WO2022074370A1 - Il-33 therapy for use in the treatment, prevention or management of age-related macular degeneration (amd) - Google Patents

Abstract

The present invention provides IL-33 for use in the treatment, prevention or management of age-related macular degeneration (AMD) and methods of treating, preventing or managing age-related macular degeneration (AMD), comprising administering IL-33 to a patient in need thereof.

Description

IL-33 THERAPY FOR USE IN THE TREATMENT, PREVENTION OR MANAGEMENT OF AGE-RELATED MACULAR DEGENERATION (AMD)

Field of Invention

The present invention relates to a novel therapy for use in the treatment, prevention or management of age-related macular degeneration (AMD).

Background to the Invention

Age-related macular degeneration (AMD) is a progressive degenerative disease of the eye and the leading cause of central vision loss. Characterised by drusen deposits, atrophy of the retinal pigment epithelium (RPE) and photoreceptor (PR) loss, AMD can progress into two acute stages of vision loss, neovascularisation AMD (nAMD) and, in the majority of patients, geographic atrophy (dry AMD). Whilst nAMD can effectively be treated in most patients with VEGF-blocking agents, dry AMD currently has no therapeutic options, in part because of the multifaceted complex pathoaetiology of the disease.

The present invention aims to provide an effective therapy for the treatment, prevention or management of AMD.

Summary of the Invention

Dysregulation of the immune system homeostasis is operative in the pathogenesis of AMD, highlighting it as an important target for early therapeutic intervention. Previous work by the inventors has revealed a protective role for the IL-1 family cytokine, interleukin-33, in choroidal neovascularisation and attenuation of wound healing (Theodoropoulou, S. et al.). IL-33 is constitutively expressed in the RPE and Muller cells of the eye and acts as an 'alarmin' under necrotic cell death. Loss of endogenous IL-33 can lead to chronic inflammation after retinal detachment. However, converse to this, under phototoxic stress endogenous IL-33 released from Muller cells has been shown to trigger an inflammatory response, leading to PR degeneration (Xi, H. et al.).

Surprisingly, the present inventors have identified a protective role for IL-33 on RPE apoptosis and maintenance of retina/RPE metabolism homeostasis, revealing IL-33 treatment as protective against the pathogenesis of AMD. The inventors have also determined that exogenous IL-33 treatment did not demonstrate toxicity in the eyes of an insidious dry AMD model, aged Cfh+/- mice on a high fat diet.

Accordingly, in a first aspect the present invention provides IL-33 for use in the treatment, prevention or management of age-related macular degeneration (AMD).

IL-33 is a member of the IL-1 superfamily of cytokines, a determination based in part on the molecule's (3-trefoil structure, a conserved structure type described in other IL-1 cytokines, including IL-lo, IL-ip, IL-lRa and IL-18. In this structure, the 12 p-strands of the (3-trefoil are arranged in three pseudorepeats of four p- strand units, of which the first and last p-strands are antiparallel staves in a six- stranded p-barrel, while the second and third p-strands of each repeat form a p- hairpin sitting atop the p-barrel. IL-33 is a ligand that binds to high-affinity receptors ST2 (also known as IL1RL1) and IL-1 Receptor Accessory Protein (IL1RAP).

The IL-33 for use in the present invention is preferably exogenous IL-33, meaning that it originates from outside of the organism to which it is being applied/introduced. The exogenous IL-33 may be naturally occurring IL-33 (i.e., obtained from another organism of the same or different type) or it may be recombinant (i.e., produced by transfecting a foreign gene into a host cell, typically a bacterial cell). In either situation the IL-33 may be a full length protein or it may be truncated.

Alternatively, the IL-33 may be overexpressed by the subject of interest, for example by introduction of a CRISPR-Cas9 activation system. One example of a suitable CRISPR-Cas9 activation system includes a deactivated Cas9 nuclease fused to an activation domain and at least one IL-33-specific guide RNA. The CRISPR-Cas9 activation system may be a CRISPR-Cas9 activation plasmid. In one particular example, the CRISPR-Cas9 activation plasmid may comprise a DIO deactivated Cas9 nuclease fused to a VP64 activation domain, in conjunction with a single guide RNA (MS2) and an IL-33-specific single guide RNA engineered to bind the MS2-p65-HSF-l fusion protein. CRISPR-Cas9 activation systems may be introduced into the ciliary muscle, e.g., by electroporation, or may be transported using a viral vector gene therapy (which are familiar to the skilled person and/or which are described herein) comprising a transgene cassette of the activation system, which can then be administered via routes which are familiar to the skilled person and/or which are described herein. Alternatively, ribonucleoproteins (RNPs) corresponding to products from the activation systems (comprising the Cas9 variant protein and the guide RNA) can be administered to the eye via routes which are familiar to the skilled person and/or which are described herein. Overexpressed IL-33 is typically naturally occurring IL-33 and may be a full length protein. The overexpressed IL-33 may comprise nuclear localised expression of the IL-33 protein.

The full length human IL-33 polypeptide sequence is 270 amino acids as shown in Figure 5A. The polypeptide is formed of a nuclear domain, a central domain containing cleavage sites for proteases and an IL-l-like cytokine domain containing cleavage sites for caspases. The N-terminal includes a nuclear domain containing a chromatin binding motif, and the C-terminal includes the IL-l-like cytokine domain, which binds to receptors ST2 (also known as IL1RL1) and IL-1 Receptor Accessory Protein (IL1RAP).

The IL-33 is preferably a single chain non-glycosylated polypeptide. Examples of suitable IL-33 sequences are shown in Figure 5, A-C. In embodiments of the invention the IL-33 may comprise an amino acid sequence having at least 80%, 85%, 90%, 95%, 96%, 97%, 98%, 99% or more sequence identity with one or more of the amino acid sequences of Figure 5, A-C.

Truncated forms of IL-33 as described herein refer to proteins in which the N- terminus nuclear domain is all or partly deleted. Preferably, truncated forms of IL- 33 conserve all or part of the IL-l-like cytokine domain (i.e., amino acids 113- 270), such that they retain the ability to specifically bind to ST2 and/or IL1RAP. In embodiments of the invention the IL-33 may comprise or consist of human IL-3395-270 or human IL-33IO9-27O, human IL-33ii3-27o or human IL-3365-268 or an amino acid sequence having at least 80%, 85%, 90%, 95%, 96%, 97%, 98%, 99% or more sequence identity with one or more of those amino acid sequences. Without being bound by theory, it is believed that human IL-3395-270 and human IL-33109-270 are likely to be major mature forms of IL-33 in vivo and are therefore preferably conserved in truncated forms of IL-33.

The IL-33 may be administered in the form of a nucleic acid encoding IL-33. The nucleic acid may correspond to the native DNA sequence in the organism (e.g., human cDNA encoding human IL-33) or it may be codon optimised. For example, the IL-33 may be administered by gene therapy. For instance, the IL-33 may be administered by a viral vector gene therapy comprising an IL-33 transgene and a promoter. The gene therapy vector can be used to target cells of the eye, such as RPE cells, in order to treat or prevent AMD.

The viral vector may be an adeno-associated virus (AAV) and suitable AAV vector serotypes include 2, 5, 8 or 9. Preferably, the AAV vector capsid serotype is selected from 2, 5, 8 or 9. Engineered capsids can also be used. Suitable engineered capsids include AAV2tYF, AAV2.7m8, R100, AAV2.GL and AAV2.NN.

AAV2 is an extensively examined serotype that has been shown to effectively transduce RPE.

Synthetic AAV vector serotypes can also be used, such as hybrid serotypes having a capsid from one strain and a genome with inverted terminal repeats (ITRs) from second type. AAV2/5 is an example of a hybrid serotype comprising the ITRs of AAV2 and the capsid of AAV5. Preferably, the AAV vector comprises the AAV2 ITRs, which can be used with other capsid serotypes such as AAV5 or AAV9 or with engineered capsids.

The present inventors have observed good expression in RPE with AAV9 packaged vectors and in the eye with AAV2 packaged vectors. Preferably, the vector is AAV2. More preferably, vector comprises AAV2 ITRs and/or AAV2 capsid or a variant thereof. For example, the vector may comprise an AAV2 ITR with an AAV2 capsid or variant thereof and a transgene encoding truncated IL-33. Variants of AAV2 capsids may comprise an amino acid sequence having at least 80%, 85%, 90%, 95%, 96%, 97%, 98%, 99% or more sequence identity with the AAV2 capsid amino acid sequence.

The AAV vector typically comprises at least one promoter. Preferably the promoter is a ubiquitous promoter, such as a CMV promoter, which is known to be a strong promoter that drives constitutive expression of genes under its control. CMV promoters also include shCMV, CMVd2 and core CMV. Another suitable ubiquitous promoter is a CAG promoter, which is a synthetic promoter comprising a CMV enhancer element; the promoter, first exon and first intron of chicken beta-actin gene; and the splice acceptor of the rabbit beta-globulin gene. CAG promoters also include mini CAG (including the SV40 intron) and mini CAG without the SV40 intron. The present inventors have observed good expression with sub-retinal injections of AAV expressing GFP under a CMV promoter and with intravitreal injections of AAV expressing IL-33 under a CAG promoter.

Other promoters that have been used in AAV-based ocular gene therapies include the cytomegalovirus early enhancer/chicken-actin promoter (CBA aka CB7; 800 bp), the human phosphoglycerate kinase (PGK) promoter, smCBA, CBh, MeCP2, SV40mini, SCP3, the elongation factor-lalpha (EF-1) promoter, PGK and UbC.

The promoter may be a occular cell specific promoter, e.g., an RPE specific promoter, such as RPE65, VMD2, NA65p or Synpii.

Other occular cell specific promoters include Muller glial cell promoters, such as CHX10, GFAP, GfaABCID, Hypoxia-induced reactive MGC promoter (HRSE-6xHRE- GfaABClD), RLBP1, Short RLBP1, Murine CD44, Murine shCD44, ProB2; photoreceptor cell promoters such as Mouse RHO, Human RHO (rhodopsin), Mouse rodopsin m0p500, Mouse rhodopsin, Human Rhodopsin kinase (RHOK/GRK1), Human blue opsin HB570, Human blue opsin HB569, PRO.5, PR1.7, PR2.1, 3LCR-PR0.5, Mouse blue opsin (mBP500), Human interphotoreceptor retinoid binding protein (hIRBP), RBPe/GNAT2, Mouse CAR/ARR3, Human CAR/ARR3, CAR/ARR3, Human red opsin, Human green red opsin (G1.7p), Crx2kb, ProAl, ProA4, ProCi, ProA6,ProB5,ProC22,ProC32,ProD2,ProD3,ProD4,ProD5,ProD6, Synpl61; bipolar cell promoters such as Mouse metabotropic glutamate receptor 6 (mGrm6), 4x mGRM6e+SV40, Grm6e-Chxl0-Cabp5, Grm6-SV40, Cabp5, Chxl0-SV40, Grm6- mGluR500P, In4s-In3e- Grm6-mGluR500P, ProB4; amacrine cell promoters such as ProC2, ProBl; horizontal cell promoters such as ProC3; and retinal ganglion cell promoters such as Synl, Nefh, hSNCGp, ProA3, Ple344, Ple345.

Optionally the AAV vector may comprise an inducible promoter, such as Zinc, cadmium or copper-inducible sheep metallothionine-Ia promoter (MT-1), dexamethasone (Dex)-inducible mouse mammary tumor virus (MMTV LTR), tetracycline On or Off system promoters (Ptet), T7 bacteriophage promoter or a riboswitch comprising a ligand-sensing aptamer, a communication module (linker), and an effector domain (ribozyme).

The transgene species is preferably matched to the patient species. For example, when treating a human patient one would typically use a human transgene. The transgene may be naturally occurring, e.g., wild-type, or it may be recombinant. The transgene is typically included in the gene therapy vector as a cDNA sequence. However, the IL-33 transgene may be any polynucleotide, such as single or double-stranded DNA or RNA, comprising a nucleic acid sequence encoding any IL-33 polypeptide as discussed above. The IL-33 transgene polynucleotide sequence may be a codon-optimised sequence.

The AAV vector may additionally comprise a Woodchuck hepatitis post- transcriptional regulatory element (WPRE). WPRE is a DNA sequence that, when transcribed, creates a tertiary structure enhancing expression. Inclusion of WPRE may increase expression of the transgene delivered by the vector. The WPRE sequence may be mutated to reduce oncogenicity without significant loss of RNA enhancement activity (Schambach et al., 2005, incorporated herein by reference). One example of a suitable WPRE sequence is shown in Figure 6.

The AAV vector may additionally comprise one or more transcriptional regulatory enhancers selected from CE (CMV early enhancer), human interphotoreceptor retinoid-binding protein proximal enhancer (IRBPe), metabotropic glutamate receptor 6 enhancer (Grm6e), Hepatitis B Virus PRE (HPRE), WPRE3 and minute virus of mice (MVM).

The IL-33 transgene may comprise a protein tag, such as a hemagglutinin (HA) tag. HA can be used as an epitope tag and has been shown not to interfere with bioactivity or biodistribution of proteins to which it has been added. The protein tag can facilitate detection, isolation, and purification of the transgene. Other suitable protein tags may include Myc tags, polyhistidine tags and flag tags.

The AAV vector gene therapy may additionally comprise a Kozak sequence between the promoter and the IL-33 transgene. The Kozak sequence is known to play a major role in the initiation of the translation process and can therefore enhance expression of the IL-33 transgene.

The AAV vector gene therapy may additionally comprise a polyadenylation signal, such as bovine growth hormone (bGH) polyadenylation signal, e.g. as shown in Figure 7. Other suitable polyadenylation signals include SV40 late, sNRPl, Rabbit P-globulin protein A, spA, hGHpolyA, HSV TK poly(A) and adenovirus (L3) USE. Polyadenylation is the addition of a poly(A) tail to a messenger RNA. The poly(A) tail consists of multiple adenosine monophosphates; in other words, it is a stretch of RNA that has only adenine bases. The poly(A) tail is important for the nuclear export, translation, and stability of mRNA. Inclusion of a polyadenylation signal can therefore enhance expression of the IL-33 transgene. The AAV vector gene therapy may comprise two or more polyadenylation signals.

The AAV vector gene therapy may additionally comprise Inverted Terminal Repeat (ITR) sequences at either end of the vector. For example, the vector structure may be, in order: ITR - promotor - transgene (with optional protein tag) - optional WRPE - polyadenylation signal - ITR.

The AAV vector gene therapy may be administered to a human patient at a dosage of about 10⁷ to about 10¹³ vector genomes per eye (vg/eye). Optionally, the AAV vector gene therapy may be administered to a human patient at a dosage of about 10⁹ to about 10¹² vg/eye or about 10⁸ to about 10¹² vg/eye.

In the description above, the term "identity" is used to refer to the similarity of two sequences. For the purpose of this invention, it is defined here that in order to determine the percent identity of two amino acid sequences, the sequences are aligned for optimal comparison purposes. The amino acids at amino acid positions are then compared. When a position in the first sequence is occupied by the same amino acid as the corresponding position in the second sequence, then the molecules are identical at that position. The percent identity between the two sequences is a function of the number of identical positions shared by the sequences (i.e., % identity = number of identical positions/total number of positions (i.e. overlapping positions) x 100). Preferably, the two sequences are the same length. A sequence comparison is typically carried out over the entire length of the two sequences being compared. The skilled person will be aware of the fact that several different computer programs are available to determine the identity between two sequences. For instance, a comparison of sequences and determination of percent identity between two sequences can be accomplished using a mathematical algorithm. Similarity of amino acid sequences may be as defined and determined by the TBLASTN program, of Altschul et al, which is in standard use in the art. In particular, TBLASTN 2.0 may be used with Matrix BLOSUM62 and GAP penalties: existence: 11, extension: 1. Another standard program that may be used is BestFit, which is part of the Wisconsin Package, Version 8, September 1994, (Genetics Computer Group, 575 Science Drive, Madison, Wisconsin, USA, Wisconsin 53711). BestFit makes an optimal alignment of the best segment of similarity between two sequences. Optimal alignments are found by inserting gaps to maximize the number of matches using the local homology algorithm of Smith and Waterman (Adv. Appl. Math. (1981) 2: 482-489). Other algorithms include GAP, which uses the Needleman and Wunsch algorithm to align two complete sequences that maximizes the number of matches and minimizes the number of gaps. As with any algorithm, generally the default parameters are used, which for GAP are a gap creation penalty = 12 and gap extension penalty = 4. Alternatively, a gap creation penalty of 3 and gap extension penalty of 0.1 may be used. The algorithm FASTA (which uses the method of Pearson and Lipman, 1988) is a further alternative.

In embodiments of the invention the IL-33 may be pegylated. In embodiments of the invention the IL-33 may be provided as a dimer having the structure IL-33- PEG-IL-33. A similar structure has been described in connection with an antibody mimetic known as Fab-PEG-Fab (FpF) (Khalili et al. 2018).

The IL-33 may be bound to or incorporated within a nanoparticle or a carrier. For example, the IL-33 may be incorporated within or bound to micelles or liposomes. The nanoparticle or carrier may be a polymeric nanocarrier, i.e., a component comprising a polymer. For example, the polymeric nanocarrier component may comprise a polyethylene glycol (PEG) group and/or a polymer based constituent, such as a poloxamer.

The IL-33 can be administered directly into an eye. The administration may be topical or may be by injection, such as by intravitreal injection, subretinal injection, suprachoroidal injection. Alternatively, the administration may comprise electroporation of the ciliary muscle. Alternatively, the administration may comprise an implant to give sustained-release of IL-33. Suitable implants may be provided in the form of a port delivery systems (PDS). Sustained-release can also be provided using nanoformulations, hydrogels and other encapsulation techniques.

The IL-33 may be provided in the form of a pharmaceutically acceptable composition. The pharmaceutically acceptable composition is preferably sterile and may comprise one or more pharmaceutically acceptable carriers or excipients. Suitable carriers and excipients will be familiar to the skilled person and may be optimised in line with the intended route of delivery. For example, compositions of the invention may include buffers, binders, preservatives, thickeners or antioxidants, such as trehalose. Pharmaceutically acceptable compositions may be in the form of solutions or suspensions in aqueous media.

The pharmaceutically acceptable composition may be suitable for topical delivery, preferably ocular delivery. Alternately, the composition may be suitable for injection, such as intravitreal injection, subretinal injection or suprachoroidal injection.

The IL-33 may be administered in combination with a penetration enhancer, i.e., a compound capable of enhancing drug permeability across cellular and/or ocular membranes. This is particularly useful for topical administration of the IL-33. The penetration enhancer may be selected from one or more of a cyclodextrin (such as o-CD, P-CD, y-CD or derivatives thereof), a chelating agent (such as ethylenediamine-N,N,N',N'-tetraacetic acid (EDTA), ethylene glycol-bis(beta- aminoethyl)-N,N,N',N'-tetraacetic acid (EGTA), l,2-bis(o-aminophenoxy)ethane- N,N,N',N'-tetraacetic acid (BAPTA), or ethylenediamine-N,N'-disuccinic acid (EDDS)), crown ethers (such as 12-crown-4, 15-crown-5, or 18-crown-6), surfactants (such as saponin or benzalkonium chloride (BAC)) or cell-penetrating peptides (such as transactivating transcriptional activator (TAT), polyarginine-7 (R7), polyarginine-9 (R9; NONA), transporter-9, polylysine-9 (K9) or penetratin (PEN)).

The IL-33 may be for a once only administration. However, more typically the IL- 33 will be delivered intermittently over a period of months or even years. For example, the IL-33 may be administered weekly, fortnightly, monthly, every six, eight, twelve or fifteen or more weeks.

AMD is characterised by a progressive accumulation of characteristic yellow deposits, called drusen (a build-up of extracellular proteins and lipids), in the macula, between the retinal pigment epithelium and the underlying choroid. This accumulation is believed to damage the retina over time. AMD can be divided into 3 stages: early, intermediate, and late, based partially on the extent (size and number) of drusen. Early AMD is typically diagnosed based on the presence of medium-sized drusen, about the width of an average human hair. Early AMD is usually asymptomatic. Intermediate AMD is typically characterised by large drusen and/or retinal pigment abnormalities. Intermediate AMD may cause some vision loss, but, like early AMD, it is usually asymptomatic. In late AMD, enough retinal damage occurs that, in addition to drusen, patients will also begin to experience symptomatic central vision loss. The damage can either be the development of geographic atrophy (dry AMD) or the onset of neovascular disease (also known as wet AMD).

The IL-33 of the present invention may be used in the treatment, management or prevention of early stage AMD, intermediate stage AMD or late stage AMD.

Dry AMD (also called nonexudative AMD) encompasses all forms of AMD that are not neovascular (wet AMD). Dry AMD includes early and intermediate forms of AMD, as well as the late stage of dry AMD known as geographic atrophy. Preferably, the IL-33 of the present invention can be used in the treatment, management or prevention of dry AMD.

Interestingly, in around 10-20% of patients dry AMD progresses to the wet/neovascular type. The IL-33 of the present invention can therefore be used to slow and/or prevent progression of dry AMD to wet AMD. Additionally or alternatively, the IL-33 of the present invention may be used as an adjunct to current therapies for wet AMD. For example, the IL-33 may be administered in sequential or simultaneous combination with one or more of Bevacizumab (Avastin), Ranibizumab (Lucentis) or Aflibercept. The present invention additionally provides a method of treating, preventing or managing age-related macular degeneration (AMD), the method comprising administering IL-33 to a patient in need thereof.

The patient is preferably a mammal, including a human, and may be a paediatric or geriatric patient.

The average age for diagnosis of AMD is around 65 years but can be only around 55 years if the patient also has a genetic risk. The patient therefore may be at least around 45 years old, or at least around 50 years old, or at least around 55 years old or at least around 65 years old.

In terms of genetic risk there are about 150 genes implicated in AMD. However, the two most high associated genes are Complement Factor H (CFH) on chromosome 1 and Age-Related Maculopathy Susceptibility 2 (ARMS2) on chromosome 10. Both of these genes are highly associated with progression of AMD but it has not yet been determined whether they are in anyway causal.

CFH is an inhibitor that functions in regulation of the alternative-complement- pathway as well as innate immunity. CFH is localized to lq32, a region found by both linkage and by GWAS to be associated with all subtypes of AMD. Specifically, several independent reports have shown the functional polymorphism Y402H (rsl061170) in CFH, where a tyrosine is substituted by a histidine, to be associated with increased risk of both early and late stages of AMD (both neovascular and geographic atrophy). These analyses further suggest that the risk allele contributes to almost half of all cases of AMD in the population.

The ARMS2 gene is found between PLEKHA1 and HTRA1 in the 10q26 locus. ARMS2 has been shown to be expressed in multiple tissues, including the retina; however, its function remains unknown and its cellular location is disputed. Within ARMS2, rsl0490924, encoding the A69S change, has been associated with severe AMD phenotypes including early onset of disease and larger choroidal neovascularization lesions. This variation has been consistently associated with AMD risk across various ethnicities and has been hypothesized to affect the interaction of ARMS2 with other proteins. In addition to the A69S variant, a complex insertion/deletion (indel) in the 3' untranslated region of ARMS2 has also been significantly associated with AMD risk. The patient may therefore be positive for one or more genetic markers of AMD, such as one or more of the CFH or ARMS2 variants or polymorphisms mentioned above.

Brief Description of the Drawings

The invention will now be described in detail, by way of example only, with reference to the figures.

Figure 1. Exogenous IL-33 is non-toxic in aged C57BL/6~HFD and Cfh+/~ ~HFD mice. (A) OCT scan of eyes from HFD fed Cfh+/- and C57BL/6 mice injected with either vehicle or IL-33 (1 ng or 200 pg dose). Scale bars are both 100 pm. (B) Graph to show retinal thickness measured from OCT scans (n = 7 - 11). (C) Transmission electron micrograph images of sub-RPE deposits in Cfh+/- ~HFD mice treated with either IL-33 (1 ng) or vehicle and of vehicle treated C57BL/6~HFD. BrM, Bruch's membrane. Scale bar is 2 pm (D) Confocal images of ZO-1 stained RPE/choroid show multinucleated cells (*; >3 nuclei/cell) in Cf77+/-~HFD and C57BL/6~HFD vehicle and IL-33 (1 ng) treated eyes. Scale bar is 20 pm. (E) Graph shows mean number of multinucleated cells (>3 nuclei/cell) per field of view (FOV).

Figure 2. IL-33 treatment provides protection against cell death in the RPE/choroid of Cf +/-~HFD eyes. (A) Confocal images of ZO-l/TUNEL stained RPE/choroid flatmounts from Cf77+/-~HFD and C57BL/6~HFD eyes treated with either vehicle or IL-33. Scale bar is 38 pm. (B) Quantitative analysis of mean TUNEL+ cells per field of view in RPE/choroid flatmounts. A significant reduction in cell death is observed in Cf77+/-~HFD IL-33 treated eyes compared to Cfjh+/-~HFD vehicle treated eyes (**p = 0.0037) and C57BL/6~HFD IL-33 treated eyes (*p = 0.040). (D) Quantitative analysis of TUNEL+ cells per field of view in retina. Cell death was significantly increased in Cf77+/-~HFD eyes compared to C57BL/6~HFD control, irrespective of treatment (*p = 0.032). IL-33 treatment did not affect cell death in the retina for both genotypes.

Figure 3. Visual function is protected by IL-33 in both Cfh+/-~HFD and C57BL/6~HFD animals. Graph shows visual responses for flash intensity of 3.9 log cd (sec/m²) pre high fat diet (HFD; week 0) at 4 weeks of HFD and after 8 weeks of HFD and 4 weeks of intravitreal injections of IL-33 or vehicle. IL-33 treated eyes in Cfh+/- HFD and C57BL/6~HFD animals have improved responses for a-wave (*p = 0.041; A) and b-wave (p = 0.055; B) amplitudes.

Figure 4. IL-33 treatment regulates proteins essential to metabolism homeostasis in the eye. Representative western blot of protein expression of Hexokinase II (HKII) and voltage dependent anion channel (VDAC) was shown for RPE (A) and retina (D) lysates. Densiometric analysis demonstrates (B) VDAC has significantly reduced expression in Cfh+/- HFD vehicle and IL-33 treated eyes compared to IL-33 treated C57BL/6~HFD eyes (*p = 0.011 and p = 0.032 respectively). A trend decrease in expression is observed in Cfh+/- F\FD vehicle eyes only when compared to C57BL/6~HFD vehicle eyes (p = 0.09). (C) HKII is significantly reduced in RPE of IL-33 treated eyes (*p = 0.024) irrespective of genotype. (E) HKII is significantly reduced in Cff)+/- HFD vehicle treated eyes compared to Cfh+/- HFD IL-33, C57BL/6~HFD veh and IL-33 treated eyes (*p = 0.017, p = 0.010 and p = 0.039 respectively). (F) VDAC expression is unchanged across genotype and treatment in retinas.

Figure 5. Examples of suitable IL-33 Sequences. (A) Human IL-33 protein, 270 amino acids (SEQ ID NO: 29). (B) Recombinant Human IL-33 protein, approximately 17.9 kDa, a single non-glycosylated polypeptide chain containing 159 amino acids (SEQ ID NO: 30). (C) Recombinant Human IL-33, a single nonglycosylated, polypeptide chain containing 160 amino acids and having a molecular mass of approximately 18 kDa (SEQ ID NO: 31).

Figure 6 shows an example DNA sequence for a WPRE sequence (SEQ ID NO: 32).

Figure 7 shows an example DNA sequence for a bGH poly(A) signal sequence (SEQ ID NO: 33).

Figure 8 shows the experiment plan to analyse secreted IL-33 gene therapy safety and efficacy.

Figure 9A. Adeno-associated virus constructs for truncated secreted IL- 33. (A) Truncated mouse IL-33 (68 - 266 aa) is expressed under a constitutively expressed CMV regulatory promoter and an equivalent null control vector on the same backbone but has no open reading frame (ORF) of coding sequence. (B) Truncated human IL-33 (65 - 270 aa) is expressed under a constitutively expressed CAG regulatory promoter and an equivalent null control vector that contains a truncated ORF stuffer sequence with no start codon (2 - 83 aa E.coli beta-galactosidase).

Figure 10. Acute light exposure leads to degenerative disease phenotypes in vivo. (A) Schematic shows method of unilateral light exposure in light-induced retinal damage (LIRD) model. (B) OCT scan through the optic disc from a LIRD and naive eye (scale bar is 100 pm). (C) LIRD mice display significant reduction in retina thickness at day 14 post light exposure, when compared to naive eyes (multiple t-test *p < 0.05, **p < 0.01). (D) LIRD leads to a significant increase of IBA1+ cells in the retina at day 24 after light exposure (t-test, *p < 0.05). (E) Representative images of retina flatmounts from LIRD and naive eyes, stained with DAPI and IBA1+ (scale bar is 80 pm).

Figure 11. Light exposure leads to significant changes in stress response associated genes in the retina and RPE/choroid. (A) PCA clustering analysis of qPCR data from 12 genes assessed in the retina reveals separation of samples day 1 after light-exposure in the LIRD model. (B) Separation of day 1 LIRD retina samples is driven by significant changes in expression of genes, including Hmoxl, C3arl, Ccl2 and Illb (*p < 0.05, **p < 0.01, ***p < 0.001). (C) PCA clustering analysis of qPCR data from 12 genes assess in the RPE/choroid identify distinct clustering and separation of samples at 6 h and 1 day after light exposure. (D) Clustering is driven by changes in gene expression including Hmoxl, Ccl2 and Illb (*p < 0.05, **p < 0.01).

Figure 12. Pre-treatment of eyes with recombinant IL-33 protects against disease phenotypes in the LIRD model. (A) Graphs show retinal thickness measured from OCT scans through the optic disc (i) Retinal thickness is significantly reduced in LIRD eyes with vehicle and IL-33 treatment at day 14 after light exposure, when compared to vehicle or IL-33 injected eyes only (*p < 0.05), however at 21 days after light exposure (ii) retinal thinning has slowed in IL-33 treated LIRD eyes, whilst vehicle treated LIRD eyes display significant reduction compared to injected only (vehicle or IL-33) controls (*p < 0.05, **p < 0.01). (Bi) Representative images of DAPI and IBA1 stained retina flatmounts (scale bar is 80 pm) and graph (Bii) to show number of IBA1+ cells in retina at day 24 post-light exposure. Vehicle treated LIRD eyes have a significant increase in IBA1 + cells compared to IL-33 treated eyes only (*p < 0.05) and similar increase compared to vehicle treated eyes only (p = 0.052). Conversely, IL-33 treated eyes start to show a reduction in IBA1+ cells. (C) PCA clustering analysis of qPCR data from 6 genes analysed in RPE/choroid 1 day after light exposure. Clustering reveals clear separation of vehicle treated LIRD eyes only.

Figure 13. AAV delivered truncated human and mouse IL-33 is expressed and secreted in vitro and confers protection against oxidative stress. (A) Treatment of cells with AAV2 or 5 expressing truncated human IL-33 or null virus is non-toxic at 3 days after transduction (i) and is secreted into the supernatant (ii) (B) Transduction of cells with AAV2 truncated mouse IL-33 or null vectors is non-toxic at 3 days after treatment (i) and secreted IL-33 is detected in the supernatant at day 4 (ii). Cells were treated with H2O2 (60 pM) 3 - 4 days after AAV transduction and cytotoxicity measured 3 days later. (C) Graph shows increase in cytotoxicity for null AAV treated cells with H2O2 compared to null only (p = 0.08) or tIL-33(human) AAV treated cells (p = 0.051). Cytotoxicity is reduced in tIL-33(human) AAV treated cells. (D) Graph shows a similar trend of protection in tIL-33(mouse) treated cells.

Figure 14. AAV2.tIL-33(human) treatment modulates gene expression changes in the retina and RPE/choroid after LIRD light-exposure in a dose dependent manner. Naive, AAV2.tIL-33(human) or AAV2.null treated eyes with LIRD or AAV2.tIL-33(human) and AAV2.null treated eyes only were taken 1 day after light exposure for analysis of gene expression in retina and RPE/choroid. (A) Graphs show qPCR data from analysis of Hmoxl, Ccl2, Illb and C3arl gene expression. Panels top to bottom show samples injected with either high (le9 gc/eye), mid (le8 gc/eye) or low (le7 gc/eye) titre virus. Expression of tlL- 33(human) at the mid-dose range recovers changes in gene expression seen after light exposure. (B) Graphs show qPCR data for Hmoxl, Ccl2 and Illb in RPE/choroid. Panels left to right show results from samples injected with high, mid and low titre virus. At the mid-range dose tIL-33(human) increases Hmoxl and Cc/2 expression with light-exposure.

Figure 15. AAV2.tIL-33(human) injected eyes express human IL-33 in the retina. Graph shows ELISA results from analysis of retina lysates taken from eyes 21 days after injection with a high (le8 gc/eye), mid (le8 gc/eye) and low (le7 gc/eye) viral titre of AAV2.tIL-33(human) or AAV2.null control vectors. Expression of human IL-33 shows a dose dependent increase.

Figure 16. AAV.tIL-33(human) gene therapy reduces lesion size in the 'wet' AMD L-CNV model in a dose dependent manner. Naive eyes or those injected with a high (2e9 gc/eye), mid (2e8 gc/eye) or low (2e7 gc/eye) were subjected to laser induced choroidal neovascularisation and lesions analysed 7 days later. (A) Representative OCT scans through lesions (scale bar is 100 pm).

(B) Graphs to show lesion size analysis. Treatment with AAV2.tIL-33(human) at a mid-range dose significantly reduced lesion size compared to equivalent AAV2.null injected eyes (*p < 0.05).

Figure 17. Role of nuclear IL-33 in mitochondrial metabolism. (A) Gene expression of IL33 following transfection of ARPE-19 with either an IL-33 activation plasmid or scrambled gRNA activation plasmid (n=3). (B) Representative immunoblot of IL-33 expression in ARPE-19 transfected with either an IL-33 activation plasmid or scrambled gRNA activation plasmid (n=3).

(C) ARPE-19 were transfected with either an IL-33 activation plasmid or scrambled gRNA activation plasmid; cell lysates were split into nuclear or cytoplasmic fractions and western blotting was used to determine the subcellular location of IL-33 (n=2). (D) Mitochondrial stress test following transfection of ARPE-19 with either an IL-33 activation plasmid or scrambled gRNA activation plasmid; XF injections were oligomycin (IpM), FCCP (0.5pM) and rotenone/antimycin A (IpM) (n=3). (E) Parameters calculated from (D) (n=3). (F) Glycolysis stress test following transfection of ARPE-19 with either an IL-33 activation plasmid or scrambled gRNA activation plasmid; XF injections were oligomycin (IpM), FCCP (0.5pM) and rotenone/antimycin A (IpM) (n=3). (G) Parameters calculated from (F) (n=3). (H) ARPE-19 were transfected either an IL- 33 activation plasmid or scrambled gRNA activation plasmid; RT-PCR was used to determine the relative gene expression of targets involved in glycolysis or the TCA cycle (n=3). (I-J) ARPE-19 were transfected either an IL-33 activation plasmid or scrambled gRNA activation plasmid; protein was extracted and immunoblot analysis was used to determine the expression of PKM2, GLUT1 and PC (n=3). (K) Representative transmission electron microscopy of ARPE-19 cells transfected with an IL-33 activation plasmid or scrambled gRNA activation plasmid. Magnification 4500x. Data are expressed as means ± SD from at least three independent experiments. (D-G) Represents the biological repeats from three independent experiments (n=3); each biological repeat is the mean of two technical repeats (two seahorse wells per experiment). (C) Represents two independent immunoblots. One-way ANOVA with Dunnet's multiple comparisons test; *p<0.05, **p<0.01, ***p<0.001.

Figure 18. Nuclear IL-33 promotes oxidative glucose metabolism. (A-B) ARPE-19 were transfected with either an IL-33 activation plasmid or scrambled gRNA activation plasmid; (A) RNA was extracted, and RT-PCR was used to determine the expression of MPC1 and MPC2 n=3) (B) protein was extracted and western blot analysis was used to determine the expression of MPC1 and MPC2 (n=3). (C-D) ARPE-19 were transfected with either an IL-33 siRNA or scrambled siRNA; (C) RNA was extracted, and RT-PCR was used to determine the expression of MPC1 and MPC2 n=3) (D) protein was extracted and western blot analysis was used to determine the expression of MPC1 and MPC2 (n=3). (E) Modified mitochondrial stress test following transfection of ARPE-19 with either an IL-33 activation plasmid or scrambled gRNA activation plasmid; XF injections were oligomycin (IpM), FCCP (0.5pM), UK5099 (5pM) and rotenone/antimycin A (IpM) (n=3). (F) Parameter calculated from (E) (n=3). (G) Modified mitochondrial stress test following transfection of ARPE-19 with either an IL-33 siRNA or scrambled siRNA; XF injections were oligomycin (IpM), FCCP (0.5pM), UK5099 (5pM) and rotenone/antimycin A (IpM) (n=3). (H) Parameter calculated from (G) (n=3). (I) Uniformly labelled C13-glucose incorporation into ARPE-19 TCA cycle metabolites following transfection with either an IL-33 activation plasmid or a scrambled gRNA activation plasmid control; relative abundance of C13 and C12 including succinate, fumarate, malate and citrate (n=3). (J) Uniformly labelled C13-glucose incorporation into ARPE-19 TCA cycle metabolites following transfection with either an IL-33 siRNA or scrambled siRNA control (n=3). Relative abundance of C13 and C12 including succinate, fumarate, malate and citrate (n=3). (K-L) Mass isotopologue distributions (MID) of ARPE-19 TCA cycle intermediates (K) M+6 citrate and (L) M+4 malate, following transfection with either an IL-33 activation plasmid or a scrambled gRNA activation plasmid (n=3). (M) ARPE-19 were transfected with either an IL-33 activation plasmid/ scrambled gRNA activation plasmid or an IL-33 siRNA/ scrambled siRNA; western blot analysis was used to determine the phosphorylation status of pyruvate dehydrogenase (n=3). (N) Citrate M+ 3/ pyruvate M+3 ratio in ARPE-19 transfected with either an IL-33 activation plasmid or scrambled gRNA activation plasmid (n=3). Data are expressed as means ± SD from at least three independent experiments. (E-H) Represents the biological repeats from three independent experiments (n=3); each biological repeat is the mean of three technical repeats (three seahorse wells per experiment). Unpaired Student's T- test; *p<0.05, **p<0.01, ***p<0.001.

Example 1

Methods

Mice and in vivo experimental procedures

C57BL/6 mice without rd8 mutation were acquired from Charles River Laboratories (Margate, UK). Cfh-/- mice were backcrossed to C57BL/6 mice to establish Cfh+/- mice within the University of Bristol Animal services Unit, UK. Animals were kept according to Home Office Regulations in the animal house facilities at University of Bristol. Procedures were performed according to University of Bristol institutional guidelines and approved under Home Office Project License 30/3281. Animal treatment was in accordance to the Association for Research in Vision and Ophthalmology (ARVO) statement.

Mice aged 18 - 22 months were switched to a high fat diet (TD.88051; Envigo, IN, USA), for eight weeks. For in vivo imaging optical coherence tomography (OCT) and electroretinogram (ERG) mice were anaesthetised by inhalation of 1.5% v/v isoflurane or 2% v/v for intravitreal injections. For all procedures, pupils were dilated with topical application of 1% v/v tropica mide, followed by 2.5% v/v phenylephrine. OCT scans of retina and fundal images were captured using Micron IV (Phoenix Research Laboratories, Pleasanton, CA, USA). For scotopic ERGs, mice were dark adapted overnight. ERGs were recorded with increasing flash intensities using the Phoenix Micron™ Focal ERG system (Phoenix Research Laboratories) and LabScribeERG v3 software (iWorx, NH, USA). Administration of recombinant mouse IL-33 (200 pg or 1 ng; ALX-522-101-C010, Enzo Life Sciences Ltd, Exeter, UK) or vehicle control was delivered via a 2 pL intravitreal injection. Vehicle consisted of RPMI, 5% v/v fetal calf serum; 0.5% penicillin/streptomyocin; 0.5% L-glutamine; 0.5 mM Sodium Pyruvate, 2.5 ng/mL recombinant murine IL-3 (all Thermo Fisher Scientific, UK) and 5 ng/mL recombinant murine SCF (Stem Cell Technologies, UK).

Whole-mount immunohistochemistry For RPE/choroid and retina whole-mounts, posterior eye cups were dissected and fixed in 2% v/v formaldehyde, (diluted from 16% formaldehyde Pierce™ [Thermo Fisher, UK] in PBS) overnight at 4 °C. Cups were petaled, retina detached from RPE/choroid and both blocked in 5% w/v bovine serum albumin with 0.3% v/v Triton X-100 (Sigma Aldrich, UK) in PBS for 1 h at room temperature. Tissue was stained with TMR Red-dUTP TUNEL reaction mix according to manufacturer's instructions (12156792910, Roche Diagnostics, Burgess Hill, UK), before incubation with ZO-1 antibody (61-7300; Thermo Fisher Scientific, UK; 1 in 50) in 1% w/v BSA with 0.15% v/v Triton X-100, overnight at 4 °C. After washes, tissue was incubated with Alexa fluor 488 (A-11034; Invitrogen, UK; 1 in 200) for 2 h at room temperature. Washed tissue was mounted with vectashield hardset mounting media (H-1500; Vector Laboratories, Peterborough, UK).

Microscopy and image analysis

Images were captured at 40x magnification as z-stacks using a Leica SP5-AOBS confocal laser scanning microscope (Leica Microsystems Ltd., Wetzlar, Germany).

Image analysis was quantified from 6 - 8 images from each whole-mount. TUNEL positive spots were identified by intensity and size using Velocity® Image Analysis Software 6.0. For analysis of multinucleated cells, the cell counter tool in FIJI (Schindelin, J. et al.) was used to count ZO-1 stained cells with three or more nuclei. Data is presented as mean number of TUNEL+ or multinucleated cells per field of view (FOV).

Retinal thickness was measured in the OCT scans using FIJI (Schindelin, J. et al.). Scale was set to image scale bar and distance measured from top of RPE layer to top of the nerve fibre layer at 200 and 400 micron from the optic nerve head.

Protein lysate and western blots

RPE lysates were prepared as previously described (Wei, H. et al.), using Pierce® RIPA buffer (89900; Thermo Fisher Scientific, UK) and protease inhibitor (5872S; Cell Signalling Technologies, UK). Whole retinas were crushed in 200 pL Pierce® RIPA buffer (89900; Thermo Fisher Scientific, UK) with protease inhibitor (5872S; Cell Signalling Technologies, UK) for protein extraction. Protein concentrations were quantified using Pierce™ BCA protein assay kit (23227, Thermo Fisher Scientific, UK). Samples were prepared for SDS-PAGE with NuPAGE LDS sample buffer (NP0007; Invitrogen, UK) and 5% v/v 2- mercaptoethanol (Sigma Aldrich, UK). Proteins were separated on Novex™ 4-20% Tris-Glycine Mini Gels (XP04202BOX; Thermo Fisher Scientific, UK), transferred to PVDF (IB24001; Thermo Fisher Scientific, UK), before blocking with 5% w/v milk in TBS+Tween 20 (TBS-T; 0.1% v/v). Blots were incubated with primary antibodies - Hexokinase II (2867), VDAC (4661) and p-actin (3700) at 4 °C overnight. After thorough washing, blots were incubated with appropriate secondary antibody, anti-rabbit HRP (7074S; 1 in 2000) or anti-mouse HRP (7076S; 1 in 2000; all Cell Signalling Technologies, UK). Chemiluminescent detection was performed using Amersham ECL reagents and developed using Hyperfilm™ ECL film (both GE Healthcare, UK). Densitometry was performed using FIJI (Schindelin, J. et al.).

Electron microscopy and image analysis

For transmission electron microscopy (TEM), eyes were fixed in 2.5% v/v glutaraldehyde for 2 h at RT. The anterior chamber and lens were removed, before further fixing in 2.5% v/v glutaraldehyde at 4°C until processed. Glutaraldehyde fixed eye cups were washed in buffer, further fixed in osmium tetroxide, washed with water, stained with uranyl acetate and then dehydrated through a graded ethanol series to 100% ethanol. Ethanol was exchanged with propylene oxide and then Epon resin. After suitable infiltration with Epon resin mixtures, eyes were embedded and polymerised at 60°C for 2 days. Survey sections (0.5 -1pm thick) from anterior to posterior showed circular profiles with the retina to choroid layers in cross-section. Regions of interested were delineated and sectioned (70-80nm) for the electron microscope, ultrathin sections were stained with lead and uranyl salts.

Images of the RPE-Bruch's membrane interface were captured using a Tecnai T12 microscope (Thermo Fisher Scientific, UK). For Cfh+/- samples 55 - 60 images were captured across four sections per eye (n = 3). For each image fifteen measurements were made along the Bruch's membrane. Each deposit measurement was taken from the central elastin region of the Bruch's membrane to the top of the deposit, using FIJI (Schindelin, J. et al.). Fifteen images were taken from one section (ora-ora) for C57BL/6 sample. Statistics

For analysis of data from treated eyes, three-way ANOVA was performed with independent factors, treatment, genotype and animal (to account for paired eyes). A Tukey multiple comparisons test was used where significant effects were found. All three-way ANOVA analysis were performed in R studio (R version 3.6.1; RStudio, Inc., Boston, MA). For analysis of sub-RPE deposits a paired t-test was performed and for pre-treatment ERG multiple t-tests were performed for each flash intensity. All t-test analyses were performed using GraphPad Prism (version 8.1.2; GraphPad Software Inc. San Diego, CA). All data is expressed as ±SEM.

Results

In vivo imaging and characterisation of pathological features in Cfh+/~ ~HFD model confirm non-toxicity of exogenous IL-33

To analyse the effect of IL-33 treatment in the Cfih+/-~HFD model of dry AMD, we determined a non-toxic dose, based on previously used effective doses (Theodoropoulou, S. et al.). Release of endogenous IL-33 after phototoxic stress leads to photoreceptor degeneration and retinal thinning (Xi, H. et al.). Conversely, exogenous IL-33 doses 200 pg and 1 ng protect against choroidal neovascularisation (Theodoropoulou, S. et al.). Here, aged Cfh+/- ~HFD and C57BL/6~HFD mice were treated with either 200 pg or 1 ng IL-33 weekly from the fifth week of HFD, for a total of four weeks. For both genotypes, no differences in retinal thickness were observed with 1 ng IL-33 (Figure 1A and B) or 200 pg treatment (Figure 1A). We therefore proceeded with the higher dose to assess IL-33 effects against disease phenotypes.

A key feature of pathology in the Cfih+/-~HFD dry AMD model is the formation of sub-RPE deposits and subsequent RPE damage. Here, we observed formation of sub-RPE deposits in both vehicle and IL-33 treated CfO+/-~HFD eyes (Figure 1C), with a mean deposit height of 721 nm in vehicle treated eyes and 676 nm in IL-33 treated eyes (n = 3,3). Large sub-RPE deposits were also noted in C57BL/6 vehicle control eyes (Figure 1C). Further analysis of RPE damage, by assessment of cells with >3 nuclei/cell, revealed presence of multinucleated RPE in both Cf77+/-~HFD and C57BL/6~HFD animals (Figure ID). IL-33 treatment had no significant effect on the number of multinucleated cells and no differences were observed between genotypes (Figure IE). Altogether, these analyses show that treatment with 1 ng of IL-33 did not demonstrate toxicity when read out over a period of eight weeks and with a total IL-33 administration of four 1 ng injections. Additionally, the data did not demonstrate any exacerbation of pathological features of the Cfh+/- HFD model. There was however no difference in the extent of RPE damage between genotypes.

IL-33 provides partial protection against apoptosis in Cfh+/-~HFD eyes

RPE atrophy, photoreceptor (PR) loss and retinal thinning are early indicators of disease in AMD. In our impaired autophagy model of dry AMD, TUNEL-positive apoptosis in RPE and PRs is increased. We therefore analysed this pathological feature in treated Cfh+/-~HFD and C57BL/6~HFD eyes. In RPE, we found TUNEL- positive apoptosis in Cfh+/- HFD animals (Figure 2A), which correlates with previous reports of RPE atrophy in this model. IL-33 treatment significantly rescued cell death in Cfh+/- HFD animals with a reduction in level of apoptosis to that seen in C57BL/6~HFD eyes (Figure 2B). There was no difference between vehicle treated C57BL/6~HFD eyes and Cfh+/- HFD eyes (Figure 2B). Combined with the observation that there was no significant interaction for treatment: genotype, this could indicate a positive effect of IL-33 treatment for both genotypes.

In the retina the Cff)+/- HFD animals had a significant increase in TUNEL-positive apoptosis compared to C57BL/6~HFD retinas, irrespective of treatment (Figure 2C). The increased cell death indicate manifestation of disease in the Cf77+/-~HFD animals and is in accordance with previous observations of moderate thinning in the ONL of this model. Further to this, these data corroborate no toxicity of IL-33 in the retina.

IL-33 treatment protects visual function in wildtype and C h+/-~HFD mice

Cf77+/-~HFD mice display a significant impairment in visual function, compared to wildtype animals of the same diet. We therefore analysed the progressive effect of disease and treatment on scotopic visual function, measured by ERG. Visual function at pre-HFD and 4 weeks of HFD showed no functional differences between C57BL/6 and Cfh+/- mice (Figure 3). After eight weeks of HFD and four weeks of treatment, vehicle treated eyes had significantly reduced visual function in both C57BL/6~HFD and Cfh+/- HFD animals, for the highest flash intensity of the a-wave amplitude of, when compared to IL-33 treated eyes (Figure 3A). A similar trend was seen at the highest flash intensity for the b-wave amplitude (Figure 3B), although the values did not reach statistical significance (p = 0.055). The reduction in visual function was greatest for the vehicle injected C57BL/6~HFD mice, which is highlighted by a significant effect of genotype at flash intensity 1.5 of a-wave response (data not shown) and the -0.9 flash intensity of b-wave response (data not shown). These data suggest a protective effect of IL-33 for visual function irrespective of genotype. No significant differences were observed before or at 4 weeks of HFD, inferring IL-33 treatment is protective against the progressive effects of HFD in both Cfh+/- and C57BL/6 mice. With age, large sub-RPE deposits in C57BL/6 eyes accrue (Figure 1C).

IL-33 regulates expression of metabolic proteins essential to RPE and retina homeostasis

Visual function requires a carefully coordinated homeostasis of energy metabolism in retina and RPE. Energetically demanding, the retina relies on aerobic glycolysis, whilst RPE limits consumption of glucose and utilises lactate excreted by photoreceptors as fuel for OXPHOS. Analysis of mitochondrial marker VDAC in our RPE lysates, as a measure of mitochondrial health, revealed a disease effect on mitochondrial function in RPE of the Cff)+/- HFD animals where a significant reduction in both IL-33 and vehicle treated Cfh+/- HFD eyes compared to C57BL/6~HFD IL-33 treated eyes was observed (Figure 4A-B). A reduction of VDAC expression in Cff)+/- HFD vehicle eyes only, when compared to C57BL/6~HFD vehicle treated eyes (Figure 4B) may indicate a general positive effect of IL-33 treatment on VDAC expression and mitochondrial health. Analysis of pro-glycolytic protein hexokinase II (HKII) in RPE lysates revealed significant reduction in expression with IL-33 treatment, irrespective of genotype (Figure 4A and C), supporting the notion that IL-33 treatment promotes the transition of glucose across the RPE layer to be utilised in PRs.

In the retina, HKII expression is an essential mediator of aerobic glycolysis. Here, in the Cfh+/- HFD vehicle control we observed significant loss of HKII compared to C57BL/6~HFD vehicle and C57BL/6~HFD IL-33 treated retinas (Figure 4D-E). IL-33 treatment in Cfh+/- HFD retinas significantly recovered HKII expression to levels similar in C57BL/6 eyes (Figure 4E). In contrast to our results in the RPE, mitochondrial marker VDAC was unchanged in retina (Figure 4D and F). Discussion

Our results highlight a non-toxic effect of exogenous IL-33 treatment in aged mice and demonstrate a protective effect of treatment for aged eyes and in a dysregulated immune-mediated insidious model of AMD (O?7+/-~HFD animals). Here, we demonstrate significant retinal cell loss and RPE atrophy in the Cfh+/- ~HFD mice, showing that IL-33 treatment protects RPE from cell death and supports the metabolic homeostasis of retina/RPE; indicating that this treatment is protective against the pathogenesis of AMD.

Example 2

Objective 1: Generate secreted IL-33 expressing AAV vector:

Truncated mouse IL-33 (68 - 266 aa) will be cloned into an AAV backbone under a broad expressing CMV promoter and packaged into AAV2 vectors (commonly used in ocular gene therapy; effectively transduces RPE). The truncated region of IL-33 is selected for expression, based on published viral expression of secreted IL-33 in the brain. The CMV promoter is chosen due to its broad and strong expression - we have seen good expression with sub-retinal injections of AAV expressing GFP under a CMV promoter.

Expression of IL-33 from construct will be tested in vitro initially. Changes to IL- 33 expression construct can be made to mitigate lack of expression. We do not anticipate any issues with transduction in vivo with the AAV2 packaging system, but in this instance our lab has previously seen good expression in RPE with AAV9 packaged vectors under CMV promoter by sub-retinal delivery.

Objective 2: Assess viral dose by titration of low and high dose in wildtype (C57BL/6J) mice

AAV-IL-33 will be injected at a low and high dose (5 x 10¹⁰ vs 1 x 10¹² gc/mL) sub-retinally into wildtype mice (C57BL/6J). Two groups will be injected; group 1 to be harvested 2 weeks post- injection and group 2 to be harvested at 4 weeks. Clinical imaging at days 3, 7, 14 and 28 (group 2 only) will allow us to monitor whether the high dose of IL-33 has any adverse effects on the retina thickness. IL-33 overexpression will be determined by western blot. Objective 3: Establish the impact of RPE secreted IL-33 overexpression on resident retinal microglia in eyes of naive and disease models (Figure 1; Group 1 and Group 2):

Using the Cx3crl^CrER;Rosa-td-Tomato mouse line, established in our lab, we will label resident microglia (tamoxifen induced microglia td-Tomato expression). After expression of tdTomato is established (+four weeks) the mice will be injected with either AAV-IL33 or AAV-empty control. An un-injected control group will be included. Two - three weeks post injection (determined in objective 2) we will induce retinal degeneration (light damage model; LD - established in the lab) through exposure to light in one eye (with the contralateral being a naive control - to be randomised between left/right eyes). Microglia will be sorted from eyes at one month post-AAV (2 weeks post-LD) and three months post-AAV (10 weeks post-LD), resulting in the following established groups:

• AAV-IL-33; LD +/- (1 month)

• AAV-empty; LD +/- (1 month)

• Un-injected; LD +/- (1 month)

• AAV-IL-33; LD +/- (3 month)

• AAV-empty; LD +/- (3 month)

• Un-injected; LD +/- (3 month)

Sorted microglia will be analysed by RNA-seq (pipeline established in lab). To establish secreted IL-33 influence in healthy and diseased eyes over time, transcriptomes will be compared between disease and control eyes and across time (compared to AAV-empty and un-injected controls). Targets of interest will be validated at protein level from sorted microglia by cytokine profiling or western blot.

The LD model established in our lab shows the greatest disease change in microglia response at day 3. Through assessment of differences in the cytokine profiles at the 2 week time point post LD we will determine whether addition of a three day time point will be important for capturing effects of secreted IL-33 overexpression on microglia over time in diseased eyes.

Determine secreted IL-33 gene therapy efficacy protecting against retinal degeneration in a light-induced oxidative stress model (Figure 8; Group 1).

As described above Cx3crl^CrER;Rosatd-Tomato;/ M-IL33 or AAV-empty control injected eyes (and un-injected controls) will be have light damage induced and eyes will be monitored in vivo before harvesting 2 weeks post-LD (one-month post-AAV). From this group AAV-IL-33 efficacy will be determined by two main readouts:

• Reduction of retinal thinning by 60% in IL-33 treated group (measured by in vivo clinical imaging)

• Reduction of TUNEL-positive cells by 60% with IL-33 treatment (measured by ex vivo immunofluorescence staining)

Example 3

Introduction

Our group has recently revealed the potential of homeostatic cytokine IL-33 for treating AMD disease phenotypes in both 'wet' and 'dry' forms of AMD (Clare et al., 2020; Scott et al., 2021; Theodoropoulou et al., 2017). Here we expand evidence for recombinant IL-33 protein in protecting against disease pathways and subsequent AMD-like pathology in a light-induced retinal damage (LIRD) model, emulating oxidative stress. Furthermore, we demonstrate a successful method for AAV gene therapy delivery of IL-33 and identify efficacy for this treatment in protecting against disease phenotypes in models of 'dry' AMD (LIRD) and 'wet' AMD (laser choroidal neovascularisation; L-CNV).

Methods

Animals and in vivo procedures

C57BL/6J mice without rd8 mutation were acquired from Charles River Laboratories (Margate, UK). A Cx3crl^CrERT2;R26-td-Tomato on a C57BL/6J background (provided by Clemens Lange, University of Freigburg, Germany) were established as homozygotes at the University of Bristol Animal services Unit, UK, and offspring further crossed with C57BL/6J mice to generate heterozygotes for experiments. Cx3crl^CrERT2;R26-td-Tomato were treated topically to the eye with tamoxifen drops (5 mg/mL; Bell et al. 2020) from 4 - 6 weeks of age. Equal numbers of male and female mice were used.

For in vivo imaging optical coherence tomography (OCT) and light exposure to induce retinal damage mice were anaesthetised by inhalation of 1.5% v/v isoflurane. To perform intravitreal injections mice were anaesthetised using intraperitoneal injection of 90 pL/10 g body weight of a solution of Ketavet (Ketamine hydrochloride 100 mg/mL; Zoetis Ireland Ltd., Dublin, Ireland) and Rompun (Xylazine hydrochloride 20 mg/mL; Bayer PLC, Newbury, UK) mixed with sterile water in the ratio of 0.6: 1 :8.4 respectively. For all procedures, pupils were dilated with topical application of 1% v/v tropicamide. OCT scans of retina and fundal images were captured using Micron IV (Phoenix Research Laboratories, Pleasanton, CA, USA). Administration of recombinant mouse IL-33 (1 ng/pL; ALX- 522-101-C010, Enzo Life Sciences Ltd, Exeter, UK) or vehicle control was delivered via a 2 pL intravitreal injection. Vehicle consisted of PBS and 5% v/v fetal calf serum (Thermo Fisher Scientific, UK) as a carrier for IL-33. Adeno- associated viral vectors were delivered by a 2 pL intravitreal injection (diluted in PBS) at doses lel2, lell, lelO gc/mL (for light damage model and expression confirmation) or 2el2, 2el0, 2ell gc/mL (laser choroidal neovascularisation model).

Light-induced retina damage (LIRD) was performed by unilateral exposure to 100k LUX for 20 mins using a topical endoscope fundal imaging system and light box. C57BL/6J mice injected with either recombinant mouse IL-33 protein or vehicle 16 - 24 h prior were exposed at 7 weeks age or Cx3crl^CrERT2;R26-td- Tomato injected with either AAV2.truncatedIL-33(human) or AAV2.null 19 days earlier were exposed at 9 - 11 weeks age.

Laser choroidal neovascularisation was performed as described in Gong et al. 2015 using 7-week-old C57BL/6J mice either naive or injected with AAV2.truncatedIL-33(human) or AAV2.null 19 days earlier.

Lesion size and retinal thickness analysis

OCT line scans were collected through each lesion 7 days after laser treatment. FIJI (Schindelin et al. 2009) was used to determine lesion size; scale was set to vertical 100 um scale bar and the lesion drawn around to measure area. Each lesion is counted as a replicate.

Retinal thickness was measured from OCT scans taken through the optic disc using FIJI. Scale was set to the vertical image scale bar and distance measured from top of RPE layer to top of the nerve fibre layer at 200 and 400 micron from the optic nerve head.

Protein lysates and ELIS As Eyes were collected from animals injected with RPE lysates were prepared as previously described (Wei et al. 2016), using Pierce® RIPA buffer (89900; Thermo Fisher Scientific, UK) and protease inhibitor (5872S; Cell Signalling Technologies, UK). Whole retinas were crushed in 150 pL Pierce® RIPA buffer (89900; Thermo Fisher Scientific, UK) with protease inhibitor (5872S; Cell Signalling Technologies, UK) for protein extraction. Protein concentrations were quantified using Pierce™ BCA protein assay kit (23227, Thermo Fisher Scientific, UK).

Expression of AAV2.tIL-33(human) in retina or RPE lysates or secretion to media supernatants of transduced ARPE-19 cells was measured using DuoSet ELISA human IL-33 kit (DY3625B-05; R&D systems) according to manufacturer's instructions). A total of 12.5 pg (RPE lysates) or 20 pg (retina lysates) protein was loaded per well and for supernatants a total of 50 pL were loaded per well. All samples were measured in duplicate.

For measurement of AAV.tIL-33(mouse) expression and secretion of IL-33 from ARPE-19 cells was the DuoSet ELISA mouse IL-33 kit (DY3626-05; R&D systems) was used with 100 pL media supernatants loaded per well (in duplicate).

RNA extraction, cDNA synthesis and qPCR

RNA extracted from retina or RPE/choroid using Trizol reagent (15596026, Invitrogen, UK) according to manufacturer's instructions. RNA measured on nanodrop and equal amounts used (300 ng RPE/choroid, 1000 ng retina) for cDNA synthesis. RNA was first treated using DNase I (amp grade 18068-015, Thermo Fisher, UK) and reverse transcribed using GoScript (A2801, Promega, UK) reverse transcriptase with random hexamer primers (according to manufacturer's instructions).

Quantitative PCR was performed using PowerUp Sybr (A25742, ThermoFisher Scientific, UK) according to manufacturer's instructions. Primer sequences are listed in Table 1. Target genes were normalised to housekeeping gene (3-actin (or P-actin and 18S) using the delta-delta Ct method (Livak and Schmittgen 2001).

Cloning and AAV production

For expression of secreted mouse IL-33, we used a previously identified truncated sequence corresponding to amino acid 68 - 266, known to secrete from the cell (Xi et al. 2016, Nguyen et al. 2019). The sequence was PCR amplified from IL-33 ORF plasmid (NM133775; MR227227, Insight) with additional start codon sequence (ATG) and cloned into an AAV plasmid backbone with CMV promoter and WPRE regulatory sequence (AAV. CMV. tIL-33(mouse). WPRE. io2; Figure 9A). A suitable control vector was generated by removal of any coding sequence to produce a null plasmid vector, AAV. CMV. WPRE. io2.

For an equivalent truncated human IL-33, a construct was designed to express the corresponding homologous sequence (65 - 268 aa; Figure 9B). Plasmid vectors AAV.CAG.tIL-33.WPRE, AAV.CAG. IL-33. WPRE and control vector AAV. CAG.ORF-stuffer. WPRE were designed using VectorBuilder design software and subsequent cloning and AAV2 and AAV5 vectors were produced by VectorBuilder Inc. Viral titres were achieved by qPCR of the ITR sequence.

The mouse tIL-33 and null AAV2 serotype particles were produced by a triple transient transfection method. The AAV plasmid construct and capsid serotype vector (AAV2/2) were transfected at a ratio of 1: 1 :3 to the HGTI helper plasmid, mixed with polyethylenimine (75 pg/mL; Sigma-Aldrich, UK) and incubated at room temperature for 15 minutes. At 56 h post transfection, cells were harvested by scraping, concentrated and lysed by freeze-thaw (3x) to release particles from cells. Viral particles were purified by iodixanol gradient and concentrated by centrifugation (using MWCO 100 kDa Amicon filter units; Sigma-Aldrich, UK). Virus was titred by qPCR of the WPRE sequence.

Table 1. Primers for qPCR analysis

Cell culture

Immortalized human retinal pigment epithelium (ARPE-19; ATCC number CRL- 2302) were maintained at 37°C with 5% CO2 in DMEM (4.5 mg/L) supplemented with 10% heat-inactivated FBS, 2 mM l-glutamine, 1 mM sodium-pyruvate, 0.5 pM 2-mercaptoethanol, 100 U/mL penicillin, and 100 pg/mL streptomycin.

For AAV transduction, cells were plated in a Corning 96-well culture plate at 4500 cells/well in normal culture media. After 12 - 24 h, cells were washed and replaced with FCS free culture media containing appropriate concentration of AAV to achieve an MOI of 50000 (in-house produced AAV) or 70000 (VectorBuilder produced AAV). Cells were washed in Dulbecco's PBS and received fresh FCS free culture media at either 48 h or 6 - 8 h after transduction for subsequent MOIs. To induce oxidative stress, cells were treated with H2O2 at the specified concentration in figures, diluted from 30% H2O2 (Sigma-Aldrich, UK), again in FCS free culture media. Cells underwent assay measurement at day 3 post H2O2 treatment, with a half media change performed at 48 h after treatment.

Lactate dehydrogenase (LDH) assay

To measure cytotoxicity, cells were assessed for LDH release into supernatants using the LDH cytotoxicity WST assay kit (ENZ-KIT157, Enzo Life Sciences Ltd, Exeter, UK), according to manufacturer's instructions with the following changes; 25 pL of media to 25 pL of working solution and 12.5 pL stop solution was used per well. Samples were run in duplicate and normalised to a control well, after subtracting media background OD reading, to give a normalised OD (abs 490 nm).

PCA clustering analysis and statistics

PCA clustering analysis was performed in R studio (R version 3.6.1; RStudio, Inc., Boston, MA). All statistical analyses were performed using GraphPad Prism (version 8.4.3; GraphPad Software Inc. San Diego, CA). Unless otherwise stated data is expressed as ±SEM.

Results

Light-induced oxidative stress causes retinal degeneration and increased monocyte recruitment in vivo

To model disease pathways of AMD, we developed a variation of the LIRD model (Cachafeiro et al., 2013; Rattner et al., 2008), which emulates oxidative stress in AMD. Mice were exposed to 100k LUX light unilaterally for 20 mins (Figure 10A) to induce light mediated oxidative stress in the retina. Subsequent disease phenotype of retinal degeneration was evident by day 14 with significant retinal thinning observed in the light-exposed eye compared to naive eyes (Figure 10B and 10C). The LIRD model also displayed increased number of monocytes (IBA1 + cells) in retinal flatmounts at day 24 (Figure 10D and E). These data demonstrate capturing of key disease phenotypes from AMD, retinal degeneration and inflammation (Kauppinen et al., 2016), in the LIRD model.

Early gene expression changes after light exposure demonstrate stress response in vivo

To understand the biological changes that drive the stress response in a light- exposed eye, we further investigated gene expression changes in the retina and RPE/choroid, focusing on complement, inflammasome and oxidative stress response pathways perturbed in AMD (Copland et al., 2018; Kaarniranta et al., 2020; Lee et al., 2021; Pool et al., 2020) and other light-damage models (Rattner et al., 2008). Retina and RPE/choroid samples were collected from eyes at 6 h, 1- day, 3- and 7-days after light exposure and expression of 12 genes were assessed, including IL-33, inflammasome associated genes Ccl2, Illb), complement genes C3arl, C3, Cis), oxidative stress related genes (Hmoxl, Sodl, Prdxl) and metabolism genes Pgcla, Mcpl, Gapdh). PCA clustering analysis of the results revealed the greatest separation from naive eyes at day 1 in retina (PCI; Figure 11A) and 6 h and day 1 in RPE/choroid (PCI; Figure 11C). In the retina, the separation of day 1 samples was driven by significant increases in complement, inflammasome and oxidative stress related genes, including C3arl, Ccl2, Illb and Hmoxl (Figure 11B). In RPE/choroid, significant changes were observed in inflammasome and oxidative stress related genes, including Cc/2, Illb and Hmoxl (Figure 11D). These data demonstrate key early changes to pathways implicated in AMD, in response to light exposure.

Treatment with recombinant IL-33 slows rescues disease phenotypes in a model of light-induced oxidative stress

We have previously shown that treatment with recombinant IL-33 in a model of outer retinal degeneration Cfh+/- mice fed a high fat diet), can rescue RPE cell death and modulate expression of metabolism associated protein HKII in the retina (Clare et al., 2020). Furthermore, we have found that pre-treatment of RPE cells in vitro with recombinant IL-33 can protect the cells against H2O2 induced oxidative stress (Scott et al., 2021). Here, we hypothesised that pre-treatment of mice with IL-33 (1 ng/pL) would confer protection against light-induced oxidative stress in our LIRD model. Mice were injected intravitreally with either a vehicle control or IL-33 (1 ng/pL) 16 - 24 h prior to light exposure and monitored for development of disease phenotypes; retinal thinning, monocyte recruitment and early gene expression changes.

The earliest signs of significant retinal thinning were seen at day 14 post lightexposure in the LIRD model, and both vehicle and IL-33 treated eyes showed significant thinning when compared to injected only controls (Figure 12Ai). However, at day 21 after light-exposure vehicle treated eyes displayed continued and significant atrophy, whilst retinal thinning in IL-33 treated eyes had slowed (Figure 12Aii). Analysis of monocytes in the retina by IBA1 immunostaining, revealed a significant increase in IBA1+ cells at day 24 in vehicle treated LIRD eyes compared to IL-33 or vehicle injected eyes only (Figure 12B). Conversely, no significant increase was seen in IL-33 treated LIRD eyes and an early decline in IBA1+ cell numbers is evident (Figure 12B). Additionally, qPCR analysis and clustering analysis in RPE/choroid of key genes Hmoxl, Sodl, Pgcla, Ccl2, Illb and Gapdh) demonstrated a clear separation of vehicle treated LIRD eyes only from naive, IL-33 or vehicle injected and IL-33 treated LIRD eyes (Figure 12C). Together these data demonstrate that pre-treatment with recombinant IL-33 can modulate disease phenotypes and ultimately protect the retina from progressive degeneration.

Truncated IL-33 is secreted and reduces stress response to hydrogen peroxide in ARPE-19 cells

Treatment with recombinant IL-33 has demonstrated huge potential in protecting against AMD-like disease phenotypes in two different models of degeneration (LIRD model and Clare et al., 2020). However, recombinant IL-33 protein has a short half-life and treatment of the Cfh+/- HFD mice required regular injections, which would be a necessary translation to clinic if provided in this format. We therefore considered the potential of utilising AAV gene therapy to deliver a secretable form of IL-33 as a 'one-time' injection. Truncating mouse IL-33 to remove the nuclear localisation signal, has previously shown translocation of IL- 33 to the cytoplasm and release from cells (Nguyen et al., 2020; Xi et al., 2016). According to these findings, we developed two AAV constructs to express either the truncated mouse IL-33 under a CMV promoter or a homologous truncated version of human IL-33 under a CAG promoter (Figure 9).

Addition of human truncated IL-33 expressing AAV, packaged with either AAV2 or AAV5 capsids, to ARPE-19 cells was non-toxic (Figure 13Ai) and resulted in secretion into the supernatant by day 3 (Figure 13Aii). Similarly, AAV treated ARPE-19 cells with truncated mouse IL-33 (AAV2 packaged) was non-toxic (Figure 13Bi) and resulted in secretion to the supernatant (Figure 13Bii). We have previously observed treatment of ARPE-19 cells with recombinant IL-33 can protect against H2O2 induced cytotoxicity (Scott et al., 2021). Here, we treated AAV transduced cells (day 3 or 4 after AAV treatment) with 60 pM H2O2 and 3 days later measured cytotoxicity by LDH presence in the supernatant. Cells receiving the AAV2.null vector had increased cytotoxicity with H2O2 exposure compared to AAV2.null and AAV2.tIL-33(human) controls, whilst cytotoxicity was reduced in AAV2.tIL-33(human) treated cells exposed to H2O2 (Figure 13C). AAV2.tIL-33(mouse) treated cells demonstrated a similar trend in protection against H2O2 (Figure 13D).

AAV delivery of human truncated IL-33 modulates in vivo stress response in retina and RPE/choroid in a dose dependent manner

Having demonstrated secreted expression and protection against oxidative stress by truncated IL-33 in vitro, we wanted to investigate the effects of AAV.tIL-33 treatment in our LIRD model. High-, mid- and low-range titres of AAV2.tIL- 33(human) or AAV2.null (le9 - le7 gc/eye) were injected intravitreally and expression confirmed in the retina by ELISA, 21 days post injection (Figure 15). The same titres were injected 19 days prior to light exposure and eyes collected for qPCR analysis of key genes one day after light exposure. Analysis of complement C3arl), inflammasome Ccl2, II lb) and oxidative stress response genes (Hmoxl) in the retina revealed modulation and recovery of the stress- induced increased expression at the mid-range dose of AAV2.tIL-33(human) (Figure 14A), which was lost at high or low-doses. The ability of AAV2.tIL- 33(human) to perturb these responses is key, with dysregulated and overactive complement and inflammasome responses driving pathogenesis in AMD (Ambati & Fowler, 2012) and excessive Hmoxl expression in the retina can promote photoreceptor degeneration (Li et al., 2021).

In the RPE/choroid, we observed no effect for AAV2.tIL-33(human) treatment on ILlb expression. However, at mid-range dose AAV2.tIL-33(human) increases Hmoxl and Cc/2 expression after light-exposure (Figure 14B). These could be critical responses to promote recovery, as Hmoxl is a key defence enzyme against oxidative stress (Gozzelino et al., 2010) and may be important for RPE recovery. Increased Cc/2 expression in RPE/choroid will promote recruitment of monocytes, which could be detrimental in the event of exacerbated inflammation. However, early recruitment may promote a faster inflammation resolution and better long-term outcome. Certainly, IL-33 is known to be essential for inflammation resolution in a retinal detachment model of retinal degeneration (Augustine et al., 2019).

Together these data demonstrate a gene therapy delivery of tIL-33 modulates the response to light-induced oxidative stress, in a dose-dependent manner, reducing expression of inflammatory genes in the retina and promoting expression of oxidative stress response genes in the RPE/choroid.

Truncated IL-33 gene therapy reduces lesion size in the laser choroidal neovascularisation model of 'wet' AMD and is dose dependent

In 15% of cases AMD patients will develop neovascularisation ('wet' AMD), which without treatment will lead to acute vision loss (Clare et al., 2021). In a laser choroidal neovascularsation (L-CNV) model, we have previously demonstrated significant reduction in lesion formation when treating with recombinant IL-33 (Theodoropoulou et al., 2017). Here, we investigated whether AAV delivery of secreted tIL-33 could provide similar protection against lesion formation in the L- CNV model. Mice were injected with either a high-, mid- or low- titre (2e9 - 2e7 gc/eye) of AAV2.tIL-33(human) or AAV2.null vectors 19 days prior to inducing the L-CNV model. Animals injected with a mid-range titre (2e8 gc/eye) of AAV2.tIL- 33(human) virus had a significant reduction in lesion size compared to the equivalent titre of AAV2.null (Figure 16A and B). This protection was lost at high and low range titres, indicating a dose dependent role. Similarly, reduction in lesion size with recombinant IL-33 treatment was dose dependent, with higher doses having no effect (Theodoropoulou et al., 2017).

Conclusions

Our results demonstrate effective protection by recombinant IL-33 against light- induced retinal degeneration, providing further support for the application of IL-33 in treating AMD-like pathology. Moreover, delivery of secreted IL-33 can be achieved in vitro and in vivo by AAV providing the opportunity to utilise gene therapy in the application of IL-33 treatment in AMD. We further provide evidence that AAV delivered secreted IL-33 effectively protects against disease phenotypes using in vivo models for both 'dry' AMD (LIRD oxidative stress model) and 'wet' AMD (L-CNV model).

Example 4

Genetic modulation of IL-33

Knockdown of IL-33 from ARPE-19 cells was achieved using the fast-forward transfection technique. Cells were seeded at a concentration of 55,000 per well of a 24-well plate in 0.5 mL of culture medium with 1% FCS and no antibiotics. Cells were incubated for 1 hour at 37°C prior to transfection. The FlexiTube Gene- Solution (QIAGEN), as a specific mixture of 4 preselected siRNA duplexes, was used to target different sequences of the human IL-33 gene. Each siRNA was diluted in 100 pL of culture medium without serum and antibiotics (final concentration 20 nM each siRNA). HiPerfect transfection reagent (6 pL) was added to the siRNA, which was then vortexed and left for 5 minutes. Transfection complex (OriGene) (100 pL) was added to the cells and left for 48 hours at 37°C.

For IL-33 overexpression in ARPE-19 cells, a CRISPR/Cas9 activation plasmid (Santa Cruz Biotechnology) was used to upregulate the expression of the human IL-33 gene. The CRISPR/synergistic activation mediator (SAM) transcription activation system enabled the upregulation of IL-33 by utilising a D10 deactivated Cas9 nuclease fused to a VP64 activation domain, in conjunction with a single guide RNA (MS2) and an IL-33-specific single guide RNA engineered to bind the MS2-p65-HSF-l fusion protein.

Cells were seeded at a concentration of 40,000 per well of a 24-well plate in 0.5 mL of culture medium with 10% FCS and no antibiotics. Cells were incubated overnight at 37°C prior to transfection. Medium was replaced just before transfection. For each transfection, 0.16 pg of plasmid DNA was diluted into 25 pL plasmid transfection medium. Separately, 0.833 pL of transfection reagent was diluted in 25 pL plasmid transfection medium. Both solutions were left for 5 minutes before being combined, mixed, and left for a further 30 minutes. Transfection complex (50 pL) was added to the cells and left for 48 hours at 37°C.

Role of nuclear IL-33 in mitochondrial metabolism.

Following the identification of metabolic perturbations associated with IL-33 absence in vitro and ex vivo, we examined whether increased IL-33 expression in the RPE would have an adverse impact on cell metabolism. Using a CRISPR/Cas9 activation plasmid, the expression of the human IL-33 gene was observed to be upregulated in ARPE-19 cells. The effectiveness of CRISPR/Cas9-mediated overexpression of IL-33 was confirmed both at the RNA and protein level in whole cell lysates (Figure 17, A and B). Because IL-33 functions as a dual-function cytokine, residing within the nucleus and acting extracellularly, it was necessary to identify the subcellular location of IL-33 following overexpression. Increased IL- 33 was observed in the nuclear subcellular fraction (Figure 17C). A mitochondrial stress test was used to identify changes in OXPHOS parameters associated with IL-33 overexpression, and this showed increased maximal (FCCP-induced) respiration rates (Figure 17, D and E). Glucose- starved cells were subjected to a glycolysis stress test to identify changes in glycolytic parameters, and IL-33 overexpression was found to significantly increase glycolytic metabolism (Figure

17, F and G). This was accompanied by increased gene expression of glycolytic and TCA cycle enzymes (Figure 17H). Protein expression of GLUT1, PC, and PKM2 was confirmed at the protein level (Figure 17, I and J). We finally confirmed alterations to mitochondrial function were accompanied by structural alterations. Image analysis indicated that mitochondria in the IL-33 overexpression group formed long elongated tubules (Figure 17K), a structural phenotype associated with increased OXPHOS in T cells.

Nuclear IL-33 promotes oxidative glucose metabolism.

Because we observed an increase in ECAR (data not shown) and extracellular lactate (data not shown) in the absence of IL-33, we investigated whether IL-33 might regulate pyruvate import into the TCA cycle. Pyruvate enters the mitochondria through the mitochondrial pyruvate carrier complex (MPC) consisting of components MPC1/2. Overexpression of IL-33 led to increased MPC1 expression, at mRNA and protein levels, but had no significant effect on MPC2 expression (Figure 18, A and B). IL-33 KD reduced the expression of both MPC components at the gene and protein levels (Figure 18, C and D). Hence, we performed a modified mitochondrial stress test with the additional injection of the MPC inhibitor UK5099 to assess the contribution of aerobic glucose metabolism to total OCR. In ARPE-19 cells with IL-33 overexpression, the increased maximal respiration was largely due to increased pyruvate-dependent respiration (Figure

18, E and F). When a similar experiment was performed on IL-33 siRNA cells, it was observed that pyruvate-dependent respiration had a significantly reduced contribution to maximal OCR (Figure 18, G and H).

To assess if pyruvate metabolism was the only component affected by the altered expression of IL-33, or if other metabolic pathways feeding into the TCA cycle were involved, a similar experiment was conducted to assess fatty acid oxidation (FAO). Etomoxir treatment after mitochondrial uncoupling significantly reduced the OCR; however, between control plasmid and IL-33 plasmid groups, this reduction was not significant (data not shown). With IL-33 loss, we observed that FAO was significantly increased (data not shown), suggesting that FAO is upregulated in the absence of IL-33 to compensate for defects in pyruvate metabolism. SITA with [U-13C]-glucose was conducted to further assess how IL- 33 altered glucose metabolism in RPE. In ARPE-19 cells with IL-33 overexpression, C13 labeling indicated increased glycolytic flux (data not shown). No significant changes were observed in glycolysis "endpoint" metabolites pyruvate or lactate (data not shown); however, substantial increases in C13 enrichment were observed in TCA metabolites (Figure 181), indicating that glucose-derived TCA cycle activity was upregulated with IL-33 overexpression. In contrast, ARPE-19 cells with IL-33 KD had a significant increase in the abundance of C13 lactate (data not shown) and decreased C13 enrichment in malate, citrate, and succinate (Figure 18J).

Overexpression of IL-33 led to an increase in the fully labeled citrate mass isotopolog (M+6) (Figure 18K). This pattern will occur in citrate when both oxaloacetate and malate are derived from glucose. Increased M+4 aspartate (data not shown) and malate (Figure 18L) indicate the increased TCA cycling, which occurs with IL-33 overexpression. Labeling patterns using C13-labeled glucose then highlighted differential flux from glycolysis and pyruvate input into the TCA cycle. Glucose-derived pyruvate can enter the TCA cycle through PDH or PC (data not shown). The citrate M+ 2/ pyruvate M+3 ratio can serve as a surrogate for PDH activity, while the citrate M+3/ pyruvate M+3 ratio is used as a surrogate of PC activity. IL-33 KD significantly reduced the citrate M+3/pyruvate M+3 (data not shown) ratio, suggesting a decrease in the activity of the PC complex. Although no significant changes were observed in the citrate M+3/pyruvate M+3 ratio with IL-33 overexpression (data not shown), there was a significant increase in the citrate M+2/pyruvate M+3 ratio (Figure 18N), suggesting that PDH activity was augmented with IL-33 plasmid treatment. Increased PDH activity was supported by reduced PDH El phosphorylation status (Figure 18M).

Glutamate labeling was unaffected by IL-33 overexpression (data not shown). However, we show that IL-33 knockdown reduced derived C13 labeling of the M+2 mass isotopolog in glutamate pools (data not shown). We found no observable C13 labeling detected in a- ketoglutarate pools (data not shown). The decrease in unlabeled a -ketoglutarate observed with IL-33 overexpression (data not shown) suggests that the glutamine metabolism is likely reduced as increased glucose-derived carbon is used to support the TCA cycle. The increase in unlabeled glutamate and a -ketoglutarate (data not shown) suggest that glutamine-derived carbon may support the TCA cycle when glucose metabolism is impaired.

Taken together these results indicate that nuclear IL-33 is a critical regulator of pyruvate oxidative metabolism in the RPE. When overexpressed, there is increased glycolytic flux into the TCA cycle most likely through increased MPC and PDH activity. The absence of IL-33 reduces the oxidative catabolism of glucose, and as pyruvate is "redirected" to lactate, FAO appears to support the TCA cycle.

Discussion

These data provide evidence for a potentially novel role of intracellular IL-33 as a regulator of RPE metabolism. IL-33 loss increases aerobic glycolysis at the expense of oxidative glucose catabolism. Cells overexpressing IL-33 display increased expression of MPC1 while activating PDH (through dephosphorylation) to facilitate increased pyruvate flux into the TCA cycle. The identification of IL-33 as a key regulator of mitochondrial metabolism suggests roles for this cytokine that go beyond its extracellular "alarmin" activities. For example, when RPE is under stress, IL-33 contributes to minimize the effects of oxidative damage to the RPE and bolster mitochondrial metabolism. IL-33 exerts control over mitochondrial respiration in RPE by facilitating pyruvate import into mitochondria via upregulation of MPC expression and may be associated with the capacity of RPE to maintain homeostasis. Therefore, as well as identifying a molecular pathway for activation of mitochondrial respiration in RPE, our results demonstrate that intrinsic cellular IL-33 acts as a metabolic regulator exerting profound effects on retinal metabolism. Our results support augmenting IL-33 expression to combat oxidative damage and bolster mitochondrial metabolism in RPE, both pathologies known to be associated with AMD.

References

Altschul et al. (1990) J. Mol. Biol. 215: 403-10

Ambati, J., & Fowler, B. J. (2012). Mechanisms of age-related macular degeneration.

Neuron, 75(1), 26-39. https://doi.Org/10.1016/j.neuron.2012.06.018

Augustine, J., Pavlou, S., Ali, I., Harkin, K., Ozaki, E., Campbell, M., Stitt, A. W., Xu, H., & Chen, M. (2019). IL-33 deficiency causes persistent inflammation and severe neurodegeneration in retinal detachment. Journal of Neuroinflammation, 16(1), 1- 15. https://doi.org/10.1186/sl2974-019-1625-y

Bell, O. H., Copland, D. A., Ward, A., Nicholson, L. B., Lange, C. A. K., Chu, C. J., & Dick, A. D. (2020). Single Eye mRNA-Seq Reveals Normalisation of the Retinal Microglial Transcriptome Following Acute Inflammation. Frontiers in Immunology, lO(January). https://doi.org/10.3389/fimmu.2019.03033

Cachafeiro, M., Bemelmans, A. P., Samardzija, M., Afanasieva, T., Pournaras, J. A., Grimm, C., Kostic, C., Philippe, S., Wenzel, A., & Arsenijevic, Y. (2013). Hyperactivation of retina by light in mice leads to photoreceptor cell death mediated by VEGF and retinal pigment epithelium permeability. Cell Death & Disease, 4, 1-11. https://doi.org/10.1038/cddis.2013.303

Clare, A. J., Copland, D. A., Nicholson, L. B., Liu, J., Neal, C. R., Moss, S., Dick, A. D., & Theodoropoulou, S. (2020). Treatment with interleukin-33 is non-toxic and protects retinal pigment epithelium in an ageing model of outer retinal degeneration. Journal of Cellular and Molecular Medicine, 24(22), 13546-13550. https://doi.org/10.llll/jcmm.16000

Clare, A. J., Liu, J., Copland, D. A., Theodoropoulou, S., & Dick, A. D. (2021). Unravelling the therapeutic potential of IL-33 for atrophic AMD. Eye, July, 1-7. https://doi.org/10.1038/s41433-021-01725-5

Copland, D. A., Theodoropoulou, S., Liu, J., & Dick, A. D. (2018). A perspective of AMD through the eyes of immunology. Investigative Ophthalmology and Visual Science, 59(4), AMD83-AMD92. https://doi.org/10.1167/iovs.18-23893

Gong, Y., Li, J., Sun, Y., Fu, Z., Liu, C. H., Evans, L., Tian, K., Saba, N., Fredrick, T., Morss, P., Chen, J., & Smith, L. E. H. (2015). Optimization of an image-guided laser- induced choroidal neovascularization model in mice. PLoS ONE, 10(7), 1-15. https://doi.org/10.1371/journal.pone.0132643

Gozzelino, R., Jeney, V., & Soares, M. P. (2010). Mechanisms of cell protection by heme Oxygenase-1. Annual Review of Pharmacology and Toxicology, 50, 323-354. https://doi.org/10.1146/annurev.pharmtox.010909.105600

Kaarniranta, K., Uusitalo, H., Blasiak, J., Felszeghy, S., Kannan, R., Kauppinen, A., Salminen, A., Sinha, D., & Ferrington, D. (2020). Mechanisms of mitochondrial dysfunction and their impact on age-related macular degeneration. Progress in Retinal and Eye Research. https://doi.Org/10.1016/j.preteyeres.2020.100858

Kauppinen, A., Paterno, J. J., Blasiak, J., Salminen, A., & Kaarniranta, K. (2016). Inflammation and its role in age-related macular degeneration. In Cellular and Molecular Life Sciences, https://doi.org/10.1007/s00018-016-2147-8

Khalili et al. RSC Adv., 2018, 8, 35787

Lee, K. S., Lin, S., Copland, D. A., Dick, A. D., & Liu, J. (2021). Cellular senescence in the aging retina and developments of senotherapies for age-related macular degeneration. In Journal of Neuroinflammation, https://doi.org/10.1186/sl2974- 021-02088-0

Li, H., Liu, B., Lian, L., Zhou, J., Xiang, S., Zhai, Y., Chen, Y., Ma, X., Wu, W., & Hou, L. (2021). High dose expression of heme oxigenase-1 induces retinal degeneration through ER stress-related DDIT3. Molecular Neurodegeneration, 16(1), 1-17. https://doi.org/10.1186/sl3024-021-00437-4

Livak, K. J., & Schmittgen, T. D. (2001). Analysis of relative gene expression data using real-time quantitative PCR and the 2-AACT method. Methods, 25(4), 402-408. https://doi.org/10.1006/meth.2001.1262

Nguyen, P. T., Dorman, L. C., Pan, S., Vainchtein, I. D., Han, R. T., Nakao-lnoue, H., Taloma, S. E., Barron, J. J., Molofsky, A. B., Kheirbek, M. A., & Molofsky, A. V. (2020). Microglial Remodeling of the Extracellular Matrix Promotes Synapse Plasticity. Cell. https://doi.Org/10.1016/j.cell.2020.05.050

Pearson and Lipman (1988) PNAS USA 85: 2444-2448

Pool, F. M., Kiel, C., Serrano, L., & Luthert, P. J. (2020). Repository of proposed pathways and protein-protein interaction networks in age-related macular degeneration. Npj Aging and Mechanisms of Disease. https://doi.org/10.1038/s41514-019-0039-5

Rattner, A., Toulabi, L., Williams, J., Yu, H., & Nathans, J. (2008). The genomic response of the retinal pigment epithelium to light damage and retinal detachment. Journal of Neuroscience, 28(39), 9880-9889. https://doi.org/10.1523/JNEUROSCI.2401-08.2008 Schambach, A., et al. 2005. Woodchuck hepatitis virus post-transcriptional regulatory element deleted from X protein and promoter sequences enhances retroviral vector titer and expression. Gene Therapy, 13, 641.

Schindelin, J., Arganda-Carrera, I., Frise, E., Verena, K., Mark, L., Tobias, P., Stephan, P., Curtis, R., Stephan, S., Benjamin, S., Jean-Yves, T., Daniel, J. W., Volker, H., Kevin, E., Pavel, T., & Albert, C. (2012). Fiji - an Open platform for biological image analysis. Nature Methods, 9(7). https://doi.org/10.1038/nmeth.2019.Fiji

Scott, L. M., Dick, A. D., Theodoropoulou, S., Scott, L. M., Vincent, E. E., Hudson, N., Neal, C., Jones, N., Lavelle, E. C., Campbell, M., Halestrap, A. P., Dick, A. D., & Theodoropoulou, S. (2021). Interleukin-33 regulates metabolic reprogramming of the retinal pigment epithelium in response to immune stressors Interleukin-33 regulates metabolic reprogramming of the retinal pigment epithelium in response to immune stressors. The Journal of Clinical Invesitgations Insight, 6(8). https://doi.Org/https://doi.org/10.1172/jci.insight.129429.

Theodoropoulou, S., Copland, D. A., Liu, J., Wu, J., Gardner, P. J., Ozaki, E., Doyle, S. L., Campbell, M., & Dick, A. D. (2017). Interleukin-33 regulates tissue remodelling and inhibits angiogenesis in the eye. Journal of Pathology, 241(1), 45-56. https://doi.org/10.1002/path.4816

Wei, H., Xun, Z., Granado, H., Wu, A., & Handa, J. T. (2016). An easy, rapid method to isolate RPE cell protein from the mouse eye. Experimental Eye Research, 145, 450- 455. https://doi.Org/10.1016/j.exer.2015.09.015

Xi, H., Katschke, K. J., Li, Y., Truong, T., Lee, W. P., Diehl, L., Rangell, L., Tao, J., Arceo, R., Eastham-Anderson, J., Hackney, J. A., Iglesias, A., Cote-Sierra, J., Elstrott, J., Weimer, R. M., & Van Lookeren Campagne, M. (2016). IL-33 amplifies an innate immune response in the degenerating retina. Journal of Experimental Medicine, 213(2), 189-207. https://doi.org/10.1084/jem.20150894

Claims

43 Claims

1. IL-33 for use in the treatment, prevention or management of age-related macular degeneration (AMD).

2. IL-33 for use according to claim 1, wherein the IL-33 is exogenous IL-33.

3. IL-33 for use according to claim 1 or 2, wherein the IL-33 is recombinant IL-33.

4. IL-33 for use according to any of claims 1 to 3, wherein the IL-33 is provided in the form of a pharmaceutically acceptable composition.

5. IL-33 for use according to any of claims 1 to 3, wherein the IL-33 is provided in the form of an adeno-associated virus (AAV) vector gene therapy, wherein the viral vector comprises an IL-33 transgene encoding a truncated IL-33 polypeptide.

6. IL-33 for use according to claim 5, wherein the AAV vector comprises AAV2 ITRs and/or AAV2 capsid or a variant thereof.

7. IL-33 for use according to claim 5 or 6, wherein the truncated IL-33 polypeptide is human IL-3395-270 or human IL-33io9-2?o, human IL-33ii3-27o or human IL-3365-268.

8. IL-33 for use according to claim 1, wherein the IL-33 is provided in the form of a CRISPR-Cas9 activation system.

9. IL-33 for use according to claim 8, wherein the CRISPR-Cas9 activation system comprises a deactivated Cas9 nuclease fused to an activation domain and at least one IL-33-specific guide RNA.

10. IL-33 for use according to any of claims 1 to 3, wherein the IL33 is provided in the form of a nucleic acid encoding IL-33. 44

11. IL-33 for use according to any of claims 1 to 10, wherein the IL-33 is to be administered into an eye.

12. IL-33 for use according to any of claims 1 to 11 wherein the IL-33 is to be delivered by injection.

13. IL-33 for use according to any of claims 1 to 11, wherein the IL-33 is to be delivered topically.

14. IL-33 for use according to any of claims 1 to 13, wherein the AMD is early stage AMD, intermediate stage AMD or late stage AMD.

15. IL-33 for use according to any of claims 1 to 13, wherein the AMD is dry AMD.

16. A method of treating, preventing or managing age-related macular degeneration (AMD), the method comprising administering IL-33 to a patient in need thereof.

17. The method of claim 16, wherein the IL-33 is exogenous IL-33.

18. The method of claim 16 or 17, wherein the IL-33 is recombinant IL-33.

19. The method of any of claims 16 to 18, wherein the IL-33 is provided in the form of an adeno-associated virus (AAV) vector gene therapy and wherein the viral vector comprises an IL-33 transgene encoding a truncated IL-33 polypeptide.

20. The method of claim 19, wherein the AAV vector comprises AAV2 ITRs and/or AAV2 capsid or a variant thereof.

21. The method of claim 19 or 20, wherein the truncated IL-33 polypeptide is human IL-3395-270 or human IL-33io9-2?o, human IL-33ii3-27o or human IL-3365-268.

22. The method of claim any of claims 16 to 18, wherein the IL-33 is provided in the form of a CRISPR-Cas9 activation system. 45

23. The method of claim 22, wherein the CRISPR-Cas9 activation system comprises a deactivated Cas9 nuclease fused to an activation domain and at least one IL-33-specific guide RIMA.

24. The method of any of claims 16 to 18, wherein the IL33 is provided in the form of a nucleic acid encoding IL-33.

25. The method of any of claims 16 to 18, wherein the IL-33 is provided in the form of a pharmaceutically acceptable composition.

26. The method of any of claims 16 to 25, wherein the IL-33 is administered into an eye of the patient.

27. The method of any of claims 16 to 26, wherein the IL-33 is delivered by injection.

28. The method of any of claims 16 to 26, wherein the IL-33 is delivered topically.

29. The method of any of claims 16 to 28, wherein the AMD is early stage AMD, intermediate stage AMD or late stage AMD.

30. The method of any of claims 16 to 29, wherein the AMD is dry AMD.

31. An adeno-associated virus (AAV) vector gene therapy, wherein the viral vector comprises an IL-33 transgene encoding a truncated IL-33 polypeptide.

32. The AAV vector gene therapy of claim 31, wherein the AAV vector comprises AAV2 ITRs and/or AAV2 capsid or a variant thereof.

33. The AAV vector gene therapy of claim 31 or 32, wherein the truncated IL- 33 polypeptide is human IL-3395-270 or human IL-33io9-2?o, or human IL-3365-268.

34. The AAV vector gene therapy of any of claims 31 to 33, further comprising a ubiquitous promoter or an ocular cell specific promoter.

35. The AAV vector gene therapy of claim 34, wherein the ubiquitous promoter is selected from CMV and CAG.

36. A CRISPR-Cas9 activation system comprising a deactivated Cas9 nuclease fused to an activation domain and at least one IL-33-specific guide RNA.

37. A CRISPR-Cas9 activation system according to claim 36, wherein the system is in the form of a CRISPR-Cas9 activation plasmid.

38. A CRISPR-Cas9 activation system according to claim 37, wherein the CRISPR-Cas9 activation plasmid comprises a DIO deactivated Cas9 nuclease fused to a VP64 activation domain, in conjunction with a single guide RNA and an IL-33-specific single guide RNA engineered to bind the MS2-p65-HSF-l fusion protein.