The interplay between fibroblast-like synovial and vascular endothelial cells leads to angiogenesis via the sphingosine-1-phosphate-induced RhoA-F-Actin and Ras-Erk1/2 pathways and the intervention of geniposide

Ran Deng1,2,3,4 | Yanhong Bu1,2,3,4 | Feng Li1,2,3,4 | Hong Wu1,2,3,4 | Yan Wang1,2,3,4 | Wei Wei5

1Key Laboratory of Xin’an Medicine, Ministry of Education, Hefei, China
2College of Pharmacy, Anhui University of Chinese Medicine, Hefei, China
3Anhui Province Key Laboratory of Chinese Medicinal Formula, Anhui University of Chinese Medicine, Hefei, China
4Anhui Province Key Laboratory of Research & Development of Chinese Medicine, Anhui University of Chinese Medicine, Hefei, China
5Anhui Medical University, Key Laboratory of Antiinflammatory and Immune Medicine (Anhui Medical University), Ministry of Education, Institute of Clinical Pharmacology, Antiinflammatory Immune Drugs Collaborative Innovation Center, Hefei, China

Correspondence
Hong Wu, College of Pharmacy, Anhui University of Chinese Medicine, Qian Jiang Road 1, Hefei 230012, China.
Email: [email protected]
Wei Wei, Institute of Clinical Pharmacology, Anhui Medical University, Key Laboratory of Antiinflammatory and Immune Medicine (Anhui Medical University), Ministry of Education, Meishan Road 81, Hefei, 230032, China.
Email: [email protected]

Funding information
National Natural Science Foundation of China, Grant/Award Numbers: 81073122, 81473400,
81874360

1 | INTRODUCTION

Rheumatoid arthritis (RA) is a chronic systemic autoimmune disease characterized by joint synovial inflammation and angiogenesis with

unknown pathogenic factors (Taams, 2020; Leblond et al., 2020). RA induced by any cause is often accompanied by an imbalance of inflam- mation microenvironment of the synovial membrane (Saeki & Imai, 2020). Recent studies on RA patients and animals have shown

that inflammation and angiogenesis are the causes of pannus forma-

†Ran Deng and Yanhong Bu contribute equally to the article.

tion. Inherent cells (fibroblast-like synoviocytes (FLSs), vascular

endothelial cells (VECs), macrophages, lymphocytes, and neutrophils), and adaptive immune cells (T/B lymphocytes) constitute RA microen- vironment (Souto-Carneiro et al., 2020; Buckley, Ospelt, Gay, & Midwood, 2021). Under pathological conditions, cells accumulate in the articular cavity and interact to form a unique inflammatory microenvironment, which is characterized by hypoxia, accumulation of pro-inflammatory and pro-angiogenic factors. Cells are exposed to adjacent cells and their secreted cytoactive substances. These cell–cell interactions aggravate the pathogenesis of RA.

As the main participants in RA inflammation microenvironment, the changes of biological functions of FLSs and VECs are closely related to the imbalance of synovial microenvironment. FLSs show abnormal proliferation and inadequate apoptosis, and release a large number of inflammatory cytokines in RA, which make hypertrophy of synovial tissue and hypoxia (Zhai et al., 2018, 2019). FLSs hyper- proliferation and angiogenesis of VECs synergistically promote char- acteristic pannus formation. In addition, the synovial membrane is infiltrated by proliferating FLSs and cytokines, so that angiogenesis continues to occur (Sun et al., 2020). FLSs have a strong secretion function, which secretes a variety of inflammatory factors such as
TNF-α, IL-6 to change the microenvironment of the articular cavity.
VECs are infiltrated by inflammatory factors, and enhance their prolif- eration and migration ability to adapt to the neogenesis of pannus. Angiogenesis provides oxygen and nutrients for the highly prolifera- tive synovial tissue to maintain the tumor like growth of FLSs, and provides a way for inflammatory cells to enter the cartilage and bone tissue structure. Angiogenesis and inflammatory proliferation of FLSs form pannus cartilage combination and erode articular cartilage and bone tissue. Therefore, synovitis hyperplasia and angiogenesis inter- weave and promote each other. Therefore, synovial hyperplasia and angiogenesis interweave and promote each other. Studies have shown that the degree of synovial angiogenesis in RA joint is posi- tively correlated with the patient’s condition, synovial hyperplasia, and inflammatory cell infiltration. Drugs that inhibit angiogenesis alle- viate the condition of RA (Lu et al., 2020).
One of the metabolites of cell membrane sphingomyelin is
sphingosine-1-phosphate (S1P). Under the strict control of sphingo- sine kinase 1 (SphK1), S1P remained low in the cell. Tissue injury or stress can induce a large number of typical intracellular signal mole- cules S1P. In turn, S1P can act as a second intracellular messenger, which can be transported to the outside of the cell in an autocrine or paracrine manner, and act as an extracellular signal molecule by bind- ing with S1PRs to regulate cell proliferation, migration, and differenti- ation (Pedowitz, Batt, Darabedian, & Pratt, 2021; Cartier & Hla, 2019). S1P induces angiogenesis by regulating the pathological processes of different cell types. Excessive proliferation of FLSs and VECs is an important feature in the pathogenesis of RA, leading to synovial hyperplasia and angiogenesis (Chen, Yang, Zhi, Yao, & Liu, 2021). Recent studies have provided a rough understanding of S1P signal as a regulator of proliferation, and the increased level of S1P in RA indi- cates that regulating the level of S1P and understanding the signal cascade of S1P in the development of RA may become a therapeutic target of RA (Zhao et al., 2019).

Geniposide (GE) is an iridoid glycoside compound obtained from the dried mature fruits of Gardenia jasminoides Ellis, which has antiinflammatory and anti-angiogenic effects (Xuejing et al., 2019). Our group has done a lot of work on the antiinflammatory activity of GE. GE regulated Ras–Mek-Erk1/2 pathway to inhibit the prolifera- tion of lymphocyte in mesenteric lymph nodes (MLNLs), restore the balance of pro/antiinflammatory factors, and improve the inflamma- tory and pathological changes of synovial tissue (Wang et al., 2018,
2017). GE blocked the activation of p38 MAPK-NF-κB-F-actin by
RhoA, which reduced the permeability of FLSs, reduced the entry of inflammatory cytokines into cells, and relieved joint inflammation (Deng et al., 2018). These studies indicated that the antiinflammatory immune activity of GE is closely related to its regulation of MAPKs pathways.
This paper aims to elucidate how the crosstalk between FLSs and VECs induces angiogenesis through in vivo and in vitro methods, and this characteristic pannus of RA has been considered as the key feature of the development of this malignant disease. Previous studies have confirmed that GE has certain anti-angiogenesis effect. Therefore, we further explore the specific mechanism of GE’s further inhibition of angiogenesis, that is, whether GE regulates the interaction between VECs and FLSs by inhibiting SphK1 and p-Erk1/2, so as to provide new targets and strategies for anti-angiogenesis therapy of RA.
2 | MATERIAL AND METHODS

2.1 | Reagents

GE was purchased from Chengdu croma Biotechnology, and its purity was more than 95% detected by HPLC. It can directly use ultrapure water to prepare high concentration GE solution, which can be diluted during administration. SCH772984 was obtained from apexbio com- pany with batch number of a3805. PF-543 and FTY-720 purchased from Selleck chemicals. Tripterygium glycosides (TG) purchased from the Affiliated Hospital of Anhui University of Chinese medicine.
2.2 | Modeling and administration

Male SD rats weighing 180–220 g, which were collected from the experimental animal center of Anhui University of Chinese medicine were used in this experiment. All animal experimental procedures were approved by animal protection and use Committee of Anhui University of Chinese medicine. The rats were randomly divided into 6 groups (6 rats in each group): control, model, GE (30, 60, 120 mg/ kg), TG (10 mg/kg), and labeled as Con, Mod, GE (30 mg/kg), GE (60 mg/kg), GE (120 mg/kg), and TG (10 mg/kg) in Figure 1. AA model was induced by subcutaneous injection of 0.1 ml complete Freund’s adjuvant (sigma) into the right foot of rats. After 15 days of modeling, the drugs were given according to the groups for 7 days.
In the experiment of “TNF-α activates S1P secretion in VECs and
intervention of GE”, HUVEC were set as control, TNF-α, GE

 

FIG U R E 1 Angiogenesis of synovium in AA rats and intervention of geniposide (GE). (a) Effects of GE on synovial pathomorphology of AA rats (n = 8), A: Conl; B: Mod; C: GE (30 mg/kg); D: GE (60 mg/kg); E: GE (120 mg/kg); F:TG (10 mg/kg) (n = 6). (b) Effect of GE on expression of CD31, SphK1, and p-Erk1/2 in synovium of AA rats (mean ± sd, n = 6).(c) Effect of GE on expressions of SphK1, p-SphK1, Erk1/2, p-Erk1/2, and Ras in synovium of AA rats (mean ± sd, n = 3) [Colour figure can be viewed at wileyonlinelibrary.com]
(25,50,100 μM), PF-543 (0.1 μM) and SCH772984 (1 μM) groups, and
labeled as Con, TNF-α, TNF-α+GE (25 μM), TNF-α+GE (50 μM), TNF-
α+GE (100 μM), TNF-α+PF-543 (0.1 μM) and TNF-α+SCH772984
(1 μM) in Figure 2. In addition to the control group, the other groups were treated with TNF-α for 24 hr.
In the part of “Effect of crosstalk between FLSs and VECs on the biological function of VECs”, HUVEC were set as control, model, GE (25, 50, 100 μM) and FTY-720 (1 μM) groups, and labeled as Con, CM, CM+GE (25 μM), CM+GE (50 μM) CM+GE (100 μM) and CM+FTY-
720 (1 μM) in Figure 3. DMEM solution containing 10% FBS was
added to the control group, while the conditioned medium was added to the other groups for 24 hr.
In the experiment of “Effect of conditioned medium on the bio-
logical function of FLSs”, FLSs were set as control, model, GE (25,50,100 μM), FTY-720 (1 μM) and SCH772984 (1 μM) groups, and labeled as Con, CM, CM+GE (25 μM), CM+GE (50 μM), CM+GE (100 μM), CM+FTY-720 (1 μM) and CM+SCH772984 (1 μM) in
Figure 4.

2.3 | Histology and immunohistochemistry

Synovial membranes were fixed with 4% paraformaldehyde for 24 hr, dehydrated with different concentrations of ethanol, transparent with
xylene, embedded in paraffin, and sectioned longitudinally (4 μm).
Serial paraffin sections were stained with hematoxylin eosin (H&E) to analyze the synovial lesions. Immunohistochemical sections were
incubated overnight with CD31, SphK1, or p-Erk1/2 antibody at 4◦C,
and then incubated with secondary goat anti-mouse IgG for 30 min the next day. The positive nuclei were brown.
2.4 | Cell cultures

HUVEC line was provided by Saiqi Bioengineering (Shanghai). The culture of HUVEC was cultivated in a medium containing fetal bovine serum (FBS, serana, Germany)-DMEM (HyClone, Logan) con- taining 5% Penicillin and Streptomycin (Beyotime Biotechnology,

 

FIG U R E 2 TNF-α activated S1P secretion in vascular endothelial cells (VECs) and intervention of geniposide (GE). (a) CCK-8 was used to detect the proliferation of HUVEC treated by different concentrations of TNF-α (mean ± sd, n = 6).(b) Effects of different concentrations of TNF- α on the secretion of S1P in HUVEC (mean ± sd, n = 3).(c) Effects of TNF-α on intracellular (A) and extracellular (B) levels for S1P of HUVEC (mean ± sd, n = 3).(d) Effect of GE on TNF-α induced the cytoplasm (A) and membrane (B) of p-SphK1 in HUVEC (mean ± sd, n = 3). (e) Effects of GE on the expression and activation of key protein in S1P pathway induced by TNF-α in HUVEC. (mean ± sd, n = 3). (f) Effects of GE on the expression and activation of key gene in S1P pathway induced by TNF-α in HUVEC. (mean ± sd, n = 3). (g) Proliferation of HUVEC treated by different concentrations of GE (mean ± sd, n= 6). (h) Effect of TNF-α on co-expression of p-Erk1/2 and SphK1 in HUVEC and the intervention of GE. (mean ± sd, n = 3). (i) Effect of GE on TNF-α induced the co-expression of p-Erk1/2 and SphK1 detected by co-IP in HUVEC. (mean ± sd,
n = 3) ##p < .01 versus control group; *p < .05, **p < .01 versus model group [Colour figure can be viewed at wileyonlinelibrary.com]

Shanghai) in a volume ratio of 10:90. When the cell density of HUVEC reached 80–90%, it could be subcultured in 1:4 or 1:5 ratio. It was observed that cells were polygonal and distributed as paving stones.

FLSs line was provided by Ginio Biological Company (Guang Zhou). The culture conditions were the same as HUVEC. When FLSs reached 80% of the bottom, it could be packed into sterile culture flasks at 1:2 or 1:3.

 

FIG U R E 3 Effect of crosstalk between FLSs and vascular endothelial cells (VECs) on the biological function of VECs. (a) Effect of geniposide (GE) on Proliferation of HUVEC treated by conditioned medium. The enhanced effect of conditioned medium on the proliferation of HUVEC was inhibited after GE administration. (mean ± sd, n = 6). (b) Effect of conditioned medium on scratch of HUVEC and the intervention of GE (×200) a: Con-CM 0 hr; b: Con-CM 24 hr; c:Mod-CM 0 hr; d: Mod-CM 24 hr; e: GE (25 μM)-CM 0 hr; f: GE (25 μM)-CM 24 hr; g: GE (50 μM)-CM 0 hr; h: GE (50 μM)-CM 24 hr; i: GE (100 μM))- CM 0 hr; j: GE (100 μM)-CM 24 hr; k: FTY-720 (1 μM)-CM 0 hr; j: FTY-720 (1 μM)-CM 24 hr. (c) Effect of
conditioned medium on secretion of angiogenesis-related factors in HUVEC (mean ± sd, n = 4). (d) Effects of GE on the distribution and morphology of F-actin in HUVEC induced by conditioned medium (×400) A: Con-CM; B: Mod-CM; C: GE (25 μM)-CM; D: GE (50 μM)-CM; E: GE (100 μM)-CM; F: FTY-720 (1 μM)-CM. (e) Effect of conditioned medium on the expression of key proteins in the S1P-S1PR1 and downstream RhoA-NF-κB-F-actin pathway of HUVEC and intervention of GE. (f) Effects of conditioned medium on the expression of S1PR1, RhoA, NF-κB p- 65 and F-actin genes in HUVEC and intervention of GE (mean ± sd, n = 3) ##p < .01 versus control group; *p < .05, **p < .01 versus model group [Colour figure can be viewed at wileyonlinelibrary.com]

2.5 | Cocultures system and conditioned medium studies

A six hole Transwell system (costar, Corning) with microporous mem- brane pore diameter of 0.4 μM was used to simulate the coculture system of FLSs and VECs. This kind of membrane prevented the
migration and contacted between them, but it allowed the soluble fac- tors secreted to diffuse between them, which simulate the crosstalk between cells. HUVEC with a density of 5 × 105 was inoculated at

the bottom of the lower chamber, and FLSs with a density of
2.5 × 105 was inoculated at the upper chamber transmembrane. Cells in both groups were cultured separately overnight and then attached to the wall. After the culture medium was changed, the insert con- taining FLSs was transferred to the hole containing HUVEC seeded at
the bottom. After 48 hr of culture, the cocultured conditioned medium was collected and centrifuged at 1500 rpm for 5 min, and fil- tered (0.22 μm) for reserve. The quality of conditioned medium needs
to be evaluated before it is used. The transparency, color change,

FIG U RE 4 Effect of crosstalk between fibroblast-like synoviocytes (FLSs) and vascular endothelial cells (VECs) on the biological function of FLSs. (a) Detection of S1P content in conditional medium (mean ± sd,
n = 4). (b) Effects of conditioned medium with different groups on cell proliferation of FLSs. (mean ± sd, n = 6). (c) Effects of conditioned
medium on production of IL-1β, IL-6,
TGF-β1, and IL-10 of FLSs (mean
± sd, n = 4). (d) Effects of conditioned medium on Ras-Erk1/2 pathway in FLSs (mean ± sd, n = 3).
(e) Effects of conditioned medium on the mRNA expression of S1PR1, Ras and Erk1/2 in FLSs. #p < .05 versus, ##p < .01 versus Control,
**p < .01 versus Model precipitation and impurities, sterility test, and pH value test of the solution should be examined.
2.6 | Cell viability assay

Cell count kit-8 (CCK-8) was used to test cell viability as previously described. HUVEC or FLSs was harvested into cell suspensions at a density of 5 × 10/ml in a 96-well plate at 100 μl per well, and was
stable overnight for adherence. Conditioned medium were added to the cells. According to the experimental summary in the literature, the action time was set to 24 hr. The absorbance (OD) of each hole was read on the enzyme label at 490 nm wavelength.
Cell proliferation was calculated by the following formula: Cell viability ð%Þ ¼ OD treatment pore=OD control pore × 100% ð1Þ
The OD treatment pore indicated the cell absorbance value after
adding the drug, and the OD control pore indicated the cell absor- bance value of the control.

2.7 | Enzyme linked immunosorbent (ELISA) assay

HUVEC or FLSs with a cell density of 1 × 10/ml was inoculated into 6-well plate and added 2 ml into each hole. After adherence, condi-
tioned medium were added for 24 hr, GE (25, 50, 100 μM) and PF- 543 (0.1 μM) were added for 48 hr. Then 5% FBS-DMEM was further cultured for 48 hr, centrifuged (2000 rpm, 20 min), the supernatant
was taken, dispensed, and stored at —80◦C for detection. According to the manufacturer’s instructions, the secretory levels of S1P, IL-1β, IL-6, IL-10 and TGF-β1, Ang-1, FGF and as were measured with the
corresponding ELISA Kit (ebioscience, San Diego, California). The absorbance of the sample fixed at 450 nm was measured in a micro tablet reader (Biotech, synergy HT). The content was calculated by standard curve.
2.8 | Cell migration assay

HUVEC was prepared at a cell suspension of 2 × 105 cells/ml in a vol- ume of 2 ml per well in a six-well plate. After continuing to culture for
24 hr, 200 μl tip was compared to the steel ruler and was as perpen- dicular as possible to the back scratch. After the scratches were com- pleted, cells were washed, and then the medium was added, and
photographed under a fluorescence microscope. The time was recorded as 0 hr. Then the cells were placed in an incubator for another 24 hr, and taken out again for photographing, and the time was recorded as 24 hr.
2.9 | Laser confocal assay

HUVEC or FLSs was adjusted to a cell density of 1 × 10/ml. The staining procedure was as follows: 2 ml of cell suspension was accu-
rately aspirated into a 6-well plate, and a sterilized slide was added to the bottom in advance. HUVEC was fixed with 4% paraformaldehyde for 10 min at room temperature; after washing, 0.5% Triton X-100 was permeabilized for 5 min. The FITC-Phalloidin working solution at a concentration of 150 nM was prepared in advance, and the stock solution was diluted with 1% BSA in PBS buffer. The preparation and
staining should be strictly protected from light. Each hole was incu- bated for 30 min with 200 μl FITC-Phalloidin working solution. The film was cleaned and sealed. The fluorescence was observed and
photographed by confocal laser microscopy.
2.10 | Western blotting

Western blotting was applied to detect the levels of S1PR1, RhoA, F- actin, NF-κB p65, p-NF-κB p65, Erk1/2, p-Erk1/2, Ras, p-SphK1, and SphK1. When HUVEC (FLSs) reached the 60–70% sub-fusion state,
then added conditioned medium and were pretreated for 24 hr. HUVEC or FLSs was washed three times with ice-cold PBS, 400 μl of Ripa lysate was dissolved, and then centrifuged (12,000 rpm, 30 min).
The supernatant was collected and the protein concentration was determined by BCA assay. The same amount of each sample was sep- arated using 12% SDS-PAGE and transferred to a nitrocellulose mem- brane (Poll, New York) using a wet transfer device. All membranes were blocked with 5% (w/v) skim milk in TBST for 2 hr at room tem- perature to prevent non-specific binding. The previously prepared pri-
mary antibody was incubated overnight at 4◦C. The membrane was
washed and then incubated with goat anti-rabbit IgG horseradish peroxidase-conjugated secondary antibody for 1 hr. The protein bands were scanned by the alpha-view-sa system (California) and ana- lyzed by image software.
2.11 | Real-time PCR

RT-qPCR was applied to detect the mRNA levels of S1PR1, RhoA, F- actin, NF-κB, Erk1/2, and Ras. Briefly, HUVEC or FLSs was pretreated with conditioned medium for 24 hr, total RNA was extracted using
TRIzol reagent (Invitrogen, Carlsbad, CA), quantified, and reverse tran- scribed into cDNA. Relative quantification of expression of the

selected genes was performed in a LightCycler 480 system (Roche, Pleasanton, CA), RT-PCR was performed using SYBR to measure mRNA levels, and then mRNA levels were calculated using the 2-ΔΔCt
method. The cycling conditions were 95◦C for 30 s, followed by 40 cycles of 95◦C for 20 s, 55◦C, 72◦C for 30 s respectively. Dissocia- tion curve was generated in a cycle of 95◦C for 15 s, 60◦C for 60 s, and back to 95◦C for 15 s.
2.12 | Co-immunoprecipitation (Co-IP) assay

HUVEC was scraped off after washing with PBS, then centrifuged (800 rpm, 3 min), and the supernatant was discarded to retain the pre-
cipitation. The supernatant was collected by adding ×1 pyrolysis
buffer (volume ratio PMSF: RIPA = 99:1) and mixing in whirlpool. The supernatant was treated by ultrasound for 30 s at 1 min interval, repeated process for three times, centrifugation (12,000 rpm, 5 min). The new centrifuge tube was incubated overnight by adding 0.7 ml
cell lysate and 1 μg purified SphK1 antibody at 4◦C. The lysed cell
lysate was transferred to a column containing washed beads and incu- bated for 2 hr at room temperature. The column was inserted into the new centrifugal tube for centrifugation (12,000 rpm, 30 s), and the supernatant was discarded. The beads coupled with protein were
cleaned with 1 ×IP buffer and 0.1 ×IP buffer and centrifuged
(12,000 rpm, 30 s). The column was inserted into a new centrifugal tube (12,000 rpm, 30 s) after 5 min of heating at 95◦C with 1 ×sample
buffer, and the eluted immunoprecipitate was stored at —80◦C for
reserve. Western blot was used to detect p-Erk1/2 and SphK1 pro- teins in immunoprecipitates.
2.13 | Statistical analysis

SPSS 23.0 statistical software was used for statistical analysis. The experimental data were expressed as mean ± SD, and one-way ANOVA was used to compare data between groups. p < .05 was con- sidered as a significant difference, which was statistically significant.
3 | RESULTS

3.1 | Angiogenesis of synovium in AA rats and intervention of GE

In order to observe the angiogenesis in AA rats, we obtained the syn- ovium of their joints for experiment. First, we evaluated the model from three aspects of foot swelling, arthritis index, and movement score. The results of pathological section showed that cells in the synovial membrane of normal rats were evenly arranged, with smooth surface and monolayer growth. While in model group, the synovial tis- sue was edema and hypertrophy, cells in the synovial tissue were abnormal proliferation, multi-layer distribution, and disordered arrangement, and a large number of inflammatory cells infiltration and
fibroblast proliferation were seen in the synovial stroma. Immunohis- tochemistry and Western blot results confirmed that the expression of SphK1 and p-Erk 1/2 in synovium increased with the aggravation of pathological symptoms, indicating that SphK1 and p-Erk1 1/2 are closely related to the angiogenesis of synovium. Previous experiments have confirmed that the content of S1P in the synovial microdialysis fluid of AA rats is significantly increased, and GE reduced the secre- tion of S1P and improve the inflammatory response of AA rats (Xuejing et al., 2019). After treatment with GE and TG for 1 week, the proliferation of synovial cells decreased and arranged orderly, but a small amount of inflammatory cells infiltrated. It is suggested that after GE treatment, synovial hyperplasia and angiogenesis can be sig- nificantly improved. Meanwhile, GE and TG reduced the protein expressions of SphK1 and p-Erk 1/2, suggesting that the therapeutic effect of GE is related to it (Figure 1a–c). In addition, Western blot results showed that the expression of Ras and p-Erk 1/2 increased in the synovium of AA rats, indicating that Erk 1/2 signaling pathway was activated.
3.2 | TNF-α activated S1P secretion in VECs and intervention of GE

The key enzyme of S1P synthesis is SphKs, in which SphK1 promotes cell survival. It phosphorylates ceramide to produce S1P, which is an important regulatory molecule of sphingolipid metabolism. Based on the research status of VECs in inflammatory microenvironment and its potential significance in the occurrence and development of RA, this study intends to establish an inflammatory injury model of HUVEC
induced by TNF-α and detect the change of S1P content in HUVEC.
TNF-α is one of the earliest and indispensable inflammatory mediators that disappears in RA synovitis. It induces FLSs proliferation and fur-
ther releases inflammatory factors, activates VECs to release angio- genic factors (Yang, Liu, Fang, & He, 2021a; Yang, Wang, Zhou et al., 2021b; Wang et al., 2021). The abnormal high expression of
TNF-α in synovial fluid of RA patients and the specific receptor
of TNF-α on the surface of VECs make TNF-α easy to induce synovial angiogenesis and inflammatory damage (Pérez et al., 2019). TNF-α, as a stimulant, can induce the inflammatory loss of endothelial cells,
which is an ideal model reagent for the research of RA. Therefore, we chose TNF-α to induce HUVEC. The results showed that 10 ng/ml TNF-α treated for 24 hr up-regulated the proliferation of HUVEC.
Meanwhile, the secretion of S1P increased significantly, and the lon- ger the treatment time was, the more significant the rising trend was (Figure 2a,b). It has been found that S1P is an important intermediate
of TNF-α mediated synovitis. Next, we used ELISA assay to detect the
effect of TNF-α on the secretory function of S1P inside and outside HUVEC’s cell membrane. As shown in Figure 2c, compared with con- trol group, the intracellular and extracellular S1P content in TNF-α group increased, indicating that inflammatory factors in the inflamma-
tory microenvironment activated VECs and promoted the secretion of S1P. Western blotting showed that the expression of p-SphK1 in the cytoplasm and membrane of TNF-α group was significantly higher

than that of control group, suggesting that TNF-α induced the transfer of p-SphK1 in the cytoplasm to the cell membrane and promoted the production of S1P (Figure 2d). Meanwhile, the protein expressions of
p-Erk1/2, p-SphK1, and SphK1 were also increased in TNF-α group, suggesting that p-Erk1/2-SphK1-S1P signaling pathway was activated in HUVEC, which promoted the release of S1P (Figure 2e). RT-PCR
confirmed that the mRNA expressions of Erk1/2 and SphK1 increased (Figure 2f).
In order to study the effect of GE on S1P secretion in inflamma- tory state of HUVEC and its possible mechanism, we conducted
follow-up experiments. First of all, CCK-8 results showed that GE had no obvious damage effect on HUVEC in the range of 12.5–100 μM.
When GE concentration was 200 μM, the cell viability was signifi-
cantly lower than that of control group, indicating that this condition had a toxic effect on cells, so the concentration of GE used in subse-
quent experiments was 25, 50, and 100 μM (Figure 2g). After adminis-
tration of GE (25, 50, 100 μM), the S1P content inside and outside the cells decreased in a concentration-dependent manner (Figure 2c). Meanwhile, the inhibitors SCH772984 (1 μM) and PF-543 (0.1 μM) also reduced the secretion of S1P. Similarly, the expression of p-
SphK1 in the cytoplasm and membrane was significantly reduced (Figure 2d). Western-bolt confirmed that the key proteins p-Erk1/2, p-SphK1, and SphK1 of the S1P pathway were all reduced after the
administration of GE (25, 50, 100 μM), SCH772984 (1 μM), and PF-
543 (0.1 μM) (Figure 2e), the mRNA expression of Erk1/2 and SphK1 also decreased (Figure 2f). These results suggested that GE’s regula-
tion of S1P secreted by HUVEC is related to the inhibition of p- Erk1/2-SphK1-S1P pathway.
In order to further clarify the mechanism of GE in angiogenesis, HUVEC was treated with different concentrations of GE. The level of p-Erk1/2 and the co-expression of SphK1 were analyzed by immuno- coprecipitation and laser confocal method. HUVEC was treated with
GE, PF-543, and SCH772984 inhibitors, which significantly blocked the increase of p-Erk1/2 induced by TNF-α and downregulated the expression of SphK1. The expression of p-SphK1 in HUVEC was ana-
lyzed by Western blot. Different concentrations of GE (25, 50, 100 μM), PF-543 (0.1 μM) and SCH772984 (1 μM) treatment signifi- cantly reduced the expression of p-SphK1 in the cytoplasm and mem- brane, suggesting that GE inhibited TNF-α induced angiogenesis by downregulating p-Erk1/2 and SphK1. p-Erk1/2 and SphK1 are poten-
tial targets of GE(Figure 2h,i).
3.3 | Effect of crosstalk between FLSs and VECs on the biological function of VECs

It has been reported that S1P promoted the abnormal proliferation of VECs by regulating RhoA-F-actin-NF-κB pathway, which suggested that S1P played an important role in RA induced angiogenesis. The
abnormal proliferation of VECs under inflammatory conditions leads to angiogenesis, but in RA, the effect of FLSs on VECs and its material basis are still unclear. In order to study the effects of cytokines released by FLSs and the crosstalk between the two cells on the
biological functions of VECs, VECs were placed in the conditioned medium. The biological function of VECs was determined by measur- ing the cell viability, migration ability, the morphological distribution of cytoskeleton protein F-actin, key pathway protein, and gene expression.
Our previous studies have confirmed that FLSs released a large
number of inflammatory factors to the culture medium in vitro, in which the contents of TNF-α, IL-1β, IL-6, IL-8, and PGE2 were increased, which indicated that there were a large number of inflam-
matory factors in conditioned medium (Wang et al., 2021). In our sub- sequent experiments, the cell viability, migration, and secretion ability of angiogenic factors of HUVEC were enhanced after being induced by conditioned medium, as shown in Figure 3a–c.
In order to further study whether the effect of conditioned medium on the biological functions of VECs were through the S1P pathway, we used the S1PR1 inhibitor FTY-720 for follow-up study. The results showed that FTY-720 inhibited the migration and tube formation of HUVEC induced by conditioned medium, and the prolif- eration of HUVEC also decreased. Meanwhile, we studied GE’s inter- vention effect on the biological function of VECs induced by conditioned medium. The results showed that GE inhibited the prolif- eration, migration ability of HUVEC induced by conditioned medium.
We also detected the key proteins and genes in the RhoA-F- actin-NF-κB pathway by laser confocal, Western blotting, and RT- qPCR. The changes of cell morphology are closely related to the cyto-
skeleton. Therefore, we further observed the effect of conditioned medium on F-actin of HUVEC cytoskeleton and the intervention of GE. F-actin in cytoskeleton of Con-CM group was mainly distributed in the dense periphery zone under the cell membrane, and a large number of F-actin was distributed in the cytoplasm. Compared with Con-CM group, the dense periphery zone under the cell membrane began to disappear, the distribution of F-actin in the cytoplasm
increased and a large number of stress fibers were formed in Mod- CM group. GE (25, 50, 100 μM) and FTY-720 (1 μM) reduced the con- tent of F-actin in the cytoplasm to varying degrees, and the volume of
stress fibers decreased to varying degrees, as its showed in Figure 3d. The above results showed that conditioned medium induced the reor- ganization of cytoskeleton, which was manifested by the change of the distribution from perimembranous to cytoplasmic and the forma- tion of stress fibers, thus increasing the permeability of HUVEC cell membranes and making it easier for harmful substances to enter cells. The results showed that GE alleviated the effect of conditioned
medium on the permeability of HUVEC. Western blotting and RT- qPCR showed that S1PR1-RhoA-F-actin-NF-κB pathway was acti- vated in HUVEC induced by conditioned medium, and this effect was
inhibited by FTY-720 and GE, as shown in Figure 3e,f.
Taken together, these results showed that the interaction between FLSs and VECs is necessary to create an inflammatory micro- environment, acts on VECs, and increases their biological functions such as proliferation, migration, and cell permeability. At the same time, the results suggested that S1P participates in S1PR1-RhoA-F-
actin-NF-κB-mediated pathway of angiogenesis, and GE inhibited the
induction.

3.4 | Effect of crosstalk between FLSs and VECs on the biological function of FLSs

It has been confirmed that S1P is involved in the induction of FLSs migration, the secretion promotion of inflammatory cytokines, and the inhibition of apoptosis response. S1P is elevated in synovial fluid of RA patients, which seems to promote the invasion and migra- tion of FLSs. Taking these observations into account, we measured the amount of S1P released from conditioned medium using an ELISA kit. The results showed that the secretion of S1P increased signifi- cantly (Figure 4a).
In order to further clarify the effect of S1P in the conditioned medium on the biological function of FLSs, we confirmed it by cell activity and ELISA assay. Conditioned medium promoted the prolifer-
ation of FLSs. It enhanced the release of cytokines (IL-1β and IL-6),
reduced the secretion of antiinflammatory factors (TGF-β1 and IL-10),
broke the dynamic balance between pro/antiinflammatory factors, and accelerated the development of RA inflammation, the results are shown in Figure 4b,c.
In order to further prove that the biological function of FLSs is caused by S1P, we used a p-Erk1/2 inhibitor (SCH772984, 1 μM) and
FTY-720 (1 μM) for subsequent experiments. The results showed that
in the presence of SCH772984 or inhibitor FTY-720, the proliferation and unbalanced secretory function of FLSs were inhibited, and GE inhibited the induction (Figure 4d,e). These results provided evidence for S1P to regulate the biological function of FLSs by activating Ras- Erk1/2 pathway. S1P mediated changes in FLSs biological function play a key role in activation of S1P dependent Ras-Erk1/2 pathway.
In addition, our data emphasized the importance of crosstalk between FLSs and VECs, and the important role of S1P released by VECs in regulating this process.
4 | DISCUSSION

Angiogenesis is an important pathological feature of RA, its formation often involves many aspects, but it is always inseparable from the increase of proliferation and migration ability of VECs. Angiogenesis runs through the whole course of RA, which is the fundamental rea- son for the occurrence and development of RA (Kim et al., 2020; Han et al., 2020). In RA, angiogenesis related cytokines include S1P, VEGF, Ang-1, FGF, among which S1P is the main cytokine involved in syno- vial neovascularization. As an extracellular stimulator, S1P binds with SIPR1 on the cell membrane, triggers intracellular signal response, and leads to VECs proliferation, migration, and tube formation (Yao, Xie, & Zeng, 2020; Chen et al., 2020). Studies have confirmed that S1P is abnormally expressed in RA patients. The concentration of S1P in synovial fluid of RA patients is significantly higher than that in plasma and serum. This is consistent with our results. The results of ELISA showed that the content of S1P secreted by VECs increased signifi-
cantly after TNF-α stimulation.
The key enzymes for S1P synthesis are SphKs, including SphK1 and SphK2. In autoimmune diseases, the body produces a largenumber of inflammatory factors such as TNF-α, IL-1β, which can be activated as Erk1/2 weak activator to produce p-Erk1/2, and then activates SphK1 to increase its activity by 14 times (Ranasinghe,Lee, & Schwarz, 2020). At the same time, it changes the conformation of SphK1 protein, promotes the transfer of p-SphK1 from cytoplasm to cell membrane, and then catalyzes the phosphorylation of Sph to produce S1P (Song, Dai, Long, Wang, & Di, 2020). Our hypothesis is further confirmed that FLSs secretes pro-inflammatory factors such as

TNF-α, PGE2, IL-1, IL-6 and IL-17, which activated as weak activator
to activate p-Erk1/2 and increase SphK1 activity, thereby increasing the secretion of S1P by HUVEC. Western blot and RT-PCR results confirmed that the expressions of p-Erk1/2 and SphK1 in VECs
increased significantly after TNF-α stimulation.
GE is an iridoid glycoside obtained, which has anti-angiogenesis and antiinflammation effects. Research has confirmed GE has strong
anti-angiogenic activity in a concentration range of 25–100 μM in a
dose-dependent manner (Sun et al., 2020). GE inhibited the produc- tion of IL-6 and IL-8 in HUVEC by blocking p38MAPK and Erk1/2 pathways (Liu et al., 2010). In vivo experiments showed that the dose of GE induced obvious liver injury in rats was 300 mg/kg, which was converted into the human dose of 60 kg, which was about 16.48 mg/ kg, while the clinical recommended therapeutic dose of GE was about 10 mg/kg, which is calculated based on the content of GE in Gardenia jasminoides Ellis of 1.8–6% (Tian, 2017). In this paper, the dose of GE for rats was 120 mg/kg for 7 day, so GE has no obvious hepatotoxic- ity at this dose.
Whether GE improves S1P secretion disorder in VECs by regulat- ing p-Erk1/2 and SphK1 pathways remains unclear. ELISA results showed that the production and secretion of S1P decreased signifi- cantly after treated with different concentrations of GE. PF-543 and SCH772984 also reduced the secretion of S1P. PF-543 is a novel SphK1 inhibitor for permeable cells, and its selectivity to SphK1 is more than 100 times higher than that of SphK2. Western blot showed that the expression of p-Erk1/2, SphK1, and p-SphK1 in cytoplasmic membrane of HUVEC decreased after GE treatment. It is suggested that the expression of p-Erk1/2, SphK1, and p-SphK1 and the trans- membrane of p-SphK1 are related to the production and secretion of S1P. SCH772984, a p-Erk1/2 inhibitor, has the same effect as PF- 5453, which indicated that p-Erk1/2 is the upstream of SphK1-S1P pathway in HUVEC. GE has similar effects with SCH772984, PF-543. Laser confocal and immunoprecipitation results showed that GE, SCH772984, and PF-543 downregulated the co-expression of p- Erk1/2 and SphK1. It is suggested that GE may inhibit the phosphory- lation and activation of SphK1 by reducing the interplay between p- Erk1/2 and SphK1 in HUVEC and further reduce the membrane trans- location of p-SphK1, which may be one of the molecular mechanisms by which GE downregulates the production and secretion of S1P in VECs. GE also downregulated the expression of p-Erk1/2 and SphK1 in vivo, which improved the angiogenesis of synovial membrane.
As the main cells that make up RA synovial tissue, VECs, FLSs are
important for the angiogenesis. It has been confirmed that the abnor- mal proliferation of FLSs releases a large number of inflammatory fac- tors to highly penetrate VECs. VECs are activated in the inflammatory

microenvironment, which significantly promotes cell proliferation and migration to form new blood vessels. The proliferation activity of FLSs is due to the two-way interaction with VECs, accompanied by the release of soluble inflammatory mediators and signal transduction molecules, which are actively involved in RA process and joint dys- function (Chen et al., 2017; Kim & Kim, 2016). Our data showed that when exposed to conditioned medium, the biological functions of HUVEC and FLSs changed significantly. These changes are accompa- nied by the activation of S1PR1 and its downstream RhoA-F-actin-
NF-κB and Ras-Erk1/2 signaling pathways. As shown in Figure 4a, the
secretion of S1P in conditioned medium increased significantly. Therefore, we speculated that the changes of biological functions of HUVEC and FLSs are closely related to the increase of S1P in inflam- matory microenvironment. When we used FTY-720, an inhibitor of S1PRs, the induction of HUVEC or FLSs by conditioned medium was weakened. It has been confirmed that the expression of S1P in RA patients is abnormal, and the concentration of S1P in synovium fluid of RA patients is significantly higher than that of plasma and serum. Pathological results showed that there was only a small amount of S1PR1 expression in normal synovial tissues, while the expression of S1PR1 was increased in synovial tissues of RA patients, and S1PR1 was more abundant in the lining cells of synovium (Choi et al., 2018). S1P can be secreted to the extracellular space by autocrine or para- crine way, which binds to the S1PRs on the cell membrane and acti- vates the classical GPCR signaling pathway, thus regulating cell survival, proliferation, and migration (Yang, Liu, Fang, & He, 2021a;
Yang, Wang, Zhou et al., 2021b). The “inside-out” relocation of S1P
activates a variety of signaling pathways, including Ras, Erk, RhoA to regulate cell proliferation, inflammatory cytokines, and cytoskeletal reconstruction (Hou et al., 2021; Abdel Rahman et al., 2021). Impor- tantly, during tissue damage, early tissue damage response signal mol- ecules (such as S1P) induce further secretion of pro-inflammatory
cytokines (such as IL-1β, IL-6 and TNF-α).
After exposure to conditioned medium, the quantity and distribu- tion of F-actin in HUVEC changed, the stress fibers increased and F- actin decreased. Together with other proteins, these proteins provide strong adhesion between VECs and control the activation of related pathways. The changes of the distribution and levels of these proteins in cell membrane often lead to the changes of cell permeability (Dasgupta, Le, Vijayan, & Thiagarajan, 2017). Further studies con-
firmed that RhoA-F-actin-NF-κB pathway in VECs were activated,
which inhibited the depolymerization of F-actin, increased the forma- tion of F-actin, promoted the production of stress fibers, thus increased the permeability of vascular endothelial cells, and subse- quently caused a series of changes in biological functions. VECs are cells with the function of proliferation and differentiation. A large number of studies have shown that it is not only involved in the for- mation of neovascularization, but also involved in cell permeability, such as the control of some substances. RhoA, a small protein family, is an upstream regulator of F-actin cytoskeleton depolymerization or polymerization, which is involved in the regulation of neovascularization and cell permeability (Li, 2020; Wu, 2020). The physiological level of S1P activates S1PRs and protects theendothelial barrier function by inducing the activation of Rac1 path- way, while excessive S1P and S1PRs combine to activate RhoA and destroy the endothelial barrier function. Therefore, the balance between RhoA and Rac1 is destroyed, which leads to the impairment of endothelial function, promotes the process of vascular inflamma- tion, and participates in the occurrence of RA (Sun, Chen, & Zhao, 2018). Previous research confirmed that the secretions of IFN-γ, IL-1, IL-6, and IL-17 were up-regulated in FLSs, and the secretionsof IL-4, IL-10, and TGF-1β were downregulated (Chen et al., 2015;

Zhang et al., 2017). The increased permeability of VECs makes it eas- ier for a large number of inflammatory factors in the joint microenvi- ronment to enter the cells, and then play a role in promoting the further proliferation of VECs, eventually leading to angiogenesis.
When FLSs was exposed to conditioned medium, its proliferation and secretion of proinflammatory factors increased abnormally, which was accompanied by the activation of Ras-Erk1/2 pathway. The cyto- kines associated with angiogenesis in RA include VEGF, S1P, Ang-1, and FGF. In the hypothetical release of inflammatory mediators, we focus on S1P, because this cytokine is usually present in the inflam- matory microenvironment, and is one of the most disordered

cytokines, which is closely related to angiogenesis in RA. In synovial tissues of RA patients, S1P -S1PR1 pathway promotes the prolifera- tion of FLSs, resulting in the increased expression of s IL-1β, IL-6, IL-8,
PGE2 (Wang et al., 2020). It leads to the proliferation and migration of VECs and changes in cell permeability. Exogenous S1P also induces FLSs migration, promotes the secretion of inflammatory cytokines, and inhibits apoptosis (Wang et al., 2021). Our results confirmed the ability of HUVEC to release S1P, which can be detected in cell culture supernatant. However, the content of S1P in supernatant of FLSs is very low, which means that VECs is the main source of S1P in inflam- matory microenvironment.
In conclusion, we proved that the crosstalk between FLSs and VECs is the key event to trigger the sharp increase of S1P release, which regulates the biological functions of two types of cells by acti- vating RhoA-F-actin and Ras-Erk1/2 pathways, as shown in Figure 5. Therefore, the regulation of S1P release is expected to be a promising strategy in the treatment of RA. However, GE inhibited the prolifera- tion of FLSs, restored the dynamic balance of pro/antiinflammatory factors, reduced the proliferation, migration, and secretion abilities of VECs, and inhibited the formation of new blood vessels, which are

FIG U R E 5 Pathways of Interaction between fibroblast-like synoviocytes (FLSs) and vascular endothelial cells (VECs) in Inflammatory Microenvironment of rheumatoid arthritis (RA). FLSs release a large number of inflammatory factors into RA synovial microenvironment. Inflammatory factors act on and activate VECs. Meanwhille, inflammatory factors act as weak activators to increase SphK1 activity by activating Erk1/2 pathway, VECs secrete S1P into the microenvironment, and activate RhoA-F-actin and Ras-Erk1/2 pathways in FLSs and VECs, resulting in changes in the biological functions of them, which ultimately lead to the aggravation of synovial inflammation and the formation of new blood vessels [Colour figure can be viewed at wileyonlinelibrary.com]
key factors for the occurrence and development of RA. Meanwhile, we confirmed that GE inhibited the angiogenesis of RA effectively, and its mechanism is to reduce the activity of SphK1 by inhibiting the activation of p-Erk1/2, thus reducing the secretion of S1P. Its targets may be p-Erk1/2 and SphK1, which are expected to become new drugs for the effective treatment of RA.

ACKNOWLEDGMENTS
This work was supported by grants from the National Natural Science Foundation of China (No 81073122, 81473400, and 81874360).

CONFLICT OF INTEREST
The authors declare that they have no conflict of interests.

AUTHOR’S CONTRIBUTIONS
Participated in research design: Ran Deng, Yanhong Bu, Professor Hong Wu, and Professor Wei Wei. Conducted experiments:Ran Deng, Yanhong Bu, Yan Wang. Contributed new reagents or analytic tools: Ran Deng, Feng Li, Yan Wang. Performed data analysis: Ran Deng, Yanhong Bu, Feng Li. Wrote or contributed to the writing of the man- uscript: Ran Deng, Yanhong Bu.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the authors upon reasonable request.
REFERENCES

Abdel Rahman, F., d’Almeida, S., Zhang, T., Asadi, M., Bozoglu, T., Bongiovanni, D., … Ziegler, T. (2021). Sphingosine-1-phosphate atten- uates lipopolysaccharide-induced Pericyte loss via activation of rho-a and MRTF-A. Thrombosis and Haemostasis, 121(3), 341–350.
Buckley, C. D., Ospelt, C., Gay, S., & Midwood, K. S. (2021). Location, loca- tion, location: How the tissue microenvironment affects inflammation in RA. Nature Reviews Rheumatology, 17(4), 195–212.
Cartier, A., & Hla, T. (2019). Sphingosine 1-phosphate: Lipid signaling in pathology and therapy. Science, 366(6463), eaar5551.
Chen, C. Y., Su, C. M., Hsu, C. J., Huang, C. C., Wang, S. W., Liu, S. C., …
Tang, C. H. (2017). CCN1 promotes VEGF production in osteoblasts and induces endothelial progenitor cell angiogenesis by inhibiting miR- 126 expression in rheumatoid arthritis. Journal of Bone and Mineral Research, 32(1), 34–45.
Chen, H. J., Yang, H. R., Zhi, Y., Yao, Q. Q., & Liu, B. (2021). Evaluation of pyrrolidine-based analog of jaspine B as potential SphK1 inhibitors against rheumatoid arthritis. Bioorganic & Medicinal Chemistry Letters, 34, 127754.
Chen, J. Y., Wu, H., Li, H., Hu, S. L., Dai, M. M., & Chen, J. (2015). Anti-
inflammatory effects and pharmacokinetics study of geniposide on rats with adjuvant arthritis. International Immunopharmacology, 24(1), 102–109.
Chen, Q., Rehman, J., Chan, M., Fu, P., Dudek, S. M., Natarajan, V., … Liu, Y. (2020). Angiocrine sphingosine-1-phosphate activation of S1PR2-YAP signaling axis in alveolar type II cells is essential for lung repair. Cell Reports, 31(13), 107828.

Choi, H. S., Kim, K. H., Jin, S., Kim, J., Yoo, I., Pack, S. P., … Jung, Y. W. (2018). Decreased expression of sphingosine-1-phosphate receptor 1 in the blood leukocyte of rheumatoid arthritis patients. Immune Net- work, 18(5), e39.
Dasgupta, S. K., Le, A., Vijayan, K. V., & Thiagarajan, P. (2017). Dasatinib inhibits actin fiber reorganization and promotes endothelial cell permeability through RhoA-ROCK pathway. Cancer Medicine, 6(4), 809–818.
Deng, R., Li, F., Wu, H., Wang, W. Y., Dai, L., Zhang, Z. R., & Fu, J. (2018). Anti-
inflammatory mechanism of geniposide: Inhibiting the hyperpermeability of fibroblast-like synoviocytes via the RhoA/p38MAPK/NF-κB/F-actin signal pathway. Frontiers in Pharmacology, 9, 105.
Han, C. C., Liu, Q., Zhang, Y., Li, Y. F., Cui, D. Q., Luo, T. T., … Wei, W.
(2020). CP-25 inhibits PGE2-induced angiogenesis by down-regulating EP4/AC/cAMP/PKA-mediated GRK2 translocation. Clinical Science, 134(3), 331–347.
Hou, L., Zhang, Z., Yang, L., Chang, N., Zhao, X., Zhou, X., … Li, L. (2021). NLRP3 inflammasome priming and activation in cholestatic liver injury via the sphingosine 1-phosphate/S1P receptor 2/Gα(12/13)/MAPK
signaling pathway. Journal of Molecular Medicine (Berlin, Germany), 99
(2), 273–288.
Kim, J. W., Kong, J. S., Lee, S., Yoo, S. A., Koh, J. H., Jin, J., & Kim, W. U.
(2020). Angiogenic cytokines can reflect the synovitis severity and treatment response to biologics in rheumatoid arthritis. Experimental & Molecular Medicine, 52(5), 843–853.
Kim, K. W., & Kim, H. R. (2016). Macrophage migration inhibitory factor: A potential therapeutic target for rheumatoid arthritis. The Korean Jour- nal of Internal Medicine, 31(4), 634–642.
Leblond, A., Pezet, S., Cauvet, A., Casas, C., Pires Da Silva, J., Hervé, R., … Avouac, J. (2020). Implication of the deacetylase sirtuin-1 on synovial angiogenesis and persistence of experimental arthritis. Annals of the Rheumatic Diseases, 79(7), 891–900.
Liu, H. T., He, J. L., Li, W. M., Yang, Z., Wang, Y. X., Yin, J., … Yu, C. (2010).
Geniposide inhibits interleukin-6 and interleukin-8 production in lipopolysaccharide-induced human umbilical vein endothelial cells by blocking p38 and ERK1/2 signaling pathways. Inflammation Research, 59(6), 451–461.
Li, X., Cheng, Y., Wang, Z., Zhou, J., Jia, Y., He, X., … Wang, J. (2020). Cal- cium and TRPV4 promote metastasis by regulating cytoskeleton through the RhoA/ROCK1 pathway in endometrial cancer. Cell death & disease, 11(11), 1009.
Lu, K., Iwenofu, O. H., Mitra, R., Mo, X., Dasgupta, P. S., & Basu, S. (2020). Chebulinic acid is a safe and effective antiangiogenic agent in collagen-induced arthritis in mice. Arthritis Research & Therapy, 22 (1), 273.
Pedowitz, N. J., Batt, A. R., Darabedian, N., & Pratt, M. R. (2021). MYPT1 O-GlcNAc modification regulates sphingosine-1-phosphate mediated contraction. Nature Chemical Biology, 17(2), 169–177.
Pérez, L., Vallejos, A., Echeverria, C., Varela, D., Cabello-Verrugio, C., & Simon, F. (2019). OxHDL controls LOX-1 expression and plasma mem- brane localization through a mechanism dependent on NOX/ROS/NF-
κB pathway on endothelial cells. Laboratory Investigation: A Journal of
Technical Methods and Pathology, 99(3), 421–437.
Ranasinghe, A., Lee, D. D., & Schwarz, M. A. (2020). Mechanistic regulation of SPHK1 expression and translocation by EMAP II in pulmonary smooth muscle cells. Biochimica et Biophysica Acta. Molecular and Cell Biology of Lipids, 1865(12), 158789.
Saeki, N., & Imai, Y. (2020). Reprogramming of synovial macrophage metabolism by synovial fibroblasts under inflammatory conditions. Cell Communication and Signaling, 18, 188.
Song, K., Dai, L., Long, X., Wang, W., & Di, W. (2020). Follicle-stimulating hormone promotes the proliferation of epithelial ovarian cancer cells by activating sphingosine kinase. Scientific Reports, 10(1), 13834.
Souto-Carneiro, M. M., Klika, K. D., Abreu, M. T., Meyer, A. P., Saffrich, R., Sandhoff, R., … Carvalho, R. A. (2020). Effect of increased lactate dehydrogenase a activity and aerobic glycolysis on the proinflammatory profile of autoimmune CD8+ T cells in rheumatoid arthritis. Arthritis Rheumatology, 72(12), 2050–2064.

Sun, M., Deng, R., Wang, Y., Wu, H., Zhang, Z., Bu, Y., & Zhang, H. (2020). Sphingosine kinase 1/sphingosine 1-phosphate/sphingosine 1-phosphate receptor 1 pathway: A novel target of geniposide to inhibit angiogenesis. Life Sciences, 256, 117988.
Sun, X. J., Chen, M., & Zhao, M. H. (2018). Rho GTPases are involved in S1P-enhanced glomerular endothelial cells activation with anti- myeloperoxidase antibody positive IgG. Journal of Cellular and Molecu- lar Medicine, 22(9), 4550–4554.
Taams, L. S. (2020). Interleukin-17 in rheumatoid arthritis: Trials and tribu- lations. Journal of Experimental Medicine, 217(3), e20192048.
Tian, J, Zhu, J, & Yi, Y. (2017). Dose-related liver injury of Geniposide asso- ciated with the alteration in bile acid synthesis and transportation. Sci. Rep, 7, 8938.
Wang, M., Wu, H., Wang, R., Dai, X., Deng, R., Wang, Y., … Zhang, H. (2021). Inhibition of sphingosine 1-phosphate (S1P) receptor 1/2/3 ameliorates biological dysfunction in rheumatoid arthritis fibroblast- like synoviocyte MH7A cells through Gαi/Gαs rebalancing. Clinical and
Experimental Pharmacology and Physiology, 48(8), 1080–1089.
Wang, R., Wu, H., Chen, J., Li, S. P., Dai, L., Zhang, Z. R., & Wang, W. Y. (2017).
Antiinflammation effects and mechanisms study of geniposide on rats with collagen-induced arthritis. Phytotherapy Research, 31(4), 631–637.
Wang, R. H., Dai, X. J., Wu, H., Wang, M. D., Deng, R., Wang, Y., … Zhang, H. (2020). Anti-inflammatory effect of geniposide on regulating the functions of rheumatoid arthritis synovial fibroblasts via inhibiting sphingosine-1-phosphate receptors1/3 coupling Gαi/Gαs conversion.
Frontiers in Pharmacology, 11, 584176.
Wang, Y., Dai, L., Wu, H., Zhang, Z. R., Wang, W. Y., Fu, J., … Zhan, X. (2018). Novel anti-inflammatory target of geniposide: Inhibiting Itgβ1/Ras-Erk1/2 signal pathway via the miRNA-124a in rheumatoid
arthritis synovial fibroblasts. International Immunopharmacology, 65,
284–294.
Wu, J., Yang, J., Yu, M., Sun, W., Han, Y., Lu, X., …, Cai, Y. (2020). Lanthanum chloride causes blood-brain barrier disruption through intracellular calcium-mediated RhoA/Rho kinase signaling and myosin light chain kinase. Metallomics : integrated biometal science, 12(12), 2075–2083.
Xuejing, D., Wenyu, W., Hong, W., Zhengrong, Z., Li, D., Jun, F., … Xiang, Z. (2019). UHPLC-MS/MS analysis of sphingosine 1-phosphate in joint cavity dialysate and hemodialysis solution of adjuvant arthritis rats: Application to geniposide pharmacodynamic study. Biomedical Chromatography, 33(7), e4526.

Yang, L., Liu, R., Fang, Y., & He, J. (2021a). Anti-inflammatory effect of phenylpropanoids from dendropanax dentiger in TNF-α-induced MH7A cells via inhibition of NF-κB, Akt and JNK signaling pathways. International Immunopharmacology, 94, 107463.
Yang, T., Wang, X., Zhou, Y., Yu, Q., Heng, C., Yang, H., … Zhang, L. (2021b). SEW2871 attenuates ANIT-induced hepatotoxicity by protecting liver barrier function via sphingosine 1-phosphate receptor- 1-mediated AMPK signaling pathway. Cell Biology and Toxicology, 10.
Yao, X., Xie, L., & Zeng, Y. (2020). MiR-9 promotes angiogenesis via targeting on Sphingosine-1- phosphate receptor 1. Frontiers in Cell and Developmental Biology, 8, 755.
Zhai, K. F., Duan, H., Cui, C. Y., Cao, Y. Y., Si, J. L., Yang, H. J., … Wei, Z. J.
(2019). Liquiritin from glycyrrhiza uralensis attenuating rheumatoid arthritis via reducing inflammation, suppressing angiogenesis, and inhibiting MAPK signaling pathway. Journal of Agricultural and Food Chemistry, 67(10), 2856–2864.
Zhai, K. F., Duan, H., Khan, G. J., Xu, H., Han, F. K., Cao, W. G., … Wei, Z. J.
(2018). Salicin from alangium chinense ameliorates rheumatoid arthri- tis by modulating the Nrf2-HO-1-ROS pathways. Journal of Agricultural and Food Chemistry, 66(24), 6073–6082.
Zhang, Z. R., Wu, H., Wang, R., Li, S. P., Dai, L., & Wang, W. Y. (2017).
Immune tolerance effect in mesenteric lymph node lymphocytes of geniposide on adjuvant arthritis rats. Phytotherapy Research, 31(8), 1249–1256.
Zhao, C., Amiable, N., Laflamme, M., Marsolais, D., Di Battista, J. A., Fernandes, M. J., & Bourgoin, S. G. (2019). Impairment of chemical hypoxia-induced sphingosine kinase-1 expression and activation in rheumatoid arthritis PF-543 synovial fibroblasts: A signature of exhaustion? Biochemical Pharmacology, 165, 249–262.