Treatment with interleukin‐33 is non‐toxic and protects retinal pigment epithelium in an ageing model of outer retinal degeneration

Abstract The leading cause of central vision loss, age‐related macular degeneration (AMD), is a degenerative disorder characterized by atrophy of retinal pigment epithelium (RPE) and photoreceptors. For 15% of cases, neovascularization occurs, leading to acute vision loss if left untreated. For the remaining patients, there are currently no treatment options and preventing progressive RPE atrophy remains the main therapeutic goal. Previously, we have shown treatment with interleukin‐33 can reduce choroidal neovascularization and attenuate tissue remodelling. Here, we investigate IL‐33 delivery in aged, high‐fat diet (HFD) fed mice on a wildtype and complement factor H heterozygous knockout background. We characterize the non‐toxic effect following intravitreal injection of IL‐33 and further demonstrate protective effects against RPE cell death with evidence of maintaining metabolic retinal homeostasis of Cfh+/−~HFD mice. Our results further support the potential utility of IL‐33 to prevent AMD progression.

In this study, we use an insidious aAMD model, aged heterozygous complement factor H knockout (Cfh+/−) mice on a high-fat diet (HFD) 6 to determine the safety of IL-33 treatment and demonstrate its potential as an early therapeutic for preventing AMD progression.

| Whole-mount immunohistochemistry and microscopy
For whole-mounts, posterior eye cups were dissected and fixed in 2% v/v formaldehyde overnight at 4°C. Retina and RPE/choroid were blocked in 5% w/v bovine serum albumin with 0.3% v/v Triton X-100 (Sigma Aldrich) in PBS before staining with TMR Red-dUTP TUNEL (Roche Diagnostics), followed by incubation with ZO-1 antibody (1 in 50; 40-2200; Thermo Fisher Scientific) overnight at 4°C. Tissue was then incubated with appropriate secondary antibody (A-11034; Invitrogen; 1 in 200) and mounted in hard-set medium (Vector Laboratories). Images were captured using a Leica SP5-AOBS confocal laser scanning microscope (Leica Microsystems Ltd.) and TUNEL spots identified using Volocity ® Image Analysis Software 6.0. Data are presented as mean number of TUNEL + cells per field of view (FOV).

| Electron microscopy
For transmission electron microscopy (TEM), eyes were fixed in 2.5% v/v glutaraldehyde initially at RT for 2 hours, before further fixing of posterior cup only at 4°C until processed. Eye cups were fixed in osmium tetroxide, stained with uranyl acetate and dehydrated through a graded ethanol series to 100% ethanol, before incubation with propylene oxide and infiltration with Epon resin. Eyes were then embedded and polymerized at 60°C for 2 days. Thin sections (70-80 nm) were prepared and stained with lead and uranyl salts. Images were captured using a Tecnai T12 microscope (Thermo Fisher Scientific). Across four sections, 55-60 images were captured per eye and fifteen measurements/image were made along the Bruch's membrane using FIJI. 7

| Protein lysate and Western blots
Whole retinas were crushed in 200 µL Pierce ® RIPA buffer (Thermo

| Statistics
Analysis of data used three-way ANOVA in R studio (R version 3.6.1; RStudio, Inc) with independent factors, treatment, genotype and animal (accounting for paired eyes). A Tukey multiple comparisons test was used where significant effects were found. If no significance was identified from further multiple comparisons, the result of three-way ANOVA is presented to demonstrate significant effect of independent factor. Mean sub-RPE deposits were analysed using a paired t test in GraphPad Prism (version 8.1.2; GraphPad Software Inc). All data are expressed as ± SEM.

| RE SULTS AND D ISCUSS I ON
The Cfh+/−~HFD model presents with distinct disease phenotype, without significant cell loss, 6 providing opportunity to measure treatment toxicity and efficacy in a mimic of early/intermediate aAMD. We further determined IL-33 toxicity by TEM analysis of sub-RPE deposits, a key pathological feature in the Cfh+/−~HFD model. 6 Here, we observed deposits in vehicle and IL-33-treated Cfh+/−~HFD eyes ( Figure 1E), with no difference in mean height (n = 3, 3, Figure 1F). Deposit formation in AMD is considered to be driven by a dysregulated immune response and local inflammation. 8 Toxicity was not apparent as no exacerbation of pathological deposit formation was observed with IL-33 treatment.
Sub-RPE deposits in Cfh+/−~HFD mice are thought to drive complement-mediated monocytosis and cytotoxicity, resulting in subsequent RPE damage and atrophy. 6 Indeed, activated monocytes and pro-inflammatory cytokines can induce reactive oxygen species (ROS) production and cell death in RPE. 9 Multiple comparisons did not reveal significant difference in extent of cell death between C57BL/6 ~ HFD vehicle eyes and any other group. Notwithstanding, there was a significant increase in cell death in the RPE of vehicle-treated Cfh+/−~HFD eyes, compared to C57BL/6 ~ HFD IL-33-treated eyes (Figure 2A-B). The rescue of cell death in Cfh+/−~HFD animals with IL-33 was to levels found in C57BL/6 ~ HFD eyes ( Figure 2B). As IL-33 treatment did not affect deposit formation, it supports protection of RPE is through re-balancing of any pro-inflammatory responses or protection from ROS. Certainly, IL-33 promotes a skew to a protective non-inflammatory type 2 immune response in EAU and CNS injury. 10,11 Additionally, our previous work has found IL-33 signalling via ST2 promotes normal bioenergetic source of oxidative phosphorylation (OXPHOS) in RPE, protecting from oxidative stress. TLR-stimulation of RPE induces a switch to glycolysis, essential to subsequent pro-inflammatory cytokine release. 12 (Figure 2A,C). In aAMD, loss of RPE precludes significant PR loss. 13 Whilst there is no effect of IL-33 treatment on cell loss in the retina early in the course, the RPE protection conferred is predicted to prevent cell loss with time.
Our results demonstrate a non-toxic effect for exogenous IL-33 treatment in aged mice. Moreover, we reveal a protective capability of IL-33 against RPE loss and for retina metabolic homeostasis in a dysregulated immune-mediated insidious model of outer retinal degeneration (Cfh+/−~HFD animals). Together, these data highlight the potential of IL-33 treatment to protect during early pathogenesis of AMD.

ACK N OWLED G EM ENTS
The authors would like to thank Judith Mantell and Sally Hobson

CO N FLI C T O F I NTE R E S T
The authors declare no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.