IL-17A Damages the Blood–Retinal Barrier through Activating the Janus Kinase 1 Pathway

Blood–retinal barrier (BRB) dysfunction underlies macular oedema in many sight-threatening conditions, including diabetic macular oedema, neovascular age-related macular degeneration and uveoretinitis. Inflammation plays an important role in BRB dysfunction. This study aimed to understand the role of the inflammatory cytokine IL-17A in BRB dysfunction and the mechanism involved. Human retinal pigment epithelial (RPE) cell line ARPE19 and murine brain endothelial line bEnd.3 were cultured on transwell membranes to model the outer BRB and inner BRB, respectively. IL-17A treatment (3 days in bEnd.3 cells and 6 days in ARPE19 cells) disrupted the distribution of claudin-5 in bEnd.3 cells and ZO-1 in ARPE19 cells, reduced the transepithelial/transendothelial electrical resistance (TEER) and increased permeability to FITC-tracers in vitro. Intravitreal (20 ng/1 μL/eye) or intravenous (20 ng/g) injection of recombinant IL-17A induced retinal albumin leakage within 48 h in C57BL/6J mice. Mechanistically, IL-17A induced Janus kinase 1 (JAK1) phosphorylation in bEnd.3 but not ARPE19 cells. Blocking JAK1 with Tofacitinib prevented IL-17A-mediated claudin-5 dysmorphia in bEnd.3 cells and reduced albumin leakage in IL-17A-treated mice. Our results suggest that IL-17A can damage the BRB through the activating the JAK1 signaling pathway, and targeting this pathway may be a novel approach to treat inflammation-induced macular oedema.


Introduction
The blood-retinal barrier (BRB) segregates the neuroretina from the periphery, thereby protecting the retina from exogenous pathogen invasion and systemic inflammatory disturbances. The BRB consists of the inner BRB (iBRB), i.e., tight junctions between retinal endothelial cells and outer BRB (oBRB), i.e., tight junctions between retinal pigment epithelial (RPE) cells. BRB dysfunction may result in the infiltration of circulating immune cells and the leakage of fluid into the neuroretina, causing retinal oedema. When oedema occurs in the macula (known as macular oedema), it can severely damage visual function. BRB dysfunction underpins many sight-threatening retinal diseases, such as diabetic macular oedema (DMO) [1], age-related macular degeneration (AMD) [2,3] and uveitis [4,5].
The mechanisms underpinning BRB dysfunction are complicated and incompletely understood. Inflammation is known to play an important role, and intravitreal injection of steroids, such as triamcinolone [6][7][8][9] and intravitreal steroid implants (e.g., Ozurdex [10] and Retisert [11][12][13]), has proven to be effective in controlling macular oedema caused by diabetic retinopathy (DR) or uveitis. However, these therapies often cause severe adverse effects, such as steroid-induced glaucoma and cataract [14][15][16]. Abnormal levels of vascular endothelial growth factor (VEGF) critically contribute to BRB leakage [17], and intravitreal injection of VEGF inhibitors (e.g., anti-VEGF neutralizing antibodies) is the standard of care for macular oedema in various conditions (e.g., DR, neovascular was sufficient for tight junction dysmorphia and barrier dysfunction [48]. Then 30 min treatment with the same dose was used to examine JAK1 phosphorylation in response to IL-17A. The bEnd.3 cells were treated with murine recombinant IL-17A (R&D Systems, Catalog No. 7956-ML-025/CF), at a concentration of 100 ng/mL. Three days of treatment with 100 ng/mL IL-17A was sufficient for tight junction dysmorphia and barrier dysfunction [49]. Thirty minutes of treatment with the same dose was used to examine JAK1 phosphorylation in response to IL-17A. Tofacitinib Citrate (Catalogue No. PZ0017, Sigma-Aldrich, St. Louis, MO, USA) (25 mg) was dissolved in 100 µL DMSO and further diluted in PBS, immediately before use, to 2.5 ug/mL (4.955 µM). Vehicle control for Tofacitinib was DMSO diluted 1:100,000 in PBS, i.e., 0.00001% DMSO.

Transepithelial Electrical Resistance (TEER)
The effects of IL-17A on TEER of both bEnd. 3 [48] and bEnd.3 [50] as described previously, following IL-17A treatment. Then 200 µL FITC-Dextran or FITC-Na (1 mg/mL) diluted in media was added to the apical chamber of transwell inserts. Fluorescence intensity of apical and basal chambers were measured at 495-570 nm by using a POLARstar Omega plate reader (BMG Labtech, Baden-Wurttemgerg, Germany). FITC concentrations were interpolated from the standard curve. Diffusion rate calculation was adapted from a previously published formula [48].

Immunocytochemistry
Following IL-17A treatment, with or without Tofacitinib Citrate, ARPE-19 and bEnd.3 cells were stained for rabbit anti-ZO-1 and rabbit anti-claudin-5, respectively, to examine tight junction alterations. Both cell types were stained with Rabbit Phospho-JAK1 (Tyr1034, Tyr1035) antibody to examine JAK1 phosphorylation, following IL-17A treatment. All the above antibodies were from Thermo Fisher. Antibody details are shown in the Table 1.

Animal Care and Housing
C57BL6/J mice aged 2 months old, of both sexes, were used for these studies; n ≥ 5 animals were assigned per experimental group. Mice were maintained in the Biological Services Unit at Queen's University Belfast, with free access to food and water on a 12 h light/dark cycle, in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. All procedures were approved by the UK Home Office Animals (Scientific Procedures) Act 1986 and the local animal welfare ethical review board of Queen's University Belfast (DOH PPL2876).

IL-17A Intravitreal and Intravenous Injection
Two-month-old C57BL/6J mice were injected intravitreally (ivt, 20 ng/µL/eye, n = 20 mice animals per group) or intravenously (i.v. 20 ng/g, n = 6 animals per group) with murine recombinant IL-17A (R&D Systems). Forty-eight hours after, animals were subjected to Micron IV examination, and eyes were collected for immunostaining or retinal protein was extracted for Western blot (see details below). For Tofacitinib Citrate treatment, 2-month-old C57BL6/J mice were treated with Tofacitinib Citrate (15 mg/kg, i.p.) immediately after IL-17A i.v. injection. Tofacitinib Citrate was administered again 24 h later, and mice were culled at 48 h (Appendix A Figure 1). Eyes were processed for Western blot for albumin (marker of BRB leakage) and immunostaining for pJAK1 and albumin.

Albumin Leakage Quantification
Paraffin-embedded eyes were sectioned at 5 µm thickness. De-waxing was carried out by immersing slides in 3 changes of clearene for 5 min each, followed by 3 changes of 100% ethanol for 3 min each, and followed by 5 min in running water. Citraconic anhydride (Sigma, Catalog No. 125318), pH 7.4, at 95 • C, for 30 min, was used for antigen retrieval. Slides were incubated with goat anti-albumin and Biotinylated Griffonia Simplicifolia Lectin I Isolectin B4 (Table 1), overnight, at 4 • C. The next day, slides were washed in PBS prior to incubation with appropriate secondary antibodies. Slides were mounted with DAPI-Vectashield and imaged by using a Leica DMi8 epifluorescence microscope. Images were analyzed by using FIJI (NIH); Isolectin B4-positive ROIs were restored on the albumin channel and measured prior to whole neuroretina measurements. Leakage ratio was calculated as follows: Leakage Ratio = (Extravascular albumin)/(Total albumin in neuroretina)

pJAK1 Quantification Neuroretina
Paraffin-embedded eyes were treated as above. After blocking, pJAK1 antibody was applied at a concentration of 1:100, at 4 • C, overnight. After several washes, secondary antibody donkey anti-rabbit 594 (Stratech Scientific Ltd, Ely, UK) was applied at 1:300 for 2 h. Slides were washed, mounted with DAPI-vectashield and imaged as above. Then pJAK1 in the neuroretina was quantified by using FIJI.

Western Blot
Samples were homogenized in RIPA buffer containing protease inhibitor cocktail (Sigma) and PhosSTOP™ (Roche). Protein concentrations were normalized by using Pierce BCA protein assay kit (Thermo Fisher, Catalog# 23225). Protein samples of equal concentrations were denatured by using NuPAGE™ 10X Reducer and LDS Sample Buffer (4X) (Thermo Fisher, Catalog No. NP0004, NP0007) as per the manufacturer's instructions. Then 10 µg (retinal protein for albumin detection) or 35 µg (cell protein for phospho-JAK1 detection) was loaded on 10% acrylamide gels. Proteins were transferred to PVDF membranes (Immun-Blot, Bio-Rad Laboratory, Hercules, CA, USA), using the wet transfer method. Membranes were blocked with 5% Bovine Serum Albumin (Sigma-Aldrich, Catalog No. A3803) for 1 h, before primary antibody incubation overnight. The next day, membranes were washed, incubated in appropriate secondary antibody washed and developed by using Clarity Western ECL Blotting Substrate (Bio-Rad, Catalog No. 1705061). Membranes were imaged by using G:BOX Chemi XRQ chemiluminescence imager (Syngene, Cambridge, UK). Secondary antibody probing was used to confirm successful stripping. Densitometry was analyzed by using FIJI software.

Statistical Analysis
Graph generation and statistical analyses were performed by using GraphPad Prism 9 (GraphPad Software Inc. San Diego, CA, USA). The differences between two groups were compared by using unpaired t-test. For the difference between three or more groups, we used One-Way ANOVA, followed by Tukey's (to compare the means of each groups with the means of every other group) or Dunnett's (to compare the means of each treatment group with the mean on of the control group) multiple comparison post hoc tests.

The Effect of IL-17A on bEnd3 Cell and ARPE19 Cell Tight Junctions and Barrier Function
Claudin-5 junctions are integral to endothelial cell tight junctions. Under normal culture conditions, claudin-5 was evenly distributed around the cell-cell junction of bEnd3 cells ( Figure 1A). IL-17A treatment (72 h) induced fragmented and clustered claudin-5 expression in bEnd.3 cells (arrowheads, Figure 1A). Measurement of trans-epithelial electrical resistance (TEER) in bEnd3 cells cultured in the transwell membrane showed that 72 h of treatment with IL-17A significantly reduced TEER ( Figure 1B). FITC-Na permeability showed a pattern indicative of leakage, but was not statistically significant in IL-17A-treated bEnd3 cells ( Figure 1C).
A previous study reported ZO-1 junction dysmorphia in ARPE-19 cells following 6 days of IL-17A treatment [48]. This was confirmed in our study ( Figure 1D). Six days of IL-17A treatment induced fragmented ZO-1 junctions and positive staining in the cytosol and nuclei in ARPE-19 cells (arrows, Figure 1D). The treatment also significantly reduced TEER ( Figure 1E) and increased FITC-Dextran (4 kDa) permeability ( Figure 1F) in ARPE19 cells cultured in transwell membranes. Together, these data suggest that IL-17A alone can induce both vascular endothelial cell (iBRB) and retinal pigment epithelial cell (oBRB) dysfunction in vitro.

The Effect of IL-17A on BRB Integrity In Vivo
Wild-type C57BL/6J mice received IL-17A (ivt, 20 ng/µL/eye) or (i.v., 20 ng/g), and BRB function was examined 48 h later. Western blot showed significantly higher levels of albumin in the retinae from IL-17A ivt treated mice compared to mice that received no injection (NI) (Figure 2A). Retinal section immunostaining also showed higher levels of albumin expression in IL-17A ivt mice ( Figure 2B). Micron IV examination revealed FITC-Dextran (4 kDa) extravasation in mice received IL-17A ivt injection compared to control PBS ivt mice (arrows, Figure 2C).

The Effect of IL-17A on pJAK1 Expression In Vitro and In Vivo
To understand the signaling pathways involved in IL-17A-induced BRB damage, we investigated the JAK/STAT pathway, as it is a master regulator of cytokine receptor signaling. As the commercially available JAK/STAT array kit was only for human tissues, we conducted a human JAK/STAT array assay in IL-17A-treated ARPE19 cells. The results show that 30 min treatment with IL-17A increased the expression of SHP1, pJAK1, pTYK2 and pSTAT6 in ARPE19 cells (Appendix A Figure A2). Further Western blot and immunocytochemistry showed that IL-17A treatment significantly upregulated pJAK1 expression in bEnd.3 cells (Figure 3A-C) but not in ARPE19 cells (Appendix A Figure 3). However, pSTAT3 expression was increased in IL-17A-treated ARPE19 cells (Appendix A Figure A4). As IL-6 and VEGF can also active the JAK/STAT pathway, we measured their concentrations in the supernatants from control and IL-17A-treated bEnd.3 cells. We found that the 3-6 days of IL-17A treatment appeared to upregulate IL-6, but not VEGFA production (Appendix A Figure A5).
Injection (i.v.) of IL-17A upregulated pJAK1 expression in mouse retina, particularly the inner layers (ganglion cell layer and inner plexiform layer, Figure 3D). The location of pJAK1 positivity in the inner retina and the fact that in vitro treatment of ARPE19 cells (oBRB) with IL-17A did not seem to increase pJAK1 expression (Figure 3) led us to focus our study on the role of JAK1 in IL-17A-mediated iBRB damage, rather than oBRB damage.

The Effect of Tofacitinib Citrate in IL-17A-Mediated Retinal pJAK1 Expression
Having shown that i.v. injection of IL-17A upregulated pJAK1 expression in mouse retina ( Figure 3A), we moved to investigate if the upregulation of pJAK1 could be reduced by JAK1/3 inhibitor Tofacitinib Citrate. IL-17A-or vehicle-injected mice were administered with Tofacitinib Citrate (15 mg/kg) once a day, for two days, and retinal pJAK1 expression was evaluated 48 h later ( Figure 5A). The results confirmed the upregulation of pJAK1 by IL-17A ( Figure 5B,C), Similar levels of pJAK1 expression were present in IL-17A and IL-17A+vehicle groups. Tofacitinib Citrate treatment slightly reduced IL-17A-mediated retinal pJAK1 expression; however, the reduction did not reach statistical significance ( Figure 5C), which may be related to the large variation and insufficient sample size.

The Effect of Tofacitinib Citrate on IL-17A-Mediated BRB Leakage In Vivo
To understand the functional role of JAK1 activation in IL-17A-mediated iBRB dysfunction, mice were treated with Tofacitinib immediately after IL-17A injection, and again 24 h later ( Figure 6A). Retinal vascular leakage was evaluated 48 h later by immunostaining of al-bumin. Extravascular albumin was successfully reduced in Tofacitinib Citrate-treated animals compared to those treated with vehicle control and IL-17A, or IL-17A alone ( Figure 6B,C), indicating that JAK1 is functionally involved in iBRB leakage in IL-17A-induced BRB dysfunction.  Values were normalized to PBS control on the same slide. Mean ± SD. ROUT outliers test 2% was used to remove outliers. Cleaned data were compared by using One-Way ANOVA with Tukey's multiple comparisons test, * p < 0.05; ns: no statistical significance. Values were normalized to PBS control on the same slide. Mean ± SD; n ≥ 5 animals per group; ROUT outliers test 2% was used to remove outliers. Cleaned data were compared by using One-Way ANOVA with Tukey's multiple comparisons test, * p < 0.05; ns: no statistical significance.

Discussion
In this study, we showed that IL-17A can induce BRB dysfunction both in vitro and in vivo. Mechanistically, IL-17A damages the structure of tight junctions in both endothelial and RPE cells. In endothelial cells, the barrier destructive effect of IL-17A was mediated by activating the JAK1 signaling pathway. We further found that the JAK1 antagonist Tofacitinib Citrate could effectively ameliorate IL-17A-mediated BRB leakage. Tofacitinib is an FDA-approved anti-inflammatory medication, used to treat autoimmune diseases such as rheumatoid arthritis [49] and ulcerative colitis [51][52][53]. Our results suggest that Tofacitinib or other JAK1 inhibitors may be re-purposed for the management of IL-17A mediated macular oedema.
This study provides novel evidence of IL-17A-mediated pJAK1 activation in iBRB cells (vascular endothelial cells) and the murine retina. The signal transduction of sev-eral cytokines, including IL-17A, IL-6 and VEGF, have been reported to converge in the JAK/STAT pathway [71,72]. We found that 3-6 days of treatment with IL-17A could increase IL-6 but not VEGF production in bEnd.3 cells, indicating that IL-17A can active the JAK/STAT pathway directly and indirectly under chronic disease conditions, such as in DR and uveoretinitis. Many reports have delineated the role of the JAK/STAT pathway in retinal diseases involving BRB dysfunction, namely in DR [73,74], AMD [75,76] and uveoretinitis [77,78]. This finding is in agreement with studies which showed that Tofacitinib Citrate is effective in refractory anterior and intermediate uveitis and scleritis [79,80] and with another study, which showed it is protective against cytokine-mediated barrier dysfunction in the gastrointestinal tract [81]. Therefore, targeting the JAK/STAT pathway is a strategy that is worthy of exploration for the management of retinal diseases involving BRB dysfunction.
The limitations of our study include the use of cell lines ARPE-19 as oBRB and bEnd.3 as iBRB models, and that these models do not take into account of the effect of other supporting cells in the neurovascular unit, such as pericytes and Muller cells. Previous studies have shown that IL-17A can activate pericytes in vitro [82] through IL-17RA/C and induce endothelial basement membrane remodeling [83]. Future studies using primary or iPS-derived RPE or retinal endothelial cells together with pericytes and/or Muller cells in 3D culture models will be helpful to better understand the mechanism of IL-17A-induced BRB damage.

Conclusions
Our results suggest that IL-17A can damage the BRB through activation of the JAK1 signaling pathway. Since JAK1 is also involved in the signal transduction of other barrierdamaging cytokines, such as IL-6 and VEGF, targeting this pathway may be a novel approach for the management of macular oedema, particularly those who are resistant to anti-VEGF therapy.

Data Availability Statement:
The data presented in this study are all contained within the main body of this article.

Acknowledgments:
The authors thank Eszter Emri, Jose Romero, Peter Barabas and Joana Sa Pessoa for helpful discussions during this project. The authors acknowledge the BSU at Queen's University Belfast for their assistance in animal maintenance.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.