Introduction

Inherited retinal diseases (IRDs) are a diverse group of rare genetic disorders associated with more than 280 different genes [1]. IRDs manifest varying degrees of clinical severity and variable inheritance patterns [2], leading to blindness in infancy/early childhood [3, 4] or a gradual and progressive vision loss during adulthood [5,6,7,8]. The development of comprehensive and effective treatment proves to be a challenge for scientists, particularly due to the diverse number of genes involved in IRDs. In 2017, Luxturna® (voretigene neparvovec), a gene therapy drug developed by Spark Therapeutics Inc., was approved by the Food and Drug Administration (FDA). Luxturna® uses adeno-associated virus serotype 2 (AAV2) as a delivery vehicle to carry the wild-type Retinal Pigment Epithelium 65 (RPE65) gene into the retinal cells with RPE65 mutation for treating patients with Leber congenital amaurosis (LCA), a rare form of inherited vision loss [9]. AAV-derived vectors have several advantages, including high biosafety, low immunogenicity, stable expression, and high infectivity in several cell types. Although AAV-derived vectors are the safest and most effective viral vectors for gene replacement therapy in the retina, they cannot accommodate genes larger than 4.7 kb, and the generation of neutralizing antibodies against AAV may attenuate the efficacy of AAV-mediated gene therapy [10, 11]. Moreover, the treatment with Luxturna® requires vitrectomy of the retina, followed by the retinal detachment using the air tamponade [12, 13]. This multi-step surgical procedure is a huge burden to patients’ fragile retinas. Another concern is the repeated treatments with Luxturna®, as a single Luxturna® dosage only lasts for five years [14] and patients are required to undergo these invasive procedures routinely. Therefore, innovative topical delivery and highly permeable gene therapy are urgently needed for the therapy of IRDs [15].

The development of the clustered regularly-interspaced short palindromic repeats (CRISPR)-CRISPR-associated protein 9 (Cas9) gene editing technique revolutionized molecular biology and showed great potential for improved gene therapy [16, 17]. The first clinical trial of CRISPR-Cas9 gene editing treatment for IRDs, delivered by the AAV, was launched for the most common cause of inherited childhood blindness, LCA type 10 (LCA10) [18]. Homology-independent targeted integration (HITI) is an advanced CRISPR-Cas9 technique that enables targeted gene insertion in non-dividing cells and presents a new approach to treating genetic disorders [19]. However, one of the greatest challenges in the therapeutic application of CRISPR-Cas9 for retinal diseases is the delivery efficiency of the CRISPR-Cas9 components into the retinal pigment epithelium (RPE) and the neurosensory retinal environment in the posterior pole of human eyeballs [19]. The gene delivery using viral vectors is efficient but associated with several disadvantages, including random insertion, mutagenesis, and biohazard concerns [20]. Recent developments in nanomedicine overcome this difficulty by introducing a nontoxic delivery of CRISPR-Cas9 components that can significantly alleviate safety concerns raised by viral vectors [20]. Today, researchers aim to engineer nanoparticles (NPs) with specialized properties to go beyond viral limitations and create new opportunities towards the application of CRISPR-Cas9 in treating IRDs. However, the translation of such technology to the clinic is hampered by several obstacles. The complexity of the retinal structure poses a significant challenge to the standard measurement of visual performance after treatment and causes unreliable diagnosis [21, 22]. Moreover, in vitro, in vivo, and species variations among disease models limit the ability to fully recapitulate the structure and functions of the retina [23,24,25]. With induced pluripotent stem cells (iPSC) technology's burgeoning field, researchers are now generating three-dimensional (3D) retinal organoids (ROs) from human iPSCs [

The complex three-dimensional (3D) architecture of the retina

During embryonic development, the retina is derived from the prosencephalon, the anterior portion of the brain [21]. The retina’s unique architecture can be classified into two distinct parts: the posterior RPE layer with the most apparent light absorption function and the anterior multilayered neuroretina (Fig. 1). This multi-layered structure of the retina comprises two synaptic layers (the outer and inner plexiform layers) and three specialized neuronal cell layers (Fig. 1). The first neuronal layer comprises rod and cone photoreceptor (PR) cells that convert the absorbed light with different intensities into electrical signals via phototransduction. PR cells then form synapses with the second layers of neuronal cells, horizontal and bipolar cells present in the inner nuclear layer (INL), through the outer plexiform layer (OPL). This eventually leads to the transmission of the signal from PRs to the third class of neuronal layer cells, the retinal ganglion cells (RGCs), through the inner plexiform layer (IPL) (Fig. 1). Finally, the axons of RGCs converge to form the optic nerve, which in turn leads to the transmission of the visual impulses from the eye to the brain [21, 22]. Any impairment in this signaling cascade can result in visual disorder. In addition, the photoreceptor dysfunction or loss can be associated with age, diabetes, and genetics [32,33,34]. The latter causes a specific category of disease described as IRDs, which is the focus of this review.

Fig. 1
figure 1

The complex architecture of the retina. The general layout of the retinal layers is shown on the left and cell types on the right. Photoreceptor cells, bipolar cells, and retinal ganglion cells constitute the signal transmission pathway that conveys vision signal to the brain. Horizontal and amacrine cells are interneurons modulating visual signal transmission. Müller glial cells perform the neuronal support functions similar to those of astrocytes in the brain

Inherited retinal diseases and the treatment obstacles

IRDs associated with several clinically and genetically heterogeneous defects belong to a group of progressive retinal degeneration diseases that may lead to vision loss. They exhibit a wide variation in genetic mutations, age of onset, and disease severity [35,36,37], with an estimated incidence of 1 in 2000 to 3000 individuals [38,39,40,41,42]. Due to recent advances toward pathogenesis and characterization of genes responsible for IRDs, with more than 270 genes identified so far, a significant progress has been made in the field of incurable IRDs [43,44,45,46,47,48]. It has led to the development of treatments aimed at restoring vision or delaying the vision loss progression; however thus far, treatment options are still limited. IRD gene variants can be transmitted in an autosomal dominant (e.g. LCA and retinitis pigmentosa (RP)) [49, 50], recessive (e.g. cone-rod dystrophy) [51], and X-linked manner (e.g. X-linked juvenile retinoschisis (XLRS)) [52,53,54,55]. As shown in Table 1, IRDs can be clinically classified into six categories based on the affected retinal regions/cell types and four categories depending on the genetic inheritance mode; other classifications are based on the monogenic and multifactorial nature of IRD and the disease progression [56]. The genetic heterogeneity observed in these diseases manifests in patients with very similar clinical phenotypes but different genetic diagnoses, demanding gene-specific therapies or gene editing treatments to develop treatment strategies [8, 57]. Nevertheless, there is no available cure for IRDs currently, and ophthalmology has been at the forefront of utilizing gene therapy to treat these disorders.

Table 1 Classification of IRDs

Overview of gene therapy techniques for IRDs

Gene therapy application for ophthalmic diseases is a blooming field of research and currently overcoming the barriers for translation to the clinic, which is extensively described in multiple reviews [58,59,60,61,62,63]. AAV vector-based gene therapy has obtained the marketing approval for treating RPE65-associated LCA [18, 64,65,66], Leber hereditary optic neuropathy (LHON) [67, 68], and choroideremia (CHM) [69,70,71]. Treatments for LHON and CHM have entered phase III clinical trials and raised hopes that these approaches might be of practical use to delay or halt disease progression in patients with these IRDs [72,73,74]. Currently, more than 30 gene therapy trials for IRDs are being conducted in the United States, and some have entered the phase III clinical trials. Positive outcomes in these clinical trials were mainly due to the advantages offered by the retina as the target organ and AAV as the carrier [135]. Also, it costs 425,000 US dollars per eye to receive Luxturna® treatment, imposing a heavy economic burden on the society or individuals [1]. Another concern in treating IRDs is the structural peculiarity of the retina, which demands a specific administration route depending on the choice of drug vehicles. With the advent of nanoparticle (NP)-mediated delivery, the unmet needs of efficient genome editing associated with viral vectors are expected to be greatly fulfilled. In therapeutics, the application of NPs as delivery carriers for genes and drugs has been profoundly investigated [136,137,138,139,140,141]. The classification of NPs based on different sizes and structures is shown in Fig. 3. Their nanoscale size enables them to interact with biological systems at the molecular level. In addition, numerous reports have documented that NPs can ensure successful targeted delivery and be transported across biological barriers that can make them an indispensable tool for the drug delivery [142,143,144,145]. The NP-mediated delivery of high molecular weight CRISPR-Cas9 complexes is one of the most significant approaches being developed for genome editing and other evolving applications [146,147,148,149]. Here, we review the promising gene delivery carriers with the potential for IRD treatment based on properties like the nanocarriers and load capacity.

Fig. 3
figure 3

Classifications of NPs. NPs are classified into organic, inorganic, and other NPs. The average sizes of the particles are shown relative to each other and the structural features are shown as discussed in the article

To effectively deliver a therapeutic agent to the retina, the particle size and charge are important parameters when develo** nanocarriers [150]. The inner limiting membrane (ILM) with a negative charge located between the vitreous and the retina is a physical and electrostatic barrier [151], which limits the diffusion of NPs [151, 152]. The diffusion of NPs or drugs into the retina varies due to the architecture differences among species, including mice, bovines, and humans [23,24,25]. For example, the pore size of human ILM is about 10 nm, with the variable thickness ranging from 100 nm in the fovea to 4 μm in the thickest area [23]. However, the thickness of the ILM in small animals (such as mice and rats) is less than 100 nm, so the pharmacokinetic distribution of drugs observed in animals often cannot be applicable in clinics [24]. In addition to ILM, the vitreous and retinal cell membranes are all negatively charged. Therefore, after the injection of NPs into the vitreous, the NP’s charge generates an electrostatic interaction which affects the diffusion rate of NPs within the vitreous. Furthermore, the charge of NPs also affects the permeability across the ILM. Therefore, excessive positive and negative charges are not conductive for the drug delivery to the retina [153]. Huang et al. compared lipid NPs with different charges (− 30mv ~  + 50mv) and found that + 35mv lipid NPs can achieve the highest distribution efficiency in the retina [154]. Here, we mainly focus on recent applications of NPs for delivering CRISPR-Cas9 to the retina while an extensive discussion of NP synthesis and design is beyond the scope of this article and has already been reviewed by others [155,156,157,158,159,160,161,162,209]. PEI-modified NPs bind with and condense DNA to form spherical structures, which were fused with the endosome and elicited the “proton sponge effect” to escape endosomal degradation [210]. Several studies have emphasized on the application of this polymer in gene therapy [211,212,213]. In 2015, PEI was utilized for delivering Cas9 and gRNA for the first time, leading to the knockout of the Ptch1 gene in the cerebellum of newborn mice [214]. Liang et al. encapsulated the CRISPR-Cas9 plasmid into an aptamer-functionalized PEG-PEI-Cholesterol nanocarrier to target and reduce vascular endothelial growth factor A (VEGF) gene expression both in vivo and in vitro [168]. However, the moderate toxicity of PEG-PEI-Cholesterol in subsequent clinical trials limited its application for drug delivery [215].

Liao et al. reported that the intravitreal injection of PEI/DNA polyplexes could deliver plasmids into retinal ganglion cells in mice. However, the efficiency of PEI-mediated gene editing was low [216]. β-cyclodextrin (β-CD), an FDA-approved drug [217], is a cyclic oligosaccharide with a diameter of 200 nm and high transfection efficiency for small plasmids [171, 218]. To increase the efficiency of gene transfection by PEI, this polymer was covalently linked with β-CD [219, 220] and showed no reported cytotoxicity in HEK293 cells [221]. This strategy increased the transfection efficiency of the luciferase gene to nearly four folds, compared with PEI alone [171]. Given that β-cyclodextrin-PEI (PEI/β-CD) can condense large plasmids at a high nitrogen-to-phosphorus ratio (N/P ratio; the ratio of positively-chargeable polymer amine (N) groups to negatively-charged nucleic acid phosphate (P) groups), Zhang et al. evaluated the efficiency of PEI/β-CD-mediated delivery of CRISPR-Cas9 system in HeLa cells in which the gene transfection efficiency was about 34%. Meanwhile, this delivery resulted in effient editing at hemoglobin subunit beat loci and rhomboid 5 homolog 1 loci of 19.1% and 7%, respectively [171]. Concerning ocular gene delivery, PEI is one of the most widely investigated polymers. Several studies have explored its potential as an alternative delivery vehicle for the eye [25, 222,223,224,225]. For example, one study showed that the intravitreally injected PEI/DNA could be successfully delivered to mouse RGCs [216]. Conceivably, these findings suggest PEI/β-CD as a potential, efficient, and safe nanocarrier for delivering the CRISPR-Cas9 gene editing system into the retina. However, more investigations and efforts are still needed to elucidate the in vivo utility of PEI/β-CD to deliver CRISPR-Cas9 machinery into the retina.

Inorganic nanoparticles

Nanodiamonds (NDs)

Nanodiamonds (NDs) are a novel class of nanomaterials that have garnered a great deal of attention for their clinical application potential due to their low cost, fluorescent capability, low cytotoxicity, and superior biocompatibility [226,227,228,229,230,231]. It has been reported that most of the NPs destroy the endosomal membrane and release the cargo via the proton sponge effect or chemical reaction [232]. Following the entry of NDs into the cells through endocytosis, the sharp structure of these nanocarriers destroys the endosomal membrane resulting in a quick escape. This unique mechanism of endosomal escape makes NDs more biologically safe and stable [233]. Notably, 2–10 nm NDs provided long-term stability without causing cell death and oxidative stress [231, 234]. In addition, DNA, protein, and drugs can be delivered through different surface modifications of NDs (e.g. carboxylation, hydroxylation, hydrogenation, amination, and halogenation) that improve their intracellular uptake and ability to target specific cells [235,236,237,238]. Despite the progress in ND technology, only one study demonstrated the potential therapeutic application of NDs in treating retinal diseases. In our laboratory, we utilized mCherry protein as a critical linker between 3 nm NDs and DNA in which the amide (–NH) and histidine groups (–His) on mCherry protein bind to the carboxylic groups (–COOH) and phosphate groups (–PO43−) on the ND and DNA molecules, respectively [239]. These chemical reactions formed a stable link between NDs and the components of the CRISPR-Cas9 genome editing system. The final size of this nanocarrier is about 5 nm, which facilitates its penetration and the delivery of CRISPR-Cas9 components to all layers of the retina, including photoreceptor and retinal pigmentation epithelium layers. The resulting NDs effectively promoted the delivery of the CRISPR-Cas9 components to initiate HDR to direct the c.625C > T mutation of RS1 gene in human iPSCs and mouse retina, generating an X-linked retinoschisis-like disease model characterized by severe perturbations of the retinal structure (Fig. 4) [239]. The regulatory approval of new drugs by FDA requires inorganic NPs to be cleared via the kidneys to minimize systemic toxicity and improve drug efficacy [240,241,242]. Since NDs are not biodegradable, using them as the delivery vehicle in therapeutics demands an articulate engineering of NPs to meet FDA standards [231, 243,244,289, 290]. However, the immune issues related to these nanocarriers are a matter of concern, therefore, their clinical applications should be further investigated [291]. To the best of our knowledge, no study explored the potential of DNA nanoclews for drug delivery into the eye and only one study reported the utility of DNA nanoclews for delivering CRISPR-Cas9 as briefly described below [292]. Since this type of NPs is made of ssDNA, a sequence complementary to gRNA can be designed to match the base-pair of the guide portion of Cas9-sgRNA [293]. Sun et al. coated the DNA nanoclews/Cas9 protein/gRNA mixture with PEI to improve its cellular uptake and the endosomal escape. The resulting DNA nanoclews were able to deliver the CRISPR-Cas9 components into the target cells in vitro and in vivo with the editing efficiencies of 36% and 25%, respectively [292].

Nanoscale zeolitic imidazole frameworks (ZIFs)

Nanoscale zeolitic imidazole frameworks (ZIFs), a subclass of metal–organic frameworks (MOFs), are composed of divalent metal cations and imidazolate bridging ligands with pH buffering capacity. These features enable ZIFs to facilitate endosomal escape [294]. ZIFs combine the advantages of the 3D network and porous structure of zeolite with traditional metal–organic clusters [295, 296] that have recently attracted more attention due to their great potential for delivering drugs, genes, and proteins [297,298,299]. Alsaiari et al. reported for the first time that ZIF-8 could encapsulate Cas9 protein and gRNA and subsequently undergo genome editing in Chinese hamster ovary (CHO) cells, with loading and editing efficiencies of 17% and 37%, respectively [294]. To the best of our knowledge, no report on the delivery of CRISPR-Cas9 system to the retina using such nanocarrier has been published.

A very recent study utilized a water-in-oil emulsion approach to fabricate a pH-responsive silica–metal–organic framework hybrid NP (SMOF NP) consisting of silica and ZIF. Subretinal injection of SMOF NPs induced efficient genome editing in mouse retinal pigment epithelium. Furthermore, both in vitro loading and delivery efficiencies of CRISPR-Cas9 components by SMOF NPs were high but varied depending on different cell lines [300]. These data introduced a promising nanoplatform that may improve gene therapy in the treatment of IRDs.

To summarize, most modern NPs use non-covalent bonding to carry plasmids expressing Cas9 and gRNA or Cas9 protein and gRNA expression plasmids. These NPs are about 100–200 nm in size with a slight positive charge, which may be suitable for intravitreal injection into the eyes of patients with IRDs. We have also described ND carriers as the only carriers smaller than 10 nm. The size may allow the NPs to pass through the human ILM barrier and effectively perform HDR-based CRISPR gene editing in all retinal layers. Also, SMNPs can adapt to various in vivo environments with their unique molecular recognition ability and therefore have the potential to be an alternative approach for the delivery of CRISPR system to the vitreous. Yet, the means of safe and efficient delivery remain to be fully investigated. The properties and advantages/disatvatages of different types of NPs are summarized in Table 2.

Table 2 Summary and comparison of properties of NPs mentioned in this review

The retinal organoids and precision medicine

The heterogeneity of IRDs hampers the development of an effective strategy to tackle a wide range of disorders. One of the major hurdles that hinders the translation of basic retinal research into clinical applications is mainly due to the poor relevance in existing preclinical models. For example, in the mouse model, more than 90% of photoreceptors are rod cells, whereas, in humans, the visual acuity is mostly dependent on cone photoreceptors [301]. Notably, many in vitro and in vivo findings could not be reproduced in humans. Drugs that proved to be safe and effective in animal studies failed to exert the same efficacy in clinical trials. In addition, the information obtained by studying two-dimensional cultures does not recapitulate the heterogeneous complexity and critical features of the microenvironment of cells in vivo [136,137,138]. This gap causes a noticeable lack of fidelity between the aforementioned experimental models and human outcomes. The advent of three-dimensional (3D) multicellular constructs, referred to as the organoid technology, offers a promising complementary model to pursue clinical translation and precision medicine applications. Human pluripotent stem cells (hPSC)-derived retinal organoids hold an excellent value for modeling the human retina features [302,303,304,305,306]. In particular, by using patient-derived cells combined with reprogramming strategies, this technique could represent an efficient pre-clinical approach toward the personalized therapeutic strategies adaptable to a broad number of IRDs and provide the link to disease-specific human drug screening models [307]. The iPSC-derived organoids strategy can provide a means for assessing the efficiency and efficacy of NP-mediated delivery of CRISPR-Cas9 system tailored to each patient's genetic makeup. In this part, we mainly focus on organoid as a technology platform for precision medicine in IRDs, especially with potential translational applications for evaluating new therapeutic drugs.

Organoids as a drug testing platform for translational research

Introducing a new pharmaceutical drug to the market is a complicated and costly process, especially when the in vivo testing result of drug candidates fail to reach the requirements initially fulfilled by the in vitro test. The gap between in vitro validation and clinical application is significant, mainly because the simplicity of the in vitro model cannot mimic the complex nature and heterogeneous characteristics of clinical patients [308]. These critical problems impose significant limitations on the translation of candidate drugs to the clinic and require advanced strategies to improve this shortage. Organoid systems show considerable reliability of recapitulating features and functions of the human system offering a great potential for testing drug efficiency in target organs. Patient-derived organoids (PDOs) can be generated particularly by reprogramming patient-derived cells to induced pluripotent stem cells (iPSCs), followed by the differentiation into the desired cell lineage and organoids [309]. Notably, several reports showed that in vitro PDOs could highly match and reproduce patients’ response to candidate drugs in most cases, highlighting the merit of this system in personalized medicine as a predictor of therapeutic outcome [310,311,312,313]. The organ-like structure technology offers a more efficient screening of candidate drugs prior to in vivo testing, which helps to reduce drug development costs. Organoids have been a powerful tool for functional drug testing, personalized therapy and disease modeling [314,315,316,317,318,319,320]. For iPSC-derived organoids, various organoid models have been generated using human iPSCs, including heart [321, 322], kidney [323], brain [324, 325], intestine [312, 313, 326], liver [327,328,329], lung [330] and retina [302,303,304,305]. Furthermore, organoids have been successfully applied to model human genetic diseases. For example, intestinal organoids derived from cystic fibrosis (CF) patients proved to be a reliable tool for effective drug treatment [314]. Of note, brain [315,316,317] and kidney [318,319,320] organoids generated from patient-derived iPSCs have also been established to model diseases. Xu et al. subjected brain organoids to Zika virus infection and used them as the platform for drug repurposing [331]. Bian et al. demonstrated that the neoplastic cerebral organoids are suitable for targeted drug testing [332]. Saengwimol et al. used retinoblastoma organoids for the evaluation of cellular response to chemotherapy drugs. [333]. However, the consistency and reducibility of this system at a scale consistent with clinically associated cell numbers is still a matter of concern [334]. For retinal organoids, Vergara et al. developed an iPSC-derived retinal organoid-based screening platform that allows the accurate quantification of fluorescent reporters [335]. Despite the progress of some organoid-based researches, the organoid technologies remains immature and not ready for the demands of high-throughput screening in drug screening [336]. It was attributed to the developmental variability and diversity of retinal organoids that may hinder the utility of retinal organoids in the evaluation of therapeutic effects and comparative analysis [337]. Nevertheless, considering that retinal organoids hold promising potential in new drug development, it would be still expected and encouraging to use retinal organoid technologies to augment the existing drug development pipelines. Collectively, these findings highlight the utility of organoids as a part of the drug testing pipeline, creating the opportunities for more effective therapies, especially for patients with rare genetic diseases in a cost-effective manner.

Retinal organoid applications in precision medicine

To establish a drug testing platform for IRD treatment based on organoid technologies, generating a representative disease model is a fundamental step. As highlighted below, several reports demonstrated the utility of organoids in eye disease modeling. Ohlemacher et al. developed a retinal organoid model using patient-specific iPSC-derived RGCs to study an inherited form of glaucoma [26]. Tucker et al. generated multi-layer optic cup-like structures representing photoreceptor precursor cells for investigating the pathogenesis of RP [27]. A separate study focused on different frameshift mutations in the RPGR gene, one of the most prevalent causes of autosomal recessive RP, and generated patient-specific retinal organoids with defects in morphology and functionality of photoreceptors accompanied with decreased cilia length as a disease model [28]. The constructed vectors for the CRISPR-Cas9 machinery were delivered into the patient-derived iPSCs via electroporation, and the mutation-corrected iPSCs were then differentiated into retinal organoids. Notably, the reversal of morphological and functional defects in retinal organoids with RPGR mutation was observed after the CRISPR-Cas9-mediated gene correction [28]. Using a similar approach, Buskin et al. demonstrated the severe RP defects observed in patient-specific retinal organoids harboring the CRISPR-Cas9-induced PRPF31 mutation [29]. This further proved the effectiveness of this combined strategy toward personalized and targeted gene therapy. Another example for coupling retinal organoids with the genome engineering technique is the application of CRISPR-Cas9 technology to correct RS1 mutation in retinal organoids derived from XLRS-patients [30]. Huang et al. successfully established the XLRS patient-derived retinal organoids that recapitulate the retinal splitting feature of the disease. Meanwhile, they delivered the CRISPR-Cas9 system using electroporation to correct the mutation and showed that CRISPR-Cas9-mediated correction of the disease-associated C625T mutation efficiently rescued the disease phenotype (Fig. 6) [30]. Parfitt et al. used LCA patient-derived iPSCs to generate 3D optic cups with the mutation of a cilia-related gene, CEP290 [31]. Introducing an antisense oligonucleotide to patient iPSC-derived organoids could effectively prevent the aberrant splicing, restore the expression of full-length CEP290 protein, and repaired the cilia defects [31]. Currently, CEP290 treatment is in phase III clinical trial (NCT03913143), paralleling the classic augmentation RPE65 trials initiated in 2007 [338, 339]. Overall, these studies demonstrated that 3D retinal organoids derived from patients with various retinal diseases are able to recapitulate the complex retinal architecture, rendering them an ideal platform for examining the safety and specificity of CRISPR-Cas9 system for the therapeutics applications.

Fig. 6
figure 6

Reproduced from our previous study [30]

Patient-derived retinal organoids recapitulate disease-specific features. A Bright-field images and B H&E staining of control and XLRS patient-derived retinal organoids exhibit schisis feature at day 150 of differentiation. C Quantification of splitting area in control and XLRS-patient-derived retinal organoids. D Bright-field images of control and XLRS patient-derived retinal organoids at days 90, 100, and 110 of differentiation. E A schematic presentation of the time course for the generation of control and XLRS patient-derived retinal organoids. Disease-specific features can be observed after applying defined differentiation stimuli and time course.

Conclusions and perspectives

The combination of NPs, CRISPR-Cas9, and retinal organoids as a promising therapeutic platform

IRDs have long been viewed as a class of disorders with no effective treatment. This maxim is now being reversed by tremendous efforts in nanomedicine and gene engineering, which results in promising clinical trials for blinding diseases. As therapeutic strategies for IRDs expand, the importance of molecular diagnosis is gaining momentum. Most IRD gene supplementation therapies are in phase I/II clinical trials, with LCA therapy approved by the FDA to treat patients carrying biallelic RPE65 mutations [59, 62, 340,341,342]. Although these efforts are still evolving, the importance of gene therapy for elevating the life quality of IRD patients has never been more apparent. In this review, we introduced a forward movement of therapy by combining the advances in CRISPR-mediated gene editing, NP-based delivery, and iPSC-derived retinal organoids technologies, to assess the potential safety and efficacy of designed CRISPR-Cas9 components and nanocarriers in a clinically relevant in vitro model. As mentioned earlier, NP-mediated delivery of high molecular weight CRISPR-Cas9 complexes combined with the advanced CRISPR-Cas9 technologies, applicable in non-dividing retinal cells (e.g. HITI), introduces a simple yet efficient approach for precise gene therapy in IRDs. In addition, patient-derived retinal organoids can mimic typical disease features, providing a reliable platform for disease modeling. Collectively, this integrated strategy is expected to facilitate the evaluation of the gene editing in preclinical tests and be a major driver towards advancing IRD’s personalized medicine (Fig. 7).

Fig. 7
figure 7

An integrative, multidisciplinary approach for future gene therapy in IRDs. IRD patient’s blood sample can be reprogrammed into induced pluripotent stem cells (iPSCs) followed by the differentiation into retinal organoids. These patient-derived retinal organoids exhibit disease-specific features and can be applied as a reliable platform for assessing disease progression and treatment outcome (e.g. XLRS-patient-derived retinal organoids exhibit severe retinoschisis-like features). Researchers can utilize optimized NPs loaded with plasmid DNA encoding CRISPR-Cas9 machinery to achieve efficient gene delivery and precise gene editing. This results in the rescue of the disease phenotypes associated with the specific IRDs (e.g. the splitting phenotype in XLRS-patient-derived retinal organoids can be rescued as shown above). Integrating patient-derived retinal organoids, CRISPR-Cas9 technologies, and NPs promotes precision gene therapy applications for IRDs

During the past few years, our preliminary research on combining CRISPR-Cas9 gene editing, patient-derived retinal organoids, and NPs has been conducted to investigate the potential of this integrated strategy in IRD therapeutics. For example, Yang et al. utilized ND-mediated delivery of CRISPR-Cas9 to introduce the mutated RS1 gene into human iPSCs and mouse retina, leading to the generation of XLRS-like disease model [239], however, the potential of this strategy for gene therapy is yet to be explored. In another study, Chou et al. successfully combined SMNP nanocarrier and CRISPR-Cas9-mediated HITI strategy to knock-in the RS1 gene in the mouse retina [119]. In the future, this strategy can be further applied in patient-derived retinal organoids model to assess its efficiency and efficacy in clinically relevant disease setting. In a separate study, Huang et al. successfully differentiated iPSCs from XLRS patients into retinal organoids presenting disease features. They further coupled this disease model with CRISPR-Cas9 technology and repaired the RS1 gene mutation with 50% efficiency [30]. However, the electroporation method used for the transfection of CRISPR-Cas9 machinery caused a massive cell death [30]. Therefore, future attempts to utilize NPs as an alternative method will be of great interest to resolve this drawback and further boost the therapeutic effects. We hope that this combined strategy will become a treatment modality for other IRDs and elevate the life quality of patients.

Technology hurdles

Although the rationale for this integrated approach is clear, several technology hurdles remain to be addressed. The development of advanced CRISPR-Cas9 systems with high specificity has armed researchers worldwide with a powerful tool to study human diseases. However, utilizing this technique for translational medicine research has inevitable concerns. Since patient-derived iPSCs not only carry the specific mutation intended to be repaired but also harbor the entire human genome, this makes CRISPR-Cas9-mediated gene editing more susceptible to undesired off-target effects. For example, gRNA may recognize sequences similar to the target loci and cause permanent sequence alterations resulting in abnormal gene function. Furthermore, no in vivo study examined how long the nuclease remains active before its degradation and what the possible adverse effect(s) might be. Anti-CRISPR proteins can be used to limit the off-target editing, however, the best time to shut off Cas activity requires optimizations [343]. Another concern is that CRISPR-Cas9-mediated knockout or overexpression of the gene of interest may be compensated by neighboring cells, which would interfere with the expected outcome [344]. As for the organoid technology platform, although the 3D retina organoid has equipped ophthalmology with a unique and relatively accurate representation of the human eye, it still lacks a high level of morphological and functional complexity demonstrated by the mammalian retina in vivo. For example, the 3D retinal organoid with retinal pigment epithelium (RPE) layer provides a more physiologically relevant disease model for photoreceptor-associated diseases. However, both simple and complex organoid models have their pros and cons; thus, the appropriate level of complexity should be designed according to the purpose of the study [345]. A more challenging issue is the disease modeling of late-onset retinal diseases by manipulating culture medium to induce ageing factors. It demands a profound knowledge of factors involved in each specific developmental stage which are yet to be investigated. More in-depth knowledge and assessment of the culture medium composition and distribution are required for modeling complex IRDs. Nevertheless, retinal organoid technologies have only been utilized for monogenic IRDs so far. Lastly, although NP-based delivery has been proposed and proven to be a promising drug delivery vehicle, improving the therapeutic application of NP-based gene therapy remains an important concern. It requires in-depth investigation on the cytotoxicity of the NPs under variable conditions and on the key factors determining the release rate of drugs from NPs to the retina. The successful drug delivery to the neuroretina, and even more specifically to photoreceptors, highly depends on the choice of the NPs most suitable for both the drug and the target tissue. Besides, different gene mutations may affect the complex retina structure and cause anatomical obstacles for nanomedicine drug delivery. Moreover, the choice of delivery route, immunoreactivity, and nucleic acid-based drug stability are critical factors that need to be addressed for a successful clinical application. Nevertheless, in this ever-evolving field, it is crucial to move scientific discoveries into clinics and new therapies for vision restoration in more patients than ever before. Ultimately, the application of such technologies in the clinic and industry should fulfill four criteria: reproducibility, standardization, validation, and quality assurance. Although NP-mediated delivery of the CRISPR-Cas9 system shows a great promise in repairing the IRDs, the time window for treatment is a critical determinant for the therapeutic outcome. For example, in XLRS, the differentiated retinal cells with the splitting phenotype are not responsive to gene therapy, indicating the delayed treatment by the time the disease is already progressed. However, more preclinical data will be required to prove this concept. Taken together, the advances and current progress of basic research hold the promises that laboratory findings can be translated into clinical applications in the near future and bring hope to patients who have blindness and other hereditary diseases. The integration of nanotechnology, CRISPR, and stem cell technologies present a novel platform and is expected to accelerate bridging the basic research and translational medicine, and further promote medical precision therapies.