Key words
IL-23 - osteoclasts - ankylosing spondylitis - Ephrin B2 - sphingosine kinase 1
Schlüsselwörter
IL-23 - Osteoklasten - Spondylitis ankylosans - Ephrin B2 - Sphingosinkinase 1
Introduction
Ankylosing spondylitis (AS) is a spinal joint disease with typical symptoms of
inflammatory back pain and affects about 0.3% of the Chinese population
[1]. The pathological features of AS
include enthesitis, erosion of bone and formation of new bone or syndesmophytes
[2]
[3]. In patients with AS, clinical radiographs
have shown unequivocally that erosions and sclerosis precede syndesmophyte
development at the identified sites [4].
However, the underlying mechanisms for syndesmophyte formation in AS have not been
fully elucidated.
IL-23 belongs to the interleukin-12 (IL-12) superfamily, and is released from
dendritic cells and macrophages when they are activated [5]. Mediated by the downstream cytokines IL-22
and IL-17, IL-23 is vital in the functioning of the IL-17/IL-23 axis by
facilitating Th17 cell differentiation. Thus, IL-23 plays a fundamental role in the
pathophysiology of AS [6]. The levels of IL-23
and IL-17 were greatly increased in the sera of AS patients [7]
[8]. Increased IL-23 levels in patients with
active AS might be caused by a subclinical gut inflammation [9]. Also, more IL-23 was secreted in AS
macrophages than healthy macrophages in response to lipopolysaccharide [10]. Moreover, IL-23 levels and Th17 cell
numbers increased in the colon of HLA-B27 transgenic rats, an animal model of AS
[11] and neutralizing anti-IL-23 antibody
treatment was demonstrated to significantly alleviate peripheral arthritis and
spondyloarthritis in the AS animal models [12]. Moreover, a critical study showed that mice with over-expression of
IL-23 spontaneously phenocopied the human AS, with the specific and characteristic
development of enthesitis and entheseal new bone formation [13]. Histological examination revealed many
osteoblasts and multinucleated osteoclasts at the sites of entheseal inflammation.
Taken together, these results indicate that IL-23 is involved in the pathogenesis
of
AS and induces entheseal new bone formation.
Osteoclasts are multinucleated cells derived from hematopoietic precursor cells in
bone marrow and possess bone-resorbing properties. The precursor cells differentiate
into mature osteoclasts when M-CSF and RANKL are present [14]. A number of studies have reported that
IL-23 could promote the differentiation of osteoclasts through RANKL and that IL-23
might indirectly inhibit osteoclastogenesis through T cells [15]
[16]
[17]
[18]. Kamiya et al. [19] reported that no remarkable changes were
observed in the expression of mRNA for type I collagen or ALP in osteoblasts after
stimulation by IL-23, indicating that IL-23 had no direct actions on osteoblasts.
We
confirmed that IL-23 had no direct effects on osteoblasts, since osteoblasts
marginally express IL-23 receptors [20]. Taken
together, the above data imply that IL-23 indirectly induces new bone formation at
the sites of entheseal inflammation.
Bone homeostasis is tightly regulated by signaling pathways that act on osteoblasts
and osteoclasts. In the physiological state, new bone formation and erosion occur
simultaneously during the dynamic equilibrium of bone metabolism. On one hand,
osteoblasts regulate the differentiation of osteoclasts by producing the essential
differentiating factors RANKL and M-CSF and also osteoprotegerin (OPG), which is an
inhibitory factor. On the other hand, osteoclasts promote the differentiation of
osteoblasts by producing osteogenic factors such as the Eph family receptor
interacting proteins B2 (Ephrin B2), sphingosine kinase 1 (SPHK1), bone
morphogenetic proteins (BMP) and Wnt10b, and inhibit the differentiation of
osteoblasts by producing anti-osteogenic factors Semaphorin 4D and high temperature
requirement protease A1 (HtrA1) [21].
Considering the paradox between the significant induction of IL-23 on the formation
of new bone and the fact that it has no direct effect on osteoblasts, we assumed
that IL-23 might participate in the indirect regulation of osteoblast
differentiation through osteoclasts. To elucidate the possible mechanisms involved,
we investigated whether IL-23 could modulate the expression of Ephrin B2, SPHK1,
BMP2, BMP6, Wnt10b, HtrA1 and semaphorin 4D in osteoclasts.
Materials and Methods
Isolation of bone marrow mononuclear cells (BMMCs)
Wild-type C57BL/6 mice (6–8 weeks) were obtained from Shanghai
Jiao Tong University Laboratory Animal Center (Shanghai, China). The mice were
humanely killed by inhalation of carbon dioxide and sterilized with 75%
ethanol. Tibiae and femora were dissected after removing the skin and associated
soft tissues. The epiphyses of the tibiae and femora were severed and bone
marrow isolated by flushing with Hanks I solution. BMMCs were harvested after
centrifugation, incubation with red blood cell lysis buffer, and washing with
Hanks I solution. The ethical committee of the Second Military Medical
University gave approval for all the experimental procedures.
Induction of osteoclasts
The isolated primary BMMCs were cultured for 3 days at 37°C in the
presence of humidified 5% CO2/95% air in
Dulbecco’s modified Eagle medium (DMEM), to which had been supplied 50
μg/mL streptomycin, 50 U/mL penicillin,
10% fetal calf serum and 50 ng/ml M-CSF (PeproTech, US).
Adherent cells were detached, plated onto 12-well plates at
5×105 cells per well and cultured for another 5–6
days supplied with RANKL (30 ng/ml, R&D, US) and M-CSF
(50 ng/ml), which was replaced every 48 h.
For osteoclast induction, murine RAW264.7 cells were also used and cultured in
DMEM containing 30 ng/ml RANKL for 5–6 days. The culture
medium was replaced at 48 h intervals.
Based on previous studies [15]
[16]
[17]
[18], the concentration of IL-23
(PeproTech, USA) was selected as 30 ng/ml in all subsequent
experiments. After identification, osteoclasts were refreshed with medium
without RANKL or M-CSF, and treated with 30 ng/ml IL-23 for the
indicated times.
TRAP staining
After washing in PBS and then fixation in paraformaldehyde solution (4%),
the cells were stained for tartrate-resistant acid phosphatase (TRAP) with a
commercial kit (Sigma, US). Using two specific osteoclast markers,
multinucleated (>3 nuclei) TRAP-positive cells were shown to be mature
osteoclasts.
RNA extraction and real-time RT-PCR
Acid guanidine-phenol-chloroform and TRIzol (BBI sciences, US) was used to
extract RNA from osteoclasts. First-strand cDNA was constructed from the
extracted RNA using a Mix reverse transcription kit (Tiangen Inc., Beijing), and
subsequently utilized as templates in PCR experiments. PCR amplification was
carried out using the specific primers listed in [Table 1]. The genes of interest were
measured using real-time PCR (ABI ViiA7, Applied Biosystems, US) with a Power
SYBR Green Master Mix (Life, US). The expression levels of genes relative to one
another were analyzed using the ΔΔCt method. An expressed gene
that encoded GAPDH was adopted as the internal control to normalize the
quantities of mRNA present in each specimen.
Table 1 Primer sequences used in quantitative real-time PCR
analysis.
|
Gene
|
Primer
|
|
BMP2
|
5’-CCTTCGGAAGACGTCCTCAG-3’
|
|
5’-CTGAGTGCCTGCGGTACAGA-3’
|
|
BMP6
|
5’-CTGCTGAGTTCCGCGTCTAC-3’
|
|
5’-TCCAGCCAACCTTCTTCTGA-3’
|
|
Ephrin B2
|
5’-CGGACAAGGCCTGGTACTAT-3’
|
|
5’-GTCTGGTCTGGCACAGTTGA-3’
|
|
GAPDH
|
5’-GGTCGGTGTGAACGGATTTG-3’
|
|
5’-TGTAGACCATGTAGTTGAGGTCA-3’
|
|
HtrA1
|
5’-AGCGACGCCAAGACCTACAC-3’
|
|
5’-GGCGATCTTCTCCACCACAT-3’
|
|
NFATc1
|
5’-CCTGGAGATCCCGTTGC-3’
|
|
5’-GGTGTTCTTCCTCCCGATGT-3’
|
|
Semaphorin 4D
|
5’-GACAGTGGCCTGGTCTTCAA-3’
|
|
5’-CACTGTGGCACTCTGCATGT-3’
|
|
SPHK1
|
5’-TTCCTGGAGGAGGCAGAGAT-3’
|
|
5’-ACCATCACCGGACATGACTG-3’
|
|
TRAP
|
5’-CGACCATTGTTAGCCACATACG-3’
|
|
5’-TCGTCCTGAAGATACTGCAGGTT-3’
|
|
Wnt10b
|
5’-GGATGGCTGTAACCACGACAT-3’
|
|
5’-GTTGTTGTGGATCCGCATTC-3’
|
PCR=polymerase chain reaction; BMP=bone morphogenetic
protein; GAPDH=glyceraldehyde-3-phosphate dehydrogenase;
HtrA1=high temperature requirement protease A1;
NFATc1=nuclear factor of activated T-cells cytoplasmic 1;
SPHK1=sphingosine kinase 1; TRAP=tartrate-resistant acid
phosphatase.
Western blotting
Proteins were extracted by lysing the osteoclasts with RIPA buffer (Aksomics,
Shanghai) and quantified by a bicinchoninic acid (BCA) assay kit (Beyotime,
Shanghai). Protein extracts (equal quantities) were separated using 10%
sodium dodecyl sulfate–polyacrylamide gel electrophoresis and were then
transferred onto a PVDF membrane (Whatman, UK). PVDF membranes were blocked with
no fat milk and incubated at 4°C overnight with primary antibodies
against GAPDH (Tianjin Sungene Biotech, China), Ephrin B2 (Abcam Inc., USA) and
SPHK1 (Cell Signaling Technology Inc., USA) at a dilution of 1:1000. After
washing 3 times, the PVDF membranes were incubated with HRP–conjugated
secondary antibodies (1:3000; Santa Cruz Biotechnology, USA) . Protein bands
were visualized using ECL (Millipore, USA)and the densitometry readings were
quantified using Image J software (NIH, USA).
Statistical analyses
Data are expressed as mean±SD and all calculations were carried out using
SPSS for Windows (ver. 22.0, IBM SPSS Software, US). A t-test was employed to
assess differences between parametric data, and non-parametric data was analyzed
using the Mann-Whitney test. P-value<0.05 was considered to represent
significant findings.
Results
Induction and identification of osteoclasts
Osteoclasts are mainly produced from hematopoietic precursor cells
(monocytes/macrophages lineage) in the bone marrow. RANKL triggers the
development and subsequent activation of osteoclasts in vitro and in
vivo
[22]. When primary osteoclast precursors
were cultured with both RANKL and M-CSF for 5–6 days, most of the
precursors fused together and became TRAP positive with>3 nuclei. These
TRAP-positive cells were much larger than the precursors, which were cultured
with M-CSF alone ([Fig. 1a] and [b]).
Fig. 1 Identification of the induced osteoclasts by TRAP
staining. a Primary mouse bone marrow mononuclear cells (BMMCs)
were obtained by flushing femurs and tibias of 6- to 8-week-old C57/BL-6
mice and cultured in DMEM containing 10% fetal calf serum (FCS)
in the presence of 50 ng/ml macrophage-colony stimulating factor
(M-CSF). b Primary mouse BMMCs were cultured in DMEM containing
10% FCS in the presence of 50 ng/ml M-CSF and
30 ng/ml receptor activator of nuclear factor-κB ligand
(RANKL) for 5–6 days. After fixation, the cells were stained for
tartrate-resistant acid phosphatase (TRAP) with a commercial kit.
c Murine RAW264.7 cells were cultured in DMEM containing
10% FCS. d Murine RAW264.7 cells were cultured in DMEM
containing 10% FCS in the presence of 30 ng/ml RANKL for
5–6 days. Then the cells were fixed and stained for TRAP. The
TRAP positive multinucleated (>3 nuclei) cells are identified as
mature osteoclasts. Scale bar=50 μm.
The RAW264.7 murine cell line is often used for in vitro studies of osteoclasts.
They do not need M-CSF for their RANKL-induced transformation to osteoclasts
[23]. Because of their ready access
and availability, we also tried to induce osteoclasts from RAW264.7 cells
cultured with RANKL ([Fig. 1c] and [d]). Due to the percentage of osteoclast
differentiation induced from RAW264.7 cells was lower than that from BMMCs, in
the following experiments we only used osteoclasts induced from primary
BMMCs.
Osteoclast bone resorption is detected by the activity of TRAP and associated
enzymes. In the late stage of osteoclastogenic signaling, it is characterized by
the amplification of NFATc1, whose function is to transcribe those genes that
control the multi-nucleation and bone resorption functions of osteoclasts [14]. To confirm further the characteristics
of osteoclasts, the expression levels of NFATc1 and TRAP mRNAs in BMMCs
pre-treated with both RANKL and M-CSF were demonstrated to be much higher than
inBMMCs pre-treated with M-CSF alone ([Fig.
2]).
Fig. 2 The mRNA expression levels of TRAP and nuclear factor of
activated T-cells cytoplasmic 1 (NFATc1) in induced osteoclasts. Primary
mouse BMMCs were cultured in the presence of 50 ng/ml M-CSF
alone (control) or in the presence of both 50 ng/ml M-CSF and
30 ng/ml RANKL (induced osteoclasts) for 5 days. Quantitative
real-time PCR expression analysis was performed using a FastStart
Universal SYBR Green Master with specific primers of TRAP a and
NFATc1 b. The constitutively expressed gene encoding
glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used for
normalization. Data are expressed as the mean±SD (n=4).
Compared with the control
cells, ***P<0.001. [Fig. 1] for other definitions.
Stimulatory effects of IL-23 on the expression of pro-osteogenic factors in
osteoclasts
Compared with unstimulated osteoclasts, IL-23 significantly increased mRNA
expression levels of Ephrin B2, SPHK1 and BMP2 ([Figs. 3a-c]). Peak levels of Ephrin B2 and
BMP2 occurred 3 h after stimulation with IL-23 and peak level of SPHK1 at 6 h.
In osteoclasts stimulated with IL-23, the maximal expression levels of mRNAs for
BMP2, Ephrin B2 and SPHK1 were increased by 1.46, 2.1 and 2.46 folds,
respectively. Increased protein levels of Ephrin B2 and SPHK1 were also
demonstrated by Western blot ([Fig. 4]).
The maximal levels of SPHK1 and Ephrin B2 proteins were increased by 1.56 and
1.65 folds, respectively. Because of the limited elevation of mRNA expression
levels of BMP2 after IL-23 stimulation, the protein level of BMP2 was not
detected in the present study.
Fig. 3 Modulatory effects of IL-23 on the mRNA expression levels
of pro-osteogenic and anti-osteogenic factors in osteoclasts.
Osteoclasts were induced from primary mouse BMMCs in the presence of
both 50 ng/ml M-CSF and 30 ng/ml RANKL. The osteoclasts
were refreshed with the medium without M-CSF or RANKL, and stimulated
with a final concentration of 30 ng/ml IL-23 for 0–48 h.
Quantitative real-time PCR expression analysis was performed using
specific primers of Ephrin B2 a, sphingosine kinase 1 (SPHK1,
b), bone morphogenetic protein 2 (BMP2, c), BMP6
d, Wnt10b e, Semaphorin 4D f and high
temperature requirement protease A1 (HtrA1, g). Data are
expressed as the mean±SD (n=4). Compared with the
osteoclasts without IL-23-stimulation,*P<0.05. [Fig. 1] for other definitions.
Fig. 4 Stimulatory effects of IL-23 on the protein expression
levels of Ephrin B2 and SPHK1 in osteoclasts. Osteoclasts were induced
from primary mouse BMMCs in the presence of both 50 ng/ml M-CSF
and 30 ng/ml RANKL. The osteoclasts were refreshed with the
medium without M-CSF or RANKL and stimulated with 30 ng/ml IL-23
for 0–48 h. Western blot analysis was performed with
specific antibodies against Ephrin B2 a, SPHK1 b and
GAPDH, and levels of Ephrin B2 c and SPHK1 d relative to
GAPDH were quantitated. The bands were subjected to densitometry
analysis with standardization to GAPDH. Results shown are representative
of 4 independent experiments. Compared with the initial phase (0 h after
stimulation),*P<0.05; **P<0.01.
[Fig. 1] for other
definitions.
No changes in the levels of mRNA expression for BMP6 or Wnt10b after IL-23
stimulation were found ([Figs. 3d,
e]).
Null effect of IL-23 on the expression of anti-osteogenic factors in
osteoclasts
When osteoclasts were stimulated with IL-23 for 3–24 h, the mRNA
expression levels of Semaphorin 4D and HtrA1 remained unchanged ([Fig. 3f, g]).
Discussion
The pathology of AS often involves the entheses i. e. the regions where
tendons, capsules and ligaments attach to bone, where involves inflammation, bone
erosion, and then syndesmophyte (spur) formation [2]. There is increasing evidence that suggests a pivotal role for IL-23
in the pathophysiology of AS [7]
[8]
[9]
[10]
[11]
[12]
[13]
[24]
[25]. However, little is known about how IL-23
regulates new bone formation (syndesmophyte formation). In the present study, for
the first time, we have demonstrated that IL-23 promotes the expression of
pro-osteogenic factors in osteoclasts including Ephrin B2, SPHK1 and BMP2. In
contrast, IL-23 had no detectable effects on the expression of Semaphorin 4D and
HtrA1, which are well recognized as anti-osteogenic factors.
The remodeling of bone is complex. When bone is resorbed by active osteoclasts, new
bone is formed by osteoblasts in the lacunae almost simultaneously. There is a
strict regulation of osteoclast and osteoblast coupling in both space and time [26]
[27]. During bone remodeling, osteoclasts and
osteoblasts communicate directly and indirectly in a process termed the
‘coupling mechanism’. Osteoclasts and osteoblasts communicate by
using diffusible (paracrine) signals, cell-to-cell contact and interactions with the
cellular bone matrix [21]
[28]
[29]. During the initial period of bone
remodeling, osteoblasts can express MCP-1 and osteoclastogenic factors including
RANKL and M-CSF, which control chemotaxis, differentiation and migration of
osteoclasts. It is noteworthy that cell signaling by RANK can activate several
downstream transcription factors of osteoclasts including AP-1 (c-Fos), NFATc1 and
NF-κB before they differentiate into mature osteoclasts and can resorb bone
[30]
[31]
[32]. Mature osteoclasts produce many
molecules that trigger differentiation of osteoblasts including Wnt10b, BMP2, BMP6,
and sphingosine-1-phosphate (produced by SPHK1), while HtrA1 is secreted during
osteoclastogenesis and functions to inhibit differentiation of osteoblasts [33]. Ephrin B2 is expressed on the cell
membrane of osteoclasts and interacts with EphB4, expressed on the membrane of
osteoblasts. Ephrin B2/EphB4 binding produces bidirectional signals that
promote differentiation of osteoblasts and inhibit osteoclastogenesis [34]
[35]. Osteoclasts also express a
membrane-bound factor called Semaphorin 4D that suppresses the formation of bone
during bone resorption [36].
BMPs are members of the TGFβ superfamily of growth signaling molecules. BMPs,
(particularly BMP-2), play vital roles in the generation of mature osteoblasts.
Another mechanism involved in osteoblastogenesis is the Wnt signaling pathway.
Canonical signaling either through Wnt10b or glycogen synthase kinase-3 (GSK3)
inhibition has been shown to promote osteoblastogenesis [37].
Eph receptors (EphRs) are tyrosine kinases that are activated by Ephrins and are
classified into EphBs (EphB1-B6) and EphAs (EphA1-A8), whose membrane-anchored
ligands are Ephrin Bs (Ephrin B1-B3) and EphrinAs (Ephrin A1-A6), respectively. The
bi-directional signaling between Ephrin B2 and EphB4 exhibit significant functions
in bone metabolism. Ephrin B2, which is expressed on the osteoclast cell surface
membrane, activates Ephrin-Eph bidirectional signaling in osteoblasts and elicits
differentiation of osteoblasts. The contra signal from EphB4 inhibits the formation
of osteoclasts. Forward signaling through EphB4 into osteoblasts also enhances
mineralization [35]
[38].
SPHK1 is the rate-limiting enzyme for the catalysis of sphingosine-1-phosphate. SPHK1
is upregulated in mature osteoclasts, but lower levels in precursor cells,
indicating that SPHK1-produced sphingosine-1-phosphate may be elevated and playing
an active role. Sphingosine-1-phosphate is an essential secreted coupling factor
[39]
[40]. In osteoblasts, sphingosine-1-phosphate
interacting with its receptors promotes osteoblast migration and survival. It also
promotes osteoblast differentiation by activating Wnt/β-catenin and
BMP2 signaling.
Semaphorin 4D expressed in osteoclasts binds to its receptor Plexin-B1 on
osteoblasts, and exerts inhibition of bone formation [36]. HtrA proteins are serine proteases of
which four distinct forms have been characterized, namely HtrA1, HtrA2, HtrA3 and
HtrA4. During osteoclastogenesis induced by RANKL, HtrA1 is markedly upregulated and
acts to limit osteoblast differentiation [33].
When HtrA1 is overexpressed it negatively regulates the mineralization of
osteoblasts.
IL-23 plays important roles in the development of experimental autoimmune disease
models and numerous afflictions affecting humans [24]. Genetic evidence links the IL-23/IL-17 pathway to
inflammation in these rheumatic diseases, and mechanistic data from mice support a
functional role for IL-23/IL-17 pathway activation in the development of
enthesitis and in entheseal bone formation [25]. Furthermore, analysis of human tissue samples, as well as data from
clinical trials, also supports a role for activation of the pathway in these
diseases. However, Two placebo-controlled, double-blind clinical trials in axial
spondyloarthritis (axSpA) of monoclonal antibodies directed against either the p40
protein or the p19 protein of the IL-23 molecule had no significant efficacy, but
whether the treatment has an effect on new bone formation is still unknown for the
slow radiographic progression unable to be observed in trial duration [41]. So, more studies are necessary to clarify
the possible mechanism of IL-23 in bone formation.
In the present study, IL-23 significantly upregulated protein and mRNA levels of
Ephrin B2, a membrane bound factor on the surface of osteoclasts. IL-23 also
significantly upregulated both mRNA and protein levels of SPHK1 in osteoclasts. Both
Ephrin B2 and SPHK1 have the potential to promote osteoblast differentiation,
migration, survival and even mineralization. Although Semaphorin 4D and HtrA1 exert
inhibitory effects on osteoblast differentiation, IL-23 had no detectable effects
on
HtrA1 and Semaphorin 4D expression. Thus IL-23 triggers osteoclasts to produce
pro-osteogenic factors, indicating that IL-23 might indirectly promote the
differentiation of osteoblasts via activated osteoclasts in AS.
Author contributions
Dan-Dan Pang: performance and data collection of the experiments; Li Cai: manuscript
drafting; Jing-Ru Zhang: statistical analysis, data collection and interpretation
of
data; Sheng-Ming Dai: conceive and design of the study, and revision of the
manuscript.
