Abstract Activating autologous stem cells after the implantation of biomaterials is an important process to initiate bone regeneration. Although several studies have demonstrated the mechanism of biomaterial‐mediated bone regeneration, a comprehensive single‐cell level transcriptomic map revealing the influence of biomaterials on regulating the temporal and spatial expression patterns of mesenchymal stem cells (MSCs) is still lacking. Herein, the osteoimmune microenvironment is depicted around the classical collagen/nanohydroxyapatite‐based bone repair materials via combining analysis of single‐cell RNA sequencing and spatial transcriptomics. A group of functional MSCs with high expression of matrix Gla protein (Mgp) is identified, which may serve as a pioneer subpopulation involved in bone repair. Remarkably, these Mgp high‐expressing MSCs (Mgp ^hiMSCs) exhibit efficient osteogenic differentiation potential and orchestrate the osteoimmune microenvironment around implanted biomaterials, rewiring the polarization and osteoclastic differentiation of macrophages through the Mdk/Lrp1 ligand–receptor pair. The inhibition of Mdk/Lrp1 activates the pro‐inflammatory programs of macrophages and osteoclastogenesis. Meanwhile, multiple immune‐cell subsets also exhibit close crosstalk between Mgp ^hiMSCs via the secreted phosphoprotein 1 (SPP1) signaling pathway. These cellular profiles and interactions characterized in this study can broaden the understanding of the functional MSC subpopulations at the early stage of biomaterial‐mediated bone regeneration and provide the basis for materials‐designed strategies that target osteoimmune modulation. Keywords: biomaterials, mesenchymal stem cells, osteoimmune microenvironment, single‐cell transcriptomics, spatial transcriptomics __________________________________________________________________ This study reveals how collagen/nanohydroxyapatite‐based bone repair materials regulate the osteoimmune microenvironment via combining single‐cell RNA sequencing and spatial transcriptomics analyses. The Mgp high‐expressing MSCs with efficient osteogenic differentiation potential are found to be aggregated around implanted materials at the early stage of bone regeneration, regulating the polarization and osteoclastic differentiation of macrophages through the Mdk/Lrp1 ligand–receptor pair. graphic file with name ADVS-11-2308986-g002.jpg 1. Introduction Bone defects suffering from trauma, infections, tumors, or congenital disorders still present a major global health concern.^[ [50]^1 ^] With the rapid development of biomaterials and in‐depth research of stem cells, it is possible to achieve bone regeneration through tissue engineering technology.^[ [51]^2 ^] Generally, stem cells, biomaterials, and growth factors are the three major elements of bone tissue engineering.^[ [52]^3 ^] Nevertheless, inadequate separation and expansion efficiency, unequal in vitro differentiation potential, as well as safety issues of stem cells hampered their further clinical translations.^[ [53]^4 ^] Uncovering the cellular and molecular mechanisms of biomaterials interacting with autologous stem cells is more alluring to devise efficient strategies for bone regeneration.^[ [54]^5 ^] Meanwhile, the bone repair microenvironment is composed of heterogeneous cell populations including mesenchymal stem cells (MSCs) and immune cells with complex phenotypes and different functions.^[ [55]^6 ^] It is necessary to reveal the spatiotemporal response of functional subsets for enhanced bone regeneration at a high resolution, which could serve as the theoretical foundation for the development of efficient bone repair materials. Over the past decade, the ongoing technological revolution has enabled to definition of the gene expression patterns of single cells and facilitates to dissecting of relevant cellular mechanisms that were previously hidden.^[ [56]^7 ^] In particular, single‐cell RNA‐sequencing (scRNA‐seq) has been developed to classify the cell heterogeneity, lineage tracing, and functions within the osteoimmune microenvironment for bone homeostasis or disease in vivo at a high resolution.^[ [57]^8 ^] To date, LepR ^+ MSCs, Pdgfra ^+ MSCs, and Ctsk ^+ MSCs have been reported to play essential roles in bone development and bone healing.^[ [58]^9 ^] Besides, it has been reported that Msx1 ^+ skeletal stem cells could be recruited by neurotrophic supplements and Krt14 ^+ Ctsk ^+ osteoprogenitors govern the maxillofacial bone homeostasis after maxillary sinus floor lifting surgery for enhanced bone regeneration, while their immunoregulatory functions have not been explored.^[ [59]^10 ^] Additionally, despite the cellular marker of functional MSCs, their spatial distributions around biomaterials are still poorly investigated. Spatial transcriptomics (ST) technology has been developed to elucidate the cellular and spatial heterogeneity within the area of interest by positioning histological sections on arrayed reverse transcription primers with unique positional barcodes.^[ [60]^11 ^] Taken together, the combination of scRNA‐seq and ST technologies could provide a comprehensive spatiotemporal landscape of the characteristic distribution and gene expression patterns of MSC subsets and immune cells, offering a broadly applicable strategy to systemically map the whole microenvironment around implanted biomaterials in vivo.^[ [61]^12 ^] In this study, the classical collagen/nanohydroxyapatite (nHA)‐based bone repair materials, which are basic components of natural bone tissue and have exhibited great potential for clinical translation, were utilized to depict the spatiotemporal transcriptomics map of microenvironment at the bone defect region after implantation of biomaterials.^[ [62]^13 ^] The cellular heterogeneity, physiological functions of MSCs and immune cells as well as their communication at the early stage of bone regeneration were comprehensively characterized via the combined analysis of scRNA‐seq and ST technologies. Remarkably, an MSCs subpopulation with a high expression level of matrix Gla protein (Mgp) was detected with significantly increased migration to the bone defect regions and progressively participated at the early stage of biomaterial‐mediated bone regeneration. They also had close cell–cell interactions with macrophages and performed potential immunomodulatory properties, which could mitigate the local inflammatory response. This study further deepens our understanding of the single‐cell spatiotemporal characterization of autologous cells regulated by implanted biomaterials, and the obtained results are expected to improve and optimize the design of cell‐targeted bone repair materials. 2. Results 2.1. The Bioactive Collagen/Nanohydroxyapatite Hydrogel Composites Enhanced New Bone Formation in Calvarial Defects Collagen and nHA are two major constituents of natural bone tissue. Herein, a bioactive hydrogel composite was fabricated via incorporating nHA with type I collagen (Col). Transmission electron microscope (TEM) analysis showed that nHA particles were 80–100 nm in length and 15–20 nm in width (Figure [63] 1A), which is similar to the hydroxyapatite nanocrystals in natural bone tissue (≈40–60 nm long, 20 nm wide).^[ [64]^14 ^] The scanning electron microscopy (SEM) images showed that Col and Col+nHA hydrogels exhibited porous structures with good connectivity (Figure [65]1A). The Col+nHA hydrogel composite embedded with nHA particles had a looser internal micro‐structure with an average pore size ≈200 µm. The micro‐computed tomography (micro‐CT) reconstructions after 12 weeks of observation indicated that Col+nHA hydrogel composite assisted bone regeneration (Figure [66]1B). The quantitative analysis of new bone formation further showed a significantly higher bone mineral density and bone volume in the Col+nHA group compared with the Blank group. Consistently, the histological analysis revealed that the implanted materials were almost completely degraded after 12 weeks of observation in both Col and Col+nHA groups (Figure [67]1C). The implantation of Col+nHA resulted in significantly more new bone formation, while the majority of the defect region was filled with fibrous connective tissue, and few bone fragments were formed in the Blank and Col groups. The nHA group exhibited some immature bone tissue and undegraded nHA particles, which also could be observed in micro‐CT images. In a word, this typical Col+nHA hydrogel composite has a reliable capability to repair bone defects. Figure 1. Figure 1 [68]Open in a new tab Expression landscape of mice calvarial bone defects repaired with or without bioactive Col+nHA hydrogel composites by scRNA‐seq. A) Schematic illustration for fabrication of Col+nHA hydrogel composites and a diagrammatic sketch of the scRNA‐seq analysis. B) Micro‐CT evaluation and quantitative analysis of bone regeneration in defect areas after 12 weeks’ observation (n = 6). BMD: Bone mineral density, BV/TV: bone volume/tissue volume, Tb.N.: trabecular number. C) Histological analysis of calvarial defects 12 weeks after implantation. Cross‐sections were stained with hematoxylin and eosin, Masson's trichrome, and Safranin O‐fast green to observe new bone formation. *: undegradable nHA particles; arrows: the newly formed bone tissue. D) Cells identified by scRNA‐seq were visualized with a t‐distributed stochastic neighbor embedding (t‐SNE) plot. Different cell populations were defined and distinguished by color. B: B cell. DC: dendritic cell. EC: endothelial cell. Mac: macrophage. MSC: mesenchymal stem cell. Myeloid prog: myeloid progenitor. Neu: neutrophil. NK: natural killer cell. Pro B: Pro B cells. T: T cells. E) Specific expression of marker genes in different cell types. F) The expression levels of marker genes were projected onto the t‐SNE atlas. 2.2. Single‐Cell Landscape of Mice Calvarial Bone Defects Repaired with or without Biomaterials To investigate the biological reactions induced by bone repair materials at the early stage of bone healing, the regenerative tissue in calvarial defect regions repaired with or without Col+nHA hydrogel composites were dissected for enzymatic digestion and subjected to scRNA‐seq (Figure [69]1D). A total number of 65722 cells were obtained. We preprocessed the dataset with the Seurat package. Uniform Manifold Approximation and Projection (UMAP) and t‐distributed Stochastic Neighbor Embedding (t‐SNE) were calculated to visualize cell heterogeneity in reduced dimensions. As shown in Figure [70]1E, cells were divided into 14 clusters based on classic cell surface markers. According to these marker genes, these clusters were annotated as myeloid progenitors (Mpo); neutrophils (S100a8); macrophages (Csfr1); MSCs (Col1a1); endothelial cells (Cdh5); osteoclasts (Ctsk); T cells (Cd3g); B cells (Cd79a); pro B cells (Vpreb1); two clusters of dendritic cells (DC, H2‐Aa), expressing higher levels of Cd74 and Siglech; natural killer cells (NK, Gzma); basophils (Prss34); neuron (Ttr) (Figure [71]1F). The top three expressed genes of each cluster were identified and compared in Figure [72]S1A (Supporting Information). As immune cells and MSCs are highly plastic and can be recruited and polarized to different states depending on the microenvironment, we analyzed the cell ratios of the above clusters among defect areas repaired with or without hydrogel composites (Figure [73]S1B,C, Supporting Information). The MSCs from the natural healing condition accounted for only 32.6% of the total cells, while 67.4% of MSCs were from samples implanted with hydrogel composites. The ratio of macrophages, NK cells, Cd74 ^hi DC cells, and osteoclasts was also higher in samples implanted with bioactive materials, and the proportion of neutrophils was higher in blank samples. The proportion of macrophages and neutrophils was higher on day 2 compared with day 7, which was consistent with the time course of innate immune response (Figure [74]S1D,E, Supporting Information). These results reflected enhanced immunoregulatory and bone remodeling processes during the biomaterial‐mediated bone regeneration. Since the MSCs have been defined as multipotent stem cells and play essential roles during the process of bone regeneration, we subsequently distinguished MSCs into nine major subsets (Figure [75] 2A). These subsets were named based on the reported marker genes of stem cells: five subsets of osteoprogenitors (OPs) expressing Cxcl12, Col1a2, Pdgfra, and Runx2; pre‐osteoblasts (pre‐OBs) expressing Sp7, Alpl, Bglap; osteoblasts (OBs) with highly expressing abundance of osteogenic genes Runx2, Alp, Sp7, Bglap and hardly expressing Cxcl12; pericytes, expressing Rgs5, Emcn, Col4a1; myeloid progenitors (MPs), expressing S100a8, S100a9, Cxcl2 (Figure [76]2B,C).^[ [77]^12 , [78]^15 ^] Figure 2. Figure 2 [79]Open in a new tab Heterogeneity of MSCs at the early stage of bone regeneration. A) Visualization of MSCs in Figure [80]1 with UMAP plot, which was divided into nine subclusters. MPs: myeloid progenitors. OPs: osteoprogenitors. OBs: osteoblasts. Pre‐OBs: pre‐osteoblasts. B) Dot plot showing the differentially expressed genes (DEGs) of different cell types. C) Violin plots showing the log‐normalized expression levels of curated feature genes in nine subclusters of MSCs. D) Enriched gene ontology (GO) terms of DEGs among the nine MSCs subcluster. E) Radar map showing the indicated function and metabolic pathway among each MSC subcluster. F) RNA velocity of MSCs estimated from unspliced and spliced transcripts of nearby cells and visualized on UMAP plot. G) Pseudotime lineage trajectory analysis indicating the direction of pseudotime and H) demonstrating the relationships of MSCs subclusters, color‐coded by subclusters. I) The expression dynamics of proliferative and osteogenic‐related genes in pseudotime. Gene ontology (GO) enrichment analysis revealed distinct functions of MSC subsets related to cell adhesion, cell proliferation, response to wounding, chemotaxis, blood vessel morphogenesis, and ossification, which were evaluated using a gene set (Figure [81]2D,E). The results showed that OPs2, OPs3, OPs4, Pre‐OBs, and OBs cells were featured by skeletal system development and ossification, and played essential roles in response to wounding. Based on the RNA velocity and pseudotime analysis, we uncovered the osteogenic differentiation trajectory of MSC subsets, originating from the OPs1 cells and then split into two main branches toward OPs2 and OBs (Figure [82]2F–H; Figure [83]S2, Supporting Information). The patterns of proliferative and osteogenic biomarkers on the trajectory axis were shown as control (Figure [84]2I). OBs were found at the terminal end of the trajectory (cell fate 2 branch) with high expression of Bglap and Ibsp. Unlike the OBs subset, the OPs2 subset located at the terminal end of cell fate 1 branch highly expressed Cxcl12 and Alpl, showing a high transcriptomic similarity to previously described osteo‐CAR (CXCL12‐abundant reticular) cells.^[ [85]^15b ^] 2.3. The Combined Analysis of scRNA‐Seq and ST‐Seq Revealed that OPs2 (Mgp ^hiMSCs) Play Essential Roles at the Early Stage of Bone Regeneration To explore the spatial characteristics of cell types during the process of bone regeneration, we applied ST‐seq to compare the spatial gene expression profiles between bone defect regions repaired with or without bioactive materials. After hematoxylin‐and‐eosin (HE) staining and brightfield imaging, the bone slides were subjected to distinguish the anatomical features, and a total of 4373 spatially barcoded spots were captured in two groups (Figure [86] 3A; Figure [87]S3, Supporting Information). Standard quality control and dimensionality reduction were performed using Seurat methods, and visualization was realized through UMAP (Figure [88]3B). The clusters were annotated based on their histological features and differentially expressed genes (DEGs), including fibrous connective tissue (FCT), regenerative tissue (RT), muscle (Mus), mature bone tissue (MBT), inflammatory connective tissue (ICT), epidermis (Epi), as well as adipose tissue (AT) (Figure [89]S4A, Supporting Information). Especially, the RT cluster also showed a higher expression level of Mgp, the marker gene of the OPs2 subset, as compared to other clusters (Figure [90]3C,D). The GO and Reactome analysis revealed that MBT, RT, Mus1, and FCT3 clusters expressed higher enrichment levels of extracellular matrix organization and collagen formation, responding to tissue damage (Figure [91]S4B, Supporting Information). The RT cluster in the Col+nHA group activated more immune responses, especially innate immune responses, compared with the blank group, positively regulated the cell migration, enhancing the antigen processing and presentation, mediating the extracellular matrix organization process (Figure [92]S4C, Supporting Information). Figure 3. Figure 3 [93]Open in a new tab Tracing the spatial distribution of cell types within microenvironment of bone regeneration. A) The unsupervised clustering indicated the calvarial bone defects were repaired with or without bioactive Col+nHA hydrogel composites. Clusters with the same annotation are merged and assigned the same color code. FCT: fibrous connective tissue. RT: regenerative tissue. Mus: muscle. MBT: mature bone tissue. ICT: inflammatory connective tissue. Epi: epidermis. AT: adipose tissue. B) UMAP plot performed on the gene expression data from spots covered by tissue. K‐means clustering analysis has identified 12 types of spots, which have been assigned a color each. C) Heatmap showing the differentially expressed genes (DEGs) of different spot types. D) Spatial feature plots showed the normalized expression of Mgp in tissue sections. E) UMAP and violin plots showing the log‐normalized expression levels of Mgp in MSCs subsets. F) The spatial feature plot showed the locations of the ST cluster and defined subclusters of MSCs in tissue sections. G) Crystal violet staining showed the migrated hBMMSCs transfected with vector, hBMMSCs transfected with MGP‐overexpressing lentivirus (MGP), as well as hBMMSCs transfected with MGP‐overexpressing lentivirus and then interfered with siMGP‐1 (MGP+siMGP). The numbers of migrated cells per microscopic field were counted via Image J software (n = 6, **: p < 0.01.). H) Immunohistochemistry staining of Mgp ^hi cells around calvarial defect regions after 7 days of hydrogel implantation. Furthermore, the combined analysis of scRNA‐seq and ST‐seq databases was performed to distinguish the functional MSC subcluster in RT. Mgp, the marker gene of the RT cluster in the ST database, was identified to be OPs2‐specific after calculating DEGs by comparing OPs2 to the rest of MSC subsets (aveLog[2]FC = 2.331, p = 5.011e−215). While, the expression distribution diagram of the MSCs showed that Mgp was highly expressed in the subsets of OPs2, pre‐OBs, partial cells of OPs1, and pericytes (Figure [94]3E). Then, we further mapped these MSC subclusters based on DEGs from the scRNA‐seq database to the calvarial defect tissue in the ST‐seq database. Scoring spots in each section showed that the OPs2 exhibited strong spatial preferences within the