Adipose-derived stem cells attenuate rheumatoid arthritis by restoring CX3CR1+ synovial lining macrophage barrier

Abstract

Background: Rheumatoid arthritis (RA) is a chronic autoimmune disease and the integrity of CX₃CR1⁺ synovial macrophage barrier significantly impacts its progression. However, the mechanisms driving the dynamic changes of this macrophage barrier remain unclear. Traditional drug therapies for RA have substantial limitations. Mesenchymal stem cells (MSCs)-based cell therapy, especially adipose-derived stem cells (ADSCs), hold therapeutic promise. Nevertheless, the underlying therapeutic mechanism of ADSCs, especially their interactions with CX₃CR1⁺ macrophages, require further investigation.

Methods: To explore the interaction between ADSCs and CX₃CR1⁺ synovial macrophages during barrier reconstruction, underlying the therapeutic mechanism of ADSCs and the mechanisms on the dynamic changes of the macrophage barrier, scRNA-seq analysis was conducted 4 days after ADSCs injection in serum transfer-induced arthritis model mice. The roles of mitochondria transfer and ADSCs transplantation were also explored. Bulk RNA-seq analysis was performed after the co-culture of ADSCs and CX₃CR1⁺ synovial macrophages. To study the in vivo fate of ADSCs, bulk RNA-seq was performed on ADSCs retrieved at 0, 2, 4, and 7 days post-injection.

Results: Intra-articular injection of ADSCs effectively attenuated the pathological progression of mice with serum transfer-induced arthritis. ADSCs gradually adhered to CX₃CR1⁺ macrophages, facilitating the restore of the macrophage barrier, while the absence of this barrier greatly weakened the therapeutic effect of ADSCs. scRNA-seq analysis revealed an Atf3^high Ccl3^high subset of CX₃CR1⁺ macrophages with impaired oxidative phosphorylation that increased during RA progression. ADSCs-mediated reduction of this subset appeared to be linked to mitochondrial transfer, and transplantation of isolated ADSCs-derived mitochondria also proved effective in treating RA. Both bulk RNA-seq and scRNA-seq analyses revealed multiple interaction mechanisms between ADSCs and CX₃CR1⁺ macrophages, including Cd74/Mif axis and GAS6/MERTK axis, which contribute to barrier restoration and therapeutic effects. Furthermore, bulk RNA-seq analysis showed that ADSCs primarily contribute to tissue repair and immune regulation subsequently.

Conclusions: Our results suggest that ADSCs ameliorated the energy metabolism signature of CX₃CR1⁺ lining macrophages and may promote barrier restoration through mitochondria transfer. In addition, we elucidated the fate of ADSCs and the therapeutic potential of mitochondria in RA treatment.

Keywords: Macrophage; Mesenchymal stem cells; Mitochondria; Rheumatoid arthritis.

Fig. 1

Intra-articular injection of ADSCs mitigated RA and restored the CX₃CR1⁺ macrophage barrier in mice STA model. (A) RA clinical score when injecting ADSCs on day 0. (B) RA clinical score when injecting ADSCs on day 2. (C) Flow diagram of the in vivo injection experiment. (D) Swelling degree of mouse hind paws on day 8. (E) HE and SO/FG stanning results of histological sections of mouse knee and ankle joints on day 8. Scale bar, 200 μm. (F) The mRNA levels of pro-inflammatory factors in knee joint synovium on day 8. (G) The levels of pro-inflammatory and anti-inflammatory factors in peripheral blood on day 8. (H) Cryosection immunofluorescence stanning of CX₃CR1⁺ lining macrophages in the knee joints of ADSCs and PBS injected mice using confocal laser scanning microscopy (CLSM) on day 8. Scale bar, 30 μm

Fig. 2

ADSCs and CX₃CR1⁺ lining macrophages co-localize both in vivo and in vitro. (A) Flow diagram of the strategy used to deplete CX₃CR1⁺ lining macrophages and in vivo injection experiment. (B) RA clinical score after the depletion of CX₃CR1⁺ lining macrophages. (C) Immunofluorescence stanning observation of the localization of ADSCs and CX₃CR1⁺ lining macrophages in knee joints using CLSM on days 1–4 after ADSCs injection (on days 3–6). Scale bar, 50 μm. (D) Flow cytometry sorting of CX₃CR1⁺ lining macrophages from animal tissues. (E) Co-culture of CX₃CR1⁺ lining macrophages and ADSCs at a 1:1 ratio. Scale bar, 400 μm. (F) The spontaneously formed linear structure of CX₃CR1⁺ lining macrophages in the co-culture system at a 1:1 ratio. Scale bar, 400 μm. (G-H). Cell length and cell area of CX₃CR1⁺ lining macrophages in different culture system

Fig. 3

Bulk transcriptome changes of ADSCs and CX₃CR1⁺ lining macrophages after co-culture. (A) Cell viability difference between separately cultured and direct co-cultured cells after 96 h at 1:1 ratio. (B) GO analysis of upregulated genes in co-cultured ADSCs vs. separately cultured ADSCs (log2FC > 1, Padj < 0.05). (C) GO analysis of upregulated genes of co-cultured CX₃CR1⁺ lining macrophages vs. individually cultured CX₃CR1⁺ lining macrophages (log2FC > 1, Padj < 0.05). (D) CellChat chord diagram of cell-cell communication between ADSCs and CX₃CR1⁺ lining macrophages based on the bulk RNA-seq data (Padj < 0.05)

Fig. 4

ScRNA-seq analysis on the heterogeneity change of CX₃CR1⁺ synovial lining macrophages in the barrier disruption and restoration process. (A) Annotated tSNE map of the subpopulations of synovial CD45⁺ CD11b⁺ LY6G⁻ cells. (B-C) tSNE map of CX₃CR1⁺ lining macrophage subpopulations on STA days 1, 2, and 5. (D) Heatmap of differentially express genes in five subpopulations of CX₃CR1⁺ lining macrophages (Padj < 0.05, top 10 most significantly differentially expressed). (E) The ratio of five cell subpopulations on STA days 1, 2, and 5. (F) GSEA analysis results showing hallmark changes of four cell subpopulations compared to Ccl24^high Fn1^low macrophages (Padj < 0.05). (G) Flow cytometry strategy for sorting synovial CD45⁺ CD11b⁺ LY6G⁻ cells. (H) The relative ratios of five cell subpopulations between the ADSCs and PBS injection groups. (I) Atf3/Cox5/Immt immunofluorescence stanning of knee joints of ADSCs and PBS injected mice on day 8

Fig. 5

Mitochondrial transfer from ADSCs during the RA treatment process. A-E used CX₃CR1⁻ synovial macrophages and F used CX₃CR1⁺ synovial macrophages. (A) Transfer of mitochondria from WT-ADSCs (no fluorescence) stained with MitoTracker-Green to CX₃CR1⁻ synovial macrophages in vitro. Scale bar, 20 μm. (B). Transfer of mitochondria from EGFP-ADSCs (EGFP fluorescence) stained with MitoTracker-Red to CX₃CR1⁻ synovial macrophages in vitro. Scale bar, 20 μm. (C) Inhibition of mitochondrial transfer from ADSCs by pretreatment with 10 µM nocodazole for 6 h. Scale bar, 40 μm. (D) Red fluorescence intensity of single CX₃CR1⁻ lining macrophage in different groups. (E) Mitochondrial transfer from EGFP-Cox8-ADSCs to CX₃CR1⁻ macrophages. Scale bar, 15 μm. (F) Mitochondrial transfer from EGFP-Cox8-ADSCs to CX₃CR1⁺ lining macrophages. Scale bar, 30 & 10 μm, from left to right. (G) CLSM observation of mitochondrial transfer from EGFP-Cox8-ADSCs to synovial cells in vivo on STA day 4. Scale bar, 20 μm. (H) Isolated mitochondria of ADSCs stained with MitoTracker-Green. Scale bar, 100 μm. (I) RA clinical score changes when the TNTs of ADSCs were inhibited using nocodazole. (J) The role of the mitochondrial transplantation from ADSCs in RA treatment

Fig. 6

In vivo fate of ADSCs during RA treatment. (A) ADSCs were sorted by flow cytometry. (B) ADSCs retrieval efficacy on days 2, 4, and 7. (C) PCA analysis of ADSCs retrieved at different timepoints. (D) Heatmap of differentially expressed genes among ADSCs retrieved at different days. (E) GO analysis of gene cluster A. (F) GO analysis of gene cluster B. (G) GO analysis of gene cluster C. (H-L) Enriched GO terms of upregulated genes in (H) 2-day-ADSCs vs. 0-day-ADSCs, (I) 4-day-ADSCs vs. 2-day-ADSCs, (J) 7-day-ADSCs vs. 4-day-ADSCs, and downregulated genes in (K) 7-day-ADSCs vs. 2-day-ADSCs, (L) 7-day-ADSCs vs. 0-day-ADSCs. Differentially Expressed genes included in E-L were screened using the criteria log2FC > 1, Padj < 0.05