Cell-cell interactions between transplanted retinal organoid cells and recipient tissues

Abstract

The transplantation of human organoid-derived retinal cells is being studied as a potentially viable strategy to treat vision loss due to retinal degeneration. Experiments in animal models have demonstrated the feasibility of organoid-derived photoreceptor transplantation in various recipient contexts. In some cases, vision repair has been shown. However, the recipient-donor cell-cell interactions are incompletely understood. This review briefly summarizes these interactions, categorizing them as synaptic structure formation, cellular components transfer, glial activation, immune cell infiltration, and cellular migration. Each of these interactions may affect the survival and functionality of the donor cells and, ultimately, their efficacy as a treatment substrate. Additionally, recipient characteristics, such as the cytoarchitecture of the retina and the immune status, may also impact the type and frequency of cell-cell interactions. Despite the procedural challenges associated with culturing human retinal organoids and the technical difficulties in transplanting donor cells into the delicate recipient retina, transplantation of retinal organoid-derived cells is a promising tool for degenerative retinal disease treatment.

Introduction

The treatment of blindness due to irreversible retinal degeneration remains among the most significant challenges in modern ophthalmic care [1]. Since the early 21^st century, when researchers established that human embryonic stem cells (hESCs) could differentiate into certain retinal cell types, interest surrounding the viability of stem cell transplantation in the treatment of retinal disease has grown exponentially. The subsequent development of mammalian organoids capable of mimicking retinal morphology and developmental biology, accompanied by their successful transplantation into animal models, underscored the feasibility of retinal stem cell therapeutics [2–4]. In animal models of retinal degeneration, multielectrode array recordings in the superior colliculi, among other tools have shown improved retinal signaling patterns and light sensitivity when compared to untreated animals with retinal degeneration. However, they are yet to demonstrate improved visual function that fully resembles non-degenerate, wild type models [13, 24, 50, 51].

In 2023, Hirami and colleagues published an account of the first documented transplantation of cultured human retinal organoid tissue into the retinas of human subjects. Their study has demonstrated the feasibility and short-term safety of the procedure and the survival of the retinal organoid graft tissue in two human recipients with end-stage retinitis pigmentosa [5**]. After decades of preclinical studies in animal models, the experimental transplantation of retinal organoid-derived photoreceptor cells into humans has become a reality. However, researchers’ understanding of the cellular interactions between the transplanted retinal organoid donor cells and the recipient retinal tissues remains incomplete. The limited range of robust assays that detail these diverse interactions in vivo contributes to the lack of complete knowledge in this area.

This review aims to summarize the current evidence concerning the cellular interactions that potentially or actually occur between transplanted organoid cells and recipient retinal neurons. For discussion herein, we have categorized these interactions as follows: 1) synaptic structure formation, 2) cellular component transfer (CCT), 3) glial activation, 4) immune cell infiltration, and 5) cellular migration. Given the translational potential for clinical treatment using human pluripotent cells, this review will briefly highlight human retinal organoid photoreceptor transplantation studies in animal models and human subjects and deemphasize studies involving non-human organoids, organoid-derived non-photoreceptor cells, and purely in-vitro experimentation.

The retina and pathogenic targets for retinal organoid transplantation

Before discussing the disease targets, a brief discussion of the anatomy and cellular composition of the retina will be useful. The mammalian retina is a multi-layered neural structure (Figure 1) that coats the back of the eye. It is bounded by the inner limiting membrane (I.L.M.) on its internal surface and a layer of photoreceptor cells on the outer surface (Figure 1) [6, 7]. The light-sensitive biochemical properties of the photoreceptor cells enable them to capture photons from the ambient environment and generate neural signals that are transmitted to structures in the brain and interpreted as sight [8]. The subretinal space is external to the photoreceptor cell layer and is the target injection site for many organoid transplantation studies [9, 10]. The outer boundary of the subretinal space is the apical surface of the retinal pigmented epithelium (RPE). A pentalaminar Bruch’s membrane borders the RPE on its basal surface. Blood-derived components from the underlying choroid diffuse through Bruch’s membrane to nourish the RPE [11].

Figure 1. Diagram of the layers and cellular components of the human retina.

Schematic diagram of human retinal organization, highlighting the location of the photoreceptor cells (i.e., the rods and cones) and their physiological structural interactions with retinal pigment epithelium (RPE) cells and inner retinal neurons (i.e., horizontal and bipolar cells). Created with BioRender.com.

The main disease targets for photoreceptor transplantation therapy using organoid transplantation are age-related macular degeneration (AMD) and inherited retinal dystrophies (IRDs). AMD and its subtypes are characterized by pathologies in the photoreceptor cell layer, RPE, Bruch’s membrane, and choriocapillaris. Although the molecular etiology of AMD is still an active area of research, both the wet and dry subtypes produce defects in the RPE integrity that can lead to degeneration of regions of the photoreceptor cell layer, most notably the macula. Dry AMD is a common cause of irreversible blindness, and its later stages compromise photoreceptor function, making it a promising target in organoid transplantation studies [12]. Similarly, rare IRDs characterized by photoreceptor degeneration and progressive blindness, such as retinitis pigmentosa, have also been targeted by organoid-based treatment approaches [13]. RPE transplantation is a vibrant field of scientific investigation that has corollaries to retinal organoid photoreceptor transplantation. However, research in that area largely lies outside the scope of this review. For a comprehensive review evaluating stem-cell-based RPE transplantation, please refer to an insightful review by the Bharti group in 2023 [14].

Human retinal organoids as a therapeutic substrate

Retinal organoids are cultured self-organizing three-dimensional structures derived primarily from hESCs or human induced pluripotent stem cells (hiPSCs). To date, very few studies have evaluated the utility of human mesenchymal stem cells (M.S.C.s) in the formation of retinal organoids [16]. Retinal organoids recapitulate optic cup formation and contain populations of nearly all significant retinal cell subtypes [2, 3, 15, 44]. Seminal work in the early 2010s established the reality of retinal organoids in modern research [2]. Retinal hiPSCs can be transplanted in an autologous or allogeneic manner. Autologous hiPSC organoids are derived from recipient tissue and transplanted into the same recipient, limiting the potential for immune rejection upon transplantation. However, culturing self-derived organoids for individual recipients is labor-intensive [17*, 18*], and the donor tissue will harbor the same causal pathogenic genetic variants, necessitating additional genetic modifications to the cells before they can be transplanted. In contrast, allogenic hiPSC organoids may be more vulnerable to immune rejection upon transplantation. Still, they promise to minimize protocol complexity [17*] and can be generated from a donor line unaffected by disease-causing genetic variants.

Protocols for culturing retinal organoids vary across the literature. However, many are variations of the 3D-2D-3D culture method or the SFEBq (3D) method [2, 3, 19]. Research has demonstrated that the exclusive 2D culture of retinal tissue may result in the incomplete functional maturity of retinal tissue, whereas 3D organoid culturing techniques have enabled researchers to grow retinal tissue of the appropriate developmental maturity with greater consistency [20, 21]. Transplantation of organoid-derived cells can be done by dissociating organoid cells into a single-cell suspension and injecting the solution into the subretinal space or by dissecting thin sheets of tissue from the organoid and transplanting them into the eye as a single graft [13, 22, 23*]. Transplantation of retinal organoid sheets has been accomplished by isolating the appropriate section of an organoid, loading it into a cannulated device, inducing a retinal detachment in the recipient, and inserting or placing the donor tissue into the targeted subretinal space [5, 24, 51, 52, 54]. Both retinal organoid sheet transplantation and dissociated organoid-derived photoreceptor injections have demonstrated donor tissue survival, differentiation, and incorporation into the recipient retina while improving visual function [13, 22, 24, 51, 52, 55*]. Retinal sheet transplantation may require larger incisions in the eye and retina, with greater risk of local hemorrhage and scarring, than retinal cell suspensions that can be delivered using smaller-gauge needles that gain access via conventional vitrectomy ports.

Current aims in transplantation

The criteria for successful organoid-derived retinal tissue transplants have evolved alongside organoid technology and developments in experimentation. Current aims include 1) maximizing survival of the donor cells in the recipient retina, 2) minimizing tumorigenesis and other adverse reactions in the recipient, 3) ensuring appropriate differentiation of transplanted retinal tissue, 4) optimizing synaptic integration between the donor tissue and the native retinal cell types, and 5) demonstrating improvements in visual function post-transplantation. Human retinal organoid transplants in murine and rat models have demonstrated that donor cell grafts can survive, differentiate, incorporate, and improve visual function [13, 22, 24]. Although the transplantation of human retinal stem cells into nonhuman primates using both two-dimensional and three-dimensional culture systems has shown survival of donor retinal organoid tissue post-transplantation and a human study has demonstrated the safety of transplantation, few functional vision assays have been performed, limiting conclusions surrounding clinical efficacy [5**, 26, 27]. Ultimately, donor-recipient cellular interactions underpin successful transplantation. What follows is a discussion of donor-recipient cellular interactions (Figure 2 and Table 1) and their impact on current experimental progress.

Figure 2. Cell-cell interactions between donor human retinal organoid cells and recipient retinal tissue.

The diagram on the left depicts the subretinal delivery of stem cells and/or their progeny. On the right, each panel depicts a representation of one of the modes of cell-cell interactions discussed in this review. (A) Synaptic structure formation, (B) Glial activation, (C) Cellular migration, (D) Cellular component transfer, and (F) Immune cell infiltration. Created with BioRender.com.

Table 1.

Summary of selected retinal organoid transplantation studies and the cell-cell interactions between donor and recipient cells.

Study	Donor cell type	Recipient species	Synaptic structuresa	Cellular components transferb	Glial activationc	Inflammatory cell infiltrationd	Donor cell migratione
Lin, 2020	hESC	Rat	Yes	Yes	Yes	Yes	Yes
Gonzalez-Cordero, 2017	hESC	Mouse	Yes	Unclear	Yes	--	--
Garita-Hernandez, 2019	hiPSC	Mouse	Yes	Unclear	--	--	-
Liu, 2023	hESC	Mouse	Yes	--	--	--	Yes
Gasparini, 2022	hiPSC	Mouse	Yes	--	Yes	Yes	--
Zou, 2019	hESC	Rat, mouse	Yes	Yes	Yes	Yes	Yes
Gagliardi, 2018	hiPSC	Rat	Yes	--	--	--	--
Seiler, 2014	hESC	Rat	--	--	Yes	--	Yes
Ribeiro, 2021	hiPSC, hESC	Mouse	Yes	--	Yes	--	--
Zerti, 2021	hESC	Mouse	Yes	--	--	--	--
McLelland, 2018	hESC	Rat	Yes	--	Yes	Yes	Yes
Yamasaki, 2022	hESC	Rat	Yes	Unclear	Yes	--	Yes
Zhu, 2017	hESC	Mouse	Yes	--	--	--	--
Lingam, 2021	hiPSC	Non-human primate	--	Unclear	--	--	--
Lin, 2024	hESC	Rat	Yes	--	Yes	Yes	Yes
Chao, 2017	hESC	Non-human primate	--	Unclear	--	--	Yes
Hirami, 2023	hiPSC	Human	--	--	--	--	--
Zhu, 2018	hiPSC	Mouse	Yes	--	Yes	--	Yes

^a.Immunohistochemistry findings or morphological evidence consistent with one of more components typical of pre- or post-synaptic structures.

^b.Cellular component transfer identified through the colocalization of sex-specific or species-specific markers found in cells of recipient or donor.

^c.Activated Müller glia detected via changes in GFAP expression.

^d.Inflammatory cells detected by positive IBA1 staining.

^e.Off-target cellular migration detected via the presence of MTCO2, HNA, HuNu, Ku80, Human NF, BrdU, and/or another donor cell specific marker in donor cells that migrated beyond the outer nuclear layer into the inner layers of the recipient retina outside of the photoreceptor layer.

Synaptic structure formation

Retinal organoid photoreceptor grafts derived from both hiPSCs and hESCs have demonstrated the capacity to incorporate into recipient retinal tissue, develop structures suggesting synaptic reconnections with recipient cells in the inner nuclear layer, and restore some degree of visual function in selected animal models. Research by Mandai et al. in 2017 was the first to demonstrate substantial evidence of synaptogenesis between transplanted mouse-derived iPSCs and recipient rd1 mouse bipolar cell terminals with a benefit to visual outcomes [50]. Further studies using genetically modified mouse-derived ES/iPSCs sheets and hESC sheets that compromised the development of donor bipolar cells were found to enhance donor cell integration and demonstrated meaningful restoration of light responsiveness in recipient ganglion cells [51, 52]. Immunohistochemistry (IHC) markers for synaptic contacts, human-specific proteins, and animal-model-specific proteins have allowed researchers to determine the location of donor and recipient tissue on samples and evaluate the potential for synaptic activity at those sites [9, 25]. Some studies have taken histological evidence of the proximity of cell-specific synaptic proteins, such as protein kinase C-alpha for rod bipolar cells and synaptophysin for transplanted photoreceptor cells, as evidence supporting synaptic formation [13]. Other groups have used glutamatergic synaptic blocker L-AP4 coupled with light stimulus testing to evaluate the functionality of synaptic structures formed between transplanted organoid-derived cone photoreceptors and recipient bipolar cells [25, 28]. Various functional assays, including behavioral light avoidance testing, multielectrode array recording, superior colliculus electrophysiology recording, and optomotor response testing among others, have been used to measure the extent of neural circuit repair in recipient tissues [10, 13, 25, 50, 51, 52]. Literature supporting the therapeutic effects of synaptogenesis in the recipient retina has been confounded by the discovery of cellular component transfer between donor and recipient photoreceptors [29]. Nonetheless, the appropriate localization and formation of novel synapses in the recipient retina are relevant to successful transplantation.

Cellular component transfer

Researchers originally postulated that the visual benefits associated with retinal precursor cell transplantation resulted from the survival, incorporation, and synaptogenesis of GFP+ retinal donor cells into a degenerated recipient outer neuronal layer. However, in a portion of the surviving post-transplant GFP+ cells in the recipient retina are native photoreceptors that received a cytoplasmic transfer of cellular components from donor photoreceptors, partially confounding the etiology of the visual function improvements in the animals studied [29–33]. To date, there is evidence to support the transfer of proteins, mRNA, and mitochondria [31, 34]. This process has been variously termed cellular materials transfer, cell-cell fusion, biomaterials transfer, and cellular components transfer (CCT). Whether CCT will confer metabolic or protein complementation benefits that facilitate improved functional vision in rodent models is not understood at present [35]. Furthermore, definitive evidence of CCT has not yet been reported in human or nonhuman primate retinas, likely due to technical considerations precluding histological discrimination between human and nonhuman primate photoreceptor-specific antigens [26, 27].

Glial activation

Retinal gliosis is a cellular phenomenon largely driven by Müller glia, and the extent of gliosis in the recipient tissue can impact donor cell incorporation and survival. In healthy retinas, Müller glial cells are thought to provide a neuroprotective benefit, stabilizing retinal networks to limit retinal cell death. However, in progressive degenerative retinal disease or instances of traumatic insult to the retina, Müller glia may transition into a hyperactive phenotype correlating to an increase in glial fibrillary acidic protein (GFAP) expression [36, 37, 56, 57]. Post-retinal detachment, hypertrophic Müller glial end feet have been shown to induce fibrotic-like phenomena in the subretinal space, creating a so-called ‘glial scar’ that may impede the incorporation of donor tissue into the recipient [38]. Given that many retinal organoid transplantation strategies involve subretinal implantation, efforts to mitigate post-transplant gliotic reactions have become a focus in current research. Strategies to filter tumorigenic/proinflammatory stem cell populations from donor samples before transplantation have shown promise in reducing levels of GFAP in recipient retinal tissue post-implantation [35].

In our research, we were interested in evaluating the extent of gliosis induced by human retinal organoid transplantation. To this end, we cultured retinal organoids using human H9 embryonic stem cells (hESC) containing a Crx:tdTomato reporter. Cellular microaggregates were collected from Crx:tdTomato⁺ retinal organoids and subsequently transplanted into the subretinal space of a mouse model (Rd1/NS) with retinal degeneration and immune deficiency. After four and a half months, we harvested the eyes of the recipient mice and performed immunohistochemistry for GFAP to evaluate gliosis in the transplanted eyes. Our data revealed significant GFAP expression in both the retinal organoid grafts and the recipient retina, likely resulting from active populations of Müller glia or astrocytes which still function in this immunodeficient mouse model [23*]. GFAP⁺ glial cell processes extended from the recipient retina into the grafted tissue and partially encapsulated the grafted retinal microaggregates (Figure 3).

Figure 3. Glial activation in retinal organoid-transplanted mice with retinal degeneration.

The figure shows representative IHC images from a recipient Rd1/NS mouse eye. Transplanted retinal organoid cells were detected by positive immunostaining for human nuclear specific antibody (H). Photoreceptor cells in the donor organoid were identified by the presence of endogenous Crx:tdTomato (Crx:tdTo) reporter. Glial cell activation was assayed through GFAP (GF) staining. DAPI (D) staining was used to visualize cellular nuclei in both the donor and recipient retina. Abbreviations: INL = inner nuclear layer of the recipient. SRS = subretinal space in which the donor cells were delivered. Scale bar, 20 microns.

Immune cell infiltration

While the retina is recognized for its relatively immune-privileged status, concerns about immune rejection in allogenic retinal tissue transplantation remain. However, research efforts have yielded mixed results. One study found that transplanted hESCs retinal organoid tissue into squirrel monkeys demonstrated no evidence of rejection in the absence of systemic immune suppression [27]. Other stem cell transplantation studies have implied that systemic immunosuppression or the use of immunodeficient models can contribute to donor cell survival [5**, 26, 39]. While transplanting major histocompatibility complex (MHC)-homozygote monkey iPSC organoid-derived retinas into monkeys with laser-induced retinal degeneration, one group found that MHC-mismatched transplantation without immune suppression showed no clear clinical signs of rejection [40*]. Together, these findings support the feasibility of using ESC/iPSC-retinal tissue in MHC-mismatched human transplantation studies and suggest that, if present, significant post-transplant immune responses in large animal models can be attenuated with current immunosuppressive medications.

Cellular migration

Recognizing that upon retinal transplantation, some retinal organoid grafts still lack terminally differentiated cell types, recent research has shown that transplanted retinal organoid tissue may differentiate into retinal or nonretinal cell types that exhibit transretinal migration into the recipient retina or potentially induce native cellular populations to migrate into untargeted retinal lamina [23*, 41, 58]. This aberrant migration may damage healthy recipient retinal tissue by mechanically displacing recipient retinal cells. They may also introduce and distribute novel antigens in the retina that could make donor tissue more vulnerable to immune rejection, assuming the migratory cells originate from the transplant [23*]. The functional consequences of organoid cell migration are unclear and require further study.

Influence of recipient characteristics on cell interactions between retinal organoid grafts and recipient cells

Cellular interactions, most notably material transfer and transplant incorporation between the graft and recipient retinal tissue, depend partly on the cytoarchitecture of the host retina and the maturity of the donor tissue. If retinal organoid-derived photoreceptors are transplanted into a recipient retina that lacks an intact external limiting membrane (ELM) but contains a notable population of remnant photoreceptors at the point of transplantation (as in Nrl^−/− and Prph2^rd2/rd2 mouse models), CCT and donor cell incorporation into the recipient retina may be more likely to occur [13, 49]. Additionally, post-mitotic rods and cones that have been derived postnatally may incorporate more readily [53] and participate in CCT more efficiently than immature retinal progenitor cells [49]. However, if retinal organoid-derived photoreceptors are transplanted into a recipient degenerating retina with a relatively intact ELM (as in the Nrl^−/−;RPE65^R91W/R91W mouse model), both CCT and donor cell incorporation may be compromised [49]. Lastly, if the donor cells are transplanted into a recipient retina with a paucity of native photoreceptors at the time of transplantation and consequently a compromised OLM (as in the Aipl1^−/− and rd1 mouse models) donor cells may be more likely to incorporate into the host retina and extend basal terminals toward host interneurons [9]. Therefore, effective clinical therapeutic interventions in the future must be tailored to the degenerative status of the recipient and account for the developmental maturity of the donor tissue. Tools like OCT imaging may help to select patients with end-stage photoreceptor degeneration for somatic augmentation transplants, and refined cell sorting protocols may enable the selection of post-mitotic donor tissue.

With regards to immunosuppression in patients, the subjects in one study received only short-term immunosuppression with tacrolimus beginning eight days prior to implantation and extending to day 42 post-operatively, followed by taper and eventual termination at day 60 [59]. In addition to the perioperative tacrolimus, an intravenous injection of methylprednisolone was administered prior to surgery on day 0. Post-mortem histological data from one subject 2 years after subretinal implantation of HLA-mismatched donor RPE cells showed no significant inflammation in the eye at the transplant site or in the retino-choroidal vasculature, and nor were humoral immune response markers detected during follow-up, even without long-term administration of immunosuppressive agents. These data shows that short-term perioperative immunosuppression may suffice to ensure long-term graft viability. Keeping in mind that some transplant recipients may be of very young or very advanced age, further minimization of the immunosuppression in terms of dose and/or duration may be advantageous in terms of systemic implications.

Future challenges

Several challenges pertain to the technical considerations of organoid culture. For example, the heterogeneity, long culture time, and limited yield of retinal organoids represent significant barriers to their laboratory study and clinical translation. Variability in culturing technique can yield inconsistent organoid density with varied yield of retinal cell populations. Furthermore, the limited size of organoids limits the mature cell yield viable for transplantation [42, 43], and there are some retinal cell subtypes that may play significant roles in optimizing donor-recipient interactions that are not recapitulated in 3D retinal organoid culture [44]. Mechanical damage incurred during fluorescence-associated cell sorting or magnetic-associated cell sorting before transplantation may compromise the health and number of transplantable cells [9, 45, 46], and research surrounding the postulated neuroprotective effects of these cells post-transplantation remains in its infancy. Lastly, the spatial distribution of “red,” “green,” and “blue” cone photoreceptors, as well as rod photoreceptors, varies across the regions of the retina [47]. However, the retinal photoreceptor organoid grafts transplanted to date have lacked a spatial photoreceptor arrangement that mimics the organization of the native retina. This likely results from the technical challenges of positioning specific cell subtypes on biocompatible scaffolds for transplantation [48]. While the progress made in the field of retinal organoid transplantation holds great promise for the clinical treatment of degenerative retinal conditions, more research is required to ensure these cell therapies provide durable and clinically meaningful benefits in target populations.