LY411575

Notch-1 and Notch-3 Mediates Hypoxia-Induced Synovial Fibroblasts Activation in
Rheumatoid Arthritis
Jianhai Chen1,2, Wenxiang Cheng1
, Jian Li1,2, Yan Wang1
, Jingqin Chen1
, Xin Shen1
, Ailing Su1
Donghao Gan6
, Liqing Ke1
, Gang Liu4
, Jietao Lin1,2, Liang Li5
, Xueling Bai1
, Peng Zhang1,2,3,4*
This study was supported by the National Key R&D Program of China (2018YFC1705205);
Foreign cooperation project of Chinese Academy of Sciences (GJHZ2063); National Natural
Science Foundation of China (92068117); Guangdong Basic and Applied Basic Research Fund
(2020B1515120052); Science and Technology Innovation Fund of ShenZhen
(JCYJ20170818153602439, JCYJ20180302150101316); Sanming Project of Medicine in
Shenzhen (SZSM201808072); Development and Reform Commission of Shenzhen Municipality
(XMHT20190106001); Shenzhen Double Chain Project for Innovation and Development Industry
supported by Bureau of Industry and Information Technology of Shenzhen (201908141541)
Jianhai Chen, PhD, Wenxiang Cheng, PhD, Jian Li, PhD, Yan Wang, PhD, Jingqin Chen, PhD,
Xin Shen, MS, Ailing Su, MS, Liqing Ke, BS, Jietao Lin, MS, Xueling Bai, PhD, Peng Zhang,
MD: Center for Translational Medicine Research and Development, Shenzhen Institutes of
Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong, China; 2
Jianhai
Chen, PhD, Jian Li, PhD, Jietao Lin, MS, Peng Zhang, MD: University of Chinese Academy of
Sciences, Beijing, China;
Peng Zhang, MD: Shenzhen Engineering Research Centre for Medical Bioactive Materials; 4
Accepted Article
This article is protected by copyright. All rights reserved
Gang Liu, MD, Peng Zhang, MD: Shenzhen Hospital, University of Chinese Academy of
Sciences; 5
Liang Li, PhD: Institutes of Biomedicine and Biotechnology, Shenzhen Institutes of
Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong, China; 6
Donghao
Gan, MD: Shandong University of Traditional Chinese Medicine.
* Address correspondence to Peng Zhang, MD, Center for Translational Medicine Research and
Development, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences,
Shenzhen, Guangdong, China (email: [email protected])
ABSTRACT
Objective. To fully understand the molecular mechanism of hypoxia-induced rheumatoid arthritis
synovial fibroblast cell (RASFC) activation via Notch-1 and Notch-3 signalling, and to evaluate its
potential as a therapeutic target.
Methods. Notch-1 and Notch-3 intracellular domain (N1ICD), Notch-3 intracellular domain
(N3ICD), and hypoxia-inducible factor-1α (HIF-1α) were detected in RA synovial tissues via
immunohistology. RASFC were cultured under hypoxic and normoxic conditions with or without
small interfering RNAs, and N1ICD and N3ICD were overexpressed under normoxic conditions.
Collagen-induced arthritis (CIA) rats were administered with LY411575 (inhibition of N1ICD and
Accepted Article
This article is protected by copyright. All rights reserved
N3ICD) for 15 and 28 days, and its therapeutic efficacy was assessed by histology, radilology and
inflammatory cytokine detection.
Results. In the study, we found that N1ICD, N3ICD and HIF-1α were abundantly expressed in
RA patient synovial tissues. Meanwhile, HIF-1α was found to directly regulate the expression of
Notch-1 and Notch-3 genes under hypoxic conditions. Moreover, hypoxia induced N1ICD and
N3ICD expression in RASFC was blocked by HIF-1α small interfering RNA (siHIF-1α). Notch-1
small interfering RNA (siNotch-1) and Notch-3 small interfering RNA (siNotch-3) inhibited
hypoxia-induced RASFC invasion and angiogenesis in vitro, whereas N1ICD and N3ICD
overexpression promoted these processes. In addition, it was revealed that Notch-1 regulates
RASFC migration and epithelial-mesenchymal transition (EMT) under hypoxia, whereas Notch-3
regulates anti-apoptosis and autophagy. Further, in vivo studies showed that N1ICD and N3ICD
inhibitor LY411575 had a therapeutic effect on CIA rats.
Conclusions. Collectively, this study has identified a functional link between HIF-1α, Notch-1,
and Notch-3 signalling in regulating RASFC activation and rheumatoid arthritis.
KEYWORDS
Rheumatoid arthritis, Hypoxia, Notch-1, Notch-3, Synovial fibroblast
Accepted Article
This article is protected by copyright. All rights reserved
INTRODUCTION
Rheumatoid arthritis (RA) is a chronic autoimmune disease characterised by synovial tissue
hyperplasia and inflammation around the joints, resulting in articular cartilage and subchondral
bone degeneration (1, 2). Normal synovium has a thin membrane comprising 2–3 layers of cells.
In contrast, aberrant rheumatoid arthritis synovial fibroblast cell (RASFC) activation leads to
significant synovial tissue hyperplasia, manifesting as 10–15 layers of cells (3). Studies have
shown that abnormal RASFC activation mediates RA pathogenesis (4). RASFC retain destructive
activation potential in the absence of inflammation in severe combined immunodeficient (SCID)
mice (5, 6). The hypoxic synovium environment is one of the key factors driving RASFC
activation (7, 8). To adapt to the unfavourable microenvironment of inflamed joints, RASFC
develop an abnormal phenotype characterised by increased invasion and impaired apoptosis (9,
10).
Hypoxia is the primary factor that induces hypoxia-inducible factor (HIF) stabilisation. HIF is a
heterodimeric transcription factor comprising an HIF-α and HIF-β subunit. The HIF-β subunit is
stably expressed in the nucleus, while HIF-α expression is regulated by oxygen (11). HIF-1α is
highly expressed in RA synovial tissues and is involved in regulating the transcription of over 60
target genes associated with cellular biological behaviours (12).
The Notch signalling pathway facilitates contact-dependent signalling between cells. In
mammals, there are four Notch receptors (Notch-1–4) and five Notch ligands (Delta-like [DLL]-1,
DLL-3, DLL-4, Jagged1, and Jagged2), all of which are membrane proteins. Following cleavage
by γ-secretase, the Notch intracellular domain (NICD) is translocated to the nucleus where it
interacts with CSL transcription factors (13, 14). Notch signalling plays a key role in the
pathogenesis of RA, and HIF-1α regulates the Notch signalling pathway in tumour cells (15, 16).
However, few studies have been conducted in RASFC.
Accepted Article
This article is protected by copyright. All rights reserved
Given the role of HIF-1α and the Notch signalling pathway in tumour cells, we hypothesised
that HIF-1α and Notch interact in RASFC under hypoxia. Therefore, we aimed to investigate the
role of Notch signalling in collagen-induced arthritis (CIA) and elucidate the molecular
mechanism of hypoxia-induced Notch signalling-mediated RASFC activation in RA. Further, we
examined the potential of Notch signalling as a pharmacological target for RA treatment.
MATERIALS AND METHODS
Patient recruitment and arthroscopy. RA and mild osteoarthritis patients (OA: control group)
were recruited from the Orthopaedic Department of the University of Chinese Academy of
Sciences Shenzhen Hospital. The median age of RA patients was 61 (42–78) years and
osteoarthritis control was 53 (37–71) years. The study was approved by the institutional ethics
committee (LL-KT-2020288) and was conducted in accordance with the Declaration of Helsinki.
Synovial tissue specimens from patients with inflammatory arthritis were obtained by arthroscopy
under local anaesthetic using a Wolf 2.7-mm needle arthroscope (Storz, Germany), as previously
described. Biopsy specimens were either embedded in formaldehyde solution (Sigma-Aldrich, St.
Louis, MO, USA) or snap-frozen in liquid nitrogen for further analysis.
Immunohistochemical analyses. Immunohistochemical staining was conducted for the N1ICD
and N3ICD antibodies their localized distribution in the synovium. The abovementioned steps for
tissue sectioning were followed to retrieve slices of the samples. The sections were placed in
xylene and alcohol concentration gradient to rehydrate the tissue. Boiling 1% citrate buffer was
used for antigen repair for 15 min. The operational instructions of a HRP/DAB (ABC) Detection
IHC Kit (Abcam; ab64264) were then followed. The antibodies used were N1ICD antibody
(1:200, Abcam, ab83232) and N3ICD (V1662) antibody (1:500, Sino Biological Inc, Beijing,
China). The mean optical density (MOD) was calculated to analyse the semi-quantitative
expression of N1ICD and N3ICD using Image-Pro plus 6.0 software.
Culture of synovial fibroblasts. Primary RASFC were isolated from the synovial tissue of RA
Accepted Article
This article is protected by copyright. All rights reserved
patients by digestion with 1 mg/mL collagenase type 1 (Sigma-Aldrich). CD34-positive and
VEGFR-2–negative cells, i.e., RASFC (17), were sorted to > 90% purity by Flow cytometer cell
sorter (BD FACSAria II, USA). RASFC were cultured in RPMI 1640 (Hyclone, Wilmington, DE,
USA) containing 10% FBS (BRL Life Technologies) and 1% penicillin/streptomycin solution
(Hyclone). RASFC were cultured under normoxia and 3% O2, reflecting the joint environment in
vivo (18). Dissociated cells were grown to confluence and used between passages 4 and 8.
Chromatin IP
RASFCs were exposed to 20% or 3% O2 for 12 h and crosslinked in the presence of 3.7%
formaldehyde for 10 min. Immunoprecipitation with HIF-1α (1:50, #14179, CST) was performed
according to manufacturer’s instructions provided for the SimpleChIP® Enzymatic Chromatin IP
kit (#9002, CST) and incubated overnight. Candidate binding sites were analysed by qPCR using
primers that flanked the potential binding site sequences. The primer sequences were as follows:
Notch-1 forward, 5ʹ-AACGAGAAGTAGTCCCAGGC-3ʹ; reverse, 5ʹ-GCACTAGTGAGGCT
CAGAGT-3ʹ. Notch-3 forward, 5ʹ-GGGCACAGGTCCTTGATGTA-3ʹ; reverse,
5ʹ-GGCATGCAGGGAAAAGTGTC-3ʹ. QPCR result analysis, input % = 2% x2 (C [T] input) – C
[T]ip).
Real-time PCR and Western blot analysis. (See supplementary file material and methods)
RNAi gene silencing and NICD overexpression. RASFC were cultured and transfected in
6-well plates with 5 μl siRNA duplexes (HIF-1α, Notch-1, Notch-3, or Scramble), and were
diluted with 200 μl Opti-MEM (Thermo Fisher) serum-free medium and 5 μl Lipofectamine 3000
(Thermo Fisher Scientific, Waltham, MS, USA). Cells were diluted with 200 ul Opti-MEM
serum-free medium and mixed gently. The cells were incubated at room temperature for 20 min in
the dark and added to the wells containing 1590 μl medium. Cells were then incubated for 24 h.
Scrambled control (a nonsense siRNA of the target sequence) and the following siRNA duplexes
were sourced from GenePharma (Shanghai, China), (Supplementary Tables). The Notch-1
Accepted Article
This article is protected by copyright. All rights reserved
intracellular domain (N1ICD amino acids: 1,770 to 2,555) and Notch-3 intracellular domain
(N3ICD amino acids: 1663 to 2312) were cloned into the eukaryotic expression plasmid
PMV-amp (BGI China, Shenzhen, China). The RASFC were cultured in a 6-well plate at a density
of 1 × 105
cells/well in 2 ml RPMI 1640 (Hyclone), containing 10% FBS (GIBCO, Life
Technologies) and 1% penicillin/streptomycin solution (Hyclone). Five micrograms of DNA
(N1ICD, N3ICD, or vector) were diluted with 250 μl Opti-MEM serum-free medium. After
adding P3000 reagent and 10 μl Lipofectamine 3000, the cells were diluted with 245 μl
Opti-MEM serum-free medium. The two solutions were combined with gentle mixing and
incubated at room temperature for 20 min in the dark. Medium (1490 μl) was added to the cell
containing the cells, and these were incubated for 24 h.
Animals. 85 female Wistar rats (9–10 weeks old) were purchased from Beijing Vital River
Laboratory Animal Technology Co., Ltd. (China). The rats were fed in a specific pathogen-free
(SPF) facility. The rat study plans were approved by the animal ethical and welfare committee of
the Shenzhen Institute of Advanced Technology, Chinese Academy of Science
(STAT-IACUC-190723-KYC-ZP-A0804). (See supplementary file material and methods)
Statistical analysis. SPSS 17.0 software was used for analysis. One-way analysis of variance
(ANOVA) was used for multiple group analysis. Student’s t-tests were used to analyse data from
two groups. Results were expressed as mean ± standard deviation (SD). P < 0.05 was considered
significant, and P < 0.01 was considered highly significant.
RESULTS
N1ICD, N3ICD, and HIF-1α are abundantly expressed in the synovial tissues of patients
with RA and RASFC. Previous studies have shown that Notch-1 and Notch-3 are highly
expressed in RA synovial tissues (19), however, it remained unclear whether their activation
fragments, N1ICD and N3ICD, are also highly expressed. To examine localised N1ICD, N3ICD,
and HIF-1α expression, immunohistological analysis was performed on synovial tissue sections
Accepted Article
This article is protected by copyright. All rights reserved
from RA and osteoarthritis (OA: control group) patients. N1ICD (Figure 1A), N3ICD (Figure 1A),
and HIF-1α (online supplementary Figure 1A) were detected in the synovial tissue of RA patients,
and minimal traces were observed in the control group. No IgG control expression was observed.
Additionally, N1ICD was highly expressed in the lining layer and pannus in RA (online
supplementary Figure 1B-i), while N3ICD was also highly expressed in the pannus (online
supplementary Figure 1B-ii). The co-staining results of endothelial tubules (CD31) and the target
antibody further revealed that N1ICD and N3ICD were highly expressed around the pannus
(Figure 1B). Notch-1, N1ICD, Notch-3 and N3ICD were detected via WB in the synovial tissue of
the RA and control groups (Figure 1C and online supplementary Figure 1D).
Primary RASFC were isolated from RA synovial tissues and cultured under normoxia or
hypoxia, and the RASFC were transfected with a scrambled control (Scr) and HIF-1α small
interference RNA (siHIF-1α). Hypoxia induced the mRNA expression of Notch-1 and Notch-3,
which was subsequently inhibited to near-basal levels following a 12 h siHIF-1α treatment (online
supplementary Figure 2A). Compared with normoxia, there was no significant difference in
Notch-2 and Notch-4 in hypoxic 12h (online supplementary Figure 2B). Meanwhile, hypoxia
induced Notch-1, N1ICD, Notch-3, and N3ICD protein expression, which were subsequently
completely inhibited by siHIF-1α (Figure 1D and online supplementary Figure 1C). Similar to
previous studies (17, 19), we found that hypoxia exposure for 12 h promoted the mRNA
expression of Notch signalling ligand Delta-like 4 (DLL4) and Jagged1 (online supplementary
Figure 2C). Additionally, knockdown of DLL4 and Jagged1 with small interfering RNA in the
hypoxia for 24h inhibited the expression of N1ICD and N3ICD (online supplementary Figure 2D
and 2E).
Next, to investigate the expression of Notch target genes, Hes1 and Hey1, we transfected
RASFCs with siNotch-1 and siNotch-3 for 12 h under 3% O2. The results showed that hairy and
enhancer-of-split-1 (Hes1) and hairy/enhancer-of-split related with YRPW motif protein 1 (Hey1)
mRNA were significantly inhibited (online supplementary Figure 2F-i, ii). Meanwhile,
overexpression of N1ICD and N3ICD in normoxic conditions significantly increased the
expression of Hes1 and Hey1 mRNA (online supplementary Figure 2F-iii, iv). To verify whether
Accepted Article
This article is protected by copyright. All rights reserved
HIF-1α directly regulates the expression of Notch-1 and Notch-3 genes, Chromatin IP detection
was performed, which revealed that HIF-1α-binding of N1ICD and N3ICD promoter sites in the
hypoxia for 12h was significantly higher than that observed in the normoxic control group (Figure
1E). Figure 1F shows a representative image of nuclear HIF-1α, N1ICD, and N3ICD staining in
RASFC following exposure to hypoxia. These data suggest that the activation of Notch-1 and
Notch-3 in RASFC under hypoxia depends on HIF-1α. In order to verify the specificity of N3ICD
(V1662) antibody, siNotch-3 was transfected and N3ICD was overexpressed in hypoxia
environment. The results showed that there was a similar result between N3ICD (V1662) and
Notch-3 (N3ICD) antibody (C-terminus, STAN) (online supplementary Figure 1E).
Hypoxia-induced angiogenesis and invasion are dependent on Notch-1 and Notch-3.
Angiogenesis is closely related to the pathological development of RA. To further explore whether
hypoxia-induced angiogenesis depends on Notch signalling, indirect co-culturing of endothelial
cells (HUVEC) and RASFCs was performed in Transwells (20). RASFC were transiently
transfected with Notch-1 small interfering RNA (siNotch-1), Notch-3 small interfering RNA
(siNotch-3), or scrambled control (Scr) (Figure 2A-i). Then co-cultured with HUVEC for 12
hours. Hypoxia significantly induced HUVEC network formation on Matrigel which was
significantly inhibited by siNotch-1 and siNotch-3 (Figure 2C). To further confirm that network
formation is regulated by Notch-1 and Notch-3, RASFC were transfected with an
N1ICD-overexpressing plasmid (N1ICD), an N3ICD-overexpressing plasmid (N3ICD), or a
control vehicle and cultured for 24 h under normoxia (Figure 2A-ii). Compared with the vehicle
group, N1ICD and N3ICD overexpression significantly increased RASFC tube formation under
normoxia (Figure 2C). Hypoxia significantly induced vascular endothelial growth factor (VEGF)
protein expression in RASFC, whereas siNotch-1 [Figure 2E (top) and online supplementary
Figure 3A-i] and siNotch-3 [Figure 2F (top) and online supplementary Figure 3B] significantly
suppressed VEGF protein expression. Under normoxia, N1ICD and N3ICD overexpression
significantly upregulated VEGF [Figure 2E, 2F (right) and online supplementary Figure 3B, 3D].
These data suggest that Notch-1 and Notch-3 regulate the expression of VEGF in RASFC under
Accepted Article
This article is protected by copyright. All rights reserved
hypoxia.
To determine whether Notch-1 and Notch-3 signalling are associated with RASFC invasion, an
invasion assay was performed. Hypoxia significantly enhanced the invasive capacity of RASFC,
but this was suppressed by siNotch-1 and siNotch-3 (Figure 2B-i and Figure 2D). Additionally,
N1ICD and N3ICD overexpression significantly enhanced the invasive capacity of RASFC under
normoxia (Figure 2B-ii and 2D). Matrix metalloproteinase2 (MMP2), MMP3, MMP9, and
MMP13 protein expression in RASFC were measured to further elucidate the molecular
mechanisms of invasion. The results showed that siNotch-1 significantly inhibited MMP2 and
MMP9 expression under hypoxia (Figure 2E and online supplementary Figure 3A-ii,iii), while
siNotch-3 inhibited MMP3 and MMP13 expression (Figure 2F and online supplementary Figure
3C-ii,iii). Under normoxia, N1ICD overexpression significantly upregulated MMP2 and MMP9
expression (Figure 2E and online supplementary Figure 3B) while N3ICD overexpression
upregulated MMP3 and MMP13 expression (Figure 2F and online supplementary Figure 3D). In
hypoxic environment, Notch-1 regulates the expression of MMP2 and MMP9, while Notch-3
regulates the expression of MMP3 and MMP13.
Hypoxia-induced migration and epithelial-mesenchymal transition (EMT) are dependent
on Notch-1. The migration of RASFC plays a key role in the destruction of cartilage and the
development of RA pathology. To examine the role of Notch-1 signalling in hypoxia-induced
RASFC migration, we dynamically monitored the random motility of Scr and siNotch-1 cells
exposed to 20% or 3% O2 for 18 h. Mean cell velocity determined at 6-h intervals revealed
increased velocity starting at 6 h of exposure to 3% O2, whereas cells exposed to 20% O2 retained
a constant velocity throughout the experiment, while siNotch-1 significantly decreased RASFC
migration [Figure 3A and B (left)]. However, N1ICD overexpression significantly increased
RASFC migration [Figure 3A and B (right)]. To confirm these effects on cell migration, a scratch
wound healing assay was performed. Compared with the control group, the migration area of
RASFC treated with siNotch-1 hypoxia for 18 hours was significantly reduced, while the
migration area was significantly increased by overexpression of N1ICD. Overexpression of
Accepted Article
This article is protected by copyright. All rights reserved
N3ICD had no significant effect on RASFC migration (online supplementary Figure 4A and 4B).
The expression of EMT markers such as Snail, Vimentin, and E-cadherin in RASFC was
measured to assess the molecular mechanism underpinning RASFC migration. It was found that
hypoxia significantly induced Snail and Vimentin expression but inhibited E-cadherin expression
in RASFC. However, these changes were reversed by siNotch-1 [Figure 3C (left) and online
supplementary Figure 4C]. Furthermore, N1ICD overexpression significantly upregulated Snail
and Vimentin expression, but downregulated E-cadherin expression in RASFC [Figure 3C (right)
and online supplementary Figure 4D]. Our data show that the migration of RASFC induced by
hypoxia is dependent on Notch-1.
Notch-3 regulates RASFC apoptosis and autophagy. A remarkable feature in the
pathogenesis of RA is the proliferation of synovial tissue. Synovial hyperplasia involves decreased
RASFC apoptosis and increased autophagy. To determine whether hypoxia inhibits RASFC
apoptosis through Notch-3, RASFC apoptosis by Annexin V and propidium iodide was measured
using flow cytometry. RASFC apoptosis was inhibited after 24 h of hypoxia, whereas siNotch-3
increased cell apoptosis [Figure 4A-i and D (left)]. N3ICD overexpression also prevented RASFC
apoptosis, while there was no significant difference in apoptosis by overexpression of N1ICD
[Figure 4A-ii and D (right)]. Hypoxia significantly inhibited cleaved caspase3 and cleaved Poly
(ADP-ribose) polymerase (PARP) protein expression in RASFC, but this was abrogated by
siNotch-3 [Figure 4C (left) and online supplementary Figure 5A]. Furthermore, N3ICD
overexpression also significantly reduced cleaved caspase3 and cleaved PARP protein expression
in RASFC under normoxia [Figure 4C (right) and online supplementary Figure 5B].
To determine whether hypoxia-induced RASFC autophagy is mediated by the Notch signalling
pathway, autophagosomes were assessed by imaging the association of monodansylcadaverine
(MDC) using an autophagy kit. RASFC cultured under hypoxia for 24 h had significantly
increased autophagic bodies. However, siNotch-3 suppressed the increase in MDC staining, and
N3ICD overexpression increased RASFC autophagy (N1ICD is not significant) (Figure 4B and
4E). LC3B is a marker for the autophagy pathway. We found that 24 h of hypoxia significantly
Accepted Article
This article is protected by copyright. All rights reserved
increased LC3B protein expression in RASFC, but this was inhibited by siNotch-3 [Figure 4C
(left) and online supplementary Figure 5A-iii]. N3ICD overexpression significantly increased
LC3B protein expression in RASFC [Figure 4C (right) and online supplementary Figure 5B].
Figure 4D shows a representative image of nuclear LC3B staining in RASFC following exposure
to hypoxia. In hypoxic environment, Notch-3 regulates RASFC to resist apoptosis and increase
autophagy.
The γ-secretase inhibitor LY411575 reduces CIA severity. As high N1ICD and N3ICD
expression was detected in human synovial tissue, we treated CIA rats with LY411575 (inhibition
of N1ICD and N3ICD). LY411575 is a potent γ-secretase inhibitor, and in vitro WB showed that
LY411575 can effectively inhibit N1ICD and N3ICD expression (online supplementary Figure 6A
and 6C). Therefore, the therapeutic efficacy of LY411575 was assessed by examining the clinical
and histopathological characteristics of CIA rats. Methotrexate (MTX) was used as a positive
control. The treatment history of the animals was kept confidential from the researcher during the
experimental phase and tissue section scoring to prevent bias.
Compared to the vehicle group, CIA rats in the MTX and LY411575 (5 mg/kg and 10 mg/kg)
groups had significantly reduced toe and ankle redness and swelling (online supplementary Figure
6B), paw thickness online supplementary Figure 6D), and arthritis scores (online supplementary
Figure 6E) after 15 and 28 days of treatment. However, these symptoms showed no significant
improvements in the LY411575 (1 mg/kg) group.
Next, serum Interleukin 1 beta (IL-1β), Interleukin-6 (IL-6), tumor necrosis factor alpha
(TNF-α) and VEGF levels were measured using ELISA to determine the effect of LY411575 on
cytokine production. TNF-α (Figure 5A-i), IL-6 (Figure 5A-ii) and VEGF (Figure 5A-iv) levels
were significantly lower in the MTX and LY411575 (5 mg/kg and 10 mg/kg) groups compared to
those in the vehicle group after 15 and 28 days of treatment. However, only MTX inhibited IL-1β
(Figure 5A-iii) and LY411575 had no inhibitory effect.
The integrity of the articular cartilage was examined by imaging, and bone quality and quantity
were analysed using micro-CT (CT) on days 15 and 28 of treatment. Three-dimensional
Accepted Article
This article is protected by copyright. All rights reserved
reconstruction of the CT data was performed using SkyScan. Bone damage was clearly visible in
the vehicle and LY411575 (1 mg/kg) groups, while improvement was seen after 15 and 28 days of
MTX and LY411575 (5 mg/kg and 10 mg/kg) treatment (Figure 5B and C). Compared to the
vehicle group, bone mineral density (BMD) (online supplementary Figure 6F-i), bone
volume/tissue volume (BV/TV) (online supplementary Figure 6F-ii) and the trabecular number
(Tb) (online supplementary Figure 6F-iii) were significantly increased after 15 and 28 days of
MTX and LY411575 (5 mg/kg and 10 mg/kg) treatment. However, there were no significant
changes in the LY411575 (1 mg/kg) group.
Haematoxylin and eosin (H&E) staining showed that the knee joints of normal group rats had
smooth surfaces and no inflammatory cell infiltration (Figure 6A). In contrast, the knee joints of
vehicle group CIA rats had extensive damage and rough surfaces, along with abnormal synovial
tissue hyperplasia and substantial inflammatory cell infiltration. These pathological changes were
significantly improved in the MTX and LY411575 (5 mg/kg and 10 mg/kg) groups, but not in the
LY411575 (1 mg/kg) group (Figure 6A-ii). Toluidine blue was used to stain the articular cartilage
(Figure 6B). Compared to the normal group, the rats in the vehicle group had thin knee joint
cartilage, which became nearly invisible by days 28. However, articular cartilage thickness was
significantly improved after 15 and 28 days of treatment in the MTX and LY411575 (5 mg/kg and
10 mg/kg) groups, but not in the 1 mg/kg LY411575 group (Figure 6B-ii). TRAP staining revealed
nearly no TRAP-positive cell accumulation in the joints of normal group rats (Figure 6C).
TRAP-positive cell accumulation was visible in the vehicle group, which was significantly
reduced after 15 and 28 days of treatment in the MTX and LY411575 (5 mg/kg and 10 mg/kg)
groups, but not in the LY411575 (1 mg/kg) group (Figure 6C-ii).
Immunohistochemistry (IHC) staining showed that N1ICD (online supplementary Figure 7A)
and N3ICD (online supplementary Figure 7B) expression was significantly downregulated in
synovial tissues after 15 and 28 days of treatment in the LY411575 (5 mg/kg and 10 mg/kg)
groups, but not in the LY411575 (1 mg/kg) group. The positive mean optical density (MOD)
values of the positive cells showed N1ICD (online supplementary Figure 7A-ii) and N3ICD
(online supplementary Figure 7B-ii) expression. The localisation of N1ICD and N3ICD expression
Accepted Article
This article is protected by copyright. All rights reserved
in CIA rat synovium is consistent with that in human RA synovium. The results of in vivo
experiments show that inhibition of N1ICD and N3ICD can effectively treat CIA rats.
DISCUSSION
Cartilage destruction and synovial tissue proliferation are the main pathological features of RA
pathogenesis (21). Cartilage destruction is mainly due to pannus formation, while the increase in
synovial tissue is mainly due to decreased RASFC apoptosis and increased autophagy (22, 23).
These events are attributed to abnormal RASFC activation (4). Hypoxic microenvironment
stimulates the activation of RASFC that contribute to the pathogenesis of RA (24, 25). However,
how hypoxia activates RASFC remains to be fully understood. Our epistasis study shows that
hypoxia, via HIF-1α, upregulates the expression of Notch-1 and 3 in RASFC activation. Nuclear
staining revealed higher expression of HIF-1α, N1ICD, and N3ICD in RA patient synovial tissue
than in control tissue, and hypoxia induces the nuclear translocation of HIF-1α, N1ICD, and
N3ICD in RASFC. Finally, CIA rats treated with LY411575 reduces N1ICD, N3ICD,
consequently the severity of RA.
The level of oxygen tension (pO2) in synovial fluid of RA patients (51. 0 ±16. 5 mmHg) was
significantly lower than that of osteoarthritis (OA) patients (79. 2 ±14. 0 mmHg) (8). At present,
there is no evidence that hypoxia regulates Notch signaling in the synovium of knee OA. With
reference to previous studies, we took patients with mild OA as the control group (26).
Hyperplastic synovial tissues contain a large number of pannus and immune cells, which can
attack adjacent articular cartilage and subchondral bone (27). The Notch signalling pathway is
critical for embryonic angiogenesis and angiogenesis (28). However pannus formation is an
important pathological feature of RA, and pannus in RA is regulated by hypoxia (9, 29). Here, we
found that hypoxia induced VEGF protein expression via Notch-1 and Notch-3. Pannus formation
can often lead to cartilage invasion, and invasion is generally associated with MMP expression.
Notch-1 can induce MMP2 and MMP9 expression in vascular endothelial cells, while Notch-3 can
induce MMP3 expression in prostate cancer cells under hypoxia (18, 30). Hypoxia also induces
MMP2, MMP3, MMP9, and MMP13 expression in RASFC (31). Consistent with these findings,
Accepted Article
This article is protected by copyright. All rights reserved
our study showed that hypoxia can promote RASFC invasion, and this increase in their invasive
capacity was inhibited by siNotch-1 and siNotch-3.
Hypoxia promotes vascular endothelial cell migration in a Notch-1-dependent manner (18).
Here, we found that hypoxia can promote RASFC migration, and this was inhibited by siNotch-1.
EMT is closely associated with Notch-1 signalling, and Notch-1 can inhibit breast cancer cell
EMT and migration (32, 33). However, the role of Notch-3 in cell migration and EMT induced by
hypoxia is still controversial (34, 35). It was previously reported that hypoxia-induced RASFC
migration usually requires EMT (36). We also found that hypoxia-induced RASFC migration was
mediated through EMT, and Snail, Vimentin, and E-cadherin expression in RASFC was regulated
by Notch-1. Together, these results indicate that hypoxia-induced RASFC migration and EMT are
mediated by Notch-1 signalling.
In RA, synovial proliferation is mainly related to a decrease in RASFC apoptosis. The apoptotic
rate of fibroblast cells below the synovial lining was reported to be 3%, while that in the synovial
lining was close to zero (37). It has been reported that Notch-3 overexpression in breast cancer
cells decreases apoptosis (34). Moreover, Notch-3 plays a critical role in antigen-specific T cell
differentiation, and Notch-3 blockade can inhibit Th1 and Th17 cell activation in CIA mice (38).
We found that hypoxia reduced RASFC apoptosis, whereas siNotch-3 promoted it. In many cases,
the cytoprotective function of autophagy is mediated by negative apoptotic regulation. Apoptosis
signalling in turn inhibits autophagy (39). Similarly, the increase in autophagic bodies in RASFC
under hypoxia can be inhibited by siNotch-3. There was no significant difference overexpressed
N1ICD on anti-apoptosis and autophagy of RASFC. The differential effect of N1ICD and N3ICD
in the RASFC may be related to their expression location in synovium. Similar to previous studies,
we also found that N1ICD was mainly expressed in perivascular parietal cells and the edge of
synovial lining layer (18), while N3ICD was mainly expressed mural cells and perivascular
fibroblasts (40). Synovial tissue is divided into different fibroblast subsets (41). So the differential
effect of N1ICD and N3ICD.
Many current RA treatments have extensive side effects and limited efficacy. Notch signalling
inhibition may therefore be an effective strategy for RA treatment. γ-secretase is a protease
Accepted Article
This article is protected by copyright. All rights reserved
complex that cleaves Notch to release NICD, and γ-secretase inhibitors (LY411575) have been
widely used for analysing and detecting Notch signalling (42).
Taken together, this is the first study to demonstrate that N1ICD and N3ICD are highly
expressed in patients with RA and in CIA rat synovial tissues. Our study provides evidence of a
functional link between HIF-1α, Notch-1, and Notch-3 signalling in the regulation of RASFC
activation in RA. The use of a γ-secretase inhibitor also indicated that the Notch signalling
pathway is a potential pharmacological target for RA treatment.
ACKNOWLEDGEMENTS
We are grateful to Professor Andrew Tan Nguan Soon (Lee Kong Chian School of Medicine,
Nanyang Technological University) for the revision of the manuscript.
Author Contributions
All authors were involved in drafting the article or revising it critically for important intellectual
content, and all authors approved the final version to be published. Jianhai Chen, Jian Li,
Wenxiang Cheng, Xin Shen and Ailing Su performed acquisition of data. Jianhai Chen, Jingqin
Chen and Donghao Gan performed the statistical analyses and interpretation of data. Peng Zhang
and Gang Liu critically interpreted the results the draft version.
Accepted Article
This article is protected by copyright. All rights reserved
REFERENCES
1. Kraan MC, Haringman JJ, Weedon H, Barg EC, Smith MD, Ahern MJ, et al. T cells, fibroblast-like synoviocytes,
and granzyme B+ cytotoxic cells are associated with joint damage in patients with recent onset rheumatoid arthritis.
Ann Rheum Dis. 2004;63(5):483-8.
2. Pitzalis C, Pipitone N, Pipitone V, Fioravanti A, Marcolongo R. [Rheumatoid arthritis. Recent findings and new
pathogenic concepts]. Recenti Prog Med. 2001;92(3):217-22.
3. Noss EH, Brenner MB. The role and therapeutic implications of fibroblast-like synoviocytes in inflammation and
cartilage erosion in rheumatoid arthritis. Immunol Rev. 2008;223:252-70.
4. Neumann E, Lefevre S, Zimmermann B, Gay S, Muller-Ladner U. Rheumatoid arthritis progression mediated by
activated synovial fibroblasts. Trends Mol Med. 2010;16(10):458-68.
5. Lefevre S, Knedla A, Tennie C, Kampmann A, Wunrau C, Dinser R, et al. Synovial fibroblasts spread rheumatoid
arthritis to unaffected joints. Nat Med. 2009;15(12):1414-20.
6. Pap T, Aupperle KR, Gay S, Firestein GS, Gay RE. Invasiveness of synovial fibroblasts is regulated by p53 in the
SCID mouse in vivo model of cartilage invasion. Arthritis Rheum. 2001;44(3):676-81.
7. Konisti S, Kiriakidis S, Paleolog EM. Hypoxia--a key regulator of angiogenesis and inflammation in rheumatoid
arthritis. Nat Rev Rheumatol. 2012;8(3):153-62.
8. Lee YA, Kim JY, Hong SJ, Lee SH, Yoo MC, Kim KS, et al. Synovial proliferation differentially affects hypoxia in the
Accepted Article
This article is protected by copyright. All rights reserved
joint cavities of rheumatoid arthritis and osteoarthritis patients. Clin Rheumatol. 2007;26(12):2023-9.
9. Akhavani MA, Madden L, Buysschaert I, Sivakumar B, Kang N, Paleolog EM. Hypoxia upregulates angiogenesis
and synovial cell migration in rheumatoid arthritis. Arthritis Res Ther. 2009;11(3):R64.
10. Ng CT, Biniecka M, Kennedy A, McCormick J, Fitzgerald O, Bresnihan B, et al. Synovial tissue hypoxia and
inflammation in vivo. Ann Rheum Dis. 2010;69(7):1389-95.
11. Wang GL, Jiang BH, Rue EA, Semenza GL. Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer
regulated by cellular O2 tension. Proc Natl Acad Sci U S A. 1995;92(12):5510-4.
12. Bustamante MF, Garcia-Carbonell R, Whisenant KD, Guma M. Fibroblast-like synoviocyte metabolism in the
pathogenesis of rheumatoid arthritis. Arthritis Res Ther. 2017;19(1):110.
13. Selkoe D, Kopan R. Notch and Presenilin: regulated intramembrane proteolysis links development and
degeneration. Annu Rev Neurosci. 2003;26:565-97.
14. Lai EC. Keeping a good pathway down: transcriptional repression of Notch pathway target genes by CSL
proteins. EMBO Rep. 2002;3(9):840-5.
15. Silvan U, Diez-Torre A, Arluzea J, Andrade R, Silio M, Arechaga J. Hypoxia and pluripotency in embryonic and
embryonal carcinoma stem cell biology. Differentiation. 2009;78(2-3):159-68.
16. Shang Y, Smith S, Hu X. Role of Notch signaling in regulating innate immunity and inflammation in health and
disease. Protein Cell. 2016;7(3):159-74.
17. Galligan CL, Fish EN. Circulating fibrocytes contribute to the pathogenesis of collagen antibody-induced
arthritis. Arthritis Rheum. 2012;64(11):3583-93.
18. Gao W, Sweeney C, Connolly M, Kennedy A, Ng CT, McCormick J, et al. Notch-1 mediates hypoxia-induced
angiogenesis in rheumatoid arthritis. Arthritis Rheum. 2012;64(7):2104-13.
19. Yabe Y, Matsumoto T, Tsurumoto T, Shindo H. Immunohistological localization of Notch receptors and their
ligands Delta and Jagged in synovial tissues of rheumatoid arthritis. J Orthop Sci. 2005;10(6):589-94.
20. Sanchez-Palencia DM, Bigger-Allen A, Saint-Geniez M, Arboleda-Velasquez JF, D'Amore PA. Coculture Assays for
Endothelial Cells-Mural Cells Interactions. Methods Mol Biol. 2016;1464:35-47.
21. Alivernini S, Fedele AL, Cuoghi I, Tolusso B, Ferraccioli G. Citrullination: the loss of tolerance and development
of autoimmunity in rheumatoid arthritis. Reumatismo. 2008;60(2):85-94.
22. Muller-Ladner U, Ospelt C, Gay S, Distler O, Pap T. Cells of the synovium in rheumatoid arthritis. Synovial
Accepted Article
This article is protected by copyright. All rights reserved
fibroblasts. Arthritis Res Ther. 2007;9(6):223.
23. Pap T, Muller-Ladner U, Gay RE, Gay S. Fibroblast biology. Role of synovial fibroblasts in the pathogenesis of
rheumatoid arthritis. Arthritis Res. 2000;2(5):361-7.
24. Fearon U, Canavan M, Biniecka M, Veale DJ. Hypoxia, mitochondrial dysfunction and synovial invasiveness in
rheumatoid arthritis. Nat Rev Rheumatol. 2016;12(7):385-97.
25. Ospelt C, Neidhart M, Gay RE, Gay S. Synovial activation in rheumatoid arthritis. Front Biosci. 2004;9:2323-34.
26. Gao W, Sweeney C, Walsh C, Rooney P, McCormick J, Veale DJ, et al. Notch signalling pathways mediate
synovial angiogenesis in response to vascular endothelial growth factor and angiopoietin 2. Ann Rheum Dis.
2013;72(6):1080-8.
27. Zvaifler NJ, Boyle D, Firestein GS. Early synovitis--synoviocytes and mononuclear cells. Semin Arthritis Rheum.
1994;23(6 Suppl 2):11-6.
28. Park M, Yaich LE, Bodmer R. Mesodermal cell fate decisions in Drosophila are under the control of the lineage
genes numb, Notch, and sanpodo. Mech Dev. 1998;75(1-2):117-26.
29. Cheng WX, Huang H, Chen JH, Zhang TT, Zhu GY, Zheng ZT, et al. Genistein inhibits angiogenesis developed
during rheumatoid arthritis through the IL-6/JAK2/STAT3/VEGF signalling pathway. J Orthop Translat.
2020;22:92-100.
30. Ganguly SS, Hostetter G, Tang L, Frank SB, Saboda K, Mehra R, et al. Notch3 promotes prostate cancer-induced
bone lesion development via MMP-3. Oncogene. 2020;39(1):204-18.
31. Hua S, Dias TH. Hypoxia-Inducible Factor (HIF) as a Target for Novel Therapies in Rheumatoid Arthritis. Front
Pharmacol. 2016;7:184.
32. Sahlgren C, Gustafsson MV, Jin S, Poellinger L, Lendahl U. Notch signaling mediates hypoxia-induced tumor cell
migration and invasion. Proc Natl Acad Sci U S A. 2008;105(17):6392-7.
33. Shariat Razavi SM, Forghanifard MM, Kordi-Tamandani DM, Abbaszadegan MR. MAML1 regulates EMT markers
expression through NOTCH-independent pathway in breast cancer cell line MCF7. Biochem Biophys Res Commun.
2019;510(3):376-82.
34. Gupta N, Xu Z, El-Sehemy A, Steed H, Fu Y. Notch3 induces epithelial-mesenchymal transition and attenuates
carboplatin-induced apoptosis in ovarian cancer cells. Gynecol Oncol. 2013;130(1):200-6.
35. Tan J, Zhang X, Xiao W, Liu X, Li C, Guo Y, et al. N3ICD with the transmembrane domain can effectively inhibit
Accepted Article
This article is protected by copyright. All rights reserved
EMT by correcting the position of tight/adherens junctions. Cell Adh Migr. 2019;13(1):203-18.
36. Li GQ, Zhang Y, Liu D, Qian YY, Zhang H, Guo SY, et al. PI3 kinase/Akt/HIF-1alpha pathway is associated with
hypoxia-induced epithelial-mesenchymal transition in fibroblast-like synoviocytes of rheumatoid arthritis. Mol Cell
Biochem. 2013;372(1-2):221-31.
37. Matsubara T, Velvart M, Odermatt BF, Spycher MA, Ruttner JR, Fehr K. The thickening of basement membrane
in synovial capillaries in rheumatoid arthritis. Rheumatol Int. 1983;3(2):57-64.
38. Jiao Z, Wang W, Xu H, Wang S, Guo M, Chen Y, et al. Engagement of activated Notch signalling in collagen
II-specific T helper type 1 (Th1)- and Th17-type expansion involving Notch3 and Delta-like1. Clin Exp Immunol.
2011;164(1):66-71.
39. Gordy C, He YW. The crosstalk between autophagy and apoptosis: where does this lead? Protein Cell.
2012;3(1):17-27.
40. Wei K, Korsunsky I, Marshall JL, Gao A, Watts GFM, Major T, et al. Notch signalling drives synovial fibroblast
identity and arthritis pathology. Nature. 2020;582(7811):259-64.
41. Mizoguchi F, Slowikowski K, Wei K, Marshall JL, Rao DA, Chang SK, et al. Functionally distinct disease-associated
fibroblast subsets in rheumatoid arthritis. Nat Commun. 2018;9(1):789.
42. Park JS, Kim SH, Kim K, Jin CH, Choi KY, Jang J, et al. Inhibition of notch signalling ameliorates experimental
inflammatory arthritis. Ann Rheum Dis. 2015;74(1):267-74.
Accepted Article
This article is protected by copyright. All rights reserved
Figure.1 The expression of N1ICD and N3ICD in RA patients and their association with hypoxia.
(A) Representative photomicrographs showing the localisation of N1ICD and N3ICD in synovial
tissue sections from RA patients and OA control synovium. (B) Representative microscopy
images of synovial tissues in which N1ICD or N3ICD (green), CD31 (red) and DAPI (blue) are
visualised by immunofluorescence staining. (C) Representative western blot showing Notch-1,
N1ICD, Notch-3, N3ICD, and HIF-1α protein in OA control and RA synovial tissue. (D)
Representative western blots of HIF-1α, Notch-1, N1ICD, Notch-3, and N3ICD in RASFC
following transient transfection with HIF-1α siRNA (Si) or Scr under normoxia or hypoxia (3%
O2) for 24 h. GAPDH was used as a loading control. (E) RASFC were exposed to normoxia or
hypoxia (3% O2) for 16 h, and ChIP analysis was performed using control IgG antibodies or those
against HIF-1α with primers targeted to the promoter region of Notch-1 and Notch-3. Primers
flanking binding sites were used for qPCR. Data are shown as mean ± SEM; n = 3. (F) Nuclear
translocation of N1ICD, N3ICD, and HIF-1α was observed by immunofluorescence.
Figure 2 Notch signalling pathway regulates RASFC angiogenesis and invasion. (A)
Representative photomicrograph of HUVEC co-cultured with RASFC showing tube formation in
a Matrigel assay. (B) Representative photomicrographs showing RASFC invasion for 24 h. (C)
The numbers of connecting branches were quantified in sequential fields (NS, not significant). (D)
Representative bar graph quantifying invasive cells (NS, not significant). (E) Representative
western blots of N1ICD, VEGF, MMP2, and MMP9 in RASFC following transient transfection
Accepted Article
This article is protected by copyright. All rights reserved
with Notch-1 siRNA (Si) or scramble control (Scr) under normoxia or hypoxia (3% O2) and
N1ICD-overexpressed (N1ICD) or control vector -overexpressing (Vector) plasmids for 24 h.
GAPDH was used as a loading control. (F) Representative western blots of N3ICD, VEGF,
MMP3, and MMP13 in RASFC following transient transfection with Notch-3 siRNA (Si) or
scramble control (Scr) under normoxia or hypoxia (3% O2), and N3ICD or control vector for 24
h. GAPDH was used as a loading control.
Figure.3 Notch-1 regulates RASFC migration and epithelial-mesenchymal transition (EMT) under
hypoxia. (A) Representative showing RASFC migration following transient transfection with
Notch-1 siRNA or scramble control (Scr) under normoxia or 3% hypoxia, and N1ICD
overexpression (N1ICD) or control vector (Vector) for 0, 6, 12, 18 h using dynamically
monitored. (B) Maximum displacement from the origin was determined for cells analyzed in A.
(C) Representative western blots of N1ICD, Snail, Vimentin, and E-cadherin in RASFC following
transient transfection with Notch-1 siRNA (Si) or scramble control (Scr) under normoxia or
hypoxia (3% O2), and N1ICD overexpression (N1ICD) or control vector-overexpressing (Vector)
plasmids for 24 h.
Figure.4 Notch-3 regulates RASFC anti-apoptosis and autophagy under hypoxia. (A)
Representative images showing RASFC apoptosis using an Annexin V-FITC and PI
double-stained flow cytometry. Cells in the lower left quadrant of each picture correspond to
normal cells (Annexin V−/PI−). Cells in the lower right quadrant correspond to early apoptotic
cells (Annexin V+/PI−). Cells in the upper right quadrant correspond to late apoptotic/dead cells
(Annexin V+/PI+). (B) Autophagic staining was measured using the cell autophagy detection
assay kit. Fluorescence micrograph showing RASFC autophagy following transient transfection
with Notch-3 siRNA (siNotch-3) or scramble control (Scr) under normoxia, or 3% hypoxia, and
N3ICD-overexpressed (N3ICD) and N1ICD or control vector (Vector). (C) Representative
western blots of N3ICD, Cleaved Caspase-3 and Cleaved PARP in RASFC following transient
transfection with Notch-3 siRNA (Si) or scramble control (Scr) under normoxia or hypoxia (3%
Accepted Article
This article is protected by copyright. All rights reserved
O2), and N3ICD-overexpressed (N3ICD) or control vector (Vector) for 24 h, GAPDH was used as
a loading control. (D) Quantitative results of RASFC apoptosis (NS, not significant). (E)
Quantitative results of RASFC autophagy (NS, not significant).
Figure.5 γ-secretase inhibitor LY411575 ameliorates collagen-induced arthritis (CIA) symptoms.
(A) TNF- , IL-6, α IL-1β, and VEGF concentration in serum at 15 and 28 days after treatment (NS,
not significant). (B) Representative 3D reconstructions of micro-CT images of knee joint. (C)
Radiological score (NS, not significant).
Figure.6 Histological staining. (A) (i) The results of haemotoxylin and eosin (H&E) staining of
knee joints 15 and 28 days after treatment. (ii) H&E semi-quantitative score (NS, not significant).
(B) (i) Articular cartilage toluidine blue staining detection. (ii) Toluidine blue semi-quantitative
score of articular cartilage (NS, not significant). (C) (i) Tissue sections were stained with TRAP to
examine osteoclasts at different time points. (ii) Statistical data of osteoclast scores in the fracture
area of different groups (NS, not significant).