Introduction

Multiple sclerosis (MS) is a chronic disease of the central nervous system (CNS), characterized by immune-mediated demyelination and neurodegeneration. Cytokines have been identified as major regulators of the immune system and have the ability to potentiate the immune response in the periphery, to activate immune cells that may cross the blood–brain barrier and to cause local neuroinflammation and/or neurodegeneration [1]. Although the mechanisms causing tissue injury in CNS are not fully understood, it seems that cross-talk between periphery and CNS is essential for the pathogenesis of MS.

One of the cytokines involved in neuroinflammation and/or neurodegeneration in MS is interleukin (IL)-23. IL-23 is a heterodimeric cytokine, composed by a specific p19 subunit and a p40 subunit which is shared with IL-12p70 (IL-12p35/IL-12p40) and may also exist as homodimeric IL-12p80, and as monomeric form [2]. IL-23 drives the expansion of an IL-17-producing CD4+ cell population, known as Th17. The axis IL-23/IL17 is involved in chronic tissue inflammation during infection and autoimmunity. Recently, we have demonstrated elevation of IL-17A in the serum of relapsing–remitting MS (RRMS) patients and the implication of IL-17A in cognitive processes [3]. Nowadays, growing body of clinical and experimental studies revealed increased level of Th17-related cytokines in serum/plasma and cerebrospinal fluid (CSF) [4,5,6,7]. In addition, recent studies demonstrated significant elevation of IL-12p40 in CSF among MS patients, pointed out this cytokine as a possible biomarker for MS-related neuroinflammation [8,9,10], showing good correlation with serum levels [11]. However, neutralization of IL-12/23p40 by ustekinumab, an IL-12/23 p40 neutralizing antibody, had no efficacy in a RRMS clinical phase II study [12]. Previously, it was suggested that the lack of the ustekinumab’s efficacy could be due to the study’s inclusion of patients with advanced disease [13]. Based on this, our current study includes RRMS patients with mild–moderate disability or with no disability as indicating by EDSS scores ranging between 1.0 and 3.5.

Genetic polymorphisms in IL12B gene (encoded IL-12p40) may also contribute to susceptibility and/or pathogenesis of MS by altering cytokines production. In this study, we investigated two polymorphisms in IL12B, the first abbreviated as IL12Bpro polymorphism due to 4 bp microinsertion/deletion combined with an AA/GC transition within the promoter region of the gene (rs17860508) and a single nucleotide polymorphism (SNP) + 16974 A/C in 3′untranslated region (UTR) of IL12B (rs3212227). The effects of these polymorphisms on RRMS risk and clinical characteristics, including the onset of the disease and disability score were explored in this study. Previously, we have shown the impact of studied polymorphisms on genetic risk of another autoimmune disease—systemic lupus erythematosus [14] and their functional effect on serum quantity [15] and on cytokine production after in vitro stimulation [16, 17], that is dependent on the stimuli used. To the best of our knowledge, the role of rs17860508 polymorphism in RRMS was not explored previously, in contrast to rs3212227 polymorphism [18,19,20,21,22,23,24]. Taking into consideration the discordant results reported previously, we may accept that the role of rs3212227 polymorphism for RRMS remains to be clarified.

In the present study, we aimed to determine alternations in IL-12p40 and IL-23 levels in RRMS cases, assess the association between two IL12B polymorphisms (rs17860508 and rs3212227) and the risk and/or degree of disability of RRMS in a Bulgarian population, define their function effects on serum levels of IL-12p40 and IL-23.

Methods

Characteristics of included cases and controls

A total group of 156 Bulgarian patients, diagnosed as having RRMS according to McDonald’s criteria [25] were included. Disability of the patients was assessed using Expanded Disability Status Scale (EDSS) [26]. Patients were recruited between 2014 and 2016 in the Department of Neurology, Medical University of Plovdiv.

The main including criteria were age between 18 and 55 years, treatment with first-line disease-modifying therapies [interferon-beta (IFN-β) or glatiramer acetate (GA)] for at least 6 months, EDSS ≤ 3.5, in remission period of the disease (defined as a period of improvement or stable clinical condition for at least 3 months).

Excluding criteria were coexisting acute or chronic inflammatory disorders, autoimmune, allergic, cardiovascular, hepatic, renal, endocrine (including obesity) diseases; cancer, depression, schizophrenia, intake of cytostatic drugs at any time, or corticosteroids 3 months prior to entering the study; exacerbation phase of MS; known history of drug or alcohol abuse.

The control group for genotyping analyses was composed of 379 clinically healthy volunteers from the same geographic region and ethnicity as cases. The control group for serological analyses was composed of 120 individuals, matched with the patients by age, gender and body mass index. All participants were Caucasians. The demographical and clinical characteristics of cases and controls are shown in Table 1.

Table 1 Characteristics of the studied groups
Full size table

Genotyping of IL12B polymorphisms

Genomic DNA was extracted from venous blood samples by standard salting out method. The quantity and purity of obtained DNA samples were defined by spectrophotometric measurement using NanoVue Spectro-photometer (GE Healthcare, Buckinghamshire, UK).

Genotyping of rs17860508 was performed by allele-specific polymerase chain reaction (PCR). The CTCTAA allele (allele-1) and GC allele (allele-2) were amplified with allele-specific primers in presence of the common (generic) primer. Genotyping of rs3212227 was performed by restriction–fragment length polymorphism–PCR. The amplicons were digested with TaqI enzyme (Thermo Fisher Scientific, Inc). Both genotyping methods were described in details previously [14]. All PCR reactions were performed in a GeneAmp PCR System 9700 (Applied Biosystems, USA). For quality control, at least 10% of random selected samples containing both cases and controls were reanalyzed.

Quantification of cytokine serum levels

Venous blood samples (5 ml) for serum separation were collected from 144 cases and 120 age–sex–BMI-matched controls. After centrifugation at 2000 rpm for 15 min, the serum aliquots were stored at − 70 °C until use. Quantification of serum IL-12p40 and IL-23 was performed by enzyme-linked immunosorbent assay (ELISA) following the manufacturer’s instructions (Invitrogen Bender MedSystems Vienna, Austria). The limit of detection of IL-12p40 and IL-23 ELISA kits were determined to be < 10 pg/ml and 4.0 pg/ml, respectively. The absorbance was measured using ELISA plate reader Thermo Scientific Multiskan EX. The results were calculated according to the standard curves, and expressed as picograms per ml (pg/ml).

Statistical analyses

Demographic characteristics were compared with the Chi square (χ2) test for categorical variables or t test for continuous variables. The data for serum cytokines levels depart from normality and were analyzed by non-parametric Mann–Whitney test for between-group comparisons. These results are presented as a median and interquartile range (IQR 25th–75th percentile). The relationships between cytokine levels and disease duration, age at onset, duration of DMT and EDSS were examined by Spearman correlation test. The testing for deviations from Hardy–Weinberg equilibrium (HWE) was performed by calculating the expected frequencies of each genotype and comparing them with the observed frequencies for cases and controls using a χ2 test, and no deviations were observed (rs17860508: p = 0.95 and p = 0.71; rs3212227: p = 0.32 and p = 0.4, respectively). The genotype distribution and allele frequency of the genetic variants were compared between cases and controls using Pearson’s χ2 test. Yates’ corrected p value (pc) and Fisher’s exact tests were used as appropriate. The strength of the association was presented as odds ratios (OR) with corresponding 95% confidence intervals (CIs) using the logistic regression method. The data were analyzed using the SPSS software, version 21 (IBM, Chicago, IL). Haplotype analysis was performed by SHEsis software [27]. In all tests, two-tailed a p value less than 0.05 was considered statistically significant.

Results

Cytokine serum levels

144 patients with RRMS, 124 treated with IFN-β and 20 with GA, were enrolled to evaluate the serum IL-23 and IL-12p40. The obtained results showed significantly elevated IL-12p40 in cases under IFN-β therapy than cases under GA therapy (203.3 pg/ml, IQR141.7–288.3 vs. 131.7 pg/ml, IQR 81.7–141.7; p = 0.0002). According to this result, further analysis of cytokine levels and characteristics of RRMS cases were made by excluding cases subjected to GA therapy (Table 2).

Table 2 IL-23 and IL-12p40 serum levels in RRMS patients under IFN-β therapy
Full size table

We have found substantially higher IL-12p40 and IL-23 serum levels in RRMS under IFN-β therapy than age–sex-matched controls (p < 0.01). In addition, significantly higher levels of both cytokines were detected in women with RRMS than in men. In our cohort of healthy controls, we observed gender-specific differences just for IL-23. Women showed higher IL-23 than men (11.6 pg/ml, IQR 7.33–30.7 vs. 7.2 pg/ml, IQR 4.36–14.1; p = 0.004), while IL-12p40 serum levels were similar in both genders (67.7 pg/ml, IQR 45–117.3 vs. 71.0 pg/ml, IQR 47.0–86.1; p = 0.61). The increase in serum IL-23 and IL-12p40 was more pronounced in cases with a higher degree of disability (EDSS 2 ÷ 3.5) when compared to the controls and a trend for higher serum IL-23 was also observed in patients with EDSS 2 ÷ 3.5 compared to the patients with EDSS ≤ 1.5 (p = 0.07).

The correlation analyses between IL-23, IL-12p40, disease duration, age at onset, duration of DMT and EDSS did not reveal significant relationships.

Genotype and allelic distribution of IL12B polymorphisms

We assessed two variants (rs17860508 and rs3212227) located in regulatory regions of IL12B in a cohort of 156 RRMS and 379 unaffected, unrelated controls. The distribution of genotypes for both polymorphisms in case and control groups did not show any significant deviation from the Hardy–Weinberg equilibrium (p > 0.3).

The genotype distribution in cases was similar to that in controls for both studied IL12B polymorphisms (Table 3). There was no significant association between any of the polymorphisms and the risk of RRMS. In respect to the well-known prevalence of RRMS among women as well as the observed higher quantity of IL-12p40 and IL-23 in women with RRMS than in men, we evaluated whether there are gender-specific associations of the IL12B polymorphisms with RRMS risk. The frequency of CC-genotype and C-allele was higher in RRMS men than healthy men. More specifically, the CC genotype was associated with higher risk of RRMS in men (OR 3.390, 95% CI 1.007–11.545, p = 0.023) referred to A-carriers (AA + AC genotype). Respectively, the C-allele was more frequent in men with RRMS than healthy men (0.368 vs. 0.24; OR 1.851, 95% CI 1.004–3.410, p = 0.033) referred to A-allele. The genotype and allelic frequencies of rs17860508 and rs3212227 were similar in women with RRMS and controls.

Table 3 Genotype distribution of IL12B polymorphisms in RRMS and controls stratified by gender
Full size table

Haplotype frequencies of IL12B polymorphisms

Following analyses of the individual variants, haplotypes distribution was analyzed (Table 4). Analysis demonstrated that both studied polymorphisms (rs3212227and rs17860508) were in linkage disequilibrium (D′ = 0.708; r2 = 0.137) and all four possible haplotypes were presented in cases and controls. The haplotype IL12Bpro-2/C (rs17860508*2, rs3212227*C) was more frequent in RRMS patients (7.6%) than in controls (3.3%). This haplotype was associated with significant RRMS risk (OR 2.422; 95% CI 1.25–4.7, p = 0.007). Additional gender-stratified analyses showed significant association of IL12Bpro-2/C with risk to RRMS in men (OR 3.740; 95% CI 1.36–10.32, p = 0.007), while the IL12Bpro-2/A haplotype was associated with a significant protective effect (OR 0.531; 95% CI 0.29–0.96, p = 0.035). Analyses among women failed to reveal any significant haplotypic associations with the risk to RRMS.

Table 4 Haplotypes of IL12B polymorphisms in RRMS patients compared to controls stratified by gender
Full size table

Association of IL12B polymorphisms with clinical characteristics in RRMS patients

To evaluate the significance of rs17860508 and rs3212227 polymorphisms for clinical manifestation of RRMS we compared the genotype distribution between cases with early (< 30 years) and later (≥ 30 years) onset of RRMS as well as between cases without functional deficit (EDSS < 1.5) and with mild–moderate disability (EDSS 2.0 ÷ 3.5). The significant gender-specific associations were observed for both studied IL12B polymorphisms (Table 5). The homozygous AA-genotype of rs3212227 was overrepresented among women with early onset of RRMS than women with later-onset (OR 2.368; 95% CI 1.007–5.608; p = 0.03). Respectively, the frequency of allele-A was higher in women with early-onset of RRMS than in women with later-onset (OR 2.078; 95% CI 1.060–4.094; p = 0.021) and men with later or early onset of RRMS. The genotype-AA was overrepresented among men with EDSS 2 ÷ 3.5, and was associated with a higher risk of disability (OR 4.714; 95% CI 0.976–24.455; p = 0.024).

Table 5 Association of IL12B polymorphisms with clinical characteristics of RRMS patients
Full size table

The rs17860508 polymorphism was associated with disability in women with RRMS. We observed higher frequency of genotype-22 and allele-2 in women without functional deficit (EDSS < 1.5) than those with mild–moderate disability (EDSS 2.0 ÷ 3.5). The genotype-22 was associated with lower disability in women with RRMS (OR 0.208; 95% CI 0.055–0.725; pc=0.01). Respectively, there was a significant protective effect (OR 0.567; 95% CI 0.321–0.998; p = 0.036) of allele-2 in relation to disability in women with RRMS. Moreover, there was a significant difference of IL12Bpro genotype distribution between women and men with minimal to moderate disability. The genotype-22 was observed in 40% of men and in 7.8% of women with EDSS 2.0 ÷ 3.5 (OR 7.833; 95% CI 1.722–38.243; pc = 0.004). Considering the onset of RRMS, the genotype-11 was significantly more frequent in men with later onset (56%) than women (24%) with OR 3.943; 95% CI 1.095–14.530; p = 0.015, and healthy men (29.7%) with OR 3.045; 95% CI 0.949–9.921; p = 0.033, and also than in men with early onset of RRMS (29.6%) without reaching statistical significance.

Association of IL12B polymorphisms with serum levels of IL-12p40 and IL-23

In an attempt to identify the functional relation of studied polymorphisms and circulating IL-23 and IL-12p40 in RRMS, serum cytokine levels were compared between patients and controls across the genotypes. With regards to rs17860508 polymorphism, our results showed more pronounced functional effect of this polymorphism on IL-23 rather than IL-12p40 levels in RRMS (Fig. 1). The lowest IL23 levels were observed among RRMS patients with genotype-22. Significant elevation of IL-23 was observed in cases with genotype-11 (p < 0.001) and genotype-12 (p = 0.022), compared to controls with the same genotypes in contrast to genotype-22 (p = 0.381). Moreover, RRMS patients with genotype-22 demonstrated significantly lower IL-23 than patients with genotype-11 (p = 0.047). This association is presented only in women, after stratification according to the gender. Women with genotype-22 showed significantly lower IL-23 levels compared with those with genotype-11 (p = 0.0345), in contrast to men. Although IL23 levels were higher in men with RRMS carriers of genotype-11 compared with genotypes-12/22, this difference was not considered significant. IL-12p40 was significantly higher in RRMS patients than in controls independently of IL12Bpro genotypes and gender.

Fig. 1
figure 1

Serum levels of IL-12p40 (left) and IL-23 (right) in RRMS patients and controls with respect to IL-12Bpro polymorphism (rs17860508), stratified according to the gender. The results are presented as median with interquartile and non-outliers ranges. The outliers and extremes are presented also. *p < 0.05; **p < 0.01; ***p < 0.001—between-group comparisons by Mann–Whitney test

Full size image

The known functional effect of the SNP in 3′UTR of IL12B (rs3212227) on the circulatory IL-12p40 levels was confirmed in RRMS cases (Fig. 2). Patients with AA- or AC-genotypes showed higher IL-12p40 than cases with CC-genotype (p < 0.01) and controls with the same genotypes (p < 0.001). In addition, our results demonstrated the lack of significant functional effect of the rs3212227 polymorphism on IL-23 serum levels among controls and cases of RRMS.

Fig. 2
figure 2

Serum levels of IL-12p40 (left) and IL-23 (right) in RRMS patients and controls with respect to SNP in 3′UTR of IL12B (rs3212227), stratified according to the gender. The results are presented as median with interquartile and non-outliers ranges. The outliers and extremes are presented also. *p < 0.05; **p < 0.01; ***p < 0.001—between group comparisons by Mann–Whitney test

Full size image

Discussion

In our current study, we presented gender dependency of circulating IL-12p40 and IL-23 and of the association of IL12B polymorphisms—rs17860508 and rs3212227 with risk of RRMS and degree of disability in RRMS patients under IFN-β therapy.

We have found substantially higher IL-12p40 and IL-23 in women than men. Moreover, the elevation of serum IL-23 was also detected among healthy women than men. These results suggested that IL-23 may be involved in molecular mechanisms influencing the prevalence of female gender in many autoimmune diseases including MS. The elevation of IL-12p40 in women with RRMS and lack of such dependency in controls could be explained with the involvement of other IL-12p40 contained cytokines in the pathogenesis of the disease, except IL-23, such as IL-12p70, homodimeric and monomeric forms of IL-12p40. The IL-12p70 is important for polarizing to Th1 cells and for decreasing the suppressive effect of Treg cells [2, 28, 29]. The homodimer of IL-12p40 has been shown that induces IL-16 in microglial and macrophages cells, an important CD4+ T cell-specific chemotactic factor [30]. The IL-12p40 monomer also has been reported to upregulate TNF-α in microglial cells [31]. Respectively, we may assume that in the condition of RRMS, the gender dependency of circulating IL-12p40 is well-presented since different IL-12p40-containing cytokines are involved.

The gender dependency was also detected in the association of IL12B polymorphisms—rs17860508 and rs3212227 with risk of RRMS and degree of disability. The carries of rs3212227*CC-genotype and haplotype IL12Bpro-2/C showed higher risk of RRMS over threefold just for men (OR 3.740; 95% CI 1.36–10.32, p = 0.007). In contrast, associative analyses among women failed to reveal any significant associations of both IL12B polymorphisms and RRMS risk. The rs3212227polymorphism was explored previously in several studies [18,19,20,21,22,23,24]. The CC genotype and C allele in SNP rs3212227 of IL12B gene were showed as risk factors for MS in Iran population [19, 21], as protective in Dutch patients [23] and was not associated with MS risk in Italian [22] and other Caucasian populations [18, 24]. The reported discordant results could be due to a different ethnic background or to gender-dependence as we have shown here.

In addition, we have showed that IL12B polymorphisms, rs17860508 and rs3212227, have an impact on RRMS course. The homozygous AA-genotype of rs3212227 was associated with early onset (before 30 years) of RRMS in women (OR 2.368; 95% CI 1.007–5.608; p = 0.03) and with higher risk of disability in men (OR 4.714; 95% CI 0.976–24.455; p = 0.024). A very recent, Benesova et al. [18] also showed an association between rs3212227 and disease severity expressed by the MSSS, in the females Czech patients. The genotype-22 and allele-2 of rs17860508 polymorphism could be accepted as protective factors for the disability in women with RRMS. To our best knowledge, this is the first study exploring the relation of IL12Bpro polymorphism (rs17860508) with RRMS-related disability.

In respect to the functional effect of both IL12B polymorphisms on the IL-12p40 circulating level, we confirmed here that rs3212227*AA-genotype is associated with highest and CC-genotype with the lowest quantity of IL-12p40 in RRMS cases. This functional effect of rs3212227 in stimulus-dependent [16], and in inflammatory media as in RRMS AA-genotype could be marked as higher IL-12p40 producing genotype. However, rs3212227 seems does not influence the serum level IL-23 significantly, opposite to IL12Bpro polymorphism (rs17860508). The genotype-11 was associated with significantly higher IL-23 in RRMS, especially among women. Collectively, we may assume that higher IL-23 level, at least partially determined by genotype-11, is associated with more severe disability in women with RRMS. While the higher quantity of IL-12p40, at least partially determined by genotype-AA is associated with onset before 30 years of RRMS in women. In men, lower-IL-12p40-producing CC-genotype, C-allele or in haplotype with allele-2 of IL12Bpro polymorphism were associated with higher risk of RRMS, while higher IL-12p40-producing AA-genotype was associated with more severe disability. The effect of sex on clinical features of MS is explored previously. Generally, it is accepted that male gender is associated with later onset of disease, poorer prognosis, and faster accumulation of disability [32, 33]. Our findings increase the weight of evidence for a significant sex effect in the pathogenesis of RRMS and raise the possibility that functional IL12B polymorphisms might be also involved in RRMS by modifying IL-12p40-containing cytokines balance.

Some limitation of our study should be mentioned. We cannot exclude the importance of other IL-12-related genetic variations including in IL12A, IL23A on IL-12p40-containing cytokine levels in RRMS. Another potential limitation was that investigated cohort of RRMS is under IFN-β therapy, that may modify IL-12/IL-23 balance. Finally, our study was focused on less severe cases of RRMS, and the observed association could be not confirmed in severely disabled cases of the disease.

In spite of these limitations, we may conclude that: (1) IL-12p40 and IL-23 are tightly involved in early events of immune-pathogenesis of RRMS, since their quantity was significantly altered in periphery of RRMS cases without or with mild–moderate disability; (2) female gender is associated with higher serum IL-12p40 and IL-23 among RRMS cases with low level of disability; (3) IL12B polymorphisms (rs17860508 and rs3212227) have gender-dependent effect on susceptibility to and diseases course of RRMS, probably mediated by their gender-dependent functional effect on IL-12p40-containing cytokines.