Introduction

Oculocutaneous albinism (OCA) is an autosomal recessive disorder characterized by partial or complete absence of melanin biosynthesis1. OCA occurs in approximately 1 in 17,000 people worldwide2, with the highest prevalence of 1 in 4000–7000 individuals reported in Africa, whereas European countries report a prevalence of 1 in 12,000–15,000 individuals3. Patients with OCA typically present with reduced pigmentation in the skin, hair, and eyes, accompanied by ocular symptoms, including abnormal retinal development, hypopigmented fundus, nystagmus, and optic chiasm defects1,4. To date, eight types of OCA (OCA1–8) have been reported, each resulting from mutations in different genes4.

OCA1 is the most common type of OCA, accounting for 50% of all cases worldwide5,6. OCA1 is caused by mutations in the tyrosinase (TYR) gene, leading to a deficiency of the enzyme responsible for melanin production in cells, including melanocytes and retinal pigment epithelium1. Specifically, the lack of ocular melanin results in developmental abnormalities in the central visual pathway, causing vision impairment and refractive errors, because melanin has been reported to influence the topographic patterning of optic nerve fibers7. Hence, patients with OCA present with excessive crossing of optic nerve fibers, leading to strabismus and reduced stereoscopic vision8.

Various mouse models have been developed for investigating OCA1 pathogenesis9, which have facilitated attempts at therapeutic interventions to address the fundamental molecular defects in patients with OCA1, including l-dihydroxyphenylalanine, nitisinone, and gene therapies10,11,12. Nevertheless, mice have significant limitations due to anatomical differences from humans, including eye size, lens shape, number and types of cone photoreceptors, absence of a fovea, and unique photoreceptor adaptations13. These differences hinder the development of effective ophthalmological treatments for OCA. In recent years, nonhuman primate models of OCA have been developed to overcome these anatomical limitations14. Although these models provide promising avenues for exploring human OCA disease mechanisms and therapeutic development, they present considerable challenges associated with high financial and husbandry costs, as well as significant animal welfare and ethical concerns15.

Pigs have emerged as powerful large-animal models for human biomedical research due to their close anatomical, physiological, and immunological similarity with humans, combined with reproductive advantages such as shorter generation intervals and larger litter sizes than those of nonhuman primates and other livestock systems16,17,18,19. In particular, the porcine eye closely parallels the human ocular globe in size, scleral thickness, and cone photoreceptor distribution in the outer retina16,20,21, a resemblance that has been reinforced by multiple comparative investigations22,23. Although TYR knockout (KO) pig models replicating OCA1 have been generated previously24,25, their specific ophthalmological phenotypes characteristic of human OCA1 remain uncharacterized. To address this gap, we conducted an exhaustive ocular assessment of TYR-deficient porcine eyes, benchmarking these findings against clinical ophthalmic presentations in patients with OCA1 to establish the translational validity of the model.

Programmable nucleases, including zinc-finger nucleases26,27, transcription activator–like effector nucleases28, and CRISPR/Cas929,30,31, are widely used for genome editing in both cultured cells and whole organisms. The landmark transplantation of a genetically modified porcine heart into a human recipient emphasizes the accelerating demand for precisely edited pig models in translational research32. To date, most genome-edited swine have been produced via somatic cell nuclear transfer (SCNT) using donor cells in porcine target genes that have been disrupted by plasmid-based nuclease delivery followed by single-cell clonal expansion24,33,34,35,36,37,38,39. Nevertheless, the prolonged proliferation of these clones can induce replicative senescence, a phenomenon that is reported in early SCNT studies that markedly diminishes the survival of the resulting piglets24. These limitations highlight the need to critically reevaluate and optimize current methodologies for generating genome-edited pigs, with the goal of improving both efficiency and animal viability in biomedical applications.

To establish a clinically relevant large-animal model of OCA1 without the limitations of clonal senescence, we used a selection-free, single-cell-expansion-independent genome-editing approach to disrupt TYR in porcine, yielding genetically diverse KO offspring manifesting the characteristic white coat and pink irises of OCA1. These TYR-deficient pigs underwent exhaustive ophthalmic evaluation, encompassing fundus imaging, retinal morphometric analyses, intraocular pressure (IOP) measurements, and electroretinography (ERG), which revealed marked structural and functional retinal abnormalities closely recapitulating human OCA1 phenotypes.

Results

Generation of TYR KO nuclear donor cells by selection-free genome editing

The effectiveness of SCNT pig production is significantly affected by the quality and nature of donor cells40. In this study, we used healthy and proliferative Jeju native pig fetal fibroblasts (JNPFFs) as primary donor cells to generate genome-edited pigs. The Jeju native pigs (JNPs), characterized by a distinctive black coat phenotype, are native to the island of Jeju, South Korea41. Unlike other domesticated pig breeds, the smaller size and robust disease resistance of JNPs make them suitable for use as preclinical models42. To generate a nuclear donor cell for the porcine OCA1 model, we generated TYR KO cells by selection-free genome editing. Using CRISPR–Cas9/sgRNA ribonucleoproteins (RNPs), we generated gene-edited donor cell populations with high efficiency, streamlining the workflow and markedly reducing the turnaround time by eliminating both flow cytometric sorting and single-cell clonal expansion (Fig. 1A). We first designed nine different Cas9 target sites targeting the TYR exon (Fig. S1). Then, we transfected the preassembled complex of Cas9 protein and sgRNA into the cells by electroporation. Mutation frequencies were determined 3 days after transfection by targeted deep sequencing (Fig. 1A). Our findings revealed a wide range of mutation frequencies of 1.4–97.9% induced by Cas9 RNPs, depending on the specific target sites (Fig. 1B,C). In particular, Cas9 nuclease in combination with sgRNA-7-targeting TYR demonstrated exceptionally high activity, resulting in a mutation frequency of 97.9% in JNPFFs.

Fig. 1
figure 1

Production of TYR KO donor cells by selection-free genome editing. (A) Schematic overview detailing the process of generating OCA1 pigs using CRISPR-Cas9 RNPs and selection-free genome editing. (B) Evaluation of Cas9-mediated mutation frequencies for each sgRNA targeting TYR in JNPFF. (C) T7E1 assay for Cas9-mediated cleavage at TYR-targeting sites in each cell. (D) Examination of off-target editing frequency induced by Cas9-mediated gene editing in TYR KO donor cells.

Full size image

We next explored whether Cas9 nuclease induced unintended off-target editing in TYR KO donor cells. To evaluate potential off-target effects, we computationally identified potential off-target sites in the porcine reference genome that contained up to three base pair mismatches to the target sequences. We then performed targeted deep sequencing to evaluate the potential off-target sites. Our results revealed no detectable indel mutations higher than the sequencing error rates in any of the TYR KO donor cells (Fig. 1D). Therefore, we used the TYR KO cell line, generated using sgRNA-7 by selection-free genome editing, as donor cells for all subsequent experiments in the production of the porcine OCA1 model.

In vitro generation of TYR KO embryos by SCNT

We conducted SCNT and in vitro developmental evaluations using TYR KO donor cells. After SCNT using TYR KO donor cells, the fusion rate was 87.9%, the cleavage rate was 54.7%, and the blastocyst formation rate was 18.8%, showing no significant difference compared with those obtained using control donor cells, indicating successful normal embryonic development (Fig. 2A,B; Table 1). After the generation of blastocysts, we measured the mutation frequency in individual blastocysts by targeted deep sequencing and found that of the 22 blastocysts examined, 13 (59.1%) had a biallelic mutation and 5 (22.7%) had a monoallelic mutation (Figs. 2C and S2). These results indicate that TYR KO donor cells produced by selection-free, genome editing exerted no detrimental effect on early embryonic development after SCNT, and the resulting blastocysts were successfully gene-edited, demonstrating various mutation types.

Fig. 2
figure 2

Evaluation of embryonic development and mutation frequency after SCNT. (A) Representative image depicting control and TYR KO donor cells and SCNT embryos at various stages during early embryonic development. Scale bar = 300 μm. (B) Graphical representation illustrating the percentage of SCNT embryos reaching each stage in the cleavage and blastocyst formation patterns. (C) Summary depicting the ratio of TYR monoallelic KO and biallelic KO frequencies observed in a single blastocyst after SCNT.

Full size image
Table 1 Summary of early embryonic development after SCNT.
Full size table

Production of Porcine OCA1 model by selection-free genome editing

We next performed SCNT of embryos to a surrogate mother to generate the porcine OCA1 model. Using TYR KO donor cells, 160 fused embryos were generated through SCNT (Fig. S3A,B). Embryonic sacs were observed by ultrasound examination on the 28th day after transfer (Fig. S3C), and four offspring were successfully produced on the 119th day post-embryo transfer (Fig. 3A). Remarkably, the born piglets exhibited features of OCA1, such as white fur coloration, compared with wild-type (WT) piglets (Fig. 3A). Genotyping of the offspring using T7 endonuclease I (T7E1) assay and targeted deep sequencing revealed a biallelic mutation at the TYR target site in all four piglets (Fig. 3B,C, and S4). Consistent with the donor cell data, no indel mutations were detected in any of the four offspring at the potential off-target sites (Fig. 3D). Moreover, their body weight was similar to that of WT pigs (Fig. 3E). We next evaluated the viability of OCA1 pig offspring and observed that all four offspring remained healthy up to 50 days after production, indicating a significantly higher survival rate than that reported in a previous study24 (Fig. 3F). These results demonstrate the greater efficacy of selection-free genome-editing-mediated SCNT than that of methodologies using SCNT mediated by single-cell-derived clonal cells.

Fig. 3
figure 3

Generation of porcine OCA1 models. (A) Depiction of WT and porcine OCA1 pig models through photographs. (B) T7E1 assay for Cas9-mediated cleavage at TYR-targeting sites in porcine OCA1 models. (C) Assessment of mutation patterns at the Cas9 target sites of porcine OCA1 models. (D) Examination of off-target editing frequency induced by Cas9-mediated gene editing in porcine OCA1 models. (E,F) Graphical representation illustrating the body weight (E) and survival rates (F) of porcine OCA1 models (n = 4).

Full size image

In addition, we confirmed that OCA1 pigs consistently exhibited the anticipated albinistic phenotype, with complete loss of melanin pigmentation and undetectable TYR expression. WT pigs displayed robust black pigmentation in both ocular and pelage tissues, whereas TYR KO pigs exhibited the white coat and pink irises (Fig. 4A). Histological analysis by hematoxylin and eosin (H&E) staining revealed dense melanin deposits throughout the choroid and iris, as well as within the hair follicles and epidermis of sections obtained from WT pigs; however, these pigments were entirely absent in specimens obtained from OCA1 pigs (Fig. 4B). Correspondingly, western blotting of skin extracts confirmed the absence of TYR protein expression in all OCA1 pigs (Fig. 4C). In the quantitative evaluation of melanogenic transcripts, TYR transcript levels were not detected in OCA1 samples, and the expression of the downstream effectors TYRP1 and KIT significantly reduced compared with that in WT controls (p < 0.05; Fig. 4D).

Fig. 4
figure 4

Phenotypic analysis of porcine OCA1 models. (A) Visual representation of observed differences in eyes and hair between WT and porcine OCA1 models. (B) Histological examination of eyes and skin by H&E staining. Ch, Choroid; Sc, Sclera; HF, Hair follicle; EP, Epidermis; DM, Dermis. (C) Evaluation of TYR protein expression in skin tissues from both WT and porcine OCA1 models. (D) Relative mRNA expression levels of genes associated with pigmentation in skin tissues from both WT and porcine OCA1 models. (E) Pedigree of crosses between the founder (F0) OCA1 pig and four WT pigs, and representative photographs of F0 (OCA1 pig-1) and F1 (offspring F1-1) OCA1 pigs.

Full size image

Establishing a stable colony harboring defined mutations is crucial for disease phenotype studies and translational research applications. Therefore, we evaluated the potential for germline transmission of targeted genetic mutations generated through selection-free genome editing in OCA1 pigs. Male OCA1 pig-1 was mated with WT female JNPs, producing 20 F1 generation offspring (Fig. 4E). Targeted deep sequencing analysis revealed that all 20 F1 progeny carried monoallelic mutations and were verified to harbor TYR mutations derived from the OCA1 pig-1 genotype (Fig. S5). Collectively, these data demonstrate that OCA1 pigs generated through selection-free genome editing exhibit albinistic phenotypes and maintain normal fertility, enabling successful germline transmission of target mutations.

Ocular clinical features of Porcine OCA1 model

We next conducted a comprehensive ophthalmic evaluation of OCA1 porcine, which revealed marked fundus hypopigmentation, with improved visualization of the choroidal vasculature characteristic of human OCA43 (Fig. 5A). Quantitative morphometry demonstrated a significant reduction in retinal thickness, with the inner nuclear layer (INL) and outer nuclear layer (ONL) measuring approximately 50% and 40% of those in WT specimens, respectively (Fig. 5B; Table 2), indicative of retinal structural compromise. Furthermore, IOP significantly increased in OCA1 pigs compared with that in WT pigs (p < 0.01; Fig. 5C).

Fig. 5
figure 5

Comprehensive ophthalmological analysis of porcine OCA1 models. (A) Fundus images of WT and OCA1 pigs. (B) H&E staining of retinal sections from WT and OCA1 pigs sampled at 4 months of age. Abbreviations for the retinal layers: nerve fiber layer (NFL), ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL), outer nuclear layer (ONL), photoreceptor inner/outer segments (IS/OS), retinal pigment layer (RPE), and epidermis (E). (C) IOP of the eyes of OCA1 pigs was higher compared with that of WT pigs. (D,E) Results for scotopic ERG (3.0 cd∙s/m2 light intensity) for the eyes of WT and OCA1 pigs. (D) Representative electroretinograms for WT and OCA1 pigs. (E) Comparison of b-wave amplitudes. (F–H) Results for OPs of the rising b-wave of scotopic ERG (3.0 cd∙s/m2 light intensity). (F) Representative graphs for OPs for WT and OCA1 pigs. (G) Comparison of the sum of OP amplitudes and (H) the ratio of the sum of OP amplitudes to b-wave amplitude between WT and OCA1 pigs. The in vivo clinical evaluations of the eyes of WT and OCA1 pigs at 10 months of age (n = 4 eyes from 2 pigs for each genotype). For bar plots, results are expressed as mean ± SD superimposed with data for individual eyes. Student’s t-test was used for comparison between groups; **p < 0.01, ns = not significant.

Full size image
Table 2 Retinal layer thickness (INL and ONL) in the central and peripheral regions of WT and OCA1 pig eyes.
Full size table

ERG recordings under scotopic conditions at 3.0 cd·s/m2, which is indicative of inner retinal function, including ON bipolar and Müller cell activity44,45,46, revealed that OCA1 porcine exhibited markedly reduced b-wave amplitudes (p < 0.01) along with accelerated implicit times compared with those in WT controls (Fig. 5D,E; Table 3). The a-wave implicit times also significantly decreased in OCA1 pigs (p < 0.05, Table 3). There was no significant difference in the mean a-wave amplitude between WT and OCA1 pigs, but there was considerable variation in the measurements (Table 3; Fig. S6). After increasing the stimulus intensity to 10.0 cd·s/m2, the b-wave amplitude deficits persisted in the OCA1 cohort; however, both a- and b-wave implicit timings were comparable to those of WT pigs (Fig. S7; Table S2).

Table 3 Summary of the ERG results measured by scotopic ERG (3.0 cd∙s/m2 light intensity) of WT and OCA1 pig eyes.
Full size table

ERG recordings under scotopic conditions at 3.0 cd·s/m2 revealed increased oscillatory potentials (OPs) on the rising phase of b-wave in OCA1 pigs compared with that in WT controls (Fig. 5F). Although the summed OP1–OP5 implicit times and amplitudes showed no significant differences between the groups (Fig. 5G; Table 3), the ratio of total OP amplitude to b-wave amplitude significantly increased in OCA1 pigs (p < 0.01), indicating an increased OP contribution to the scotopic b-wave (Fig. 5H; Table 3). In the photopic ERG recording at the same stimulus intensity, OCA1 pigs exhibited a significantly shortened a-wave implicit time, whereas a- and b-wave amplitudes and flicker ERG parameters remained comparable to those in WT pigs (Fig. S8; Table S3).

Discussion

Pigs are considered animals that can recapitulate human phenotypes and are extensively used in diverse disease modeling research. In this study, we successfully produced porcine OCA1 models by selection-free genome editing, eliminating the need for single-cell-derived clonal expansion and demonstrating the efficacy of this streamlined approach. Our findings differ from those of previous methods that depended on clonal expansion from a single-cell24,33,34,35,36,37,38,39, which frequently resulted in cellular senescence and subsequently decreased survival rates in genome-edited pigs24. By avoiding this step, we not only increased the efficacy of producing porcine OCA1 models but also generated KO pigs with varying genotypes (Fig. 3C). Furthermore, our OCA1 pig model successfully recapitulated the OCA1 phenotype, demonstrating the complete loss of melanin in ocular and skin tissues along with comprehensive ophthalmological abnormalities, including retinal structural changes and visual dysfunction (Figs. 4 and 5).

In this study, we used the Cas9 RNP method, which is known for its ability to exert minimal off-target effects47. To evaluate the off-target effects of Cas9 nuclease, we identified potential off-target sites through computational analysis and measured the mutation frequencies at each site (Fig. 1D). Although our off-target assessment relied on amplicon-based deep sequencing of candidate sites, which is primarily sensitive to small indel and may not reliably detect larger deletions, insertions, or chromosomal rearrangements, the results indicated an absence of detectable off-target effects, thereby emphasizing the precision of Cas9 nuclease when targeting TYR. We also examined the general health of animals born as porcine OCA1 models and found no significant changes. The absence of health issues in porcine OCA1 models further supports the safety and efficacy of our genome editing strategy. An alternate method for generating KO pigs involves microinjection of Cas9 mRNA and sgRNA25; however, using this method for Cas9 delivery via porcine embryo microinjection could result in the formation of chimeric pigs25. Typically, these chimeras exhibit a mixture of biallelic mutant cells, monoallelic mutant cells, and WT cells. In contrast, the piglets generated by SCNT in our study demonstrated a remarkable advantage in uniform genotype. This distinction emphasizes the precision and consistency achieved by SCNT and its superiority in producing pigs with a consistent KO phenotype compared with the potential mosaic outcomes associated with microinjection methods.

Our findings extend those of previous TYR KO studies by demonstrating not only the anticipated melanin depletion in porcine ocular tissues but also profound structural retinal alterations that were not previously evaluated24,25. Zhang et al. reported negligible ERG changes in an OCA2 porcine model48, whereas the OCA1 pigs generated in our study exhibited a dramatic thinning of the retina, with the INL and ONL reduced by approximately 50%, in addition to pigment loss. This degree of retinal attenuation far exceeds the ~ 10% perifoveal thinning documented in human patients with albinism49,50, emphasizing a potentially more severe phenotype in the TYR-deficient porcine and the critical species- and gene-specific differences in the pathology of albinism.

ERG evaluations revealed selective inner retinal dysfunction in OCA1 pigs. Specifically, b-wave amplitudes significantly decreased under scotopic conditions at 3.0 and 10.0 cd·s/m2, whereas a-wave metrics remained comparable to those in WT controls (Figs. 5E, S6, and S7). Even at the low-intensity scotopic stimulus of 0.01 cd·s/m2, OCA1 pigs exhibited a modest reduction in rod-driven b-wave responses (Fig. S7), whereas photopic b-wave amplitudes remained unaffected (Fig. S8; Table S3). The preservation of a-wave responses along with b-wave attenuation implicates the dysfunction of INL, primarily bipolar and Müller cells, rather than photoreceptor pathology in ONL44,45,46. Furthermore, the decreased implicit times suggest improved retinal light scatter and increased anterior retinal stimulation due to hypopigmentation51,52. Altogether, these findings indicate a selective compromise of the rod bipolar pathway in TYR-deficient porcine, necessitating further research into the cellular mechanisms underlying this impairment.

In addition to inner retinal deficits, porcine OCA1 models exhibited supernormal OPs under scotopic stimulation, with marked OP amplitudes on the ascending limb of the b-wave at 3.0 cd·s/m2 (Fig. 5H). Although the total OP amplitude at 3.0 cd·s/m2 did not reach statistical significance, the elevated OP-to-b-wave amplitude ratio in TYR KO pigs underscores an increased OP contribution to the overall response. Such supernormal OPs have been related to disturbances in synaptic transmission within the inner retina53, as demonstrated in rhodopsin-mutant retinitis pigmentosa models in rabbits54 and pigs55. Overall, these data indicate that TYR loss stimulates developmental and synaptic alterations in the rod bipolar circuitry of the porcine retina12,56,57.

Moreover, OCA1 pigs exhibited elevated IOP compared with WT controls (Fig. 5C). The mean IOP measured in WT pigs (12.4 ± 1.67 mmHg) was consistent with values reported in live porcine studies58,59. The etiology of increased IOP in the context of TYR deficiency remains unclear. Although human OCA1 is not typically related to ocular hypertension, isolated case reports describe concurrent glaucoma in patients with albinism60,61. Scrutinizing whether this hypertensive phenotype develops directly from OCA1 pathology or reflects secondary mechanisms in the porcine model will require focused longitudinal and mechanistic studies.

Overall, the in vivo phenotypes of the OCA1 pigs were significant but relatively mild compared with the stark reduction in retinal thickness observed in the histological sections. This discrepancy in observations is complicated by the fact that histological sections and in vivo readings were obtained from different animals at different ages (histology: 4 months after birth; in vivo readings: 10 months after birth), due to constraints in available equipment and limited sample sizes. The substantial variation observed in some of the parameters in the ERG measurements could also be attributed to the limited sample size. Additionally, individual variations among OCA1 pigs based on their specific genotypes may also contribute to these observations.

Nevertheless, we used age-matched WT controls for both histological and ERG analyses, which allowed us to identify ophthalmological phenotypes associated with OCA1 disease. Furthermore, mutations of OCA1 pigs can be transmitted to subsequent generations through germline transmission (Figs. 4E and S5). Therefore, this discrepancy requires further investigation utilizing long-term tracking of subsequent generations using both functional (ERG) and morphological (in vivo optical coherence tomography) measurements.

This study represents the inaugural use of selection-free genome editing to establish porcine OCA1 models, accompanied by the first comprehensive ophthalmic phenotyping of TYR-deficient pigs. Our streamlined editing strategy achieved greater efficiency and consistency than traditional single-cell cloning approaches. Although porcine eyes mimic human ocular anatomy in terms of globe size and photoreceptor composition, they lack a macula and fovea, featuring instead a cone-rich foveal streak, which is an inherent limitation for modeling the foveal hypoplasia central to human OCA162,63. Nonetheless, the extensive structural and functional retinal impairments we documented support the translational utility of these porcine models. By integrating rapid, selection-free editing with detailed phenotypic evaluation, our platform provides a robust foundation for dissecting the pathogenesis of OCA1 and advancing the preclinical evaluation of targeted therapies.

Materials and methods

Animals and ethics declaration

The Jeju native black pigs used in this study were raised at Cronex Inc. (Cheongju, South Korea). All animal experiments were approved by the Committee on Ethics of Animal Experiments of the Chungbuk National University (Permit Number: CBNUA-2176-23-01). All surgeries and experiments were conducted under anesthesia to minimize animal suffering. Additionally, all experimental procedures were conducted in accordance with relevant institutional and national guidelines and were designed, performed, and reported following the ARRIVE guidelines 2.0 (https://arriveguidelines.org). Animals were anesthetized via the auricular vein with a ketamine–xylazine mixture (1:3 v/v; total dose 10 mg/kg ketamine and 3.3 mg/kg xylazine), and anesthesia was maintained with 1–2% isoflurane in oxygen. Under deep anesthesia, euthanasia was then performed by intravenous administration of succinylcholine at 1 mL per 10 kg body weight.

Generation of JNPFFs

After 50 days of fertilization, the JNPs were euthanized, and the fetus heads and internal organs were removed. Then, each fetus was minced and separately incubated in 0.25% trypsin for 30 min at 37 °C. The digested cells were washed with Dulbecco’s modified Eagle medium supplemented with 10% (v/v) fetal bovine serum and 2% (v/v) antibiotic–antimycotic. After an overnight incubation, adherent cells were subcultured and cryopreserved for further experiments.

Cas9 and in vitro transcribed sgRNA

Cas9 protein was purchased from integrated DNA technologies. sgRNAs were synthesized by in vitro transcription using T7 RNA polymerase (New England Biolabs). Each 20-nucleotide sgRNA sequence is positioned between the 5′-GAAATTAATACGACTCACTATA (T7-Promoter) sequence and 3′-GTTTTAGAGCTAGAAATAGCAAGTTAAAATAAGGCTAGTCCG (homology to scaffold template), which is the forward primer for in vitro transcription. sgRNA templates were generated by two complementary oligonucleotides each forward primer and Universal Reverse primer by PCR. sgRNA templates were incubated with T7 RNA polymerase (NEB) in a reaction buffer at 37 °C for 4 h according to the manufacturer’s instruction. Transcribed sgRNAs were incubated with calf intestinal alkaline phosphatase (New England Biolabs) in a 10× rCutSmart reaction buffer at 37 °C for 2 h to remove phosphates and purified using PCR purification kits (GeneAll Biotechnology). All primer sequences are listed in Table S1.

Transfection

JNPFFs were detached using trypsin/EDTA and collected at the desired cell number. The collected cells were resuspended in 16.4 µL of buffer P3 and 3.6 µL of supplement 1. The transfection-ready cell mixture was electroporated with 17 µg of spCas9 nuclease and 5 µg of in vitro transcribed sgRNA using the P3 Primary Cell 4D-Nucleofector (Lonza) with pulse code CM-137. After nucleofection, the cells were incubated in a 5% CO2 incubator. Genomic DNA was isolated using a DNeasy Blood & Tissue kit (Qiagen) at 72 h after transfection.

SCNT and embryo transfer

In vitro maturation of porcine oocytes, SCNT, and embryo transfer were performed according to methods described in our previous studies64,65. Cell colonies with the highest editing efficiency were used as nuclear donors for SCNT. The donor cells were injected into the perivitelline space of enucleated oocytes and fused by an electric pulse. Reconstructed embryos were transferred to porcine zygote medium (PZM) with 6-DMAP (0.4 µg/mL demecolcine and 6-dimethylaminopurine) for activation. After 4 h, the activated embryos were transferred to fresh PZM and cultured in a humidified atmosphere with 5% CO2 and 5% O2 for 6–7 days at 39 °C. Each blastocyst was individually sampled for genotyping. For embryo transfer, reconstructed embryos were transferred into the oviduct (at the ampullary–isthmic junction) of a surrogate. Pregnancy was confirmed by ultrasound examination 28 days after transplantation. All piglets were delivered naturally.

Targeted deep sequencing

Genomic DNA was purified using the DNeasy Blood & Tissue kit (Qiagen) according to the manufacturer’s instructions. Genomic DNA that included the on-target and potential off-target sites was amplified using Taq DNA polymerase (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocols. The amplified products, which contained Illumina TruSeq HT dual index adapter sequences, were subsequently subjected to 150-bp paired-end sequencing using Illumina iSeq 100. To calculate the frequencies of indel, we used MAUND, which is available at https://github.com/ibs-cge/maund.

Quantitative reverse transcription PCR

Total RNA was extracted using TRIzol reagent (TaKaRa Bio, Inc., Otsu, Shiga, Japan) according to the manufacturer’s protocol. Subsequently, 1 µg of total RNA was converted into complementary DNA (cDNA) using SuperScript IV VILO Master Mix (Thermo Fisher Scientific). PCR assays were performed using the synthesized cDNA, 2× SYBR Premix Ex Taq (TaKaRa Bio, Inc.), and 5 pmol of specific primers (Macrogen, Inc., Seoul, Republic of Korea). PCR amplification was performed at 95 °C for 5 min, followed by 40 cycles of 95 °C for 15 s, 56 °C for 15 s, and 72 °C for 30 s using the CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). Relative quantification was normalized to RN18S and calculated using the 2−ΔΔCT method66. All primer sequences are listed in Supplementary Table 1.

Western blot analysis

Skin samples from pigs of similar age were homogenized and lysed in RIPA buffer (Thermo Scientific) supplemented with Halt™ Protease and Phosphatase Inhibitor Cocktail (Thermo Scientific). SDS-PAGE separation of an equal amount of proteins from skin tissue lysates was conducted using 10% sodium dodecyl sulfate–polyacrylamide gel. Then, the proteins were transferred to a PVDF membrane (Millipore Corporation, Billerica, MA, USA), blocked with EveryBlot Blocking Buffer (Bio-Rad) for 10 min at room temperature, and incubated overnight at 4 °C with the following antibodies: anti-TYR (1:1000, Abcam, ab170905) and anti-GAPDH (1:1000, Cell Signaling, #2118). After washing in TBST, the blot membranes were incubated with HRP-conjugated secondary antibodies (anti-rabbit, 1:3000) in a blocking buffer for 1.5 h at room temperature. Subsequently, the protein bands were detected using SuperSignal™ West Pico PLUS Chemiluminescent Substrate (Thermo Scientific) and visualized using a Lumino Graph II (ATTO Corporation, Tokyo, Japan).

Histological analysis

Whole eyes from pigs at 4 months of age were fixed in TB-Fix solution after anesthesia and stored overnight at 4 °C, as reported previously67. Skin tissues were sampled in 10% buffered formaldehyde. Subsequently, the samples were paraffin-embedded and sectioned at a thickness of 3 μm. After deparaffinization, the sections were subjected to H&E staining and imaged using the Thunder Imaging Systems (Leica). The mean thickness of retinal layers (INL and ONL) was measured using the Fiji Image J (v.1.54f) software, which is available at https://imagej.net/ij/. For each sample, retinal layer measurements were obtained from two central areas (near the optic nerve head) and three peripheral areas.

Clinical ocular evaluation in pigs

Two male OCA1 pigs and two male WT pigs at 10 months of age were observed for clinical evaluation. The animals were clinically evaluated using dark-adapted (scotopic) and light-adapted (photopic) methods via the ERG test. Before ERG examination, the pigs were first anesthetized using a mixture of zoletil and xylazine. Anesthesia was maintained using isoflurane throughout the examination. To achieve pharmacologic mydriasis, 1% tropicamide (Alcon) was applied to the eye. The pigs were then dark-adapted for 20 min before ERG testing.

For the ERG setup, needle electrodes (12-mm ground/reference needle electrode, Electrode Store) were used as the ground electrode, which was inserted subcutaneously in the forehead, and the reference electrode, which was placed at the base of the ipsilateral ear. The contact electrode (ERG-Jet, Fabrinal) was placed on the cornea with a drop of sodium carboxymethylcellulose eyedrops (Jeil Healthcare Science, Inc.). The ground, reference, and contact electrodes were connected to the Retevet device (LKC Technologies, Inc.). ERG was performed using a preset ECVO 5-step, long protocol. The protocol started with 0.01, 3.0, and 10.0 cd∙s/m2 flash steps (0.2, 1/15, and 0.05 Hz stimulus time, respectively) in the scotopic state in both eyes. After a 20-min light adaptation, 3.0 cd∙s/m2 flash and flicker steps (0.2 and 28.3 Hz stimulus time, respectively) were performed in the photopic state in both eyes.

Fundus images of WT and OCA1 pigs were obtained using Cirrus photo 600 (Zeiss). Pigs were anesthetized using a mixture of zoletil and xylazine, and 1% tropicamide (Alcon) was applied to the eye for pharmacologic mydriasis before fundus imaging. Fundus images were obtained at a separate session to the ERG measurements.

For IOP measurements, the pigs were anesthetized using a mixture of zoletil and xylazine, and anesthesia was maintained using isoflurane throughout the examination. IOP was measured using a tonometer (Tono-Pen AVIA, Reichert Technologies). For each eye, five measurements were obtained and averaged.

Statistical analysis

Statistical analysis was conducted using SPSS 21.0 (SPSS Inc., Chicago, IL, USA). Data are expressed as mean ± SEM. Comparisons between two groups were analyzed using Student’s t-test. Statistical significance was considered at *p < 0.05 and **p < 0.01.