UNC93B1 Is Widely Expressed in the Murine CNS and Is Required for Neuroinflammation and Neuronal Injury Induced by MicroRNA let-7b

Abstract

The chaperone protein Unc-93 homolog B1 (UNC93B1) regulates internalization, trafficking, and stabilization of nucleic acid-sensing Toll-like receptors (TLR) in peripheral immune cells. We sought to determine UNC93B1 expression and its functional relevance in inflammatory and injurious processes in the central nervous system (CNS). We found that UNC93B1 is expressed in various CNS cells including microglia, astrocytes, oligodendrocytes, and neurons, as assessed by PCR, immunocyto-/histochemistry, and flow cytometry. UNC93B1 expression in the murine brain increased during development. Exposure to the microRNA let-7b, a recently discovered endogenous TLR7 activator, but also to TLR3 and TLR4 agonists, led to increased UNC93B1 expression in microglia and neurons. Microglial activation by extracellular let-7b required functional UNC93B1, as assessed by TNF ELISA. Neuronal injury induced by extracellular let-7b was dependent on UNC93B1, as UNC93B1-deficient neurons were unaffected by the microRNA's neurotoxicity in vitro. Intrathecal application of let-7b triggered neurodegeneration in wild-type mice, whereas mice deficient for UNC93B1 were protected against injurious effects on neurons and axons. In summary, our data demonstrate broad UNC93B1 expression in the murine brain and establish this chaperone as a modulator of neuroinflammation and neuronal injury triggered by extracellular microRNA and subsequent induction of TLR signaling.

Keywords: UNC93B1; let-7b; microRNA; microglia; neurodegeneration; neuroinflammation; neurons; toll-like receptor.

Figure 1

UNC93B1 is expressed in various CNS cell types. (A) Total RNA was extracted from purified cortical neurons, astrocytes, and microglia, as well as from whole brain, spleen, liver, and lung, derived from C57BL/6 (wild-type, WT) and Unc93b1^-/- mice. In addition, RNA from N1E-115 cells, Oli-neu cells (left panel), as well as human-derived HMC3 and SH-SY5Y cells (right panel) was isolated. Samples were analyzed by RT-PCR, using primers for Unc93b1. β-actin served as loading control. NTC, non-template control; RT-, condition without reverse transcriptase. (B) Purified WT cortical neurons, microglia, and astrocytes were immunolabeled with an antibody directed against the C-terminus of UNC93B1 (anti-UNC93B1-C) and co-stained with IB4 to mark microglia, with GFAP antibody to mark astrocytes, and with MAP-2, NeuN, and neurofilament (NF) antibodies to label neurons. Oli-neu cells were immunolabeled with the UNC93B1-C antibody and the oligodendrocyte marker APC. Scale bar, 50 µm. (C, D) Brain sections from WT and Unc93b1^-/- mice were immunolabeled using antibodies directed against UNC93B1-C, NeuN, GFAP, and CD11b, the latter serving as markers for neurons, astrocytes, and microglia/brain macrophages, respectively. An overview (C) and higher magnification images (D) of the cerebral cortex displaying the respective CNS cell populations are shown. Arrows indicate cells co-labeled with UNC93B1 antibody and the respective cell type marker. Scale bar, 50 µm. (E) WT brain tissue was dissociated, and cells were stained with both anti-UNC93B1-N antibody, recognizing the N-terminal region of UNC93B1, and different surface markers, as indicated. Microglia were defined as CD11b⁺GLAST^-, astrocytes as CD11b^-GLAST⁺, and neurons as CD11b^-CD45^-βIII-tubulin⁺ cells. Subsequently, cells were analyzed by flow cytometry. One representative experiment out of 3 with similar results is shown.

Figure 2

Expression of UNC93B1 in the mouse brain increases during development. Brain homogenates from C57BL/6 mice at various embryonic (E), postnatal (days after birth, P), and adult (age of months, Pm) stages, as indicated, were assayed by endpoint RT-PCR (A) and quantitative real-time PCR (B) using primers for Unc93b1 and IRAK1. Relative quantification was assessed using the formula 2^–ΔCT and by normalizing the amount of the target gene to the housekeeping genes β-actin (A) and GAPDH (B). The mean value of E13 was set to 1-fold induction, and mean values of all other developmental stages were related to E13 (n = 3). Results are presented as mean ± SD. (C) Immunolabeling of the C57BL/6 cerebral cortex at different developmental stages, as indicated, with the UNC93B1-C antibody. Scale bar, 200 µm.

Figure 3

UNC93B1 expression in microglia and neurons is regulated by TLR activation. (A) Microglia and neurons isolated from C57BL/6 mice, were exposed to poly(I:C) (150 µg/ml), LPS (1 µg/ml), loxoribine (1 mM), CpG-ODN 1668 (50 µM), or let-7b oligoribonucleotide (20 µg/ml) for 6 h Subsequently, cells were dissociated, labeled with both anti-UNC93B1-N and the respective antibodies for cell-specific markers, and were analyzed by flow cytometry. (B) Results were expressed as fold change of UNC93B1 PerCP median fluorescence intensity of CD11b⁺ (microglia) and betaIII-tubulin⁺ (neurons) cells (relative MFI; TLR agonist treatment vs. control, ctrl; n = 3) and are presented as mean ± SEM. (C) qPCR analysis of the cell cultures treated as described above using primers against UNC93B1 was performed. β-actin served as loading control. Relative Unc93b1 mRNA expression levels (TLR agonist treatment vs. control, ctrl) are shown (n = 3). Results are presented as mean ± SEM. P values for relevant groups were determined by Student’s t test. p* < 0.05; **p < 0.01; ***p < 0.001; ns, not significant.

Figure 4

Neuroinflammatory response and neuronal injury in response to extracellular let-7b are dependent on UNC93B1 in vitro. Microglia derived from C57BL/6 (wild-type, WT) and Unc93b1^-/- mice were incubated with various doses of let-7b oligoribonucleotide, as indicated, for 12 h (A), or with 10 µg/ml let-7b for various durations, as indicated (B). Mutant oligoribonucleotide (20 µg/ml, 12 h) and the transfection agent LyoVec served as negative control. LPS (100 ng/ml, 12 h) was used as positive control for microglial TNF-α release through TLR4 activation, while loxoribine (1 mM, 12 h) served as control for TLR7-dependent TNF-α release. Subsequently, supernatants were analyzed by TNF-α ELISA. Results are presented as mean ± SD. n = 4; n.d., not detectable. (C–E) WT neurons co-cultured with either WT or Unc93b1^-/- microglia were treated with various let-7b doses (C, D: 10 µg/ml for 72 h; E: as indicated for 5 d). Mutant oligoribonucleotide (5 µg/ml) served as negative control, LPS (100 ng/ml) served as positive control for microglial activation, and loxoribine (10 mM) served as positive control for TLR7-dependent microglial activation. Subsequently, co-cultures were immunolabeled with Iba1 (C) and NeuN (D) antibodies to mark microglia and neurons, respectively. Scale bar, 50 µm. (E) Quantification of NeuN-positive cells in co-cultures treated as described above. Untreated control was set to 100%. For each condition, experiments were performed in duplicates. At least 3 independent experiments were performed. Data are expressed as mean ± SEM. Kruskal-Wallis test was used to determine global significance over all conditions (p = 0.0406 over WT and p = 0.0897 over Unc93b1^-/- ). P values of relevant groups (indicated condition vs. unstimulated control, ctrl) were determined by Dunn’s multiple comparison test. P values between the respective groups of WT vs. Unc93b1^-/- (in brackets) were assessed by unpaired Student’s t test. *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 5

UNC93B1 is required for let-7b-induced cell-autonomous neuronal apoptosis in vitro. (A, B) Enriched cortical neurons isolated from C57BL/6 (wild-type, WT) and Unc93b1^-/- mice were exposed to 10 µg/ml let-7b for 72 h or were left untreated (control). Mutant oligoribonucleotide (5 µg/ml) served as negative control. LPS (100 ng/ml) was used to indicate functionally relevant microglia cell numbers in the enriched neuronal cell cultures. Loxoribine (10 mM) served as positive control for TLR7-dependent neuronal injury. Subsequently, cells were immunolabeled using Neurofilament (A) and NeuN (B) antibodies to assess axonal damage and neuronal viability, respectively. (C, D) Cell cultures derived from WT and Unc93b1^-/- mice described above were incubated with various doses of let-7b, as indicated, for 72 h (C) or treated with 5 µg/ml let-7b for different time periods, as indicated (D). (E–H) WT cortical neurons were incubated with let-7b (5 µg/ml) for various time periods, as indicated, and were subsequently analyzed by (E) TUNEL assay and (G) immunostaining using active caspase-3 antibody. DAPI staining marked all present cells. Subsequently, TUNEL-positive and caspase-3-positive cells were quantified and normalized against DAPI-positive cells (F and H, respectively). Mutant oligoribonucleotide served as negative control, while loxoribine and LPS were used as positive control. (C–H) At least three independent experiments were performed. Data are expressed as mean ± SEM. Kruskal-Wallis test was used to determine global significance over all conditions [(C) p = 0.0042 over WT and p = 0.9557 over Unc93b1^-/- ; (D) p = 0.0036 over WT and p = 0.9971 over Unc93b1^-/- ; (F) p < 0.0001 over WT and p = 0.3514 over Unc93b1^-/- ; (H) p = 0.0010 over WT and p = 0.0036 over Unc93b1^-/- ]. P values of relevant groups (indicated condition vs. unstimulated control, ctrl) were determined by Dunn’s multiple comparison test. P values between respective groups of WT vs. Unc93b1^-/- (in brackets) were assessed by unpaired Student’s t test. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Scale bar, 50 µm.

Figure 6

Neurodegeneration triggered by intrathecal let-7b and loxoribine requires UNC93B1 expression. We intrathecally injected 10 μg of let-7b, 10 μg of mutant oligoribonucleotide, 136 µg loxoribine, or PBS (control) into C57BL/6 (WT; let-7b, n = 7; mut. oligo, n = 6; loxoribine, n = 6; PBS, n = 4) or Unc93b1^-/- (let-7b, n = 5; mut. oligo, n = 6; loxoribine n = 6; PBS, n = 4) mice. After 3 d, brain sections were immunolabeled with neurofilament (A) and NeuN (B) antibodies, and representative images of the pericallosal area and cerebral cortex, respectively, are shown [scale bar, 200 µm in (A); 50 µm in (B)]. (C) Density of neurons in the cerebral cortex of WT and Unc93b1^-/- mice was assessed quantifying NeuN-positive cells. Brain sections described above were stained with TUNEL assay (D), and TUNEL-positive cells in the cerebral cortex were quantified (E). (C, E) One-way ANOVA was used to determine statistical significances across all tested conditions (C) p < 0.0001 over WT and p = 0.1232 over Unc93b1^-/- ; (E) p < 0.0001 over WT and p = 0.571 over Unc93b1^-/- ]. Unpaired Student’s t-test was used to assess significant differences between pairs (indicated group vs. control, ctrl; WT vs. Unc93b1^-/- ; **p < 0.01; ***p < 0.001). Scale bar, 100 µm (overview D) and 50 µm (inlay in D).