Bullous keratopathy is characterized by corneal endothelial decompensation associated with irreversible stromal edema. Conditions or events that give rise to bullous keratopathy include pseudophakic bullous keratopathy (PBK), a sterile or infectious inflammation, or FECD (Morishige and Sonoda, 2013). Endothelial decompensation in PBK is thought to be multi-factorial and can be caused by intra-operative surgical trauma or post-operative events (inflammation, high intra-ocular pressure) following intra-ocular surgery. Contrary to PBK (where the endothelium is not diseased), FECD manifests by the progressive thickening of Descemet’s membrane, and the development of extracellular matrix-related excrescences, called guttae, on its endothelial surface, that can be embedded by a fibrillar layer at later stages (Figure 1). Concurrently, there is a decrease in cell density. Due to this cell loss, the corneal endothelium loses its integrity and is not able to maintain stromal deturgescence, which will create stromal edema and painful epithelial bulla (Elhalis et al., 2010). Evidence that FECD Descemet’s membrane constitutes a toxic environment for the cells are numerous (Rizwan et al., 2016; Kocaba et al., 2018; Hribek et al., 2021), which supports the importance of removing the diseased Descemet’s membrane, and thus also increases the need for the development of a Descemet’s membrane substitute.

PBK and FECD are the first common causes of corneal transplantation worldwide (Ghosheh et al., 2008; Gain et al., 2016; Matthaei et al., 2017; Iselin et al., 2021). The current treatment is to replace the dysfunctional corneal endothelium with a healthy one through corneal transplantation. Descemet’s stripping automated endothelial keratoplasty (DSAEK) and Descemet’s membrane endothelial keratoplasty (DMEK) are currently the two preferred surgical techniques for corneal endothelial transplantation, with both techniques presenting fewer risks of surgical complications and permitting faster visual recovery with a lower rate of graft rejection than full thickness grafting techniques such as penetrating keratoplasty (Turnbull et al., 2016).

Despite the high success rate of DSAEK/DMEK, important issues remain, such as graft dislocation (Price and Price, 2006; Trindade and Eliazar, 2019), immunological rejection (Price et al., 2009; Hos et al., 2017; Price et al., 2018) and primary donor failure (Deng et al., 2018), which may require surgical revisions, prolonged corticosteroid treatment or re-grafting. DMEK/DSAEK surgical procedures are invasive, and their long-term clinical results remain unclear. Furthermore, corneal transplantation and eye banking facilities are not equally available around the world and most countries suffer from a severe shortfall of donated tissues. Globally, 12.7 million patients await transplantation, with only 1/70 grafted (Gain et al., 2016). The limited quantity of high-quality donor tissue and their high demand are limiting factors justifying the need for the development of alternative therapies. A promising approach is to expand donor CECs in vitro (allowing cells from one donor cornea to treat many patients), seed them on a compatible biomaterial, and transplant them as a functional endothelial monolayer.

As a proof of concept, immortalized CECs (Bednarz et al., 2000; Schmedt et al., 2012) can be used to verify that cells adhere and reform a monolayer on the tested biomaterial. However, moving towards human implantation, untransformed primary cells should be privileged, as immortalized CECs have been modified to increase their proliferative potential.

Proliferation of human CECs can be induced in vitro by disrupting contact inhibition and adding growth factors (Joyce and Zhu, 2004). Unfortunately, this expansion is limited, and cells tend to lose their endothelial phenotype following repeated passaging (Roy et al., 2015), a process attributed to endothelial-to-mesenchymal transition (Engelmann et al., 1988; Engelmann et al., 2004; Zhu and Joyce, 2004; Okumura et al., 2013; Roy et al., 2015). This change in phenotype upon culture can be explained by the partial mesenchymal nature of CECs. Indeed, native corneal endothelial cells express mesenchymal markers, such as vimentin (Foets et al., 1990; He et al., 2016) and n-cadherin (Zhu et al., 2008). On top of mesenchymal markers, patches of cells in the corneal endothelium also display epithelial markers, such as keratins 8 and 18 (Foets et al., 1990; Merjava et al., 2009; He et al., 2016), and neuronal markers, such as neuron specific enolase (NSE) (Hayashi et al., 1986). The heterogenous nature of the corneal endothelium was recently been brought to light using scRNAseq analysis, where 2 to 4 distinct subpopulations were identified in the native tissue (Català et al., 2021; Wang et al., 2022), and up to seven in primary cultures of CECs, including some clusters with senescent or fibrotic phenotypes (Català et al., 2023). CECs with a fibroblastic phenotype will not be able to form a homogenous monolayer capable of forming the cellular junctions essential for the functionality of the endothelium (Roy et al., 2015). Many studies have thus addressed the influence of cell culture conditions in order to maintain an endothelial phenotype, including different cell isolation techniques (Li et al., 2007; Choi et al., 2014; Santerre and Proulx, 2022), cell culture media formulation (reviewed in (Peh et al., 2011)), and dual media approaches (Peh et al., 2015; Beaulieu Leclerc et al., 2018). Maintaining an endothelial phenotype is key in regenerative medicine for preserving the cell monolayer’s physiological function. Cell morphology is a good indicator of an endothelial phenotype. It can be measured by calculating cell circularity based on the area and the perimeter of the cells (Circularity index = 4π × Area/Perimeter²) (Peh et al., 2013; Ang et al., 2016). Using this formula, a round cell will have a circularity index of 1, a hexagonal cell will have a circularity index of 0,88, and a fibroblastic cell will have a circularity index below 0,55 (Xu et al., 2021).

However, cell morphology is not sufficient to determine cell quality. Using immunofluorescence and flow cytometry, mature CECs can be distinguished according to their expression of specific cell markers. Cells can be evaluated by their expression profile of specific clusters of differentiation (CD). Mature differentiated CECs should express the following profile of CD; CD166⁺, CD44^−/dull, CD24⁻, and CD105⁻ (Kitazawa et al., 2023). Some membrane proteins also can be used as markers as they are only expressed in mature CECs, such as SLC4A11, which is only expressed in a low mitogenic environment and low passage in vitro (Frausto et al., 2020).

The limited expansion of primary cells can also be circumvented using stem cells. Many types of stem cells can be differentiated into CEC-like cells. Unipotent stem cells can be isolated by sphere-forming assay (Yokoo et al., 2005; Amano et al., 2006; Mimura et al., 2010). However, these progenitor cells may acquire epigenetic modifications through passages leading to inhibition of proliferation and differentiation (Navaratnam et al., 2015). Pluripotent embryonic stem cells (ESCs) (Hanson et al., 2017) and induced pluripotent stem cells (iPSCs) (Wagoner et al., 2018) can also be used but are associated with ethical concerns (King and Perrin, 2014) and teratoma formation (Lee et al., 2009). Alternatively, CEC-like cells can be differentiated from adult multipotent mesenchymal stem cells (MSCs) isolated from umbilical cord blood (Joyce et al., 2012), bone marrow (Shao et al., 2011), corneal stroma (Ju et al., 2012; Hatou et al., 2013), and skin (Inagaki et al., 2017; Shen et al., 2017; Gutermuth et al., 2019), therefore negating the ethical concerns and teratoma formation risen by pluripotent cells. The use of stem cells eliminates the need for donor ocular tissue. It opens the door for autologous treatment, but further research is needed to characterize the precursor cells and their derivates. Furthermore, genetic editing techniques such as CRISPR could allow autologous treatment of patients with genetic endotheliopathies, such as FECD.

Collagen is an abundant protein in most extracellular matrices (Ricard-Blum, 2011), and Descemet’s membrane is no different (de Oliveira and Wilson, 2020). Collagen and its denatured form, gelatin, are widely used in the tissue engineering of many organs because of their biocompatibility, biodegradability, hydrophilicity, and affordability (Rezvani Ghomi et al., 2021). However, they tend to be fragile and swell when maintained in aqueous conditions, leading to decreased transparency (Watanabe et al., 2011). Cross-linking increases chemical bonds between the collagen fibrils, therefore improving mechanical strength (Jumblatt et al., 1980; McCulley et al., 1980; Mimura et al., 2004b; Kimoto et al., 2014; Rizwan et al., 2017). Unfortunately, many cross-linking techniques require chemical compounds that may cause toxicity to the cultivated cells or the host. Alternatively, collagen can be compressed, making the biomaterial sturdier without chemical cross-linking (Levis et al., 2012; Cen and Feng, 2018). Another derivate of collagen used in corneal endothelium bioengineering is Vitrigel (Koizumi et al., 2008; Yoshida et al., 2014; Yoshida et al., 2017). The latter is produced by gelation, vitrification, and rehydration of collagen type I, creating a resistant thin transparent membrane (Takezawa et al., 2004).

Since collagen is a large protein, that is, hard to manipulate, collagen-like peptides (CLPs) are being studied as short peptides able to self-assemble into triple helices mimicking full-length collagen hydrogels (Wang et al., 2008; Stahl et al., 2010; O'Leary et al., 2011; Luo and Kiick, 2013). CLPs combined with polyethylene glycol (CLP-PEG) have already been grafted into mini-pig models as corneal stromal replacement and demonstrated good biocompatibility and sufficient mechanical strength (Jangamreddy et al., 2018). Pilot studies have demonstrated that this biomaterial allows adhesion and proliferation of CECs (Sasseville et al., 2022). This material showed excellent potential for future graft but still requires optimization as long-term culture was difficult. An advantage of this approach is that short extracellular matrix (ECM) sequences can be added in the hydrogel. For example, the laminin-derived peptide sequence YIGSR can be cross-linked into the hydrogel, which could help CECs to remain onto the biomaterial.

Silk fibroin membranes have also been extensively used for their biodegradability, non-immunogenicity, and inherent optical and mechanical attributes (Cao and Wang, 2009; Lee et al., 2016). In vitro, silk membranes can support human (Madden et al., 2011; Ramachandran et al., 2020) and rabbit (Kim et al., 2018; Song et al., 2019) CEC maturation. Silk has also been used in an in vivo rabbit model and could recover corneal transparency over a 6-week follow-up period (Vazquez et al., 2017). Furthermore, Hazra et al. demonstrated that despite the much greater tensile strength of silk membrane compared to native Descemet’s membrane, human CECs secreted similar ECM proteins and exhibited endothelial phenotype in long-term cell culture (Hazra et al., 2023). Long-term in vivo studies are still needed to evaluate safety and performance.

Chitosan is a polysaccharide derived from crustaceans. It is biocompatible, biodegradable, antibacterial, antifungal, and easy to fabricate, but it is fragile and hydrophobic. For those reasons, chitosan is often used combined with other compounds. Polycaprolactone (PCL) is a polymer that can be added to chitosan without extensive chemical modifications (Sarasam and Madihally, 2005) and allows CEC culture (Wang et al., 2012; Young et al., 2014; Wang et al., 2019). However, chitosan-PCL scaffolds can be fragile. Supplementing the model with chitosan nanoparticles can amplify its mechanical strength (Tayebi et al., 2021).

A mature monolayer of CECs has also been achieved using a self-assembled corneal stromal substitute (Proulx et al., 2010; Jay et al., 2015; Bourget and Proulx, 2016). The biocompatibility of the corneal stromal substitute was assessed by intrastromal transplantation in a cat model and showed no sign of rejection, inflammation, and toxicity over a four-month follow-up period (Boulze Pankert et al., 2014). The self-assembly approach is based on the inherent capacity of fibroblasts to secrete and assemble their own ECM when cultured in the presence of serum and ascorbic acid. This technique has been used to produce tissue-engineered bone, skin, connective and adipose tissues, cornea, and choroidal substitutes (Proulx et al., 2010; Labbé et al., 2011; Boa et al., 2013; Galbraith et al., 2017; Djigo et al., 2019). It is a great alternative that does not require exogenous biomaterials and that could be produced using the patient’s own cells, which would avoid graft rejection.

In an effort to be as biocompatible as possible, native thin membranes have also been proposed as biomaterials for the engineering of the corneal endothelium. Unlike biomimetic substrates, they require organ donation but have a more complex composition that can better mimic the Descemet’s membrane composition. However, their allogenic/xenogeneic nature and their mechanical resistance can be issues.

Denuded amniotic membranes have been used as a corneal endothelium biomaterial. While the transparency is sub-optimal, endothelialized amniotic membranes have been grafted in rabbits (Ishino et al., 2004) and cats (Fan et al., 2013), leading to decreased corneal edema and increased transparency as opposed to ungrafted eyes. Recently, Zhao et al. (2022) explored cross-linking as a way of enhancing the mechanical strength of amniotic membranes and ECM protein coating to improve cell culture. They found increased mechanical strength with cross-linking, improved adhesion, and proliferation with ECM coating, and return of transparency when grafted into cats and monkeys (Zhao et al., 2022).

The anterior lens capsule is a transparent thick basement membrane with a similar collagen composition (type IV collagen and laminin α5 and β2) as the posterior surface of the Descemet’s membrane (Nakayasu et al., 1986; Marshall et al., 1991; Ronci et al., 2011) and is compatible with integrin binding of CECs (Choi et al., 2013). Human (Yoeruek et al., 2009) and pig (Spinozzi et al., 2019) CECs can reform an endothelial monolayer and can sustain corneal transparency when grafted into pigs (Spinozzi et al., 2019). The main downfall with anterior lens capsule as a biomaterial is donor dependant variability.

Tilapia is a fish that possesses scales composed of type I collagen (Chen et al., 2011). Once decalcified and decellularized, fish scales form a biodegradable (van Essen et al., 2013), transparent (Ikoma et al., 2003) material that can be cellularized with or without ECM protein coating (Parekh et al., 2018; Hsueh et al., 2019; Li et al., 2020).

Descemet’s membrane itself can be used as a carrier. Indeed, Descemet’s membrane was the first carrier used for the engineering of the corneal endothelium (Jumblatt et al., 1978; Gospodarowicz et al., 1979). At the time, corneal buttons were excised and deendothelialized using a cotton swab. The cornea with the remaining denuded Descemet’s membrane was seeded with cells, maintained in culture, and grafted through full thickness keratoplasty. Since then, improvements have been made and the reconstructed endothelium is grafted using only Descemet’s membrane (Lu et al., 2020). Nonetheless, human Descemet’s membranes are limited due to donor shortage and animal Descemet’s membranes represent a risk of xenograft rejection.

Biologic and biomimetic biomaterials can have high variation between batches. Synthetic substrates could therefore be a great alternative for low variability and highly pure and defined materials. They can also be great off-the-shelf options that are mechanically strong and have tunable degradation time and porosity. Many have been tested; Poly (methyl-methacrylate) (PMMA), polycaprolactone (PCL), poly (lactic-co-glycolic acid) (PLGA) (Kruse et al., 2018), poly-ε-lysine (Kennedy et al., 2019), and poly (ethylene glycol) (PEG) (Ozcelik et al., 2014) and have been shown to allow endothelial monolayer formation in vitro. However, many synthetic polymers are hydrophobic. This can be an issue since hydrophobic substrates tend to have low permeability to small hydrophilic molecules such as glucose, little to no degradation in aqueous conditions and low biocompatibility (Kruse et al., 2018). Biomaterials with low degradation can raise long-term toxicity concerns. For example, PMMA is non-degradable and shows long-term cytotoxicity (Kruse et al., 2018). Albeit, PCL has a slow degradation rate but breaks down in small non-toxic molecules that can be evacuated form the aqueous humor, causing no toxicity (Sun et al., 2006). To improve biocompatibility, synthetic scaffolds can be crosslinked to ECM proteins or peptides to enhance chemical interactions between the cell membrane and the substrate (Palchesko et al., 2015; Kennedy et al., 2019). They can also be amalgamed with biomimetic compounds to improve overall mechanical strength and biocompatibility (Young et al., 2014; Chen et al., 2015; Van Hoorick et al., 2020; Himmler et al., 2021; Tayebi et al., 2021). A way to incorporate synthetic materials to biomimetic compounds is by layering. Van Hoorick et al. used a thin layer of poly (D,L-lactic acid) (PDLLA), a FDA-approved biodegradable polymer, to mechanically enhance gelatin sheets (Van Hoorick et al., 2020). They were able to obtain a transparent, glucose permeable 1 µm thick membrane with a Young’s modulus comparable to the native Descemet’s membrane. Song et al. (2022) layered poly (lactide-co-caprolactone) (PLCL) on top of decellularized ECM synthetized by umbilical cord blood mesenchymal stem cells. They were able to obtain a robust transparent 20 µm thick membrane. Both layered synthetic-biological membranes were able to sustain immortalized and primary human CEC culture exhibiting proper morphology and function-related proteins. Nonetheless, further studies are still necessary to ensure long term safety and efficiency.

In parallel to tissue engineering of the corneal endothelium, other groups have focused their efforts on reforming a functional corneal endothelium by injecting CECs into the anterior chamber, thereby bypassing the need for a scaffold. While most pre-clinical studies have focused on the injection of cultured CECs (Mimura et al., 2005; Bostan et al., 2016; Okumura et al., 2018), it has also been suggested that injection of noncultured CECs could reform a functional endothelium and restore corneal transparency. Interestingly, this process, described as simple noncultured endothelial cells injection (SNECi), represents a simplified method when compared to tissue engineering or cultured endothelial cell injection techniques, as it does not necessitate the proliferation of cells that need to maintain an endothelial phenotype throughout the cell culture process (Ong et al., 2020). Moreover, it was noted that with both cultured and noncultured cells injection techniques, results were comparable to those of studies using tissue engineered corneal endothelium (Peh et al., 2019). However, it was also noted that to promote adhesion of injected cells to Descemet’s membrane, strict prone positioning had to be maintained for at least 3 h after surgery (Mimura et al., 2007; Peh et al., 2019) and techniques to enhance CEC delivery and adhesion to the inner cornea such as the use of magnetic particles might be necessary (Moysidis et al., 2015; Xia et al., 2019).

On the other hand, although it is generally accepted that both endothelial cell injection and tissue engineered endothelium techniques will help alleviating the global shortage of donor corneas as compared to more traditional techniques such as DSAEK/DMEK, it is unclear yet if either of these techniques will be more cost-effective than the other (Soh et al., 2021; Kaufman and Jun, 2023). Nonetheless, tissue engineering offers some theoretical advantages, such as a more finely tuned targeted delivery to the diseased area. It is also likely that patients with a more advanced disease, for example, patients with advanced FECD in whom the Descemet’s membrane is severely compromised, might benefit more from Descemet’s membrane stripping, guttae removal and replacement of the endothelium using a tissue engineered corneal endothelium grown on a scaffold, rather than using direct intracameral injection of CECs.

Age-related macular degeneration (AMD) is a multifactorial disease that progressively causes irreversible central vision loss in the elderly population (≥65 years old) and accounts for about 10% of blindness worldwide (Al-Khersan et al., 2019; Sarkar et al., 2022). AMD can be classified into two forms, dry and wet AMD, and is characterized by the deposition of extracellular deposits, called drusen, between Bruch’s membrane and the RPE (Figure 1), which over time can cause RPE and photoreceptors degeneration in the center of the macula (Thomas et al., 2021). Bruch’s membrane structure and biomechanical properties encounter age-related changes contributing to AMD. Elasticity, permeability, and diffusion reduction are some of these changes that occur due to increased collagen cross-linking, thickness and stiffness, and drusen deposition, respectively (Booij et al., 2010; Tisi et al., 2021).

Dry AMD contributes to 90% of AMD cases, while 10% of patients present wet AMD symptoms. AMD commonly starts with the dry form, then can progress to the wet form (Sarkar et al., 2022). As an impact of choroidal neovascularization, abnormal sub-RPE leaky blood vessels grow from the choriocapillaris into the retina and can lead to the accumulation of fluid or hemorrhage in the retina, which can dramatically reduce vision. (Gryziewicz, 2005; Chakravarthy and Peto, 2020). AMD results from various genetic, environmental and lifestyle risk factors, including aging, oxidative stress, smoking, comorbidity diseases (such as hypertension, hypercholesterolemia, arteriosclerosis, obesity), and single nucleotide polymorphisms (SNPs) in genes like complement factor H and ARMS2/HTRA1 (Lim et al., 2012; Shaw et al., 2016). Oxidative stress is a major contributor of AMD (Jarrett and Boulton, 2012). The RPE resides in a high oxidative stress environment because of their high metabolic activity, photo-oxidative stress from light exposure, presence of photosensitizers like rhodopsin and lipofuscin, and high polyunsaturated fatty acids level. This environment favors ROS production and increases the probability of oxidative damage and cell death of RPE and photoreceptor cells (Hanus et al., 2015; Abokyi et al., 2020; Zhang et al., 2020).

The current gold standard treatment for wet AMD is intravitreal injections of anti-vascular endothelial growth factor (anti-VEGF). This treatment option has proven to slow exudative AMD progression and improve visual outcomes by decreasing retinal fluid and subretinal neovascularization (Hernández-Zimbrón et al., 2018). For dry AMD, nutritional supplements, such as antioxidants and minerals (AREDS and AREDS2 studies) (Age-Related Eye Disease Study Research, 2001; Age-Related Eye Disease Study 2 Research, 2013; Chew et al., 2022), have been reported to prevent the progression from intermediate to more advanced AMD in about 25% of cases. However, despite the higher prevalence of dry AMD, there is currently no effective treatment for reversing dry AMD or for preventing GA (Schmidl et al., 2015). The investigated treatment approaches for dry and wet AMD are summarized in Table 2. This lack of effective treatment options for patients suffering from dry AMD has led to remarkable efforts to identify novel treatment strategies. Tissue engineering of the RPE is one of them.

ARPE-19 is a human RPE cell line with polarized epithelial cell morphology that expresses genes specific to native RPE cells. ARPE-19 can exhibit functions equivalent to human RPE, like photoreceptor outer segment phagocytosis and epithelial monolayer formation on supportive scaffolds (Dunn et al., 1996; Samuel et al., 2017). Since ARPE-19 cells do not completely resemble human RPE cells transcriptome and functions, do not express certain RPE-related proteins, and might lose some differentiated properties after multiple passages, these cells are valuable only for in vitro studies (Samuel et al., 2017; Gupta et al., 2023).

Various cell sources for RPE cell replacement have been suggested. Autologous RPE cells referred to RPE-choroid sheet or RPE cell suspension harvested from the nasal subretinal area of a patient’s eye and manipulated in vitro before transplantation is one of them (Assadoullina et al., 2004; Binder et al., 2004; Boulton et al., 2004). Autologous RPE cells may carry the same genetic mutations related to AMD (Binder et al., 2007). Allogeneic fetal RPE cells derived from human fetuses in 15–17 weeks of gestational age have also been used for RPE cell transplantation in the foveal area after surgical removal of the defective RPE layer and the Bruch’s membrane. Immune rejection, however, is the limitation of allogeneic transplantation (Algvere et al., 1994; Algvere et al., 1999).

Pluripotent stem cells like human embryonic stem cells (hESCs) and induced pluripotent stem cells (iPSC) can be differentiated towards an RPE fate using a rapid and directed method described by Foltz and Clegg (Foltz and Clegg, 2017). In brief, stem cells were seeded in plates coated with an ECM-based hydrogel, and cultured for 14 days in the presence of growth factors (insulin growth factors, basic fibroblast growth factor, transforming growth factor beta) and WNT pathway agonist. This protocol generated RPE-like cells that were isolated and cultured in RPE cell maintenance medium for further passage and cryopreservation. The generated RPE cells showed specific polarized, pigmented, and polygonal morphology, expressed RPE markers, and secreted RPE growth factors. This protocol has been used extensively to obtain RPE cells from different stem cell sources (Foltz and Clegg, 2017; Parikh et al., 2022; Regha et al., 2022). Autologous and allogeneic iPSC-derived RPE cells have been employed in human studies. Mandai et al. generated iPSC from dermal fibroblasts of two AMD patients, differentiated them to RPE cells, and transplanted the autologous iPSC-derived RPE cell sheet subretinally (Mandai et al., 2017). In another clinical trial, allogeneic iPS-RPE cells of an HLA homozygous donor were transplanted to 5 HLA-matched patients with wet AMD (Sugita et al., 2020).

Human umbilical cord mesenchymal stem cells (hUCMSC) differentiation to PRE cells using CRX, NR2E1, C-MYC, LHX2, and SIX6 transcription factors is outlined recently by Zhu et al. (2022) hUCMSC-derived iRPE cells were implanted into subretinal space of rat models and demonstrated desirable therapeutic effects. Furthermore, Bone marrow mesenchymal stem cells (BMSCs) treated with ciliary neurotrophic factor were injected intravitreally to diabetic rat eyes. Differentiation of BMSCs to RPE cells was then assessed to evaluate retinal regeneration (Huang et al., 2021).

Collagen types I, III, IV, and V are major components of Bruch’s membrane, and most of the studies are devoted to develop scaffolds using these ECM proteins. Collagen substrates have been delivered to the subretinal space as crosslinked and non-crosslinked collagen, collagen film, and ultrathin collagen membranes (Susanne, 2011; Murphy et al., 2020). Bhatt et al. (1994) transplanted human fetal RPE cells cultured on type I collagen into the subretinal space of rabbits in two forms: non-crosslinked (soft and flexible) or crosslinked with ultraviolet light (rigid). During the 6 weeks of examination, they observed that RPE cells remained as a monolayer on the non-crosslinked collagen, which complied with the shape of Bruch’s membrane. They suggested that this approach could possibly transfer growth factors or cytokines and prevent retinal degeneration. Other studies investigated RPE cell differentiation characteristics, viability, and morphology after being cultured on thin (10 μm-thick) and ultrathin (7 µm-thick) collagen type I membranes which are approved for clinical use. This easily degradable membrane could be an anchorage for RPE cells to maintain their specific characteristics after subretinal transplantation in vivo (Thumann et al., 2006; Thumann et al., 2009). Recently, Yamashita et al. (2023) generated a biodegradable hybrid PCL/collagen nanosheet and grafted it into the subretinal space of a rat model after loading RPE cells on it. This nanosheet showed supportive properties to increase RPE cells survival and monolayer formation with tight junctions. Due to its great mechanical strength, controlled drug delivery, and strong cell adherent property, this RPE cell-loaded PCL/collagen nanosheet could be considered a potent design for cell monolayer transplantation, and retinal degenerative diseases treatment.

Chitosan is a positively charged polysaccharide derived from chitin deacetylation. It is a biocompatible, biodegradable, and mucoadhesive polymer with antibacterial, wound healing, and penetration enhancement properties. Noorani et al. (2018) designed an electrospun gelatin/chitosan nanofibrous substrate that provided a suitable surface for RPE cell adhesion in vitro. They disclosed that a mixture of gelatin and chitosan could imitate Bruch’s membrane composition and nanofibrous structure that help RPE cells to preserve their phenotype. Chitosan conjugation with a peptide (serine-threonine-tyrosine) induced tyrosine kinase activity in RPE cells which in turn enhanced photoreceptor outer segment phagocytosis by RPE cells. This conjugated nano chitosan, synthesized by Jayaraman, et al., could therefore serve as a carrier for retinal drug delivery and AMD treatment (Jayaraman et al., 2012).

Silk is mainly produced in the glands of silkworms, commonly Bombyx mori, during their metamorphosis. Bombyx mori silk has been widely used as a biomaterial for tissue engineering purposes (Kundu et al., 2013; Koh et al., 2015). This material has several advantages including great biocompatibility, chemical modifiability, slow degradation, excellent permeability to water and oxygen, and mechanical strength due to the presence of fibroin proteins (Rockwood et al., 2011). Silk fibroin membranes have been readily exploited as a Bruch’s membrane substitute to support ocular cells attachment, growth, and differentiation (Shadforth et al., 2012). As demonstrated in several studies, RPE cells from different sources were cultured on an ultrathin (mostly 3 µm) and porous B. mori silk fibroin membranes to explore RPE cell behavior, adhesion, proliferation, survival, and maturation after cultivation. Considering the results, RPE cell function on a scaffold manufactured from silk fibroin were equivalent to their function on native Bruch’s membrane (Shadforth et al., 2012; Shadforth et al., 2017; Galloway et al., 2018). To optimize the physical properties of silk fibroin membranes, porogens like PEG, crosslinking agents like HRP, polycaprolactone (PCL), gelatin, and elastin (in the form of human recombinant tropoelastin) were incorporated into these scaffolds (Xiang et al., 2014; Shadforth et al., 2015; Suzuki et al., 2019; Belgio et al., 2020). Jeong et al. (2020) produced hydrogels from silk fibroin and mixed it with a synthetic biomaterial, PEG, at different sonication times. RPE cells were isolated from rabbit eyes and seeded on these hydrogels. Scaffold’s porosity, strength, degradation, RPE cell attachment, and proliferation were then measured. Based on the results, the hybrid silk fibroin hydrogel and PEG scaffold prepared with a 20-s sonication period could be a promising material for retinal tissue engineering regarding its beneficial effects on RPE cell growth and gene expression (Jeong et al., 2020).

The self-assembly approach of tissue engineering has been employed to construct tissues without the need for synthetic or exogenous biomaterials (see Section 3.1). Djigo et al. (2019) developed a self-assembled choroidal substitute using choroidal fibroblasts and demonstrated similarities between the engineered tissue and the native tissue concerning ECM composition, thickness, biomechanical properties, and biocompatibility for RPE cells, vascular endothelial cells and choroidal melanocytes. This choroidal stroma, therefore, could become a biomaterial to transplant the RPE reformed on a choroidal substitute.

Other biomimetic biomaterials, such as gelatin (Shakibaie et al., 2018), alginate (Wikström et al., 2008), laminin (Peng et al., 2016), fibrin glue (Ahmed et al., 2015), bioprinted Bruch’s membrane (Kim et al., 2021; Kim et al., 2022), and autologous cryoprecipitate (Farrokh-Siar et al., 1999) have been assessed by several studies as substrates for RPE transplantation into the subretinal space.

Human amniotic membrane is a thin (20–50 μm), translucid, and elastic tissue that represents the innermost layer of the placenta. Owing to its ECM structural composition (IV and VII collagen, hyaluronic acid, laminin, fibronectin, and proteoglycans), this membrane has exhibited favorable mechanical properties like permeability, cell adhesion, and elasticity (Tosi et al., 2005; Leal-Marin et al., 2021). Additionally, since this tissue can produce different growth factors and cytokines, and is immune-privileged, anti-angiogenic, anti-inflammatory and anti-microbial, various ophthalmic and tissue engineering studies have widely used this natural membrane for decades (Rizzo et al., 2020; Fénelon et al., 2021). Ohno-Matsui et al. (2005) examined RPE cell characteristics cultured on human amniotic membranes and confirmed the preserved epithelial morphology and differentiation phenotype of these cells. Despite its advantages, limitations like low biomechanical consistency and rapid biodegradation have restricted its application (Ma et al., 2010). In this respect, Majidnia et al. (2022) blended human amniotic membranes with a synthetic biomaterial and revealed high porosity, hydrophilicity, cell adhesion, and lack of toxicity attributed to this ultrathin amniotic membrane powder/PCL scaffold generated by electrospinning technique. This hybrid substrate could potentially support RPE cell adhesion, viability, and morphology.

Descemet’s membrane is another natural substrate with features resembling Bruch’s membrane characteristics concerning the abundant presence of type IV collagen, laminin and fibronectin. Daniele et al. (2023) used this membrane to produce a polarized monolayer of human embryonic stem cell-derived retinal pigment epithelial (hESC-RPE) cells in vitro.

Discussed biomimetic and natural biomaterials have some drawbacks that make them inefficient for clinical usage and need to be overcome in future studies. Poor biomechanical strength, high biodegradability, difficult processing condition, infection probability, and the possibility of constituents and property changes associated with processing procedures are certain disadvantages related to natural biomaterials. Although some methods like cross-linking have been introduced to enhance their mechanical strength, they have the disadvantage of producing a thicker, non-permeable, and non-biodegradable biomaterial (Karwatowski et al., 1995; Nair et al., 2021; Alvarez Echazú et al., 2022; Wang, 2023).

Hydrogels imitate Bruch’s membrane structure and can be physically and chemically cross-linked to natural (collagen, chitosan, alginate, gellan gum, hyaluronic acid) and synthetic polymers (poly (ethylene glycol) (PEG), poly-L-lysine (PLL), poly (3-caprolactone) (PCL), poly (methacrylate-co-meth acrylamide) (PMMA), poly (vinyl alcohol) (PVA), and poly (hydroxyethyl methacrylate) (PHEMA)) to encapsulate cells, stem cells, and drugs (Ballios et al., 2010; Wang and Han, 2017; Park et al., 2018; Cooper and Yang, 2019; Lin et al., 2021; Youn et al., 2022).

Poly (glycolic acid) (PLGA) is an FDA-approved biomaterial with remarkable properties including high biocompatibility, non-toxicity, high cell adhesion, good mechanical properties, and adjustable biodegradation rate. Sharma et al. isolated CD34⁺ cells from peripheral blood of three dry AMD patients to produce oncogenic mutation-free clinical-grade iPSC-RPE patches. They seeded iPSC-RPE cells on PLGA substrates and transplanted them into the subretinal area of immunocompromised rats and pigs. These PLGA iRPE patches significantly improved RPE monolayer integration into the host Bruch’s membrane, differentiation, and functionality (Sharma et al., 2019).

Poly (chloro-para-xylylene) (parylene C) is a FDA approved class IV biocompatible polymer, that is, widely adopted for biomedical applications and drug delivery thanks to its mechanical strength, flexibility, transparency, chemical inertness, conformability, and good encapsulation efficiency (Golda-Cepa et al., 2020; Coelho et al., 2023). In a long-term study, Rajendran Nair et al. (2021b) transplanted human iPSC-RPE cultured on an ultrathin parylene C membrane into the subretinal space of immunodeficient RCS rat models. They assessed iPSC-RPE survival, RPE-specific marker expression, phagocytosis function, and vision restoration after 1, 4, and 11 months. The results revealed survival and functionality of the polarized iPSC-RPE monolayer on this substrate, where RPE survival was observed in 50% of the rats after 11 months. Lu et al. (2012) reduced the thickness of parylene C to the submicron scale (0.15–0.30 μm), which simulated human Bruch’s membrane permeability. The ultrathin (0.30 μm) parylene C was permeable to nutrients and macromolecules needed to nourish RPE cells and could preserve RPE cell features like adherence, proliferation, monolayer formation with tight junctions, and polarization with microvilli. These findings denote that this structure can be an artificial Bruch’s membrane for RPE cell transplantation. Acting as an artificial Bruch’s membrane, the mesh-supported submicron parylene C membrane was used by Koss et al. (2016) to surgically implant polarized human embryonic stem cell-derived RPE (hESCRPE) monolayer in 14 minipig eyes. Retinal preservation, normal RPE and choroid morphology, RPE monolayer survival, lack of local inflammatory response or peri-implant fibrosis were favorable results described in this study.

Polycaprolactone (PCL) is another FDA approved synthetic polymer for controlled intraocular drug delivery. This polymer is identified by its notable slow degradation rate, great in vivo biocompatibility, low melting temperature, hydrophobicity, and high porosity nature (Jiang et al., 2020; Waterkotte et al., 2022). Belgio et al. (2021) developed a 3D in vitro model of a Bruch’s membrane substitute and an RPE monolayer. They fabricated the Bruch’s membrane substitute using silk fibroin and PCL. The produced Bruch’s membrane showed 44 µm thickness as well as structural and mechanical behavior, biocompatibility, porosity, and permeability comparable to physiological Bruch’s membrane. They concluded that the seeded RPE cells could adhere and grow to this 3D in vitro substrate (Belgio et al., 2021). Liu et al. (2022a) employed the electrohydrodynamic jet (EHDJ) printing method to fabricate ultrathin EHDJ-printed PCL scaffolds with small pore sizes for RPE cell culture. This EHDJ-printed PCL substrate was similar to human Bruch’s membrane in terms of biomimetic features such as thickness, biomechanical properties, and permeability, and promoted RPE cells maturation to form a polarized and functional monolayer with tight junctions. Poly (trimethylene carbonate (PTMC) (Sorkio et al., 2017), poly (l-lactic acid) (PLLA) (Thomson et al., 2011), poly (glycerol-sebacate) (PGS) (Neeley et al., 2008), and poly (methyl methacrylate) (PMMA) (Tao et al., 2007) are other FDA approved polymers being used for Bruch’s membrane mimetic substrates with the aim of RPE cells culture and transplantation.

Although synthetic biomaterials are greatly advantageous for ocular tissue engineering applications due to their modifiable properties, reproducibility, good mechanical features, and longer shelf life, there are certain disadvantages associated with these materials, such as lack of cell adhesion on the surface that reflects a need for chemical modification, and immune response or inflammatory reactions induction (Rajendran Nair et al., 2021a; Reddy et al., 2021; Rohiwal et al., 2021).

In a prospective, interventional, FDA–cleared phase I/IIa clinical trial (NCT02590692), Kashani et al. (2018) developed a composite subretinal implant, named the California Project to Cure Blindness–Retinal Pigment Epithelium 1 (CPCB-RPE1), that is, being injected to advanced dry AMD patients. This composite is made of a polarized monolayer of hESC-RPE on an ultrathin perylene substrate that mimics the Bruch’s membrane. In 2018, 16 patients had been enrolled in this study. OCT imaging results have shown hESC-RPE and photoreceptor anatomic integration. Implanted eyes showed 17-letter improvement in visual acuity and improved functional activity without vision loss progression. Results support this implant as a potential treatment for patients with severe vision loss from dry AMD (Kashani et al., 2018). In a further study, this group explored in detail the intraoperative surgical procedures and anatomic results (Kashani et al., 2020).