Kinsenoside screening with a microfluidic chip attenuates gouty arthritis through inactivating NF-κB signaling in macrophages and protecting endothelial cells

Gouty arthritis is a rheumatic disease that is characterized by the deposition of monosodium urate (MSU) in synovial joints cause by the increased serum hyperuricemia. This study used a three-dimensional (3D) flowing microfluidic chip to screen the effective candidate against MSU-stimulated human umbilical vein endothelial cell (HUVEC) damage, and found kinsenoside (Kin) to be the leading active component of Anoectochilus roxburghi, one of the Chinese medicinal plant widely used in the treatment of gouty arthritis clinically. Cell viability and apoptosis of HUVECs were evaluated, indicating that direct Kin stimulation and conditioned medium (CM) from Kin-treated macrophages both negatively modulated with MSU crystals. Additionally, Kin was capable of attenuating MSU-induced activation of nuclear factor-κB/mitogen-activated protein kinase (NF-κB/MAPK) signaling, targeting IκB kinase-α (IKKα) and IKKβ kinases of macrophages and influencing the expressions of NF-κB downstream cytokines and subsequent HUVEC bioactivity. Inflammasome NLR pyrin domain-containing 3 (NALP3) and toll-like receptor 2 (TLR2) were also inhibited after Kin treatment. Also, Kin downregulated CD14-mediated MSU crystals uptake in macrophages. In vivo study with MSU-injected ankle joints further revealed the significant suppression of inflammatory infiltration and endothelia impairment coupled with alleviation of ankle swelling and nociceptive response via Kin treatments. Taken together, these data implicated that Kin was the most effective candidate from Anoectochilus roxburghi to treat gouty arthritis clinically.

Gouty arthritis is characterized by increased hyperuricemia level, which is triggered by disorders of the purine catabolic pathway, and leads to the deposition of monosodium urate (MSU) crystals within synovial joints and tissues. 1 This process of an acute inflammatory attack with intense pain, swelling and skin reddening is highlighted by the influx of neutrophils into articular or periarticular tissues and the release of inflammatory cytokines after the phagocytosis of deposited MSU microcrystals and membranolysis by monocyte-macrophages, 2 which further magnifies the inflammation cascade. The resultant stimulation of inflammatory cytokines is a cardiovascular risk factor, 3 which contributes to the significant damage against human umbilical vein endothelial cells (HUVECs) via an increased apoptosis rate 4 and thrombus formation. 5 Colchicine and non-steroidal antiinflammatory drugs (NSAIDs) (e.g., indomethacin, Indo), which are first-line treatments that inhibit this process clinically, are frequently used against acute gouty arthritis. Nonetheless, the inevitable side effects involving gastrointestinal bleeding, gastrointestinal toxicity and renal toxicity of such pharmacological reagents 6 restrict their further utilization. 6 Therefore, active extracts from natural herbal plants are gradually catching the attention of researchers because of the distinct merits of extensive sources, desirable efficacy and reduced adverse effects. 7 Anoectochilus roxburghii, a widely used remedy for gouty arthritis in Traditional Chinese Medicine clinically, is supposed to show extensive pharmacologic effects against liver disease, 8 hyperglycemia, 9 hyperlipidemia 10 and cancers. 11,12 So far, researchers have unraveled five major categories that constitute the active component of A. roxburghii, which are flavonoids, phenols, steroids, organic acids and glycosides. 13 Despite the widely application in antigout treatment clinically, the primary functioning element of A. roxburghii and the underlying mechanisms still remain elusive. Furthermore, recent pharmaceutical studies of gouty arthritis primarily focused on two-dimensional static testing in vitro and ignored the real three-dimensional circulated circumstances that endothelial cells underwent in vivo. These researches highlighted the direct influence of certain drugs on inflammation or endothelial cells and neglected the intimate crosslink between inflammation and vascular complaints. 2,5,6,14,15 Therefore, a microfluidic device, in which functionalized cells are grown on a three-dimensional flowing system instead of two-dimensional static culturing dishes, emerged to identify the major nutraceutical of Anoectochilus roxburghii. This device exhibits the following advantages: high throughput, sufficient flexibility 16 and accurate modeling of an in vivo-like coculture context of cell-cell and cell-ECM in vitro, 17 providing a promising toolkit for effective drug screening, especially for traditional medical plants.
In summary, the present study identified the major antigout candidate isolated from the widely used A. roxburghii under three-dimensional (3D) microfluidic conditions. The underlying protective mechanisms of this agent on MSU-induced endothelial cell injury should be investigated.

Results
Determination of drug concentrations in chips. Figure 1A shows that the entire microfluidic chip (gradient culture chip, GCC) primarily contained medium inputs, parallel microfluidic channels (concentration generator channel, CGCs) and cellculturing chambers (CCCs) to evaluate cell viability in threedimensional flowing conditions under the stimulations of five representative components derived from A. roxburghii: gastrodin (Gas), kaempferol (Kae), stigmasterol (Sti), ursolic acid (Ua), kinsenoside (Kin). Figure 1Ba illustrates that the design of the GCC enabled a desirable mixture of media from two separate inputs with a gradient color background from red to blue in the CGCs. Crystal violet and PBS were used to evaluate disparate concentrations in the five CCC/outlets ( Figure 1Bb). Figures 1B and C shows that the color shades gradually change from purple to transparent, and outlets 1-4 accounted for~0%, 25%, 50% and 75% of the transformed gray scale in outlet 5, respectively, which indicates that the concentrations in outlets 1-4 also reached 0%, 25%, 50% and 75%, respectively, of the drug concentration in outlet 5. The complete view of microfluidic chip was shown in Figure 1C. Therefore, we pumped medium containing 100 μM of the five candidate extracts from Input 2 and nodrug medium from Input 1, which contributed to the 0, 25, 50, 75 and 100 μM in outlets 1-5, respectively.
Microfluidic screening of effective candidate from A. roxburghii. High concentrations of MSU crystals (150, 225 and 300 μg/ml) were able to induce significant dosedependent damage against HUVECs (Figures 2A and C). Therefore, 300 μg/ml of MSU was further deployed to investigate the drugs protective efficacy against MSU crystals. Figures 2Ba, b  Kin protects HUVECs from MSU crystals. To determine the effects of Kin on MSU-stimulated HUVECs, further experiments were performed in 96-well plates. Kin did not exhibit cytotoxicity against HUVECs at concentrations from 6.25 μg/ml (25 μM) to 25 μg/ml (100 μM) (Figure 3Ba). When HUVECs were exposed to 300 μg/ml of MSU, cell viability was inhibited significantly (Figure 3Bb). Nonetheless, despite that Indo (20 μg/ml) failed to enhance HUVEC proliferation after MSU intervention, Kin exhibited protective effects on MSU-stimulated HUVECs with increased viability compared with MSU treatment alone. Flow cytometry ( Figure 3C) further demonstrated that Kin treatments (6.25-25 μg/ml) exerted beneficial effects on MSU-augmented cell apoptosis compared with the Indo group 14.46 ± 0.62%, as exemplified by the 10.67 ± 0.66%, 9.52 ± 0.64% and 4.61 ± 1.12% of apoptosis rates from 6.25, 12.5 to 25 μg/ml in the Kin groups, compared with 16.32 ± 0.46% in the MSU group and 2.21 ± 0.13% in the control group. These results were also confirmed using western blotting ( Figure 3D), with the expressions of Bax, caspase-3, cleaved caspase-3 (proapoptotic proteins) increased and Bcl-2 (antiapoptotic protein) decreased in a dose-dependent manner in the MSU group, signifying that the enhancement on apoptosis was achieved via MSU stimulation. Besides, various treatments of Kin decreased the expressions of Bax and cleaved caspase-3 and increased the level of Bcl-2, implicating the potential antiapoptosis effects of Kin on MSU-induced HUVECs.
Kin prevents the impairment of HUVECs from MSU-CM of macrophage cells. Figure 4a reveals that conditioned medium (CM) from merely MSU-treated macrophage cells was able to inhibit the growth of HUVECs. Nonetheless, when RAW264.7 cells were administered with various concentrations of drugs (Indo 20; Kin 6.25-25 μg/ml) and MSU, the CMs collected were considered to enhance cell proliferative abilities of HUVECs at 12.5 and 25 μg/ml of Kin even though low concentration of Kin. Indo failed to demonstrate protective properties. Annexin V/PI double staining revealed that high concentrations of Kin (12.5 and 25 μg/ml) decreased MSU-CM-induced apoptosis of HUVECs. However, a low concentration of Kin (6.25 μg/ml) and Indo (20 μg/ml) exhibited antiapoptotic capabilities despite the negligible protective effects on cell growth (Figure 4b) (MSU: 17.82 ± 0.96%; Indo: 12.55 ± 0.57%; 6.25 μg/ml Kin: 13.63 ± 0.52%; 12.5 μg/ml Kin: 10.18 ± 1.03%; 25 μg/ml Kin: 4.86 ± 0.65%). Additionally, the expressions of Bax, cleaved caspase-3 and caspase-3 decreased and Bcl-2 increased compared with MSU-CMtreated group when HUVECs were added with both Kin-and MSU-treated CM from RAW264.7 cells, which suggests that the CM from inflammatory macrophages had a vital role in exhibiting protective properties against HUVECs (Figure 4c). Thus, we proposed that the efficacious regulative mechanisms of Kin might be ascribed to the adjusted CM from macrophages and thereby affecting HUVECs and the subsequent evolution of gouty arthritis.
Kin inhibits the MSU-induced activation of NF-κB and JNK/Erk. Nuclear factor-κB (NF-κB) luciferase reporter gene activity assay (Figure 5a) indicated that NF-κB transcription increased significantly following MSU stimulation. Kin administrations decreased NF-κB activation dose-dependently. Moreover, treatment with Kin barely affected the MSUstimulated upstream TRAF-6, p-TAK-1 and p-IKK proteins expressions ( Figure 5b) while decreasing the expressions of p-IκBα plus p-p65 and increasing the production of inhibitor of NF-κB (IκBα) (Figure 5c), implying that Kin probably targeted IKK-α and IKK-β kinases, thus inhibiting downstream activation of NF-κB. These results were further confirmed using a molecular docking assay, which demonstrated that Kin bound to ASP151, CYS84 and TYR83 in IKK-α and ILE151, ARG92, TYR84 and CYS85 in IKK-β (Figure 5d), providing potential inhibition sites that Kin functioned against the NF-κB pathway. Similarly, administration of MSU increased the expression of p-JNK (c-Jun N-terminal kinase), p-Erk (extracellular signalregulated kinase) and p-p38 in macrophages. Nonetheless, Kin was able to attenuate the expression of MSU-induced activation of p-JNK and p-Erk dose-dependently irrespective of p-p38 ( Figure 5e).
Inhibition of NF-κB downstream inflammatory cytokine expressions and inflammasomes. As shown in Figure 6a, the mRNA expressions of NF-κB-related inflammatory cytokine genes (IL-1β, IL-6, TNF-α, iNO and COX-2) were significantly promoted with MSU stimulation. Nonetheless, the inductions of these genes were repressed markedly by the addition of Kin in a dose-dependent manner. Enzymelinked immunosorbent assay (ELISA) (Figure 6b) illustrated that expressions of IL-1β, IL-6, TNF-α, NO and prostaglandin E 2 (PGE 2 ) in CM from macrophage cells were enhanced owing to MSU stimulation, with a dose-dependent decrease in MSU-induced inflammatory cytokines from macrophages after Kin treatment.
The effects on inflammasome of macrophages after Kin administration were also evaluated. We found that Kin treatments inhibited the NLR pyrin domain-containing 3 (NALP3) expression in MSU-stimulated macrophages. Also, TLR2 expression was significantly enhanced after MSU administration and various Kin treatments attenuated MSUstimulated TLR2 expressions dose-dependently ( Figure 6c).    Figure 8Ab shows that without MSU stimulation, we observed no obvious macrophagocytes migration into vascular ring muscle, as well as the relatively thin interlayer of vessel. Nonetheless, a large number of neutrophils migrated into the vascular ring muscle in the synovial membrane of MSU-treated rats, which contributed to the thicker interlayer of blood vessels. In contrast, no obvious evidence of inflammatory leukocyte penetration in Indo-treated gouty arthritis rats, which exhibited thinner vessel walls, was observed. Furthermore, Kin administration decreased the thickness of the vessel ring muscle dose-dependently with diminished inflammatory infiltration into the medial layer of the vessel wall of neutrophils. Particularly, the endothelial-protective effect of high-dose Kin (10 mg/kg) was also superior to Indo treatment, which indicates its potential as a promising pharmaceutical reagent against gouty arthritis. Additionally, as shown in Figure 8Ba, b, the gastric gland arrangement and renal structure remained intact after Kin treatments.
Attenuation of MSU-induced ankle swelling, tail-flick response and writhing reaction. The ankle perimeter of urate-induced right hindlimbs increased significantly within 3 days compared with control rats. Indo (5 mg/kg) plus 10 mg/ kg Kin administration markedly reduced the MSU-induced ankle swelling compared with the model group, but 2.5 mg/kg Kin barely exerted an antiarthritic effect ( Figure 8C). Figure 8Da shows that the Indo group and the 10 mg/kg Kin group significantly increased the incubation period of the tail-flick response compared with the control group and increased the pain threshold 42.08 ± 14.44% and 28.02 ± 4.43%, respectively, irrespective of 2.5 and 5 mg/kg of Kin-treated rats. Indo and Kin treatments (2.5, 5 and 10 mg/kg) significantly decreased the cumulative number of acetic acid-induced writhing responses. The analgesic rates were 28.8 ± 5.6%, 18.5 ± 4.4%, 35.3 ± 4.7% and 48.5 ± 2.9%, respectively, relative to control rats (Figure 8Db).

Discussion
In this study, Kin was used to treat MSU-stimulated HUVECs based on a flowing three-dimensional microfluidic system that was deployed to screen the leading active candidate from five representative elements of Anoectochilus roxburghii. Kin has been proven to suppress LPS-stimulated inflammatory response in mouse peritoneal lavage macrophages, 19 inhibiting RANKL-stimulated osteoclastogenesis via NF-κB pathway 20 and decreasing blood glucose levels in diabetic mice, indicating a promising potential for the treatment of osteoporosis and diabetic vascular diseases. 21 Moreover, the administration of Kin significantly repressed inflammatory IL-1β, IL-17, IFN-γ, TNF-α and MMP-9 cytokines production and bone destruction in mice with collagen-induced  22 However, no existing study shows the antigouty arthritis effects of Kin through modulation of macrophages. Herein, we pioneered to unravel that Kin was capable of targeting IKK kinases in macrophages, thus inhibiting the activation of NF-κB signaling and ameliorating MSU-induced HUVECs impairment. Endothelial cells inhabit in 3D architectures in host microenvironments. Nonetheless, currently, the majority of in vitro drug-screening researches were conducted in twodimensional culturing systems. [23][24][25] The genotypic and phenotypic bioactivity of stromal cells in 2D system may diverge tremendously from authentic 3D state owing to the absence of essential modulation of extracellular microenvironments. 26,27 The 3D systems vividly model the interior coculture context, by which flowing environments can be recreated for antigouty arthritis research. So far, many studies have shown that microfluidic system was able to not only simulate the shear stress, interstitial and blood flow that endothelial network exposed to 28 but also assay the efficacy of anticancer drug sensitivity under such vivo-resemble microenvironment. 29 Therefore, a 3D culturing system that HUVECs endured was constructed with consecutively flowing drug stimulation, providing authentic screening results from Anoectochilus roxburghii. We discovered that Kin screening from microfluidic chips was more effective than other representative candidates against MSU-induced impairments of endothelial cells with the lowest value of EC 50 in 3D vivo-like condition. Further results confirmed that non-toxic Kin doses also prevented MSU-intervened proliferation and apoptosis of HUVECs directly. MSU hass an important role during acute arthritis and vascular disorder of joint, which represents a well-established prognosis marker in gouty arthritis patients with obesity, hypertension, diabetes and chronic kidney disease. 30 MSU crystal deposition in synovial tissues produces an influx and phagocytosis of monocyte-macrophages, and the release of cytokines further influences the bioactivity of endothelial cells. 2,31,32 Kin improved HUVECs viability and inhibited HUVECs apoptosis caused by the CM of MSUtreated macrophage cells. Various Kin concentrations directly and indirectly inhibited MSU-induced growth arrest and apoptosis of HUVECs with better efficacy compared with Indo treatment. Therefore, Kin may target inflammatory cells and safeguard the regular biological activities of endothelial cells.
As noted above, the activation of NF-κB signaling serves as the central mediator of inflammation response that impresses the productions of inflammatory enzymes and cytokines. 33 TLR2 at the plasma membrane interacts with MSU crystals, 18 and CD14 serves as adaptor molecule that enable the recognition of MSU to TLR2 and further stimulated crystals uptake and downstream canonical signal pathway. 34 The activation of TLR2 upon MSU binding induces an upregulation of TRAF-6 and TAK-1 phosphorylation. 35,36 Subsequently, the IKK complex, which is comprised of catalytic IKKα, IKKβ and IKKγ subunits, phosphorylates IκBα, which activates the phosphorylation of IκBα and its subsequent degradation. This action is followed by the increased translocation and phosphorylation of downstream p65. 37 Similarly, the activation and phosphorylation of MAPKs (p-JNK/JNK, p-Erk/Erk and p-p38/p38) enact crucial roles during various cell behaviors, involving proliferation, apoptosis and inflammation response. 38,39 We found that the transcription of NF-κB increased markedly in the presence of MSU, and the addition of Kin dose-dependently attenuated NF-κB activation. Kin remarkably inhibited the degradation of IκBα and phosphorylation of IκBα and downstream p65 without influencing the upstream IKK subunits. These results suggested that Kin bound IKK kinases and inhibited corresponding function, which affected the downstream IκBα/p-IκBα. Computational analyses of molecular docking indicated that Kin bound IKKα and IKKβ at specific docking sites, which confirmed our hypothesis that IKKs are targets for Kin. Additionally, the activation of MSU-activated MAPK pathway was repressed with Kin treatment significantly, which might contribute to the regulative inflammation mechanisms upon HUVECs partially.
The expression of IL-1β, IL-6, TNF-α, NO and PGE 2 is mediated via NF-κB signaling. 40-42 NALP3 inflammasome, a vital innate immune pathway that modulates the secretion of inflammatory cytokines, is activated by a variety of bacterial toxins and crystals. 43,44 We found that Kin treatment dosedependently attenuated IL-1β, IL-6, TNF-α, NO and PGE 2 via suppression of NF-κB, which diminished HUVEC damage from Kin-treated macrophages. 14,31 Also, Kin inhibited MSUinduced expressions of NALP3 and TLR2, which contributed to the attenuated inflammation in gouty disease.
NSAIDs are ubiquitously prescribed in clinics to provide anti-inflammatory and analgesic properties via COX inhibition in gout arthritis treatments. 45 COX that supposes to determine the production of prostanoid from arachidonic acid (AA) comprises two isoforms: COX-1, the housekeeping enzyme that is responsible for hemostatic integrity and gastric cytoprotection, and COX-2, the inducible enzyme that reacts to external stimuli and promotes the inflammatory reaction. 46,47 The severe side effects such as gastric bleeding, mucosa erosion and ulcers are unavoidable of NSAIDs such as Indo, which are primarily the results of suppression of the COX-1-derived protective PGE 2 . Kin dose-dependently abrogated MSU-induced COX-2 expression in macrophages without affecting the MSU-stimulated COX-1 decrease in HUVECs. Despite that gout patients take COX-2 inhibitor only for few days, and that the activation of COX-2 has a vital role during other inflammatory arthritis and cancer progression, 48 the selective COX-2 inhibition of Kin could be used in many aspects besides gout arthritis such as anti-inflammation and anticancer strategies.
The influx of a vast number of mononuclear phagocytes into joint synovial tissue is the key feature of acute gouty arthritis. 49 To evaluate more closely the role of Kin against gouty arthritis, we established a murine gouty arthritis model through intraarticular injection of MSU crystals. This procedure triggered a series of inflammatory reactions resemble to that of acute gouty arthritis. 50 In MSU-injected ankle joints, Kin treatments alleviated inflammatory infiltration and ankle perimeter. The impairment of HUVECs produces a thicker interlayer of blood vessels with inflammatory neutrophil infiltration. 51 Kin administration dose-dependently decreased the thickness of the vessel ring muscle with diminished inflammatory infiltration into the medial layer vessel wall of neutrophils, which was better than the efficacy of Indo in vivo. As for intense pain control, 52 the tail-flick response reflects central acting analgesic effect. 53 Results showed that high concentration of Kin extended reaction duration, whereas low concentration of Kin failed to show antinociceptive effects. The writhing response evaluates peripheral acting analgesic effect, 54 and results illustrated that Kin attenuated acetic acid-induced intraperitoneal response, with increased analgesic rate. The antinociceptive effects of Kin were superior to Indo. Our animal experiments demonstrated an obvious suppression of MSUinduced inflammatory infiltration and endothelium damage with evident amelioration of ankle swelling and nociceptive reaction via Kin treatment in vivo.
Undoubtedly, the progress of cancer, microbial infection and other chronic arthritis require the active participation of macrophages. Previous researches revealed that inflammatory cells could serve as potential targets to achieve desirable efficacy during the treatments of tumor, microbial infection and inflammatory bone diseases. [55][56][57][58] Therefore, we consider that the treatment of Kin could be applied in the therapies of cancer, microbial infection and inflammatory diseases, which require future in-depth work.
In conclusion, vivo-like 3D flowing microfluidic chip screening in vitro demonstrated that Kin was the most effective candidate from A. roxburghii against MSU-induced gouty disease. Various Kin treatments suppressed the MSUstimulated proliferation inhibition and apoptosis induction of HUVECs in vitro and in vivo via targeting of IKKα and IKKβ kinases in macrophages, which repressed NF-κB signaling and related cytokine expressions and subsequent endothelium bioactivity. Our study provides a deeper understanding of gouty arthritis and suggests that Kin is a novel COX-2selective inhibitor for the clinical treatment of gouty arthritis.

Materials and Methods
Cells, media and reagents. HUVECs and murine RAW264.7 macrophage cells were cultured with Dulbecco's minimum essential medium (DMEM; HyClone, Logan, UT, USA) containing 10% fetal bovine serum (HyClone), penicillin 100 U/ml and streptomycin 100 μg/ml (Gibco, Invitrogen Ltd, Carlsbad, CA, USA) at 37°C in humidified conditions with 5% CO 2 . CM contained serum-free DMEM medium and 5% bovine serum albumin (Sigma-Aldrich, St. Louis, MO, USA). MSU crystals and Indo were purchased from Sigma-Aldrich and stored in dark at room temperature. Inclusion criterion of selected drugs (Sti, Ua, Kae, Kin and Gas) were: (1) constitute the majority of A. roxburghii; (2) represents each specific type of A. roxburghii (Gas from flavonoids, Kae from phenols, Sti from steroids, Ua from organic acids and Kin from glycosides). Sti, Ua and Gas 13 were purchased from Sigma-Aldrich. Kae and Kin 13 were purchased from Shifeng Biological Technology Company (Shanghai, China).
Fabrication of microfluidic chip. The entire microfluidic chip (GCC) mainly contained medium inputs, parallel microfluidic channels (CGC) and CCC. Two Harvard injection pumps that connected to the inputs were used to manipulate the medium flow in chip. The GCC was manufactured with polydimethylsiloxane (PDMS) by repeated molding on a master, which was fabricated with a 100 mm layer of SU8-2035-negative photoresist onto a glass wafer and shaped by photolithographic methodology. 59 PDMS base and curing agent were mixed violently (mass ratio at 10:1), degasified under vacuum and poured onto the master. After 1 h heating at 80°C and subsequent cooling, PDMS layer was peeled carefully from the master and stick irreversibly to a glass substrate (1.2 mm thick) after biong treated with oxygen plasma for 2 min. Such fabricated microfluidic GCC chip was immersed into ddH 2 O and sterilized under ultraviolet before use.
Determination of drug concentration distribution in chips. The determination of drug concentration in microfluidic GCC was based on the flow field distribution. 60,61 The CGC zone deployed zigzag design so as to mix different medium from inputs sufficiently. Crystal violet and PBS were used to establish varying drug concentrations that could be reflected by the color depth of crystal violet.
Drug administration via microfluidic chip. HUVECs were harvested and resuspended at a concentration of 2.5 × 10 5 /ml in cell-basement membrane extract (BD Bioscience, San Jose, CA, USA) to create a 3D format. 62 A cell-gel mixture (20 μl) was seeded into each CCC before the covering of syringe needle, which allowed the effusion of the inner medium to the outer collecting pipes. 3D endothelial cells were initially treated with various concentrations of MSU crystals (0-300 μg/ml) to validate the optimal urate stimulation dose. Cells pre-treated with culturing medium with and without drugs (100 μM) were further stimulated with MSU medium (300 μg/ml) with and without drugs (100 μM), and then pumped through Input 1 and Input 2, respectively (10 μl/min).
3D flow drug efficacy assay. Ethidium homodimer (Invitrogen, Carlsbad, CA, USA) that binds to nucleic acid of the dead cells and produce red fluorescence was used through the chip for 30 min. 63 The apoptotic cell staining was observed under confocal laser scanning microscope (Leica, Wetzlar, Germany). For quantitative analysis, equal volume of Cell Counting Kit-8 (CCK-8) solution was pumped into the chip for 3 h through two inputs (3 μl/min) and the collected CCK-8 from five outlets were measured at the optical density (OD) of 450 nm.
2D drug efficacy assay. HUVECs (6000 cells per well) were then seeded in a 96-well plate, pre-treated with culturing medium or CM containing Kin (6.25-25 μg/ml) and Indo (20 μg/ml) and stimulated with MSU (300 μg/ml) for 24 h. The CCK-8 assay was performed, and the OD 450 values were measured. [64][65][66] Flow cytometry. For Annexin V/propidium iodide (PI) immunofluorescence, HUVECs were pre-treated directly with DMEM added with drugs and MSU crystals for 24 h. Meanwhile, RAW264.7 macrophages were pre-treated with DMEM containing Kin (6.25-25 μg/ml) and Indo (20 μg/ml) and stimulated with MSU (300 μg/ml) for 24 h. The CM supernatants of macrophages were collected to further administer with HUVECs for 24 h. Cell apoptosis levels of HUVECs from both administrations were determined with flow cytometry via Annexin V/PI staining according to the manufacturer's instructions (BD Bioscience). 67 The percentages of apoptotic HUVECs were presented.
NF-κB luciferase reporter gene assay. RAW264.7 macrophages stably transfected with a p-NF-κB-TA-Luc luciferase reporter construct were used to evaluate the effects of Kin on MSU-induced NF-κB activation according to a previous report. 70 The cells were pre-treated with Kin and MSU for 7 h, and luciferase activities were determined using the Promega Luciferase Assay System (Promega, Madison, WI, USA), following the manufacturer's instructions.
Molecular modeling. The three-dimensional homology models of mouse IKKα/β kinase domains were established with Modeler 9.12 using the architectures of human IKKα/β as templates. The stereochemical constructs of IKKα/β were confirmed using PROCHECK. MolConverter and MarvinSketch (ChemAxon, Budapest, Hungary) were used to generate and refine the spatial coordinates of the Kin drug. Kin was linked to the ATP docking pockets of IKKα and IKKβ using the Lamarckian genetic algorithm based on AutoDock and AutoDock Vina. 71 The consequential molecular modeled figures of binding activities were prepared using the PyMOL visualization software (Schrodinger LLC, New York, NY, USA).
Animal model of MSU-induced ankle gouty arthritis. A total of 80 male Sprague-Dawley rats (200 ± 20 g) were purchased from Shanghai Slac Laboratory Animal Company (Slac, Shanghai, China) and fed commercial food and water in specific pathogen-free conditions approved by Animal Ethical Committee of Shanghai Ninth Hospital. Thirty rats were randomly divided into six groups (five rats per cage). Group 1 rats were injected daily with 100 μl of 0.9% physiological saline into the ankle of their right hind legs consecutively for 3 days. For Groups 2-6, 100 μl MSU crystals (10 mg/ml) was administered into the rat ankle joint of the right lower limb 1 h before an intraperitoneal injection of 500 μl Indo (5 mg/kg per body weight (bw)) in Group 3 and 500 μl Kin (2.5, 5 and 10 mg/kg per bw) in Groups 4-6, respectively. Drugs and MSU crystals were administered daily for 3 days, and the circumferences of injected ankle joints were measured before each MSU injection.
Histological observation. All of the animals were killed 3 days after the drug and MSU interventions, and the injected ankle joints of the right hindlimbs as well as kidney and stomach tissues were harvested and fixed in 4% paraformaldehyde. Synovial tissues from ankle joints and kidney and stomach tissues were paraffinembedded and sectioned for hematoxylin and eosin staining.
Behavioral testing: tail-flick response and writhing response. According to the well-established methods, 72 50 rats were randomly assigned into five groups (five rats per cage): control (Group 1, 0.9% physiology saline, intraperitoneally), Indo (Group 2, 5 mg/kg bw, intraperitoneally), Kin at low (Group 3, 2.5 mg/kg bw, intraperitoneally), medium (Group 4, 5 mg/kg bw, intraperitoneally) and high (Group 5, 10 mg/kg bw, intraperitoneally) concentrations (n = 10). All animals received intra-articular injection of MSU crystals before further behavioral testing. All the required drugs were given intraperitoneally daily for consecutive 7 days. The tail-flick latency was evaluated 1 h after the last drug administrations on the seventh day. A constant hot water of 55 ± 1°C was retained and the distant end of rat tail (5 cm) was submerged into hot water and the latent interval between tail submergence and tail flick was recorded with a stopwatch (s). Also, on the first day before drug interventions, the basal pain threshold (s) was marked to compute the pain latency and the pain threshold increase ratio (%) with below formula: 52 Painthresholdincrease % ð Þ ¼ ðLatencyoftailflickresponseindrugsgroups= The writhing response was assessed with acetic acid (0.7%) (15 ml/kg bw) that was injected into each rat intraperitoneally 15 min after final drugs administration on the seventh day. The intensity of writhing response was measured with the accumulative number of writhing activities within duration of 15 min since acetic acid administration. Analgesic ration was determined with following formula: 52 Analgesicratio % ð Þ ¼ 1FðThenumberofwrithingactivitiesindrugsgroup= thenumberofwrithingactivitiesincontrolgroupÞ 100% Statistical analysis. The data collected were analyzed using SPSS 13.0 (Statistical Package for the Social Sciences, Chicago, IL, USA), and the results are expressed as the means ± S.D. using one-way analysis of variance and post hoc analysis. P-values o0.05 were considered statistically significant.