The IL‐23/IL‐17 axis promotes the formation of retinal neovascularization by activating the NLRP3 inflammasome in macrophages in an experimental retinopathy mouse model

Abstract

Retinal neovascularization (RNV), a pathological process shared among diabetic retinopathy, retinopathy of prematurity and other retinopathies, has been widely studied, but the mechanism remains unclear. In this study, the mechanism by which the interleukin (IL)‐23/IL‐17 axis regulates RNV in oxygen‐induced retinopathy (OIR) model mice and in cell experiments in vitro was characterized. In the retinas of OIR mice, IL‐23/IL‐17 axis activation was increased and regulated RNV formation, and this effect was accompanied by increased macrophage recruitment and nucleotide‐binding domain leucine‐rich repeat and pyrin domain containing receptor 3 (NLRP3) inflammasome activation. Moreover, inhibiting the IL‐23/IL‐17 axis reduced the number of macrophage and the expression and activation of NLRP3 inflammasome. On the other hand, recombinant (r) IL‐23p19 and rIL‐17A promoted the expression and activation of NLRP3 inflammasome, and the proliferation and migration of macrophages. Furthermore, macrophage elimination decreased the activation of IL‐23/IL‐17 axis and the expression and activation of NLRP3 inflammasome. In summary, our experiments showed that the IL‐23/IL‐17 axis promoted the formation of RNV by activating the NLRP3 inflammasome in retinal macrophages of an OIR mouse model.

The IL‐23/IL‐17 axis activation improved RNV formation, and this effect was accompanied by increased macrophage recruitment and NLRP3 inflammasome activation in OIR model mice. Taken together, these findings indicate that the IL‐23/IL‐17 axis promotes RNV by activating the NLRP3 inflammasome in retinal macrophages in OIR model mice. All of these data improve the understanding of the RNV mechanism

Abbreviations

ASC

an apoptosis‐associated speck‐like protein containing caspase activation and recruitment domain

CAS1

caspase‐1

DMEM

Dulbecco's modified Eagle's medium

IL‐23

interleukin‐23

Nab

neutralizing antibody

NLRP3

the nucleotide‐binding domain leucine‐rich repeat and pyrin domain containing receptor 3

NV

neovascularization

OIR

oxygen‐induced retinopathy

PMA

phorbal‐12‐myristate‐13‐acetate

PVDF

polyvinylidene fluoride

RNV

retinal neovascularization

VEGF

vascular endothelial growth factor

INTRODUCTION

Retinal neovascularization (RNV) is a pathological process shared among diabetic retinopathy, retinopathy of prematurity and other retinopathies [1]. Currently, RNV‐related diseases are treated by mainly invasive techniques such as laser photocoagulation and vitrectomy, which may lead to serious complications; result in high recurrence rates; impose very large medical, social and economic burdens; and exhibit efficacy in only some patients. The long‐term therapeutic effects, safety, complications and side effects of agents such as VEGF antagonists injected intravitreally need to be verified in the clinic [2, 3, 4, 5, 6, 7. Therefore, thorough study of the RNV mechanism is urgently needed to provide an experimental basis for the development of safe and effective alternative or combination therapeutic drugs.

Interleukin (IL)‐23 is a heterodimer protein containing P19 and P40 subunits that plays significant roles in autoimmune diseases, such as mandatory spondylitis, reactive arthritis and tumorigenesis [8]. IL‐23 is mainly expressed in activated dendritic cells and monocyte macrophages, binds to the surface receptors IL‐23R and IL‐12Rβ1 [9] and then activates JAK2/ STAT3 signalling molecules in cells, to induce RORγt expression, thus activating Th17 cells [10, 11, 12 and promoting the expression of IL‐17 and other inflammatory factors [13, 14. Previous studies have also confirmed that IL‐23 plays significant roles in the differentiation, survival and amplification of human Th17 cells [12, 13, 15. Intradermal injection of recombinant (r) IL‐23 protein can promote the expression of IL‐17A and IL‐17F cytokines and lead to psoriasis‐like inflammatory changes [13, 14, 15. Therefore, the role of the IL‐23/IL‐17 axis in chronic inflammation has attracted increasing attention. In our previous studies, we also found IL‐23 and IL‐17A to be essential for RNV in oxygen‐induced retinopathy (OIR) model mice [16, 17. However, how IL‐23/IL‐17 axis activation regulates RNV remains unclear. Therefore, we aimed to investigate the mechanism of the IL‐23/IL‐17 axis in RNV by using an OIR model and in vitro cell experiments to further improve the understanding of the occurrence and development of RNV.

MATERIALS AND METHODS

Mice and ethics statement

The animal experiments (mice, C57BL/6J background) used in this study were approved (SYXK‐2011–0026) by the Animal Care Committee and performed in accordance with the Guide for the Care and Use of Laboratory Animals at the Shanghai Jiao Tong University School of Medicine.

Mouse model of oxygen‐induced retinopathy

The OIR mouse model was built as follows: mice at postnatal day 7 (P7) and their nursing mothers were housed at 75 ± 2% oxygen for 5 days and thereafter housed in a normal environment [18]. OIR mice at P12 were divided into a few groups at random. One eye was treated by intravitreal injection with an anti‐IL‐23p19 neutralizing antibody (NAb, 1 μl, 5 μg/μl, R&D Systems, Minneapolis, MN) [16], an anti‐IL‐17A NAb (1 μl, 1 μg/μl, R&D Systems) [17], clodronate liposomes (0·5 μl, FormuMax Scientific, CA, USA) [19] or MCC950 (an inhibitor of NLRP3, 1 μl, 100 μM, Selleck Chemicals, Houston, TX, USA) [20]. The contralateral eye was treated by intravitreal injection with phosphate‐buffered saline (PBS, 1 μl).

Immunofluorescence staining

Frozen cross sections (10 μm) of eyes from normal and OIR mice were fixed, permeabilized and blocked in 5% bovine serum albumin. After incubation with a rabbit anti‐IL‐23p19 antibody (1:200, Santa Cruz Biotechnology, TX, USA), rabbit anti‐IL‐17A antibody (1:200, Santa Cruz Biotechnology), rat anti‐NLRP3 antibody (1:50, Thermo Fisher Scientific. Inc., MA, USA), rabbit anti‐ASC antibody (1:200, Santa Cruz Biotechnology) and rabbit anti‐caspase‐1 (CAS1) antibody (1:200, Santa Cruz Biotechnology) overnight, these cross sections were incubated with a fluorescein isothiocyanate (FITC)‐conjugated anti‐rabbit IgG antibody (1:200, Santa Cruz Biotechnology), a mixture of FITC‐labelled Griffonia simplicifolia Isolectin B4 (GSA‐lectin‐FITC) (1:50, Vector Laboratories, Inc., Burlingame, CA, USA) and F4/80‐PE (1:50, eBioscience, Inc., San Diego, CA, USA), a mixture of the Alexa Fluor 488 anti‐rat secondary antibody (1:500, Thermo Fisher Scientific) and F4/80‐PE (1:50, eBioscience), a mixture of the Alexa Fluor 555 anti‐rat secondary antibody (1:500, Thermo Fisher Scientific) and (FITC)‐conjugated anti‐rabbit IgG antibody (1:200, Santa Cruz Biotechnology) or a mixture of the FITC‐conjugated anti‐rabbit IgG antibody (1:200, Santa Cruz Biotechnology) and F4/80‐PE (1:50, eBioscience) for 1 h in the dark. The THP‐1 cells were fixed, permeabilized, blocked and incubated with a mixture of the rat anti‐NLRP3 antibody (1:50, Thermo Fisher Scientific) and rabbit anti‐ASC antibody (1:200, Santa Cruz Biotechnology) at 4℃ overnight and then incubated with a mixture of the Alexa Fluor 555 anti‐rat secondary antibody (1:500, Thermo Fisher Scientific) and (FITC)‐conjugated anti‐rabbit IgG antibody (1:200, Santa Cruz Biotechnology) at room temperature for 1 h. After staining with 4′, 6‐diamidino‐2‐phenylindole (1:50, Beyotime Biotechnology, Shanghai, China) for 5 min, pictures were obtained by fluorescence microscopy (Nikon Instruments, Inc., Melville, New York, USA).

qRT‐PCR

RNA was acquired and converted into cDNA with a cDNA synthesis kit (Roche, Basel, Switzerland). SYBR Green Mix (Roche) and an ABI 7500 Real‐time PCR System (Applied Biosystems, CA, USA) were used to perform qRT‐PCR. Relative expression was calculated by the 2^−ΔΔCt method. The primers for IL‐23p19 [16], IL‐17A [17], F4/80, NLRP3 [21], ASC [21], CAS1 [21] and cyclophilinA [17] are listed in Table1.

TABLE 1.

Primer sequences for qRT‐PCR

Gene	GI Number	Forward primer (5′–3′)	Reverse primer (5′–3′)	Product length
IL‐23p19	7706702	GAAGGGCAAGGACACCATTA	CCAAGGGCTCGAGACTTTATT	105
IL‐17A	6754324	GACGCGCAAACATGAGTCC	TTTGCGCCAAGGGAGTTAAAG	190
F4/80	33859546	CCACCCTGGCTTTGCATCTA	TCCATATCCTTGGGAGCCTTC	196
NLRP3	94574401	TCCTGGTGACTTTGTATATGCGT	TTCTCGGGCGGGTAATCTTC	284
ASC	165377185	GCTGAGCAGCTGCAAACGA	ACTTCTGTGACCCTGGCAATGA	137
CAS1	86198305	TGCCTGGTCTTGTGACTTGGA	CCTATCAGCAGTGGGCATCTGTA	94
Cyclophilin A	6679439	CAGACGCCACTGTCGCTTT	TGTCTTTGGAACTTTGTCTGCAA	133

Western blotting

Proteins were isolated from retinas and cell lysates and measured with a BCA protein assay kit (Pierce, Rockford, IL, USA). The proteins were separated by sodium dodecyl sulphate polyacrylamide gel electrophoresis and then transferred onto polyvinylidene fluoride (PVDF) membranes (EMD Millipore, Bedford, MA, USA) [22]. The primary antibodies used in the experiment were rabbit anti‐IL‐23p19 antibody (1:500, Santa Cruz Biotechnology), rabbit anti‐IL‐17A antibody (1:500, Santa Cruz Biotechnology), rabbit anti‐NLRP3 antibody (1:1000, Cell Signaling Technology, MA, USA), rabbit anti‐ASC antibody (1:500, Santa Cruz Biotechnology), rabbit anti‐CAS1 antibody (1:500, Santa Cruz Biotechnology), rabbit anti‐IL‐1β antibody (1:1000, Abcam, Cambridge, UK), mouse anti‐β‐actin antibody (1:1000, Cell Signaling Technology) and rat anti‐F4/80 antibody (1:200, Bio‐Rad Laboratories, CA, USA). The membranes were then incubated with secondary antibodies conjugated to horseradish peroxidase for 2 h. The Western blots were imaged by a Tanon 5200 Multichemiluminescent imaging system (Tanon, Shanghai, China) [23, 24 and quantitatively analysed by ImageJ. β‐actin served as an internal control for band intensity normalization.

Retinal flat‐mounts

The cornea, lens and vitreous were removed from the eyes and fixed with 4% paraformaldehyde. The retinas were then carefully separated from the choroids, washed twice with PBS and incubated with GSA‐lectin‐FITC, F4/80‐PE or a mixture of GSA‐lectin‐FITC and F4/80‐PE for 45 min. After washing with PBS, the retinas were mounted on glass slides and photographed under a fluorescence microscope. The area of RNV and the number of macrophages were determined by Image‐Pro Plus software (Media Cybernetics, Silver Spring, MD, USA).

Cell culture and stimulation

A human monocytic cell line (THP‐1, Cell Bank of Chinese Academy of Science, Shanghai, China) was cultured in RPMI‐1640 medium (Gibco, Grand Island NY, USA) containing penicillin‐streptomycin (100 U/ml) and fetal bovine serum (FBS, 10%, Gibco) at 37℃ and 5% CO₂. THP‐1 cells at the logarithmic growth stage were plated into 6‐well plates for several groups and then differentiated into M0 macrophages with 100 ng/mL phorbol‐12‐myristate‐13‐acetate (PMA) (Sigma‐Aldrich, St. Louis, MO, USA) for 24 h. The medium of some group was then replaced with fresh medium containing rIL‐23p19 (10 ng/ml, 50 ng/ml), rIL‐17A (10 ng/ml, 50 ng/ml) (R&D Systems) or PBS [16, 17. The medium of the rest was then replaced with fresh medium containing IL‐17A siRNA or negative control for 6 h and then replaced with fresh medium containing rIL‐23p19 (50 ng/ml). The cells were collected after 24 h for Western blot analysis.

Cell proliferation assay

After digestion with 0·25% trypsin and neutralization with Dulbecco's modified Eagle's medium (DMEM; Gibco) containing 10% FBS (Gibco), RAW264·7 cells (2 × 10⁴ cells/well) in logarithmic growth phase were plated into 96‐well plates. After 24 h, the original medium was replaced with fresh medium containing PBS, rIL‐23p19 (50 ng/ml) or rIL‐17A (50 ng/ml). The cells were cultured for 0, 24 or 48 h at 37℃. After the addition of CCK‐8 reagent incubation for 2 h at 37℃, the absorbance of the plates at 450 nm was detected by a microplate reader (Thermo Fisher Scientific). RAW264·7 cells were inoculated onto Millicell EZ SLIDE 8‐well glass slides (Millipore) at a density of 2 × 10⁴ cells/well. After 24 h, the medium was replaced with fresh medium containing PBS plus BrdU (10 μM), rIL‐23p19 (50 ng/ml) plus BrdU (10 μM), or rIL‐17A (50 ng/ml) plus BrdU (10 μM), and the cells were cultured for 24 h at 37℃. After fixation, denaturation, staining with a BrdU primary antibody and a corresponding secondary antibody and nuclear staining with DAPI, the proportion of BrdU‐positive cells in 16–22 different regions was calculated (100×).

Cell migration assay

RAW264·7 cells (2 × 10⁵) in the logarithmic growth phase were placed into the upper chamber of a Transwell insert with serum‐free DMEM. Then, 500 μl of DMEM containing PBS, rIL‐23p19 (50 ng/ml) or rIL‐17A (50 ng/ml) (R&D Systems) was added to the lower chamber. After incubation at 37℃ for 24 h, the membranes filters were washed twice with PBS, fixed with 4% paraformaldehyde for 20 min and then washed twice with PBS. After staining with 0·1% crystal violet, the adherent cells (on the upper membrane) were removed with cotton swabs, and the cells in the lower layer were quantified with Image‐Pro Plus 6·0 software after observing and photographing 14–17 different regions (100×) at random positions under a microscope.

Statistical analysis

Two‐tailed Student's t‐test and the Student‐Newman‐Keuls test were used to determine statistical significance. Differences were regarded as statistically significant at p < 0·05. All data are presented as the mean ± SEM and were analysed by SAS 9·0 software (SAS Institute Inc., Cary, NC, USA).

RESULTS

Activation of the IL‐23/IL‐17 axis regulated RNV formation in the retinas of OIR mice

The immunofluorescence staining results showed that IL‐23p19 and IL‐17A expression was increased in the retinas of OIR mice at P18 (Figure 1a,b). At P15, IL‐23p19 and IL‐17A mRNA expression was significantly higher in the retinas of OIR mice than in those of age‐matched controls (Figure 1c,d). Injection of IL‐23p19NAb into OIR mice markedly reduced the mRNA expression of IL‐17A (Figure 1e). Similar results were obtained by Western blot (Figure 1f–h). The areas of RNV in the IL‐23p19NAb and IL‐17ANAb groups were markedly decreased (Figure 1i–n). These results suggest that IL‐23/IL‐17 axis activation regulates RNV formation in the retinas of OIR mice.

FIGURE 1.

Activation of the IL‐23/IL‐17 axis was increased in OIR mice, and inhibiting the IL‐23/IL‐17 axis decreased the formation of RNV in these mice. (a–b) Immunofluorescence staining for IL‐23p19 and IL‐17A in the retinas of mice at P18. The arrowheads point to positive areas. (c–d) qRT‐PCR was performed to quantify the mRNA expression of IL‐23p19 and IL‐17A in the retinas of OIR mice and age‐matched controls at P15 (n = 6–8 mice/group). (e) qRT‐PCR was performed to quantify the mRNA expression of IL‐17A in the retinas of OIR mice treated with PBS and IL‐23p19NAb at P15. (f–g) Western blot analysis of the protein expression of IL‐23p19 and IL‐17A in the retinas of OIR mice and age‐matched controls at P15 (n = 3 mice/group). (h) Western blot analysis of the protein expression of IL‐17A in the retinas of OIR mice treated with PBS and IL‐23p19NAb at P15 (n = 3 mice/group). The data are presented as the mean ± SEM from three independent experiments. (i–k) Samples from OIR mice injected with PBS (n = 8 eyes) or IL‐23p19NAb (n = 8 eyes) were stained with lectin‐FITC and flat‐mounted at P18. (l–n) Samples from OIR mice injected with PBS (n = 7 eyes) or IL‐17ANAb (n = 7 eyes) were stained with lectin‐FITC and flat‐mounted at P18. The magnified pictures shown are representative images, and the arrowheads point to positive areas. The data are presented as the mean ± SEM and were analysed by two‐tailed Student's t‐test. *p < 0·05, **p < 0·01, ***p < 0·001

Increased recruitment of macrophages aggravated the levels of RNV in the retinas of OIR mice

Whole retina staining revealed that the areas of RNV and numbers of F4/80‐labelled macrophages were significantly increased in the OIR group (Figure 2a–i). Analysis of frozen sections yielded similar results (Figure 2k–p). The mRNA expression of F4/80 in the retinas of OIR mice at P15 was also significantly increased (Figure 2g). The injection of macrophage eliminators into the vitreous cavity significantly decreased macrophage quantity (Figure 2q–s) and the areas of RNV (Figure 2t–v). These results suggested that the increased recruitment of macrophages aggravated the levels of RNV in the retinas of OIR mice.

FIGURE 2.

The recruitment of macrophages was increased in OIR mice, which promoted the formation of RNV. (a–c, e–g) The retinas of OIR (n = 6 eyes) and normal (n = 7 eyes) mice were stained with lectin‐FITC plus F4/80‐PE and flat‐mounted at P18. (d, h) The magnified images more clearly show the relationship of RNV and macrophages. (i) The numbers of F4/80‐positive macrophages in the retinas of OIR and normal mice are presented as the mean ± SEM and were analysed by two‐tailed Student's t‐test. (j) qRT‐PCR was performed to quantify the mRNA expression of F4/80 in the retinas of OIR and normal mice at P15. (k–p) Immunofluorescence staining for lectin‐FITC and F4/80‐PE in eye cross sections of normal and OIR mice at P18. The arrowheads point to positive areas. (q–v) Samples from OIR mice intravitreally injected with liposomes containing PBS (n = 7 eyes) or clodronate (n = 7 eyes) were stained with F4/80‐PE or lectin‐FITC and flat‐mounted at P18. The number of F4/80‐positive macrophages and the areas of RNV are presented as the mean ± SEM and were analysed by two‐tailed Student's t‐test. The magnified pictures shown are representative images, and the arrowheads point to positive areas. ***p < 0·001

NLRP3 inflammasome activation promoted the formation of RNV in the retinas of OIR mice

Immunofluorescence staining revealed higher expression of NLRP3, ASC and CAS1 in the inner layer of the retina, more colocalization between these molecules and F4/80 and obvious colocalization of NLRP3 and ASC in the OIR mice compared with normal control mice (Figure 3a). This finding indicates that macrophages are the main source of the NLRP3 inflammasome. qRT‐PCR and Western blot analyses showed that the mRNA and protein levels of NLRP3, ASC, CAS1 and activated IL‐1β expression markedly upregulated in the retinas of OIR mice (Figure 3b–j). Intervention with MCC950 decreased the NLRP3 protein expression, increased the ASC protein expression, decreased the pro‐CAS1 and activated CAS1 expression and decreased the activated IL‐1β expression, indicating that MCC950 inhibited NLRP3 inflammasome activation (Figure 3k–p). The areas of RNV were significantly decreased after the intravitreal injection of MCC950 (Figure 3q–s), indicating that the activated NLRP3 inflammasome is involved in RNV formation.

FIGURE 3.

The activation of the NLRP3 inflammasome was increased in OIR mice, which contributed to the formation of RNV. (a) Immunofluorescence staining for F4/80 (red), NLRP3, ASC, CAS1 (green) and NLRP3 (red) plus ASC (green) in the retinas of normal and OIR mice at P18. The arrowheads point to positive areas. (b–d) qRT‐PCR was performed to quantify the mRNA expression of NLRP3, ASC and CAS1 in the retinas of OIR mice and age‐matched controls at P15 and P18 (n = 6–8 mice/group). (e–j) Western blot analysis of the protein expression of NLRP3, ASC, pro‐CAS1, CAS1 p10 and IL‐1β in the retinas of OIR mice and age‐matched controls (n = 3 mice/group). (k–p) Western blot analysis of the protein expression of NLRP3, ASC, pro‐CAS1, CAS1 p10 and IL‐1β in the retinas of OIR mice treated with PBS or MCC950 (n = 3 mice/group). The data are presented as the mean ± SEM from three independent experiments. (q–s) Samples from OIR mice injected with PBS (n = 11 eyes) or MCC950 (n = 13 eyes) were stained with FITC‐lectin and flat‐mounted at P18. The magnified pictures shown are representative images, and the arrowheads point to positive areas. The data were analysed by two‐tailed Student's t‐test. *p < 0·05, **p < 0·01, ***p < 0·001

Inhibiting the IL‐23/IL‐17 axis reduced the number of macrophages and NLRP3 inflammasome activation

Intravitreal injection of IL‐23p19NAb or IL‐17ANAb significantly reduced the numbers of macrophages as determined by whole‐retina immunofluorescence staining (Figure 4a–d, p < 0·01). Western blot assays also showed that the F4/80 protein levels were markedly decreased after treatment with IL‐23p19NAb or IL‐17ANAb (Figure 4e–f, p < 0·001). Compared with the PBS group, the IL‐23p19NAb and IL‐17ANAb intervention groups exhibited decreased F4/80 and NLRP3 staining, increased ASC staining and decreased CAS1 staining (Figure 4g–o). The Western blot results showed that the expression levels of NLRP3, ASC, pro‐CAS1, activated CAS1 and activated IL‐1β were decreased after IL‐23p19NAb or IL‐17ANAb intervention (Figure 4p–s).

FIGURE 4.

Inhibiting the IL‐23/IL‐17 axis reduced the number of macrophages and the expression of the NLRP3 inflammasome. (a–d) Samples from OIR mice intravitreally injected with PBS (n = 8 eyes), IL‐23p19NAb (n = 8 eyes) or IL‐17ANAb (n = 8 eyes) were stained with F4/80‐PE and flat‐mounted at P18. The number of F4/80‐positive macrophages is shown as the mean ± SEM and was analysed by Student‐Newman‐Keuls test. (e–f) Western blot analysis of the protein expression of F4/80 in the retinas of OIR mice treated with PBS, IL‐23p19NAb or IL‐17ANAb (n = 3 mice/group) at P18. (g–o) Immunofluorescence staining for F4/80 (red), NLRP3, ASC and CAS1 (green) in the retinas of mice treated with PBS, IL‐23p19NAb or IL‐17ANAb at P18. The arrowheads point to positive areas. (p–s) Western blot analysis of the protein expression of NLRP3, ASC, pro‐CAS1, CAS1 p10 and IL‐1β in the retinas of OIR mice treated with PBS, IL‐23p19NAb or IL‐17ANAb (n = 3 mice/group) at P18. The data are presented as the mean ± SEM and were analysed by two‐tailed Student's t‐test. *p < 0·05, **p < 0·01, ***p < 0·001

IL‐23 promoted the expression and activation of the NLRP3 inflammasome via IL‐17A in macrophages

To further explore the relationships among IL‐23, IL‐17A and the NLRP3 inflammasome, rIL‐23p19 and rIL‐17A were used to stimulate THP‐1 cells. We observed significantly increased expression and activation of the NLRP3 inflammasome, the colocalization of NLRP3 and ASC, and the formation of ASC specks in the rIL‐23p19‐ and the rIL‐17A‐treated groups, and rIL‐23p19 promoted the expression of IL‐17A in THP‐1 cells. Compared to those in the Negative control plus rIL‐23p19 group, the expression levels of NLRP3 and ASC were significantly decreased, and the ASC specks disappeared in the IL‐17A siRNA plus rIL‐23p19 group (Figure 5).

FIGURE 5.

rIL‐23p19 and rIL‐17A promoted the expression and activation of the NLRP3 inflammasome in macrophages. (a–g) Western blot analysis of the protein expression of IL‐17A, NLRP3, ASC, pro‐CAS1, CAS1 p10 and IL‐1β in THP‐1 cells treated with rIL‐23p19 (0, 10, 50 ng/ml). (h–m) Western blot analysis of the protein expression of NLRP3, ASC, pro‐CAS1, CAS1 p10 and IL‐1β in THP‐1 cells treated with rIL‐17A (0, 10, 50 ng/ml). (n) Immunofluorescence staining of NLRP3 and ASC in PBS group, rIL‐23p19 group and rIL‐17A group. (o) Immunofluorescence staining of NLRP3 and ASC in Negative control (NC) plus rIL‐23p19 group and IL‐17A siRNA plus rIL‐23p19 group. Merged images showed the formation of ASC specks and the colocalization of NLRP3 and ASC, and the magnified pictures shown are representative images. (p) Western blot analysis of the protein expression of NLRP3 and ASC in THP‐1 cells treated with Negative control (NC) plus rIL‐23p19 or IL‐17A siRNA plus rIL‐23p19. The data are presented as the mean ± SEM and were analysed by Student‐Newman‐Keuls test or two‐tailed Student's t‐test. *p < 0·05, **p < 0·01, ***p < 0·001

rIL‐23p19 and rIL‐17A promoted the proliferation and migration of macrophages

We next investigated wondered whether IL‐23 and IL‐17A directly affect the proliferation and migration of macrophages. CCK‐8 assays showed that rIL‐23p19 and rIL‐17A promoted the proliferation of macrophages compared with that in the PBS treatment group (Figure 6a–b). Similar results were observed in the BrdU labelling assay (Figure 6c–h). Transwell assays showed that both rIL‐23p19 and rIL‐17A promoted the proliferation of macrophages (Figure 6i–n).

FIGURE 6.

rIL‐23p19 and rIL‐17A promoted the proliferation and migration of macrophages. (a) A CCK‐8 assay was performed to detect the proliferation of RAW264·7 cells treated with PBS (0 ng/ml) and rIL‐23p19 (50 ng/ml). (b) A CCK‐8 assay was performed to detect the proliferation of RAW264·7 cells treated with PBS (0 ng/ml) and rIL‐17A (50 ng/ml). (c–e) A BrdU labelling assay was performed to detect the proliferation of RAW264·7 cells treated with PBS (0 ng/ml) and rIL‐23p19 (50 ng/ml). (f–h) A BrdU labelling assay was performed to detect the proliferation of RAW264·7 cells treated with PBS (0 ng/ml) and rIL‐17A (50 ng/ml). The ratio of BrdU‐positive cells was calculated in 16–22 different regions. (i–n) Transwell assays were used to detect the migration of RAW264·7 cells (n = 14–17 regions/group). The data are presented as the mean ± SEM from three independent experiments and were analysed by two‐tailed Student's t‐test. *p < 0·05, **p < 0·01, ***p < 0·001

Macrophage elimination resulted in decreased IL‐23/IL‐17 axis activity and NLRP3 inflammasome expression

After the intravitreal injection of clodronate liposomes, the IL‐23p19 and IL‐17A protein expression was significantly decreased in the retinas of OIR mice (Figure 7a–c), and the expression of NLRP3, ASC, pro‐CAS1, activated CAS1 and activated IL‐1β was also significantly decreased (Figure 7d–i). Thus, macrophage elimination resulted in decreased IL‐23/IL‐17 axis activity and NLRP3 inflammasome expression.

FIGURE 7.

Macrophage abrogation decreased the activity of the IL‐23/IL‐17 axis and the expression of the NLRP3 inflammasome. Western blot analysis of the protein expression levels of IL‐23p19, IL‐17A, NLRP3, ASC, pro‐CAS1, CAS1 p10 and IL‐1β in the retinas of OIR mice treated with PBS or clodronate at P18 (n = 3 mice/group). The data are presented as the mean ± SEM from three independent experiments and were analysed by two‐tailed Student's t‐test. ***p < 0·001

DISCUSSION

IL‐23, as a cytokine derived from activated monocyte macrophages and dendritic cells, plays significant roles in autoimmune diseases, tumours and ocular NV [16]. Moreover, IL‐23 is necessary for the production of IL‐17 in cells [25]. IL‐23 binding to the cell surface receptors IL‐23R and IL‐12Rβ1 [17] and subsequent activation of Th17 cells [10, 11, 12, 26 promotes the expression of IL‐17 and causes several chronic inflammatory responses [13, 14, 27. The IL‐23/IL‐17 axis and IL‐23R play significant roles in various autoimmune and inflammatory diseases [27, 28, 29. In OIR mice, we found that IL‐23p19 and IL‐17A expression was significantly increased, while IL‐17A expression was significantly decreased after intervention with IL‐23p19NAb, which indicated that the IL‐23/IL‐17 axis is closely associated with RNV. Injection of NAb against IL‐23p19 or IL‐17A into the vitreous cavity significantly decreased the area of RNV, which indicated that IL‐23/IL‐17 axis activation promoted the formation of RNV. However, the mechanism by which the IL‐23/IL‐17 axis regulates RNV formation is not clear. In our previous studies on IL‐17A, IL‐17A was shown to promote the polarization of macrophages to the M1 type, forming a microenvironment that promoted the immune inflammatory response and aggravated the formation of RNV [17]. This finding indicates that a certain correlation exists between IL‐23/IL‐17 axis activation and macrophages. Previous studies have shown that macrophages are closely related to angiogenesis and that macrophage depletion significantly inhibits angiogenesis [30, 31. In our study, we found that the quantity of macrophages and the areas of RNV were significantly increased in OIR model mice but decreased in OIR mice treated with clodronate liposomes, which suggested that the increased recruitment of macrophages promoted the formation of RNV. This finding is consistent with previous results [32]. Some scholars believe that macrophages infiltrate tissues and secrete high levels of IL‐23p19 in the acute phase (6–24 h) of cerebral haemorrhage, while δ T cells infiltrate into local tissues in the delayed phase (24 h later) and release IL‐17, thereby damaging tissues. In addition, IL‐23^−/− mice exhibit significantly reduced numbers of IL‐17+ cells after cerebral haemorrhage, further suggesting that the production of IL‐17 depends on the expression of IL‐23 [33]. In our current study, the quantity of retinal macrophages and the protein expression of F4/80 were significantly decreased after the intravitreal injection of IL‐23p19NAb or IL‐17ANAb. Moreover, the proliferation and migration of macrophages were significantly enhanced by rIL‐23p19 or rIL‐17A, which suggested that IL‐23/IL‐17 axis activation increased the number of retinal macrophages. Macrophages are involved in phagocytosis, chemotaxis, retinal homeostasis maintenance and immune monitoring, and they secrete large numbers of cytokines [34, 35.

As a part of the induction of the adaptive immune response to nod like receptors, the NLRP3 inflammasome is a significant component of the aseptic immune inflammatory response and an inflammatory protein complex that includes the innate immune receptor NLRP3, the adaptor protein ASC and the protease CAS1 [36, 37. In our study, NLRP3 inflammasome and F4/80 colocalization were increased, suggesting that the NLRP3 inflammasome is closely related to macrophages. The expression and activation of the NLRP3 inflammasome were significantly increased in the retinas of OIR model mice. Treatment with the specific inhibitor MCC950 reduced NLRP3 inflammasome activity and significantly reduced the RNV areas, which indicated that the NLRP3 inflammasome derived from macrophages was involved in RNV formation. After macrophages were eliminated from OIR mice, the expression and activation of the NLRP3 inflammasome were significantly decreased, which further suggested that macrophages are at least one main source of the NLRP3 inflammasome. After intravitreal injection of a NAb against IL‐23p19 or IL‐17A, the expression and activation of the NLRP3 inflammasome were significantly reduced. In vitro assays also showed that the expression and activation of NLRP3 inflammasome were markedly increased and the formation of ASC specks in the rIL‐23p19 and rIL‐17A‐treated groups, while the expression levels of NLRP3 and ASC were significantly decreased and the ASC specks disappeared in the IL‐17A siRNA group which implied that the IL‐23/IL‐17 axis regulated the activation of the NLRP3 inflammasome in macrophages. These results demonstrated that NLRP3 inflammasome activation plays an essential role in the regulatory effect of the IL‐23/IL‐17 axis in RNV formation in OIR mice.

In summary, our experiments elucidated that IL‐23/IL‐17 axis activation improved RNV formation, and this effect was accompanied by increased macrophage recruitment and NLRP3 inflammasome activation in OIR model mice. Taken together, these findings indicate that the IL‐23/IL‐17 axis promotes RNV by activating the NLRP3 inflammasome in retinal macrophages in OIR model mice. All of these data improve the understanding of the RNV mechanism.

CONFLICT OF INTEREST

All authors declare no competing financial interests.

AUTHOR CONTRIBUTIONS

Xi Shen performed result interpretation and finalized manuscript. Ailing Sui, Yanji Zhu and Bing Xie conceived and designed experiments. Yiyun Yao and Yixuan Yao involved in data analysis. Ailing Sui and Xiuping Chen performed the experiments, wrote and modified the paper.

Sui A, Chen X, Yao Y, Yao Y, Shen X, Zhu Y, et al. The IL‐23/IL‐17 axis promotes the formation of retinal neovascularization by activating the NLRP3 inflammasome in macrophages in an experimental retinopathy mouse model. Immunology. 2021;164:803–816. 10.1111/imm.13402