Abstract Background Maple syrup urine disease (MSUD) is an inherited metabolic disorder caused by a deficiency in the activity of the hepatic branched-chain α-ketoacid dehydrogenase (BCKDH) complex, which leads to the toxic accumulation of three branched-chain amino acids (BCAAs) and their respective α-ketoacid, resulting in severe neurotoxicity, coma and even death without effective therapeutic measures. Methods In this study, we established the patient induced pluripotent stem cells (iPSC)-derived hepatic organoids (HOs), analyzed the characteristics, and applied adenine base editor (ABE8e) to correct a mutation (T322I) of the BCKDHB (branched chain keto acid dehydrogenase E1, beta polypeptide) gene in patient induced pluripotent stem cells (iPSC)-derived hepatic organoids (HOs). qRT-PCR and western blot analysis were performed to assess the expression level of BCKDHA (branched chain keto acid dehydrogenase E1, alpha polypeptide) and BCKDHB. The effects of base editing were comprehensively analyzed using both bulk RNA sequencing and single-cell RNA sequencing (scRNA-Seq). Results Immunofluorescence and RT-PCR arrayed the high expression of hepatoblast specific proteins in HOs, such as α-1-anti-trypsin (A1AT), hepatocyte nuclear factor-4-alpha (HNF4A), cytokeratin18 (CK18), albumin (ALB), cytochrome P450 family 3 subfamily A member 4 (CYP3A4) and cytochrome P450 family 3 subfamily A member 7(CYP3A7). Functional experiments indicated that these HOs recapitulated characteristics of hepatocytes like glycogen accumulation, low-density lipoprotein (LDL) uptake, indocyanine green (ICG) uptake and release as well as quantitation of ALB and urea from HOs. The levels of BCKDHA and BCKDHB were dramatically decreased in MSUD-HOs compared with control-HOs (P < 0.01) detected by qRT-PCR, western blot and immunofluorescence. Deep sequencing and whole genome sequencing (WGS) demonstrated that the correction of BCKDHB mutation in patient iPSC-derived HOs caused high on-target gene editing without any detectable off-target effects. Moreover, the corrected MSUD-HOs exhibited restored BCKDH enzymatic function and reduced BCAAs level. The transcriptome analysis indicates that the MSUD-HOs with BCKDHB mutation reduced mRNA level of regulating metabolism associated with liver mitochondrial function, while the corrected MSUD-HOs rescued those processes after ABE8e correction. The scRNA-Seq analysis further validated the rescue effects of BCKDH function after gene editing. Conclusion Our study provides reliable evidence that ABE8e is highly efficient and safe in correcting patient-derived HOs from MSUD, indicating the feasibility to be a transformative treatment for genetic hepatic diseases like MSUD. Graphical abstract [40]graphic file with name 13287_2025_4630_Figa_HTML.jpg Supplementary Information The online version contains supplementary material available at 10.1186/s13287-025-04630-w. Keywords: Base edit, Adenine base editor, Maple syrup urine disease, Human induced pluripotent stem cells (iPSCs), Hepatic organoids (HOs), RNA-sequencing, Single cell RNA sequencing, Whole genome sequencing Introduction Maple syrup urine disease (MSUD, OMIM #248600) first described in 1954, is an autosomal recessive metabolic disorder with about 1 in 185,000 of worldwide incidence, characterized by the toxic accumulation of three branched-chain amino acids (BCAAs) and their corresponding branched-chain α-ketoacids (BCKAs), which leads to severe neurotoxicity, coma and even death, in the lack of effective therapeutic measures [[41]1, [42]2]. MSUD resulted from the deficiency of the branched-chain α-ketoacid dehydrogenase (BCKDH) complex, which is a key enzyme catalyzing the irreversible and rate-limiting process in BCAAs catabolism, consisting of three catalytic subunits, namely E1 (a heterotetramer of 2E1α and 2E1β) coded by BCKDHA and BCKDHB genes respectively, E2 coded by the DBT gene, and E3 coded by the DLD gene [[43]3, [44]4]. The function of BCKDH is to convert BCKAs into acetyl-CoA and succinyl-CoA, entering the tricarboxylic acid (TCA) cycle. The genetic defects in one of four genes result in decrease or absent enzyme activity of the BCKDH complex causing substantial accumulation of BCAAs (including leucine, isoleucine, and valine) in plasma and corresponding BCKAs in urine [[45]5, [46]6]. MSUD can be classified into classic, intermediate, intermittent and thiamine-responsive subtypes, while approximately ~ 75% of MSUD patients tend to be the classic MSUD type that mostly was caused by BCKDHA and BCKDHB gene mutations, which is the most severe form of MSUD characterized by less than 2% residual enzyme activity [[47]2, [48]7]. Most infants with classic MSUD emerge subtle non-specific symptoms within 2–3 days, and further manifest neurological signs including hypertonia and spasticity progressing to seizures and coma even death [[49]1, [50]6]. Treatment options for the disease currently remain unsatisfactory, and liver transplantation, as an effective long-term treatment to restore BCKDH enzyme activity, has been applied to treat patients with classic MSUD but is subjected to the availability of ideal donors, high cost and potential risk of graft failures, while the development of novel therapeutic strategies faces challenges due to the shortage of a suitable MSUD model [[51]8]. Rodents and 2D human monolayer cultures have been the most common models used for modeling liver diseases to study biomolecular mechanisms and therapeutic strategy, but are poorly reliable because they could not thoroughly recapitulate the architecture and functions of human liver [[52]9, [53]10]. Organoids are in vitro cell culture systems exhibiting three-dimensional structures that mimic the key functional, structural and biological complexity of corresponding organs in vivo, which can be derived from pluripotent stem cells. Human induced pluripotent stem cells (iPSCs)-derived hepatic organoids (HOs) have recently been promising tools for a wide range of biomedical applications, from disease modeling of rare disorders to personalized medicine or cell therapy [[54]10–[55]12]. More recently, advances in genome editing technology have brought great promise for fundamentally eliminating the occurrence of genetic diseases. Base editing is a new genome editing technique that enables more precise, direct, and secures conversion of single-point mutations without introducing DNA double-stranded breaks (DSBs) generated by using CRISPR-Cas9 system [[56]13–[57]15]. Adenine base editors (ABEs), mediating A•T to G•C conversions, have shown remarkable on-target effects compared to those of CRISPR/Cas9 or cytosine base editors (CBEs), and are capable of correcting approximately 48% of pathogenic point mutations based on the ClinVar database, indicating tremendous potential for gene therapy of genetic diseases [[58]16, [59]17]. So far, ABEs have been utilized to correct genetic mutations in patient-derived iPSCs with high efficiency and the edited iPSCs could be differentiated into different target cells like mesenchymal stromal cells and hepatocytes [[60]18–[61]20]. At almost same time, Hans Clevers’ group first applied a ABE on four selected cystic fibrosis (CF) organoid samples, obtained genetic and functional repairmen in all four cases without detectable off-target effect, and later Amistadi et al. used a specific ABE (NG-ABEmax) to recover the CFTR function by correcting a splicing mutation which was validated in CF patient-derived rectal organoids and bronchial epithelial cells, indicating a prominent potential for treatment of genetic disorders in a clinical setting [[62]21, [63]22]. According to the Human Gene Mutation Database (HGMD), more than 280 pathogenic variants have been recorded in BCKDHA, BACKHB, and DBT genes and most variants with point mutational types were reported in the BCKDHB gene [[64]23, [65]24]. The ideal treatment for curing an autosomal recessive disorder like MSUD is the restoration of the pathogenic mutants to the wild-type on bi-alleles, and the correction of mutation on one allele could raise the BCKDH enzymatic activity that is sufficient to maintain normal condition. Hence, we hypothesized that base editors might provide a potential treatment for MSUD and tested the approach in vitro with iPSCs-derived hepatic organoids (HOs) from MSUD patients carrying pathogenic point mutations in the BCKDHB gene. By differentiating iPSCs derived HOs from the peripheral blood of the patient and performing gene editing directly using ABE8e, we rescued disrupted BCKDH enzymatic function and reduced BCAAs level with high editing efficiency, without any detectable off-target effects and toxicity. Our work provides a reliable proof that ABE8e correction for MSUD patient-derived HOs is a valuable strategy with high efficiency and safety to treat genetic hepatic diseases like MSUD. Materials and methods Peripheral blood-derived induced pluripotent stem cells (iPSCs) culture and maintenance Human iPSC lines of MSUD patient and healthy individual with matched sex and age involved in this study were generated from peripheral blood [[66]25, [67]26]. iPSCs were maintained on Matrigel (BD Biosciences) in ncTarget hPSC Medium (Nuwacell Biotechnologies Co., Ltd) and dissociated with Accutase cell detachment solution (Sigma-Aldrich) at a 1:6 split ratio at 37 °C in 5% CO[2]. Hepatic organoid generation The generation of hepatic organoids directly from the iPSCs by a combination of a series of growth factors has been described previously [[68]27]. Briefly, iPSCs were differentiated into endoderm in endoderm differentiation medium from day 0 (iPSCs) to day 3 (definitive endoderm, EN). For foregut (FG) induction from days 3 to 6, cells were cultured in the medium containing RPMI 1640 plus B27 (GIBCO) and FGF10 (Peprotech). For hepatoblast (HB), FGF10 and BMP4 were added to the medium from days 6 to day 9. After day 9, cells were inducted in the medium of organoid growth and differentiation (HCM, Lonza) including HGF, oncostatin M, and dexamethasone (Sigma-Aldrich). On day 20, the primary HOs (HO1) were formed and collected. For the secondary HOs (HO2), the HO1 were reseeded and Maintained in the growth medium, which consisted of RPMI 1640 plus B27 medium with the following growth factors: 250 nM LDN-193189 (iBMP), 3 µM CHIR99021, 10 µM A83-01 (iTGF), 100 ng/ml EGF, 10 ng/ml FGF10, and 20 ng/ml HGF. HOs were separated from Matrigel (BD 354230) by cell recovery solution (CRS, BD 354253) and dissociated by digestion with TrypLE™ Express (Gibco 12605010). Liver function assay LDL Uptake Cell-Based Assay Kit (Cayman) was used to analyze the LDL uptake of HO according to the manufacturer’s protocol. Glycogen storage of HO was determined by PAS staining kit (Solarbio) following the instructions. ICG uptake assay was performed by incubating 1 mg/ml ICG (Cayman) for 6 h in cell incubator and then washed with PBS. Images of ICG uptake and release were taken by a microscope. The supernatant of culture medium was corrected and the production of albumin proteins was measured using the Human Albumin ELISA Kit (Proteintech), according to the manufacturer’s protocol. Immunofluorescence Hepatic organoids were extracted from Matrigel with Cell Recovery Solution (Corning) in 4℃ for 1 h and fixed with 4% PFA for 30 min. Then, organoids were washed with PBS for three times and blocked with 3% BSA supplemented with 0.2% Triton X-100 at room temperature for 1 h. Primary antibodies were added and incubated at 4℃ overnight. After washing with PBS, secondary antibodies were treated to organoids at room temperature for 1 h and stained with DAPI for 15 min. Images were taken by Zeiss LSM 800. RT-PCR analyses Total RNA was extracted by TRIzol^® RNA Isolation Reagents (Ambion, Grand Island, NY), and 1 µg RNA was reverse-transcribed by the PrimeScript RT reagent Kit (Takara RR047A) according to the manufacturer’s instructions. The expression level of genes was calculated using the comparative threshold (2 − ΔΔCt) method. The results were normalized to the expression of GAPDH. Primer sequences are listed in Table [69]1. Table 1. Genome-Wide sequencing analysis for the Off-Target Total SNPs SNPs after Excluding dbSNPs Total InDels InDels after Excluding dbSNPs A > C A > G A > T C > A C > G C > T G > A G > C G > T T > A T > C T > G Possible Off-Target Site (1332 Sites) Control-1 3,510,051 23,274 508,247 52,844 994 3210 875 1449 1,136 3949 3,850 1192 1,407 990 3,229 993 - Control-2 3,545,912 23,848 533,748 58,132 1,037 3290 909 1509 1,182 4026 3,945 1210 1,455 981 3,280 1024 - Edited Organoid-1 3,508,195 23,591 518,927 57,324 1,011 3256 913 1449 1,163 4012 3,892 1203 1,440 994 3,271 987 0 Edited Organoid-2 3,508,837 23,639 510,872 53,373 1,042 3306 887 1482 1,174 3979 3,875 1184 1,407 991 3,331 981 0 [70]Open in a new tab Edited organoids-1 & edited organoids-2 represent that organoids were nucleofected with ABE8e and sgRNA plasmid, and control-1 & control-2 represent that organoids were nucleofected with ABE8e but without sgRNA Western blot analysis The iPSCs or HO2 were lysed with lysis buffer (RIPA with 0.1% cocktail). The protein concentration of the cell lysate was measured by BCA kit (Beyotime). Samples with 30ng were loaded into wells of an 8% SDS–polyacrylamide running gel in running buffer at 80 V for 30 min and then turned into 120 V for 1 h. Proteins were transferred by running at 90 V for 1 h in transfer buffer with the polyvinylidene difluoride membrane. After the membranes blocked with 5% skim milk in PBS containing 0.05% Tween-20 for 1 h, primary antibody diluted with 5% skim milk in PBST were incubated at 4℃ overnight. The membranes were then washed with PBST for 3 times, and incubated with secondary antibody for 1 h at room temperature. Signals were detected using ChemiScope Series 5300 (Clinx Science Instruments Co., Ltd). Branched chain amino acid assay The branched chain amino acids (leucine, isoleucine, and valine) of supernatant were assayed with Branched Chain Amino Acids (BCAAs) Assay Kit (Sigma-Aldrich) according to the manufacturer’s instruction. Briefly, 50 µl supernatant was added to a 96-well plate. Next, 46 µl of assay buffer was added to each well, along with 2 µl of branched-chain amino acid enzyme mix and 2 µl WST substrate mix for a final volume of 100 µl per well. Blanks were performed by mixing lysate with BCAAs assay buffer and WST substrate mix but omitting enzyme mix. BCAAs concentration was measured using a coupled enzyme reaction, which results in a colorimetric (450 nm) product, proportional to the BCAAs present. Plasmid construction The plasmids ABEmax-NG, ABE8e and pGL3-U6-sgRNA-PGK-puromycin were ordered from Addgene ([71]https://www.addgene.org/, #124163, #138489 and #51133). ABE8e-NG involved in this study was constructed by introducing the amplified Cas9 (N) from ABEmax-NG plasmid to the backbone of ABE8e using ClonExpress II One Step Cloning Kit (Vazyme, C112-01). The target-specific sgRNA re-BCKDHB-c.965C > T sequence containing the DNA oligo (5’-GCCCTATTTT GATCACAGCA-3’) for constructing sgRNA expression vectors were synthesized, annealed, and cloned into BsaI-digested pGL3-U6-sgRNA- PGK-puromycin. Base editing HOs ABE8e was applied for the correction with NG as PAM sequence and the target site was at position 6 within the editing window. At the 35 day of HO2, organoids were collected and dissociated into a single-cell suspension using CRS and TrypLE. The single cells were electroporated the plasmids of ABE8e and pGL3-U6-sgRNA–PGK-puromycin with the 4D Nucleofector X-unit System (Lonza) by using the P3 Primary Cell Kit with the optimization scheme (DN100) provided for 4D Nucleofector systems. The quantity of plasmids of ABE8e and sgRNA used in this study was 7.2 and 2.25ug, respectively. Then cells were resuspended with Matrigel and formed as dome every 50 µl which seeded on 12-well plates. One day after transfecting the base editor vectors, puromycin with a concentration of 1 µg/ml was added to the plated wells and incubated for 24 h allowing the selection of positive cells grown into single organoid structures. The culture medium containing puromycin was then removed and the growth medium was added for further culture. Sanger sequencing for BCKDHB targeted amplicon Organoids were separated from Matrigel by CRS and washed with DPBS. DNA was extracted from organoids by TIANamp Genomic DNA Kit (TIANGEN, Beijing, China) following the manufacturer’s instructions. Primers were constructed to amplify the region containing the identified mutation in BCKDHB. PCR reactions were performed with primers and Taq Master Mix (CWBIO, Beijing, China) for 35 cycles (94 °C for 30 s, 58 °C for 30 s and 72 °C for 30 s). Sanger sequencing was carried out on a 3130XL DNA Analyzer (Applied Biosystems, Foster City, CA, USA). Editing efficiency was analyzed using EditR software ([72]http://baseeditr.com). Off-target analysis and deep sequencing The hepatic organoids were harvested for off-target analysis at the 10th day after transfection. The on-target and potential off-target sites were predicted by an online algorithm Cas-OFFinder, 1 on-target site and 20 off-target sites with up to 5-nt mismatch for the correctional sgRNA were amplified for 3 corrected and 3 control samples (cells that were nucleofected without sgRNA) using the related primers (Table [73]S1). The purified PCR products were sequenced on Illumina platforms with PE150 strategy in Novogene Bioinformatics Technology Co., Ltd (Beijing, China). Approximately 1 G clean data were generated for each sample. The Genome Center platform (National Human Genetic Resources Center) was used to process the data. Whole-genome sequencing (WGS) WGS analysis was performed on ABE8e treated and untreated hepatic organoids to search for base editing induced SNPs and InDels variants. A total amount of 0.2 µg DNA per sample was used as input material for the DNA library preparations and the sequencing (depth of 30×) was performed on Illumina HiSeq X Ten (2 × 150 PE) at the Novogene Bioinformatics Technology Co., Ltd (Beijing, China). Valid sequencing data was mapped to the reference genome (GRCh37/hg19) by Burrows Wheeler Aligner (BWA) software to get the original mapping result in BAM format. SAMtools were used to do variant calling and identify SNPs and InDels. The common SNPs between the corrected and the control organoids were excluded depending on the database of single nucleotide polymorphisms (dbSNPs). RNA-seq Total RNA was isolated using TRIZOL method following the protocol recommended by the manufacturer (Life Technologies, USA). The sample purity was assayed using a NanoPhotometer^® spectrophotometer (Nanodrop Technologies Inc. DE, USA). Agilent 2100 Bioanalyzer (Agilent, Palo Alto, CA, USA) was used to detect the RNA integrity and concentration. The whole library preparation was completed by RT-PCR. Samples were sequenced according to the product protocol using the VAHTS Universal V6 RNA-seq Library Prep Kit for Illumina (NR604-01/02) following the manufacturer’s recommendations. The libraries were sequenced on the Illumina sequencing platform (Novaseq6000) generating 150 bp paired end reads. The data was analyzed after sequencing. The genomes and the annotation file as reference were obtained from ENSEMBL database. Genome index was built using Bowtie2 v2.2.3, and Clean Data was then aligned to the reference genome using HISAT2 v2.1.0. Gene expression was calculated by RPKM (Reads Per Kilobase per Million mapped reads). Gene ontology (GO) and KEGG (Kyoto Encyclopedia of Genes and Genomes database) pathway analysis was carried out with DEGs. scRNA-seq HOs of P1 and P1-corr were collected and separated from Matrigel by CRS at 4 °C for 30 min. Then CRS was abandoned, and HOs were dissociated to single cells by Triple at 37 °C for 10 min. Single-cell suspension was washed with 1× PBS (0.04% BSA), filtered through 70 μm nylon strainer (BD Falcon), collected by centrifugation (330×g, 10 min, 4 °C) and resuspended in 1× PBS (0.04% BSA). Cells are manually counted three times by Trypan blue exclusion after each centrifugation and resuspended at a concentration of ≥ 2 × 10^6/ml. Single cells were run on a 10× Chromium system (10× Genomics) and then through a Library preparation by LC Sciences, following the recommended protocol for the Chromium Single Cell 30 Reagent Kit (v3 Chemistry). Libraries were run on the HiSeq4000 for Illumina sequencing. Postprocessing and quality control were performed using a 10× Cell Ranger package (v7.2.0; 10× Genomics). Reads were aligned to the homo sapiens reference assembly (Ensembl v105). Primary assessment with the 10× Cell Ranger for the sample treated with saline reported 12,876 − 13,823 cell-barcodes with 1,656–2,155 median genes per cell sequenced to 24.1% − 21.3% sequencing saturation with 30,005–30,689 mean reads per cell. Primary assessment with this software for the sample treated reported 7,213–8,769 cell-barcodes with 2,131–2,305 median genes per cell. scRNA-seq was conducted by LC-bio (Hangzhou, China). Briefly, Gene expression Matrices generated by the 10× Cell Ranger aggregate option were further analyzed with R package Seurat (version 3.0) with default parameters. The data were filtered according to the following thresholds: greater than 500 as unique expressed genes (nFeature_RNA) and greater than 25% as the per- centage of mitochondrial genome content. The data were then normalized by converting with a scale factor (default as 10,000) and log-transformed with the Seurat embedded function. A correlation analysis was performed by employing the RunPCA function of the Seurat package, followed by an integrated analysis of the three datasets. Clustering analysis was carried out with standard Seurat package procedures with a resolution at 0.8. The identified clusters were then visualized using t- distributed Stochastic Neighbor Embedding (UMAP) of the principal components in Seurat. Average gene expression matrixes were then retrieved for each cluster, and differential expression among clusters was performed to identify the top markers at a high level by each cluster with the FindAllMarkers implemented function (parameters: only.pos = FALSE, min.pct = 0.2, thresh.use = 0.2). Statistics Statistical analyses were carried out. Unpaired two-tailed Student’s tests or one-way ANOVA were used to compare two or more groups. Data were presented as mean ± SD. P-values < 0.05 were considered to indicate statistical significance. More detailed information has been provided in each figure legend which is annotated as follows: *p < 0.05, **p < 0.01, ***p < 0.001. Results MSUD-HOs derived from iPSCs of MSUD patients presented disrupted BCKDH and metabolic phenotypes The establishment and characterization of iPSCs derived from MSUD patients and heathy donors have already been reported in our previous studies [[74]28, [75]29]. Then, we generated and characterized hepatic organoids (HOs) from iPSCs line derived from the MSUD patient (P) caused by BCKDHB mutations. Experimental control HOs (C) were from a wild type healthy individual with Matched sex and age. Both HOs presented same expression patterns of cell Markers. The representative analysis of cell markers during the differentiation process from iPSCs to HOs at the main stages on day 0 (iPSCs), day 3 (definitive endoderm, EN), day 6 (foregut, FG), day 9 (hepatoblast, HB), day 20 (primary HOs, HO1) and day 28 (secondary HOs, HO2) were presented in Figure [76]S1A. Further, qRT-PCR verified the expression of key genes in the differentiation of HOs (see Figure [77]S1B). Immunofluorescence and RT-PCR detected high expression of hepatoblast specific proteins, such as α-1-anti-trypsin (A1AT), hepatocyte nuclear factor-4-alpha (HNF4A), CK18, albumin (ALB), CYP3A4 and CYP3A7 (Fig. [78]1A). Western blot analysis of hepatoblast specific proteins including A1AT, CK18, CYP3A4, HNF4A and ALB were presented in Figure [79]S3. Functional experiments indicated that these HOs recapitulated characteristics of hepatocytes like glycogen accumulation (PAS staining, Fig. [80]1B), LDL uptake (Fig. [81]1C), ICG uptake and release (Fig. [82]1D) as well as quantitation of ALB (Fig. [83]1E) and urea (Fig. [84]1F) from HOs, which were illustrated in Fig. [85]1. Fig. 1. [86]Fig. 1 [87]Open in a new tab Human iPSCs-derived hepatic organoids (HOs) recapitulate morphological and functional features of human liver. A Immunostaining of HOs for A1AT, HNF4A, CK18 and ALB. The nuclei are counterstained with DAPI (scale bar, 100 μm). Functional experiments show Periodic Acid Schiff (PAS) staining (B) (scale bar, 50 μm), LDL uptake (C) (scale bar, 100 μm), and the ability to uptake and release indocyanine green (ICG) (D) (scale bar, 200 μm). E Quantification of albumin (ALB) secretion in HOs is showed after 24 h compared with culture medium. F Urea production of HOs is compared with culture medium. G Relative expression levels of liver specific markers: ALB, CK18, HNF4A and CYP3A4 in control (C) and patient (P) Hos The levels of BCKDHA and BCKDHB were dramatically decreased in MSUD-HOs compared with control-HOs (P < 0.01) detected by qRT-PCR (Fig. [88]2A, B), Western blot (Fig. [89]2D, E, F) and Immunofluorescence (Fig. [90]2G). Compared with the control-HOs (C), the BCAAs level in MSUD-HOs (P) was dramatically increased (P < 0.001) (Fig. [91]2C), demonstrating that the disruption of BCKDHA and BCKDHB decreased the enzymatic function of BCKDH, thereby increasing BCAAs level and causing toxicity in MSUD patients, indicating the HOs can be used as a successful in vivo model for MSUD study. Fig. 2. [92]Fig. 2 [93]Open in a new tab The MSUD patient hepatic organoids (P) present typical metabolic phenotypes of MSUD. A Relative lower expression level of BCKDHA in patient HOs (P) than that in control HOs (C) by qRT-PCR analysis. **P < 0.01. B Relative lower expression level of BCKDHB in patient HOs (P) than that in control HOs (C) by qRT-PCR analysis. **P < 0.01. C The higher levels of BCAAs in patient HOs (P) than that in control HOs (C). ***P < 0.001. D Expression levels of BCKDHA and BCKDHB in patient HOs (P) were compared with the control HOs (C) and both groups of iPS cells by Western blot. E Relative protein expression of BCKDHA in patient HOs (P) were compared with that in control HOs (C) and both groups of iPS cells. **P < 0.01. F Relative protein expression of BCKDHB in patient HOs (P) were compared with that in control HOs (C) and both groups of iPS cells. **P < 0.01, ***P < 0.001. G The decreased levels of BCKDHA and BCKDHB in patient HOs (P) were compared with those in control HOs (C) by immunofluorescence staining. The nuclei are counterstained with DAPI (scale bar, 100 μm) Adenine base editor corrected the BCKDHB mutation in the MSUD-HOs ABE8e was applied for the correction. After electroporation and puromycin treatment, only 4 individual organoids clones were obtained, one of which was discarded as it no longer grew. Sanger sequencing was then carried out to estimate the editing efficiency of those edited organoids (Fig. [94]3A). We observed that a complete reversal at the mutation of c.965 C > T in one of HOs (P-corr 1), the proportion of the C allele became 100% and that of mutated T allele decreased to 0% with an editing efficiency of 100% in P-corr 1. The editing efficiency of the other two organoids (P-corr 2, P-corr 3) was 62% and 84%, respectively (Fig. [95]3B, C). Thus, we chose the P-corr 1 for the subsequent experiments. Fig. 3. [96]Fig. 3 [97]Open in a new tab Adenine Base-Editing-mediated functional repair of BCKDHB on Canonical PAM (NGG). AThe location of the mutation c.965 C > T(T322I) on the BCKDH gene. B Bright image of three hepatic oranoids after base editing and puromycin treatment. C Sanger validation of BCKDHB mutation c.965 C > T(T322I) repair using ABE8e in three clonal HOs. D The editing efficiency on c.965 C > T was analyzed by Edit R and highlighted in red No detectable off-target effect for the mutation of c.965 C > T/pT322I in the MSUD-HOs Deep sequencing was first performed to identify the off-target sites the off-target sites and confirmed the results above (Fig. [98]4A and B). We further performed WGS to assess the off-target effects for corrected and control organoids. Totally 1332 potential off-target sites predicted by Cas-OFFinder with up to 5 mismatches were analyzed. There were no off-target sites detected after excluding the false-positive sites depending on the database of single nucleotide polymorphisms (dbSNPs) (Table [99]1; Fig. [100]4C, D). These results demonstrated the safety of ABE8e correction for the pathogenic mutation in iPSC-derived HOs. Fig. 4. [101]Fig. 4 [102]Open in a new tab Off-Target analysis of the edited and control organoids by deep-sequencing and WGS. A Cas-OFFinder identified off-target (OT) sites in the BCKDHB c.965 C > T sgRNA and associated genomic location. B Deep-sequencing for on-target editing at the BCKDHB c.965 C > T site and for bystander edits within the editing window. C and D Numbers of total SNPs (C) and indels (D) identified in edited and control organoids by WGS. Letters in lowercase represent mismatches between the on and off target sites, the underlined bases in the sequences are those capable of being edited by ABE8e and the PAM is highlighted in blue. Data are from three edited samples and two control samples. All values represent means ± SD ABE8e rescued BCKDH function and reduced BCAA toxic accumulation in corrected MSUD-HOs Next, we performed expression analysis of representative hepatoblast markers and functional assay of corrected MSUD-HOs by immunofluorescence, which indicated consistent characteristics of hepatocytes as that of control HOs (Fig. [103]5). To further determine the restored effect, mRNA levels of BCKDHA and BCKDHB were detected using qRT-PCR and showed significant elevation in corrected MSUD-HOs (P-corr) compared with MSUD-HOs (P) (P < 0.01) which was close to the level of control HOs (Fig. [104]6A, B). Moreover, western blot analysis showed that expression levels of E1α and E1β subunits coded by BCKDHA and BCKDHB in corrected MSUD-HOs (P-corr) were higher than those in MSUD-HOs (P) (Fig. [105]6D, E, F). The rescued effect of BCKDHA and BCKDHB in the corrected MSUD-HOs (P-corr) was also confirmed using immunofluorescence staining in comparison with that of the MSUD HOs (P) (Fig. [106]5G). We further analyzed the BCAAs levels in both MSUD-HOs (P) and corrected MSUD-HOs (P-corr) and found the BCAAs concentrations in corrected MSUD-HOs (P-corr) were dramatically decreased compared to MSUD-HOs (P < 0.05, Fig. [107]6C), demonstrating that function of E1α and E1β subunits was rescued and thereby reduced BCAAs toxic accumulation. Fig. 5. [108]Fig. 5 [109]Open in a new tab Corrected MSUD-HOs (P-corr) by ABE8e show morphological and functional similarities to the liver. A Immunostaining of HOs for A1AT, HNF4A, CK18 and ALB. The nuclei are counterstained with DAPI (scale bar, 100 μm). Functional experiments show the ability to uptake and release indocyanine green (ICG) (B) (scale bar, 200 μm), Periodic Acid Schiff (PAS) (C) staining (scale bar, 50 μm) and LDL uptake (D) (scale bar, 100 μm). (E) Quantification of albumin (ALB) secretion in P-corr is showed after 24 h compared with P (patient MSUD-HOs). F Urea production of P and P-corr are presented Fig. 6. [110]Fig. 6 [111]Open in a new tab ABE8e corrected MSUD-HOs (P-corr) are protected from disease signature. A qRT-PCR analysis of BCKDHA expression in P-corr compared with that in HOs of P and C. **P < 0.01. B qRT-PCR analysis of BCKDHB expression in P-corr compared with that in HOs of P and C. **P < 0.01. C The declined levels of BCAAs in P-corr compared with that in HOs of P and C. **P < 0.01, ***P < 0.001. D Western blot of BCKDHA and BCKDHB levels in P-corr compared with those in HOs of P and C. E Relative protein expression of BCKDHA in P-corr compared with that in HOs of P and C. **P < 0.01. F Relative protein expression of BCKDHB in P-corr compared with that in HOs of P and C. **P < 0.01. G The increased levels of BCKDHA and BCKDHB in P-corr compared with those in HOs of P by immunofluorescence staining. The nuclei are counterstained with DAPI (scale bar, 100 μm) RNA-seq demonstrating distinct transcriptomic profile in the corrected HOs We further investigated corrected effects to gain an in-depth understanding of transcriptome level in MSUD-HOs after correction and observed 18,284 differentially expressed genes (DEGs; fold change > 2), in which 2,126 genes were up-regulated and 3,635 genes were down-regulated (Fig. [112]7A and B). Canonical pathway enrichment analysis showed that the changes of genes were significant after ABE8e correction and the number of differentially expressed genes (about 2000 genes) closely related to metabolic pathways was the highest, followed by organic substance metabolic process (about 1900 genes) and cellar metabolic processes (about 1800 genes) (Fig. [113]7C), as well as localization, regulation of biological quality, transport, positive regulation of cellular metabolic process, biosynthetic process and so on. KEGG pathways analysis is closely related to metabolic pathways, motor proteins, cell cycle, oxidative phosphorylation, and so on (Fig. [114]7D). Gene set enrichment analysis (GSEA) on differentially expressed genes showed the enrichment specifically in genes of drug metabolism-cytochrome P450, intracellular cholesterol transport, regulation of autophagosome assembly, regulation of toll-like receptor 4 signaling pathway, and positive regulation of tissue remodeling (Fig. [115]7E). Thus, the transcriptome analysis indicates that the MSUD-HOs with BCKDHB mutation reduced mRNA level of regulating metabolism associated with liver mitochondrial function, while the corrected MSUD-HOs rescued those processes after ABE8e correction. Fig. 7. [116]Fig. 7 [117]Open in a new tab Global transcriptome functional analysis of the edited and patient organoids. A Venn diagram showed genes changed in the edited and patient organoids. (P < 0.05). B Volcanic map showed genes changed in the edited and patient organoids. (P < 0.05). C Canonical pathway enrichment analysis of the genes showed significant changes in the edited and patient organoids. D KEGG analysis showed enrichment of canonical pathways the edited and patient organoids. E GSEA showed the enrichment in genes of in drug metabolism-cytochrome P450, intracellular cholesterol transport, regulation of autophagosome assembly, regulation of toll-like receptor 4 signaling pathway, and positive regulation of tissue remodeling Sc RNA seq affirmed the rescue of base editing A multiplex analysis of scRNA-Seq data obtained from the MSUD-HOs (P) and corrected MSUD-HOs (P-corr) was performed. As a whole, UMAP plot of 26,699 cells in P organoids and P-corr organoids showed that dimensionality reduction clustering generated 13 clusters (Fig. [118]8A) and each of the different cell types was tightly clustered. Two groups of cells were annotated as four types of cells, including hepatocytes (clusters 0, 5, 9, 11), fetal hepatocytes (cluster 7), epithelial cells (cluster 4) and cholangiocytes (cluster 1, 2, 3, 6, 8, 10, 12) using functional gene localization method (Fig. [119]8B) and the proportion of different cell types in both groups were showed via stacking diagram (Fig. [120]8D). The cell subtypes in the hepatic organoids were confirmed by immunofluorescence staining and presented in Figure S4. The expression of three representative marker genes in each cell type was demonstrated using dot plot (Fig. [121]8C). GO and KEGG enrichment analysis compared the differential genes in both groups and a large number of up-regulated genes were found to be associated with cytoskeleton construction and molecular metabolic functions in P-corr group (Fig. [122]8E, F, and Figure [123]S2). The expression heatmap showed that the expression levels of key genes involved in BCKDH and BCAAs metabolism were significantly changed after gene correction, such as BCKDHA, BCKDHB and DBT encoding BCKDH subunits, BCAT1 and BCAT2 key enzymes in the initial step of BCAAs metabolism, which were increased in the P-corr group; while ECHS1, a key enzyme promoting BCAAs catabolism, reduced expression in the P-corr group (Fig. [124]8G). Therefore, our scRNA-Seq analysis further validated the rescue effects of BCKDH function after gene editing. Fig. 8. [125]Fig. 8 [126]Open in a new tab Differences between liver organoid samples as detected by single-cell RNA sequencing A. UMAP plot of 26,699 cells in 2 specimens shows that dimensionality reduction clustering generated 13 clusters; B. Two samples were annotated as four types of cells (P group on the left and P-corr group on the right) using functional gene localization method; C. Dot plot showing the expression of three representative marker genes in each cell type; D. Stacking diagram shows the proportion of different cell types in two groups; E,F. Gene enrichment analysis of scRNA-seq dataset in two groups comparing the differences before and after gene editing; A large number of up-regulated genes in the repair group were found to be associated with molecular metabolic functions in the KEGG database, as well as with cytoskeleton construction in the GO database; G. Expression heatmap shows key genes involved in branched chain amino acid metabolism, with higher expression levels observed in the P-corr group; H. UMAP plot displays the expression levels of key genes involved in branched chain amino acid metabolism in various cell types Discussion In this study, we applied an ABE8e to correct the pathogenic point mutation in the BCKDHB gene of iPSCs-derived hepatic organoids (HOs) from a MSUD patient, thereby rescuing BCKDH function and reducing BCAAs toxic accumulation with a high editing efficiency of 100%, without any detectable off-target effects and toxicity. To the best of our knowledge, this is the first report of gene-edited iPSCs-derived HOs from MSUD patient and has great potential for efficient and safe personalized treatment of genetic and metabolic diseases like MSUD. Genetic and metabolic diseases are important indicators of liver transplantation in children [[127]30]. With the advent of next generation sequencing technology, more inherited liver diseases can be diagnosed earlier than ever before, so prompt treatment becomes even more critical to save lives and alleviate symptoms in extra-hepatic organs, especially the nervous system [[128]31, [129]32]. Classic MSUD is the most common and severe type, which is fatal in the neonatal period, so liver transplantation is often needed to prevent MSUD from developing life-threatening metabolic crises [[130]1, [131]6, [132]33]. Human fetal liver has long been considered the most promising and available source of transplantation for liver diseases, but availability is limited for ethical and practical reasons [[133]34]. Human iPSCs-derived HOs can recapitulate key features of hepatic organogenesis and function in vitro holding great potential for personalized medicine and liver transplantation [[134]35–[135]38]. In this study, we generated and characterized iPSCs-derived HOs from a MSUD patient caused by BCKDHB mutations and observed typical liver tissue-like characteristics along with disruption of the BCKDH enzyme and accumulation of BCAAs, which recapitulated the phenotype of MSUD and thereby could be a reliable in vitro model for further studying personalized therapy of MSUD. ABE8e, a recently evolved adenine base editor mediating AT to GC conversions, contains eight additional mutations compared with ABE7.10 and has exhibited the expanded targeting scope, elevated editing efficiency and overall utility of ABEs [[136]39–[137]41]. In the present study, we chose the theoretically editable mutation site of c.965 C > T (p.T322I) in the BCKDHB gene by ABEs, and applied ABE8e to directly correct the mutation in iPSCs-derived MSUD HOs with the optimization scheme provided for the 4D Nucleofector system. After screening and identification, we found that the c.965 C > T mutation site was reverted to the wild type, achieving an editing efficiency of up to 100% for this mutation. Moreover, the corrected MSUD-HOs significantly increased the expression of E1α and E1β subunits encoded by BCKDHA and BCKDHB, rescued BCKDH function and reduced BCAAs accumulation, indicating the correcting one allele of BCKDHB sufficient to maintain the BCKDH function in iPSCs-derived MSUD HOs, which is consistent with previous reports [[138]3, [139]7]. Off-target effect in gene editing, which refers to unintended changes made to the DNA sequence or even adverse alterations to the genome, is often sgRNA-dependent and a major concern [[140]42]. To detect the sgRNA-dependent off-target effect of the base editor, we used an in silico online software (Cas-OFFinder) to predict potential off-target sites and performed off-target sites specific PCR for deep sequencing [[141]43]. Whole genome sequencing (WGS) is an unbiased approach that is widely used to detect off-target effects across the genome in vivo, while it has sensitivity problem [[142]44–[143]46]. In our study, no detectable off-target sites were found in the genome by using WGS and deep-sequencing analysis, manifesting the potential for safe application of individualized therapy. To further analyze the changes in gene expression, we carried out RNA-seq and scRNA-seq, and observed remarkable changes in gene expression of MSUD-HOs after ABE8e gene correction, including different pathways associated with the metabolism of BCKDH and BCAAs as well as Liver mitochondrial function. BCKDH enzyme is situated on the inner mitochondrial membrane in various tissues, such as skeletal muscle, Liver, kidney and brain, while it demonstrates the highest activity in liver that is responsible for the expression of 10% BCKDH and catabolism of 10% BCAAs [[144]1, [145]47]. Some researchers focused on the relationship of BCAAs and metabolic homeostasis and mitochondrial function in liver, proposing that the increase of BCAAs could cause a variety of metabolic disorders and mitochondrial dysfunction [[146]48–[147]50]. It has been known that Liver mitochondria play a key role in fatty acid metabolism, drug metabolism, oxidative phosphorylation, intracellular cholesterol transport, ferroptosis, autophagy, toll-like receptor 4 (TLR4) as well as energy homeostasis and cellular processes [[148]25, [149]26, [150]51–[151]56]. Consistently, the RNA-seq result in our HOs in vitro model reconfirmed that the above pathways were impaired in MUSD-HOs and rescued in corrected MSUD-HOs. The scRNA-seq analysis also indicated that correction of the BCKDHB mutation in MSUD-HOs not only rescued the expression of BCKDHB and BCKDHA, but increased the level of mRNAs involved in BCKDH and BCAAs associated metabolic pathways. The limitations of this study are the lack of large samples and in vivo tests. First, due to PAM sequences and editing windows of ABE8e, only one mutated site of BCKDHB was selected for genome editing. Second, delivery platforms for ABE8e-mediated correction of BCKDHB mutations require systematic screening and in vivo development. Conclusion: In conclusion, our study provides proof-of-principle that ABE8e correction on the MSUD-HOs can rescue the BCKDH function and reduce BCAAs toxic accumulation with high on-target gene editing efficiency without any detectable off-target effects, indicating potential to be a transformative treatment for genetic hepatic diseases like MSUD. Supplementary Information Below is the link to the electronic supplementary material. [152]Supplementary Material 1.^ (849.4KB, tif) [153]Supplementary Material 2.^ (248.6KB, tif) [154]Supplementary Material 3.^ (522.4KB, tif) [155]Supplementary Material 4.^ (795.9KB, tif) [156]Supplementary Material 5.^ (21.3KB, docx) Acknowledgements