Glucocorticoid-induced loss of beneficial gut bacterial extracellular vesicles is associated with the pathogenesis of osteonecrosis

Abstract

Osteonecrosis of the femoral head (ONFH) commonly occurs after glucocorticoid (GC) therapy. The gut microbiota (GM) participates in regulating host health, and its composition can be altered by GC. Here, this study demonstrates that cohousing with healthy mice or colonization with GM from normal mice attenuates GC-induced ONFH. 16S rRNA gene sequencing shows that cohousing with healthy mice rescues the GC-induced reduction of gut Lactobacillus animalis. Oral supplementation of L. animalis mitigates GC-induced ONFH by increasing angiogenesis, augmenting osteogenesis, and reducing cell apoptosis. Extracellular vesicles from L. animalis (L. animalis-EVs) contain abundant functional proteins and can enter the femoral head to exert proangiogenic, pro-osteogenic, and antiapoptotic effects, while its abundance is reduced after exposure to GC. Our study suggests that the GM is involved in protecting the femoral head by transferring bacterial EVs, and that loss of L. animalis and its EVs is associated with the development of GC-induced ONFH.

Fig. 1.. Cohousing with healthy mice prevents ONFH in GC-treated cage mates.

(A) Schematic diagram of the experimental design for assessing the effects of cohousing with the healthy or MPS-treated mice on the femoral heads of their cage mates. (B to F) μCT-reconstructed images (B) and quantification of Tb. BV/TV (C), Tb. Th (D), Tb. N (E), and Tb. Sp (F). Scale bar, 1 mm. n = 10 per group. (G) H&E staining images of femoral heads. The arrows indicate empty osteocytic lacunae. Scale bars, 200 μm (blue) or 50 μm (black). (H) Blood vessels in femoral heads visualized by μCT-based microangiography. Scale bar, 500 μm. (I) Quantification of total vessel volume. n = 5 per group. (J) CD31 immunostaining images of femoral heads. Scale bar, 25 μm. (K) Quantification of the number of CD31⁺ cells. n = 5 per group. (L) Osteocalcin (OCN) immunostaining images of femoral heads. Scale bar, 25 μm. (M) Quantification of the number of OCN⁺ cells. n = 5 per group. (N) Terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) staining images of femoral heads. Scale bar, 50 μm. (O) Quantification of the number of TUNEL⁺ cells. n = 5 per group. *P < 0.05, **P < 0.01, and ***P < 0.001. i.m., intramuscularly.

Fig. 2.. Colonization with GM from normal mice attenuates GC-induced ONFH.

(A) Schematic diagram of the experimental design for exploring the effects of oral treatment with GM from MPS-treated mice (GM^MPS) or vehicle-treated healthy mice (GM^Vehicle) on the femoral heads of the vehicle- or MPS-treated mice. (B to F) μCT-reconstructed images of femoral heads (B) and quantification of Tb. BV/TV (C), Tb. Th (D), Tb. N (E), and Tb. Sp (F). Scale bar, 1 mm. n = 10 per group. (G) H&E staining images of femoral heads. Scale bars, 200 μm (blue) or 50 μm (black). (H) μCT-reconstructed images of the Microfil-perfused blood vessels in femoral heads. Scale bar, 500 μm. (I) Quantification of total vessel volume. n = 5 per group. (J) CD31 immunostaining images of femoral heads. Scale bar, 25 μm. (K) Quantification of the number of CD31⁺ cells. n = 5 per group. (L) OCN immunostaining images of femoral heads. Scale bar, 25 μm. (M) Quantification of the number of OCN⁺ cells. n = 5 per group. (N) TUNEL staining images of femoral heads. Scale bar, 50 μm. (O) Quantification of the number of TUNEL⁺ cells. n = 5 per group. *P < 0.05, **P < 0.01, and ***P < 0.001. i.g., intragastric

Fig. 3.. Cohousing with healthy mice rescues the GC-induced loss of L. animalis, and transplantation with L. animalis protects against GC-induced ONFH.

(A) Observed number of OTUs and estimated OTU richness (Chao1 and ACE) in fecal microbiota from the cohoused (Ch) or non-cohoused (Non-Ch) vehicle- or MPS-treated mice. n = 3 per group. (B) Relative and absolute abundance of the identified fecal microbiota at the genus level tested by 16S rRNA gene sequencing. n = 3 per group. (C and D) Relative and absolute abundance of the genus Lactobacillus (C) and the species L. animalis and L. intestinalis (D) in fecal microbiota from mice in (A). n = 3 per group. (E) qRT-PCR analysis of L. animalis abundance in fecal microbiota from mice receiving vehicle or MPS for 1, 2, and 3 weeks. n = 5 per group. (F) Schematic diagram of the experimental design for testing the effects of oral treatment with L. animalis on the femoral heads of vehicle- or MPS-treated mice. (G to K) μCT-reconstructed images of femoral heads (G) and quantification of Tb. BV/TV (H), Tb. Th (I), Tb. N (J), and Tb. Sp (K). Scale bar, 1 mm. n = 8 per group. (L) H&E staining images of femoral heads. Scale bars, 200 μm (blue) or 50 μm (black). (M to P) Quantification of total vessel volume (M) and the numbers of CD31⁺ (N), OCN⁺ (O), and TUNEL⁺ (P) cells in femoral heads. n = 5 per group. *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 4.. L. animalis-EVs directly promote angiogenesis, augment osteogenesis, and reduce cell apoptosis.

(A) Morphology of L. animalis-EVs under a transmission electron microscope. Scale bar, 50 nm. (B) Diameter measurement of L. animalis-EVs by nanoparticle tracking analysis. (C) Quantification of vesicle numbers in 100 μg of L. animalis-EVs from five different batches (E1, E2, E3, E4, and E5) by nanoparticle tracking analysis. n = 3 per group. (D) Uptake of the DiO (green)– or DiI (red)–labeled L. animalis-EVs by HMECs, MLO-Y4, mouse preosteoblast MC3T3-E1 cells, and BMSCs. Scale bar, 10 μm. (E) Tube formation images of HMECs treated with vehicle, MPS, or MPS + different concentrations of L. animalis-EVs (L-EVs). Scale bar, 200 μm. (F and G) Quantification of total loops (F) and total tube length (G). n = 3 per group. (H) Tube formation images of HMECs treated with vehicle, vehicle + L. animalis-EVs, vehicle + L. reuteri-EVs, MPS, MPS + L. animalis-EVs, or MPS + L. reuteri-EVs. Scale bar, 200 μm. (I and J) Quantification of total loops (I) and total tube length (J). n = 3 per group. (K) Alizarin red S (ARS) staining images of BMSCs with different treatments under osteogenic induction. Scale bar, 200 μm. (L) Quantification of the percentage of ARS⁺ areas. n = 3 per group. (M to P) Cell counting kit-8 (CCK-8) analysis of HMECs (M), MLO-Y4 (N), MC3T3-E1 (O), and BMSCs (P) with different treatments. n = 3 per group. (Q and R) TUNEL staining images of different cell types receiving different treatments (Q) and quantification of the ratio of TUNEL⁺ apoptotic cells (R). Scale bar, 20 μm. n = 3 per group. *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 5.. L. animalis-EVs enter the femoral head and mitigate GC-induced ONFH.

(A) Distribution of the DiR-labeled L. animalis-EVs in the femurs, tibias, and femoral heads of mice detected by ex vivo fluorescent imaging after oral administration for 3, 24, and 72 hours. Scale bar, 6 mm. (B) Quantification of the fluorescence intensity. n = 3 per group. (C) Distribution of the DiO-labeled L. animalis-EVs within the mouse femoral head after oral administration for 3 hours. Scale bars, 50 μm (white) or 20 μm (red). (D) Schematic diagram of the experimental design for testing the effects of oral treatment with L. animalis-EVs on the femoral heads of vehicle- or MPS-treated mice. (E) Detection of L. animalis-EVs in the mouse femoral heads using the antibodies targeting L. animalis-EVs (L-EVs Ab). Scale bars, 50 μm (white) or 20 μm (red). (F) Quantification of the mean intensity for L-EV⁺ areas. n = 3 per group. (G to K) μCT-reconstructed images of femoral heads (G) and quantification of Tb. BV/TV (H), Tb. Th (I), Tb. N (J), and Tb. Sp (K). Scale bar, 1 mm. n = 8 per group. (L) H&E staining images of femoral heads. Scale bars, 200 μm (blue) or 50 μm (black). (M to P) Quantification of total vessel volume (M) and the numbers of CD31⁺ (N), OCN⁺ (O), and TUNEL⁺ (P) cells in femoral heads. n = 5 per group. *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 6.. Proteomic analysis of L. animalis-EVs and L. animalis.

(A) Summary of tandem mass spectrometry (MS/MS) database search results for L. animalis-EVs and L. animalis. (B) Volcano plot showing the numbers of differentially expressed proteins between L. animalis-EVs and L. animalis with the cutoff of P < 0.05 and |fold change| ≥ 1.5. (C) The expression ratios of the top 10 most abundant proteins in L. animalis-EVs (L-EVs) relative to L. animalis (L). n = 3 per group. (D) GO annotation of the up-regulated proteins in L. animalis-EVs relative to L. animalis in terms of their subcellular localizations. (E) GO biological process enrichment analysis of the up-regulated proteins in L. animalis-EVs relative to L. animalis. (F) KEGG pathway enrichment analysis of the differentially expressed proteins in L. animalis-EVs relative to L. animalis.