Abstract Background Acute graft-versus-host disease (aGVHD) remains a major life-threatening complication of allogeneic haematopoietic cell transplantation (allo-HSCT), often limiting the therapeutic efficacy of allo-HSCT. Recent studies have suggested that mesenchymal stem cells (MSCs) may be beneficial for the treatment of aGVHD. However, the therapeutic potential of MSCs is often negatively impacted by their heterogeneity. Methods To investigate MSCs heterogeneity, we conducted single-cell transcriptomic analysis of human umbilical cord-derived MSCs (HUC-MSCs) and identified key feature genes that distinguish MSCs subpopulations. The function of the newly discovered biomarker CRISPLD2 was also explored. We engineered human umbilical cord-derived MSCs (HUC-MSCs) to overexpress the CRISPLD2 gene using lentiviral vectors. The downstream regulatory effects of CRISPLD2 overexpression were assessed through bulk RNA sequencing. Additionally, we evaluated its impact on cellular senescence using Western blotting and β-galactosidase (SA-β-gal) staining. The immunoregulatory capability of HUC-MSCs was tested through coculture experiments with T cells and liver organoids in vitro. Mitochondrial function was analysed via flow cytometry and electron microscopy. The in vivo therapeutic effects of HUC-MSCs on aGVHD were evaluated using an aGVHD murine model. The graft-versus-leukaemia (GVL) effect was measured via the inoculation of luciferase-positive A20 cells, and tumour growth was monitored via bioluminescence imaging. Results Our findings indicated that the CRISPLD2 gene is heterogeneously expressed in HUC-MSCs subsets characterized by stemness and immunosuppressive properties. Transcriptomic analysis revealed that CRISPLD2 overexpression suppressed calcium ion binding and G protein-coupled receptor signalling. In vitro studies demonstrated a marked increase in IL-10 secretion, which enhanced T-cell suppression in CRISPLD2-modified HUC-MSCs. The in vivo results demonstrated that transfusion of CRISPLD2-overexpressing HUC-MSCs ameliorated aGVHD while maintaining GVL activity. Mechanistically, CRISPLD2 overexpression overcomes the mitochondrial damage mediated by extracellular ATP and LPS in HUC-MSCs by inhibiting P2Y11 receptor signalling, thereby preserving their stemness and IL-10-mediated immunosuppressive functions. Conclusions Our study revealed that CRISPLD2 is a novel marker for identifying HUC-MSCs subpopulation with enhanced immunosuppressive functions. CRISPLD2 overexpression enhances the immunosuppressive function of HUC-MSCs by inhibiting P2Y11 receptor signalling. Targeting CRISPLD2 is a promising strategy to improve the therapeutic efficacy of HUC-MSCs in aGVHD while maintaining GVL activity. Graphical abstract [54]graphic file with name 13287_2025_4321_Figa_HTML.jpg Supplementary Information The online version contains supplementary material available at 10.1186/s13287-025-04321-6. Keywords: Acute graft-versus-host disease (aGVHD), Mesenchymal stem cells (MSCs), CRISPLD2, P2Y11R Background Allogeneic haematopoietic stem cell transplantation (allo-HSCT) is a curative therapy for haematologic malignancies [[55]1, [56]2]. The success of allo-HSCT depends partly on alloreactive T-cell-mediated graft-versus-leukaemia (GVL) activity [[57]1]. However, T cells from the same donor also cause graft-versus-host-disease (GVHD) [[58]3]. Acute GVHD (aGVHD) is initiated by tissue damage caused by pretransplantation conditioning regimens such as chemotherapy and irradiation [[59]3]. The release of danger signals such as pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs) following tissue damage further activates donor immune cells and exacerbates the inflammatory response [[60]4, [61]5]. Human umbilical cord-derived mesenchymal stem cells (HUC-MSCs) have been shown to have advanced efficacy in the prevention and second-line treatment of GVHD [[62]6–[63]9]. HUC-MSCs, which have immunosuppressive properties, exert their therapeutic effects via the secretion of immunoregulatory cytokines such as IL-10 and IDO, which reduce chemotaxis, ROS production, and NADPH oxidase levels [[64]10–[65]13]. In addition, MSC-derived exosomes have been shown to be key factors in the delivery of immunomodulatory agents [[66]14–[67]16]. However, the immunosuppressive effects of HUC-MSCs can be impaired by extracellular PAMPs, such as lipopolysaccharides (LPS); DAMPs, such as high mobility group box 1 (HMGB-1); and proinflammatory cytokines, such as interleukin (IL)−1β, IL-6, and ATP, which are released during preconditioning in allo-HSCT [[68]5, [69]17]. ATP impairs the function of MSCs by inhibiting their differentiation, proliferation, migration, and tissue homing [[70]18–[71]23]. The subsequent activation of P2X/P2Y receptor-mediated calcium signalling is associated with ATP-mediated MSCs dysfunction [[72]24]. Notably, various P2X and P2Y receptors are expressed in MSCs from different tissues and species, although their expression levels and functional roles significantly vary [[73]25–[74]28]. Our previous study highlighted significant tissue-specific characteristics among MSCs derived from different tissues and identified many previously unknown tissue-specific functional subpopulations and associated molecular markers. Among these molecular markers, CRISPLD2 was identified as a potential marker of HUC-MSCs subsets. However, the role of CRISPLD2 in MSCs remains poorly understood [[75]29]. The cysteine-rich secretory protein LCCL domain-containing 2 (CRISPLD2) protein is encoded by the CRISPLD2 gene. Recent studies have indicated its potential to enhance the regenerative properties of human bone marrow stromal cells (hBMSCs). A positive correlation between CRISPLD2 expression and the progression of three-dimensional chondrogenesis in hBMSCs has been reported [[76]30]. Notably, CRISPLD2 levels are significantly lower in late-passage hBMSCs than in earlier-passage hBMSCs, leading to impaired osteogenic differentiation of MSCs in vitro [[77]31, [78]32]. Furthermore, elevated CRISPLD2 levels can attenuate the production of proinflammatory cytokines induced by the danger signal HMGB1 [[79]33], suggesting a protective role of CRISPLD2 in HUC-MSCs against damage caused by DAMPs and PAMPs. However, the role of CRISPLD2 in regulating the therapeutic efficacy of HUC-MSCs in the context of aGVHD, particularly its impact on downstream P2X/P2Y receptor signalling, is largely unknown. Therefore, we tested whether modification of CRISPLD2 expression on HUC-MSCs could enhance their immunosuppressive function and improve the efficacy of HUC-MSCs in the prevention of aGVHD. Methods Single-cell transcriptomics data quality control and clustering In June 2024, high-quality single-cell transcriptomic datasets of HUC-MSCs were systematically collected through Google Scholar and PubMed searches. Three datasets containing data on umbilical cord tissues ([80]GSE182158, f4b2ykfv56-1) were downloaded for subsequent data analysis. The single-cell RNA sequencing (scRNA-seq) data were processed and analysed via the Scanpy toolkit (v1.10.2) according to our previous reports [[81]34]. Briefly, for data quality control, each sample was filtered to remove cells expressing fewer than 200 genes and genes expressed in fewer than three cells. Additionally, cells expressing more than 7,000 genes or having more than 10% mitochondrial gene content were excluded from further analysis. Doublets were predicted via DoubleFinder (v2.0) and removed before performing principal component analysis (PCA) for dimensionality reduction and batch effect correction. Cell cycle effects and mitochondrial genes were also excluded during the identification of highly variable genes (HVGs). The corrected nearest neighbours were then used to perform UMAP for nonlinear dimensionality reduction and visualization. For clustering, we applied the Leiden algorithm using the corrected neighbours, with a resolution parameter set to 2.0 to achieve finer subpopulation granularity. Developmental trajectory analysis Developmental trajectory analysis was conducted via Monocle 2 (v2.32.0) [[82]35]. The cells were ordered in pseudotime on the basis of highly variable genes (HVGs) identified during preprocessing. Dimensionality reduction with the DDRTree method was used to reconstruct the trajectory, which was visualized along a pseudotime axis to identify key developmental transitions. Differentially expressed genes along the trajectory were analysed to reveal the molecular changes driving progression. Transcription factor activation prediction Transcription factor (TF) activation was predicted via the pySCENIC toolkit (v0.12.1) [[83]36] in three steps. First, a coexpression matrix was generated from the gene expression data. Next, enriched motifs were identified, and cis-regulatory networks were refined to optimize target TFs. Finally, regulon activity was calculated across individual cells to quantify TF activation and its regulatory impact on downstream gene expression. Gene network analysis Pathway enrichment analysis of the top differentially expressed genes was performed via Metascape [[84]37]. Networks were visualized and modified in Cytoscape (v3.7.1) [[85]38], and protein‒protein interaction networks were constructed via the BioGrid, InWeb_IM, and OmniPath databases with default settings. For further details, refer to previous research [[86]29]. Mice BALB/c (H-2 Kd) and C57BL/6 J (H-2 Kb) mice were purchased from the VITALRIVER Experimental Animal Company (Beijing, China). All the mice were maintained in a specific pathogen-free room at Xinqiao Hospital of Army Medical University (Chongqing, China). All murine experiments were performed in accordance with protocols approved by the animal care and use committee of Army Medical University. Cell culture HUC-MSCs were purchased from iCell (Chongqing, China) and cultured in OriCell® basal medium (Guangzhou, China) supplemented with OriCell® foetal bovine serum and culture supplement for human umbilical cord mesenchymal stem cells in an incubator with 5% CO[2] at 37 °C. A20-luc cells were purchased from Zhong Qiao Xin Zhou Biotechnology Co., Ltd. (Shanghai, China) and cultured in RPMI-1640 (GIBCO, USA) supplemented with 10% FBS and 0.05 mM 2-Mer (GIBCO, USA) an incubator with 5% CO[2] at 37 °C. When the cells reached approximately 80% confluence, they were dissociated with trypsin and inoculated into T25 cell culture flasks for other experiments. HUC-MSCs were identified by detecting cell surface phenotypes and their multipotent potential for differentiation into adipogenic, osteogenic, and chondrogenic lineages and maintained as previously described [[87]39]. Preparation of human umbilical cord mesenchymal stem cells Briefly, fresh umbilical cords were obtained and washed with phosphate-buffered saline (PBS) containing 1% penicillin and streptomycin. The cord vessels were removed to prevent endothelial cell contamination, and the cords were rinsed several times in phosphate-buffered saline (PBS) and cut into small pieces (1–2 mm^3), which were immediately placed in six-well plates in 2 mL of OriCell® medium for HUC-MSCs culture expansion. After 3 days, nonadherent cells were discarded by washing with PBS, and adherent cells were defined as passage zero. The HUC-MSCs were subcultured for further experiments. The osteogenic, adipogenic, and chondrogenic differentiation of HUC-MSCs The induction of osteogenic, adipogenic, and chondrogenic differentiation was performed according to previously reported methods [[88]40] with some modifications. To induce osteogenic differentiation, HUC-MSCs were seeded into six-well plates at 2 × 10^4 cells/well, with 2 mL of complete medium added to each well. The cells were cultured in an incubator with 5% CO[2] and saturated humidity. When the cell confluence reached 70%, the complete medium was replaced with 2 mL of osteogenic induction medium, which was changed every three days. After 2–4 weeks of induction, osteoblast differentiation was assessed via Alizarin Red staining. To induce adipogenic differentiation, after the cells reached 100% confluence, the medium was removed and replaced with 2 mL of adipogenic induction differentiation medium (Medium A). After three days, Medium A was removed and replaced with Medium B for adipogenic differentiation induction. Medium A and medium B were used alternately. After 2–4 weeks, Oil Red O staining was performed to assess lipid formation. To induce chondrogenic differentiation, 4 × 10^5 HUC-MSCs at P3 were transferred to a 15 mL centrifuge tube and centrifuged at 250 × g at 20 °C for 4 min. A total of 500 μL of chondrogenic differentiation induction premix was added to the cell pellet obtained from centrifugation, and the mixture was resuspended. The centrifuge tube cap was loosened to allow for gas exchange and placed upright in an incubator at 37 °C with 5% CO₂. When the cells appeared to clump together, the bottom of the centrifuge tube was flicked to release the cell clumps, which were then suspended in the fluid. After their inoculation into flasks, the cells were provided with fresh chondroblast induction complete differentiation medium every 2–3 days. Lentiviral vector construction, packaging, and transduction of HUC-MSCs The full-length coding sequence of CRISPLD2 ([89]NM_031476.4; 4583 bp) was amplified from human MSC cDNA via polymerase chain reaction (PCR), gel-purified, and ligated with T4 DNA ligase into the [90]H29272 vector (GL180 pcSLenti-EF1-EGFP-P2 A-Puro-CMV-MCS-3Xflag -WPRE; OBiO Technology, Shanghai, China), which carries the enhanced green fluorescent protein (EGFP) gene, to develop a construct coexpressing human CRISPLD2 and EGFP. The ligation mixture was transformed into competent Escherichia coli bacteria, which were subsequently plated onto an ampicillin plate. Colonies were selected 16 h later and inoculated onto Luria–Bertani medium. Bacteria were cultured on a shaker at 37 °C for 16 h. The plasmid was extracted via the alkaline lysis method, confirmed via PCR and sequenced. All the constructs were transfected into 293 T packaging cells (OBiO Technology, Shanghai, China.) with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) to produce lentiviruses. MSCs were transduced with viral supernatant (multiplicity of infection = 50), and the mRNA expression of CRISPLD2 was measured via reverse transcription‒polymerase chain reaction (RT‒PCR) 3 days after transduction to verify CRISPLD2 overexpression. The transduction efficiency was evaluated by detecting the expression of GFP with an Olympus IX51 fluorescence microscope (Olympus Co., Tokyo, Japan). HUC-MSCs carrying either EGFP (HUC-MSCs-GFP) alone or both CRISPLD2 and EGFP (HUC-MSCs-CRISPLD2) were harvested after selection using puromycin at the minimal lethal concentration (1.5 μg/ml). The puromycin-resistant cells were then collected for further use (Table [91]1). Table 1. Primer sequences for CRISPLD2 amplification Forward primer (5′–3′) AACCTGAAACGGACGAGATGAAT Reverse primer (5′–3′) CTTTGCACCTGTCCTTCATCTTG [92]Open in a new tab Cell viability assay HUC-MSCs were seeded in a 96-well plate at a concentration of 100 μl/well (5 × 10^4 cells). The 96-well plates were incubated at 37 °C for 24 h or 48 h. CCK-8 solution (10 μl; CCK-8; Dojindo, Japan) was added to each well, and the cells were incubated at 37 °C for 4 h. The OD values of the different groups at 450 nm were determined. Colony formation assay HUC-MSCs (1 × 10^3) were seeded in a 6-well plate, and 2 mL of human umbilical cord mesenchymal stem cell medium was added. Medium changes were performed every 3 days, and crystal violet staining was performed on the 14 th day. The number of colonies formed was calculated, and the ratio of the number of cells seeded was the colony formation rate. Bulk RNA sequencing A total of 1.5 × 10^6 HUC-MSCs were collected for RNA extraction. Total RNA was isolated using TRIzol reagent (Invitrogen, USA) according to the manufacturer's protocol. The purity and concentration of the extracted RNA were assessed using a NanoDrop 2000 spectrophotometer and a Qubit 2.0 fluorometer (Thermo Fisher Scientific). RNA integrity was evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies), ensuring that the RNA integrity number (RIN) was greater than 7 prior to library preparation. Library construction was performed with a TruSeq Stranded mRNA LT Sample Prep Kit (Illumina, San Diego, CA, USA) following the manufacturer's instructions. Paired-end 150 sequencing reads were generated on an Illumina HiSeq X10 platform by OE Biotech Co., Ltd. (Shanghai, China). Low-quality reads and adapters were removed via Trimmomatic to ensure high-quality data. The cleaned reads were aligned to the human reference genome (UCSC genome browser, hg38) via HISAT2. Subsequently, HTSeq-count was employed to obtain read counts for each gene on the basis of the annotated gene reference (hg38). The fragments per kilobase of transcript per million mapped reads (FPKM) for each gene were calculated via Cufflinks, with additional correction for fragment bias to increase expression estimation. Finally, data visualization was performed via the R package ggplot2 (version 3.0). Quantitative RT‒PCR Total RNA was extracted from the samples using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Reverse transcription was conducted with a PrimeScript RT Reagent Kit (Takara, Kusatsu, Japan) to synthesize complementary DNA (cDNA). The expression levels of IDO, CXCR5, TLR2, TLR4, TLR6, IL-10, HGF, VEGF, CCL18, GM-CSF, and PGE2 in HUC-MSCs were quantified via RT–PCR using a SYBR Green Master Mix in accordance with the manufacturer’s guidelines, with GAPDH serving as the internal control. The primers used are shown in Table [93]2. The 2^−ΔΔCT method was used to determine the relative mRNA expression. Each assay was performed in triplicate. Table 2. Primer sequences for RT‒qPCR Gene Forward primer (5′–3′) Reverse primer (5′–3′) GAPDH GGAGTCCACTGGCGTCTTCA GTCATGAGTCCTTCCACGATACC IDO CAAATGCAAGAACGGGACACTTT GTCTTCCCAGAACCCTTCATACA CXCR5 CCTACAGCCTCATCTTCCTCCTG TTGACTTTGTGCAGGGCAATCA TLR4 CATTGGTGTGTCGGTCCTCA CCAGTCCTCATCCTGGCTTG IL-10 AGCTCCAAGAGAAAGGCATCTAC GTCTATAGAGTCGCCACCCTGAT HGF GCAGCTACAAGGGAACAGTATCT CTTCGTAGCGTACCTCTGGATTG VEGF CGAGTACATCTTCAAGCCATCCT CCTTTCCCTTTCCTCGAACTGAT TLR6 CAAACGTGGGCTCTTTTGGG AGGCACCTCCAGACAGTTAC CCL18 GGTGTCATCCTCCTAACCAAGAG GGCATAGCAGATGGGACTCTTAG GM-CSF GGCTAAAGTTCTCTGGAGGATGT TCATCTCAGCAGCAGTGTCTCTA PGE2 CACCTCATTCTCCTGGCTATCAT ACGCATTAGTCTCAGAACAGGAG TLR2 AAGCACTGGACAATGCCACA ACCATTGCGGTCACAAGACA [94]Open in a new tab Western blotting analysis HUC-MSCs were lysed in RIPA buffer (Beyotime, Shanghai, China) supplemented with a protease inhibitor cocktail (Thermo Fisher, Waltham, USA). The protein concentration of the extracted lysates was measured using a BCA assay (Beyotime, Shanghai, China). Equal amounts of protein were then loaded onto 10% SDS‒PAGE gels and subjected to electrophoresis before being transferred to polyvinylidene difluoride (PVDF) membranes. Each membrane was blocked with 5% BSA and incubated overnight at 4 °C with the following primary antibodies: anti-CRISPLD2 (Abcam, UK), anti-p-p53, anti-p53, anti-p16INK4 A, anti-p21, and GAPDH (CST, USA). After washing, the membranes were treated with peroxidase-conjugated goat anti-rabbit or goat anti-mouse secondary antibodies (Invitrogen, Carlsbad, CA, USA). Protein detection was accomplished using a chemiluminescent horseradish peroxidase (HRP) substrate system (Millipore, Massachusetts, USA) to visualize the protein bands on the blots. The antibodies used for immunostaining and their dilutions are listed in Supplemenatry Table [95]1. Enzyme-linked immunosorbent assay (ELISA) The concentration of IL-10 in the conditioned medium of HUC-MSCs was measured using ELISA kits (mlbio, Shanghai, China) following the manufacturer's protocol. SA-β-gal staining Cellular senescence was evaluated with an SA-β-gal staining kit (Beyotime, Shanghai, China). Briefly, HUC-MSCs were fixed with 4% paraformaldehyde and stained with SA-β-gal. Two independent donor-derived HUC-MSCs were tested. Six random fields for each donor-derived HUC-MSCs were captured at 200 × magnification. The average percentage of SA-β-gal-positive HUC-MSCs was calculated and used for statistical analysis. T-cell proliferation assay CD90.2^+ T cells from the spleens of C57BL/6 mice were purified with CD90.2 Microbeads (Miltenyi Biotec, Germany). The purity was evaluated via flow cytometry, and > 95% purity was obtained. Then, the purified CD90.2^+ T cells were labelled with CFSE (BD, New Jersey, USA) and plated in 96-well plates (1 × 10^5 cells per well) with or without CD3/CD28 Dynabeads (Gibco, Carlsbad, CA, USA)[[96]41]. HUC-MSCs were subsequently added at a density of 1 × 10^5 cells per well, and the coculture was maintained for 72 h. Following coculture, the cells were collected and stained for flow cytometry analysis. Human peripheral blood mononuclear cells (PBMCs) were directly stimulated and cocultured with HUC-MSCs without prior T-cell purification. Antibodies and FACS analysis (human antibodies) mAbs against mouse TCR-β (H57-597), CD4 (RM4-5), and CD8α (53–6.7) were purchased from Thermo Fisher Bioscience. mAbs against human CD3 (SK7), CD4 (SK3), and CD8a (RPA-T8) were purchased from Biolegend. The cells were stained with antibodies according to the manufacturer's protocol. Flow cytometry analyses were performed with a BD LSRFortessa flow cytometer (Franklin Lakes, NJ), and the resulting data were analysed with FlowJo software V11 (Tree Star, Ashland, OR). Ca^2+measurements Mitochondrial Ca^2+ was measured with Indo-1 AM (MCE, USA), and cellular Ca^2+ was measured with Rhod-1 AM (MCE, USA). Briefly, the cells were stimulated with ATP (MCE, USA) at a concentration of 2 mM for 5, 10, or 15 min. Following stimulation, the HUC-MSCs were incubated with 1 mL of Indo-1 AM working solution at room temperature for 30 min. After incubation, the cells were washed and analysed via flow cytometry. Cellular and mitochondrial reactive oxygen species (ROS) and mitochondrial membrane potential measurements Cellular and mitochondrial ROS levels were quantified with dihydroethidium (MCE, USA) and MitoSOX™ Red (Thermo Fisher Scientific, USA), respectively. For mitochondrial membrane potential measurement, TMRM (MCE, USA) was used to stain the cells. HUC-MSCs were stimulated with different conditions: ATP (MCE, USA) for 1 h or LPS (MCE, USA) for 1 h, followed by an hour of ATP stimulation. After stimulation, the HUC-MSCs were collected and washed twice with PBS. The cells were then incubated with dihydroethidium or MitoSOX™ Red at 37 °C for 45 min or TMRM for 30 min. The fluorescence intensity was subsequently measured via flow cytometry performed with a BD LSRFortessa (Franklin Lakes, NJ), and the data were analysed with FlowJo software (Tree Star, Ashland, OR). Construction of the liver organoid-T-cell coculture model Fresh samples were placed in tissue preservation solution (Daxiang Technology, Beijing, China) and transported to the laboratory at low temperatures. After the tissue was washed with washing solution (Daxiang Technology, Beijing, China), it was cut into pieces that were 1–2 mm^3 using ophthalmic scissors and collected in a centrifuge tube. A tissue dissociation enzyme mixture (Daxiang Technology, Beijing, China) was added, and the mixture was incubated in a temperature-controlled shaking incubator at 37 °C for 30–60 min. After the cell suspension was resuspended in organoid washing solution (Daxiang Technology, Beijing, China), it was filtered through a 100 μm cell strainer. The filtered cell suspension was centrifuged to obtain a cell pellet, which was then resuspended in matrix gel (Daxiang Technology, Beijing, China) and seeded as gel droplets on a 24-well cell culture plate. After solidifying in the incubator, 500 μL of cancer or tissue organoid culture medium (brand, product number) was added. For the first two days of culture, the organoid medium was supplemented with anti-apoptotic factors (Daxiang Technology, Beijing, China). The medium was completely changed every 3‒4 days, and subculturing was performed every 7‒14 days depending on the viability of the organoids. Photographs of the organoids under bright field conditions were captured using a microscope (Nikon). Liver organoid culture and subculture were conducted following the above organoid subculture procedures. After digestion, liver organoid culture medium (Daxiang [97]HK100101) was added to the liver organoids, which were collected, seeded into matrix gel-modified cell culture wells at a standard density of 1000–2000 cells/60 μL/well, and cultured for 4–6 days, with the medium changed every other day. A cell tracing dye (Daxiang [98]LS100105) was used to stain the T cells. The cells were subsequently seeded into the same cell culture well at a 3:1 effector-to-target ratio. When T cells were inoculated, the culture medium was supplemented with CD3/28 T-cell activators, fluorescent dyes that stain apoptotic cells (Daxiang [99]LS100106), and IL-10. During coculture, high-content imaging was performed daily to capture the fluorescent signals of apoptotic cells in the organoids and the T-cell-tracing fluorescent signals, which were further analysed via MetaXPress high-content analysis software. The fluorescence intensity of the corresponding areas of the images was quantified to obtain the total fluorescence intensity of organoid apoptotic cells. Finally, the total fluorescence intensity of organoid apoptotic cells under different treatment conditions was calculated to evaluate the efficacy of IL-10. Mouse serum LPS and ATP measurements Male BALB/c mice, aged 8–12 weeks, were subjected to 750 cGy total body irradiation (X-ray), which was administered in two fractions on Day −1 and Day 0. The mice without X-ray irradiation were used as controls. The mice were euthanized following cervical spine dislocation after being anaesthetized with isoflurane on Days 3 and 4. Serum was collected from these mice to measure the levels of extracellular ATP (Qisong, China) and lipopolysaccharide (JiONLNBIO, China). Murine GVHD and GVL models Male BALB/c mice, aged 8–12 weeks, were subjected to 750 cGy lethal irradiation (X-ray)administered in two sessions on Day −1 and Day 0. Six to eight hours after the second irradiation session, bone marrow cells (5 × 10^6) and splenocytes (1.25 × 10^6) from C57BL/6 donors were injected into the recipient mice via the tail vein. Apyrase (4 units; Sigma) was administered daily from Days 0 to 4. The survival of the recipients was monitored continuously. In the GVL experiment, lethally irradiated BALB/c mice were infused with 5 × 10^6 T-cell-depleted bone marrow (TCD-BM) cells and 1.25 × 10^6 splenocytes from C57BL/6 donors, along with 5 × 10^6 luciferase-positive A20 cells at the time of transplantation. Assessment of aGVHD severity and GVL activity The severity of aGVHD was assessed via aGVHD scoring criteria as described previously [[100]42]. To evaluate the histopathology of aGVHD target organs, the liver, small intestine, and colon were evaluated. The mice were euthanized following cervical spine dislocation after being anaesthetized with isoflurane. Tissue sections were prepared and stained with haematoxylin and eosin (H&E). Pathologic changes were analysed as described previously [[101]43]. In brief, liver aGVHD was scored by evaluating the degree of lymphocytic infiltration, the number of involved tracts, and the severity of liver cell necrosis; the maximum score was 9. Gut aGVHD was scored by evaluating the degree of mononuclear cell infiltration and the severity of morphological aberrations (e.g., hyperplasia and crypt loss), with a maximum score of 8. To assess GVL activity, the mice were anaesthetized with isoflurane. Bioluminescence imaging (BLI) was performed to measure the tumour burden of the recipients using the IVIS Lumina S5 system (PerkinElmer, Waltham, MA). Statistics The data are presented as the means ± SEMs. The survival curves were compared via the log-rank test, and the luciferin intensity curves were compared via two-way ANOVA for multiple comparisons. Three means were compared via one-way ANOVA for multiple comparisons, whereas two means were compared via an unpaired two-tailed Student’s t test (Prism, version 9.5; GraphPad Software); p values less than 0.05 were considered significant. (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). Results Single-cell transcriptomics analysis revealed CRISPLD2 + HUC-MSC subsets with potential immunosuppressive functions In our previous study, we conducted a comprehensive analysis of MSCs heterogeneity across four tissue sources (bone marrow, umbilical cord, dermis, and adipose tissue) at the single-cell level, revealing distinct tissue-specific traits [[102]29]. Functional assays confirmed that HUC-MSCs have superior immunosuppressive properties compared with other tissue-derived MSCs [[103]44]. Clinical studies also highlight the enhanced efficacy of HUC-MSCs in treating aGVHD [[104]45, [105]46]. However, this therapeutic effectiveness can vary significantly across different medical centres because of the inherent heterogeneity of MSCs. Therefore, there is an urgent need to identify unique markers that can distinguish between various subsets of MSCs. To identify and characterize the subpopulations of HUC-MSCs that contribute to their functional advantages, single-cell transcriptome analysis was performed. The data revealed three distinct subgroups on the basis of CRISPLD2 expression: two CRISPLD2-high subpopulations (CRISPLD2 + _1 and CRISPLD2 + _2) and one CRISPLD2-low subpopulation (CRISPLD2-) (Fig. [106]1A). Developmental trajectory analysis indicated that the CRISPLD2 + subpopulation exhibited greater stemness than the CRISPLD2- subpopulation did (Supplemenatry Figure S1). Further transcription factor activation analysis revealed significant activation of stemness-related transcription factors in both CRISPLD2 + subpopulations, including MEIS2 in the CRISPLD2 + _2 subpopulation and SOX4, SOX6, SOX9, and HMBOX1 in the CRISPLD2 + _1 subpopulation (Fig. [107]1B). Gene Ontology (GO) analysis further validated the significant functional heterogeneity between the subpopulations. In the CRISPLD2 + _1 subpopulation, genes involved in the"positive regulation of the cell cycle process"were downregulated, whereas in the CRISPLD2 + _2 subpopulation, genes related to the"negative regulation of cell differentiation"were downregulated. Notably, the CRISPLD2 + _2 subpopulation also presented significant enrichment of the"Toll-like receptor 3 signalling pathway"(Supplemenatry Figure S2). Recent studies have demonstrated that TLR3 activation in human adipose-derived MSCs enhances their immunosuppressive properties and drives phenotypic changes in the subpopulation expressing TLR3 [[108]47]. Further analysis of differentially expressed genes revealed that both CRISPLD2 + subpopulations notably overexpressed TGFB1, with the CRISPLD2 + _2 subpopulation showing particularly high levels of CXCL12 and GPNMB (Fig. [109]1C). Previous research has demonstrated that TGFB1-mediated MSCs exhibit superior therapeutic effects in modulating corneal allograft immune rejection and that MSCs overexpressing TGF-β1 can improve treatment outcomes in patients with organ injury and inflammation by reducing macrophage infiltration [[110]48, [111]49]. Moreover, CXCL12 and GPNMB are recognized for their role in suppressing immune responses and inflammation [[112]50, [113]51]. Our study also revealed that the expression of CXCL12, GPNMB, and TGF-β1 tends to decrease as MSCs stemness decreases (Supplemenatry Figure S1). These findings suggest that CRISPLD2 + subpopulations may offer functional advantages in therapeutic applications owing to their increased stemness and secretion of therapeutic factors and their ability to inhibit immune responses. Fig. 1. [114]Fig. 1 [115]Open in a new tab CRISPLD2 is a novel marker for immunosuppressive subsets of HUC-MSCs. (A) UMAP plots of scRNA-seq data from HUM-MSCs across multiple tissue sources, with each tissue type colour-coded. (B) Transcription factor activation in each subpopulation was predicted from gene expression, and a heatmap illustrates the transcription factors uniquely activated in each subpopulation. (C) UMAP and violin plots showing signature gene expression across all MSC populations Engineering the CRISPLD2 gene overexpression in HUC-MSCs To further examine the function of CRISPLD2 in HUC-MSCs, we engineered CRISPLD2-overexpressing HUC-MSCs through lentivirus transfection with the EGFP (HUC-MSCs-GFP) and CRISPLD2 (HUC-MSCs-CRISPLD2) genes. This transfection was successful, as indicated by a strong green fluorescent signal in more than 90% of the cells (Fig. [116]2A). Quantitative PCR confirmed a dramatic increase in CRISPLD2 mRNA levels (Fig. [117]2B), and western blot analysis revealed a significant increase in CRISPLD2 protein expression (Fig. [118]2C). The adipogenic, osteogenic, and chondrogenic differentiation capacities of HUC-MSCs-CRISPLD2 were preserved (Fig. [119]2D). Furthermore, flow cytometry analysis revealed that the modified cells retained positive markers for CD90, CD105, and CD73 and were negative for CD34, CD11b, CD19, CD45, and HLA-DR, indicating that the fundamental characteristics of HUC-MSCs were unchanged (Fig. [120]2E). Fig. 2. [121]Fig. 2 [122]Open in a new tab Characterization of HUC-MSCs-CRISPLD2. (A) Fluorescence microscopy images showing the green fluorescence intensities of HUC-MSCs-GFP and HUC-MSCs-CRISPLD2. (B) Relative mRNA expression levels of target genes in HUC-MSCs-GFP and HUC-MSCs-CRISPLD2 were quantified via qRT‒PCR; n = 3. (C) Western blot analysis of CRISPLD2 protein expression in HUC-MSCs-GFP and HUC-MSCs-CRISPLD2. Uncropped blots are shown in Additional file [123]1: Fig. [124]3A. (D) Differentiation potential of HUC-MSCs was assessed via lineage-specific staining: Oil Red O for adipogenesis, Alizarin Red for osteogenesis, and Alcian Blue for chondrogenesis. (E) Flow cytometry analysis of MSC markers (CD73, CD90, and CD105) in HUC-MSCs-GFP and HUC-MSCs-CRISPLD2. (F) Western blot analysis of senescence-associated proteins (p-p53, p53, p21, and p16INK4 A) in HUC-MSCs-GFP and HUC-MSCs-CRISPLD2. GAPDH was used as a loading control. Uncropped blots are shown in Additional file [125]1: Fig. [126]3B. (G) Senescence-associated β-galactosidase (SA-β-gal) staining was performed to evaluate cellular senescence. Representative images of SA-β-gal staining in HUC-MSCs-GFP and HUC-MSCs-CRISPLD2 are shown. (H) Quantification of SA-β-gal-positive HUC-MSCs-GFP and HUC-MSCs-CRISPLD2; n = 6. (I) Cell proliferation was assessed via a CCK-8 assay at 24 h and 48 h; n = 6. (J) Colony formation assay. Representative images of crystal violet-stained dishes and quantification of the colony formation rate (colony numbers per dish normalized to the number of cells); n = 3. The data are presented as the means ± SEMs. The results are combined from 2–3 independent experiments. Statistical significance was determined via an unpaired two-tailed Student’s t test. **p < 0.01, ****p < 0.0001 Since HUC-MSCs senescence often results in altered cellular responses and functional decline, to test whether CRISPLD2 overexpression impacts the senescence of HUC-MSCs, we examined senescence-associated protein expression and found that there was a slight reduction in p-p53 levels but no difference in p53, p21 and p16INK4 A levels in HUC-MSCs-CRISPLD2 and HUC-MSCs-GFP (Fig. [127]2F). The SA-β-gal assay results further suggested that CRISPLD2 overexpression has no effect on the senescence of HUC-MSCs (Fig. [128]2G-H). Finally, to evaluate the effects of CRISPLD2 overexpression on HUC-MSC proliferation, we performed CCK-8 and colony formation assays. The results of the CCK-8 assay revealed that CRISPLD2 overexpression significantly increased HUC-MSC proliferation at the 12-h time point; however, no significant difference in proliferation was observed at the 24-h time point compared with that in the control group (F[129]ig. [130]2I). Furthermore, CRISPLD2 overexpression did not affect the colony-forming ability of HUC-MSCs (Fig. [131]2J). Taken together, these results suggest that CRISPLD2 overexpression does not significantly impact the overall proliferation capacity of HUC-MSCs. Genomic and functional insights into the role of CRISPLD2 in HUC-MSCs To investigate the genomic impact of CRISPLD2 on HUC-MSCs, we performed bulk RNA sequencing on both HUC-MSCs-CRISPLD2 and control HUC-MSCs-GFP. This analysis identified 289 differentially expressed genes, comprising 150 genes whose expression was upregulated and 139 genes whose expression was downregulated in HUC-MSCs-CRISPLD2 (Fig. [132]3A). Fig. 3. [133]Fig. 3 [134]Open in a new tab Bulk RNA sequencing of HUC-MSCs-GFP and HUC-MSCs-CRISPLD2. (A) A total of 289 differentially expressed genes were identified; 150 genes were upregulated, and 139 genes were downregulated. Red represents upregulated genes, and blue represents downregulated genes. (B) Gene Ontology (GO) enrichment analysis of three main domains: biological process (BP), molecular function (MF), and cellular component (CC). (C) GO enrichment analysis revealed significantly altered genes within specific functional categories. (D, E) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of significantly altered genes. (F–I) Gene set enrichment analysis (GSEA) revealed enrichment of ATP-dependent chromatin remodelling activity (F), mitochondrial respiratory chain complex I (G), aerobic respiration (H), and mitochondrial respiratory chain complex I assembly (I). (J) Volcano plots showing different genes encoding transcription factors. Three replicate samples per group were combined Gene Ontology (GO) enrichment analysis revealed that CRISPLD2 overexpression significantly altered genes related to protein binding (LRRCC1, TNFRSF4, LCA5, KLHL4, and NRCAM), membrane (P2RY11, TNFRSF4, TMEM63 C, BEST1 and GPR75), signal transduction (P2RY11, SPX, GPR75 and GPR158), calcium ion binding (CAPN3, CRB1, PCDHA3, THBS4, PLCB2 and PRRG4), and G protein-coupled receptor activity and signalling (P2RY11, GPR37L1, OPRD1, NPBWR1, PLCB2 and GRM5) (Fig. [135]3B, [136]C). Moreover, KEGG pathway analysis revealed that the differentially expressed genes were predominantly associated with the cytokine‒cytokine receptor interaction pathway (e.g., TNFRSF4, IL36G, IL20RA, IL37, and GDF1), the Wnt signalling pathway (e.g., CXXC4, PRKCG, WNT9 A, and PLCB2), and the calcium signalling pathway (e.g., PRKCG, NTRK3, PLCB2, and GRM5) (Fig. [137]3D, [138]E). Importantly, gene set enrichment analysis (GSEA) revealed enriched pathways related to ATP-dependent chromatin remodelling in the upregulated genes (Fig. [139]2F) and enriched pathways associated with mitochondrial respiratory chain complex I, aerobic respiration, and mitochondrial respiratory chain complex I assembly in the downregulated genes (F[140]ig. [141]3G–I). These data suggest that CRISPLD2 potentially plays a role in modulating the function of HUC-MSCs through the remodelling of mitochondrial function. Next, we explored the impact of CRISPLD2 on transcription factor dynamics. We observed an upregulation of CSRNP3 (cysteine-rich protein 3) accompanied by a downregulation of LHX1 (LIM homeobox 1), HOXA2 (homeobox A2), and GRHL3 (grainyhead-like 3) in HUC-MSCs (Fig. [142]3J). The elevated expression of CSRNP3 has been shown to play a role in the stemness and undifferentiated state of HUC-MSCs [[143]52]. In contrast, the downregulation of LHX1, HOXA2, and GRHL3 was associated with reduced differentiation signalling [[144]53, [145]54]. Collectively, these results indicate that CRISPLD2 potentially enhances the immunoregulatory capabilities of HUC-MSCs via the regulation of the GPCR, Wnt, and calcium pathways. Furthermore, the observed changes in transcription factor dynamics suggest that the overexpression of CRISPLD2 suppressed the commitment of HUC-MSCs to specific cell lineages, which is critical for the multipotent characteristics of HUC-MSCs. CRISPLD2 plays an important role in the regulation of the immunosuppressive function of HUC-MSCs in vitro To investigate the impact of CRISPLD2 overexpression on the immunomodulatory function of HUC-MSCs, we examined the expression levels of molecules associated with the immunoregulatory function of HUC-MSCs and inflammatory cytokines via RT‒qPCR. Consistent with the bulk RNA-seq data, CRISPLD2 overexpression in HUC-MSCs significantly increased the expression levels of IL-10 and IDO but decreased those of HGF, VEGF, CCL18, GM-CSF, and PGE2 compared with those in HUC-MSCs expressing GFP (Fig. [146]4A, [147]B). Further analysis revealed a notable increase in the IL-10 concentration in the culture media of HUC-MSCs-CRISPLD2 (Fig. [148]4C). Fig. 4. [149]Fig. 4 [150]Open in a new tab HUC-MSCs-CRISPLD2 exhibit enhanced immunosuppression function in vitro. (A, B) Real-time PCR analysis of the upregulated (A) or downregulated (B) genes in HUC-MSCs-GFP and HUC-MSCs-CRISPLD2. n = 3. (C) The concentrations of IL-10 in the cell culture media of HUC-MSCs-GFP and HUC-MSCs-CRISPLD2 were measured via ELISA. n = 4. (D, E) Naïve mouse T cells were purified using Thy1.2 beads, labelled with CFSE, and then either left unstimulated or activated with anti-CD3/CD28 beads. These T cells were subsequently cocultured with HUC-MSCs-GFP or HUC-MSCs-CRISPLD2 for 72 h. Representative flow cytometry profiles and the percentage of CFSE-negative cells were analysed in four experimental groups: unstimulated T cells, activated T cells, activated T cells cocultured with HUC-MSCs-GFP, and activated T cells cocultured with HUC-MSCs-CRISPLD2 (n = 4–5). (F, G) PBMCs from healthy donors were labelled with CFSE and then either left unstimulated or activated with anti-CD3/CD28 beads. These cells were subsequently cocultured with HUC-MSCs-GFP or HUC-MSCs-CRISPLD2 for 72 h. Representative flow cytometry profiles and the percentage of CFSE-negative cells were analysed across four experimental groups: unstimulated T cells, activated T cells, activated T cells cocultured with HUC-MSCs-GFP, and activated T cells cocultured with HUC-MSCs-CRISPLD2, n = 4–7. The data represent the means ± SEMs. The results from two replicate experiments were averaged. P values were calculated via multiple t tests (A & B), unpaired two-tailed Student’s t tests (C), and one-way ANOVA for multiple comparisons (E & G). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 Given that MSCs can inhibit the proliferation and cytokine production of immune cells such as T cells and B cells [[151]55, [152]56], we explored how CRISPLD2 overexpression affects murine T-cell activation by coculturing CD90.2^+ T cells from C57BL/6 mice with either HUC-MSCs-CRISPLD2 or HUC-MSCs-GFP. Both types of HUC-MSCs inhibited the proliferation of αCD3/CD28-activated CD90.2^+ T cells. However, T-cell proliferation was significantly lower in the HUC-MSCs-CRISPLD2 group than in the HUC-MSCs-GFP group (Fig. [153]4D, [154]E). Moreover, we tested the suppressive effect of HUC-MSCs on human T cells. Accordingly, human PBMCs were cocultured with HUC-MSCs-CRISPLD2 or HUC-MSCs-GFP. Consistent with murine T cells, HUC-MSCs-CRISPLD2 had a greater ability to suppress human T-cell proliferation than HUC-MSCs-GFP did (Fig. [155]4F, [156]G). The enhanced immunosuppressive function of HUC-MSCs-CRISPLD2 is IL-10 dependent Since the liver is a primary target organ in aGVHD, we further investigated whether the increased secretion of IL-10 contributed to the suppression of T-cell cytotoxicity and alleviated alloreactive T-cell-mediated liver damage. Organoid technology has been proven to be instrumental in mimicking alloreactive T-cell-mediated GVHD target organ damage in vitro [[157]57–[158]59]. Therefore, we employed a coculture system comprising human liver organoids and human CD3 + T cells. This innovative setup enabled real-time monitoring of cellular interactions and potential cytotoxicity, providing a dynamic and detailed view of the HLA-mismatched T-cell-mediated immune responses in the liver [[159]60]. Alloreactive T cells caused apoptosis in the liver organoids, as suggested by a significant increase in the caspase signal within the liver organoids at 48 h postcoculture initiation (Fig. [160]5A, [161]B). Compared with that of the organoid-only control group, the observed caspase fluorescence intensity in the treated organoid group markedly increased at 72 and 96 h (Fig. [162]5C, [163]D). These data are consistent with cytotoxic T-cell-mediated liver tissue damage in the context of aGVHD. In contrast, when liver organoids were treated with IL-10, there was a notable reduction in caspase fluorescence intensity compared with that in untreated organoids at 72 and 96 h post coculture initiation (Fig. [164]5E–G). This protective effect of IL-10 was consistent across the various concentrations tested (10 and 40 ng/mL) at 96 h post coculture (Fig. [165]5G), indicating the robust capacity of IL-10 to reduce alloreactive T-cell-mediated apoptosis in liver organoids. Fig. 5. [166]Fig. 5 [167]Open in a new tab The enhanced immunosuppressive function of HUC-MSCs-CRISPLD2 is IL-10 dependent. (A–D) Representative micrographs showing cocultures of human T cells and liver organoids treated with various concentrations of human IL-10 (10, 20, and 40 ng/mL) at different time points (24, 48, 72, and 96 h). Scale bars: 500 µm. n = 2–10. (E) Fluorescence intensity of caspase signals under different IL-10 concentrations (10, 20, and 40 ng/mL) at 24, 48, 72, and 96 h. (F, G) Bar graphs of caspase fluorescence intensity at 72 and 96 h. (H, I) Flow cytometry analysis of human T-cell proliferation. Representative flow cytometry profiles and the percentages of CFSE-negative cells are shown for unstimulated T cells, activated T cells, and activated T cells treated with varying IL-10 concentrations (250 ng/mL, 500 ng/mL, 2 µg/mL, 10 µg/mL, and 20 µg/mL). n = 3–6. The data are presented as the means ± SEMs from two independent experiments. Statistical significance was determined via an unpaired two-tailed Student’s t test (*p < 0.05, **p < 0.01, ****p < 0.0001) To further elucidate the role of IL-10 in suppressing T-cell cytotoxicity, we examined its effects on T-cell proliferation across a range of concentrations (250 ng/mL to 20 µg/mL). While lower concentrations (250 ng/mL, 500 ng/mL, and 2 µg/mL) had no significant inhibitory effects, higher concentrations (10 µg/mL and 20 µg/mL) markedly suppressed T-cell proliferation (F[168]ig. [169]5H, [170]I), indicating a concentration-dependent inhibitory effect. These findings support the hypothesis that IL-10 suppresses alloreactive T-cell activity, contributing to its protective effects against liver organoid damage, underscoring the therapeutic potential of CRISPLD2-overexpressing HUC-MSCs in aGVHD. Injection of HUC-MSCs-CRISPLD2 ameliorated aGVHD while preserving GVL activity Next, we investigated whether the in vivo administration of HUC-MSCs-CRISPLD2 could ameliorate aGVHD and preserve GVL activity [[171]61]. Accordingly, irradiated BALB/c recipients were engrafted with splenocytes (1.25 × 10^6) and T-cell-depleted bone marrow (TCD-BM) cells (5 × 10^6) from major histocompatibility complex (MHC)-mismatched C57BL/6 donors. Recipients given TCD-BM cells alone were used as GVHD-free controls. Recipients were treated with HUC-MSCs-CRISPLD2 or HUC-MSCs-GFP control on Days 0, 9, 16 and 23 after allo-HSCT. Notably, with this dosage of donor splenocytes, recipients treated with HUC-MSCs-GFP did not show a significant increase in overall survival rates; however, they did exhibit a marked reduction in tissue damage (Fig. [172]6A–C). In contrast, the administration of HUC-MSCs-CRISPLD2 significantly improved overall survival, with 80% of the recipients surviving longer than three weeks (Fig. [173]6A). Moreover, HUC-MSCs-CRISPLD2 effectively reduced damage to aGVHD target organs, including the liver, small intestine, and colon (Fig. [174]6B–C). Fig. 6. [175]Fig. 6 [176]Open in a new tab Administration of HUC-MSCs-CRISPLD2 alleviated aGVHD while preserving GVL activity. BALB/c mice were lethally irradiated and transplanted with bone marrow cells and splenocytes from either allogeneic donor C57BL/6 mice or syngeneic donor BALB/c mice. Recipients of allogeneic transplants were further infused with HUC-MSCs-GFP or HUC-MSCs-CRISPLD2 on Days 0, 9, 16, and 23 post-HSCT. Survival was monitored throughout the experiment. (A) Survival curve analysis comparing survival rates among the experimental groups: Syngeneic, GVHD, HUC-MSCs-GFP, and HUC-MSCs-CRISPLD2. n = 15. (B) Haematoxylin and eosin (H&E) staining was used to assess tissue pathology. Representative images of the liver, small intestine, and colon on Day 21 post-transplantation (original magnification: × 200; n = 4). (C) Histopathological scores of the liver, small intestine, and colon. (D) Bioluminescence imaging (BLI) was performed using an IVIS Lumina S5 system (PerkinElmer, Waltham, MA) to measure the tumour burden in the recipients. Representative BLI images are shown. n = 10. (E) Tumour-free survival curves of the experimental groups: Syngeneic, GVHD, HUC-MSCs-GFP, and HUC-MSCs-CRISPLD2. n = 10. (F) Tumour fluorescence intensity quantified as photons/second. The data are presented as the means ± SEMs from 2–3 independent experiments. Statistical analyses were performed via one-way ANOVA (C), the log-rank test for survival comparisons (A & E), or two-way ANOVA for fluorescence intensity comparisons (F). Significance levels: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 To evaluate the impact of HUC-MSCs-CRISPLD2 on GVL activity, BALB/c recipients were inoculated with luciferase-positive A20 cells (A20/Luc; 5 × 10^6 cells/mouse, administered intraperitoneally) concurrently with allo-HSCT. In recipients receiving TCD-BM, A20/Luc tumour cells proliferated rapidly, resulting in the death of 8 out of 10 recipients within 25 days post-HSCT (Fig. [177]6D–F). Without HUC-MSCs administration, mice engrafted with splenocytes and TCD-BM successfully eliminated the tumours; however, most of these mice died from aGVHD within the same 25-day period (Fig. [178]6D–F). Recipients treated with HUC-MSCs-GFP exhibited transient growth of A20/Luc tumour cells but experienced 50% mortality due to aGVHD within 25 days after allo-HSCT (Fig. [179]6D–F). Conversely, HUC-MSCs-CRISPLD2-treated recipients demonstrated only transient A20/Luc tumour cell growth, with successful tumour clearance by Day 24. Remarkably, 80% of these mice survived beyond 25 days without tumour relapse (Fig. [180]6D–F). Together, these findings suggest that HUC-MSCs-CRISPLD2 treatment effectively mitigates aGVHD while maintaining GVL activity. Extracellular ATP and LPS regulated P2Y11R-associated loss of function in HUC-MSCs In the context of aGVHD, tissue damage can lead to the translocation of bacterial products and the release of extracellular ATP, which initiates the innate and adaptive immune response through the MyD88/TLR signalling pathway [[181]62–[182]68]. Elevated extracellular ATP levels, which are mediated by purinergic receptors, such as P2X7R and P2Y2R, have been shown to worsen inflammation and the severity of disease [[183]66, [184]69]. Our bulk RNA sequencing analysis revealed significant downregulation of P2RY11 in HUC-MSCs modified with CRISPLD2 compared with HUC-MSCs modified with the GFP control (Fig. [185]3C). Since P2RY11 encodes the purinergic receptor P2Y11R, we further validated the P2Y11R protein level and found that it was reduced in HUC-MSCs-CRISPLD2 (Fig. [186]7A). Therefore, we hypothesized that ATP and LPS impair HUC-MSC function through the P2Y11R signalling pathway. To address this, we measured changes in ATP and LPS levels following irradiation. We found that extracellular ATP levels in the serum peaked at 4 days postirradiation, paralleling an increase in LPS levels (Fig. [187]7B,C). Consistent with previous studies [[188]69], we confirmed increased serum LPS levels in aGVHD mice, although no significant differences in ATP levels were observed between aGVHD and non-GVHD recipients (Fig. [189]7D,E). Fig. 7. [190]Fig. 7 [191]Open in a new tab Extracellular ATP and LPS regulate P2Y11R-associated loss of function in HUC-MSCs. (A) Western blot analysis of P2Y11R expression in HUC-MSCs-GFP and HUC-MSCs-CRISPLD2. Uncropped blots are shown in Additional file [192]1: Fig. [193]3C. (B, C) Serum was collected from the mice on Days 3 and 4 after X-ray irradiation. Extracellular LPS (B) and ATP (C) concentrations in the serum were quantified via ELISA. Data are presented for irradiated and nonirradiated mice; n = 3–4. (D, E) Serum was collected from mice on Days 3 and 4 post-HSCT. Extracellular LPS and ATP concentrations were measured in the serum of syngeneic HSCT and GVHD mice via ELISA; n = 3–4. (F) The P2Y11R antagonist NF157 was added to the culture medium of HUC-MSCs-GFP. After 48 h, the supernatant was collected, and the IL-10 concentration was quantified via ELISA. The data are shown for HUC-MSCs-GFP, HUC-MSCs-CRISPLD2, and HUC-MSCs-GFP treated with NF157; n = 5–10. (G, H) Naïve mouse T cells were purified using Thy1.2 beads, labelled with CFSE, and either left unstimulated or activated with anti-CD3/CD28 beads. These T cells were cocultured with HUC-MSCs-GFP, HUC-MSCs-CRISPLD2, or HUC-MSCs-GFP and treated with NF157 for 72 h. Representative flow cytometry profiles and the percentages of CFSE-negative cells are shown for five experimental conditions: unstimulated T cells, activated T cells, activated T cells cocultured with HUC-MSCs-GFP, activated T cells cocultured with HUC-MSCs-CRISPLD2, and activated T cells cocultured with HUC-MSCs-GFP and treated with NF157; n = 3–9. (I, J) PBMCs from healthy donors were labelled with CFSE and either left unstimulated or activated with anti-CD3/CD28 beads. These cells were cocultured with HUC-MSCs-GFP, HUC-MSCs-CRISPLD2, HUC-MSCs-GFP treated with NF157, or HUC-MSCs-GFP treated with the P2Y11R agonist NF546 for 72 h. Representative flow cytometry profiles and the percentages of CFSE-negative cells are shown for six experimental conditions: unstimulated T cells, activated T cells, activated T cells cocultured with HUC-MSCs-GFP, HUC-MSCs-CRISPLD2, HUC-MSCs-GFP treated with NF157, and HUC-MSCs-GFP treated with NF546; n = 5. (K) Survival curves of mice treated with HUC-MSCs alone or in combination with apyrase compared with untreated controls; n = 10. The data are presented as the means ± SEMs. All the experiments were performed in duplicate. Statistical analyses were performed via multiple t tests (B‒C), two-way ANOVA with multiple comparisons (D‒E), one-way ANOVA with multiple comparisons (F‒J), and the log-rank test for survival analysis (K). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 In vitro experiments revealed that the selective P2Y11R inhibitor NF157 augmented IL-10 secretion from HUC-MSCs (Fig. [194]7F) and improved their suppressive effects on both murine (Fig. [195]7G,H) and human T cells (F[196]ig. [197]7I,J). Conversely, the application of the P2Y11R agonist NF546 diminished these suppressive effects on T-cell proliferation (F[198]ig. [199]7I,J). Moreover, targeting extracellular ATP in the peritoneal space directly via the intraperitoneal administration of apyrase from Days 0 to 4 significantly improved survival compared with that of the groups without apyrase treatment (Fig. [200]7K). Together, these data suggest that ATP and LPS released during tissue damage from aGVHD impair the functionality of HUC-MSCs through the P2Y11R signalling pathway. Targeting P2Y11R is sufficient to improve the efficacy of HUC-MSCs in ameliorating aGVHD. Inhibition of P2Y11R rescued intracellular calcium-mediated mitochondrial fitness in HUC-MSCs The mitochondria of quiescent MSCs typically present with punctate, fragmented, or spherical morphologies and fewer cristae due to minimal energy demands [[201]70]. This morphology aligns with the fundamental role of stem cells in preserving both nuclear and mitochondrial genomes, as well as their epigenomes. In addition, it is generally believed that the mitochondria in MSCs are in an immature state, as suggested by the low levels of OXPHOS, ATP, and ROS [[202]70]. Mitochondrial calcium uptake plays an important role in MSCs functionality, growth, differentiation, and metabolic activity [[203]71–[204]73]. Calcium ions are second messengers that govern MSCs proliferation and differentiation [[205]73]. Calcium influx via the mitochondrial calcium uniporter (MCU) can stimulate processes such as ATP synthesis and trigger signalling pathways essential for stem cell maintenance and function [[206]71]. Dysregulation of mitochondrial calcium uptake impairs the immunoregulatory functions of MSCs [[207]73]. The presence of high calcium concentrations increased mitochondrial respiration and compromised stem cell functional maintenance, along with increased expression of differentiation markers [[208]74]. In contrast, low calcium concentrations promote stem cell maintenance in vitro [[209]75]. The downstream signalling pathway of P2Y11R is the mitochondrial calcium flux pathway [[210]76]. To investigate whether ATP- or LPS-induced P2Y11R activity impacts mitochondrial function in HUC-MSCs, we first assessed the cellular and mitochondrial calcium concentrations. Compared with HUC-MSCs-GFP, HUC-MSCs-CRISPLD2 significantly reduced the cytosolic calcium level (Fig. [211]8A) and modestly decreased the mitochondrial calcium level (Fig. [212]8B). Furthermore, inhibition of P2Y11R in HUC-MSCs-GFP resulted in significant reductions in both cytosolic and mitochondrial calcium concentrations (Fig. [213]8A,B). Morphological analysis of the mitochondria revealed that HUC-MSCs-GFP had densely packed structures, whereas HUC-MSCs-CRISPLD2 mitochondria were perinuclearly localized, spherical, fragmented, and punctate, with fewer cristae (Fig. [214]8C–F). Under ATP/LPS stimulation, the mitochondria in HUC-MSCs-GFP maintained their original shapes, whereas those in HUC-MSCs-CRISPLD2 presented fewer tubular forms and adopted shorter or fragmented shapes (Fig. [215]8C–F). Fig. 8. [216]Fig. 8 [217]Open in a new tab P2Y11R inhibition rescued intracellular calcium-mediated changes in mitochondrial fitness in HUC-MSCs. (A, B) Cells were stimulated with 2 mM ATP for 5, 10, or 15 min. Cytosolic (A) and mitochondrial (B) calcium concentrations were measured in HUC-MSCs-GFP, HUC-MSCs-CRISPLD2, and HUC-MSCs-GFP treated with a P2Y11R inhibitor; n = 3–6. (C) Representative transmission electron microscopy (TEM) images of mitochondria in HUC-MSCs-GFP under unstimulated conditions or following stimulation with ATP and LPS. (D) Quantification of mitochondrial morphological phenotypes in HUC-MSCs-GFP; n = 10. (E) Representative TEM images of mitochondria in HUC-MSCs-CRISPLD2 under unstimulated conditions or following stimulation with ATP and LPS. (F) Quantification of mitochondrial morphological phenotypes in HUC-MSCs-CRISPLD2; n = 9. (G–I) Cells were stained with 1 µM TMRM for 30 min to assess the mitochondrial membrane potential (ΔΨm). The mean fluorescence intensity (MFI) of TMRM was measured in HUC-MSCs-GFP, HUC-MSCs-CRISPLD2, and HUC-MSCs-GFP treated with a P2Y11R inhibitor under the following conditions: unstimulated (G), stimulated with ATP for 1 h (H), or stimulated with LPS for 1 h followed by ATP for 1 h (I). (J–O) Cellular and mitochondrial reactive oxygen species (ROS) levels were assessed using dihydroethidium (DHE) and MitoSOX™ Red staining, respectively. The cells were stained for 30 min, and the MFI of cellular ROS (J, L, N) and the percentage of mitochondrial ROS (K, M, O) were measured in HUC-MSCs-GFP, HUC-MSCs-CRISPLD2, and HUC-MSCs-GFP treated with a P2Y11R inhibitor under the following conditions: unstimulated (J-K), stimulated with ATP for 1 h (L, M), or stimulated with LPS for 1 h followed by ATP for 1 h (N, O). The data are presented as the means ± SEMs. All the experiments were performed in duplicate; n = 6–12. Statistical significance was determined via one-way ANOVA with multiple comparisons (A, B, G–O) and two-way ANOVA (D, F). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 We then stained the cells with MitoTracker Red to assess the mitochondrial membrane potential (ΔΨm). Our results revealed a significant reduction in TMRM levels in both HUC-MSCs-CRISPLD2 and HUC-MSCs-GFP under P2Y11R inhibition under unstimulated conditions (Fig. [218]8G). These findings suggest that both CRISPLD2 overexpression and P2Y11R inhibition promote the maintenance of stemness in HUC-MSCs [[219]77]. However, upon ATP treatment to mimic the early events of aGVHD in vivo, both HUC-MSCs-CRISPLD2 and HUC-MSCs-GFP with P2Y11R inhibition resulted in slight increases in TMRM levels (Fig. [220]8H). Even under more intense ATP and LPS combination treatment, there was not an overall rise in TMRM levels compared with the unstimulated state or ATP treatment alone. Notably, P2Y11R inhibition led to an increase in TMRM levels, while HUC-MSCs-CRISPLD2 maintained TMRM levels comparable to those of HUC-MSCs-GFP (F[221]ig. [222]8I). In addition, while appropriate levels of ROS are necessary for essential cellular functions, excessive ROS can cause oxidative damage, autophagy, and apoptosis [[223]78, [224]79]. Therefore, we examined the cellular and mitochondrial ROS levels in HUC-MSCs. Under unstimulated conditions, HUC-MSCs-CRISPLD2 presented significantly lower cellular and mitochondrial ROS levels than HUC-MSCs-GFP did (Fig. [225]8J,K). The P2Y11R inhibitor NF157 reduced mitochondrial ROS levels without affecting cellular ROS levels (Fig. [226]8J,K). ATP treatment for one hour did not increase the overall cellular or mitochondrial ROS levels, but these levels remained lower in HUC-MSCs-CRISPLD2 than in HUC-MSCs-GFP (Fig. [227]8L,M). Interestingly, HUC-MSCs-GFP with P2Y11R inhibition further reduced mitochondrial ROS without altering cellular ROS (Fig. [228]8L,M). However, intense stimulation with a combination of ATP and LPS led to a dramatic increase in both the cellular and mitochondrial ROS levels in HUC-MSCs-GFP (Fig. [229]8N,O). Notably, HUC-MSCs-CRISPLD2 maintained lower levels of cellular and mitochondrial ROS (Fig. [230]8N,O). P2Y11R inhibition reduced the level of cellular ROS but did not impact the level of mitochondrial ROS (Fig. [231]8N,O). Collectively, our findings demonstrated that CRISPLD2 overexpression in HUC-MSCs reduced mitochondrial calcium accumulation and ROS production. This highlights the mechanism by which the ATP, LPS, and P2Y11R signalling pathways regulate mitochondrial fitness and cellular identity in HUC-MSCs. Discussion HUC-MSCs can reduce inflammation in mice with aGVHD; however, the ability of HUC-MSC infusion to increase survival is inconsistent due to the heterogeneity of MSCs [[232]80]. The lack of a specific lineage marker to distinguish the immunosuppressive HUC-MSCs subset is the major obstacle. Here, we identified CRISPLD2 as a biomarker for HUC-MSCs subsets characterized by advanced immunosuppressive function through systematic analysis of MSCs heterogeneity across various tissue sources. The enrichment of the negative regulation of immune process pathways in the CRISPLD2-positive HUC-MSCs subset emphasizes the potential therapeutic efficacy of CRISPLD2-positive HUC-MSCs in the prevention of aGVHD. These findings align with our earlier findings that HUC-MSCs possess superior immunosuppressive capacities compared with those of MSCs from other tissues [[233]29]. For the first time, we showed that CRISPLD2 is a specific marker for a unique subset of HUC-MSCs with advanced immunosuppressive function and provided compelling evidence of its translational potential via both in vitro and in vivo studies. To define the mechanisms through which CRISPLD2 regulates HUC-MSCs function, we initially focused on genomic analysis. We found that CRISPLD2 gene overexpression in HUC-MSCs led to significant alterations in gene expression profiles related to immune modulation, cellular signalling, and metabolic pathways. The downregulation of the P2RY11 gene and calcium signalling-associated genes suggests a potential link between CRISPLD2 and ATP- and LPS-mediated signalling in the context of aGVHD. Radiation conditioning–induced extracellular ATP is a primary driver of the innate immune response. Acute GVHD injury increases the levels of LPS and ATP [[234]62, [235]69]. However, whether danger signals released early after allo-HSCT cause HUC-MSCs dysfunction has not been explored. Here, we demonstrated that conditioning regimen-induced ATP and LPS release drives HUC-MSCs dysfunction through P2Y11R engagement and ROS activation. We further showed that overexpression of the CRISPLD2 gene in HUC-MSCs significantly reduced P2Y11R expression, thereby prohibiting ATP binding to its receptor P2Y11R. Apyrase administered into the peritoneum, the ATP release site, posttransplant to degrade ATP synergized with HUC-MSCs and significantly improved the survival of allo-HSCT recipients compared to those that did not receive apyrase treatment. Downregulated P2Y11R expression in HUC-MSCs via CRISPLD2 gene overexpression resulted in significantly greater survival than unmodified HUC-MSCs. In vitro pharmacologic targeting of P2Y11R promoted the immunosuppressive function of HUC-MSCs and exhibited effects comparable to those of CRISPLD2 gene editing, indicating its biological effects. However, whether P2Y11R-deficient HUC-MSCs could have a superior effect on the prevention of aGVHD in vivo requires future study. In addition, our in vitro HUC-MSCs and T-cell coculture assays demonstrated the enhanced immunosuppressive capacity of HUC-MSCs-CRISPLD2, as suggested by the significantly lower T-cell proliferation rate in HUC-MSCs-CRISPLD2 than in HUC-MSCs-GFP. Moreover, we observed that CRISPLD2 overexpression in HUC-MSCs increased IL-10 secretion. The addition of exogenous IL-10 to T-cell and liver organoid cocultures significantly reduced the cytotoxic effects of T cells and increased the viability of liver organoid cells. Importantly, inhibiting P2Y11R increased IL-10 secretion from HUC-MSCs and enhanced their capacity to suppress alloreactive T-cell proliferation. These data indicate that IL-10 release is required for optimal CRISPLD2 + HUC-MSCs function. We directly added the cytokine IL-10 rather than HUC-MSCs to the T cells and liver organoid coculture system because HUC-MSCs and liver organoids grow in distinct culture media. In addition, we hypothesized that HUC-MSCs-CRISPLD2 do not reduce GVL activity because of their specific migration pattern towards GVHD target organs but not lymphoid organs. Therefore, future studies to determine the direct protective effect of HUC-MSCs on liver organoids and the migration pattern in vivo are needed. Finally, CRISPLD2 overexpression affected HUC-MSCs Ca^2+ and ROS levels, suggesting a connection between CRISPLD2 and mitochondrial health. An increase in external Ca^2+ levels downregulates the expression of the cell cycle inhibitor p27kip1, therefore promoting the switch of stem cells from the G1 phase to the S phase [[236]73]. Compared with unmodified HUC-MSCs, CRISPLD2-overexpressing HUC-MSCs presented reduced cellular and mitochondrial calcium and lower ROS levels, indicating that CRISPLD2 promotes a healthier mitochondrial profile that is more resistant to damage mediated by ATP and LPS. However, the mechanism by which cellular and mitochondrial calcium channels impact the functions of HUC-MSCs has still not been fully identified. Further investigations into how CRISPLD2 regulates HUC-MSCs mitochondrial function in this regulatory framework will be essential for developing more effective MSCs-based therapies. While our findings highlight the therapeutic potential of CRISPLD2-overexpressing HUC-MSCs, several critical considerations regarding the long-term implications and potential off-target effects of this genetic modification warrant further investigation. First, the safety of genetically modified HUC-MSCs, particularly with respect to their tumorigenicity, must be rigorously evaluated. Although our bulk RNA sequencing data revealed enrichment of cancer-related pathways, no signs of malignant transformation were observed in our experimental murine models. Second, the potential consequences of chronic IL-10 overexpression, such as systemic immune suppression or tolerance, require careful assessment. These effects could increase susceptibility to infections or compromise the GVL effect. While our data suggest preserved GVL activity, extensive longitudinal studies with multiple tumour cell lines and murine models are needed to confirm these findings. In addition to immune modulation, our data indicate that CRISPLD2-overexpressing HUC-MSCs exhibit enhanced mitochondrial integrity under stress conditions, suggesting their potential for prolonged functional persistence in vivo. This characteristic could translate into more durable therapeutic effects. However, the long-term consequences of altered mitochondrial dynamics, including potential effects on cellular metabolism and energy production, remain to be fully elucidated. Conclusions Our study revealed CRISPLD2 as a key regulator of HUC-MSC immunosuppressive function and mitochondrial fitness. By increasing IL-10 secretion, suppressing T-cell proliferation, and mitigating ATP/LPS-induced damage, CRISPLD2 overexpression represents a promising strategy to improve the therapeutic efficacy of HUC-MSCs in aGVHD. These findings not only provide mechanistic insights into the role of CRISPLD2 in modulating immune responses and cellular resilience but also highlight its potential as a biomarker for identifying highly immunosuppressive HUC-MSC subsets. To fully realize the therapeutic potential of CRISPLD2-modified HUC-MSCs, future studies should focus on evaluating their long-term safety, durability, and clinical translation. Additionally, further exploration of the P2Y11R pathway and mitochondrial dynamics may reveal complementary strategies to enhance the functional persistence and therapeutic outcomes of HUC-MSCs in cell-based therapies. Supplementary Information [237]Supplementary material 1^ (2.4MB, docx) Acknowledgements