Review 18 September 2022 , , and 1 Sorbonne Université, INSERM, CNRS, Department of Therapeutics, Institut de la Vision, 75012 Paris, France 2 Université Paris-Saclay, Univ Evry, Inserm, Genethon, Integrare Research Unit UMR_S951, 91000 Evry-Courcouronnes, France 3 Institut de la Vision, INSERM UMR S968, 17 rue Moreau, 75012 Paris, France * Author to whom correspondence should be addressed.
Inherited retinal diseases (IRDs) are a leading cause of blindness in industrialized countries, and gene therapy is quickly becoming a viable option to treat this group of diseases. Gene replacement using a viral vector has been successfully applied and advanced to commercial use for a rare group of diseases. This, and the advances in gene editing, are paving the way for the emergence of a new generation of therapies that use CRISPR–Cas9 to edit mutated genes in situ. These CRISPR-based agents can be delivered to the retina as transgenes in a viral vector, unpackaged transgenes or as proteins or messenger RNA using non-viral vectors. Although the eye is considered to be an immune-privileged organ, studies in animals, as well as evidence from clinics, have concluded that ocular gene therapies elicit an immune response that can under certain circumstances result in inflammation. In this review, we evaluate studies that have reported on pre-existing immunity, and discuss both innate and adaptive immune responses with a specific focus on immune responses to gene editing, both with non-viral and viral delivery in the ocular space. Lastly, we discuss approaches to prevent and manage the immune responses to ensure safe and efficient gene editing in the retina.
Inherited retinal degenerations (IRDs) are conditions of the retina caused by mutations in genes mostly expressed in the retinal pigment epithelium or photoreceptors. IRDs can result in severe visual impairment or complete vision loss, and are a social and economic burden [1]. There are ongoing efforts to develop therapies for IRDs that include pharmacotherapy, neuroprotection, gene therapy, optogenetic therapy, retinal prostheses and stem cell therapy. Many of these approaches have entered the clinical phase of development and some of them have translated into approved products. Currently, gene therapy is the most promising approach to treat IRDs caused by a single recessive gene defect [2]. Gene therapy refers to the use of genetic material to counteract an inherited or complex disease. If applied early in the disease, it can take the form of gene correction, gene replacement or augmentation, which have a common goal in replacing the mutant gene with a healthy copy either at the chromosomal site in situ or by providing an additional healthy copy that is maintained extra-chromosomally [3]. In the past decade, the Clustered Regularly Interspaced Short Palindromic Repeat (CRISPR)-Cas9 system has become one of the most powerful tools for precise gene editing and is being considered for retinal gene therapy applications [4,5]. While this holds immense therapeutic promise, it also presents some tool-specific challenges, including off-target editing and immune responses to the Cas9 protein and/or to the guide RNAs [4,6].
The therapeutic gene can be supplied to the eye by viral vector delivery or non-viral delivery. Recombinant versions of a common virus are generated by replacing the viral genome with the therapeutic gene of interest [7]. Many engineered viral vectors have been successfully used for gene therapy applications [8]. Among these, the Adeno-associated virus (AAV) is the vector of choice for the majority of retinal gene therapies, owing to their excellent transduction and safety profile. AAVs provide long-term transgene expression, induce a low host immune response, infect both dividing and quiescent cells and do not integrate into the host genome, thereby reducing the risk to patients [9]. However, the large size of IRD genes is currently a limiting factor for AAV-mediated gene augmentation as the AAV has a packaging limit of ~4.5 kb [10]. There are other viral vectors that have a larger packaging capacity, such as Lentiviruses (LVs), which can carry genes of up to 8 kb [11]; and Adenoviruses (Ads), which can reach a packaging capacity of 37 kb [12]. LVs are enveloped retroviruses containing a positive, single-stranded RNA genome, capable of infecting both dividing and non-dividing cells. Their major drawback is that their large size limits their diffusion and transduction capacity of the neural retina. Additionally, they are genome-integrating, and their integration can at times occur at an oncogenic locus [13]. Ads have a large packaging capacity, but they are highly immunogenic, thereby decreasing both the efficiency and safety of gene therapy [14,15,16].
Non-viral delivery is a broad-spectrum term used for many delivery methods that do not use viral packaging. These may include, but are not limited to, strategies that functionalize (by an addition of peptides or charges) proteins or DNA for direct delivery, use nanoparticles or use lipid-based coating. The intention of developing such techniques is to have larger packaging capacities or to deliver genes without packaging to avoid viral components—and hence, be less immunogenic—and for the ease of large-scale productions [17]. However, so far most of these strategies have reported low efficiency and could not accomplish transgene expression at therapeutic levels [18]. Additionally, as elaborated in the following sections of this review, immune responses induced by non-viral vectors have also been reported [19].
Most of the early advances in gene therapies were achieved in the eye, owing to the convenient tissue access and immune-privilege of the eye. In addition, this region also ensures the ability to monitor non-invasively, and ensures the existence of the contralateral eye as an in vivo control [8]. However, adverse outcomes from some clinical trials and evidence from animal studies have demonstrated that the immune-privilege of the eye is not absolute [20]. The introduction of foreign substances in the form of DNA, protein, chemical compounds and nanoparticles can elicit both innate and adaptive immune responses [17]. Innate immune responses induced by the vector, transgene or protein products can promote local inflammation in the eye, leading to deterioration of visual acuity. Innate immune responses can further boost adaptive immune responses to generate antibodies or cytotoxic T cells, which limit the transduction efficiency of the vector and/or clear the transduced cells (Figure 1) [21].
Figure 1. Schematic representation of immune responses to ocular gene therapy. Some components of viral and non-viral delivery methods present danger signals to the immune system. Exposure to these components prior to gene therapy can result in pre-existing immunity. Individuals without prior exposure can activate innate and adaptive mechanisms after receiving ocular gene therapy. Individuals with pre-existing immunity can generate neutralizing antibodies (NAbs) that eliminate the vector and further trigger innate and adaptive responses. Together, these responses can result in an adverse outcome of ocular gene therapy, such as clearance of transduced cells, inflammation and reduction in visual acuity.
In this review, we first discuss studies that have reported on pre-existing immunity towards viral vectors and the Cas9 protein. We then evaluate both innate and adaptive immune responses with a specific focus on immune responses to gene editing, both with non-viral and viral delivery in the ocular space, and finally discuss approaches to prevent and manage the immune responses to ensure an increase in safety and efficacy of viral or non-viral vector-mediated gene editing in the retina.
Pre-existing immunity generally refers to the presence of antibodies in a host against components of therapy which can interfere with the therapeutic intervention. The most widely used Cas9 orthologs are derived from Staphylococcus aureus (SaCas9) and Streptococcus pyogenes (SpCas9) [4], which are common human commensals that could sometimes become pathogenic [22,23]. Therefore, it is reasonable to assume the presence of pre-existing immunity against Cas9 proteins in the human population that would affect the efficacy and safety of therapies involving Cas9 proteins. A study involving 125 healthy adult human blood donors in the USA with a median age of 43 reported pre-existing anti-SaCas9 antibodies and anti-SpCas9 antibodies in 78% and 58% of the total samples, respectively. In this study, apart from the antibodies, pre-existing cellular immune response was also evaluated. The authors reported an increase in the cells releasing IFN-γ in peripheral blood mononuclear cells (PMBS) from 18 donors stimulated with SaCas9 and SpCas9. Among the tested donors, 78% had SaCas9-specific T-cells and 67% were positive for SpCas9-specific T-cells. Cytokine positive cells were observed by Intracellular Cytokine Staining (ICS) targeting IFN-γ, TNF-α and IL-2. The presence of activated T-cells was reported by the FACS sorting T-cell activation markers—CD137 and CD154 [24]. Another study conducted on 48 healthy donors also confirmed this finding of a high prevalence of T-cells against SpCas9 protein, demonstrating that SpCas9 stimulation could activate CD137+ and CD154+ T-cells [25]. A study that compared the pre-existing anti-SaCas9 and anti-SpCas9 antibodies between serum and vitreous liquid from 13 patients who received vitreoretinal surgery with vitrectomy concluded that all the three serum samples were positive for SaCas9 and SpCas9, while only two vitreous fluid samples tested positive for SaCas9 and two others for SpCas9. This study also showed that the antibodies’ levels were higher in serum than in vitreous liquid [26].
An animal study conducted on 78 adult canines (60 WT and 18 models for Duchenne muscular dystrophy (DMD)) reported pre-existing anti-SpCas9 antibodies in all animals, but not in 16 newborn puppies, which only had moderate levels of maternally-derived Cas9 antibodies that dropped between two and six weeks of age. This study indicated that older animals in a study are more likely to be positive for anti-Cas9 antibodies than the younger ones [5].
AAVs are widely spread in the natural environment, and it has been reported that about 50–90% of the human population have been exposed to different AAV serotypes [27]. Exposure results in the presence of anti-AAV antibodies, which can further trigger a stronger response upon a second exposure, such as in the case of AAV-mediated gene therapy (Figure 1). A study conducted to evaluate pre-existing anti-AAV antibodies in 226 healthy donors between the ages of 25 and 64 years demonstrated that there is a prevalence of serum IgG to AAV in the healthy population. The highest seroprevalences were observed with AAV2 (72%) followed by AAV1 (67%), AAV9 (47%), AAV6 (46%), AAV5 (40%) and AAV8 (38%) [28]. Another study evaluated NAbs in 888 human serum samples from healthy volunteers from 10 countries. NAb assays were performed to detect antibodies against AAV1, AAV2, AAV7, AAV8 and AAVrh32.33 (a novel, structurally distinct AAV variant). The results showed that the highest prevalence was against AAV2, even though there was viability in samples from different continents. The anti-AAV1 and anti-AAV7 NAb levels were similar, followed by anti-AAV8, and not surprisingly, the lowest NAb titers were against the novel variant AAVrh32.33 [29]. In most human studies, the prevalence was highest against AAV2, while the lowest prevalence was against AAV8. However, a study conducted on 41 non-human primates (NHP) revealed that the highest level of antibodies was against AAV8 and AAV9 [30]. In humans as well as NHPs, the pre-existing antibodies against AAV5 tend to be lower [28,30]. Hence, there is pre-existing immunity against AAV among humans and animals used in research. However, their levels show variations by species, region and age. These findings are plausible, since the older the individual, the higher the probability is of them encountering the corresponding wild-type microorganism, thus, triggering the pre-existing adaptive immunity. Moreover, it is not surprising that the concentration of anti-Cas9 or anti-AAV immunoglobulins is lower in intraocular fluids due to the existence of ocular barriers, such as the blood-retinal barrier. However, a lower antibody concentration in the vitreous humor should not be underestimated because it is not yet clear whether a low level of Cas9 or AAV-specific antibodies can have a negative impact on the transduction efficiency and/or on the triggering of a local inflammatory response.
CRISPR-associated proteins (Cas) are present in most archaea and bacteria as adaptive immune systems [31] and provide sequence-specific resistance against phages [32]. A CRISPR–Cas system contains Cas proteins and small guide RNA sequences (sgRNA) that guide the Cas proteins to specific DNA binding sites that contain specific protospacer adjacent motif (PAM) sequences. This system has been adapted for a wide array of gene-editing applications, including gene therapy [33]. As an exogenous protein, Cas9 can be regarded as a foreign antigen in animals, and hence, can be presented by antigen-presenting cells (APC) to lymphocytes to induce host immune responses (Figure 1) [34].
The presence of ocular barriers allows the eye to limit inflammation, and therefore, makes it a good target for CRISPR/Cas9-based gene therapies [33]. A study aimed at comparing the serum and vitreous anti-Cas9 antibody levels in wild-type (WT) C57BL/6J mice immunized intramuscularly wild-type (WT) C57BL/6J mice with Cas9 intramuscularly, and reported a lower expression of antibodies in vitreous fluid compared to the serum, implying that preexisting immunity to Cas9 may present a lower risk in human eyes than systemically administered ones [26]. Angiogenesis is the process of the formation of new blood vessels from existing ones and is a hallmark of certain stages of eye diseases, such as wet age-related macular degeneration (AMD) [35]. A study aimed to use genome editing to treat angiogenesis-assoc
