Molecular Vision 2022; 28:412-431
Received 09 November 2021 | Accepted 10 December 2022 | Published 12 December 2022
1National Institute of Genetic Engineering and Biotechnology, Tehran, Iran; 2Ophthalmic Research Center, Research Institute for Ophthalmology and Vision Science, Shahid Beheshti University of Medical Sciences, Tehran, Iran; 3Blood Transfusion Research Center High Institute for Research and Education in Transfusion Medicine, Tehran, Iran; 4Molecular Biomarkers Nano-imaging Laboratory, Brigham & Women's Hospital, Department of Radiology, Harvard Medical School, Boston, MA; 5Gynecologist and obstetrician, Aban hospital, Tehran, Iran
Correspondence to: Zahra-Soheila Soheili, Ministry of Science, Research and Technology, National Institute of Genetic Engineering and Biotechnology, P.O. Box: 14965/161, Pajoohesh Boulevard, 17th Kilometers, Tehran-Karaj Highway, Tehran-Iran, Postal Code: 1497716316; Phone: +98-21-44787379; FAX: +98-21-44787399; email: email@example.com.
Purpose: The development of biomaterials provides potent promise for the regeneration of neuroretinal cells in degenerative eye diseases and retinal tissue engineering. Biomimetic three-dimensional (3D) microenvironments and specific growth factors motivate the differentiation of human retinal pigment epithelial (hRPE) cells toward a retinal neural lineage. In this study, we evaluated alginate/gelatin (A/G) as a substrate for the culture of hRPE cells.
Methods: hRPE cells were isolated from neonatal human cadaver globes and cultivated on A/G substrate under different culture conditions, including 30% human amniotic fluid (HAF), 10% fetal bovine serum (FBS), and serum-free Dulbecco’s modified Eagle’s medium/nutrient mixture F-12 (DMEM/F12). The proliferation of cells in different culture conditions was determined using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide and a cell proliferation assay. Immunocytochemistry and real-time PCR were performed to evaluate the effect of the substrate on hRPE cell differentiation.
Results: A significant increase in the cell proliferation rate was observed in hRPE cells cultivated on an A/G substrate. Continuous observations demonstrated that hRPE cells formed densely packed, suspended spheroids in DMEM/F12 culture conditions, with dominant transdifferentiation into amacrine cells. Small adherent clusters of hRPE cells in HAF- and FBS-treated cultures represented dedifferentiation toward retinal progenitor cells. These cultures generated amacrine, rod photoreceptors, and bipolar cells.
Conclusions: These findings indicated that A/G substrate induced neural retinal cell propagation in cultures and would therefore be promising for RPE-based tissue engineering studies.
Retinal degenerative disorders, such as age-related macular degeneration (AMD) and retinitis pigmentosa (RP), are the leading causes of untreatable blindness worldwide. Several strategies are under investigation for the treatment and repair of the aforementioned disorders. Among these strategies, retinal cell transplantation is one of the most promising therapeutic approaches . A wide range of cells, including embryonic stem cells, adult stem cells, induced pluripotent stem cells, retinal progenitor cells, and RPE cells, have been recruited for retinal regeneration studies [2-6]. The RPE forms a monolayer of highly specialized pigmented cells located at the interface between the photoreceptors of the neural retina and the vascular choroid. Due to the vital role of the RPE in normal retinal structure and function, RPE-based regenerative therapies have been extensively investigated in previous studies [7,8].
AMD was the first target disease for PSC-derived RPE replacement cell therapy. Ongoing phase I clinical trials confirmed the safety and some aspects of the efficacy of these potential therapeutic approaches .
However, there are several challenges associated with retinal cell transplantation. Growth factors and scaffolds provide promising opportunities to improve RPE-based therapies. Scaffolds can address several issues in cell transplantation strategies, such as ensuring precise cell delivery, enhanced cell survival, reduced cell death, robust cell integration, and controlled cell differentiation . Human amniotic fluid (HAF) is a natural source of various growth factors, such as epidermal growth factor (EGF), fibroblast growth factor (FGF), nerve growth factor (NGF), hepatocyte growth factor (HGF), transforming growth factor alpha (TGF-a), transforming growth factor beta (TGF-b1), insulin-like growth factor-I (IGF-I), insulin-like growth factor-II (IGF-II), erythropoietin (EPO), granulocyte colony stimulating factor (G-CSF), and macrophage colony stimulating factor (M-CSF). These factors represent a complex physiologic and protective liquid that surrounds the fetus. Amniotic fluid possesses broad immunosuppressive properties and has been proposed for the treatment of chronic inflammation and immune responses . Our previous studies demonstrated that HAF affects the dedifferentiation and transdifferentiation of human retinal pigment epithelial (hRPE) cells to neural retinal cells [12-14].
Tissue engineering is a rapidly expanding interdisciplinary field that combines biology, medicine, and engineering. It uses
a combination of cells, biologic, or synthetic materials
To more precisely study developmental biology, scientists have designed 3D cell culture matrices. In tissue engineering therapy, scaffolds are used as 3D matrices that induce cell adhesion, proliferation, and differentiation. These scaffolds should have high biocompatibility and should not cause serious systemic or immunogenic issues during in vivo degradation and absorption. Scaffold-guided tissue engineering is used to create functional and transplantable tissue constructs. To date, a large library of synthetic and natural scaffold materials has been investigated for tissue engineering applications . Alginate is composed of a family of natural linear anionic polysaccharides derived from the cell walls of marine brown algae. Gelatin is a purified protein that emerges after the thermal, chemical, or physical denaturation of collagen. Due to gelatin’s valuable biologic origin, commercial availability, low cost, excellent biocompatibility, nonimmunogenicity, and appropriate biodegradability, it has been successfully employed in various biomedical applications .
Previous studies have shown that the negative charges of alginate reduce the proliferation and adhesion of RPE cells on purified alginate films . Alginate has been applied in combination with gelatin, chitosan, elastin, agarose, or hyaluronic acid [20, 21]. This increases its biologic activity and improves its properties, ensuring cell proliferation and differentiation in vitro. Furthermore, its mechanical properties imitate native tissues [22-24]. Gelatin is a type of collagen of animal origin  that is biocompatible, with no induction of immune responses in the body and an absence of toxicity. Therefore, it is used in the food industry, pharmacy, wound healing, drug delivery, and gene therapy. It is frequently used in cell cultures due to its bioactive sequences, which enable a suitable microenvironment for cell adhesion, migration, proliferation, and differentiation [23, 26]. However, gelatin is soluble at 30–40 °C and passes from a gel state to a solution, limiting its application in transplantation. To prolong its degradation time and increase its water resistance, it is necessary to apply a cross-linking procedure to improve the half-life of the meshwork .
Cells adequately interact with the ECM in a 3D state on alginate–gelatin hydrogel scaffolds, and cocultures of multiple cell types mimic the microenvironment. The biocompatibility, biodegradability, and nontoxicity of alginate–gelatin hydrogels depend on their interactions and tissue or body fluids. The low interfacial free energy of the hydrophilic surface leads to a low tendency for proteins and cells to adhere to this surface, which makes it satisfactorily biocompatible [28,29]. Flexibility and soft structures enable this hydrated polymeric biomaterial to be used in biomedical fields as an ECM substitute in various forms, such as hydrogels, microcapsules, microspheres, fibers, foams, and sponges [30, 31]. Porosity is an important aspect, since substantial surface areas promote the attachment and growth of cells. Further, pores and channels enable the transport of gases, nutrients, and waste . Alginate-hydrogel-based cell constructs are promising candidates for tissue engineering and regenerative medicine. Therefore, clinical applications are expected in the near future.
Previously, we reported an improved growth rate of hRPE cells on alginate/gelatin (A/G) scaffolds at a ratio of 20:80 . The purpose of the present study was to evaluate the viability, proliferation, and death of hRPE cells cultured on an A/G film and to survey the effect of the substrate on the de- and transdifferentiation of hRPE cells under different culture conditions (30% HAF; 10% fetal bovine serum (FBS), and serum-free Dulbecco's modified Eagle’s medium/nutrient mixture F-12 (DMEM/F12).
Human cadaver eyes from individuals younger than 2 years old were individually obtained from the Central Eye Bank of Iran and transferred to the National Institute of Genetic Engineering and Biotechnology (NIGEB). The Ethics Committee of the Ophthalmic Research Center approved the use of human globes for this study. The hRPE cells were isolated using enzymatic methods within the first 24 h after death. The eyes were transferred to PBS (1X; 120 mM NaCl, 20 mM KCl, 10 mM NaPO4, 5 mM KPO4, pH 7.4) in 10 cm2 cell culture plates, and the extraocular tissues were completely removed. Each globe was cut along the ora serrata, and the anterior segment was discarded. After making four radial cuts from the edge of the eyecup, the neural retina was peeled off. The exposed RPE layer was washed with PBS, and the dark-brown pigmented RPE monolayer was completely dissected from the surrounding tissue and cut into small pieces. Then, isolated tissues were incubated in 1 U/ml dispase I solution (Roche, Mannheim, Germany) at 37 °C in a humidified atmosphere of 10% CO2 for 50 min. The cell suspension was subsequently centrifuged at 300 × g for 5 min. After the supernatant was removed, isolated hRPE cells were plated in 25 cm2 cell culture flasks (Nunc, Roskilde, Denmark). Primary cultures at 10% confluency were incubated at 37 °C in a humidified 5% CO2 incubator until they became approximately 90% confluent (1 × 106 cells in T25 flask). The cultures were re-fed every 3–4 days with fresh DMEM/F12 media supplemented with 10% FBS. Subconfluent cell cultures were subcultured at a 1:3 ratio (3× 105 cells in T25 flask) using 0.25% trypsin-1 mM EDTA (T/E) solution.
Standard immunostaining was performed using rabbit polyclonal anti-human RPE65 antibody (Santa Cruz Biotechnology, sc-32893) to verify the identity of the isolated cells according to the manufacturer’s protocol. Negative controls (secondary antibodies only) and background controls (no antibodies) were also included in the experiments (Appendix 1). Three globes (donors aged between 2 months and 2 years old) were included to provide variety in the experimental cultures.
To prepare a fitting A/G substrate, a sodium alginate solution (2%; Aldrich, Munich, Germany) was prepared in distilled water at room temperature. An aqueous solution of 8% (w/v) gelatin (Sigma, Munich, Germany) in distilled water was prepared by stirring at 50 °C. A 20:80 weight ratio of A/G blend was prepared by mixing alginate and gelatin stock solutions. To prepare an A/G substrate of 1 mm thickness, 320 µl of A/G blend was transferred into each well of a polystyrene 12-well microplate and chilled at −70 °C for 20 min. The resulting A/G scaffold was crosslinked by adding 500 µl 2% (w/v) CaCl2-2H2O (pH 7.2; Merck, Darmstadt, Germany) and incubating for 1 h at room temperature. Subsequently, the substrate was washed with sterilized water to remove free calcium ions and then equilibrated by the addition of 500 µl DMEM/F12 (Sigma, Munich, Germany) media, followed by incubation for 4 h in a cell culture incubator.
Amniotic fluid samples were obtained from 30 pregnant women who underwent amniocentesis between the 14th and 16th weeks of pregnancy for the assessment of genetic deficiencies. Amniotic fluid cells were removed for karyotype analysis by centrifugation at 300 × g and 4 °C for 5 min. The resulting supernatants were sterilized using a 0.2 µm membrane filter (Orange Scientific, Brussels, Belgium) and stored at −70 °C until analysis. Cell-free supernatants, in cases with no evidence of chromosomal abnormalities, were pooled and used in downstream experimental procedures, including cell viability, cell death, cell proliferation, quantitative real-time PCR, and immunostaining assays. The procedure used to collect the samples was approved by the Ophthalmic Research Center (ORC).
A previous study demonstrated that 30% HAF, compared to 20% HAF and 10% HAF, effectively stimulated the proliferation and differentiation of hRPE cells into retinal neurons . Cultures from Passage 4 were used (1 × 105 cells/well) to seed cells on A/G substrate in 12-well microplates (Nunc). DMEM/F12 culture media supplemented with either 30% HAF or 10% FBS (Gibco, Karlsruhe, Germany) were added (1 ml) to nourish the cells for 7 days. Cultures were continuously examined and imaged using phase-contrast microscopy (Axiophot Zeiss, Germany), and morphological changes were reported.
Cells from Passage 4 were seeded (1 × 105 cells/well) on A/G substrate in 12-well microplates (Nunc), and 1 ml DMEM/F12 culture media was added and supplemented with either 30% HAF or 10% FBS. Cells cultivated on polystyrene substrates were used as controls. Cultures were incubated at 37 °C in a humidified 5% CO2 incubator. After 2 days, 100 µl of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; 5 mg/ml; Sigma) was added to each culture vessel and incubated for an additional 16 h. When violet crystals developed, 1 ml of SDS-HCl (10% SDS/1% 1 N HCl; Merck) was added, and the cultures were incubated for another 18 h. Lastly, 200 µl from the supernatant of the vessels was transferred to each well of a 96-well microplate, and absorbance was measured at 580 nm with reference to 692 nm (Titertek Multiscan ELISA reader, Labsystems Multiskan, Roden, Netherlands) . Three independent experiments were performed with at least three replicates each.
Cell proliferation and cell death enzyme-linked immunosorbent assays (ELISAs) were performed to evaluate the effect of the A/G substrate on the proliferation or unfavorable apoptosis of hRPE cells in culture. As mentioned previously, A/G substrates were placed in each well of a 96-well microplate. At Passage 4, cells were seeded (1 × 104 cells) on A/G substrate in the presence of 200 µL DMEM/F12 culture media without serum or supplemented with either 30% HAF or 10% FBS and incubated at 37 °C in a 5% CO2 incubator for 2 days. Cultures on polystyrene substrates that underwent similar treatments were assessed as controls. Cell proliferation and cell death assays were performed according to the manufacturer’s instructions (cell proliferation ELISA, BrdU colorimetric, and cell death detection ELISA kits, Roche, Grenzach-Wyhlen, Germany). The results are presented as the means of three independent experiments performed at least in triplicate.
Using QIAzol (Qiagen, Hilden, Germany) RNA extraction reagent, total RNA was isolated from hRPE cells (1 × 105 cells/well in a six-well microplate) that had been cultured on A/G substrate for 7 days in medium supplemented with 30% HAF, 10% FBS, or serum-free conditions. Subsequently, first-strand cDNA was synthesized using a cDNA synthesis kit (Qiagen) according to the manufacturer’s instructions. Real-time PCR was performed using SYBR Green master mix (Roche) to quantitatively assess RNA expression levels of the PAX6, SOX2, VSX-2 (visual system homeobox 2), cellular retinoic acid-binding protein (CRABPI), rhodopsin (RHO), protein kinase C(PKCα), Thy-1, and Nestin genes. As an internal control, the housekeeping gene glyceraldehyde phosphate dehydrogenase (GAPDH) was used to normalize the gene expression data. Specific primers were used from the QuantiTect Primer Assays – QIAGEN; Table 1).
Real-time PCR experimental conditions were as follows: an initial activation step at 95 °C for 10 min followed by 40 cycles of amplification that consisted of a denaturation step at 95 °C for 10 s, an annealing step (35 s at 60 °C), and finally, an extension step (15 s at 72 °C). Amplification efficiency was estimated by plotting a standard curve using appropriate serial dilutions of cDNA samples for the genes of interest and housekeeping genes. Relative gene expression was calculated using Bio-Rad software (Rel Quant UpDate for relative quantification) according to the 2-ΔΔCt method based on threshold cycle (Ct) values. To evaluate the expression of genes of interest, real-time PCR was performed for three series of treated and control cDNA samples (corresponding cultures that had been developed on polystyrene) collected from distinct experiments. Each sample was assessed at least in triplicate.
Cells from Passage 5 were cultured at 1×105 hRPE cells/well on the A/G substrate and treated with 30% HAF, 10% FBS, or serum-free media. After 7 days, the cultures were harvested and reattached to glass coverslips using cell spin cytocentrifugation for 5 min at 160 × g. Standard immunostaining protocols were performed according to the manufacturer’s instructions (Santa Cruz Biotechnology). Each coverslip was individually transferred to a well in a 6-well microplate, washed three times with PBS, and fixed in ice-cold methanol (−20 °C; Merck) for 5 min at room temperature (RT). Air-dried coverslips were blocked using 1% BSA (BSA, Merck) in 1% PBST (Triton X-100, Sigma, in PBS) for 1 h at RT and incubated overnight at 4 °C with specific primary antibodies at a dilution of 1:50 in 1.5% (w/v) BSA in 1% PBST (v/v). Goat polyclonal anti-human SOX2 (Santa Cruz Biotechnology, sc-17320), rabbit polyclonal anti-human Nestin (Santa Cruz Biotechnology, sc-20978), goat polyclonal anti-human CRABPI (Santa Cruz Biotechnology, sc-10061), rabbit polyclonal anti–human PKCα (Santa Cruz Biotechnology, sc-10800), mouse polyclonal anti-human rhodopsin (Santa Cruz Biotechnology, sc-53991), mouse polyclonal anti-human Thy-1 (Santa Cruz Biotechnology, sc-59396), goat polyclonal anti-human PAX6 (Santa Cruz Biotechnology, sc-53108), and goat polyclonal anti-human VSX2 (Santa Cruz Biotechnology, sc-21690) antibodies were applied on different coverslips. The coverslips were then thoroughly washed with PBS to reduce nonspecific antibody binding. Slides were incubated with FITC–conjugated goat anti-rabbit immunoglobulin (Santa Cruz Biotechnology, sc-2012), goat anti–mouse immunoglobulin (Santa Cruz Biotechnology, sc-2010), or donkey anti-goat immunoglobulin (Santa Cruz Biotechnology, sc-3853) at a dilution of 1:100 in 1.5% BSA in 1% PBST for 45 min at RT in the dark. Cell nuclei were counterstained with 4ʹ,6-diamidino-2-phenylindole dihydrochloride (DAPI; Santa Cruz Biotechnology) for 10 min, and coverslips were mounted on slides using a fluorescence mounting medium (glycerol 90%, PBS 10%, and phenyl diamine 10% [w/v]). Slides were imaged using an Axiophot Zeiss fluorescence microscope equipped with a 460-nm filter and a 520-nm filter for DAPI- and FITC-labeled antibodies, respectively. Negative controls (secondary antibodies only) and background controls (no antibodies) were also included in all experiments (Appendix 2). Immunostaining tests were performed in three independent experiments.
A two-tailed Student t test (Microsoft Excel 2010) was used to evaluate the statistical significance of the data, and p values less than 0.05 were considered statistically significant.
As they reached confluence, individual hRPE cells exhibited a variety of morphologies in the culture. Throughout the first 24 h to 48 h in the primary culture, the cells adhered to the flask surface and exhibited a star-like morphology with highly pigmented cytoplasm (Figure 1A).
After 3 to 4 days, cell numbers gradually increased, and a monolayer of elongated fusiform cells formed (Figure 1B). At full confluence, the cells became hexagonal and formed small, tightly packed cobblestone structures with distinct phase-bright borders (Figure 1C). During consecutive passages, the cells gradually lost their melanin granules. By Passage 5, cultures had a high proliferation rate, but eventually, cell proliferation and growth rates reduced, at which time cellular morphology changed from an elongated fusiform shape to a flattened amorphous appearance shape (data not shown). Newly developed cultures were examined by immunostaining for RPE65, which confirmed greater than 90% purity for the isolated cells and demonstrated that the cultures were RPE driven (Figure 1D).
The hRPE cells exhibited the same morphology under treatment with FBS and DMEM/F12 on polystyrene. They displayed a flattened, spindle-shaped phenotype in the control uncoated wells during the inspection of the cultures (Figure 2).
HAF treatment induced hRPE cells to form spheroids when they were cultured on polystyrene. After 24 h and 48 h, the hRPE cells displayed a flattened, spindle-shaped phenotype on polystyrene (Figure 2). After 3 days, the cells began to aggregate and formed sphere-like structures that tended to detach from the surface of the plate and float into the medium. The spheroids maintained their morphology over the next 4 days (Figure 2 and Table 2).
The hRPE spheroids in FBS- and HAF-treated cultures were entrapped within the A/G substrate. Individual cells and small nodular spheroids were observed on A/G substrates. The cells did not attach to the substrate on the first day. Rather, they adhered to the A/G substrate 2 to 7 days after cultivation (Figure 3). Inspection of the cultures over consecutive days revealed that single cells and spheroids grew larger, their numbers increased, and dense spheroids became entrapped within the A/G substrate.
Suspended dense spheroids developed on A/G substrates in DMEM/F12-treated cultures (Figure 3). When the cultures were maintained in DMEM/F12, single cells and small nodular spheroids were detected on the A/G substrate during the first 24 h. The spheroid structures did not engage with the A/G substrate during the 7 days of consecutive inspections. The high rate of cell proliferation led to the formation of sizable cellular clumps in the cultures, which tended to aggregate and form densely packed, rounded spheroids (Table 2).
According to the colorimetric MTT assay, no significant difference was observed in the viability of hRPE cells between media supplemented with FBS (p = 0.11), or DMEM/F12 (p = 0.8), regardless of the substrate type. Cellular viability was increased in the HAF-treated cultures compared to the control (p = 0.04; Figure 4A-B).
According to the cell proliferation ELISA, proliferation rates increased by approximately twofold when cells were cultivated on A/G substrates, irrespective of the type of nutrient media used (HAF, p = 0.0009; FBS, p = 0.002; DMEM/F12, p = 0.008). The highest rate of hRPE cell proliferation was observed in the HAF-treated cultures, presumably due to their considerable growth factor content (Figure 4C).
The cell death ELISA demonstrated that the A/G substrate did not have a significant impact on the apoptotic rate of the hRPE cell culture when compared to the polystyrene substrate (FBS, p = 0.61; HAF, p = 0.10; DMEM/F12, p = 0.55). Cell death was significantly lower than that of the positive control included in the kit (Figure 4D).
According to RT–PCR data and immunostaining, the A/G substrate increased the expression levels of neural retinal progenitor (PAX6), neural stem (SOX2), late retinal progenitor (VSX2), neural progenitor cell (Nestin), amacrine (CRABPΙ), rod photoreceptor (RHO), bipolar (PKCα), and retinal ganglion (Thy-1) cell markers compared to the polystyrene substrate.
Compared to corresponding cultures on polystyrene, the A/G substrate, did not induce SOX2 gene expression in DMEM/F12-treated cultures, whereas considerable amounts of SOX2 gene expression were detected in HAF-treated cultures (382-fold, p = 0.00002) and FBS- (43-fold, p = 0.0001) (Figure 5). SOX2 protein expression was detected in HAF-, FBS-, and DMEM/F12-treated cultures (Figure 6). Although SOX2 protein expression is generally restricted to the nucleus, cytoplasmic staining was detected in this study.
An approximately twofold increase in Nestin gene expression was observed in HAF- or DMEM/F12-treated cultures on A/G compared to polystyrene. A 1.3-fold increase was detected in FBS-treated cultures (Figure 5). Immunostaining of hRPE cells cultured in HAF-, FBS-, or DMEM/F12-containing media showed that the positive cells expressed neural progenitor cell markers (Figure 6).
Assessment of A/G cultures with respect to cultures on polystyrene revealed that PAX6 gene expression levels in HAF-, FBS-, and DMEM/F12-treated cultures were enhanced 695-, 177-, and 12-fold, respectively. Relative expression of PAX6 in FBS-treated cultures was significantly higher compared to DMEM/F12- (p = 0.0002) and HAF- (p = 0.00015)-treated cells (Figure 5). Immunostaining of hRPE cells cultured in media containing HAF, FBS, and DMEM/F12 showed that the cultures were positive for PAX6, a neural retinal progenitor cell marker (Figure 7).
Lastly, VSX-2 gene expression in HAF-, FBS-, and DMEM/F12-treated cultures was enhanced 303-, 259-, and 16-fold in A/G cultures compared to controls plated on polystyrene. The expression levels of the VSX-2 gene in hRPE cells treated with FBS (p = 0.00029) and HAF (p = 0.00024) were much higher than those in cells treated with DMEM/F12 (Figure 5). Immunostaining results for VSX-2 revealed that hRPE cells expressed late retinal progenitor cell markers in HAF-, FBS-, and DMEM/F12-containing media (Figure 7).
The hRPE cells in A/G demonstrated a fourfold increase in Thy-1 gene expression under FBS treatment compared to cells grown on polystyrene (Figure 5). Immunostaining results revealed the presence of Thy-1 ganglion cell markers in DMEM/F12-treated cultures (Figure 8).
The expression levels of PKCα were increased fourfold in HAF-treated cultures on A/G compared to the polystyrene substrate (Figure 5). Immunostaining results for PKCα protein showed that the cells expressed bipolar cell markers in all three culture conditions (Figure 8).
The same levels of RHO gene expression were observed in HAF- and FBS-treated cultures, which were approximately 50-fold higher than for DMEM/F12 treatment (Figure 5). Immunostaining for rhodopsin demonstrated its expression in all culture conditions (Figure 9).
In A/G cultures, compared to polystyrene, CRABPΙ gene expression under FBS-, HAF-, and DMEM/F12-treatments was enhanced 177-, 18-, and 76-fold, respectively (Figure 5). A large number of cells were positively labeled with the CRABPI antibody in the examined culture conditions (Figure 9). Schematic presentations of hRPE cell gene expression are presented in Figure 10 .
Since the cells mostly formed spheroids on the A/G substrate and in DMEM free of serum, accurately counting positive cells expressing individual markers was not possible. Therefore, positive cells for the investigated markers are represented qualitatively, not quantitatively.
Herein, we evaluated the competence of the A/G substrate for hRPE cell cultures under treatment with HAF- or FBS-containing media or in DMEM/F12 growth factor-free media. We tested the hypothesis that the proposed culture conditions would enhance retinal proliferation and differentiation with respect to traditional cultures on polystyrene. The hRPE cells were routinely cultivated in DMEM/F12 supplemented with 10% FBS and displayed a flattened, spindle-shaped phenotype on polystyrene during early passages (Figure 1B). Liao and Stanzels showed that RPE cells displayed hexagonal morphology in fully confluent cultures [35, 36]. We also observed this characteristic morphology in fully confluent cultures (Figure 1C). The hRPE cells exhibited the same morphology in FBS-containing or FBS-free DMEM/F12 media when cultured on polystyrene (Figure 2). The cells exhibited normal morphology during the first 48 h in the HAF-treated culture (Figure 2). However, by day 3 in culture, they joined together and formed sphere-like structures. The spheroids detached from the surface of the plate, floated into the medium, and maintained their spheroid morphology throughout subsequent days (Figure 2). Similar to the HAF culture conditions, hRPE cells exhibited a 3D spheroid morphology on the A/G substrate (Figure 3). Several reports have indicated that spheroid formation is a desirable characteristic for applications in biomedical fields. Chen et al. demonstrated that human osteoblast cells that formed 3D clusters on Ca-alginate scaffolds provided an invaluable environment for bone-like tissue formation . A study also showed that a 3D culture efficiently promoted osteogenic differentiation of mesenchymal stem cells (MSCs) . It seems that spheroid formation by hRPE cells on A/G substrates establishes a potential carrier for retinal cell transplantation and retinal-like tissue formation. Spheroid formation is the first phase of cell reorganization during the fabrication of an engineered organ .
Studies have reported two types of cell spheroid cultures due to spontaneous interactions between cells (cell–cell) and substrates (cell–ECM). In addition to the role of growth factors in guiding cellular differentiation, signals derived from cell–cell or cell–ECM interactions also play an important role in neural cell differentiation. These messages can be mechanical, physical and/or chemical stimuli, which are received through contact with surrounding cells or the ECM, and drive differentiating cells toward particular cell lineages .
A/G substrate promoted cell–ECM interactions under HAF or FBS treatments. However, DMEM/F12 led to hRPE cells’ interactions and finally cellular aggregates formed large suspended spheroids in cultures. In a DMEM/F12-treated culture, the A/G substrate stimulated hRPE cells to aggregate into large suspended spheroids that established cell–cell interactions (Figure 3).
MSCs exhibited adhered spheroids on alginate hydrogel similar to HAF- and FBS-treated hRPE cultures . However, a human breast cancer cell line (MCF-7) formed suspended nonadherent spheroids on silicon substrate , similar to DMEM/F12-treated hRPE cell cultures. Under such conditions, the cells did not exhibit anchorage-dependent growth, and the size of the multicellular spheroids was dependent on the number of seeded cells [42,43].
The MTT assay results demonstrated that the A/G substrate promoted hRPE cell survival. The cells exhibited 98% viability at the end of the culture period in all experimental conditions (Figure 4A–B). Cell proliferation ELISA results revealed an increased rate of proliferation on the A/G substrate compared to polystyrene in all three culture conditions (twofold increase; Figure 4C). Spheroid formation on the A/G substrate promoted hRPE proliferation by inducing cell–cell or cell–ECM interactions. Previous studies have confirmed that the alginate substrate supports the growth of RPE cell lines, bone marrow cells, NB2a neuroblastoma cells, and H9C2 cardiomyoblasts [44-47].
Cell death ELISA demonstrated that A/G substrates did not induce significant cell death in hRPE cell cultures (Figure 4D), indicating that the A/G substrate allows the cells to absorb sufficient amounts of oxygen and nutrients. Previous studies have shown that encapsulated hRPE cells in alginate hydrogels do not display significant levels of apoptosis .
The highest proliferation rate was observed in HAF-treated cultures, which may be due to HAF being composed of several growth factors, e.g., transforming growth factor beta (TGF-β), complement C3, and plasminogen, which play important roles in inducing cell proliferation. The role of HAF in inducing cell proliferation in cultured hRPE cells, hematopoietic stem cells, lens cells, and human skin fibroblast cells has been previously demonstrated [12,49-51].
As cells differentiate, the proliferation rate usually decreases, and cells arrest in G0 phase of the cell cycle . Undifferentiated retinal progenitor cell (RPC) cultures are typically characterized by their ability to proliferate in vitro [53, 54]. Our data revealed that HAF-treated cultures harbored RPCs with a high proliferation rate, while FBS- and DMEM/F12-treated cultures included neural differentiated cells with a low proliferation rate.
Real-time PCR and immunostaining data confirmed that SOX2 was expressed under all examined conditions (Figures 5–6). Dominant SOX2 expression was detected in the HAF-treated cultures (Figure 5). Substrate-induced cell–ECM interactions and growth factors increased the expression of neural stem cell markers. Similarly, Gao et al. reported the differentiation of multipotent human adipose-derived stromal/stem cells into neural stem cells on chitosan and gelatin 3D scaffolds . SOX2 is usually expressed in the nuclei of undifferentiated and proliferating cells. We identified cytoplasmic expression of the SOX2 protein in the immunostaining experiments. Previous studies have also reported SOX2 cytoplasmic localization [56-59].
In 2009, Baltus et al. suggested that SOX2 moves from the nucleus to the cytoplasm in response to differentiation signals . In neuronal cells, cytoplasmic localization of SOX2 has also been observed . Cytoplasmic localization of SOX2 in differentiated neurons seems to result from the process of differentiation-induced proteolytic degradation . By contrast, nuclear SOX2 expression is lost upon differentiation .
In 2018, Okolicsanyi et al. indicated that SOX2 localization was distinctly cytoplasmic in MSC-induced neurosphere cultures, with a more neural progenitor-like population than the nuclear localization observed in undifferentiated human MSC and neural stem cell (NSC) cultures .
The same level of Nestin gene and protein expression was observed in all three culture conditions (Figures 5–6). Cell–ECM interactions, such as cell–cell interactions, played a similar role in the expression of neural progenitor cell markers. The effect of cellulosic hydrogels on the differentiation of MSCs into neural progenitor cells has been previously confirmed .
Remarkable expression of PAX6 at the mRNA transcript and protein levels was detected in all culture conditions. The highest level of PAX6 expression was observed in FBS-treated cultures (Figures 5 and 7). This suggests that cell–ECM interactions induced the appearance of neural retinal progenitor cells (NRPCs) under the corresponding conditions. NRPCs can differentiate into glial cells, various types of photoreceptors, and neurons . Jasty et al. (2014) isolated stem cells from the ciliary pigment epithelium and cultivated them on nanofiber scaffolds. They demonstrated that stem cells became entrapped and formed colony-like clusters in the scaffolds. The adjoining spheroids expressed neural retinal progenitor markers, such as PAX6, and neural progenitor cell markers, such as Nestin .
A/G substrate induced remarkable increases in VSX-2 expression in FBS-, HAF-, and DMEM/F12-treated cultures (Figure 5 and Figure 7). Substrates played an important role in the expression of late retinal progenitor cell markers by inducing cell–ECM interactions. Late retinal progenitor cells tended to differentiate into bipolar cells, horizontal cells, amacrine cells, rod photoreceptors, and Müller glial cells. It has also been previously shown that alginate hydrogel encapsulation improves the differentiation of human pluripotent stem cells into neural retinal cells .
In DMEM/F12-containing media, a few cells in large suspended spheroids exhibited Thy-1 protein expression (Figure 8), whereas mRNA expression was not detected in those cultures.
Similarly, Schlamp et al. demonstrated that ganglion 452 cell death caused a decrease in Thy1 mRNA level, with no change in the number of Thy1-expressing cells. When mouse eyes were subjected to optic nerve crush, leading to induced ganglion cell death, Thy1 mRNA level decreased within 24 h and steadily declined to nearly undetectable level by 2 weeks; however, the number of Thy1-expressing cells did not decrease for 7 days . Only traces of Thy1 mRNA were detectable in N-Methyl-D-aspartate- and NMDA-injected eyes after 48 days, but Thy1 protein immunoreactivity slowly decreased over 8 days . Thy-1 protein was not detected under either FBS or HAF treatment.
It is possible that the substrate, which induced cell–cell interactions, assisted in the expression of retinal ganglion cell markers. Previous studies have demonstrated the effective role of ﬁbrin hydrogels and polybenzyl glutamate scaffolds in the differentiation of stem cells into retinal ganglion cells [69, 70].
PKCα gene expression increased fourfold in HAF-treated cultures on A/G substrates. Other treatments did not induce a remarkable increase in gene expression (Figure 5). PKCα protein expression was detected in all culture conditions (Figure 8). These findings indicate that the expression of bipolar cell markers requires an appropriate substrate to induce cell–ECM interactions in combination with growth factor supplements in culture. This finding is consistent with Tomita and colleagues’ report that polylactic-coglycolic acid/poly-L-lactide acid (PLGA/PLLA) substrates increased the differentiation rate of RPCs into bipolar cells .
This likely indicates that substrate-induced cell–ECM interactions were sufficient for the expression of rod photoreceptor cell markers. Steedman et al. cultivated RPCs on polycaprolactone (PCL) thin film scaffolds and demonstrated cell–cell communication in clusters that appeared on the substrate and induced the differentiation of RPCs into rod photoreceptor cells .
A considerable increase in the protein and transcript levels of CRABPΙ was detected in response to different treatments on the A/G substrate (Figure 5 and Figure 9). In 2017, human fetal-derived RPCs were transplanted into the eyes of rat models with retinal degeneration. Based on the results, a phase I clinical trial was performed to confirm the safety and tolerability of transplantation in eight patients with advanced RP.
For the first time Yong Liu, et.al  confirmed the long-term safety and feasibility of vision repair by stem cell therapy in retinitis pigmentosa patients. Their results provided applicable information for studies related to cell-based therapies for retinitis pigmentosa and other inherited retinal degenerative disorders.
This study indicates that the proliferation and differentiation of hRPE cells on the A/G substrate relies on cell–cell interactions, cell–ECM interactions, and growth factor supplementation. In the HAF-treated culture, hRPE cells formed small, adherent spheroids similar to FBS-treated cultures. Recruitment of growth factors predominantly induced differentiation of cells toward neural stem cells and late retinal progenitor cells with high proliferative potential. The highest proliferation rate of the cultures depended on the augmentation of HAF, which is a source of various nutrient growth factors.
Under routine culture conditions of DMEM/F12 media supplemented with FBS, hRPE cells exhibited small adherent spheroids on the A/G substrate. These cell–ECM interactions predominantly induced hRPE cells to differentiate into neural retinal progenitor cells and late retinal progenitor cells, as well as terminally differentiated neural retinal cells.
Large suspended spheroids of hRPE cells were formed in DMEM/F12 culture on A/G substrate. Cell–cell interactions predominantly induced hRPE cell differentiation toward amacrine cells. The low proliferation rate in DMEM/F12-treated cultures confirmed that the culture featured terminally differentiated cells. As evidenced in this study, the A/G substrate can be recruited to generate different lineages of retinal cells in retinal regenerative studies.
A/G substrate improved hRPE cell survival, proliferation rate, and biologic integration with the coordination of growth factor resources. It directed the de- and transdifferentiation of hRPE cells to the neural retinal lineage through signals derived from cell (cell–cell) and substrate (cell–ECM) interactions. Our findings propose an A/G substrate for tissue engineering of the retinal nexus in the context of cell replacement therapies and future retinal regenerative research.
This work was supported by Iran National Science Foundation through grant number of 97009298, National Institute for Medical Research Development (NIMAD) through grant number of 964465 and National Institute of Genetic Engineering and Biotechnology (NIGEB), Tehran, Iran through grant number of 715. We wish to acknowledge Meisam Lesani and Maliheh Davari for their contribution in completing this work. The manuscript was presented in the 9th annual meeting of the Iranian research association for vision and ophthalmology. Authorship: H.SHN. performed the experiments, participated in the design of the experiments and data analysis, and wrote the manuscript. Z-S.S. participated in the design of the experiments and data analysis, edited the manuscript, provided the budget and supervised the study. SH.S. performed the experiments. H.A. participated in the preparation and editing of the manuscript. E.RP. participated in the design of the experiments. M.H. participated in HAF preparation. Funding: This work was supported by Iran National Science Foundation through grant number of 97,009,298, National Institute for Medical Research Development (NIMAD) through grant number of 964,465 and National Institute of Genetic Engineering and Biotechnology (NIGEB), Tehran, Iran through grant number of 715. Data Availability Statement: The data that support the findings of this study are available from the corresponding author upon reasonable request.