Serum Interleukin-23 in Polish Patients with Systemic Lupus Erythematosus: Association with Lupus Nephritis, Obesity, and Peripheral Vascular Disease

Objectives To analyze the correlation between the serum concentration of interleukin- (IL-) 23 and atherosclerotic changes, traditional atherosclerotic risk factors, the autoantibody profile, and involvement of selected organs in systemic lupus erythematosus (SLE) patients. Patients and Methods We studied 94 SLE patients and 27 controls. We analyzed the IL-23 serum concentration, autoantibodies, carotid intima-media thickness and atherosclerotic plaque, the ankle-brachial index, atherosclerotic risk factors, and organ manifestations. Results Concentrations of IL-23 significantly differed between SLE patients and the controls (p = 0.0015). On the basis of multivariate stepwise analysis, we revealed that high levels of IL-23 were associated with atherosclerotic plaque in common femoral arteries (OR = 12.67; 95% CI: 1.41–113.84), lupus nephritis (OR = 3.69; 95% CI: 1.16–12.22), and obesity (OR = 4.21; 95% CI: 1.40–12.67). Autoantibodies related to IL-23 were anti-phosphatidylethanolamine antibodies (OR = 11.06; 95% CI: 1.24–98.65) and anti-SS-B/La antibodies (OR = 15.43; 95% CI: 1.73–137.25). Conclusions IL-23 may be involved in lupus nephritis pathogenesis. Through its association with obesity and selected antiphospholipid antibodies, IL-23 might promote a hypercoagulable state contributing to atherothrombosis development in SLE patients.


Introduction
Systemic lupus erythematosus (SLE) is a chronic, inflammatory, multisystem autoimmune disease characterized by an irreversible break in immunologic self-tolerance and successive immune-mediated tissue damage [1]. A remarkable feature of the disease is the clinical heterogeneity, which indicates variations in underlying etiologic factors. The multifactorial etiology of SLE includes genetic susceptibility and hormonal and environmental triggers [2,3]. Defective function of T cells and overactivation of B cells as well as defective clearance of apoptotic debris cause the production of autoantibodies, activation of complement, formation and deposition of immune complexes, and, consequently, tissue and organ damage [4,5]. The crucial role in SLE pathogenesis plays innate and adaptive immune dysregulation, and it has been confirmed that certain cytokines are closely linked to SLE pathogenesis [6].
However, to our knowledge, no complex data are available considering the additional relationship between serum concentrations of IL-23, atherosclerosis, and atherosclerotic risk factors in SLE patients.

Objectives
The aim of this study was to evaluate levels of IL-23 and their association with atherosclerotic changes, traditional atherosclerosis risk factors, disease characteristics including autoantibody profiles, and the involvement of selected organs in SLE patients.

Materials and Methods
This study was approved by the Local Ethics Committee of Pomeranian Medical University in Szczecin. Informed consent was obtained from all patients.
All patients were Caucasian. We studied 94 SLE patients and 27 healthy volunteers as the controls. The diagnosis of SLE was made according to the 1997 American College of Rheumatology Revised Criteria for Classification of Systemic Lupus Erythematosus [24].
All SLE patients and matched controls underwent noninvasive imaging investigations in the Department of Diagnostic Imaging and Interventional Radiology Pomeranian Medical University in Szczecin. All analyses were performed by the same experienced ultrasonographist with HDI 3500 (ATL) using a 5-12 MHz linear transducer. Carotid intimamedia thickness (cIMT) measurements were performed with B-mode ultrasound in common carotid artery, bifurcation, and internal carotid artery on the right and left sides according to procedures previously described [26]. Due to the high variability of this parameter in populations [27,28], the normal and pathological ranges of cIMT values were established on the basis of measurements in the controls [29]. The B-mode ultrasound was also used as a screening for atherosclerotic plaque presence in carotid and lower extremity arteries (the iliac, common femoral, deep femoral, superficial femoral, popliteal, and tibial arteries) [30]. The ankle-brachial index was assessed using Doppler ultrasonography and calculated as a ratio of systolic pressure measured in the posterior tibial and dorsal arteries of both feet to the systolic pressure in the brachial artery. Abnormal values were considered at ABI < 1.0 [31].
We assessed the presence of traditional risk factors of atherosclerosis: hypertension (systolic blood pressure ≥ 140 mmHg, diastolic pressure ≥ 90 mmHg), dyslipidemia (total cholesterol > 190.0 mg/dL, LDL cholesterol > 115.0 mg/dL, HDL cholesterol in males < 40.0 mg/dL and in females < 45.0 mg/dL, and triglycerides > 150.0 mg/dL), overweight and obesity based on body mass index (BMI) (overweight when BMI was 25 to <30, obesity when BMI was ≥30), diabetes, smoking habits, oral contraceptive use, and positive family history of cardiovascular diseases.
Serum was stored at −80°C until analysis for IL-23 using the sensitive ELISA method with the Quantikine Human IL-23 Immunoassay ELISA kit (minimum detectable dose less than 6.8 pg/mL). Kits were from R&D Systems, Minneapolis, USA. The system uses microplates coated with a monoclonal antibody and an enzyme-linked polyclonal antibody specific for IL-23. All analyses and calibrations were performed in duplicate and were read using BioTek PowerWave XS, BioTek Instruments, Winooski, USA.
All continuous variables were checked for equality distribution with Kolmogorov-Smirnov test. Data are described as the mean ± standard deviation and median (Q1, Q3). Comparison of continuous variables was performed by the Mann-Whitney U test. For categorical variables, differences were assessed by the logistic regression model and multivariate stepwise analysis. The logistic regression model and multivariate stepwise analysis probability (P) were assessed by a chi-square testing or Fisher's exact test. Results were shown as a P value, odds ratio (OR), and 95% confidence interval (95% CI) adjusted for sex and age. Findings were considered statistically significant at P < 0 05. Additionally, Bonferroni correction was used for related variables. All statistical analyses were performed with STATISTICA version 8.0, StatSoft Inc., Tulsa, OK, USA.

Results
The clinical and laboratory characteristics of the patients and healthy controls, including classical atherosclerotic risk factors and serologic profile, are presented in Tables 1, 2, 3, 4, and 5. The study group consisted of 82 women and 12 men suffering from SLE. The mean age was 44.5 years. The disease duration ranged from 1 to 30 years. The majority of patients presented low and medium SLE activity indices (52 and 33, resp.). The coexistence of APS was confirmed in 31, kidney involvement in 24, stroke in 10, and CAD in 11 SLE patients. The analysis of classical atherosclerosis risk factors revealed abnormal levels of total cholesterol in 52, LDL cholesterol in 40, HDL cholesterol in 5, and triglycerides in 25 patients with SLE. Hypertension was present in 37, overweight in 15, obesity in 23, diabetes in 11, smoking habits in 31, and positive family history of cardiovascular diseases in 4 SLE patients. The use of oral contraceptives confirmed 4 women suffering from SLE. Table 5 shows the serological profile of SLE patients at the time the study was conducted. The frequency of ANA was 77.7%, which is lower than usually reported in SLE patients. However, the prevalence of ANA can change over time. In our patient group, the duration of the disease ranged from 1 to 30 years. Moreover, various immunosuppressive therapeutic schemes were applied in the patients and the majority of them went into remission which also might be the case of ANA disappearance in some of the SLE patients. The analysis of disease histories revealed that at the moment of diagnosis, all of them were positive for ANA.
Serum levels of IL-23 were significantly higher in SLE patients than in the controls (P = 0 0015 after Bonferroni correction). The maximum level of IL-23 in the controls was 3.1 pg/mL and was established as the cut-off value ( Table 6 and Figure 1).
The reference range of cIMT was established on the basis of the measurements in the controls. Intima-media thickness were considered moderately increased at values of 0.66-0.86 mm (51 SLE patients) and highly thickened at >0.86 mm (15 SLE patients). Values of ABI < 1.0 were found in 22 patients and <0.9 in five patients with SLE.
In the multivariate stepwise analysis, the OR for the serum level of IL-23 > 0.0 pg/mL in SLE patients with atherosclerotic plaque in the right common femoral artery was 12.67 (95% CI 1.  (Table 8).
There was no correlation between IL-23 and other analyzed laboratory and serologic characteristics in SLE patients (data not shown).

Discussion
The analysis of the role of selected proinflammatory and angiogenic cytokines in SpA patients [14,15,32,33] showed higher levels of IL-23 in ankylosing spondylitis in comparison with the controls and revealed the association with selected clinical characteristics such as psoriatic onycholysis and the use of nonsteroid anti-inflammatory drugs [14,15].
In addition, our initial reports [34] on atherosclerotic risk in SLE patients showed that risk factors for atherosclerosis    in SLE differ from those observed in the general population. Chronic inflammation, the presence of selected autoantibodies (e.g., aPLs and AECA), and applied treatment (especially high cumulative dosage of glucocorticosteroids) are particularly important in atherosclerosis development. However, there is still the need to search for other pathomechanisms underlying this process.
In this study, we evaluated the relationship between serum levels of IL-23 and clinical and laboratory characteristics of SLE with the special focus on atherosclerosis and atherosclerotic risk factors. We found a significant correlation between high serum levels of IL-23 and atherosclerotic changes in the arteries of the lower extremity arteries. This finding stays in accordance with the reports on the IL-23  serum levels in patients affected by peripheral arterial disease (PAD) [22]. The authors analyzed 29 patients with advanced PAD: 28 had stage IV and one had stage III disease. All patients were qualified for lower extremity bypass (LEB) surgery using either autogenous vein or synthetic polytetrafluoroethylene (PTFE) graft material. Blood for measurement of IL-23 levels was taken at three points: 24 h before, 24 h after, and 5 days after the surgical intervention. IL-23 levels were higher at all three time points compared to the control group. Interestingly, 22 patients undergoing LEB with the autogenous vein had IL-23 levels greater than the seven patients who had synthetic PTFE graft material at all three time points. In our study, we did not find a significant correlation between serum levels of IL-23 and ABI values.   This might be the result of the low number of patients with clinical manifestation of PAD. However, the number of SLE patients with ABI < 1.0 was 22, but significantly decreased values of this index <0.9 were found only in three patients, and among them, two confirmed intermittent claudication symptoms and only one underwent surgical intervention. In any case, the significant association between IL-23 levels and the presence of atherosclerotic plaque in lower extremity arteries, especially common femoral arteries, supports the hypothesis that IL-23 might be involved in PAD development in the course of SLE. Furthermore, there are data in the literature that address the possible role of the IL-23-IL-17 axis in carotid atherosclerosis [21] and in CAD [35]. Abbas et al. [21] showed that patients with carotid atherosclerosis had markedly raised plasma levels of IL-23 compared with healthy controls. The highest levels were present in patients with the most recent symptoms within the last two months. Moreover, IL-23 concentrations at follow-up after several years (mean 3.5) were associated with increased mortality mainly caused by cerebrovascular and cardiovascular events.
Khojasteh-Fard et al. [35] found a significant decrease of IL-23 gene expression in unstimulated peripheral blood lymphocytes of patients with CAD compared to those without CAD reinforcing the potential role of IL-23 in the complex mechanisms associated with the development of atherosclerosis. Data from patients who underwent diagnostic catheterization showed imbalances between Th17 and regulatory T cells and significantly higher levels of serum IL-23 along with IL-17 and IL-6 in those with acute MI and unstable angina compared with concentrations in patients with stable angina and chest pain syndrome [36]. Our study did not reveal a significant correlation of IL-23 serum levels and CAD, MI, stroke, and TIA as well as cIMT in SLE patients. To our knowledge, there are no available data in the literature concerning this issue. Some studies showed increased levels of IL-6 in SLE patients and the association with the burden of atherosclerosis [37]. Similarly, SLE patients with atherosclerosis showed increased levels of IL-6 and IL-17 compared with SLE alone and the controls [38]. The exact role of IL-23 and the IL-23-IL-17 axis in atherosclerosis and the development of complications in SLE patients still remains the issue that should be the matter of further studies.
In this study, we also analyzed the correlation between serum levels of IL-23 and classical atherosclerotic risk factors in SLE patients. We only found a significant association between IL-23 levels and obesity. Data from the study in 26 obese women and 20 lean volunteers showed significant increases of IL-23 levels as well as IL-17, leptin, and macrophage migration inhibitory factor (MIF) in obese subjects compared with the controls. Interestingly, there was no relationship between IL-12 and interferon γ and obesity. Moreover, the authors did not find the correlation between IL-23 and IL-17 and central obesity markers like BMI and waist circumference in obese women that suggests that the accumulation of IL-23-IL-17-producing cells in adipose tissue might not increase in proportion with the buildup of abdominal fat. It is also possible that blood mononuclear cells or some other cell type, rather than leukocytes infiltrating the adipose tissue, could be the main source of IL-17 and IL-23 in obese subjects. The role of the IL-23-IL-17 axis seems to be increased in obese women independently of proinflammatory mediators MIF and leptin [23]. These observations together with the data on the possible role of the IL-23-IL-17 proinflammatory cytokine axis in atherosclerosis development may indicate the important link between obesity and cardiovascular complications in SLE patients, and further research should be conducted to investigate this problem.
The association between SLE activity and clinical characteristics of the disease and IL-23 levels was also studied. We found significantly higher levels of IL-23 in SLE patients than in the controls, and there was also an important correlation of IL-23 serum concentrations and lupus nephritis, which is in accordance with other data. The relationship between the elevated serum levels of IL-23 and renal involvement in SLE patients was confirmed by several studies [39][40][41]. High serum IL-23 mRNA levels were also found in patients with SLE and renal disease [20]. Similarly, higher levels of serum IL-23 were found in SLE patients compared with the controls [17,36,42]. We did not confirm the relationship between SLE activity and other clinical characteristics of the disease. In the literature, contrary data are available on the association of IL-23 and SLE activity. Some reports showed a correlation between high levels of IL-23 and activity of the disease [18] as well as a relationship between IL-23 levels and selected manifestations like serositis and cutaneous involvement among patients with active disease [19]. Some reports presented discrepant results and did not show any association of serum IL-23 and SLE activity [12,17,36]. In our study, only nine patients had high activity of the disease according to the SLEDAI scale. The majority had low (52) or medium (33) activity, which may partly explain the lack of a relationship between IL-23 and activity of SLE.
Another interesting finding in our study was the significant correlation between serum IL-23 and selected autoantibodies including aPLs and aSS-B/La. Little data are available on the association between IL-23 and the serologic profile in SLE patients. Some studies did not find any correlation of IL-23 with autoantibodies typical for lupus like anti-double stranded DNA or anti-Sm [18,19]. In fact, only the basic profile of autoantibodies was analyzed in these studies. One paper focused on IL-17 single-nucleotide polymorphisms and increased activity of the IL-23-IL-17 cytokine axis including serum measurements of IL-17 and IL-23, and tissue growth factor β in PAPS confirmed higher levels of IL-23 than that in controls. A significant correlation between IL-17 and IL-23 was reported. IL-17 was more frequently found in patients with deep vein thrombosis and thrombocytopenia. No significant associations between the analyzed cytokines and aPLs or clinical manifestations of PAPS were found [13]. In our previous observations in SLE patients with atherosclerotic changes, we found significant relationships between aPLs, secondary antiphospholipid syndrome, and atherosclerotic lesions [34,43]. In this regard, the current finding is important to explain the possible mechanism of their involvement in atherosclerosis pathogenesis in SLE patients. Further studies are warranted to fully analyze the role of the IL-17-IL-23 cytokine axis in this phenomenon.

Conclusion
In conclusion, there are data, including ours, that support the involvement of IL-23 in SLE pathogenesis. Higher levels of IL-23 in patients compared with the controls and a significant correlation between IL-23 levels and selected clinical characteristics such as lupus nephritis were documented. The limitation of our study was that we analyzed only IL-23 serum levels, without the broad spectrum of proinflammatory cytokines. A relatively low number of patients presented cardiovascular and cerebrovascular manifestations. The majority of patients showed low activity of the disease. Even though we were able to confirm the interesting relationship between atherosclerotic lesions in the arteries of the lower extremities, obesity as well as selected aPLs and IL-23 levels in lupus patients also suggests the possible contribution of this cytokine in vascular involvement in the course of SLE, revealing new insights into atherothrombotic risk assessment in SLE patients. Further studies are necessary to fully elucidate this issue.