Reporter-guided photoreceptor differentiation in 2D culture using PGP1-hIPSCs: A comparable alternative to 3D organoid systems?

Human induced pluripotent stem cells

The PGP1-hIPSC reporter line [19] was obtained from Robinson & Del-Tsonis (Miami University, Oxford, OH). This CRISPR/Cas9-engineered, triple transgenic line expresses fluorescent markers for early retinal cell types: VSX2-cerulean (neural progenitor), recoverin (RCVRN)–mCherry (pan-PhR), and BRN3b-GFP (retinal ganglion cell). This system enables real-time tracking of cell fate transitions from pluripotency through eyefield progenitor cells, neural retinal progenitor cells, and PhR commitment. Use of hIPSCs was approved by the Bond University Human Research Ethics Committee (BUHREC; protocol RO1740).

PGP1-hIPSCs at passage (p) 46 (experiment 1), p49 (experiment 2), and p60 (experiment 3) were cultured feeder-free on hESC-qualified Matrigel (BioStrategy, Macquarie Park, Australia) and maintained in mTeSR complete medium (StemCell Technologies, Rhodes, Australia) with daily media changes. Prior to differentiation, cells were transitioned to growth factor reduced Matrigel. Cells were seeded at ~150,000 cells/cm² and transitioned to primary differentiation medium (day [D] 1; Appendix 1) upon reaching ~90% confluence [20,21]. On D7, cells at ~70% to 80% confluence (~50,000 cells/cm²) were passaged with Collagenase IV (ThermoFisher Scientific, Scoresby, VIC, Australia). Rock inhibitor (10 μM) (Focus Bioscience, Tingalpa, QLD, Australia) was added 1 h before each passage to enhance yield, minimize stress, and support survival [22].

Given the postmitotic nature of PhR cells, passage timing was guided by cell morphology assessment and proliferation rates. Gentle Cell Dissociation Reagent (StemCell Technologies) was used for enzyme-free dissociation. As single-cell passage proved ineffective, cultures were passaged as clumps, limiting accurate cell counting during secondary differentiation.

Murine 661W PhR cells as control

The immortalized murine PhR 661W cell line [23,24], provided by Dr. Muayyad Al-Ubaidi (University of Oklahoma, Norman, OK) and obtained through Dr. Krisztina Valter-Kocsi (Australian National University, Canberra, Australia), was used as a control for immunocytochemistry (ICC), flow cytometry, and microscopy analysis. Cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium with HEPES (ThermoFisher Scientific), 10% fetal bovine serum (Cytiva, Macquarie Park, NSW, Australia), 0.5% Glutamax (ThermoFisher Scientific), and 1 µl/ml 2-Mercaptoethanol (Merck, Bayswater, VIC, Australia).

Cells were maintained on uncoated Costar tissue culture (TC)-treated plates and passaged at ~80% confluency using TrypLE (ThermoFisher Scientific). Detached and resuspended cells were centrifuged at 300g for 3 min, reseeded, and cultured at 37 °C, 5% CO₂, and 90% relative humidity.

Differentiation medium and supplementation

From D0, primary differentiation medium guided the cultures toward the eyefield and neural retina, with medium changes every 2 to 3 days. On D8, secondary differentiation medium was introduced to enhance neural induction and PhR lineage commitment. Supplements included triiodothyronine (T3, D9; Sigma-Aldrich, St. Louis, MO), DAPT (D19; Sigma-Aldrich), taurine (D23; Sigma-Aldrich), and retinoic acid (D33; Sigma-Aldrich) and were maintained until D66 and D111. The complete list of reagents is detailed in Appendix 1 and Appendix 2.

Immunocytochemistry

Cells were fixed in 4% paraformaldehyde (PFA) for 10 min, permeabilized in blocking buffer (95% DPBS^−/−, 0.3% goat serum, 0.3% Triton X-100) for 1 h, and incubated overnight at 4 °C with primary antibodies, 1:100 prepared in Antibody Dilution Buffer (ADB: DPBS; Sigma-Aldrich, supplemented with 10% bovine serum albumin and 0.3% Triton-X). Antibodies are listed in Appendix 3. Secondary antibodies prepared in ADB (1:1,000; Appendix 4) were applied for 1 h at room temperature. Nuclei were stained with 20 μM Hoechst 34,580 for 10 to 15 min. Coverslips were mounted using Dako (Agilent Technologies, Mulgrave, VIC, Australia) Fluorescence Mounting Medium.

Confocal and live-cell microscopy

Following immunocytochemical staining of cultured cells, samples were imaged using an Olympus FV3000 confocal microscope (Evident, Macquarie Park, NSW, Australia) with Fluoview FVS315-SW analysis software (version 2.3.2.169 Evident), a Nikon Eclipse Ti-DH fluorescent microscope (Nikon, Rhodes, NSW, Australia), or a Live Cell Nikon Ti-E Inverted microscope (Nikon DMi1–468798) with NIS-Elements Advanced Research Software (version 5.02.03; Nikon). Live cells were monitored throughout the differentiation process and imaged using an Olympus DP23M digital microscope (Evident) with CellSens Standard software (version 4.2; Evident). Images of comparative samples were acquired under related settings within each experiment and exported as raw OIR (Olympus) or ND2 (Nikon) files, respectively, and converted to TIFF or JPEG files after analysis. Images of comparative samples were acquired under similar settings within each experiment and exported as raw OIR (Olympus) or ND2 (Nikon) files, respectively. Images were processed using ImageJ software (version 1.4f; National Institutes of Health, Bethesda, MD, USA). Individual color channels were first separated and saved, after which they were recombined to generate the final TIFF composite image.

Flow cytometry

Cells were dissociated with TrypLE and processed in FACS buffer (DPBS^−/−, 2% bovine serum albumin, 0.01% sodium azide). After permeabilization in blocking buffer (30–60 min, on ice), cells were stained with primary antibodies (1:100, 60–90 min) and secondary antibodies (1:1,000, 30 min; Appendix 2 and Appendix 3). For experiment 2, D52 cultures were stained for RCVRN; 661W cells served as positive controls. Negative controls were incubated with secondary antibody only (goat anti-mouse IgG [H⁺L], Alexa Fluor 488; ThermoFisher Scientific). In experiment 3, D42 and D49 samples were stained for OPSIN-S, OPSIN-M/L, or RCVRN, with 661W and unstained samples as positive or negative controls, respectively. Data were acquired on a BD FACSAria Fusion and analyzed with FACSDiva v9.0 (Macquarie Park, NSW, Australia). Detailed flow cytometric gating strategies, including cell population selection, control samples, and marker-specific analyses, are shown in Appendix 5, Appendix 6 and Appendix 7.

RT-qPCR for the identification of development from hIPSC-PhR progenitors

RNA was extracted using the Quick-RNA Miniprep Kit (Zymo Research Integrated Sciences, Chatswood, NSW, Australia) per manufacturer instructions, with additional concentration and clean-up using RNA Clean & Concentrator-5 (Zymo Research) as needed. First-strand complementary DNA (cDNA) synthesis was performed using the iScript cDNA Synthesis Kit (Bio-Rad, Gladesville, NSW, Australia 1,708,890) with 200 ng input of RNA material.

RT-qPCR reactions were run using the SsoFast EvaGreen Supermix (Bio-Rad, 1,725,201) with 500 nM forward and reverse primers, at 95 °C for 30 s, 40 cycles of (1) denaturation at 95 °C for 5 s and (2) annealing and extension at 60 °C for 5 s with a final melt curve. Data analysis used Bio-Rad CFX Maestro v2.3 software, with fold changes calculated using the ΔΔCT method. GAPDH served as the reference gene [25], and fold change values were calculated based on maximum expression for each gene target. Gene targets were selected based on their role in the transition from pluripotency to PhR progenitor fate, consistent with previous in vitro studies. Primer sequences were obtained as predesigned KiCqStart primers (Merck KSPQ12012 or Integrated DNA Technologies, Coralville, IA), detailed in Appendix 8 with results presented in Appendix 9.

Topological analysis

Scanning electron microscopy

On D99 of experiment 3, cells were fixed in 3% glutaraldehyde, dehydrated in graded ethanol, rinsed with hexamethyldisilazane (Sigma-Aldrich), and air-dried overnight. Samples were mounted on stubs, gold-coated (JEOL Smart Coater), and imaged using a JCM-5000 NeoScope scanning electron microscope (Peabody, MA). Images were analyzed with NeoScope v1.2.3.

Summary of time course cell analysis

To monitor cell differentiation and development in culture, samples were taken at key developmental time points for analysis of marker expression using ICC, gene expression using RT-qPCR analysis, and subset marker analysis using flow cytometric analysis. Additionally, cells were pelleted and preserved for bulk RNA sequencing. For ICC, cells were collected at six time points (experiment 1: D0-IPSC, D6, D30, D46, D52, and D66). Cells were stained with specific antibodies followed by secondary fluoresceinated antibodies. Controls were exposed to secondary antibodies only. Control antibody staining is shown in Appendix 10. For gene expression analysis, cells were harvested at five time points (experiment 1: D0-IPSC, D6, D30, D40, and D59) for RT-qPCR analysis. GAPDH was used as a housekeeping reference gene [25]. Cells for bulk RNA sequencing were collected at five time points (experiment 1: D0-IPSC, D9, D30, D40, and D53).

To monitor cell differentiation and assess the reproducibility of the original protocol developed, samples were taken as close to the previous time points in differentiation as possible for ICC analysis, RT-qPCR, flow cytometric analysis, and collection and pelleting of cells for bulk RNA sequencing. For ICC, cells were collected at six time points for each subsequent experiment (experiment 2: D9, D31, D42, D52, D61, and D66; experiment 3: D9, D36, D43, D52, D66, and D111). For RT-qPCR, cells were harvested for analysis at five time points (experiment 2: D0-IPSC, D7, D30, D45, and D66; experiment 3: IPSC, D9, D30, D45, and D66). Flow cytometric analysis was performed at one time point, D52, for experiment 2 and two time points, D42 and D49, for experiment 3. Cells for bulk RNA sequencing were collected at five time points for experiment 2 (D0-IPSC, D9, D30, D45, and D66) and seven time points for experiment 3 (D0-IPSC, D7, D30, D45, D50, D66, and D111; time points analyzed and assays used are listed in Appendix 11, Appendix 12, and Appendix 13).

Transcriptome analysis

Raw sequencing reads for monolayer cultures were assessed for quality and adaptor or index contamination using FastQC, then aligned to the human reference genome (Hg38) using the Rsubread package with default parameters. Gene-level count matrices were generated by mapping aligned reads to the GENCODE comprehensive gene annotation (GRCh38.p14) using featureCounts. Count matrices for organoid cultures were downloaded from GSE119274. Raw counts were normalized using the trimmed mean of M-values method to account for library size differences. Genes with low expression, defined as fewer than 1 count per million in more than 70% of samples, were excluded from downstream analyses. Dimensionality reduction was first performed via principal component analysis (PCA) on gene-wise scaled expression values, followed by uniform manifold approximation and projection to facilitate visualization of global transcriptional patterns. Differential expression analysis was conducted using the limma–voom pipeline. Briefly, the voom function was used to estimate the mean–variance relationship of log-counts, which were then modeled using linear modeling (lmFit) and moderated with empirical Bayes (eBayes) shrinkage. Significantly differentially expressed genes were identified using topTable, adjusting for multiple testing. Gene set enrichment and pathway analyses were performed on differentially expressed gene lists to identify functional annotations and regulatory pathways.

PGP-1–reporter live fluorescence and morphological changes observed during differentiation

The PGP1-hIPSC line, engineered to express fluorescent reporters for retinal development markers, was used in this study. An experimental outline is illustrated in Figure 1A. This line, previously validated for differentiation into neural retina and PhR progenitors [26], enables real-time monitoring via live fluorescence, eliminating the need for cell sacrifice for endpoint assays. Daily culture assessments were conducted by tracking cerulean (VSX2) and mCherry (RCVRN) signals; GFP (BRN3B) was not detected and is not discussed.

The earliest stages in driving progenitors toward a PhR commitment involve the inhibition of NOTCH signaling. Figure 1B summarizes the small molecules used in the primary and secondary medium. Here, DAPT (γ-secretase inhibitor) is used as a NOTCH inhibitor [27] and was added to the medium on D19 to drive PhR commitment. Similarly, taurine is a vital signaling molecule for PhR development [28] and is prevalent across the entire retina. Endogenous taurine is a cysteine derivative that has been shown to induce neural development and maintenance and promote differentiation to PhR terminal fate [29,30]. Taurine deficiencies in rodents, cats, and primates reveal the inability of PhRs to mature [31]. T3, an activator of thyroid hormone receptors, was used from the onset of secondary differentiation (D9) as it is thought to influence and promote specifically cone fates [32]. When used in vivo and in vitro in organoid studies, T3 demonstrates a propensity to determine cone fates as well as diversity within cone subtypes [33]. Retinoic acid was added as a final supplement at later stages of secondary differentiation (D33). It regulates NRL expression and, therefore, rod fate [34]. Live morphological changes were observed throughout, with rosette formation and neurite outgrowth observed around D15 to D20 (Figure 1C). Cells were harvested upon fluorescence detection, with mCherry (RCVRN) expression first observed at D27 of the initial pilot experiment (Figure 1D).

Emergence of a PhR phenotype

Retinal progenitor cell development began with the emergence of eyefield markers characteristic of the anterior neural ectoderm, visualized via ICC (Figure 2A). At D6 of differentiation, cultures continued to express the pluripotency-associated markers, OCT4 and SOX2, as shown by nuclear localization in most cells. Concurrently, upregulation of early eyefield transcription factors RAX, PAX6, and LHX2 was observed, indicating the onset of retinal lineage specification in response to exogenous patterning cues present in the differentiation medium.

Following transfer to neural induction medium, cultures exhibited marked morphological changes, with sequential supplementation of T3 (D9), DAPT (D19), taurine (D23), and retinoic acid (D33). By D30, ICC revealed the presence of neural retina progenitor and proliferation marker CHX10 in the presence of NR2F2 (Figure 2A). By D46, coexpression of RCVRN and NR2F2 indicated the initiation of PhR lineage commitment, while CHX10 expression was notably reduced. RAX expression persisted alongside PRDM1, consistent with a transitional progenitor state and early PhR identity. By D66, cells exhibited markers of mature PhRs, including OPSIN-M/L and RCVRN, signifying further differentiation toward a committed PhR-like phenotype.

Gene expression changes from pluripotency to a PhR-like identity

Gene expression dynamics from pluripotency toward a PhR-like fate were assessed by RT-qPCR, normalized to GAPDH [25] and expressed as a percentage of each gene’s maximum expression (Figure 2B). The RT-qPCR plot shows biological trends; accordingly, the results are descriptive trends only. OCT4 declined steadily from D0, remaining low until a slight increase from D40. OTX2 is involved in early ocular genesis and in subpopulations of progenitor cells that primarily produce cones [35]. Here we report a steady expression until D30, after which we observed an apparent increase. The neural retina progenitor marker VSX2 increased to D6, followed by a transient decrease at D30, whereafter it resumed its increase. PRDM1 was undetectable until D6, increased sharply to D30, then declined gradually after D40. The PhR commitment marker CRX was noticeably upregulated as early as D6 with peak expression seen at D30, followed by a transient decline by D40 with a subsequent increase toward the final time point. NR2F2 and RCVRN followed similar trends, with low early expression and steady upregulation from D30 onward, indicating progression toward a PhR lineage. NR2F2 and RCVRN, both encoding PhR progenitor genes, showed an apparent increase in expression from D30.

Reproducible and efficient differentiation toward a PhR lineage

Reproducibility of PhR differentiation was confirmed in experiments 2 and 3 using stage-specific protein markers. From D9 of both experiments, as illustrated in the representative images in Figure 3A, SOX2, a key marker that supports retinal progenitor proliferation, was observed. Eyefield transcription markers PAX6, SIX3, CHX10, OTX2, LHX2, and RAX were also detected at this time point [36,37]. By D31 (experiment 2; Figure 3B), OTX2 expression was maintained, with expression of markers of committed PhRs, CRX. OPSN-S expression was not clearly visible. In contrast, NR2F2 was abundantly expressed. RHO expression was not evident, whereas RCVRN was readily detected, along with PRDM1 and the neural marker TUJ1. A similar pattern was observed in experiment 3 on D36 (Figure 3C), and OTX2 expression was once again maintained with the emergence of PhR commitment marker CRX. OPSN-S was clearer at this time point with RCVRN. NR2F2 was again expressed with RxRg appearance, which some studies have shown to be localized to cone PhRs [38]. Expression of PRDM1 and TUJ1 was also noted at this time point.

At D42 of experiment 2, expression of CRX OPSN-S, RCVRN, LHX2, OTX2, PRDM1, and TUJ1 was maintained (Figure 3D). By D43 of experiment 3, concurrent expression of RHO and OPSIN-M/L indicated the emergence of both rod and cone PhR lineages (Figure 3E). By D52 of experiment 2 (Figure 4A), expression of RCVRN and ARR3 was evident, while CRX was not. OPSN-S was present, and TUJ1 and LHX2 remained expressed. Ki67-positive nuclei within the differentiating cell population demonstrated ongoing proliferative activity. Colocalization with RCVRN was not observed, consistent with the distinct subcellular localization of these respective markers.

By D52 of experiment 3 (Figure 4B), cultures expressed THRβ, ARR3, and OPSN-M/L, while NRL and RCVRN were not detected. At D61 in experiment 2 (Figure 4C), staining for PRDM1, CRX, and RCVRN was reduced at this time point, whereas cone-specific markers ARR3, OPSN-S, GNAT2, and THRβ were more prominent. At the final time point, D66 of experiment 2 (Figure 4D), ARR3, OTX2, and OPSN-S expression persisted alongside diminishing TUJ1 and LHX2. Detected CRX levels varied between experiments. By D66 of experiment 3 (Figure 4D), all three PhR markers, RCVRN, OPSIN-M/L, and ARR3, were clearly expressed. The culture duration for experiment 3 was extended to D111, revealing sustained expression of both rod and cone PhR markers, RHO and OPSIN-S, via immunocytochemistry. Additionally, expression of the gap junction protein CX36 and the connecting cilium marker TMEM138 was observed. These findings, integrated with gene expression and ultrastructural data are discussed collectively in the Results section.

Flow cytometric detection of RCVRN and mature PhR marker expression

Live mCherry fluorescence, indicative of RCVRN expression, was detected as early as D24 in experiment 2 and D21 in experiment 3 (Appendix 14). By D52 of differentiation, 97.4% of cells in experiment 2 were RCVRN-positive by flow cytometry, compared to 100% in the 661W control population (Figure 5A).

Immunocytochemical staining revealed consistent expression of RCVRN, OPSN-S, and OPSN-M/L as early as D31 in experiment 2 and D36 in experiment 3 (Figure 3B,C). To quantify expression, flow cytometry was performed on cells collected at days 42 and 49 of experiment 3. At D42 (Figure 5B), 92.6% of cells were RCVRN-positive, 96.2% expressed OPSIN-M/L, and 70.3% expressed OPSIN-S. The 661W cells showed 99.9% positivity for all markers. By D49, 93.3% of cells were RCVRN-positive, compared to 99.3% in the 661W control. These markers were selected as they have previously been validated in the 661W cell line corresponding to PhR-like cell lineage [39–41].

Integrated morphological and transcriptomic analysis of PhR differentiation

PCA in Figure 6A illustrates dynamic changes in gene expression as cells transition from hIPSCs to PhR-like cells. Clustering of technical replicates confirms reproducibility and consistency within individual experiments (Figure 6B). Transcriptomic differences were observed across experimental groups. Experiment 1 showed greater variability in gene expression trends, while experiments 2 and 3 exhibited tighter clustering along the differentiation trajectory. Temporal gene expression patterns revealed pronounced shifts between hIPSC and early differentiation stages (D7–D9), particularly in experiments 1 and 2, followed by relative stabilization at later time points (Figure 6C).

Transcriptomic identification of photoreceptor pathways and morphological correlates

Heatmap visualization in Figure 6D illustrates dynamic gene expression patterns associated with PhR development across differentiation stages. Gene Ontology biological process enrichment analysis identified key pathways and genes associated with PhR development and maturation. Early time points showed expression of retinal progenitor genes, including OTX2, PAX6, RAX, and LHX2. By approximately D30, expression shifted toward committed PhR markers such as RCVRN, CRX, THRβ, OPSN-S, and RORB. Additional PhR-specific genes not previously examined (i.e., ARMC9, EYS, MAP2, and IMPG2) were also detected, supporting their involvement in PhR development. These data highlight the temporal activation and suppression of key regulatory genes, which likely correlate with distinct phases of PhR maturation. These transcriptomic findings informed subsequent morphological and ultrastructural analyses.

Scanning electron microscopy of PhR-like cells

Scanning electron microscopy (SEM) analysis of D99 cultures from experiment 3 (Figure 6E) revealed distinct morphological features consistent with PhR-like cells. The cells exhibited well-defined cell bodies (cb) and extended processes resembling connecting cilium (ml) observed in their neurite-like structure. Surface textures ranged from smooth to granular, suggesting different stages of differentiation or structural specialization. Cells appeared organized in clusters and interconnected networks, potentially reflecting in vivo–like maturation and increased cell-to-cell interactions. Some images showed connecting processes (potential plasma membrane projections [pp]) between cells, indicative of early structural communication or support mechanisms.

Transcriptomic and ultrastructural evidence of PhR maturation

Genes associated with PhR connecting cilium were upregulated during differentiation, suggesting tightly regulated timing for cilium assembly and function (Figure 7B). Transcriptomic analysis identified clusters of genes linked to outer segment development, indicative of advancing PhR maturation (Figure 7C). Although these genes were not the primary focus of earlier datasets, their expression patterns align with ultrastructural features observed via transmission electron microscopy (TEM; Figure 7A, D). The combined detection of connecting cilium and outer segment–related transcripts, along with immunocytochemical evidence of protein localization (Figure 7E), supports the emergence of putative outer segment structures at later time points. These findings provide molecular and morphological evidence for the maturation of PhR-like cells.

Ultrastructural analysis of differentiated PhR-like cells

TEM of hIPSC-derived PhR-like cells at days 50 and 64, alongside 661W controls (Appendix 15), revealed multiple ultrastructural features consistent with PhR-like morphology. At D50 (Figure 7A, B), cells demonstrated linear organization with identifiable inner segment–like and outer segment–like regions, calyceal process-like structures (cp), and abundant mitochondria (m). Vesicular structures were observed in proximity to the Golgi apparatus and within the cytoplasm. Whorl-like membranous structures were also present at later time points, particularly evident at D106 (Figure 7D).

Appendix 15 shows a cross section of D50 cells from experiment 3 arranged in a semi-organized layer with clearly identifiable nuclei (Nu) and prominent nucleoli. Cytoplasmic compartments containing mitochondria and Golgi apparatus were readily observed, together with membrane-bound vesicles and polarized cell morphology.

Additional TEM imaging (Appendix 15) confirmed similar ultrastructural features in D50 cultures, including well-defined nuclei, cytoplasmic organelles, Golgi complexes, mitochondria, calyceal process–like structures, and putative inner segment regions. Fine membrane extensions resembling early neurite projections were also present between adjacent cells.

Comparative TEM analysis of D64 experiment 3–derived PhR-like cells and immortalized 661W cells (Appendix 15) revealed notable morphological differences. In both cell types, nuclei were clearly identifiable, but 661W cells exhibited more uniform nuclear morphology and denser cytoplasmic organization. Mitochondria appeared more numerous and densely packed in 661W cells, whereas stem cell–derived PhR-like cells displayed a looser cytoplasmic arrangement with a more dispersed organelle distribution. These observations suggest that, while PhR-associated ultrastructural features are present in differentiated cultures, the overall cellular architecture differs from that of the immortalized 661W cell line.

At later differentiation stages (D106), TEM imaging revealed prominent whorl-like membranous structures consistent with ongoing putative outer segment–like membrane organization (Figure 7D). Integration of TEM and SEM datasets provides a framework for tracking morphological progression of hIPSC-derived PhR-like cells across differentiation time points.

Early transient photoreceptor commitment in 2D culture compared to organoid-based differentiation

A comparative analysis of selected genes was used to compare this culture paradigm to the more widely used organoid method of culturing in vitro retinal cells. Figure 8A demonstrates a diverging pattern across PC1. The monolayer method approximates a linear trajectory while the organoid cultures follow a more irregular average gene expression trend. Figure 8B presents a comparative analysis of gene expression profiles over time for selected transcription factors and markers associated with PhR differentiation, using this developed protocol and the more traditional organoid culture system. Organoid comparison data were obtained from [42]. The figure displays the temporal dynamics of multiple genes, including CRX, LHX2, and NR2F2, throughout the differentiation process, providing insights into the developmental trajectory under each culture condition.

The results highlight the temporal changes in gene expression across the differentiation process, with each subplot illustrating the distinct activation patterns of individual genes. These temporal patterns suggest different roles and activation timings during PhR differentiation, reflecting the complex regulatory processes involved. Both culture methods exhibit similar overall expression trends for most genes, indicating that both protocols can effectively induce PhR differentiation. However, differences in timing and amplitude of expression reveal important distinctions between the two approaches.

For genes such as CRX and RAX, the monolayer culture reaches peak expression earlier than the organoid culture, suggesting a faster commitment in the monolayer condition to a PhR final fate. Conversely, genes like SIX3 and PRDM1 show more sustained expression in organoid cultures, indicating prolonged activation and potentially a more gradual or complete differentiation pathway. This prolonged expression pattern in organoids could be attributed to the 3D structure, which provides a more in vivo–like environment, supporting better cell–cell interactions and niche signaling. The shaded areas surrounding the expression curves represent statistical variance in expression levels. Broader variance in some plots, such as TMEM138 and RCVRN, might reflect higher heterogeneity in gene expression among replicates or cell populations within the culture system. Such variability could be due to differences in differentiation efficiency or the intrinsic heterogeneity of the cells being cultured. Gene-specific differences are particularly notable in this comparative analysis. Genes such as NR2F2 and RXRG show distinct profiles between the two culture methods, with organoid cultures maintaining higher overall expression levels. This suggests that these genes might be more influenced by the 3D environment. Conversely, genes such as SIX3 and SOX2 exhibit highly synchronized patterns across both methods, implying that these genes are robustly regulated during differentiation, regardless of the culture system. The findings from this analysis indicate that while monolayer and organoid cultures follow divergent temporal differentiation kinetics, they ultimately converge on similar transcriptional programs, leading to a PhR fate.