Notch signaling in activated human hepatic stellate cells and macrophages

Since HSCs and macrophages are the key cells in the fibrous areas²², we investigated the activation of Notch pathway activation in these cells in vitro. In TGFβ-activated human HSCs, we found the expression of fibrosis-related parameters (collagen I and α-SMA), and Notch pathway related genes (Notch3, Jag1 and Hes1) were significantly upregulated as compared to non-activated control cells (Fig. 2A).

Open in a separate window

Figure 2

Expression of Notch signaling pathway activation in human HSCs and macrophages.

(A) Quantitative gene expression analysis of fibrotic marker (Col1A1), HSC activation marker (α-SMA) Notch receptors (Notch-1,-2 and -3), Notch ligands (Dll1, Dll4 and Jag1) and downstream Notch signaling molecule (Hes1) in TGFβ-activated HSCs versus control HSCs (expressed as dotted line). (B) Representative microscopic images of M1-differentiated (LPS/IFNγ treated) and M2-differentiated (IL-4/IL-13 treated) macrophages (scale bars, 200μm). Quantitative gene expression analysis of (C) M1-specific markers (IL-1β and IL-6) and M2-specific markers (Arginase I, Arg1 and Mannose receptor C type 1, MRC1); (D) Notch pathway related genes: Notch receptors (Notch-1, -2 and -3); Notch ligands (Dll1, Dll4 and Jag1) and downstream Notch signaling molecule (Hes1), in M1-differentiated and M2-differentiated RAW macrophages. Untreated/undifferentiated RAW cells are denoted as dotted line. Bars represent mean±SEM of atleast three independent experiments. *p<0.05 and **p<0.01 denotes significance versus respective control cells.

To examine the Notch signaling pathway in macrophages, we differentiated mouse macrophages to M1 or M2 macrophages using LPS and IFNγ (for M1) or IL-4 and IL-13 (for M2). As shown in Fig. 2B, the polarization to M1 and M2 macrophages exhibited morphological differences. After stimulation with LPS/IFNγ, M1 macrophages preferentially showed long and spindle-shaped morphology while IL4/IL13 (M2) differentiated macrophages were largely round-shaped (Fig. 2B). The M1 and M2 phenotypes were further confirmed by gene expression analysis. As compared to undifferentiated RAW cells, M1-differentiated macrophages showed significant upregulation of M1-specific inflammatory genes (IL-1β and IL-6) while M2-specific genes (Arginase 1 and Mannose receptor C type 1, MRC1) were substantially induced in M2-differentiated macrophages (Fig. 2C).

Thereafter, we investigated Notch signaling pathway related genes in RAW, M1- and M2-differentiated macrophages. We found that M1 macrophages expressed higher levels of Notch receptors (Notch1>Notch3>Notch2) and ligands (Dll1, Dll4, Jag1) and the downstream signaling molecule Hes1 as compared to un-differentiated or M2 macrophages. These data suggest that Notch pathway is strongly activated in M1-differentiated macrophages in comparison to un-differentiated and M2-differentiated macrophages (Fig. 2D).

Notch inhibitor Avagacestat inhibits differentiation and contractility of HSCs

To understand the role of Notch signaling pathway in fibrogenic cells, we inhibited Notch signaling pathway in human HSCs using a selective arylsulfonamide γ-secretase Notch inhibitor, Avagacestat. We observed that following TGFβ activation, there was a significant increase in protein and gene expression of fibrotic parameters (collagen-I, α-SMA and vimentin) (Fig. 3A,C). Inhibition of Notch signaling pathway with Avagacestat, which was confirmed by significant decrease in Hes1 expression (Fig. 3B), led to the dose-dependent reduction in collagen-I, α-SMA and vimentin protein expression and also a significant inhibition in the respective mRNA expression levels was observed (Fig. 3A,C). No significant effect on cell viability was found at different concentrations of Avagacestat, indicating no effect on cell death or proliferation (Fig. 3D).

Open in a separate window

Figure 3

Effect of Notch inhibition on the activation and contractility of TGFβ-activated human HSCs.

(A) Representative images (scale bars, 200μm) of collagen I and vimentin stained LX2 cells treated with medium alone (control), TGFβ (5ng/ml)±5μM or 10μM Notch inhibitor (Avagacestat). Quantitative gene expression analysis (normalized with GAPDH) of (B) Notch signaling molecule Hes1, (C) Fibrotic parameters: collagen I, α-SMA and Vimentin on LX2 cells treated with medium alone (control expressed as dotted line), TGFβ (5ng/ml)±10μM Avagacestat. (D) Graph showing % cell viability performed on LX2 cells incubated with increasing concentrations of Avagacestat. (E) Graph depicts % 3D-collagen-I gel contraction after 24, 48 and 72h of treatment with medium alone (control), TGFβ (5ng/ml)±10μM Avagacestat. (F) Representative microscopic images of the contracted collagen gels after 72h of treatments. (G) Representative pictures (scale bars, 200μm) of collagen I stained mouse 3T3 fibroblasts treated with medium alone (control), TGFβ (5ng/ml)±5μM or 10μM Avagacestat. Bars represent mean±SEM of atleast three independent experiments. ^#p<0.05 denotes significance versus respective control cells; *p<0.05 denotes significance versus TGFβ-treated cells.

Furthermore, we examined the effect of Avagacestat on the contractility of human HSCs in 3D collagen gel. We found that 10μM Avagacestat significantly inhibited TGFβ-induced contractility of HSCs after 48h and 72h as shown in Fig. 3E,F. In addition, Avagacestat showed inhibition of collagen-I expression in activated mouse 3T3 fibroblasts (Fig. 3G) indicating its inhibitory effect in both mouse and human fibroblasts.

Blocking of Notch signaling inhibits M1-polarization

We then investigated the role of Notch signaling in macrophage polarization to M1 and M2. We found that incubation with the Notch inhibitor, Avagacestat completely inhibited the M1-specific Notch signaling gene Hes1 (Fig. 4A). Furthermore, treatment with Avagacestat dose-dependently inhibited the M1-induced nitric oxide (NO) release (an indicator of M1 activation) (Fig. 4B). Of note, at these doses, there was no effect on the cell viability of macrophages (data not shown). We further found that Avagacestat led to the substantially reduced M1 polarization as depicted in M1-related genes i.e IL-1β, IL-6 and nitric oxide synthase 2 (NOS2) (Fig. 4C). No differences in M2 polarization were observed following treatment with the Notch inhibitor, as shown with mRNA expression analyses (Fig. 4D).

Open in a separate window

Figure 4

Effect of Notch inhibition on the polarization of M1- and M2-differentiated macrophages.

(A) Quantitative gene expression analysis of Hes1 in M1-differentiated and M2-differentiated RAW macrophages incubated with 10μM Avagacestat. Untreated/undifferentiated RAW macrophages are expressed as dotted line. (B) Graph depicts nitric oxide release (expressed in %) by M1-differentiated macrophages incubated with increasing concentrations of Avagacestat. (C) Quantitative gene expression analysis of M1 markers (IL-1β, IL-6 and NOS2) in M1-differentiated RAW macrophages incubated with 10μM Avagacestat. (D) Quantitative gene expression analysis of M2 markers (Arg1 and MRC1) in M2-differentiated RAW macrophages incubated with 10μM Avagacestat. Untreated/undifferentiated RAW macrophages are expressed as dotted line. Bars represent mean±SEM of atleast three independent experiments. ^#p<0.05 denotes significance versus respective untreated RAW cells. *p<0.05 denotes significance versus M1-differentiated RAW macrophages.

Effect of Avagacestat on the crosstalk between macrophages and fibroblasts

Since fibrosis is driven by the cross-talk between macrophages and fibroblasts (or activated HSCs), we designed experiments to determine the effect of Avagacestat on the interplay between macrophages and fibroblasts. We performed co-culture studies using mouse 3T3 fibroblasts and mouse RAW macrophages, and analyzed nitric oxide release. We found that co-culture of 3T3 and RAW macrophages alone (without TGFβ and M1 cytokines LPS and IFNγ) did not induce nitric oxide release (data not shown), but treatment with TGFβ (3T3 activation) and LPS/IFNγ (M1-differentiation) led to highly significant increase in NO release as compared to M1 macrophages mono-culture suggesting activated fibroblasts stimulates M1 activation. This 3T3-driven M1 activation was significantly down-regulated by Notch Inhibitor, Avagacestat (Fig. 5A). In co-culture studies, it might be possible that Avagacestat inhibited both M1 macrophages differentiation and 3T3 activation. Furthermore, 3T3-driven M1 activation may have resulted from direct contact and/or secreted fibroblast factors.

Open in a separate window

Figure 5

Effect of Avagacestat on the cross-talk between 3T3 fibroblasts and RAW macrophages.

(A) Graph depicting nitric oxide release (expressed in %) from RAW macrophages and 3T3 fibroblasts cultured either alone or together (co-culture). 3T3 (TGFβ-activated) and M1-differentiated macrophages were treated with 10μM Avagacestat either separately or in co-culture. (B) Graph depicts nitric oxide release (expressed in %) from undifferentiated RAW macrophages and M1-differentiated macrophages treated with medium alone (control), 3T3-conditioned medium (3T3CM) and Avagacestat (10μM) treated 3T3-conditioned medium (3T3CM+Avagacestat). (C) Quantitative gene expression analysis of iNOS (normalized with GAPDH) in RAW macrophages treated with medium alone (control), 3T3-conditioned medium (3T3CM) and Avagacestat (10μM) treated 3T3-conditioned medium (3T3CM+Avagacestat). (D) α-SMA gene expression analysis (normalized with GADPH) in 3T3 fibroblasts treated with medium alone (control), undifferentiated RAW macrophages conditioned medium (RAW CM), Avagacestat (10μM) treated RAW macrophages conditioned medium (RAW CM+Avagacestat), M1-differentiated macrophages conditioned medium (M1CM) and Avagacestat (10μM) treated M1-differentiated macrophages conditioned medium (M1CM+Avagacestat). Bars represent mean±SEM of atleast three independent experiments. ^#p<0.05, ^##p<0.01 denotes significance.

To determine whether macrophage can activate 3T3 fibroblasts or vice versa, we established conditioned medium studies. Undifferentiated RAW cells and M1-differentiated macrophages were treated with medium alone, conditioned medium from 3T3 fibroblasts or Avagacestat-treated 3T3 fibroblasts. We found that 3T3 conditioned medium alone stimulated M1-differentiation in undifferentiated RAW macrophages and further activated M1-differentiated macrophages as analyzed by NO release (Fig. 5B). This fibroblasts-driven macrophage activation was blunted when Avagacestat-treated fibroblast conditioned medium was added on macrophages (both undifferentiated and M1-differentiated) suggesting role of Notch pathway in 3T3-driven macrophage activation. These results were further confirmed by quantitative PCR where 3T3 conditioned medium stimulated iNOS gene expression, which was inhibited by Avagacestat-treated 3T3 conditioned medium (Fig. 5C).

On the other hand, we determined effect of RAW macrophages or M1-differentiated macrophages on 3T3 fibroblasts activation. Undifferentiated macrophage-conditioned medium alone induced 3T3 activation as analyzed by α-SMA gene expression (Fig. 5D). However, M1-differentiated macrophage-conditioned medium induced drastic (>2-fold higher) activation of fibroblasts suggesting role of macrophages and M1-macrophages in fibroblasts activation. This macrophage-driven fibroblasts activation was inhibited by Avagacestat-treated macrophage conditioned medium suggesting role of Notch pathway in 3T3 activation (Fig. 5D).

These co-culture and conditioned medium studies demonstrate cross-talk between macrophages and fibroblasts via direct contact (juxtacrine effect) and/or via secreted stimulating factors (non-contact/paracrine effect) and suggests an important regulatory role of Notch signaling in this cross-talk or interplay.

Cell Lines

Mouse NIH3T3 fibroblasts and murine RAW264.7 macrophages were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Human hepatic stellate cells (LX2) were kindly provided by Prof. Scott Friedman (Mount Sinai Hospital, New York). RAW macrophages and 3T3 were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium (Lonza, Verviers, Belgium) and Dulbecco’s modified Eagle’s (DMEM) medium (Lonza) respectively supplemented with 2 mM L-glutamine (Sigma, St. Louis, MO), 10% fetal bovine serum (FBS, Lonza) and antibiotics (50U/ml Penicillin and 50μg/ml streptomycin, Sigma). Human LX2 were cultured in DMEM-Glutamax (Invitrogen, Carlsbad, CA) supplemented with 10% FBS and antibiotics (50U/ml Penicillin and 50μg/ml streptomycin, Sigma).

In vitro effects of Avagacestat in human HSC and mouse fibroblasts

Cells were seeded in 24 well plates (30,000 cells/well for stainings) and 12 well plates (80,000 cells/well for quantitative PCR analysis) and cultured overnight. To assess the effects on fibrotic parameters, cells were starved for 24h and then incubated with starvation medium alone, Avagacestat (Selleckchem Boston, USA; 5μM and 10μM for stainings and 10μM for quantitative PCR) and 5ng/ml of human recombinant TGFβ1 (Roche, Mannheim, Germany) for 24h. Cells (24 well plates) were then fixed with chilled acetone:methanol (1:1) for 20min, dried and stained for collagen I, α-SMA and vimentin (antibodies and dilutions are summarized in supplementary Table 1). In addition, cells (12 well plates) were lysed with RNA lysis buffer constituted with β-mercaptoethanol (Sigma) to perform quantitative real time PCR analyses for fibrotic parameters (Collagen 1α1, α-SMA and vimentin) and Notch pathway genes (Notch-1, -2 and -3, Dll-1 and -4, Jag1 and Hes1). Stainings and quantitative PCR analysis was performed on three independent experiments.

To assess effects on cell viability, cells plated in 96 well plates were serum-starved for 24h and incubated with different concentrations of Avagacestat (0.5, 1, 5 and 10uM) for 24h. Cell viability assay was performed using Alamar Blue reagent (Invitrogen). The results are represented as % cell viability normalized to untreated control cells (at 100%). All measurements were performed in triplicates in three independent experiments.

3D collagen I gel contraction assay

A collagen gel contraction assay was performed as described previously⁴¹ with minor modifications. We used this assay to examine the inhibitory effect of Avagacestat on the contractile activity of LX2 cells. A collagen suspension (5ml) containing 3.0ml Collagen G1 (5mg/ml, Matrix biosciences, Morlenbach, Germany), 0.5ml 10x M199 medium (Sigma), 85ul 1N NaOH (Sigma) and sterile water was mixed with 1.0ml (2×10⁶ cells) human LX2 cells. Collagen gel-cells suspension (0.6ml/well) was plated in a 24‐well culture plate and allowed to polymerize for 1h at 37°C. Once polymerized, 1ml of 0.5% serum containing medium was added with or without TGFβ (5ng/ml) together with 10μM Avagacestat or PBS followed by detachment of the gels from the culture wells. Digital images were made at 0, 24, 48 and 72h using a Nikon Coolpix digital camera (Nikon, Mississauga, ON, Canada). Measurement of collagen gel diameter at the indicated time points was performed using Image J imaging software (NIH, Bethesda, MD). Gels were measured by tracing around the edges of the gel disk and normalized with their respective well size in each image. Gel contraction experiments were performed in duplicates in three independent experiments.

In vitro effects of Avagacestat in differentiated mouse RAW Macrophages

For differentiation of RAW macrophages, cells were plated (1×10⁶ cells/well) in 12 well plates and cultured overnight. The cells were then incubated with LPS (10ng/ml) and IFNγ (10ng/ml) for M1 differentiation, and IL-4 (10ng/ml) and IL-13 (10ng/ml) for M2 differentiation for 24h. To study the effect of Notch inhibitor, 10μM Avagacestat was added with M1- and M2-differentiating cytokines and incubated for 24h. Cells were then lysed with RNA lysis buffer and samples were analyzed for M1-specific genes (IL-1β, NOS2 and IL-6) and M2-specific genes (MRC1 and Arginase 1) by quantitative PCR. All the experiments were performed in three independent experiments.

Co-culture and conditioned medium studies

Nitric Oxide (NO) release bioassay

The effect of Avagacestat on M1 macrophages was assessed by measuring the inhibition in nitrite NO₂ release, a stable NO metabolite produced by M1 macrophages. RAW cells (2×10⁵ cells/200ul/well) seeded in 96-well plates were cultured with IFNγ (10ng/ml) and LPS (10ng/ml) together with different concentrations of Notch inhibitor Avagacestat (0, 0.5, 1, 5 and 10μM). After 24h, 100μl culture supernatant were added to 100μl of Griess reagent (1% sulfanilamide; 0.1% naphthylethylendiamine dihydrochloride; 3% phosphoric acid) and absorbance at 540nm was measured with a microplate reader.

Animal Experiments

All the animal experiments in this study were performed in strict accordance with the guidelines and regulations for the Care and Use of Laboratory Animals, University of Twente, The Netherlands. The protocols were approved by the Institutional Animal Ethics Committee of the University of Twente, The Netherlands. Male 6- to 8-week old C57BL/6 mice were purchased from Harlan (Zeist, Netherlands) and kept at 12h light/12h dark cycles with ad libitum normal diet.

CCl₄-induced 4 weeks and 8 weeks liver fibrosis mouse models

Male Balb/c mice (6–10 weeks old; n=5 per group) were treated with increasing concentration of carbon tetrachloride (CCl₄) (week 1: 0.5ml/kg; week 2: 0.8ml/kg and week 3–8: 1ml/kg prepared in olive oil) twice weekly by intra-peritoneal injections for 4 or 8 weeks. Olive oil treated non-fibrotic mice are used as controls. The mice were sacrificed at week 4 and week 8 by cervical dislocation and livers were collected for subsequent analysis.

CCl₄-induced acute liver injury mouse model

To study the inhibitory effect of Notch inhibitor, male C57BL/6 mice (20–22g) were treated with a single intraperitoneal injection of CCl₄ (1ml/kg in olive oil) at day 1.Olive-oil treated mice were used as controls. At day 2 and day 3, CCl₄-treated mice intraperitoneally received 10mg/kg Avagacestat prepared in 1% DMSO (Sigma) and 5% β-hydroxycyclodextrin (Sigma) or vehicle treatment (1%DMSO/5%β-hydroxycyclodextrin/PBS) (n=5 per group). At day 4, all mice were sacrificed and livers were harvested for subsequent analysis.

Progressive 4 and 8 weeks model has been established in Balb/c mice while C57BL/6 mouse strain was used for acute liver injury model. From our experience, C57BL/6 mice have high mortality, following 4–8 weeks CCl₄ treatment and leads to multi-organ fibrosis. While CCl₄ administration in Balb/c mice, leads to slow and progressive fibrosis and hence are not suitable for acute liver injury model. Therefore, we have established acute liver injury model in C57BL/6 to have early liver fibrogenesis and Balb/c for progressive (advanced) fibrosis. The difference between these mice strains lies in the differences in Th1 and Th2 responses⁴². Th1-dominance response in C57BL/6 strain and heightened Th2-responses in Balb/c mice determine the severity of liver fibrosis in these mice strains.

Immunohistochemistry and Immunoflourescence

Liver tissues were harvested and transferred to Tissue-Tek OCT embedding medium (Sakura Finetek, Torrance, CA), and snap-frozen in 2-methyl butane chilled in a dry ice. Cryosections (4μm) were cut using a Leica CM 3050 cryostat (Leica Microsystems, Nussloch, Germany). The sections were air-dried and fixed with acetone for 10min. Cells or tissue sections were rehydrated with PBS and incubated with the primary antibody in appropriate dilution (refer to supplementary Table 1) for 1h at room temperature. Cells or sections were then incubated with horseradish peroxidase (HRP)-conjugated secondary antibody for 1h at room temperature. Then incubated with HRP-conjugated tertiary antibody or donkey anti-goat Alexa 594 labeled tertiary antibody (Life Technologies, Gaithersburg, Md) for 1h, after which these were washed thrice with 1x PBS. Thereafter, peroxidase activity was developed using AEC (3-amino-9-ethyl carbazole) substrate kit (Life Technologies) for 20min and nuclei were counterstained with hematoxylin (Fluka Chemie, Buchs, Switzerland). For tissue sections, endogenous peroxidase activity was blocked by 3% H₂O₂ prepared in methanol. Cells or sections were mounted with Aquatex mounting medium (Merck, Darmstadt, Germany). The staining was visualized and the images were captured using light microscopy (Nikon eclipse E600 microscope, Nikon, Tokyo, Japan). For immunofluorescence, sections were mounted with DAPI containing mounting medium and examined using Hamamatsu NanoZoomer Digital slide scanner 2.0HT (Hamamatsu Photonics, Bridgewater NJ).

Quantitative histological analysis

For quantitation, stained sections were scanned at high resolution using Hamamatsu NanoZoomer Digital slide scanner 2.0HT (Hamamatsu Photonics). High resolution scans were viewed using NanoZoomer Digital Pathology (NDP2.0) viewer software (Hamamatsu Photonics). About 20 images (100x) of each entire section (from NDP) were imported into NIH ImageJ software (NIH, Bethesda, MD) and were analyzed quantitatively at a fixed threshold.

RNA extraction, reverse transcription and quantitative real time PCR

Total RNA from cells and liver tissues was isolated using GenElute Total RNA Miniprep Kit (Sigma) and SV total RNA isolation system (Promega Corporation, WI, USA) respectively according to manufacturer’s instructions. The RNA concentration was quantitated by a UV spectrophotometer (NanoDrop Technologies, Wilmington, DE). Total RNA (1 μg) was reverse-transcribed using iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). All the primers were purchased from Sigma-Genosys (Haverhill, UK). Real-time PCR was performed using 2x SensiMix SYBR and Fluorescein Kit (Bioline, QT615-05, Luckenwalde, Germany), 20 ng cDNA and pre-tested gene-specific primer sets (listed in supplementary Table 2 and 3). The cycling conditions for the BioRad CFX384 Real-Time PCR detection system were 95°C for 10min, 40 cycles of 95°C/15sec, 58°C/15sec and 72°C/15sec. Finally, cycle threshold (Ct) values were normalized to reference gene GAPDH and fold changes in expression were calculated using the 2^−ΔΔCt method.

Western blot analysis

Liver tissues were homogenized in cold RIPA buffer (50mM Tris–HCl, 150mM NaCl, 0.1% SDS, 0.1% Igepal in 0.5% sodium deoxycholate with 1 tablet of protease inhibitor cocktail and 1 tablet of phosphatase inhibitor (Roche Diagnostics, Mannheim, Germany) in 10ml) on ice with a tissue homogenizer and the lysates were centrifuged at 12,000 rpm for 1h at 4°C. The supernatants were stored at −70°C until use. The samples were boiled in standard protein sample buffer (Life Technologies) and subjected to SDS-PAGE with 10% Tris-glycine gels (Life Technologies) followed by protein transfer onto PVDF membrane. The membranes were developed according to the standard protocols using primary and secondary antibodies as mentioned in supplementary Table 4. The bands were visualized using ECL detection reagent (Perkin Elmer Inc., Waltham, MA) and photographed using FluorChem M Imaging System (ProteinSimple, Alpha Innotech, San Leandro CA). Intensity of individual bands was quantified using ImageJ densitometry software, and expressed in % relative to β-actin.

This work was supported by the VENI grant from Netherlands Organization for Scientific research (NWO) and University of Twente.