Endocrine modulation of brain-skeleton axis driven by neural stem cell-derived perilipin 5 in the lipid metabolism homeostasis for bone regeneration

Abstract

Factors released from the nervous system always play crucial roles in modulating bone metabolism and regeneration. How the brain-driven endocrine axes maintain bone homeostasis, especially under metabolic disorders, remains obscure. Here, we found that neural stem cells (NSCs) residing in the subventricular zone participated in lipid metabolism homeostasis of regenerative bone through exosomal perilipin 5 (PLIN5). Fluorescence-labeled exosomes tracing and histological detection identified that NSC-derived exosomes (NSC-Exo) could travel from the lateral ventricle into bone injury sites. Homocysteine (Hcy) led to osteogenic and angiogenic impairment, whereas the NSC-Exo were confirmed to restore it. Mecobalamin, a clinically used neurotrophic drug, further enhanced the protective effects of NSC-Exo through increased PLIN5 expression. Mechanistically, NSC-derived PLIN5 reversed excessive Hcy-induced lipid metabolic imbalance and aberrant lipid droplet accumulation through lipophagy-dependent intracellular lipolysis. Intracerebroventricular administration of mecobalamin and/or AAV-shPlin5 confirmed the effects of PLIN5-driven endocrine modulations on new bone formation and vascular reconstruction in hyperhomocysteinemic and high-fat diet models. This study uncovered a novel brain-skeleton axis that NSCs in the mammalian brain modulated bone regeneration through PLIN5-driven lipid metabolism modulation, providing evidence for lipid- or bone-targeted medicine development.

Keywords: bone regeneration; homocysteine; lipid metabolism; mecobalamin; neural stem cell; perilipin 5.

Graphical abstract

Figure 1

Nestin⁺ NSC-derived exosomes traveled into bone injury sites (A) Illustration for surgery and intervention methods. (B) Representative images of fluorescence distribution at 10 min and 48 h after intracerebroventricular injections of PBS or Dil-Exos. n = 3 per group. (C) Representative fluorescent images of removed brain and tibia at 48 h after intracerebroventricular injections of PBS, Dil, or Dil-Exos. n = 3 per group. (D) Identification of Dil-Exos located at DG and SVZ. Scale bar, 100 μm. DG, dentate gyrus; SVZ, subventricular zone. (E) Representative images of the distributions of Dil-Exos from brains within bone fracture sites after PBS or Dil-Exos administration at 24 and 48 h. Scale bar, 100 μm. (F) Representative images of the distributions of Dil-Exos from brains within DRG (L6 and S1) after PBS or Dil-Exos administration at 24 and 48 h. Scale bar, 100 μm. (G and H) Representative images of the distributions of Dil-Exos within bone and quantitative analysis after PBS, Dil-Exos, or Dil-Meco-Exos administration at 72 h. Scale bar, 50 μm. n = 4 per group. (I) Representative images of DiO-Exos internalized by BMSCs. Scale bars, 20 μm. (J) Representative images of DiO-Exos internalized by HUVECs. Scale bars, 10 μm. (K) Quantitative analysis for DiO-Exos internalization by BMSCs and HUVECs. n = 4 per group. (L) Representative western blot images of FN1 with or without treatment of mecobalamin. n = 3 per group. (M) Summary of the transport process of NSC-derived exosomes homing to the regenerative bone zone. Data are presented as mean ± SD. Unpaired two-tailed Student’s t test. ∗p < 0.05, ∗∗∗p < 0.001.

Figure 2

NSC-derived exosomal cargos reverse Hcy-induced retarded cell proliferation and exacerbated apoptosis in vitro (A and B) The proliferation curve of BMSCs and HUVECs at days 0, 1, 3, and 5 after different treatments. n = 5 per group. (C) Illustration for image arrangement and color representations of following panels in this figure. (D–G) The results of EdU assay and quantitative analyses of BMSCs and HUVECs after 4- or 3-h cultures with EdU, respectively. Scale bars, 50 μm. n = 4 per group. (H–K) Representative images and quantitative analyses of flow cytometry of BMSCs and HUVECs detected by Annexin V-FITC/PI. n = 4 per group. (L) Representative western blot images of apoptosis-related proteins in BMSCs and HUVECs after different treatments. n = 3 per group. (M) Schematic diagram of NSC-Exo protecting cells from apoptosis caused by Hcy. Data are presented as mean ± SD. One-way ANOVA with Tukey’s tests. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 3

NSC-derived exosomal cargos improve Hcy-related impaired osteogenesis and angiogenesis in vitro (A) Illustration for image arrangement and color representations of the following panels in this figure. (B and C) Representative ALP staining images of BMSCs and quantitative analysis after 7 days of osteogenic induction. Scale bar, 100 μm. n = 4 per group. ALP, alkaline phosphatase. (D and E) Representative alizarin red S staining images of BMSCs and quantitative analysis after 21 days of osteogenic induction. Scale bar, 100 μm. n = 4 per group. (F) Immunofluorescence staining for transcriptional factor Runx2 in BMSCs after different treatments. Scale bar, 50 μm. (G) Representative western blot images of osteogenic proteins OCN and COL I extracted from BMSCs after different treatments. n = 3 per group. (H and I) Representative images from Transwell assay of HUVECs and quantitative analysis in various groups. Scale bar, 100 μm. n = 4 per group. (J and K) Representative images from tube formation assay of HUVECs and quantitative analysis after different treatments. Scale bar, 100 μm. n = 4 per group. (L) Immunofluorescence staining against VEGFA in HUVECs after different treatments. Scale bar, 50 μm. (M) Representative western blot images of angiogenic markers CD31 and MMP9 extracted from HUVECs after different treatments. n = 3 per group. (N) The relative expression levels of angiogenic genes Pdgfb, Ang2, and Sdf1 were detected by qRT-PCR. n = 3 per group. (O) Schematic diagram of NSC-Exo improving osteogenesis and angiogenesis impaired by Hcy. Data are presented as mean ± SD. One-way ANOVA with Tukey’s tests. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 4

Upregulated PLIN5 after mecobalamin stimulation potentially contributes to enhanced capacities of NSC-Exo (A) Schematic illustration of workflow for screening differentially expressed genes. (B) Heatmap of upregulated and downregulated genes, including Plin5 in the presence or absence of mecobalamin. n = 3 per group. (C) The dominant groups of differentially expressed genes by gene set enrichment analysis. (D) The mRNA expression levels of Plin5 in NSCs and NSC-Exos were detected by qRT-PCR. n = 3 per group. (E–G) Representative images of PLIN5 expression of NSCs or NSC-Exo and quantitative analyses in various concentrations of mecobalamin detected by western blot. n = 3 per group. (H and I) Immunofluorescence staining and quantitative analysis for Nestin/PLIN5 of brain sections with or without mecobalamin stimulation. Scale bar, 100 μm. n = 5 per group. DG, dentate gyrus; SVZ, subventricular zone. (J and K) Representative western blot images of PLIN5 contents and quantitative analysis under various culture conditions in BMSCs and HUVECs. n = 3 per group. Data are presented as mean ± SD. One-way ANOVA with Tukey’s tests for (F, G, and K); unpaired two-tailed Student’s t test for (I). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 5

PLIN5 restores intracellular lipid metabolism homeostasis through lipophagy (A) Illustration for image arrangement and color representations of the following panels in this figure. (B and C) Representative alizarin red S staining images of BMSCs and quantitative analysis after 14 days of osteogenic induction. Scale bar, 100 μm. n = 4 per group. (D and E) Representative oil red O staining images of BMSCs and quantitative analysis after 14 days of adipogenic induction. Scale bar, 50 μm. n = 4 per group. (F and G) Representative images from tube formation assay of HUVECs and quantitative analysis among various groups. Scale bar, 100 μm. n = 4 per group. (H and I) Intracellular neutral lipid droplets of HUVECs stained by bodipy493/503 and quantitative analysis. Scale bar, 10 μm. n = 4 per group. (J) The expression levels of lipid anabolic enzyme Seipin and lipid catabolic enzymes HSL and ATGL were examined by western blot. n = 3 per group. (K) Representative TEM images of BMSCs and HUVECs, where lipid droplets (∗) and autophagosomes/autolysosomes (#) were indicated. Scale bars, 200 nm. (L) BMSCs were transfected with mCherry-GFP-LC3 plasmids and then submitted to further experiments, in which yellow dots indicate autophagosomes and red dots represent autolysosomes. Scale bar, 10 μm. (M) Quantitative analyses of autophagy flux by calculating the total amount of autophagosomes/autolysosomes and the ratio of autolysosomes to autophagosomes. n = 4 per group. (N and O) HUVECs were transfected with mCherry-GFP-LC3 plasmids and then submitted to further experiments and quantitative analyses. Scale bar, 5 μm. n = 4 per group. (P) The protein levels of autophagy-related markers, including phospho-ATG16L1, total ATG16L1, LC3, ATG5, and Beclin1. n = 3 per group. (Q) Schematic diagram of PLIN5-induced lipophagy and lipid metabolic homeostasis. Data are presented as mean ± SD. One-way ANOVA with Tukey’s tests. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 6

Endocrine factors from the brain partially alleviate impaired callus growth of hyperhomocysteinemic rats through PLIN5 (A) Workflow of animal experiments and intervention methods. (B) Illustration for grouping of the following panels in this figure. (C) Representative three-dimensional images of the regenerative bone in the distraction zone detected by μCT on day 43. Scale bar, 1 mm. (D) Quantitative analysis for BV/TV of regenerative tissues within distraction areas. n = 5 per group. (E–G) Mechanical properties of suffered bone, including ultimate load, stiffness, and energy to failure evaluated on day 43. n = 5 per group. (H) Histological images of undecalcified bone sections stained with von Kossa, Goldner, Masson, and Safranin O among various groups. Scale bar, 1 mm. (I) Dynamic histomorphometric assessments by subcutaneous injection of calcein (day 15) and xylenol orange (day 40). Scale bar, 50 μm. (J and K) Quantitative analyses by calculating mineral apposition rate and ratio of mineralizing surface to the bone surface. n = 4 per group. (L and M) Histological structures of regenerative bone as shown by H&E staining and the BV/TV of regenerative tissues in various groups. Scale bar, 500 μm. n = 5 per group. (N and O) Immunohistochemistry staining for OCN within newly formed callus followed by quantitative analysis. Scale bar, 100 μm. n = 5 per group. Data are presented as mean ± SD. One-way ANOVA with Tukey’s tests. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 7

Exosomal PLIN5 from the brain facilitated newly vascular reconstruction under hyperhomocysteinemia through lipid metabolism modulation (A) Workflow of animal experiments and intervention methods. (B) New vessels within regenerative areas evaluated by microfil perfusion. Scale bar, 1 mm. (C) Total vessel volume based on reconstructive vessel images was calculated. n = 3 per group. (D–G) Immunohistochemistry staining and quantitative analyses to detect the expression of CD31 and VEGFA within regenerative zones. Scale bar, 100 μm. n = 5 per group. (H) The distribution images of type H vessels labeled by antibodies against CD31 and Emcn. Scale bar, 50 μm. (I) The total volume of CD31^hiEmcn^hi vessels based on immunofluorescence staining. n = 5 per group. (J and K) Immunostaining for the expression of PLIN5 within DG followed by quantitative analysis. Scale bar, 100 μm. n = 5 per group. DG, dentate gyrus. (L and M) Immunohistochemistry staining for the expression of PLIN5 within regenerative zones followed by quantitative analysis. Scale bar, 100 μm. n = 5 per group. (N) Representative western blot images of PLIN5, ATGL, and ATG5 expression levels within whole distraction zones. n = 3 per group. (O) Schematic diagram of the connection of lipophagy and lipid metabolism homeostasis within the distraction zone. Data are presented as mean ± SD. One-way ANOVA with Tukey’s tests. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Figure 8

Schematic diagram of PLIN5 transported by brain-derived extracellular vesicles for remote lipid metabolism modulation and bone regeneration