Skin specimens were obtained from two lar gibbons (H. lar) [GAIN-ID: 0023 (Jessica, a 24-year-old female) and GAIN-ID: 0247 (Kyutaro, a 21-year-old male)], one Abbott’s gray gibbon (H. abbotti) [GAIN-ID: 0159 (Anna, an estimated 46-year-old female)], and two siamangs (S. syndactylus) [GAIN-ID: 0105 (Matsu, an estimated 32-year-old female) and GAIN-ID: 0297 (Peach, an 11-year-old female)]. These samples were provided by the Japan Monkey Centre, Toyohashi Zoo and Botanical Park, and the Great Ape Information Network (GAIN) for primary cultures. Each skin specimen was minced into small pieces and transferred onto tissue culture dishes (353,002; Corning, Glendale, AZ, United States) with 15% fetal bovine serum (FBS)/Dulbecco’s modified Eagle’s medium (DMEM), which consists of DMEM with high glucose (044-29765; Wako, Osaka, Japan) supplemented with 15% FBS, 1× non-essential amino acids (139-15651; Wako), 1× GlutaMAX Supplement (35050061; Gibco, Waltham, MA, United States), 1 mM sodium pyruvate (190-14881; Wako), 0.11 mM 2-mercaptoethanol (21,985-023; Gibco), and 100 units/mL penicillin and 100 μg/mL streptomycin (168-23191; Wako). Fibroblast outgrowth from the skin pieces was observed after approximately 1 week of culture. The fibroblasts were subsequently passaged though single-cell dissociation with 0.25% trypsin/ethylenediaminetetraacetic acid (EDTA) and expanded in 15% FBS/DMEM medium on 0.1% gelatin-coated culture dishes at 37°C with 5% CO₂.

Small ape skin fibroblasts (passage number 2–4) were seeded at 1.0 × 10⁵ cells/well on 0.1% gelatin-coated culture plates. The following day, the cells were transduced with six human reprogramming factors (OCT3/4, KLF4, SOX2, c-MYC, LIN28, and NANOG) using SRV iPSC-3 (S1011626A; Tokiwa Bio, Tsukuba, Japan) at a multiplicity of infection (MOI) of three for GAIN-ID: 0247, 0159, 0105, and 0297. For GAIN-ID: 0023, transduction was performed with four human reprogramming factors (OCT3/4, KLF4, SOX2, and c-MYC) using the SRV iPSC-1 (S1011624A; Tokiwa Bio) at an MOI of 3. The transduction procedure involved incubating the cells with the respective SRVs at room temperature for 2 h, followed by overnight incubation at 37°C with 5% CO₂. Twenty-four hours after transduction, the cells were dissociated with 0.25% trypsin/EDTA and reseeded onto iMatrix-511-coated 24-well culture plates at 1.0 × 10⁴ cells/well. The medium was changed every other day under two different conditions. For condition A, cells were cultured in 15% FBS/DMEM medium from day 1 to day 7, followed by StemFit AK02N medium (AK02N; Ajinomoto, Tokyo, Japan) without supplements from day 7 to approximately day 21. Thereafter, the cells were maintained in StemFit AK02N medium. For condition B, StemFit AK02N medium without supplements was used from day 1 to day 7 or 14, followed by continued culture in StemFit AK02N medium. Emerging EGFP⁺ reprogrammed colonies were manually picked and cultured in StemFit AK02N medium on iMatrix-511-coated culture plates.

To remove SRVs from expanded EGFP⁺ reprogrammed cells, small interfering RNA (siRNA) transfection was performed in accordance with the manufacturer’s instructions (S1011626A; Tokiwa Bio). Briefly, EGFP⁺ reprogrammed cells, after several serial passages, were dissociated with TrypLE Express (12604021; Gibco) and seeded at 0.75 × 10⁴ cells/well in StemFit AK02N medium supplemented with 10 μM Y-27632 (036-24023; Wako) on iMatrix-511-coated 24-well culture plates. The next day, the medium was replaced with fresh StemFit AK02N medium. The cells were transfected with siRNA (sense: CAAUAGUUCACGCUGAAAGUG; anti-sense: CUUUCAGCGUGAACUAUUGCU) using Lipofectamine RNAiMAX (13778100; Thermo Fisher Scientific, Waltham, MA, United States) on days 1 and 4 after plating. The siRNA-transfected cells were passaged upon reaching sub-confluency and maintained in StemFit AK02N medium on iMatrix-511-coated culture plates. For subcloning SRV-free iPSCs, the siRNA-transfected cells were seeded at 0.5 × 10³ cells/cm², and EGFP⁻ colonies were manually picked up and expanded as iPSCs in StemFit AK02N medium on iMatrix-511-coated culture dishes.

Small ape iPSCs were maintained in StemFit AK02N medium on iMatrix-511-coated culture dishes. The medium was refreshed with StemFit AK02N medium on days 1, 3, 4, and 5 following passage. On day 6, the iPSCs were dissociated with TrypLE Express and reseeded at 1.4 × 10³ cells/cm² onto iMatrix-511-coated culture plates in StemFit AK02N medium supplemented with 10 μM Y-27632.

Metaphase spreads of each small ape iPSC line were prepared following the standard procedure (Tanabe et al., 2000). Briefly, KaryoMAX Colcemid Solution (15210-040; Gibco) was added to iPSC cultures at a final concentration of 0.05 μg/mL for 1.5–3 h. The harvested cells were treated with a hypotonic solution of 0.075 mol/L KCI containing 0.1% (w/v) trisodium citrate for 22 min at 37°C, followed by fixation with Carnoy’s fixative. The cell suspensions were then dispensed onto glass slides using a HANABI metaphase spreader (ADSTEC), air-dried, and stored in a freezer until further analyses. Chromosome numbers and structural rearrangements were evaluated using the QFH-banding technique (Yoshida et al., 1975), which involves staining with Quinacrine mustard and Hoechst 33258. Over 20 QFH-banded metaphase images were captured and analyzed for each iPSC line using a Leica DM5000B fluorescence microscope equipped with a CCD camera and CW4000 image analysis system (Leica Microsystems). Chromosomes of the lar gibbon (H. lar) and siamang (S. syndactylus) were identified and assigned based on the Atlas of Mammalian Chromosomes (Graphodatsky et al., 2020) and related references (Jauch et al., 1992; Capozzi et al., 2012; Stanyon, 2013).

Total RNA was extracted with RNeasy Plus Mini Kit (74,104; Qiagen, Hilden, Germany) and reverse-transcribed into cDNA with the PrimeScript FAST RT Reagent Kit with gDNA Eraser (RR092A; TaKaRa, Kusatsu, Japan). RT-PCR analyses were performed semi-quantitatively with ExTaq Hot Start Version (RR006A; TaKaRa). Quantitative RT-PCR analyses were performed with THUNDERBIRD Next SYBR qPCR Mix (QPS-201; Toyobo, Osaka, Japan) on a StepOne Real-Time PCR System (Applied Biosystems, Foster City, CA, United States). Primer sequences are provided in Supplementary Table S1.

Total RNA was extracted using ISOGEN (317-02503; Nippon Gene, Tokyo, Japan) and Direct-Zol RNA Miniprep (R2050; Zymo Research, Irvine, CA, United States). cDNA library preparation and sequencing were performed by Azenta Japan (Tokyo, Japan). Briefly, cDNA libraries were constructed using NEBNext Ultra Directional RNA Library Prep Kit for Illumina (E7420; New England Biolabs, Ipswich, MA, United States) and sequenced on an Illumina NovaSeq with a 2 × 150 bp paired-end configuration. Each sample yielded approximately 20 million sequencing reads.

For gene expression profiling of small ape iPSCs, Trim Galore was used for adapter and quality trimming with the -fastqc option (https://zenodo.org/records/7598955). Sequencing reads from small ape cells were mapped to nomLeu3 using STAR (Dobin et al., 2013), and uniquely mapped reads were retained for further analyses. BAM files were converted to BED format, and PCR duplicates were removed. The intersectBed tool in Bedtools was used to sum reads of overlapping exons of genes homologous to those of hg38 (Quinlan and Hall, 2010). Only the first read of paired-end reads was included in the count. Transcripts per million (TPM) values were calculated for normalization using Pandas (McKinney, 2010). Principal component analysis (PCA) and hierarchical clustering were conducted using Python 3.10.13.

For transposable element (TE) expression profiling, the consensus sequence of LAVA was obtained from a previous study (Carbone et al., 2012), and input into repeatmasker with the default parameters. Full-length (1,300 bp) and nearly full-length copies were selected on the basis of the following criteria: repStart <100 and repEnd >1,200. These copies were added to the RepeatMasker track of nomLeu3, while repeat annotations overlapping with LAVA were discarded. TE expression analysis was performed in accordance with our previous study (Hirata et al., 2022). Briefly, sequencing reads were mapped to nomLeu3 using Hisat2 with the--no-spliced-alignment option. For genomic regions with multiple hits, one candidate region was randomly selected. Only reads aligned in the sense orientation with the RepeatMasker track were used and normalized as reads per million reads mapped to the genome (RPM) to evaluate the expression levels.

For cross-species analyses, previously published fastq files of primate iPSC transcriptomes (PRJNA206563, PRJEB65856, PRJNA329031, PRJNA218873, PRJNA649770, PRJNA830828, PRJNA987625) were used (Marchetto et al., 2013; Wunderlich et al., 2014; Thoma et al., 2016; Hirata et al., 2022; Roodgar et al., 2022; Li et al., 2024; Ruiz-Orera et al., 2024). Preprocessed reads were mapped to the reference genomes of each species (hg38, panTro6, panPan3, gorGor6, ponAbe3, nomLeu3, rheMac10, macFas5) using STAR. Annotations from hg38 were transferred to other reference genomes using LiftOver (Hinrichs et al., 2006) to generate reference annotations based on hg38 (Hinrichs et al., 2006). Transcripts transferred across all seven non-human primate reference genomes (131,175 transcripts) and protein-coding genes commonly annotated across all eight reference genomes (13,543 genes) were used. For samples derived from the same individuals, reads mapped to the same genes were combined. Genes differentially expressed between each BioProject were identified using EdgeR [false detection rate (FDR) < 0.05] and excluded from further analysis to reduce batch effects (Robinson et al., 2010). This filtering resulted in a dataset containing 5,310 genes for cross-species analyses. Hierarchical clustering was performed using TPM value and the hclust function in the stats R package. FDR was calculated using EdgeR, and genes with FDR <0.05 and |log₂FC| > 1 were classified as being differentially expressed. Gene ontology analysis was performed using DAVID (Huang et al., 2009; Sherman et al., 2022).

Alkaline phosphatase staining was performed using Leukocyte Alkaline Phosphatase kit (86R-1KT; Sigma-Aldrich, St. Louis, MO, United States), in accordance with the manufacturer’s instructions. For immunofluorescence staining, cells (passage number over 9) were fixed with 4% paraformaldehyde (PFA), permeabilized with 0.5% Triton X-100, and blocked with 20% ImmunoBlock (CTKN001; KAC, Kyoto, Japan). The cells were then incubated overnight with primary antibodies: OCT3/4 (611,202; BD Biosciences; 1:150), SOX2 (MAB 2018; R&D Systems, Minneapolis, MN, United States; 1:150), NANOG (ab109250; Abcam, Cambridge, United Kingdom; 1:150), SSEA4 (MAB4304; Millipore, Burlington, MA, United States; 1:150), R-10G (011-25811; Wako; 1:150), tubulin β3 (801201; BioLegend, San Diego, CA, United States; 1:150), HAND1 (AF3168-SP; R&D Systems; 10 μg/mL), PRRX1 (ZRB2165; Sigma-Aldrich; 1:200), and goat anti-SOX17 (AF 1924; R&D Systems; 1:150). After washing with PBS, the cells were incubated with secondary antibodies: Alexa Fluor 488 goat anti-mouse IgG (A32723; Thermo Fisher Scientific; 1:500), Alexa Fluor 488 donkey anti-goat IgG (A11055; Thermo Fisher Scientific; 1:500), and Alexa Fluor 555 donkey anti-rabbit IgG (A31572; Thermo Fisher Scientific; 1:500). Nuclei were counterstained with 1 μg/mL 4′,6-diamidino-2-phenylindole (DAPI) (D523; Dojindo, Kumamoto, Japan). Each experiment included negative controls without the primary antibodies. Images were acquired using a BZ-X700 (Keyence, Osaka, Japan).

Direct neurosphere formation culture was performed following modified versions of previously reported procedures (Nakai et al., 2018; Kitajima et al., 2020; Nakai et al., 2022). Small ape iPSCs were dissociated with TrypLE Express and plated at 9.0 × 10³ cells/well into 96-well low-attachment culture plates (174,929; Thermo Fisher Scientific) in KBM Neural Stem Cell medium (16050200; Kohjin Bio, Sakado, Japan) supplemented with 10 μM Y-27632. The next day, the medium was further supplemented with 2 μM dorsomorphin (11967; Cayman Chemical Company, Ann Arbor, MI, United States), 10 μM SB431542 (13,031; Cayman Chemical Company), 1 ng/mL Stembeads FGF2 (SB500; StemCultures, Rensselaer, NY, United States), and 1× B-27 supplement (17504044; Gibco). On day 4, half of the medium was replaced with fresh medium. For neuronal differentiation, day 7 neurospheres were attached onto Geltrex (A1569601; Gibco)-coated culture plates and cultured for 2 weeks in KBM Neural Stem Cell medium without fibroblast growth factor (FGF)2 and epidermal growth factor (EGF), but supplemented with 1× B-27 supplement.

Limb bud mesoderm differentiation was performed with reference to previous reports (Loh et al., 2016; Yamada et al., 2021; Smith et al., 2023). Small ape iPSCs were dissociated with TrypLE Express and seeded at 1.5 × 10³ cells/cm² onto iMatrix-511-coated culture plates in StemFit AK02N medium supplemented with 10 μM Y-27632. The medium was replaced with fresh StemFit AK02N medium the next day. After another 2 days, the iPSCs were washed with DMEM and cultured in chemically defined CDM2 basal medium supplemented with 30 ng/mL Activin A (014-27621; Wako), 6 μM CHIR99021 (13,122; Cayman Chemical Company), 40 ng/mL human BMP4 (120-05ET; Peprotech, Cranbury, NJ, United States), 20 ng/mL human FGF2 (13,256-029; Gibco), and 100 nM PIK90 (S1187; Selleck, Houston, TX, United States) for 24 h to induce mid primitive streak differentiation (days 0–1). The next day, the mid primitive streak cells were washed with DMEM and cultured in CDM2 basal medium with 1 μM A83-01 (S7692; Selleck), 1 μM Wnt-C59 (S7037; Selleck), and 30 ng/mL human BMP4 for 24 h to induce lateral plate mesoderm differentiation (days 1–2). Then, the lateral plate mesoderm cells were washed with DMEM and cultured in CDM2 basal medium with 3 μM CHIR99021, 30 ng/mL human BMP2 (120-02; Peprotech), 10 ng/mL human FGF2, and 1× B-27 supplement for 48 h to induce limb bud mesoderm differentiation (days 2–4). The limb bud mesoderm differentiation medium was replaced daily with fresh medium. The CDM2 basal medium was composed of a 1:1 mixture of Iscove’s Modified Dulbecco’s Medium (IMDM, 31980030; Gibco) and Hams’ F-12 (31765035; Gibco) supplemented with 1 mg/mL polyvinyl alcohol (P8136; Sigma-Aldrich), 1% CD lipid concentrate (11905031; Gibco), 450 nM monothioglycerol (195-15791; Wako), 7 μg/mL human insulin (093-06351; Wako), 15 μg/mL transferrin (201-18081; Wako), and 100 units/mL penicillin and 100 μg/mL streptomycin.

Definitive endoderm differentiation was performed with reference to previous reports (Loh et al., 2014; Ang et al., 2018). Small ape iPSCs were dissociated with TrypLE Express and seeded at 1.5 × 10⁴ cells/cm² onto iMatrix-511/111 (1:3 mixture of iMatrix-511 and iMatrix-111)-coated culture plates in StemFit AK02N medium supplemented with 10 μM Y-27632. The medium was replaced with a fresh StemFit AK02N medium the next day. One day later, the iPSCs were washed with DMEM and cultured in endoderm induction medium supplemented with 100 ng/mL Activin A, 3 μM CHIR99021, 10 ng/mL human FGF2, and 100 nM PIK90 for 24 h to induce anterior primitive streak differentiation (days 0–1). On day 1, anterior primitive streak cells were washed with DMEM and cultured in endoderm induction medium supplemented with 100 ng/mL Activin A, 500 nM LDN193189 (S2618; Selleck), and 100 nM PIK90 for 24 h to induce definitive endoderm differentiation (days 1–2). The endoderm induction medium was composed of DMEM/F-12 (048-29785; Wako) supplemented with 0.5× B-27 supplement minus vitamin A (12587010; Gibco), 0.5× N-2 supplement (141-08941; Wako), 1×GlutaMAX supplement, and 100 units/mL penicillin and 100 μg/mL streptomycin.

Small apes (family Hylobatidae) occupy distinct phylogenetic branches positioned between hominoids (humans and great apes) and catarrhine monkeys. Their rapid diversification has led to their classification into four genera: Nomascus, Hoolock, Hylobates, and Symphalangus (Mootnick and Groves, 2005; Carbone et al., 2014; Roos, 2016; Shao et al., 2023) (Figure 1A). For the induction of iPSCs, primary fibroblast cells were cultured from skin specimens collected from five individuals representing three species within two genera: two lar gibbons (H. lar, 0023F and 0247M) and one Abbott’s gray gibbon (H. abbotti, 0159F) in the genus Hylobates, and two siamangs (S. syndactylus, 0105F and 0297F) in the genus Symphalangus (Figure 1B; Supplementary Figure S1).

The SRV, derived from SeV, was employed to deliver reprogramming factors into the small ape fibroblasts. SeV targets sialic acids on cellular membranes, facilitating efficient integration across diverse mammalian species (Bitzer et al., 2003). SRV offers the additional advantages of high and stable transgene expression because it replicates within the cytoplasm of transduced cells unless actively eliminated by small interfering RNA (siRNA). Using SeV vectors and SRVs, we have established iPSCs in non-human primates such as chimpanzees and Japanese macaques (Nakai et al., 2018; Imamura et al., 2024). For this study, SRVs encoding either four (OCT4, KLF4, SOX2, and c-MYC) or six (OCT4, KLF4, SOX2, c-MYC, NANOG, and LIN28) human reprogramming factors, along with an enhanced green fluorescent protein (EGFP) reporter gene (Nakanishi and Iijima, 2024), were used to reprogram small ape fibroblasts (Figure 1C). Three days after SRV transduction, the successful delivery of transgenes was confirmed via EGFP fluorescence, which indicated efficient gene transduction into small ape fibroblasts. Although most EGFP⁺ cells underwent cell death over the course of subsequent culture, a small number of tightly packed cell colonies, exhibiting EGFP fluorescence, emerged between days 10 and 21 (Figure 1D). These EGFP⁺ colonies continued to proliferate and were manually picked up around day 30. Eventually, 1–14 proliferating EGFP⁺ cell lines (designated as SRV⁺ reprogrammed cells) were independently established from each of the five small ape individuals (Table 1).

The SRV⁺ reprogrammed cells derived from small ape fibroblasts could be expanded in StemFit AK02N medium under feeder-free conditions with the laminin 511 E8 fragment (iMatrix-511) coating (Nakagawa et al., 2014) (Figure 1E). After several passages, selected SRV⁺ reprogrammed cell lines underwent siRNA transfection to eliminate the transduced SRVs, resulting in the emergence of EGFP⁻ colonies (data not shown). Subsequent cloning of these EGFP⁻ colonies led to the establishment of five small ape iPSC lines: one line (0023F-1) from a lar gibbon and four lines (0105F-B1, 0105F-B2, 0297F-A5, and 0297F-C1) from two siamangs. These small ape iPSCs were maintained in StemFit AK02N medium on iMatrix-511, under conditions identical to those used for human and other non-human primate iPSCs (Nakagawa et al., 2014; Kitajima et al., 2020; Imamura et al., 2024), and exhibited no detectable EGFP fluorescence (Figure 1F). The absence of SRVs was further validated by RT-PCR (Figure 1G). Chromosome QFH-banding confirmed normal karyotypes for the iPSC lines, with lar gibbons exhibiting 2n = 44 and siamangs 2n = 50 (Figure 1H).

To evaluate the pluripotency characteristics of the established small ape iPSCs, we analyzed the expression of key pluripotency-associated genes in iPSCs derived from a lar gibbon and siamangs. All of the small ape iPSC lines demonstrated alkaline phosphatase activity and were immunopositive for core pluripotency-associated transcription factors (OCT3/4, NANOG, and SOX2), as well as cell surface antigens (SSEA4 and R-10G) (Figure 2A; Supplementary Figure S2). To further characterize their transcriptional profiles, we conducted RNA-seq analyses on parental fibroblasts, SRV⁺ reprogrammed cells, and iPSCs. Principal component analysis (PCA) revealed a clear separation between cells before reprogramming (fibroblasts) and after reprogramming (SRV⁺ reprogrammed cells and iPSCs) along the first principal component (PC1) (Figure 2B). The top 100 genes contributing to PC1 included core pluripotency genes (e.g., LIN28A, SOX2, POU5F1, PODXL, ZIC2, OTX2, NANOG, MYCN, and SALL4) as well as fibroblast-related genes (e.g., ASPN, COL3A1, COL6A3, POSTN, CDH13, DCN, LOX, S100A4, and COL8A1) (Supplementary Table S2). These results confirmed the successful reprogramming of small ape fibroblasts through SRV-based gene transduction. The separation along the second principal component (PC2) appeared to reflect interspecies differences between lar gibbons and siamangs (Figure 2B). This distinction was further supported by unsupervised hierarchical clustering, which revealed a similar grouping pattern (Figure 2C).

We further examined the pluripotency state of the small ape iPSCs by analyzing the expression of genes associated with naïve or primed pluripotency using RNA-seq profiles (Takashima et al., 2014; Endoh and Niwa, 2022; Kunitomi et al., 2022). Consistent with the PC1 components identified in the PCA, genes associated with “shared pluripotency”—those common to both naïve and primed states (e.g., EPCAM, POU5F1, and NANOG)—were expressed in SRV⁺ reprogrammed cells and iPSCs, but were absent from fibroblasts (Figure 2D). Conversely, fibroblast-related genes (e.g., ACTA2, CD44, and COL1A2) were expressed in fibroblasts but not in SRV⁺ reprogrammed cells or iPSCs. Primed pluripotency genes (e.g., DNMT3B, OTX2, and PODXL) exhibited an expression pattern similar to that of the shared pluripotency genes, further indicating the transition to a primed pluripotent state. In contrast, naïve pluripotency genes (e.g., DNMT3L, DPPA2, and DPPA3) showed minimal expression in the iPSCs, reinforcing their classification as primed pluripotent cells. Interestingly, partial expression of naïve pluripotency genes was observed in the SRV⁺ reprogrammed cells, potentially attributable to the constitutive high expression of SRV-derived reprogramming factors. This observation aligns with findings in other species, where forced expression of reprogramming factors induces naïve-like pluripotency signatures (Fang et al., 2015; Honda et al., 2017a; Honda et al., 2017b; Kisa et al., 2017; Shiozawa et al., 2020).

To further dissect the molecular characteristics of the small ape iPSCs, we conducted RNA-seq analysis with a specific focus on transposable elements (TEs) (Supplementary Table S3). PCA of overall TE expression revealed distinct separation along PC1 based on reprogramming status and along PC2 based on species differences (Figure 3A). Notably, overall TE expression was upregulated following reprogramming (Figure 3B). This upregulation aligns with findings in early development, reprogramming, and pluripotent stem cells, where TE expression is typically elevated, contrasting with its suppression in somatic cells (Feng et al., 2010; Göke et al., 2015; Grow et al., 2015). The observed increase in TE expression further supports the assertion that the small ape fibroblasts were successfully reprogrammed, as corroborated by the upregulation of pluripotency-associated genes (Supplementary Table S2; Figure 2D). To identify TEs differentially expressed across reprogramming states and species, we extracted the top 10 TEs with the highest positive and negative loadings from PC1 and PC2. This analysis identified 36 TEs, with four overlapping between the components (HERV-K-int: PC1-negative/PC2-positive, MER65-int: PC1-positive/PC2-positive, HERVF-H21-int: PC1-negative/PC2-positive, HERV-H48-int: PC1-negative/PC2-positive) (Figure 3C). Among these, several endogenous retrovirus (ERV) elements exhibited higher expression in SRV⁺ reprogrammed cells and iPSCs across both species. Notable examples included HERV-H and LTR7, which are also highly expressed in human pluripotent stem cells (Santoni et al., 2012; Ohnuki et al., 2014), indicating evolutionary conservation of these expressional regulatory mechanisms among primates. Conversely, certain ERV and ancient repeat elements, such as LTR5A and Mamrep564, displayed higher expression in fibroblasts than in SRV⁺ reprogrammed cells and iPSCs. Additionally, species-specific expression patterns were identified for some ERV elements, with consistent expression levels before and after reprogramming (e.g., MLT1E1, MER66-int, MER34-int, HERVE-int). Finally, we examined the expression of LAVA, a TE unique to small apes and absent from other primate genomes (Carbone et al., 2012). LAVA was upregulated in SRV⁺ reprogrammed cells and iPSCs compared with the level in fibroblasts (Figure 3D), consistent with the behavior of other TEs conserved among primate genomes (Figure 3B).

To investigate the differentiation potential of the small ape iPSCs, we performed directed differentiation to induce lineage-specific differentiation into the three germ layers. First, to evaluate ectodermal differentiation, we employed a direct neurosphere formation culture, which recapitulates early neural development in human and non-human primate iPSCs (Nakai et al., 2018; Kitajima et al., 2020; Nakai et al., 2022) (Figure 4A). Small ape iPSCs formed neurospheres following a 7-day suspension culture, subsequently producing βIII tubulin⁺ neurons after an additional 14 days in adherent culture conditions (Figures 4B,C). For mesodermal and endodermal differentiation, we employed stepwise protocols to induce limb bud mesoderm or definitive endoderm differentiation from iPSCs. Referring to previous reports (Loh et al., 2014; Loh et al., 2016; Ang et al., 2018; Yamada et al., 2021; Smith et al., 2023), we optimized the protocol for limb bud mesoderm differentiation, guiding sequential differentiation into the mid primitive streak (day 1), lateral plate mesoderm (day 2), and limb bud mesoderm (day 4) (Figure 4D). Immunofluorescence analyses demonstrated the efficient differentiation of HAND1⁺ lateral plate mesoderm cells on day 2 and PRRX1⁺ limb bud mesoderm cells on day 4 (Figure 4E). Quantitative RT-PCR further confirmed these findings, showing the downregulation of pluripotency markers (POU5F1, NANOG) and stepwise upregulation of lateral plate mesoderm markers (HAND1, FOXF1, NKX2.5, ISL1) and limb bud mesoderm markers (PRRX1, MSX1, PITX1) over the course of differentiation (Figure 4F). In a separate differentiation protocol, small ape iPSCs were directed towards definitive endoderm cells. Efficient differentiation into definitive endoderm cells was demonstrated by the expression of SOX17 protein in immunofluorescence analyses (Figure 4G) and by quantitative RT-PCR detection of definitive endoderm-associated genes (SOX17, FOXA2, GSC, CER1, HHEX, GATA6) (Figure 4H). These results demonstrate that small ape iPSCs possess robust differentiation potential, capable of generating derivatives of all three germ layers.

Finally, to address the species-specific expression signatures, we compared the gene expression profiles of small ape iPSCs with those of other primate iPSCs reported in previous studies. We constructed a dataset comprising 90 sets of sequencing data from 42 individuals across nine primate species, namely, humans (Homo sapiens) (Marchetto et al., 2013; Wunderlich et al., 2014; Thoma et al., 2016; Hirata et al., 2022; Roodgar et al., 2022; Ruiz-Orera et al., 2024), great apes [chimpanzees (Pan troglodytes), bonobos (Pan paniscus), gorillas (Gorilla), and orangutans (Pongo pygmaeus)] (Marchetto et al., 2013; Wunderlich et al., 2014; Hirata et al., 2022; Roodgar et al., 2022; Li et al., 2024; Ruiz-Orera et al., 2024), small apes [lar gibbons (H. lar) and siamangs (S. syndactylus)], and macaques [rhesus macaques (Macaca mulatta) and cynomolgus macaques (Macaca fascicularis)] (Wunderlich et al., 2014; Thoma et al., 2016; Roodgar et al., 2022; Ruiz-Orera et al., 2024) (Supplementary Table S4), from eight independent studies. To minimize the confounding factors unrelated to interspecies differences, hierarchical clustering was performed to identify and exclude outlier samples that clustered separately from the majority (height >0.8) (Supplementary Figure S3). Batch effects were mitigated by identifying and removing genes exhibiting study-specific expression biases (FDR <0.05) (Supplementary Table S5). This preprocessing yielded a gene expression matrix comprising 38 individuals and 5,310 genes (Supplementary Table S6). PCA of this matrix demonstrated a clear separation between human and non-human primate iPSCs along PC1. Non-human primate iPSCs were distributed along PC2 in a pattern reflecting their phylogenetic relationships (Figure 5A). Hierarchical clustering further segregated the samples into distinct clusters corresponding to humans, great apes, small apes, and macaques (Figure 5B), confirming the utility of this matrix for exploring the species-specific gene expression patterns. Differential expression analysis identified 160 upregulated and 304 downregulated genes specific to the small apes (Figure 5C; Supplementary Table S7, S8). Notably, the downregulated genes were significantly enriched for functions related to genome stability and cell death (e.g., BUB1, DRAM1, RAD21) (Pati et al., 2002; Jeganathan et al., 2007; Guan et al., 2015) (Figure 5D; Supplementary Tables S9, S10). These findings raise the possibility of the presence of small ape-specific regulatory mechanisms and responses to DNA damage and chromosomal rearrangements that are distinct from those observed in other primate species.