In my experience, there’s still work to be done to get good retinal gene expression, especially in the macula, following suprachoroidal delivery. A lot of the expression seems to be targeted at the RPE, which can be great for an RPE targeted program, but for photoreceptors, finding the right mix of vector and delivery method to get gene expression within the macula is pretty difficult from the suprachoroidal method. Whereas for subretinal, you can place that bleb right under the macula or fovea. You are risking doing damage to the retina with a subretinal injection, but that gene therapy is going to be directly exposed to the photoreceptors and RPE. So generally, you get pretty strong expression in both tissues.

That’s where the scoring in life is important on eye exams, the SPOT scoring or uveitis scoring, standardization of uveitis nomenclature scoring systems are important. It is a subjective assessment, but they try to make it objective by incorporating the scoring systems. But really the more objective evaluation of inflammation severity can be done histopathologically where you’re actually looking on a cellular level at how many lymphocytes are infiltrating into the tissues or how severe that inflammation is. It’s really a combination of having a scoring system for your eye exams, doing OCTs, and then histopathology really help tie together severity.

For gene therapy studies, collecting ADA samples, it’s pretty common to also do PBMC testing. So you collect blood and isolate peripheral mononuclear cells and then challenge those T cells to the antigen to see that. That way you can determine which specific antigen the immune system is reacting to. Generally, those samples will be taken at baseline and then at around 2 weeks to 4 weeks post injection when that adaptive immune response is occurring. One thing to note is, for both ADAs and PBMCs, if you do see inflammation, it would be ideal to collect a sample prior to initiating immunosuppressive treatment, because those immunosuppressives are blocking those pathways for antibody development and for T cell activation proliferation. So if you collect those samples two weeks or four weeks after immunosuppressive treatment, they may be impacted by the treatment itself.

It can really confound both safety and efficacy signals. One example being my master’s project evaluating the tropism of different AAV vectors following intravitreal injection in dogs. Some dogs developed severe retinal inflammation, and had very strong GFP signal the week prior to development of inflammation, with very little fluorescence remaining following development of inflammation. The inflammatory cells directly damage the target tissues that you’re trying to deliver your drug to or have expressed your protein in the case of gene therapy. This inflammation really has a huge impact on both efficacy endpoints and safety endpoints.

The suprachoroidal space permits higher volumes than intravitreal administration does. One thing to consider is the distribution of where the drug goes after administration. Generally, it tends to stay more peripheral. Generally volumes are 100 microliters or 150 microliters, but I’ve been involved in some studies that have pushed the volume up into the multiple hundreds, two to 300 microliter range. And in those cases, you start to see some leakage of the test material outside of the eye along the posterior pole. So say you do a suprachoroidal injection with a dye at that volume and then you dissect the eye right away, even though you don’t see the dye coming out where you injected it, there’s some leakage likely from where the vortex veins exit the sclera. So increasing the volume may not directly relate to increased intraocular exposure in a linear manner. Higher volumes can cause a substantial transient increase in IOP, but this typically returns to the normal range within 5-10 minutes.

ACAID is part of the natural ocular immune response, so I don’t think it really needs to be accounted for specifically from a translation perspective. If you have a low dose, a mid dose, and a high dose, and your mid dose and high dose is having signs of an adaptive immune response and your low dose has nothing, then you likely did not exceed the threshold for ACAID in your low dose, and you probably have some degree of immune tolerance that developed in your low dose group if you didn’t see any signs of inflammation.

The SPOTs system is basically a modification of the McDonald-Shadduck or Hackett McDonald systems combined with the SUN (standardization fo uveitis nomenclature) scoring system. It does a nice job of combining those preclinical animal model systems developed to assess ocular surface irritation with the human uveitis grading systems. It’s comprehensive. For practical application, I prefer to split the SPOTs system into two grading scales depending on the dose route. For example, if we are doing an intravitreal or subretinal gene therapy study, what we see most often is inflammation in the posterior segment. So, I might not do the scoring for conjunctival redness or swelling or corneal opacity or the anterior segment findings, and focus on scoring signs of intraocular inflammation. I like to tailor the inflammatory scoring to the tissue or compartment of interest. I might do something similar to the Hackett McDonald on a topical study, but for a retinal gene therapy study, we’ll use the posterior segment portions of the SPOT system along with the anterior chamber cell and flare scoring.

A big topic of conversation that I have with clients is local versus systemic immune suppression. If your drug is going to cause activation of the innate immune system only, the local environment reaction to the drug is causing inflammation, then a topical or a subconjunctival immunosuppressive may be all that you need on your study. But most of the studies that I work on end up with some form of an adaptive immune response, which is driven peripherally in the spleen and lymphoid organs, and those lymphocytes that are causing inflammation in the eye are moving there from the spleen and lymph nodes. In those cases, you really need a systemic immunosuppression regimen to prevent that proliferation of the antigen specific response.

As far as specific drug regimens, in the clinic, corticosteroids are still the mainstay treatment used, but it is common to combine corticosteroids with a non steroid immunosuppressive. And there are more and more non steroid immunosuppressives coming on the market. But despite that, corticosteroids are usually used in that initial phase after drug delivery. And that’s what we mimic in animal models as well. Corticosteroids are the mainstay. One thing to consider is that humans will most often voluntarily take oral medications, while animals are not often as compliant, so an injectable corticosteroid with a long duration of action is usually most appropriate in the animal model to achieve a consistent immunosuppressive effect. Adding in a non steroidal immunosuppressive sometimes is appropriate, especially cell therapy where you really need a good immunosuppressive effect. You might add something like cyclosporine or rapamycin or abatacept, something that specifically targets lymphocyte proliferation.

There’s not a great way to reproducibly perform visual acuity tests in large animals used on toxicology studies. Pattern ERGs are a possible way to assess visual acuity, but they provide lower amplitude responses and tend to have a high rate of variability, so it’s hard to get a definitive answer. A full field ERGis a measurement of the entire retina’s response to light, but does not provide information about visual perception or acuity. It’s really inferred from the full field flash ERG and histopathology and OCT assessments of the macula. If you see some photoreceptor loss within the fovea and macula or RPE degeneration or atrophy in that region, then you would expect an impact on visual acuity in the clinic.

The cynomolgus macaque has been the standard for retinal gene therapy because it has both a macula and fovea and most gene therapies are targeting the macula in humans. Minipigs are becoming more popular as an alternative to using NHPs. They don’t possess a macula, they instead have a visual streak, which is a region of high cone density that can serve as a surrogate for the macula, but doesn’t directly represent what happens in the human. So while NHPs are still preferred, Minipigs are an acceptable surrogate. Minipigs tend to be more immunologically reactive, that should be taken into consideration.

That’s been a bit of a topic of debate in the industry. Rabbits have a very thin retina to begin with, and following subretinal injection, outer retinal degeneration commonly occurs over the dose site, just from the procedure itself. So it’s not often used as a subretinal model, except in cases where you may want to see some retinal degeneration or photoreceptor loss within the dose site. Say you’re doing a subretinal photoreceptor cell replacement therapy, then it may be an appropriate model.

They should be more or less aligned even though they’re not the same. In humans, we can sit a human down and ask them to read an eye chart, or we can walk them through a maze and see how they do. With preclinical models, we don’t have those opportunities. We’re running a full field ERG, whereas the clinical trial might have visual acuity testing or multifocal ERG. The preclinical studies generally have less definitive testing. But eye exams and IOPs are going to be included in most clinical packages. And obviously, clinical packages aren’t going to have necropsy and histopathology.

Eye exams and IOPs are generally the minimum required endpoints. OCTs are a really nice surrogate for serial histopathology evaluations. In the past, prior to using OCT on preclinical studies, you’d often have serial endpoints, so a two week cohort and a one month necropsy cohort on a six month study to see histopathology at multiple timepoints to assess for toxicity. OCT allows you to get that assessment of retinal structure over time, so you may not need those earlier necropsy cohorts, and it decreases the number of models you need on studies. For drugs that may impact retinal function, full field ERGs are going to be required as well.

That depends on which arm of the FDA or regulatory body that you’re submitting to. If you’re submitting to CBER, then each package can be different as far as what’s minimally required. For example, a GLP safety study is going to be the minimal requirement for CBER in one large animal species. For CDER, having a GLP safety study in two species [that can be small and large animal or two large animal species] is going to be the minimum requirement. And combined with that GLP study (which will determine your NOAEL) is your preclinical non-GLP efficacy studies, which might be run at an academic institution. There are some cases where a GLP study might just be in a small research model, especially for gene editing studies where it may be impossible to have editing activity in a large animal model, so you might have a single transgenic rodent GLP study with or without then a smaller, large animal non GLP or GLP safety study.