xative (methanol/acetic acid = 3/1) was added and mixed. Incubated the tubes in a 37°C for 15 min and centrifuged at 300 × g at 4°C. Then, added 30 mL of fresh xative and incubated overnight. Centrifuged the xed cells and aspirated the xative, leaving 2 mL in the tube. Precooled slides were placed in the humidity chamber and the resuspended cells were added to a slide. We then allowed the slides to dry overnight.

Microscopy
PNA-FISH assays were recorded on an AxioPlan microscope from ZEISS, equipped with a Plan-Apochrom at 63×, NA 1.4, oil immersion lens, and a cooled CCD camera (CoolSNAP HQ, Photometrics). Image acquisition, processing, and analysis software were from MetaMorph (Molecular Devices). Images of immuno uorescence were recorded using a confocal microscope from Leica.

RNA-seq and bioinformatics analyses
RNA-seq was performed according to the manufacturer's guidelines and previous protocols (C-10365, Life Technologies) [28]. RNA deep-sequencing analyses were performed at BGI-Tech (Shanghai, China). For bioinformatics analyses, transcript structure and abundance were estimated using Cu inks software and differential expression analysis was performed using Cuffdiff software. [29] The cutoff value of differential expression gene was: | log2 (fold change) | > 1, p-value < 0.05. Gene ontology (GO) enrichment analysis was performed using DAVID ver. 6.7 (Database for Annotation, Visualization and Integrated Discovery), which is a web-based application (https://david.ncifcrf.gov/)

Real-time PCR (RT PCR) validation
RT-PCR primer sequences were designed using Primer3 web software (version 4.0.0). The primer sequences used are provided in Additional le 2. The GAPDH gene was used to calculate the relative folddifferences based on comparative cycle threshold (2 -ΔΔCt ) values. The RT-PCR procedure was as follows: 1 μL of cDNA in H 2 O was added to 5 μL of 2× SYBR Green buffer, 0.1 μM each primer, and H 2 O to a nal volume of 10 μL. Differences between the two samples were calculated using Student's t-test at a signi cance level of 0.05 in Graphpad Prism 6.0 software. All expression analysis was performed for three biological repeats and the average values of three repeats values were shown in the gures.

Statistical analysis
Based on the univariate test, continuous normal variables were expressed as the mean value ± SD. Parametric variables of normal distribution were analyzed either by the two-tailed t-test or the F test of ANVOA, followed by the Duncan test for each two group comparison. Results were considered signi cant at p < 0.05. Statistical analysis was performed with Graphpad Prism 6.0 software.

Chrysin suppressed laser-induced CNV succesfully
It has been demonstrated that intravitreally injected chrysin could signi cantly inhibit angiogenesis in laser-induced experimental CNV model [12]. However, the precise mechanism of this is unknown. Considering that intravitreally injection may cause other side effects [13], thus we wanted to know whether intragastrically administered chrysin would take positive effect in laser-induced CNV leision. Chrysin was administered intragastrically at 25 mg/kg daily for 3 days or one week after laser-treated, then we observed the CNV areas and CNV thickness via immuno uorescence staining of choroidal atmount, SD-OCT and uorescein angiograghy, respectively (Figs. 1A, C, E). The results showed that chrysin decreased both CNV area and thickness signi cantly (Figs. 1B, D). It indicated that administered intragastrically chrysin did help to inhibit angiogenesis in laser-induced CNV leision.

Chrysin alleviated DNA damage of RPE cells in CNV lesion
Since DNA damage is a common cause of retinal disease. It has been showed that the 8-OHdG level in aqueous humor was signi cantly higher in exudative AMD patients and correlated with macular lesion size [8], which indicated that oxidative DNA damage is associated with CNV. Therefore, we next explored whether chrysin suppressed CNV lesion via alleviating DNA damage. After 7 days intragastrically administered with chrysin at 25 mg/kg daily in laser-induced CNV mice, we compared the expression of γ-H2AX protein by immuno uorescence in CNV lesion, which is a sensitive indicator of double-strand DNA breaks. Interestingly, we found that the expression of γ-H2AX was mainly in RPE cells rather than vascular endothelial cells, and chrysin could signi cantly decrease the number of γ-H2AX foci per cell and the percentage of γ-H2AX positive cell (Figs. 2A-B, Additional le 1). Subsequently, we investigated the level of γ-H2AX expression in retina and choroidal-RPE complex respectively. It showed that the expression of γ-H2AX was mainly originated from choroidal-RPE complex and chrysin could reduce the γ-H2AX expression, which was consistent with the expression of VEGFA expression. These results indicated that chrysin could compromise DNA damage in CNV lesion and stress-induced DNA damage of RPE cells played an important role in this pathogenesis.
Chrysin rescued stress-induced DNA damage, especially telomere deletion, to suppress in ammation in RPE cells in vitro To further con rm the effect of chrysin on stress-induced DNA damage in RPE cells, we used the in vitro model of RPE cells with photosensitization of A2E, which has been showed to extremely re ect stressinduced DNA damage level and could mimic the early pathogenesis of AMD [9]. To determine the effect of chrysin on the viability of RPE cells, cells were incubated with 3µmol/L chrysin for 24 hours rstly and then cell viability was examined by MTT assay. It showed that the viability of RPE cells increased with chrysin incubation under photosensitization of A2E, and the result was consistent with the treatment of NAC, which has demonstrated to reduce DNA damage effectively (Fig. 3A).
Further, we explored whether chrysin could compromise the stress-induced DNA damage based on the level of γ-H2AX expression by western blotting. We observed a decrease in the level of γ-H2AX in RPE cells with 3 µmol/L chrysin incubation under photosensitization of A2E, which is consistent with the effect of NAC (Fig. 3B). Since telomeres, located at the end of chromosome, are special repeat DNA fragments, which are particularly sensitive to stress. We have demonstrated photosensitization of A2E induced telomere loss (both single and double strand) in the previous study. Thus, we explored whether chrysin could have protective effect on telomere deletion. We evaluated the telomere deprotection by monitoring co-localization of the shelterin TRF1 (used as a telomere marker) with γ-H2AX (named TIF for telomere dysfunction-induced foci). It showed that chrysin alleviated the number of γ-H2AX foci per nucleus. Importantly, chrysin decreased the number of TIF per nucleus (Figs. 3C-E). Then, we examined telomeres for abnormalities using the telomeric peptide nucleic acid (PNA) probe in metaphase spread staining. Interestingly, chrysin signi cantly rescued telomere erosion caused by photosensitization of A2E (Figs. 4A-B).
Since chrysin successfully decreased DNA damage (especially telomere damage), which is associated with cell senescence. Meanwhile, it has been demonstrated that photosensitization of A2E could accelerate RPE senescence in our previous study [9]. Then, we determined whether chrysin could compromise RPE senescence under photosensitization of A2E. So, we monitored SA-β-galactosidase staining as a marker for cellular senescence. Our results showed that chysin signi cantly decreased the percentage of SA-β-galactosidase positive cell (Figs. 4C-D). Considering that cellular senescence usually coupled with the secretion of various pro-in ammatory molecules (known as the senescence associated secretory phenotype (SASP)), we evaluated the expression of pro-in ammatory cytokines by RT-PCR. As expected, we found that chrysin decreased the expression of IL6 and VEGF (Figs. 4E-F). We conclude that chrysin compromise stress-induced DNA damage, especially telomere erosion, simultaneously alleviate secretion of pro-in ammatory cytokines. Also, this may indicated that chrysin may play a role in reduction of in ammatory response.

thus inhibited laser-induced angiogenesis in vivo
To investigate the mechanisms involved in the anti-angiogenesis effect of chrysin, we explored how chrysin in uences gene expression in CNV lesion using RNA-Seq and bioinformatics analyses. We identi ed 897 genes that were differentially expressed with chrysin treatment, including 149 genes up regulated and 748 genes down regulated. To understand the involved pathway of the differentially expressed genes (DEGs), we performed KEGG pathway annotation. We found that the "IL-17 signaling pathway" and "cytokine-cytokine receptor interaction" pathway were signi cantly enriched (Fig. 5A).
Interestingly, it has been found the signi cantly increase of IL-17 in the sera of AMD patients, suggesting that IL-17 may contributes to CNV and AMD [14]. Hence, we focused on the genes enriched to the IL-17 signaling pathway, which was shown in Fig. 5B. Further, we selected 9 genes for validation by RT-qPCR. The results were agreement with those found in the RNA-Seq analysis (Fig. 6A). The transcript level of CCL17 was markedly increased in CNV lesion and decreased with chrysin treated, which indicated that CCL17 may play role in CNV pathogenesis. To our knowledge, this was the rst study that reported the possible relationship of CCL17 and CNV, but the underlining mechanism need to be further investigated. As expected, the expression of IL17A, IL1β and VEGF was increased signi cantly and the increasement was suppressed after chrysin treatment, which is consistent with the precious study.
To increase our understanding of the mechanism evolved in suppressing in ammation effect of chrysin, we further explored the transcription regulators. It has been demonstrated that DNA damage response induces in ammation by inhibiting GATA4 via activating transcription factor NFκB [15]. Thus, we explored the expression of phospho-p65 (Ser536) and total p65. Surprisingly, the expression of p65 changed slightly both in CNV and chrysin treated group (Fig. 5C). Otherwise, the expression of phospho-p65 was signi cantly increased in CNV lesion, but changed lightly after chrysin treatment. These results suggested that chrysin may effect slightly on NFκB translocation in laser-induced CNV lesion. Notably, CNV generation was accompanied by STAT3 activation [16] and one previous study had indicated the crosstalk between STAT3 and IL17 pathway [17]. We further tested the expression level of p-STAT3, and it showed a remarkably decrease after chrysin treatment (Fig. 5C). These results suggest that chrysin downregulated IL17 pathway mainly through inactivation of STAT3.

Discussion
In this study, we found that chrysin suppressed angiogenesis by alleviating DNA damage mainly in RPE cells, further decreasing the secretory of proin ammatory cytokines and improving the disturbed immunohomeostasis in retina. Our results indicate that RPE cell dysfunction plays an important role in the pathogenesis of CNV. In addition, chrysin decreased DNA damage, especially telomere erosion, resulting in the percentage of senescence cell reduced and alleviating the parain ammation in a stress-induced RPE damage model in vitro. Also, our results suggest that chrysin mainly down regulate the IL17 pathway in this in ammatory imbalance environment. Totally, our results emphasize the critical role of interplay between DNA damage, RPE cell dysfunction, in ammatory imbalance and angiogenesis in CNV development. Besides, chrysin may be of promising therapeutic value for the treatment of neovasvular diseases.
Over the past several years, observation shows that activation of other pathogenic pathways have compromise the bene cial of anti-VEGF agents, resulting in poor response or no response [1]. Indeed, apart from VEGF, multiple factors involve in pathological angiogenesis, including a combination of parain ammation as well as heightened in ammasome activation and chronic in ammatory responses [18,19]. RPE cells are critically important in maintaining retinal immuno-homeostasis, which located as a monolayer of polarized cell, sustaining the outer blood-retinal barrier while it regulates nutrient and oxygen delivery to the outer retina and removal of metabolic waste from the photoreceptors [20]. Our previous study has shown that dysfunctional RPE cells create para-in ammation environment [9]. In previous study, we found that acute injury of RPE cells increase the secretory of pro-in ammatory cytocins through activation of STAT3 and NFκB. These suggest that besides vascular endothelial cells, the role of RPE dysfunction in pathological angiogenesis must be taken into account. Thus, to better understanding the mechanism underling the interplay of RPE dysfunction and in ammation response may further help to identify new and better antiangiogenic regents.
In this work, we found that DNA damage is a notable inducer of RPE alterations in cellular phenotypes. In our study, we demonstrated that the expression of γH2AX was signi cantly increased both by westernblot and immuno uorescence mainly in RPE cells in CNV lesion. Also, chrysin successfully decreased the level of DNA damage. Particularly, we found chrysin could especially protect telomere from deletion in stress-induced RPE cells. To our limited knowledge, this is the rst report that identi ed the antiangiogenesis effect of chrysin in laser-induced CNV model through alleviating DNA damage in RPE cells. Importantly, we rst prove the protective effect of chrysin on telomere. In the previous study, chrysin was identi ed to be able to drastically deprotect telomeres against DNA damage response in a highthroughput screening assay for drugs altering telomeres, which is inconsistent with our results [21]. In detail, we found the concentration of chrysin was extremely different, which in our research the dose of chrysin was much lower. This indicates that the effect of chrysin on telomeres may be dose-dependent, which need to be further evaluated.
Apart from the effect of DNA damage on RPE cell phenotype, it has been found that DNA damage of RPE cells was the primary cause of disturbed immuno-homestasis. This nding is substantiated by several lines of evidence. One recent study indicated that DNA damage is associated with in ammation and parain ammation, which is an adaptive response of the immune system to low levels of tissue stress [22]. It has been showed that DNA damage induced a secretory program in quiescent TME, including proin ammation cytokines IL-1β, IL-6 and IL-8, which fostered adverse cancer phenotypes [23]. Additionally, DNA damage response induced in ammation and senescence by inhibiting GATA4, resulting in NFκB activation and SASP induction [15]. In our study, results of RNA-Seq revealed that acute damage of RPE induced secretion of numerous proin ammation cytokines in CNV lesion and chrysin dramatically down-regulated these cytokines by inactivation of STAT3. In vitro, chrysin compromised cellular senescence, simultaneously, decreased pro-in ammatory cytokines in stress-induced DNA damage of RPE cells. These results underlined the protective effect on immuno-homeostasis of chrysin.
Especially, KEGG pathway analysis showed that chrysin mainly suppresses the IL17 pathway in laserinduced CNV. And we con rmed chrysin could signi cantly decrease the expression of CCL17, IL17A and other related genes. The previous study showed that IL17 is signi cantly increased in human eyes with AMD [14], and blocking IL17 in eyes of mice was found to be neuroprotective [24]. Recent publications have indicated that the IL17 present in the eye during age-dependent degeneration as well as in mouse choroidal neovascularization (CNV) generated by γδT-rather than Th17-cells [25]. Beside, IL17 has been shown to increase VEGF production and VEGF can promote IL17 producing γδT-cell accumulation [25]. In the tumor microenvironment, Chung et al has shown that IL-17A is responsible for mediating resistance to the antiangiogenic effects of VEGF blockade [26]. Totally, these evidences indicate that IL-17 pathway may play critical role in the disturbed immuno-homeostasis environment and pathological angiogenesis in laser-induced CNV.

Conclusions
In conclusion, we reveals that the DNA damage of RPE cells is responsible for pathological angiogenesis, causing disturbed immuno-homeostasis and pro-angiogenic environment, indicating the important crosstalk between DNA damage, RPE dysfunction, in ammatory imbalance and angiogenesis play a critical role in pathology of laser-induced CNV. Our results also suggest that chrysin may be a promising therapeutic supplement for the treatment of CNV.     Validation of the expression of VEGF in RPE cells treated with 3 μM chrysin or 25 μM A2E under photosensitization by RT-qPCR . ** indicates p value < 0.01, *** indicates p value < 0.001, **** indicates p value < 0.0001 compared to the control. The experiment was performed independently at least three repeats. F. Validation of the expression of the pro-in ammatory IL6 in RPE cells treated with 3 μM chrysin or 25 μM A2E under photosensitization by RT-qPCR. ** indicates p value < 0.01 compared to the control.
The experiment was performed independently at least three repeats.