Schwann cell-secreted frizzled-related protein 1 dictates neuroinflammation and peripheral nerve degeneration after neurotrauma

Abstract

Neurotrauma in limbs can induce sustained neuroinflammation, resulting in persistent disruption of nerve tissue architecture and retardation of axon regrowth. Despite macrophage-mediated inflammation promoting the removal of necrotic neural components and stimulating neo-vessel ingrowth, detrimental shifts in macrophage phenotype exacerbate nerve degeneration. Herein, we find that peripheral nerve injuries (PNIs) result in abundant secreted frizzled-related protein 1 (sFRP1) expression, particularly by Schwann cells (SCs). Heat shock protein 90 (HSP90) in macrophages recognizes sFRP1 and triggers a dysregulated secretion of inflammatory mediators. Single-cell atlas of human injured peripheral nerves reveals the appearance of sFRP1-expressing SCs with mesenchymal traits and macrophages with a proinflammatory genetic profile. Deletion of either SC-specific sFRP1 or macrophage-specific HSP90 alleviates neuroinflammation and prevents the progression of nerve degeneration. Together, our findings implicate the response of macrophages to SC-derived sFRP1 in exacerbating nerve damage following PNIs.

Keywords: axon regeneration; heat shock protein 90; neuroinflammation; peripheral nerve injury; secreted frizzled-related protein 1.

Graphical abstract

Figure 1

sFRP1 was abundantly produced in nerve ECM following injury and associated with nerve degeneration (A) The isolation of sciatic nerve samples and proteomic analysis process. (B) The clustering distribution of injured and uninjured nerve samples as plotted by PCA analysis. (C) Differentially expressed proteins between injured and uninjured nerve samples are displayed in volcano plot. N = 3 mice. Proteins regulated over 1.5-fold changes (adj. p < 0.05) are highlighted in blue (downregulated) and red (upregulated). (D) GO enrichment analysis indicating the classification of differentially expressed proteins related to the biological process category. (E) Differentially expressed proteins in the GO category of ECM are displayed as a heatmap. (F) Western blotting analysis demonstrating increased production of sFRP1 in the injured nerve tissue. (G) Quantification of sFRP1 protein level in sciatic nerves isolated from uninjured and injured mice as indicated by western blot analysis. N = 3 mice. (H) Representative TEM, HE, and TB images of injured nerves isolated from mice treated with WAY-316606 and PBS at 3 weeks post injury. N = 6 mice. (I and J) Quantification of myelinated axon diameter and g-ratio as indicated in TEM images. (K) Quantification of myelinated axon density as indicated in HE-stained images. (L) Representative IHC images of human nerves stained for sFRP1 at 12 h after injury. Statistical significance was determined using two-tailed unpaired Student’s t tests; ∗∗p < 0.01; ∗p < 0.05; ns, no difference. Data were presented as mean ± SD.

Figure 2

SCs sensed injury signals to release sFRP1 and elicited mesenchymal traits (A) t-distributed stochastic neighbor embedding (t-SNE) plot shows clustering of nerve cells based on gene expression. Single-cell sequencing datasets are analyzed from GSE120678. BC, B cell; TC, T cell; EC, endothelial cell; Macro, macrophage; SC, Schwann cell; Endo, endoneurial fibroblast; Epi, epineurial fibroblast; Peri, perineurial fibroblast. (B) sFRP1 expression is mainly distributed in SCs and fibroblasts in both uninjured and injured sciatic nerves. (C and D) Double IF staining of S100β (red)/sFRP1 (green) and Fibro (red)/sFRP1 (green) on both longitudinal (C) and transverse (D) sections of sciatic nerves. (E and F) Percentage of sFRP1-positive SCs and fibroblasts in uninjured and injured nerves. N = 6 mice. (G) Illustration of the in vivo LPS treatment design. (H) sFRP1 protein level in SCs isolated from PBS or LPS (15 mg/kg) intraperitoneally treated mice. (I) Quantification of sFRP1 protein level in sciatic nerves isolated from LPS-treated and PBS-treated mice as indicated by western blot analysis. N = 3 biological replicates. (J) Illustration of the in vitro LPS treatment design. (K) Western blot analysis of sFRP1 protein level in SCs treated with different concentrations of LPS. (L and M) Quantification of fluorescence intensity of sFRP1 and PDGFRα staining in SCs. (N) sFRP1 (green) and PDGFRα (red) double staining on LPS-treated and PBS-treated SCs. N = 3 biological replicates. Two fields were quantified as technical replicates in each biological replicates. Statistical significance was determined using two-tailed unpaired Student’s t tests; ∗∗∗∗p < 0.0001; ∗∗p < 0.01; ∗p < 0.05. Data were presented as mean ± SD.

Figure 3

Mice with deletion of sFRP1 in SCs profoundly reduced macrophage infiltration and improved nerve regeneration (A) Sfrp1^flox/flox mice were bred with Plpcre^Ert1 mice to generate tamoxifen-inducible SC-specific sFRP1 knockout (Sfrp1^flox/floxPlpcre^Ert1) and littermate control (Sfrp1^flox/flox) mice. (B and C) Representative SCG10 immunostaining and related quantification of sciatic nerves at 14 days post transection. N = 6 mice. The dashed line indicates the transection site. Scale bar, 500 μm. (D and E) Representative F4/80 immunostaining (red) of sciatic nerves taken from the injury site, 1,000, 2,000, and 3,000 μm distal to the injury site and related quantification of infiltrated macrophages. Scale bar, 100 μm. N = 6 mice. (F and G) Western blot analysis and related quantification of TNF-α level in injured nerves at 24 h post transection. N = 3 mice. (H and I) Triple staining of CCL2 (green), F4/80 (red), and NeuN (pink) on sciatic DRG sections from Sfrp1^flox/flox and Sfrp1^flox/floxPlpcre^Ert1 mice and related quantification of CCL expression level in DRGs. N = 6 mice. No significant difference of CCL2 expression is observed between DRGs of Sfrp1^flox/flox and Sfrp1^flox/floxPlpcre^Ert1 mice. (J–L) Representative TUBB3 immunostaining (green) of sciatic DRG neurons isolated from Sfrp1^flox/flox and Sfrp1^flox/floxPlpcre^Ert1 mice (n = 6 mice) and related quantification. DRG neurons were cultured in vitro for 4 days or 7 days. (M and N) Representative immunostaining and related quantification of ATF3 (red) and the neuronal marker NeuN (green) in sciatic DRGs at 24 h after nerve injury. N = 6 mice. Scale bar, 50 μm. Statistical significance in (C) and (E) was analyzed by two-way ANOVA followed by Sidak’s post hoc analysis. Statistical significance was determined using two-tailed unpaired Student’s t tests; ∗∗∗∗p < 0.0001; ∗∗∗p < 0.001; ∗∗p < 0.01; ∗p < 0.05; ns, no significance. Data were presented as mean ± SD.

Figure 4

SFRP1 induces the F4/80⁺ CD86⁺ proinflammatory macrophage phenotype and inhibits oxidative metabolism (A and B) The axon length of sciatic DRG neurons demonstrates no significant difference in response to sFRP1 treatment. N = 6 biological replicates. (C) Representative TEM images reveal that the morphology and structure of mitochondria were well preserved in sFRP1-treated neurons. (D and E) Representative TEM images and related quantification of nerve transections (N = 6 mice). The suppressing effect of sFRP1 on axon regrowth is alleviated in a macrophage-deficient condition. (F) Double staining of IL-1β (red) and TNF-α (green) on sFRP1-treated BMDMs. (G) sFRP1-induced phenotypic switch is revealed by flow cytometric quantification. FITC reflects F4/80-positive cells. PE reflects CD206-positive cells. APC reflects CD86-positive cells. (H and I) Quantification of the percentage of IL-1β and TNF-α-positive cells as reflected by Figure 4F. Biological replicates n = 3 with two technical replicates each. (J) Double staining of Arg-1 (red) and Wnt3a (green) on sFRP1 and PBS-treated BMDMs. (K) The internalizing capacity of BMDMs was measured by incubating with 100 μg/mL pHrodo BioParticles (green). BMDMs were visualized by F4/80 (red) staining. (L and M) Quantification of the percentage of proinflammatory and pro-resolving macrophages as reflected by Figure 4G. Biological replicates n = 3 with two technical replicates each. (N) Heatmap of differentially expressed genes between sFRP1-treated and PBS-treated macrophages (control) based on RNA sequencing. N = 3 biological replicates. (O) GO classification of differentially expressed genes related to the biological process, cellular component, and molecular function categories. (P) KEGG enrichment analysis of differentially expressed genes based on RNA sequencing. (Q) Schematic diagram of the detection of mitochondrial respiration and glycolysis of macrophages by measuring OCR and ECAR. (R) OCR of macrophages at baseline and after serial administration with oligomycin, FCCP, and rotenone plus Antimycin A. Macrophages were treated with 50, 100, 200, and 500 nM sFRP1. (S) ECAR was compared at baseline and after serial administration with glucose, oligomycin, and 2-DG. Statistical significance in (E), (H), and (I) was determined using one-way ANOVA followed by Tukey’s multiple comparisons tests; ∗∗∗∗p < 0.0001 versus PBS group; ∗∗p < 0.01 versus PBS group. Statistical significance in (B), (L), and (M) was determined using two-tailed unpaired Student’s t tests; ∗∗∗p < 0.001 versus PBS group; ns, no significance. Data were presented as mean ± SD.

Figure 5

Identification of HSP90 as a binding protein to mediate the proinflammatory effect of sFRP1 on BMDMs (A) List of candidates with top 10 scores in LC-MS/MS analysis of BMDM-derived proteins with incubation of His-labeled sFRP1. (B) IP-MS analysis identifies HSP90 as an interacting protein that binds sFRP1. (C and D) IP analysis of Myc-sFRP1 (C) and HA-HSP90 (D) binding. (E) BMDMs were treated with sFRP1 plus HSP90-siRNA or control. HSP90 and sFRP1 interactions are confirmed in BMDM lysates by IP with anti-HSP90, followed by western blot analysis with anti-HSP90 and anti-sFRP1 antibody, respectively. (F) Representative IHC images of human nerves stained for HSP90 at 12 h after injury. (G) t-SNE plots of injured nerves marked by genes of HSP90 isoforms. Color key from orange to yellow indicated relative gene expression levels from high to low. (H–J) Representative IF staining and related quantification of TNF-α (red), F4/80 (red), and p65 (green) staining on BMDMs treated with sFRP1 plus HSP90-siRNA and controls. Biological replicates n = 3 with two technical replicates each. (K–M) BMDM phenotypic switch as revealed by flow cytometric quantification. FITC reflects F4/80-positive cells. PE reflects CD206-positive cells. APC reflects CD86-positive cells. N = 6 biological replicates. (N–P) TUBB3 staining on sciatic DRG neurons cocultured with macrophages for 4 days and 7 days and related quantification of axonal length. Scale bar, 100 μm. The start and the end of an axon were marked by red arrows. Biological replicates n = 3 with two technical replicates each. Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparisons tests; ∗∗∗∗p < 0.0001; ∗∗∗p < 0.001; ∗∗p < 0.01; ns, no significance. Data were presented as mean ± SD.

Figure 6

Depletion of HSP90 in macrophages attenuated neuroinflammation and nerve degenerative changes exerted by sFRP1 (A) Hsp90aa^flox/+ mice were bred with Lyz2-cre mice to generate macrophage-specific HSP90-deficient (Hsp90aa^flox/+Lyz2-cre) and littermate control (Hsp90aa^flox/+) mice. (B and C) Representative IF images of SCG10 staining and related quantification of sciatic nerves at 2 weeks post injury. The dashed line indicates the transection site. Scale bar, 500 μm. N = 6 mice. (D and E) Representative IF images of F4/80 staining (red) of sciatic nerves and related quantification of macrophages at 2 weeks post injury. Scale bar, 100 μm. N = 6 mice. (F–I) Double staining of TNF-α (red) and IL-1β (green) on nerve longitudinal sections and related quantification. (J–L) Representative TUBB3 staining (green) and related quantification of sciatic DRG neurons isolated from Hsp90aa^flox/+ and Hsp90aa^flox/+Lyz2-cre mice after 4 days and 7 days of culture. Biological replicates n = 3 with two technical replicates each. Statistical significance was determined using two-way ANOVA followed by Sidak’s post hoc analysis in (C) and (E), and using two-tailed unpaired Student’s t tests in (F), (G), (K), and (L); ∗∗p < 0.01; ∗∗∗p < 0.001; ∗p < 0.05; ns, no significance. Data were presented as mean ± SD.

Figure 7

SFRP1-neutralizing antibody treatment improved axon regeneration in vivo and in vitro (A and B) Representative SCG10 immunostaining and related quantification of murine injured nerves at 2 weeks after nerve transection. The dashed line indicates the transection site. Scale bar, 500 μm. N = 6 mice. (C) Schematic diagram of DRG neuron and macrophage microfluidic coculture chamber assay. (D) Representative optical images of macrophages in the neuron-macrophage coculture chambers. (E and F) Representative TUBB3 immunofluorescent images of neurons in the neuron-macrophage co-culture chambers and related quantification of average axon length in microfluidic channels. Biological replicates n = 3 with two technical replicates each. (G) Schematic diagram of DRG neuron and macrophage direct coculture assay. (H and I) Representative IF images stained for TUBB3 (green) on sciatic DRG neurons, and quantification of average axon length per cell in the direct coculture dishes. Biological replicates n = 3 with two technical replicates each. Statistical significance was determined using two-way ANOVA followed by Sidak’s post hoc analysis in (B) and (I) and using two-tailed unpaired Student’s t tests in (F); ∗∗∗p < 0.001; ∗∗p < 0.01; ∗p < 0.05. Data were presented as mean ± SD.