Bioactive siRNA-Based Liposomes Promoted Tendon-Bone Healing in Osteoporotic Mice by Recovering the Stemness of CD248+ TSPCs

Abstract

Osteoporosis significantly impairs tendon-bone healing, increasing the risk of rotator cuff tears and postoperative retearing. Tendon stem/progenitor cells (TSPCs) are vital for tendon repair, with their stemness crucial to healing outcomes. This study investigated the role of CD248 in regulating TSPC stemness and assessed the therapeutic potential of si-CD248-loaded liposomes in promoting tendon-bone healing under osteoporotic conditions. Single-cell RNA sequencing (scRNA-seq) identified increased TSPC populations, particularly a unique subcluster, TSPC-0, with elevated CD248 expression in osteoporotic tendon samples. CD248⁺ TSPCs displayed reduced proliferation, increased apoptosis, and impaired migration, driven by altered FAK-JAK-STAT1 signaling. si-CD248-loaded liposomes were formulated and characterized, demonstrating efficacy in inhibiting CD248 expression, restoring TSPC stemness, and promoting tendon-bone healing. In osteoporotic mice, liposome treatment significantly enhanced tissue regeneration, improving histological scores, collagen organization, and biomechanical properties. This study reveals that elevated CD248 expression negatively impacts TSPC stemness and impairs healing under osteoporotic conditions. Targeting CD248 using si-CD248-loaded liposomes effectively restores TSPC regenerative potential, representing a promising therapeutic strategy to enhance tendon-bone healing in osteoporotic patients.

Keywords: CD248; TSPCs; liposomes; osteoporosis; single cell; stemness; tendon‐bone healing.

Figure 1

Single‐cell RNA sequencing analysis of overall cell types. a) UMAP plot showing cell types. This UMAP plot illustrates the clustering of different cell types identified through single‐cell RNA sequencing. Different colors represent various cell types, including tendon stem/progenitor cells (TSPCs), endothelial cells, macrophages, neutrophils, monocytes, fibroblasts, smooth muscle cells, nerve cells, and erythrocytes. b) Feature plots of subtypes in the UMAP plot. c) UMAP plot comparing cell distributions between the control and osteoporotic groups. This UMAP plot compares the distribution of cell types between the control group (NC_T) and the osteoporotic model group (OP_T), with a particular focus on the changes in TSPCs. d) Quantification of TSPC numbers. This bar chart quantifies the number of TSPCs in the control and osteoporotic model groups, showing a fourfold increase in TSPCs in the osteoporotic model group. e) Interaction network of TSPCs with other cell types. This network graph shows the potential ligand‒receptor interactions between TSPCs and other cell types. The direction of the arrows indicates the direction of signal transmission, with color and width representing the strength of interactions. f) Heatmap of potential ligand‒receptor pairs between cell types. This heatmap displays the number of potential ligand‒receptor pairs between different cell types, highlighting the interactions between TSPCs and other cell types. g) Detailed interaction map of TSPCs with other cell types. This interaction map provides a detailed view of the specific ligand–receptor pairs between TSPCs and other cell types, offering insights into cellular communication mechanisms. h) Dot plot of specific ligand–receptor pairs between cell types. This dot plot details the specific ligand–receptor pairs identified between TSPCs and other cell types, with dot size and color representing expression levels and significance.

Figure 2

Subcluster analysis of single‐cell RNA sequencing data. a) t‐SNE plot showing subclusters of TSPCs. This t‐SNE plot illustrates the identification of distinct subclusters within the dataset, with each color representing a different subcluster (0–13). b) Feature plots of subclusters in t‐SNE. c) Heatmap of gene expression profiles across subclusters. The heatmap shows the gene expression profiles of the different subclusters, highlighting the distinct transcriptional signatures associated with each subcluster. d) Pie charts comparing the proportion of subclusters between the control and osteoporotic groups. These pie charts illustrate the relative abundance of each subcluster in the control group (NC_T) and the osteoporotic model group (OP_T), indicating significant shifts in cellular composition. e) Violin plots of gene expression in subclusters. The violin plots display the distribution of the top 10 gene expression levels for selected genes across the subclusters, providing insights into the functional heterogeneity within each subcluster. f) Gene Ontology (GO) and pathway enrichment network. This network diagram highlights the enriched GO terms and pathways associated with the DEGs in the subclusters, revealing key biological processes and signaling pathways. g) Bar charts of pathway enrichment analysis. These bar charts show the results of pathway enrichment analysis for the subclusters, including GO molecular function (GO‐MF), KEGG, Reactome, and hallmark pathways, underscoring the biological relevance of the identified subclusters.

Figure 3

Pseudotime analysis of single‐cell RNA sequencing data. a) UMAP plot of cell clusters from the HPA database. This UMAP plot displays the clustering of cells into distinct groups based on their gene expression (CD248) profiles. Each color represents a different cell type, with annotations indicating specific clusters (e.g., C0, C1, C2). b) Violin plots of gene expression across TSPC subtypes. These violin plots illustrate the distribution of gene expression levels for the top 10 marker genes across different TSPC subtypes. The median expression in each group is highlighted to show variations in gene expression. c,d) Pseudotime trajectory analysis. e) Pseudotime trajectory analysis (colored by sample group). This trajectory plot is colored by sample group (e.g., NC_T and OP_T), showing how cells from different sample groups progress through pseudotime. f) Ridge plot of cell density along pseudotime. This ridge plot represents the density of cells along the pseudotime trajectory for each TSPC subtype, highlighting the distribution and transitions of subtypes over pseudotime. g) Pseudotime trajectory colored by lineage. This plot illustrates the pseudotime trajectory of cells colored by their inferred lineage, showing the developmental pathways of different cell lineages. h) Heatmap of gene expression along pseudotime. The heatmap displays the dynamic changes in gene expression along the pseudotime trajectory. Each row represents a gene, and each column represents a position along pseudotime, with colors indicating expression levels. i) Scatter plot of Col1a1 expression over pseudotime. This scatter plot shows the expression levels of the Col1a1 gene over pseudotime. Each dot represents a cell, colored by pseudotime, with a trend line indicating the overall expression trend. j) Scatter plot of Col1a1 expression by subcell type over pseudotime. Similar to panel i, this scatter plot displays Col1a1 gene expression over pseudotime, with cells colored according to their respective cell types, highlighting differences in expression patterns among cell types.

Figure 4

CD248+ TSPCs Showed Low Stemness In Vitro. a) a1‐Schematic diagram of cell processing. a2‐EdU staining for evaluating the proliferation rate of TSPCs and CD248‐positive TSPCs. The scale bars represent 100 µm. b) The proliferation rate of TSPCs was statistically analyzed (n = 3). The data are presented as the means ± SDs. c) Cell cycle analysis showing the proliferation status of TSPCs and CD248‐positive TSPCs. d) S and G2 phases of TSPCs and CD248‐positive TSPCs were statistically analyzed (n = 3). The data are presented as the means ± SDs. e) The percentages of apoptotic TSPCs and CD248‐positive TSPCs were assessed via live/dead staining. The scale bars represent 100 µm. f) The percentages of PI‐positive TSPCs and CD248‐positive TSPCs were statistically analyzed (n = 3). The data are presented as the means ± SDs. g) The percentages of apoptotic TSPCs and CD248‐positive TSPCs were assessed via apoptosis flow cytometry. h) The percentages of TSPC‐ and CD248‐positive TSPCs in the upper right and lower right quadrants were statistically analyzed (n = 3). The data are presented as the means ± SDs. i) Cell scratch analysis for evaluating the migration ability of TSPCs and CD248‐positive TSPCs. The scale bars represent 200 µm. j) Changes in apoptosis‐related protein expression in TSPCs and CD248‐positive TSPCs were analyzed. k) The relative protein expression of Bax, Bcl‐2, and cleaved caspase 3 was calculated after normalization to that of GAPDH (n = 3). The data are presented as the means ± SDs. l) The protein expression of CD248 and the protein phosphorylation of FAK, JNK, and STAT1 in TSPCs and CD248‐positive TSPCs were determined by Western blotting. m) The relative protein expression of CD248, p‐FAK, p‐JNK, and p‐STAT1 was calculated after normalization to that of GAPDH, FAK, JNK, and STAT1 (n = 3). The data are presented as the means ± SDs. n) Changes in the intensity of immunofluorescence staining for CD248 and p‐JNK (green) in TSPCs and CD248‐positive TSPCs. The scale bars represent 50 µm. o) The relative fluorescence intensities of CD248 and p‐JNK were calculated after normalization to those in the TSPCNC group (n = 3). The data are presented as the means ± SDs.

Figure 5

The Stemness of CD248+ TSPCs was Controlled by the CD248/FAK/JNK/Stat1 Signaling Axis. a) a1‐Schematic diagram of cell processing. a2‐Cell cycle analysis showing the proliferation status of TSPCs in the four different groups. b) The sum of the S and G2 phases of TSPCs in the four different groups was statistically analyzed (n = 3). The data are presented as the means ± SDs. c) EdU staining for evaluating the proliferation rate of TSPCs in the four groups. The scale bars represent 100 µm. d) The proliferation rate (red/blue) of TSPCs in the four groups was statistically analyzed (n = 3). The data are presented as the means ± SDs. e) The percentages of apoptotic cells among the four groups were assessed via apoptosis flow cytometry. f) The percentages of cells in the upper right (late apoptotic) and lower right (early apoptotic) quadrants among the four groups were statistically analyzed (n = 3). The data are presented as the means ± SDs. g) The percentages of apoptotic cells in the four groups were assessed via live/dead staining. The scale bars represent 100 µm. h) The percentage of PI‐positive cells (red) among the four groups was statistically analyzed (n = 3). The data are presented as the means ± SDs. i) Changes in the expression of apoptosis‐related proteins among the four groups were analyzed via Western blotting. j) Relative protein expression of Bax, Bcl‐2, and cleaved caspase 3 was calculated after normalization to that of GAPDH (n = 3). The data are presented as the means ± SDs. k) Cell scratch analysis for evaluating the migration ability of TSPCs in the four different groups. The scale bars represent 200 µm. l) The protein expression of CD248 and the protein phosphorylation of FAK, JNK, and STAT1 in the four groups were determined by Western blotting. m) The relative protein expression of CD248, p‐FAK, p‐JNK, and p‐STAT1 in the four groups was calculated after normalization to that of GAPDH, FAK, JNK, and STAT1 (n = 3). The data are presented as the means ± SDs. n) Changes in the intensity of CD248 and p‐JNK (green) immunofluorescence among the four groups are presented. The scale bars represent 50 µm. o) The relative fluorescence intensities of CD248 and p‐JNK were calculated after normalization to those in the TSPCNC group (n = 3). The data are presented as the means ± SDs.

Figure 6

Characterization and efficacy of siRNA‐loaded liposomes for CD248+ TSPC delivery. a) Schematic of the Lipo‐siRNA Assembly and Delivery Mechanism. The schematic illustrates the assembly process of the siRNA‐loaded liposomes (Lipo@si‐CD248), where lipids are combined with siRNA to form the delivery vehicle. siRNA targets CD248 mRNA in CD248+ tendon stem/progenitor cells (TSPCs), inhibiting its expression. b) TEM images of liposomes. TEM images showing the morphology of Lipo@si‐Ctrl (left) and Lipo@si‐CD248 (right). Insets provide higher magnification images demonstrating the spherical shape and uniform size distribution of the liposomes. Scale bars represent 100 nm. c) NTA parameters and images of liposomes. Fluorescence microscopy images indicating successful loading of the siRNAs into the liposomes, as shown by the green fluorescence signals in both the Lipo@si‐Ctrl and Lipo@si‐CD248 samples. The scale bars represent 200 nm. d) Table of Liposome Characterization Parameters. The hydrodynamic diameter, PDI, zeta potential, and entrapment efficiency of Lipo@si‐Ctrl and Lipo@si‐CD248. Both liposome formulations show similar characteristics with high entrapment efficiency (>91%). e) Histological analysis of tissue sections treated with liposomes. Histological images showing tissue sections stained with H&E, indicating the biocompatibility and distribution of Lipo@si‐Ctrl and Lipo@si‐CD248 in different tissues. Scale bars represent 50 µm. f) The in vitro fluorescence imaging data of si‐CD248‐Flu and Lipo@si‐CD248‐Flu after incubation for 12h. The scale bars represent 25 µm. g) The in vivo fluorescence imaging data of si‐CD248‐Flu and Lipo@si‐CD248‐Flu after injection for 12h/24h/48h/5d/10d.

Figure 7

Lipo@si‐CD248 Enhanced the Stemness of CD248+ TSPCs by Inhibiting CD248 In Vitro a) a1‐Schematic diagram of cell processing. a2‐Cell cycle analysis showing the proliferation status of TSPCs in the four different groups. b) The sum of the S and G2 phases of TSPCs in the four different groups was statistically analyzed (n = 3). The data are presented as the means ± SDs. c) EdU staining for evaluating the proliferation rate of TSPCs in the four groups. The scale bars represent 100 µm. d) The proliferation rate (red/blue) of TSPCs in the four groups was statistically analyzed (n = 3). The data are presented as the means ± SDs. e) The percentages of apoptotic cells among the four groups were assessed via apoptosis flow cytometry. f) The percentages of cells in the upper right (late apoptotic) and lower right (early apoptotic) quadrants among the four groups were statistically analyzed (n = 3). The data are presented as the means ± SDs. g) The percentages of apoptotic cells in the four groups were assessed via live/dead staining. The scale bars represent 100 µm. h) The percentage of PI‐positive cells (red) among the four groups was statistically analyzed (n = 3). The data are presented as the means ± SDs. i) Changes in the expression of apoptosis‐related proteins among the four groups were analyzed via Western blotting. j) Relative protein expression of Bax, Bcl‐2, and cleaved caspase 3 was calculated after normalization to that of GAPDH (n = 3). The data are presented as the means ± SDs. k) Cell scratch analysis for evaluating the migration ability of TSPCs in the four different groups. The scale bars represent 200 µm. l) The protein expression of CD248 and the protein phosphorylation of FAK, JNK, and STAT1 in the four groups were determined by Western blotting. m) The relative protein expression of CD248, p‐FAK, p‐JNK, and p‐STAT1 in the four groups was calculated after normalization to that of GAPDH, FAK, JNK, and STAT1 (n = 3). The data are presented as the means ± SDs. n) Changes in the intensity of CD248 and p‐JNK (green) immunofluorescence among the four groups are presented. The scale bars represent 50 µm. o) The relative fluorescence intensities of CD248 and p‐JNK were calculated after normalization to those in the TSPCNC group (n = 3). The data are presented as the means ± SDs.

Figure 8

Lipo@si‐CD248 Promoted Tendon‐bone Healing after Rotator Cuff Injury in Osteoporotic Mice. a) Animal experiment flow chart. b) Histological staining results. The scale bars represent 200 µm. c) Quantification of histological staining. The data are presented as the mean ± SD. *p < 0.05, ***p < 0.001, ****p < 0.0001. d) Immunofluorescence staining results of CD248 on day 7. The scale bars represent 100 µm. e) Quantification of immunofluorescence staining. The data are presented as the mean ± SD. *p < 0.05, ***p < 0.001, ****p < 0.0001. f) Immunofluorescence staining results of α‐SMA on day 7. The scale bars represent 100 µm. g) Quantification of immunofluorescence staining. The data are presented as the mean ± SD. *p < 0.05, ***p < 0.001, ****p < 0.0001.

Figure 9

Lipo@si‐CD248 promoted shoulder functional recovery after rotator cuff injury in osteoporotic mice. a) Gait analysis results. b) Quantification of Gait. The data are presented as the mean ± SD. c) Results of biomechanical experiments. d) Quantification of biomechanical experiments. The data are presented as the mean ± SD.

Figure 10

Schematic depiction of this work. In this study, we found that osteoporosis leads to disturbances in the ECM of tendons and thus causes a decrease in TSPC stemness, resulting in a rotator cuff that is easily damaged and prone to retearing after repair. Targeted intervention with si‐CD248‐loaded liposomes can restore TSPC stemness, promote tendon‐bone healing, and restore shoulder function.