biomedicines Article Effects of the Topical Administration of Semaglutide on Retinal Neuroinflammation and Vascular Leakage in Experimental Diabetes Rafael Simó 1,2, * , Patricia Bogdanov 1,2 , Hugo Ramos 1,2 , Jordi Huerta 1 , Olga Simó-Servat 1,2 and Cristina Hernández 1,2 1 Diabetes and Metabolism Research Unit, Vall d’Hebron Research Institute, Universitat Autònoma de Barcelona, 08035 Barcelona, Spain; patricia.bogdanov@vhir.org (P.B.); hugo.ramos@vhir.org (H.R.); jordi.huerta@vhir.org (J.H.); olga.simo@vhir.org (O.S.-S.); cristina.hernandez@vhir.org (C.H.) 2 Centro de Investigación Biomédica en Red de Diabetes y Enfermedades Metabólicas Asociadas (CIBERDEM), Instituto de Salud Carlos III (ISCIII), 28029 Madrid, Spain * Correspondence: rafael.simo@vhir.org Abstract: Background: An unexpected increase in the rate of severe diabetic retinopathy was ob- served in the Semaglutide in Subjects with Type 2 Diabetes (SUSTAIN)-6 clinical trial. Although this effect was attributed to a rapid decrease in blood glucose levels, a direct deleterious effect of semaglutide on the retina could not be ruled out. In order to shed light on this issue, we have per- formed a study aimed at testing the direct effect of semaglutide administered by eye drops on retinal neuroinflammation and microvascular abnormalities using the db/db mouse model. Methods: Eye  drops containing semaglutide (0.33 mg/mL; 5 µL once/daily) or vehicle (PBS; 5 µL once daily) were  administered for 15 days. Results: We found that semaglutide significantly reduced glial activation, Citation: Simó, R.; Bogdanov, P.; as well as the retinal expression of Nuclear factor kB (NF-κB), proinflammatory cytokines (IL-1β, Ramos, H.; Huerta, J.; Simó-Servat, O.; IL-6, IL-18) and Intercellular Adhesion Molecule (ICAM)-1. In addition, semaglutide prevented the Hernández, C. Effects of the Topical apoptosis of cells from the retinal ganglion layer and activated the protein kinase B (AKT) pathway. Administration of Semaglutide on Finally, a dramatic decrease in vascular leakage was observed in db/db mice treated with semaglu- Retinal Neuroinflammation and tide. All these findings were observed without any change in blood glucose levels and, therefore, can Vascular Leakage in Experimental be directly attributed to semaglutide. Conclusions: These experimental findings point to a beneficial Diabetes. Biomedicines 2021, 9, 926. rather than a deleterious effect of semaglutide on the retina of subjects with diabetes. https://doi.org/10.3390/biomedicines9080926 Academic Editor: Erich Knop Keywords: diabetic retinopathy; semaglutide; retina; db/db mouse Received: 25 June 2021 Accepted: 27 July 2021 Published: 31 July 2021 1. Introduction Neurodegeneration has emerged in recent years as a key element involved in the Publisher’s Note: MDPI stays neutral pathogenesis of early stages of diabetic retinopathy (DR) [1,2]. In fact, the American Dia- with regard to jurisdictional claims in betes Association (ADA) has defined DR as a highly specific neurovascular complication published maps and institutional affil- and, therefore, it is no longer acceptable to consider DR as merely a microvascular com- iations. plication [3]. However, the Consortium for the Early Treatment of Diabetic Retinopathy (EUROCONDOR) study has shown that a significant proportion of patients present retinal microangiopathic abnormalities without any sign of neurodegeneration [4,5]. This finding suggests that neurodegeneration does not always play a role in the development of DR, Copyright: © 2021 by the authors. and points to drugs that protect both neurons and microvessels as the best candidates Licensee MDPI, Basel, Switzerland. for treating early stages of DR. In recent years, experimental evidence has shown that This article is an open access article glucagon-like peptide 1 (GLP-1) and its analogues (GLP-1R agonists) are among those distributed under the terms and drugs with a dual effect: neuroprotective and vasculotropic action [6,7]. conditions of the Creative Commons However, the results of the Semaglutide in Subjects with Type 2 Diabetes (SUSTAIN) Attribution (CC BY) license (https:// 6 clinical trial, seemed to contravene the beneficial effects of GLP-1R agonists observed creativecommons.org/licenses/by/ in preclinical studies [8]. The SUSTAIN-6 clinical trial showed clear beneficial effects of 4.0/). Biomedicines 2021, 9, 926. https://doi.org/10.3390/biomedicines9080926 https://www.mdpi.com/journal/biomedicines Biomedicines 2021, 9, 926 2 of 10 semaglutide (a GLP-1R agonist) in terms of cardiovascular disease, but an unexpectedly higher rate of advanced DR was observed in patients treated with semaglutide in compar- ison with those treated with placebo. Among the potential explanations of this finding, the rapid lowering of blood glucose levels was the most plausible. In fact, a post-hoc analysis confirmed this hypothesis and concluded that the magnitude of the reduction of HbA1c and the presence of pre-existing DR were the main factors involved in the risk of DR worsening [9]. Nevertheless, a direct deleterious effect of semaglutide on the retina could not be ruled out. In order to shed light on this issue, we performed an experimental study aimed at testing the direct effect of semaglutide on early stages of DR using eye drops. The rationale for using eye drops was that we had previously demonstrated that by this route GLP-1R agonists reach the retina without modifying blood glucose levels. Therefore, with this approach, we could examine the direct action of semaglutide independently of its effect on glycemic values. The specific aim was to evaluate the effect of semaglutide administered by eye drops on retinal neurodegeneration, neuroinflammation and early microvascular abnormalities (vascular leakage) in an experimental mouse model. 2. Materials and Methods 2.1. Animals A total of 20 male db/db (BKS.Cg-Dock7m +/+ Leprdb/J) mice were purchased from Charles River Laboratories (Calco, Italy). Ten non-diabetic male mice (db/+) matched by age served as a control group. The animals were randomly housed under tight en- vironmental conditions of humidity (60%), temperature (20 ◦ C) and cycles of 12h/12h light/darkness. They had free access to filtered water and “ad libitum” food (ENVIGO Global Diet Complete Feed for Rodents, Mucedola, Milan, Italy). All animal procedures were conducted in accordance with the Animal Care and Use Committee of VHIR (Vall d’Hebron Research Institute, Passeig de la Vall d’Hebron 119–129, Barcelona, Spain): approval code 75/15, approval date 2 December 2015. In addition the study was performed following the recommendations of the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research. 2.2. Interventional Study When the diabetic mice were aged 10 weeks, semaglutide eye-drops (n = 10) or vehicle eye drops (n = 10) were randomly administered. For this purpose, a micropipette (Eppendorf Research® plus pipette 0.5–10 µL Ref. 3123000098; Hamburg, Germany) was used. One drop (5 µL) of semaglutide (0.33 mg/mL), or vehicle (5 µL PBS, pH 7.4) was administered directly onto the superior corneal surface once daily for two weeks in each eye. On the last day, one drop of either semaglutide or vehicle was administered to the eyes 1 h before euthanasia. The evaluation of the results was performed by investigators unaware of the treatment received by the mice. 2.3. Neurovascular Damage Assessment Glial activation was evaluated by Laser Scanning Confocal microscopy using specific antibodies against GFAP (Glial fibrillary acidic protein) as previously reported [7,10]. To evaluate the degree of glial activation, a scoring system based on the extent of GFAP staining previously described was used [11]. The scoring system was as follows: Müller cell end-feet region/ganglion cell layer (GCL) only (score 1); Müller cell end-feet region/GCL plus a few proximal processes (score 2); Müller cell end-feet plus many processes, but not extending to the inner nuclear layer (INL) (score 3); Müller cell end-feet plus processes throughout with some in the outer nuclear layer (ONL) (score 4); Müller cell end-feet plus many dark processes from the GCL to the outer margin of the ONL (score 5). Retinal ganglion cells (RGC) death is characteristic of DR. Thus, the number of cells in the RGC layer was quantified by hematoxylin and eosin (H&E) staining. Biomedicines 2021, 9, 926 3 of 10 Retinal vascular permeability was examined by the assessment of albumin leakage from the blood vessels into the retina using the Evans Blue albumin method (ex vivo). Evans Blue was intraperitoneally injected (E2129 SIGMA, Sant Louis, MO, USA): 17 mg/kg body weight, in a concentration of 5 mg/mL dissolved in PBS pH 7.4. Immediately after injection, the animals visibly turned blue, confirming dye uptake and distribution. After 120 min, the mice were euthanized by cervical dislocation and the eyes were enucleated. Flat-mounted slides were obtained, and coverslipped with a drop of mounting medium Prolong Gold antifade reagent (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). Digital images from different random fields of all the retinas were acquired using a confocal laser scanning microscope (FV1000; Olympus, Hamburg, Germany) at 60× using the 561-nm laser line, and each image was recorded with identical beam intensity at a size of 1024 pixels × 1024 pixels. For analysis of the albumin-bound Evans Blue, Z-stack retinal images (step size 1.16 µm) of different regions of the vascular tree were acquired. For quantitative analysis of the albumin-bound Evans Blue, the number of extravasations per field of 60× was counted. 2.4. Mechanism of Action Pro-survival signaling (AKT/p-AKT) and inflammation (NFκB, proinflammatory cy- tokines and ICAM-1) were analyzed by RT-PCR, Western blot, and immunohistochemistry. 2.5. RNA Isolation and Quantitative RT-PCR assay Total RNA from mice was extracted using Trizol® reagent (Invitrogen, Madrid, Spain) according to the manufacturer’s protocol. Then, RNA samples were treated with DNase (Qiagen, Madrid, Spain) to remove genomic contamination and purified on an RNeasy MinElute column (Qiagen, Madrid, Spain). RNA quantity was measured on a Nanodrop spectrophotometer, and integrity was determined on an Agilent 2100 Bioanalyzer. The single-strand cDNA was synthesized as described in Prime Script® RT Master Mix kits. Real-time RT-PCR was performed using SYBR Green PCR Master Mix (Applied Biosystems, Warrington, UK) using the 7.900 HT Sequence Detection System in 384-well optical plates with specific primers displayed in Table 1. Table 1. Primers used for RT- PCR. Gen Symbol Sequence Primer Il 1 β Forward (50 -30 ) 50 -GCAACTGTTCCTGAACTCAACT-30 Il 1 β Reverse (50 -30 ) 50 -ATCTTTTGGGGTCCGTCAACT-30 Il 6 Forward (50 -30 ) 50 -TAGTCCTTCCTACCCCAATTTCC-30 Il 6 Reverse (50 -30 ) 50 -TTGGTCCTTAGCCACTCCTTC-30 Il 18 Forward (50 -30 ) 50 -AGCAGTCCCAACTAAGCAGTA-30 Il 18 Reverse (50 -30 ) 50 - CAGCCAGTAGAGGATGCTGA-30 Icam-1 Forward (50 -30 ) 50 - GGCATTGTTCTCTAATGTCTCCG-30 Icam-1 Reverse (50 -30 ) 50 - TGTCGAGCTTTGGGATGGTAG-30 B2M Forward (50 -30 ) 50 - GGCCCATCTTGCATTCTAGGG-30 B2M Reverse (50 -30 ) 50 - CAGGCAACGGCTCTATATTGAAG-30 Each sample was assayed in triplicate. Applied Biosystems 7900HT Fast Real-Time PCR System (ABI SDS 2.0 RQ, Applied Biosystems, Waltham, MA, USA) software and the 2−∆∆Ct analysis method were used for relative quantification (R.Q.) calculation with Beta-2-Microglobin (B2M) as endogenous control. Biomedicines 2021, 9, 926 4 of 10 2.6. Western Blotting A lysis buffer was prepared (RIPA buffer: phenylmethanesulfonylfluoride [PMSF], 1 mM; Na3VO4, 2 mM; NaF, 100 mM. Sigma, St Louis, MO, USA) with a 1× protease inhibitor cocktail (Sigma, St Louis, MO, USA). Then, proteins were extracted from the neuroretinas in 80 µL of the aforementioned buffer. A total of 25 µg protein was resolved by 10% SDS-PAGE and transferred to a PVDF membrane (Bio-Rad Laboratories, Madrid, Spain). The primary antibodies (Table 2) were incubated overnight at 4 ◦ C. Table 2. Primary antibodies and specific dilutions used in Western Blot analysis. Antibody Description NF-κB (p65) 1:1000; sc-8008; Santa Cruz Biotechnology Inc, Dallas, Texas, USA p-AKT 1:1000; #2965; Cell Signaling, Leiden, The Netherlands AKT 1:1000; #9272; Cell Signaling, Leiden, The Netherlands Cyclophilin A 1:10000; BML-SA296; Enzo, NY, USA Vinculin 1:100,000; sc-73614; Santa Cruz Biotechnology Inc, Dallas, TX, USA The following day, the suitable HRP secondary antibody anti-rabbit or anti-mouse (Dako Agilent, Santa Clara, CA, USA) was incubated for 1 h at room temperature. Cy- clophilin (CYP) and vinculin were used to normalize protein levels. Densitometric analysis of the western blot bands was performed with Image J software (National Institutes of Health, Bethesda, MD, USA). 2.7. Immunohistochemical Analysis Firstly, paraffined sections were deparaffinized in xylene (VWR, Barcelona, Spain) and rehydrated in ethanol (Sigma, St Louis, MO, USA). Sections were fixed in acid methanol (−20 ◦ C) for 1 min and washed with 0.01 M 4 phosphate-buffered saline (PBS; Biowest, Lab- clinics, Barcelona, Spain) at pH 7.4. Then, sections were incubated in blocking solution (3% BSA, Tween 0.05% PBS; Sigma Aldrich, St Louis, MO, USA) for 1 h at room temperature and afterwards, they were incubated overnight at 4 ◦ C with a specific primary antibody (anti- GFAP rabbit monoclonal; 1:500; ab7260; Abcam, Cambridge, UK). The following day, after washing, sections were incubated with a fluorescent ALEXA 488 as a secondary antibody (anti-rabbit or anti-mouse) (Life Technologies S.A, Madrid, Spain) in blocking solution (Protein Block Serum-Free Ready-To-Use DAKO Agilent X0909, Agilent Technologies, Inc., Santa Clara, CA, USA) for 1 h and washed again. Nuclei were counterstained using Hoechst 33,342 (20 -[4-ethoxyphenyl]-5-[4-methyl-1-piperazinyl]-2,50 -bi-1H-benzimidazole trihydrochloride trihydrate), an organic compound that emits blue fluorescence when bound to DNA (Thermo Fisher Scientific, OR, USA). Finally, samples were mounted in Mounting Medium Fluorescence (Prolong, Invitrogen, Thermo Fisher Scientific, OR, USA) with a coverslip. Images were acquired with a confocal laser scanning microscope (FV1000; Olympus, Hamburg, Germany) at a resolution of 1024 × 1024 pixels. Five fields (three corresponding to the central and two to the peripheral retina) from each section were analyzed using ImageJ software (National Institutes of Health, Bethesda, MD, USA). 2.8. Statistical Analysis The results are expressed as the mean ± Standard Error (SE). Statistical compar- isons were performed with a Student’s unpaired test. When multiple comparisons were performed, one-way ANOVA followed by the Bonferroni test was used. For statistical purposes, the GFAP extent score was categorized as “normal” (scores 1 and 2) and “patho- logical” (scores 3, 4 and 5), and for comparisons between groups the Fisher’s exact test was used. Levels of statistical significance were set at p < 0.05. Biomedicines 2021, 9, 926 5 of 10 Biomedicines 2021, 9, x FOR PEER REVIEW 5 of 12 3. Results We did not find any difference in blood glucose concentrations and body weight Weduring did notthe study find betweenindb/db any difference mice treated blood glucose with semaglutide concentrations and db/db and body weight dur- mice treated ing thewith studyvehicle between (Figure 1). treated with semaglutide and db/db mice treated with db/db mice vehicle (Figure 1). Figure Figure 1. 1. Evolution Evolution of blood of blood glucose (a)glucose and body(a) and body weight (b) in weight (b) in thegroups. the experimental experimental groups. 3.1.ofEffect 3.1. Effect of Semaglutide Semaglutide on Neurovascular on Neurovascular Unit Unit As expected, As expected, in non-diabetic in non-diabetic control micecontrol (db/+)mice GFAP(db/+) GFAP expression wasexpression mainly con- was mainly con- fined to the retinal ganglion cell layer (GCL) (Figure 2a,b). The diabetic mice (db/db) mice (db/db) fined to the retinal ganglion cell layer (GCL) (Figure 2a,b). The diabetic treatedtreated with presented with vehicle vehicle presented significantlysignificantly higher GFAP higher GFAPthan expression expression than non-diabetic non-diabetic mice matched mice matched by age (Fisher’s by age (Fisher’s exact exact test: p < test: p