Design of triple HLA gene KO iPSCs
To generate hypoimmunogenic iPSCs without immune rejection, we used YiP3 PBMC-derived iPSCs with heterozygous alleles for the HLA genes on chromosome 6. The YiP3 line carried the following alleles: HLA-A 11:01:01:01 and HLA-A 29:01:01:01, HLA-B 13:02:01:01 and HLA-B 58:01:01:01, HLA-C 03:02:02 and HLA-C 02:02:02, and HLA-DRA 01:02:01 and HLA-DRA 01:01:02. Our strategy aimed to KO HLA-A and HLA-B, which represent polymorphisms in class I of the HLA locus, using CRISPR/Cas9, and KO HLA-DRA to prevent the expression of HLA-DR, which represents polymorphism in class II, while leaving HLA-C, which has a minor polymorphism (Fig. 1a, b). The CRISPR/Cas9 system was used for gene KO. For each gene, two guide RNAs (gRNAs) were designed to target regions excluding heterogeneous regions within the HLA-A gene, where a protospacer adjacent motif (PAM) site (NGG) existed and may act biallelically on a 20 bp sequence (Supplementary Fig. 1a-c). For example, using immuno polymorphism database-international immunogenetics information system (IPD-IMGT)/HLA database 42, each allele was aligned We designed a gRNA (G0002-HLA-A-g1, ACAGCGACGCCGCGAGCCAG, PAM:AGG) targeting codons 37–43 within Exon 2 for HLA-A. We designed a gRNA (G0002-HLA-B-g1, GCTGTCGAACCTCACGAACT, PAM:GGG) targeting codons 31–38 within Exon 2 for HLA-B and a gRNA (G0002-HLA-DRA-g2, TGGCAAAGAAGGAGACGGTC, PAM:TGG) targeting codons 36–42 within Exon 2 for HLA-DRA (Fig. 1c, d, e).
Assessment of designed gRNA
The efficiency of CRISPR/Cas9 guide RNAs (gRNAs) designed for the HLA-A, HLA-B, and HLA-DRA genes was assessed by electroporating bulk iPSCs with each gRNA and measuring the transfection efficiency (Fig. 2a). Sanger sequencing and inference of CRISPR edits (ICE) analysis of the transfected iPSC pool revealed that, for HLA-A gRNA, G0002-HLA-A-g1 and G0002-HLA-A-g2 demonstrated 91 and 47% efficiency, respectively. Therefore, G0002-HLA-A-g1 was selected as the final gRNA. For HLA-B gRNA, G0002-HLA-B-g1 and G0002-HLA-B-g2 demonstrated 78 and 0% efficiency, respectively, indicating a lack of match. Therefore, G0002-HLA-B-g1 was selected as the final gRNA. For HLA-DRA, G0002-HLA-DRA-g1 and G0002-HLA-DRA-g2 demonstrated 99 and 86% efficiency, respectively. However, considering that the HLA-DR polymorphism predominantly occurs within Exon 2, G0002-HLA-DRA-g1, targeting this region, was selected as the final gRNA (Fig. 2b, c, d).
Confirmation of triple HLA gene KO iPSCs
To obtain engineered YiP3 cells with triple HLA gene KO, single clones were obtained through electroporation-mediated transfection using selected gRNAs that target the HLA-A, HLA-B, and HLA-DRA genes (Fig. 2e). To establish the concentration conditions for each gRNA for transfection, 40 and 80 µg of each gRNA were used to form the RNP complex. The efficiency of KO score for each gene, determined through ICE analysis, was 53, 37, 44, 58, 45, and 56%. Subsequently, transfection was performed using 80 µg of each gRNA to generate the RNP complex (Supplementary Fig. 2a). Single-cell cloning was conducted on the transfected iPSCs, followed by EP pool analysis and genotyping using next-generation sequencing (NGS). Subsequently, 48 clones were collected by seeding them into a 96-well plate, and each clone was screened and genotyped using Sanger sequencing. Six potential triple-KO clones were selected and further assessed through confirmatory sequencing and/or NGS (Supplementary Fig. 2b).
Clone B4 had mixed alleles in the HLA-B region, whereas clone C6 had unresolved issues in the HLA-A region and was excluded from further assays (Supplementary Fig. 3b). Genotyping results for the triple-KO clones underwent further sequencing verification for final confirmation. Clone A7 exhibited homozygosity with a 1-bp insertion, 28-bp deletion, and 1-bp deletion in the HLA-A, HLA-B, and HLA-DRA regions, respectively. Clone B2 exhibited homozygosity with a 1-bp insertion, 34-bp deletion, and 2-bp deletion in the HLA-A, HLA-B, and HLA-DRA regions, respectively (Fig. 2f, Supplementary Fig. 2b). Additionally, clones B3 and B11 demonstrated incomplete HLA-A with a 1-bp insertion and 1-bp deletion near g1 (Supplementary Fig. 2b, c). Therefore, based on the genotyping results, clones A7 and B2 were selected as strong candidates for the triple HLA gene KO, and further comparative analysis was conducted along with clones B3 and B11 through additional assays.
Engineered triple-KO iPSCs retain the pluripotency
To assess the pluripotency, quality testing was conducted on the triple-KO iPSC clones in comparison to the control YiP3. Morphologically, clones A7 and B2 exhibited colony formation (Fig. 3a), and positive staining for alkaline phosphatase (AP) confirmed their undifferentiated state (Fig. 3b). At the mRNA level, the expression of pluripotency markers, such as octamer-binding transcription factor 4 (OCT4), SRY (sex-determining region Y)-box 2 (SOX2), Krüppel-like factor 4 (KLF4), Lin-28 homolog A (LIN28), and Nanog homeobox (NANOG) was confirmed, whereas the endoderm differentiation marker (SOX17), mesoderm differentiation marker (BRACHYURY), and ectoderm marker (paired box 6 [PAX6]) were not expressed (Fig. 3c). Similar patterns were observed in clones B3 and B11 (Supplementary Fig. 3a, b, c).
Subsequently, to compare the expression of pluripotency markers at the protein level using pre-engineered YiP3, flow cytometry was performed for OCT4, stage-specific embryonic antigen 4 (SSEA4), NANOG, tumor rejection antigen 1–60 (TRA-1-60), and the negative marker (CD34). The results demonstrated that clones A7 and B2 expressed OCT4, SSEA4, NANOG, and TRA-1-60 at levels exceeding 95% in cell populations, comparable to YiP3 (Fig. 3d). Clone B11 exhibited expression levels exceeding 99% when compared to YiP3, whereas clone B3 exhibited a lower expression of NANOG at 86% (Supplementary Fig. 3d).
To assess the ability of each clone to differentiate into the three germ layers (endoderm, mesoderm, and ectoderm), lineage differentiation was induced, and immunofluorescence staining for SOX17, BRACHYURY, and paired box 6 (PAX6) markers was performed. Expression of these markers was confirmed in all clones, indicating that engineered iPSC clones A7 and B2 retained their differentiation capacity similar to that of YiP3 without any alterations (Fig. 3e). Similar results were observed for clones B3 and B11, indicating no effect on the differentiation of three germ layers (Supplementary Fig. 3f).
In conclusion, among the selected triple-KO clones, clones A7 and B2 exhibited normal iPSC properties regarding pluripotency and the differentiation ability of three germ layers, similar to YiP3. However, clone B3 was excluded because of its significantly lower NANOG expression at the protein level.
Genetic stability of HLA-triple KO iPSCs in clones A7 and B2
The genetic stability of HLA-triple KO clones A7 and B2 was assessed by investigating karyotypes, CNVs, CRISPR/Cas9 off-targets, and expression alterations of cell differentiation potential-associated genes. Normal karyotypes were identified for clones A7 and B2 (Fig. 5a), whereas clone B11 exhibited a chromosomal abnormality with a deletion on chromosome 6p (Supplementary Fig. 3e), thereby precluding its consideration as an HLA-triple KO iPSC candidate clone. No CNV was identified in clone A7 through SNP genotyping using the CytoScan HD array (Fig. 5b). However, copy number (CN) losses at two loci, 2q22.1 (138.17–138.82 Mbp on chromosome 2) and 6p21.33–6p21.32 (31.32–32.41 Mbp on chromosome 6), were identified (Fig. 5b). Specifically, CN loss at 6p21.33 and 6p21.32 indicates the genomic instability of HLA-B and HLA-DRA in clone B2 (Fig. 5b). After predicting CRISPR/Cas9 off-targets in the human reference genome using Cas-OFFinder, potential off-target Cas9 activity was detected from WGS data of clones A7 and B2. Among the 21 predicted off-target sites using Cas-OFFinder, none corresponding to off-target sites was observed in the WGS data of clones A7 and B2 (Fig. 5c). However, four structural variants (SVs) that may be induced through on-target and/or off-target activity were identified: a 28-bp deletion in HLA-A for clone A7, and a 34-bp deletion in HLA-A and two CNV losses for clone B2 (Fig. 5c). The SVs identified in clones A7 and B2 were observed within coding sequences adjacent to the on-targets for HLA-A and HLA-B KO (Supplementary Table 5), indicating the possibility of additive induction in HLA-A and HLA-B KO. Additionally, somatic mutations, including single nucleotide variants (SNVs) and insertions and/or deletions (InDels), were analyzed from the whole-genome sequence data of A7 and B2 clones by comparing them with YiP3 iPSCs. Somatic coding variants that may cause functional effects were not observed in clones A7 and B2 (Supplementary Table 6). In addition to assessing the genomic stability of the HLA-triple KO clones, we assessed their transcriptome alterations. First, it was confirmed that the KO genes (HLA-A, HLA-B, and HLA-DRA) were downregulated in clones A7 and B2 compared to those in YiP3 (Fig. 5d), indicating the consequences of HLA-A, HLA-B, and HLA-DRA KOs. Second, despite the KO events, a strong correlation in overall gene expression was observed among YiP3, A7, and B2: a Pearson correlation of ≥ 0.99 between YiP3 and A7, and ≥ 0.98 between YiP3 and B2 (Fig. 5e). This indicates that the genome-wide gene expression pattern of the HLA-triple KO clones A7 and B2 closely resembles that of the wild-type YiP3. Third, we also assessed the potential role of the cell differentiation process through gene set enrichment analysis (GSEA). In contrast to clone A7, genes associated with the development of the three germ layers (endoderm, mesoderm, and ectoderm) in pluripotent stem cells were significantly downregulated in clone B2 (p = 0.000 ~ 0.029) (Fig. 5f), which was consistent with the low EC differentiation ability in clone B2 (Fig. 7b). These results indicate the potential genomic instability of clone B2 owing to the downregulation of genes associated with cell development and off-target effects, such as CNV loss. Considering these results, our findings demonstrate that only the HLA-triple KO clone A7 attains genomic stability.
Assessment of the HLA expression
We assessed the mRNA expression levels of HLA-A, HLA-B, and HLA-DRA in the genetically edited and selected clones (A7 and B2). Real-time PCR analysis revealed a significant reduction in the delta cycle threshold (dCt) values of HLA-A and HLA-B compared to YiP3 (p < 0.01 vs. YiP3), and a reduction in HLA-DRA expression (p < 0.05 vs. YiP3). Normalization using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) relative to YiP3 demonstrated a significant reduction in the mRNA levels of HLA-A and HLA-B (p < 0.01 vs. YiP3), alongside a reduction in HLA-DRA expression (p < 0.05 vs. YiP3) (Fig. 4a).
Flow cytometry analysis was conducted to assess whether gene editing and selection targeting HLA-A, B, and DRA in clones A7 and B2 resulted in the absence of HLA-DR protein expression. Without IFN-γ stimulation, YiP3 and clones A7 or B2 iPSCs did not express HLA-A, B, DR, and C (unedited region). Following IFN-γ stimulation for two days, HLA-A, B, and C protein expression increased in YiP3, with HLA-A, B, C, and DR increasing by 99.03, 91.61, 88.35, and 0.04%, respectively (Fig. 4b). In contrast, under IFN-γ stimulation, the triple-KO clone A7 exhibited a significant reduction in HLA-A (0.07%), HLA-B (0.15%), and HLA-DR (0.02%) protein expression. However, HLA-C protein expression (97.34%) remained unaffected, because it was not within the gene-edited region. Similar results were observed for clone B2 (Fig. 4b). These results demonstrate selective KO of the targeted HLA-A, B, and DRA regions in YiP3 at the protein level, as confirmed through our analyses.
In vitro immunogenicity test
To assess whether gene-edited iPSCs exhibit immunogenicity, we conducted co-culture experiments using PBMCs from a donor whose HLA types differed from those of YiP3. We isolated TEM and TCM cells and assessed their proliferation levels (Fig. 4a). Initially, we selected a donor with HLA types different from those of YiP3 for each allele. Allele 1 carried HLA-A 02:01, HLA-B 15:01, HLA-C 01:02, and HLA-DRB1 11:01, and allele 2 carried HLA-A 02:07, HLA-B 46:01, HLA-C 04:01, and HLA-DRB1 15:02 (Fig. 4b). Before co-culturing with PBMCs, both YiP3 and clones A7 and B2 were stimulated with IFN-γ for two days. We analyzed the proliferation of activated T cells in response to antigen presentation by antigen-presenting cells within PBMCs that were depleted of T cells. Co-cultures were initiated using PBMCs and carboxyfluorescein succinimidyl ester (CFSE)-labeled CD4 + T cells, to assess the proliferation of CD4 + TCM (CD3 + CD4 + CD45RO + CD62L+) and CD4 + TEM (CD3 + CD4 + CD45RO + CD62L-) cells using flow cytometry. Harvesting PBMCs and CFSE-labeled CD4 + TCM cells seven days after the initiation of co-culture revealed a slight increase in the cell population for YiP3 and clones A7 and B2, with average increases of 4.8, 4.5, and 4.7%, respectively. However, upon restimulation of YiP3 and clones A7 and B2 using IFN-γ-stimulated PBMCs after 14 and 21 days of co-culture, respectively, the proliferation of CFSE-labeled CD4 + TCM cells increased to 12.8% (day 14, mean) and 25.2% (day 21, mean) for YiP3. In contrast, proliferation was reduced to 9.6% (day14, mean) and 19.3% (day 21, mean) for clone A7, which was significantly lower than that of YiP3 (p < 0.05 vs. YiP3). Clone B2 demonstrated a similar trend of increased proliferation, but the reduction was less pronounced compared to that of clone A7, with no significant difference in p-values observed (Fig. 4c, d). Similar results were observed in CFSE-labeled CD4 + TEM cells. On the 21st day of co-culture, proliferation increased to 22.7% (day 21, mean) for CD4 + TEM cells in response to YiP3 stimulation, while clone A7 exhibited a significantly lower proliferation of 17.1% (day 21, mean) compared to that of YiP3 (p < 0.05 vs. YiP3). Clone B2 did not exhibit significant alterations compared with YiP3 (Fig. 4c, d).
In conclusion, triple-KO clone A7, with edited HLA-A, HLA-B, and HLA-DRA genes, exhibited reduced immunogenicity when co-cultured with PBMCs of different HLA types, indicating its potential for immunological compatibility.
Assessment of the HLA protein expression in EC
We differentiated clones A7 and B2 into ECs of mesoderm lineage and compared their differentiation capacities. Both YiP3 and clone A7 differentiated into ECs through the hemogenic mesoderm, displaying morphology similar to that of primary ECs. However, we observed a slightly different morphology for clone B2. Flow cytometry analysis of CD31 and VE-cadherin double-positive cells, markers for ECs, revealed percentages of 95.69, 91.9, and 0.36% for YiP3, clone A7, and clone B2, respectively. These results indicated that clone A7, with a differentiation rate of over 90% in ECs, exhibited normal differentiation ability. In contrast, clone B2 exhibited a diminished capacity to differentiate into ECs.
Western blot analysis revealed that before IFN-γ stimulation, HLA-A, B, and C were not expressed in ECs derived from YiP3 and clones A7 and B2. However, following IFN-γ stimulation, HLA-A, B, and C were expressed in YiP3, whereas only HLA-C was expressed in clones A7 and B2, which indicated that HLA-A and B were not expressed. This confirmed that clone A7 retained its differentiation capacity into ECs, and under IFN-γ stimulation, HLA-A and B were not expressed, whereas HLA-C, which was not the target of gene editing, was expressed.