IL‐38: A novel cytokine in systemic lupus erythematosus pathogenesis

Abstract IL‐38 is a newly identified cytokine that belongs to the IL‐1 family. In our previous study, we found elevated plasma levels of IL‐38 in patients with systemic lupus erythematosus (SLE). However, the clear relationship of IL‐38 expression in plasma, peripheral blood mononuclear cells (PBMCs) and clinical and laboratory features needs elucidation. Additionally, we evaluated the possible role of IL‐38 in regulating production of inflammatory cytokines in PBMCs in vitro. A pristane‐induced murine lupus model was used to further demonstrate the effects of IL‐38 on cytokines in vivo and discuss the significance of IL‐38 in lupus development. The results showed that mRNA expression of IL‐38 in PBMCs of patients with SLE was elevated compared with volunteers, and expression of IL‐38 in both plasma and PBMCs was strongly related to clinical features, such as haematuria and proteinuria, and correlated with a SLEDAI score. Plasma levels of TNF‐α, IL‐1β, IL‐6 and IL‐23 were elevated in patients with SLE and were related to plasma levels of IL‐38. In vitro, PBMCs of patients with SLE stimulated with IL‐38 showed a decreased expression of the four inflammatory cytokines compared with PBMCs of patients without treatment. Interestingly, IL‐38 administration in lupus mice significantly reduced the development of lupus, such as reduced proteinuria, improved histological examinations of the kidneys and down‐regulated inflammatory cytokines. In conclusion, IL‐38 may suppress synthesis of pro‐inflammatory cytokines and therefore regulate lupus pathogenesis.

IL-1Ra and IL-36Ra. 4 IL-38 has a weight of 17-18 kDa. It is expressed in several immune cells such as B cells, monocytes and macrophages and some tissues, including the skin, spleen and thymus. 5,6 IL-38 binds to receptor chains, IL-1R1, IL-36R and IL-1RAPL1, 7,8 and then induces an immune response. Recent findings have shown that skin and serum levels of IL-38 were reduced in psoriatic patients, 9 whereas IL-38 expression in patients with Sjogren's syndrome (SS) was higher as compared with controls. 10 Interestingly, IL-38 concentration was elevated in serum from patients with SLE. 11 In our previous study, patients with SLE revealed higher plasma levels of IL-38. 12 To further discuss the relationship of IL-38 and lupus, we first elucidated the association of IL-38 expression in plasma, peripheral blood mononuclear cells (PBMCs) with disease activity, and clinical and laboratory characteristics. We then discussed the role of IL-38 in the production of inflammatory cytokines both in vitro and in vivo, as well as the significance of IL-38 in lupus development.

] years) fulfilling the American
College of Rheumatology revised criteria for SLE were recruited. 13 Thirty-one age-and sex-matched volunteers were assigned to be normal controls ( Table 1.

| PBMC isolation and plasma preparation from patients
Venous blood (10 mL) from each participant was obtained and handled within 3 hours. The PBMC fraction was obtained by density-gradient centrifugation. Isolated human PBMCs were washed twice in phosphate buffer saline and then used for cell culture or total RNA extraction.
Plasma for each participant was stored at −80°C until determination.

| Cell culture
PBMCs were fostered in RPMI 1640 supplemented with 10% FCS, 100 IU/mL penicillin and 100 μg/mL streptomycin (all from HyClone, Thermo) under the condition of 37°C and 5% CO 2 . The cells (5 × 10 5 cells/mL/well) were first treated for the presence or absence of recombinant human IL-38 at 200 ng/mL for 24 hours and then stimulated with lipopolysaccharides (LPS, 1 μg/mL) for 6 hours.
Finally, supernatants were gathered and stored at −80°C until enzyme-linked immunosorbent assay (ELISA) analysis.

| Animal
Eight-week female C57BL/6 mice were purchased from SPF Biotechnology. The mice were treated according to federal and institutional guidelines on animal welfare according to Animal Ethics Committee of Southwest Medical University and had free access to food and water in a temperature-controlled room with a 12-hour light/dark cycle. The mice were divided into three groups. Five mice received 500 uL saline intraperitoneally injected once. Ten mice received 500 uL pristane (Sigma-Aldrich) intraperitoneally once, which were further divided into two groups at the 20th week after injection of pristane (5 mice/group): pristane-induced lupus mice and pristane-induced lupus + IL-38 mice. The lupus + IL-38 group was injected with recombinant murine IL-38 (AdipoGen) for 7 days intravenously (iv) every day at the base of the tail (2.5 ng/uL), and the lupus group received saline (iv) daily for 7 days. All the mice were monitored and killed at the 24th week.

| Biochemical and physiological characteristics of the mice model
Urinary protein and skin lesions were assessed monthly beginning at the age of 8 weeks. Urinary protein expression was determined freshly by collecting morning urine using a semi-quantitative test for each mouse, evaluated as 0-4 according to the manufacturer (Bayer Clinitek).

| Histochemical staining
Two kidneys from each individual mouse were separately used for morphometrical and immunofluorescence assays and ultrastructural analysis. 15 One kidney was fixed in 10% formaldehyde solution, dehydrated in ethanol and processed for paraffin embedding.
Serial 5-um tissue sections of each sample were cut. Haematoxylin and eosin (H&E) and Masson stain were used to check the pathological stage of the specimen tissue. The kidney tissue was stained with FITC-conjugated anti-mouse IgG (Abcam) for immunofluorescence analyses, and a fluorescence microscope (FV1000, Olympus) was used for observing distribution of collagen and immunoglobulin deposition. 16 Another kidney was fixed in 2.5% glutaraldehyde and post-fixed in 1% osmium tetroxide, dehydrated in acetone and em-

| ELISA for IL-38, inflammatory cytokines, ANA and anti-dsDNA
Plasma IL-38 expression was tested by ELISA as previously described. 12 In this study, we discuss the role of IL-38 in PBMCs and mainly evaluated effects of IL-38 on cytokines TNF-α, IL-1β, IL-6 and IL-23. Therefore, we assessed the plasma concentrations of above inflammatory cytokines in both patients and controls and determined the cytokines in supernatant, which were performed using single ELISA. ELISA kits for the pro-inflammatory cytokines were purchased from Neobioscience Inc, Shenzhen, China. Additionally, the inflammatory cytokines were determined in plasma from the mice models using multiplex assay (RayBiotech). The plasma concentrations of ANA and anti-dsDNA antibodies were purchased from Alpha Diagnostic using single ELISA. All samples were measured in duplicate.

| Real-time polymerase chain reaction
Total RNA was drawn from PBMCs using TRIzol (Invitrogen).
Qualified RNA from each individual patient and control was reversely transcribed to cDNA (Bio-Rad). Primers for IL-38 and the internal control β-actin are summarized in Table S1. Relative quantification of the gene transcript for IL-38 was determined with an SYBR Green PCR Kit from TaKaRa on an Applied Biosystems 7900 Real-Time PCR System as reported previously. 12

| Statistical analysis
Parametric and non-parametric statistical analyses were used for calculating the mean ± standard deviation (SD) and median (P 25

| IL-38 in relation to clinical and laboratory characteristics in patients with SLE
In the previous study, we discussed elevated plasma IL-38 in patients Other parameters did not show a significant relationship of plasma IL-38 with patients with SLE ( Table 2; Table S2). Furthermore, to validate the dysregulated expression of IL-38 in patients with SLE, we determined the IL-38 mRNA expression in PBMCs of patients with SLE. We found that mRNA levels of IL-38 were significantly higher in patients with SLE (N = 41) compared with those in healthy controls (N = 31, P < .005, Figure 1H). Subgroup analysis showed that patients with haematuria, proteinuria and pyuria features revealed elevated mRNA levels of IL-38 as compared with the patients without the features (all P < .005, Figure 1I-K). The levels of IL-38 were negatively related to expression of C 3 (r s = −0.333, P < .05, Figure 1L) and C 4 (r s = −0.336, P < .05, Figure 1M). It is notable that mRNA levels of IL-38 are strongly related to SLEDAI score (r s = 0.855, P < .005, Figure 1N). Other parameters did not report a significant relationship of IL-38 mRNA levels with patients with SLE (data not shown). These data suggest up-regulated levels of IL-38 in patients with SLE and are correlated with the disease activity.

| Plasma concentrations of inflammatory cytokines in patients with SLE and correlation with IL-38
Inflammatory cytokines have been demonstrated to contribute to SLE pathogenesis. 3

| Injection of IL-38 relieved nephritis and skin inflammation in C57BL/6 mice
Urinary protein was tested to examine for renal injury. There were low levels of urine protein score in the control group from the 8th week to the 24th week. On the contrary, pristane injection aggravated production of proteinuria, by which the urine protein score in both the lupus group and the lupus + IL-38 group was significantly higher compared with that in the control group both at the 16th week and at the 20th week (P < .005, Figure 5A). The urine protein score in the lupus group and the lupus + IL-38 group was comparable. Urine protein score in the lupus + IL-38 group was significantly reduced after injection of IL-38 compared to that in the lupus group (P < .005) at the 24th week and was similar to the score in control mice ( Figure 5A). It is known that lupus mice develop inflammatory skin lesions such as hair loss in the dorsal neck region. The severity of skin lesions was evaluated before and after IL-38 or saline injection. As shown in Figure S1, the mice in the lupus group treated with saline showed severe skin lesions at the 24th week, whereas mice in the lupus + IL-38 group displayed reduced skin lesions after treatment with IL-38. Mice in the controls showed no skin lesions before and after treatment with saline.
To discuss the effects of IL-38 treatment on histopathology of the kidney, tissues of the kidney were collected at the 24th week.

| D ISCUSS I ON
In our previous study, we found higher expression of IL-38 in patients with rheumatoid arthritis (RA), which also correlated to RA disease activity, showing the potential to be a disease marker for RA and might involve in disease development. 12 Similarly, we observed elevated plasma expression of IL-38 in patients with SLE; however, its clinical association and disease manifestation need elucidation. Several lines of evidence reported that inflammatory cytokines destroyed the immunologic balance in lupus, including TNF-α, IL-1β, IL-6 and IL-23. [17][18][19] Patients with SLE reported elevated serum concentration for TNF-α, 16 and TNF-α signalling contributed to lupus development in lupus-prone mice. 20 TNF-α genetic polymorphisms were strongly related to renal disorders, haematological manifestation and elevated prevalence of positive anti-dsDNA antibody in patients with SLE. 21 Additionally, serum expression for IL-1β was strongly increased in patients with SLE, 22 and IL-6 regulated lupus nephritis development. 23 IL-23 receptor deficiency in lupus mice showed attenuated nephritis, accompanied by reduced accumulation of inflammatory cells in the kidney. 24 We observed significantly higher plasma concentration for the above four inflammatory cytokines in patients with SLE in our study, demonstrating that these pro-inflammatory components are dysregulated in SLE and may perform significantly in the pathogenesis of SLE.
In our study, we reported elevated expression of inflammatory cytokines in patients with SLE. Interestingly, our data also indicated that up-regulated IL-38 significantly correlated with disease activ- Several studies have indicated the roles of IL-38 in inhibition of autoimmune disease development (eg arthritis, psoriasis and lupus). 27,31 IL-38 gene knockout mice treated with K/BxN serum revealed significant arthritic features such as increased disease activity up-regulated synthesis of IL-1β and IL-6 expression in the joints. 31 Collagen-induced arthritic mice injected with IL-38 down-regulated clinical inflammatory scores and reduced the synthesis of IL-23 and TNF-α. 27 Moreover, psoriatic patients reported an elevated expression of IL-38 in skin biopsies after secukinumab treatment. 9 On the contrary, release of IL-38 was strongly reduced by IL-22, IFN-γ and IL-36γ stimulation in human keratinocytes. The effects of IL-38 on psoriasis development were demonstrated through IL-38 injection into imiquimod-induced psoriatic mice, showed down-regulated levels of loricrin, CXCL8, CXCL20 and IL-6, and ameliorated the manifestations, such as reduction of acanthosis and dermal inflammatory infiltrate. 9 Furthermore, injection of IL-38 into lupus-prone mice strongly ameliorated clinical symptoms, for instance, reduced proteinuria, leukocyturia and skin lesions. Levels of IL-22 in mice serum were down-regulated after IL-38 injection. 32 IL-38 injection inhibited phosphorylation of ERK1/2, P38 MAPK and the subunit P65 of NF-κB in human dermal microvascular endothelial cell (HDMEC). 9 It is known that MAPK (ERK1/2, p38) and NF-κB signalling pathways are of importance in lupus pathogenesis. 33,34 These cascade proteins are able to regulate synthesis of inflammatory components.
For example, classical NF-κB signalling knockout in lupus-prone mice strongly down-regulated levels of IL-1α, IFN-γ and IL-6 in the kidneys. 33 Normalizing ERK1/2 in lupus-prone mice reduced expression of IL-1β, TNF-α and IL-6 and relieved development of lupus. 35,36 Therefore, it is possible that inhibiting the activity of downstream signallings by IL-38 may cause less inflammatory component generation in PBMCs during inflammatory response, which will further result in inhibition of dysregulated immunity, and prevent development of lupus.
To confirm the role of IL-38 in generation of inflammatory cytokines, we conducted the lupus mice model by injection of C57BL/6 mice with pristane. The findings showed that wild-type mice that were injected with pristane at the 24th week had severe features of lupus, such as high score of urine protein, damaged histological examination of kidney sections, skin lesions and elevated levels of ANA, anti-dsDNA and IgG deposition. On the contrary, IL-38 administration to pristane-induced lupus mice reduced the effects by pristane. It is notable that IL-38 injection significantly reduced the levels of inflammatory cytokines IL-1β, TNF-α, IL-6 and IL-23 in plasma, demonstrating that IL-38 is able to protect mice from lupus development, and has a negative role in regulating inflammatory cytokines both in vitro and in vivo. Considering the correlation between histopathology of the kidney, SLE and IL-38, the current study found a reduced IgG renal deposition in lupus mice treated with IL-38, and the plasma ANA and anti-dsDNA antibody were reduced as well after IL-38 administration, indicating that IL-38 may suppress the production of autoantibodies in autoimmunity for mice. In another study by Chu et al, MRL/lpr mice showed a decreased percentage of splenic CD3 + CD4-CD8-double-negative (DN) T cells after IL-38 treatment. 32 It is known that DN T cells are a subpopulation of T cells that are highly expressed in patients with SLE and may be involved in tissue damage. 37 DN T cells are able to promote IgG and anti-DNA antibody generation in patients with SLE. 38,39 Therefore, a reduced percentage of DN T cells by IL-38 treatment may help to down-regulate the autoantibody accumulation in lupus.
In conclusion, this study indicates that IL-38 expression correlated with activity of SLE and took part in the development of lupus by regulating inflammatory cytokine generation.

DATA AVA I L A B I L I T Y S TAT E M E N T
Datasets are available from the corresponding author on reasonable request.