Myeloid-specific blockade of Notch signaling ameliorates nonalcoholic fatty liver disease in mice

Abstract

Rationale: Macrophages play a central role in the development and progression of nonalcoholic fatty liver disease (NAFLD). Studies have shown that Notch signaling mediated by transcription factor recombination signal binding protein for immunoglobulin kappa J region (RBP-J), is implicated in macrophage activation and plasticity. Naturally, we asked whether Notch signaling in macrophages plays a role in NAFLD, whether regulating Notch signaling in macrophages could serve as a therapeutic strategy to treat NAFLD. Methods: Immunofluorescence staining was used to detect the changes of macrophage Notch signaling in the livers of human patients with NAFLD and choline deficient amino acid-defined (CDAA) diet-fed mice. Lyz2-Cre RBP-J^flox or wild-type C57BL/6 male mice were fed with CDAA or high fat diet (HFD) to induce experimental steatohepatitis or steatosis, respectively. Liver histology examinations were performed using hematoxylin-eosin (H&E), Oil Red O staining, Sirius red staining and immunohistochemistry staining for F4/80, Col1α1 and αSMA. The expression of inflammatory factors, fibrosis or lipid metabolism associated genes were evaluated by quantitative reverse transcription (qRT)-PCR, Western blot or enzyme-linked immunosorbent assay (ELISA). The mRNA expression of liver samples was profiled by using RNA-seq. A hairpin-type decoy oligodeoxynucleotides (ODNs) for transcription factor RBP-J was loaded into bEnd.3-derived exosomes by electroporating. Mice with experimental NAFLD were treated with exosomes loading RBP-J decoy ODNs via tail vein injection. In vivo distribution of exosomes was analyzed by fluorescence labeling and imaging. Results: The results showed that Notch signaling was activated in hepatic macrophages in human with NAFLD or in CDAA-fed mice. Myeloid-specific RBP-J deficiency decreased the expression of inflammatory factors interleukin-1 beta (IL1β) and tumor necrosis factor alpha (TNFα), attenuated experimental steatohepatitis in mice. Furthermore, we found that Notch blockade attenuated lipid accumulation in hepatocytes by inhibiting the expression of IL1β and TNFα in macrophages in vitro. Meanwhile, we observed that tail vein-injected exosomes were mainly taken up by hepatic macrophages in mice with steatohepatitis. RBP-J decoy ODNs delivered by exosomes could efficiently inhibit Notch signaling in hepatic macrophages in vivo and ameliorate steatohepatitis or steatosis in CDAA or HFD mice, respectively. Conclusions: Combined, macrophage RBP-J promotes the progression of NAFLD at least partially through regulating the expression of pro-inflammatory cytokines IL1β and TNFα. Infusion of exosomes loaded with RBP-J decoy ODNs might be a promising therapy to treat NAFLD.

Keywords: Notch signaling; RBP-J; exosomes; macrophage; nonalcoholic fatty liver disease; transcription factor decoy.

Figure 1

Notch signaling in hepatic macrophages was activated in human with NASH or in CDAA-fed mice. (A) H&E staining of liver tissues from normal or patients with NASH. (B) Immunofluorescence staining for Hes1 and macrophage marker CD68 in liver tissues from normal or patients with NASH, nuclei were counterstained with DAPI, and Hes1⁺ CD68⁺ were quantified and compared. (C) H&E staining of liver tissues in chow-fed or CDAA-fed mice. (D) Immunofluorescence staining for Hes1 and F4/80 in liver tissues from chow-fed or CDAA-fed mice, and Hes1⁺ F4/80⁺ were quantified. (E) The mRNA levels of Notch ligands, Notch receptors, and downstream genes in liver extracts from chow-fed or CDAA-fed mice were determined by qRT-PCR. Bars = means ± SD; * P < 0.05, ** P < 0.01, *** P < 0.001.

Figure 2

Myeloid-specific RBP-J deficiency attenuated experimental steatohepatitis in CDAA-fed mice. Lyz2-Cre⁺ RBP-J^flox/+ (Control) or Lyz2-Cre⁺ RBP-J^flox/flox (KO) mice were fed with CDAA diet for 10 weeks. (A) Liver sections were stained by H&E staining. The lower row of micrographs were a higher magnification of the yellow frames in the upper row. (B) Quantitative comparison of steatosis areas, and histological scores about steatosis, inflammation and NAS in (A). (C) Liver-to-body weight ratio. (D) Detection of ALT, AST and LDL in serum. (E) Liver sections were stained with Oil Red O. (F) Quantitative comparison of positive signals for Oil Red O in (E), and detection of triglyceride content in livers and serum. (G) Liver sections were subjected to immunohistochemical staining for F4/80. The lower row of micrographs were a higher magnification of the red frames in the upper row. (H) The positive areas of F4/80 staining in (G) were quantitatively compared. (I) The mRNA levels of IL1β, TNFα and iNOS in livers were determined by qRT-PCR. (J) The protein levels of IL1β and TNFα were determined by Western blot, with Tubulin as a reference control. Bars = means ± SD; * P < 0.05, ** P < 0.01, *** P < 0.001.

Figure 3

Notch blockade in macrophages attenuated lipid accumulation in hepatocytes through inhibiting the expression of IL1β and TNFα in vitro. (A) Schematic showing that RAW264.7 macrophages were treated with γ-secretase inhibitor DAPT (5 μM) for 3 days, changed and cultured with the fresh medium containing LPS (100 ng/ml) for 1 day, and the conditioned medium (CM) were harvested. Then AML12 hepatocytes were cultured with CM, in the presence or absence of TNFα (10 ng/ml), and/or IL1β (10 ng/ml) in palmitic acid medium (PA, 10 mM) for 2 days. (B) The mRNA levels of Hes1, Hey1, IL1β and TNFα in RAW264.7 macrophages treated with DMSO or DAPT were determined by qRT-PCR. (C) The levels of IL1β and TNFα in RAW264.7 culture supernatants were measured by ELISA. (D) Lipid accumulation in AML12 cells was assessed with Oil Red O staining. (E) The positive areas of Oil Red O staining in (D) were quantitatively compared. (F) Triglyceride content in AML12 hepatocytes was measured. Bars = means ± SD; * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001.

Figure 4

Exosomes loaded with RBP-J decoy oligodeoxynucleotide (ODNs) inhibited Notch signaling in hepatic macrophages in a murine model of steatohepatitis. (A-C) Characterization of exosomes derived from bEnd.3 cells. (A) The levels of CD9, Alix, flotillin-1, and VDAC1 in exosomes and bEnd.3 cells lysates were determined by western blot. (B) The size distribution of exosomes was determined by nanoparticle tracking analysis (NTA). (C) Exosome morphology was observed by transmission electron microscopy (TEM). (D) Exosomes were stained with DiI and approximately 200 µg (protein equivalent) of exosomes in 150 µL of PBS were injected via the tail vein into mice with CDAA-induced steatohepatitis. After 6 h, DiI signals in the liver, lung, spleen, intestine, kidney, and heart were examined using bioluminescence imaging. (E) 6 h after injection of DiI-labeled exosomes, liver sections were stained with an anti-mouse F4/80 antibody and analyzed by fluorescence microscopy. Nuclei were counterstained with DAPI. (F) Mice were fed with CDAA diet for 10 weeks, and exosomes loaded with RBP-J decoy or control decoy ODNs (exosomes/decoy ODNs = 200 μg/2.5 nmol) were injected into mice four times via tail vein at the last 2 weeks. F4/80⁺ hepatic macrophages were isolated using magnetic-activated cell sorting (MACS). The mRNA levels of Hes1, Hey1, TNFα, and IL1β in F4/80⁺ hepatic macrophages were measured by qRT-PCR. Bars = means ± SD; * P < 0.05.

Figure 5

Exosomes loaded with RBP-J decoy ODNs attenuated experimental steatohepatitis in CDAA-fed mice. (A) Schematic illustration of the procedure used for exosomes loaded with RBP-J decoy ODNs or control decoy ODNs to treat steatohepatitis in mice fed with CDAA diet. (B) Liver sections were stained by H&E staining. The lower row of micrographs were a higher magnification of the yellow frames in the upper row. (C) Quantitative comparison of steatosis areas, and histological scores about steatosis, inflammation and NAS in (B). (D) Detection of ALT, AST and LDL in serum. (E) Liver sections were stained with Oil Red O, and areas positive were quantified and compared. (F) Liver sections were subjected to immunohistochemical staining for F4/80. The lower row of micrographs were a higher magnification of the red frames in the upper row. (G) The positive areas of F4/80 staining in (F) were quantitatively compared. (H) The mRNA levels of IL1β, TNFα and iNOS in livers were determined by qRT-PCR. (I) The protein levels of IL1β and TNFα were determined by Western blot, with Tubulin as a reference control. Bars = means ± SD, * P < 0.05, ** P < 0.01.

Figure 6

Exosomes loaded with RBP-J decoy ODNs attenuated fatty liver in HFD-fed mice. (A) Schematic illustration of the procedure used for exosomes loaded with RBP-J decoy ODNs or control decoy ODNs to treat steatosis in HFD diet-fed mice. Mice were fed with HFD diet for 22 weeks, and exosomes loaded with RBP-J decoy or control decoy ODNs (exosomes/decoy ODNs = 200 μg/2.5 nmol) were injected into mice four times via tail vein at the last 2 weeks. (B) Liver sections were stained by H&E staining. (C) Quantitative comparison of steatosis areas, and histological scores about steatosis, inflammation and NAS in (B). (D) Detection of triglyceride content and LDL in serum. (E) Liver sections were stained with Oil Red O, and areas positive were quantified and compared in (F). Bars = means ± SD; * P < 0.05.

Figure 7

Schematic summary of this study. During the progression of NAFLD, injured hepatocytes caused by over-accumulated lipid toxicity, released damage-associated molecular patterns (DAMPs), and then stimulated hepatic macrophages to release pro-inflammatory cytokines such as IL1β, TNFα. Meanwhile, Notch signaling, mediated by transcription factor RBP-J, was activated in hepatic macrophages and further promoted the expression of IL1β and TNFα. And then IL1β and TNFα in turn enhanced lipid accumulation in hepatocytes. Next, we used exosomes targeting delivery of RBP-J decoy ODNs and inhibited the activation of Notch signaling in macrophages. Finally, the infused exosomes-RBP-J decoy ODNs via tail vein resulted in the amelioration of NAFLD in mice.