**Background:** Age-related macular degeneration and retinitis pigmentosa cause irreversible photoreceptor loss, leading to blindness. Current treatments cannot replace lost photoreceptors. Stem cell-derived photoreceptor transplantation offers a potential therapy, but existing methods using retinal organoids are complex, lengthy, and rely on undefined animal-derived components. Laminins are basement membrane proteins that provide tissue-specific cues. The retina-specific laminin isoform LN523 (α5β2γ3) is enriched in the inter-photoreceptor matrix. This study hypothesized that LN523 could recapitulate the retinal niche to drive differentiation of human embryonic stem cells (hESCs) into photoreceptor progenitors.
**Methods:** Recombinant human LN523 and LN323 were produced in HEK293 cells and purified. Four hESC lines (H1, H9, HS980, HADC106) were cultured on LN523+LN521 (2:1 ratio) in NutriStem medium. Differentiation was induced by switching to neural induction medium (day 2–9) and then photoreceptor differentiation medium (day 9–32). Single-cell RNA-seq was performed at days 0, 2, 9, 22, and 32. Immunofluorescence quantified CRX and RCVRN co-expression. Tumorigenicity was assessed in nude mice. For in vivo studies, day 32 progenitors were cryopreserved and transplanted subretinally into rd10 mice (P20) and rabbits with induced degeneration. Functional recovery was evaluated by full-field electroretinogram (ERG) at 1, 2, and 4 weeks post-transplantation in mice, and by water maze swimming test at 4 weeks. Histology and electron microscopy assessed engraftment, maturation, and synaptic connectivity at 2, 4, and 20 weeks in mice and at 1 month in rabbits.
**Key Results:** LN523+LN521 drove differentiation into photoreceptor progenitors within 32 days. Single-cell RNA-seq showed downregulation of pluripotency markers (Oct3/4, Nanog) by day 9, upregulation of eye-field markers (Pax6, Rax) at day 9, optic cup markers (Vsx2, MITF) at day 22, and photoreceptor progenitor markers (CRX, NRL, RCVRN, PDE6H) at day 32. Approximately 17% of cells were CRX+ by transcriptomics, and 33.5% co-expressed CRX and RCVRN by immunofluorescence. Differentiation was highly reproducible across four hESC lines (Spearman ρ = 0.91–0.96, p < 2.2×10⁻¹⁶). LN523+LN521 yielded 23.2% and 32.4% higher efficiency than LN323+LN521 and LN521 alone, respectively. No teratomas formed in nude mice injected with day 32 cells. In rd10 mice, at 2 weeks post-transplantation (P34), cell-transplanted eyes showed preserved outer nuclear layer (ONL) thickness (p < 0.005) and significantly higher ERG a-wave (photoreceptor) and b-wave (bipolar cell) amplitudes compared to sham (p < 0.05). At 4 weeks, ERG differences were not significant, but water maze testing showed that 73% of cell-transplanted mice swam <4 m vs. 33% of sham, and 80% swam <40 s vs. 47% of sham (p < 0.05). At 20 weeks, engrafted cells survived, expressed rhodopsin, and associated with host PKCα+ bipolar cells and synaptophysin+ synapses. In rabbits, engrafted cells expressed CRX, RCVRN, rhodopsin, and s-opsin, and showed colocalization of synaptophysin and Bassoon, indicating ribbon synapse formation. Host PKCα+ bipolar cell processes extended into the graft.
**Clinical Implications:** This study presents a chemically defined, xeno-free, and reproducible method to generate photoreceptor progenitors from hESCs in 32 days using retina-specific laminin LN523. The progenitors engraft, mature in vivo, form synaptic connections with host bipolar cells, and partially restore visual function in two animal models without teratoma formation. This approach overcomes limitations of organoid-based methods (batch variability, long duration, animal-derived components) and may pave the way for safe and scalable stem cell therapies for retinal degenerative diseases. However, the study notes limitations: lack of purification of photoreceptor progenitors, inability of full-field ERG to detect localized functional recovery at later time points, and the need for further validation in fully degenerated retinas.