Increased group 2 Innate lymphoid cell (ILC2) activity in pediatric familial mediterranean fever patients with spondyloarthropathy

Abstract

Familial Mediterranean fever (FMF), an autoinflammatory disease caused by MEFV mutations, frequently coexists with immune-mediated conditions such as spondyloarthritis (SpA). However, the contribution of innate lymphoid cells (ILCs) to FMF-associated SpA remains poorly defined. Our objective was to characterize the phenotype and functional activity of ILCs in pediatric patients with FMF, SpA, and overlap FMF/SpA. Peripheral blood mononuclear cells (PBMCs) and synovial fluid mononuclear cells (SFMCs) from FMF (n = 15), SpA (n = 11), FMF/SpA (n = 13), and age-matched healthy controls (n = 14) were analyzed by flow cytometry to quantify ILC subsets. Plasma and synovial cytokines were measured by ELISA, and expression of ILC-associated cytokines in PBMCs and sorted ILCs was assessed by quantitative PCR. Total ILC frequencies and subset distributions in PBMCs and SFMCs were comparable across groups (p > 0.05). However, FMF/SpA patients demonstrated increased IL4 and IL5 expression in sorted ILCs (p < 0.05), consistent with a type 2–skewed ILC activation profile. PBMC thymic stromal lymphopoietin (TSLP) expression was elevated in SpA (p < 0.01). Plasma IL-6, IL-8 and IP-10 concentrations were higher in the full cohort of FMF/SpA patients; however, these differences were sensitive to outlier exclusion and indicate substantial inter-individual heterogeneity. Synovial fluid from SpA patients showed higher IP-10 and IL-17 levels (p < 0.05). Pediatric FMF/SpA is characterized by altered ILC functional polarization rather than numerical expansion, with a prominent type 2 ILC signature. In contrast, SpA displays synovial IP-10 and IL-17 enrichment and increased PBMC TSLP expression. These findings suggest distinct innate immune activation axes in overlap versus isolated SpA and support further investigation of ILC-mediated pathways in FMF-associated spondyloarthritis in larger cohorts.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10238-026-02260-x.

Highlights

ILC2 activation (IL4/IL5↑) marks pediatric FMF/SpA despite comparable ILC counts.

SpA shows elevated TSLP and synovial IP-10/IL-17, implicating local drivers.

Circulating IL-6/IL-8/IP-10 rise in a subset of FMF/SpA patients.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10238-026-02260-x.

Introduction

Familial Mediterranean fever (FMF) is a prototypic autoinflammatory disease caused by pathogenic variants in the MEFV gene encoding pyrin, leading to dysregulated pyrin inflammasome activation and increased production of innate inflammatory mediators during recurrent attacks [1, 2]. Although FMF is classically defined by self-limited episodes of fever and serositis with elevated acute-phase reactants, a growing body of clinical evidence shows that FMF frequently coexists with immune-mediated inflammatory comorbidities, including spondyloarthritis (SpA) [3, 4–8]. In a large population-based analysis, FMF was strongly associated with MHC class I–related inflammatory diseases and ankylosing spondylitis, supporting a shared “MHC-I-opathy” landscape in which innate immune dysregulation can amplify tissue inflammation and interface with adaptive immune pathways [8].

SpA pathogenesis is increasingly understood as a continuum of autoinflammation and immune-mediated inflammation rather than a purely autoimmune process [9, 10]. The IL-23/IL-17 axis is central to joint and entheseal inflammation, with IL-23 produced by antigen-presenting cells supporting the expansion and effector programming of IL-17–producing lymphocytes [11, 12]. In addition to conventional Th17 cells, multiple innate-like lymphocyte populations—including γδ T cells, MAIT cells, and innate lymphoid cells (ILCs)—can contribute to IL-17/IL-22–dominated inflammation in SpA [10–12]. In pediatric SpA and related entities, synovial IL-17 can be readily detected and may be enriched locally compared to circulation, underscoring the importance of tissue-level immune circuits [9].

ILCs are antigen receptor–negative lymphocytes derived from common lymphoid progenitors that rapidly respond to tissue “alarmins” and cytokines and are commonly enriched at barrier sites and inflamed tissues [13–15]. ILC1, ILC2, and ILC3 subsets mirror Th1, Th2, and Th17-like programs, respectively, and are defined by distinct transcriptional regulators and effector cytokines [13–15]. Of particular relevance to inflammatory arthritis, ILC2s can be activated by epithelial/stromal signals (including IL-25, IL-33, and thymic stromal lymphopoietin [TSLP]) and can produce type 2 cytokines such as IL-4 and IL-5 [13, 16]. Experimental models further support functional roles for ILC2s in shaping joint inflammation; for example, enhancing ILC2 abundance or activity can attenuate inflammatory arthritis in mice, highlighting that ILC2 biology can influence disease severity even when overall lymphocyte numbers are not dramatically altered [16–18].

Within the SpA spectrum, human studies have most consistently implicated ILC3s, particularly IL-22 and/or GM-CSF–producing ILC3 populations, in inflamed joints and mucosal sites [19–21]. Expansion of ILC3s has been reported in the gut, peripheral blood, synovial fluid, and bone marrow of ankylosing spondylitis patients, supporting the concept of a gut–joint axis in which innate lymphoid programs contribute to tissue inflammation [21]. Synovial ILC analyses in inflammatory arthritis have also shown that the joint environment can be enriched for chemokines and cytokines that promote lymphocyte recruitment and effector function. For example, CXCL10/IP-10 is strongly expressed within inflamed synovium and detectable in synovial fluid in pediatric inflammatory arthritis, consistent with a role in CXCR3-dependent immune cell trafficking into joints [22]. Despite these insights, the contribution of ILCs to FMF-associated SpA—particularly in children—remains insufficiently defined, and it is unclear whether disease overlap reflects changes in ILC abundance, subset distribution, or (critically) activation state.

Based on the convergence of inflammasome-driven autoinflammation in FMF, tissue-focused innate-like immune circuits in SpA, and the ability of ILCs to rapidly translate cytokine/alarmin cues into effector programs, we hypothesized that pediatric FMF/SpA is associated with an altered ILC activation signature. Therefore, we performed a combined phenotypic and functional assessment of circulating and synovial ILCs in pediatric FMF, SpA, and FMF/SpA patients. Using flow cytometry to quantify ILC subsets, cytokine measurements in plasma and synovial fluid, and transcriptional profiling of ILC-associated cytokines in PBMCs and sorted total ILCs, we aimed to define whether FMF/SpA is marked by distinct innate lymphoid effector programs, particularly type 2 cytokine polarization, and to identify candidate upstream signals and tissue chemokine axes that may shape local inflammatory circuits in pediatric SpA.

Materials and methods

Patient selection

This study included 53 pediatric patients followed at the Department of Pediatric Rheumatology, Erciyes University Faculty of Medicine: 15 with FMF, 11 with SpA, 13 with FMF/SpA, and 14 age-matched healthy controls. The controls were recruited from the healthy age/sex matched donors who come general pediatrics. FMF patients were selected based on the 2019 diagnostic criteria [23] and confirmed pathogenic MEFV gene mutations. SpA patients met the International League of Associations for Rheumatology (ILAR) criteria [24].

Patients were evaluated during disease flares to minimize the influence of ongoing treatments. For FMF, flares were defined by clinical symptoms (e.g., fever, abdominal pain, arthritis) and elevated acute-phase reactants, with severity assessed using the Pras scoring system [25]. For SpA, flares were defined by new-onset arthritis or active sacroiliitis confirmed by magnetic resonance imaging (MRI). In FMF/SpA patients, arthritis persisting for ≥ 6 weeks in an affected joint or active sacroiliitis on MRI was considered a flare. SpA disease activity was evaluated using the Juvenile Spondyloarthritis Disease Activity Index (JSpADAS) [26]. Peripheral blood was collected from all patients during flares for complete blood count, acute-phase reactants, and cellular analysis. Synovial fluid was obtained from patients with peripheral arthritis prior to intra-articular steroid injection. All procedures were conducted at Erciyes University Genome and Stem Cell Center. The study was approved by the Erciyes University Clinical Research Ethics Committee (Decision No: 2022/91) and adhered to the Declaration of Helsinki. Informed consent was obtained from all individual participants included in the study. Written informed consent was obtained from the parents.

Isolation of peripheral blood and synovial fluid mononuclear cells

Ten mL of blood was collected from patients in each group into sterile EDTA-filled hemogram tubes and was diluted with equal volume of phosphate-buffered-saline (PBS). Diluted sample was slowly overlayed onto 5 mL Ficoll-Hypaque to form a gradient. Blood was centrifuged at 400 g for 30 min at room temperature without brake and Peripheral blood mononuclear cells (PBMCs) were collected from the mid-phase in a 15 mL falcon, washed and counted with Tyrpan Blue on Thoma slide.

To isolate synovial fluid mononuclear cells (SFMCs), synovial fluid was centrifuged at 300 g for 7 min at 4 °C. Two ml of the supernatant phase was frozen for ELISA experiments at -80º C. The cell pellet was resuspended, counted with Trypan Blue, 1/3 of it was taken and after centrifuge lysed with Lysis buffer from RNeasy kit (Invitrogen) and frozen for gene expression analyses. The remaining were used for surface staining to perform total ILC analysis and total ILC sort in FACSAria III.

Surface staining and sorting of innate lymphoid cells

Isolated PBMCs and SFMCs were dispersed in wells of a 96-well plate as 1 × 10⁶ cells/well, in 25 µl stain buffer (PBS, supplemented with 2% FBS) incubated with 1µL Fc blocking/25 µL volume for 5 min and the staining antibodies were added. For surface staining, fluorochrome-conjugated antibodies specific to the surface antigen and stained for 30 min on ice in the dark. The cells were then centrifuged at 400 g for 3 min and washed twice with stain buffer. Total ILCs were sorted from PBMCs and SFMCs by FACSAria III (BD™) based on Eken et al. as Lineage^-CD161⁺CD127⁺ cells [27, 28]. After checking the sorting quality, sorted total ILCs were spun at 400 g for 5 min at 4in 500µL RPMI medium, the cells were transferred to lysis buffer from RNeasy kit (Invitrogen) for RNA isolation. The remaining PBMCs were centrifuged at 400 g for 5 min 4 °C and lysed with Trizol. Samples were stored at -80 °C until use.

The following antibodies were used staining of ILC cells. PE anti-human CD161 Antibody (Biolegend#339903), PE/Cy7 anti-human CD127 Antibody (Biolegend#560822), PE/Dazzle 594 anti-human CD117 (ckit) Antibody (Biolegend#313226), PerCP/Cy5.5 Rat anti-human CD294(CRTH2) Antibody (Biolegend#561660), FITC anti-human Lineage Cocktail (CD3, CD14, CD16, CD19, CD20, CD56) (BioLegend#348801). Lineage-CD161 + CD127+cKit+ cells were sorted as ILC3, Lineage-CD3-CD161 + CD127+cKit- cells were sorted as ILC1 and Lineage-CD3-CD161 + CD127+CRTH2 + cells sorted as ILC2, and Lineage-CD161 + CD127+ cells as total ILCs.

Real-time quantitative PCR

Due to the extremely low frequency of circulating ILCs and the limited RNA yield obtained after sorting, cells from two donors were pooled prior to RNA isolation for selected experiments, same procedure was applied to PBMC samples for consistency. Statistical analyses were performed using the pooled samples as independent experimental units. Pooled cells were lysed with Lysis buffer from RNeasy kit (Invitrogen). RNA isolation was performed with Invitrogen RNA Mini Kit following the manufacturer’s instructions. cDNA was synthesized with iScript cDNA synthesis kit (Bio-Rad) following the manufacturer’s instructions. LightCycler^® 480 (Roche) Device and SYBRGreen (Bio-Rad) method were used for the measurement of gene expression. Relative gene expression was calculated by the ^ΔΔCT method and normalized over the 18 S ribosomal RNA message.

The mRNA expression levels of genes were determined by using primer sequences specific for IFN-ɣ, IL-4, IL-5, IL-13, IL-22, GMCSF, TNF-α genes from cDNAs obtained from blood and TNF-α, IL-6, IL-1β, IL-17, IL-23p19, TSLP, IL-33 genes from cDNAs obtained from synovial cells (Supplementary Table S1).

Enzyme-linked immunosorbent assay

Cytokine levels in plasma and synovial fluid were measured using enzyme-linked immunosorbent assay (ELISA). IL-2, IL-6, IL-17, IL-22, GMCSF cytokine levels were measured in the plasma obtained from blood and synovial fluid. BioLegend Human IL-2 (#431804), IL-6 (#430504), IL-17 (#433914), IL-22 (#434504), GM-CSF (#432004) ELISA kit was used according to the protocols specified by the manufacturer. Additionally, IL-1b, IL-8, IL-10, IP-10, IFN-α2 cytokine levels were measured with the LEGENDPlex Multiplex ELISA kit (Human Anti-Virus Response Panel (13-plex) w/VbP BioLegend (#740390). The values below limit of detection (LOD) were entered as LOD value, or the lowest calculated value on LegendPlex Platform and data points were not omitted.

Statistical analysis

Data normality was assessed using the Shapiro-Wilk test. For pairwise comparisons, Student’s t-test was used for normally distributed data, and the Mann-Whitney U test was applied for non-normally distributed data. For multiple comparisons, one-way analysis of variance (ANOVA) with Tukey’s post-hoc test was used for normally distributed data, while the Kruskal-Wallis test with Dunn’s post-hoc correction was applied for non-normally distributed data. A p-value ≤ 0.05 was considered statistically significant. Flow cytometry data were analyzed using FlowJo software (version 10.0.6; BD Biosciences, San Jose, CA, USA). Statistical analyses were performed using GraphPad Prism software (version 6.0; GraphPad Software, San Diego, CA, USA). Outliers were identified using the ROUT method (Q = 1%). ROUT analysis was performed as a sensitivity analysis to assess the robustness of plasma cytokine findings rather than as a criterion for excluding biologically valid observations from the primary analyses.

Results

Clinical and laboratory findings

The study cohort comprised 53 pediatric participants followed at the Department of Pediatric Rheumatology, Erciyes University Faculty of Medicine: 15 with familial Mediterranean fever (FMF), 11 with spondyloarthritis (SpA), 13 with both FMF and SpA (FMF/SpA), and 14 age-matched healthy controls. Demographic, clinical, and laboratory characteristics, including treatments, are summarized in Supplementary Table S2. Age, sex, and age at FMF onset did not differ significantly across groups (p > 0.05). Clinical features of FMF were similar between FMF and FMF/SpA groups (p > 0.05). Similarly, SpA clinical features, showed no significant differences between SpA and FMF/SpA groups (p > 0.05) C-reactive protein (CRP) levels were significantly higher in the FMF group compared to SpA, FMF/SpA, and healthy control groups (p < 0.05).

Comparable number and frequency of total ILCs and their subtypes in the peripheral blood cells of patients with FMF, SpA, FMF/SpA patients, and healthy controls and in the synovial fluids of patients with SpA and FMF/SpA patients

PBMCs from FMF, SpA, FMF/SpA and healthy controls were stained with surface markers of ILC subtypes. The gating strategies of total ILCs and their subtypes are shown in Fig. 1A. The results showed that the percentages and absolute numbers of total ILCs and subtypes in the patients with FMF, SpA and FMF/SpA were similar to those in the control group (Fig. 1B, C, D, E, p > 0.05) suggesting the number and frequency of ILCs are stable across the disease groups and the healthy donors in the blood.

Fig. 1.

ILC subsets in peripheral blood of patients with FMF, SpA, FMF/SpA and healthy controls Comparable ILC subsets frequency and absolute number in peripheral blood of patients with FMF(n:15), SpA(n:11) and FMF/SpA (n:13) and healthy controls (n:14). (A) Gating strategy for total ILCs. Lineage: CD3/CD14/CD16/CD19/CD20/CD56. Percentage and absolute numbers of (B) total ILC (Lineage-CD161 + CD127+). (C) ILC1 (Lineage-CD161 + CD127+cKit-CRTH2-). (D) ILC2 (Lineage-CD161 + CD127+cKit-/+CRTH2+), (E) ILC3(Lineage-CD161 + CD127+cKit+CRTH2-), (p > 0.05)

Synovial fluid mononuclear cells were isolated from the patients, SpA(n:3) and FMF/SpA(n:4). ILCs subsets were analyzed in the synovial fluids. The gating strategy of ILCs and subtypes is shown in Fig. 2A. The results revealed an increasing trend for the absolute number of ILCs in patients with SpA compared to FMF/SpA patients, this did not reach significance (Fig. 2B, p > 0.05). ILC1,2,3 percentages and numbers were not significantly different between the two groups (Fig. 2C, D, E, p > 0.05) Collectively, these results suggested unchanged ILC number and frequency across healthy control, FMF, FMF/SpA and SpA in the blood and as well as in the synovial fluid of SpA and FMF/SpA patients.

Fig. 2.

Total ILCs and subsets in the synovial fluid of patients with SpA and FMF/SpA Total ILCs and subsets in the synovial fluid of patients with SpA(n:3) and FMF/SpA(n:4). (A) Gating strategy. Lineage: CD3/CD14/CD16/CD19/CD20/CD56. (B) total ILC (Lineage-CD161 + CD127+). (C) ILC1 (Lineage-CD161 + CD127+cKit-CRTH2-). (D) ILC2 (Lineage-CD161 + CD127+cKit-/+CRTH2+), (E) ILC3(Lineage-CD161 + CD127+cKit+CRTH2-), (p > 0.05)

Comparison of proinflammatory cytokines in the plasma of all study groups and in the synovial fluid of SpA and FMF/SpA patients

Levels of cytokines, including those associated with the activation or effector function of ILCs, in the plasma of FMF, SpA, FMF/SpA patients, and healthy controls, were measured by ELISA (Fig. 3). IL-6 concentration was higher in the FMF/SpA group compared to the control when all patients were included (Fig. 3B, *p < 0.05), however two patients expressed very high levels of IL-6 and marked as outlier with ROUT method (Q = 1%), thus, when removed from the cohort, IL-6 levels became comparable across the groups (Supplemental Fig. 1A). IL-8 concentration was higher in FMF/SpA group compared to all other groups when all patients in the cohort are included (Fig. 3C, *p < 0.05), similarly after ROUT exclusion, the levels marginally lost the significance (Supplemental Fig. 1B). IP-10 concentration was higher in FMF/SpA group compared to all other groups, when all patients are included (Fig. 3E, *p < 0.05) which also became comparable after outlier exclusion ((Supplemental Fig. 1C). IL-22 level was higher in FMF group compared to control and SpA groups (Fig. 3H, *p < 0.05), which became comparable after outlier removal with ROUT method (Supplemental Fig. 1D). In the healthy control group, compared to the FMF and FMF/SpA groups IL-2 levels were higher (Fig. 3I, *p < 0.05, ***p ≤ 0.001). IL-1β, IL-10, IFN-α2, GM-CSF, and IL-17 concentrations did not show any significant difference across the groups before and after outlier removal (Fig. 3A, D, F, G, J, p > 0.05). Together, these data point to considerable biological heterogeneity amongst patients with respect to cytokine levels in the plasma.

Fig. 3.

Proinflammatory cytokine production in the peripheral blood plasma of FMF, SpA, FMF/SpA patients and healthy control Proinflammatory cytokine production in the peripheral blood plasma of FMF(n:15), SpA(n:11), FMF/SpA(n:13) patients and healthy controls(n:14). Comparison of (A) IL-1β (pg/mL), (B) IL-6 (pg/mL) (*p < 0.05), (C) IL-8 (pg/mL) (*p < 0.05), (D) IL-10 (pg/mL), (E) IP-10 (pg/mL) (*p < 0.05), (F) IFN-α2(pg/mL), (G) GM-CSF (pg/mL), (H) IL-22(pg/mL) (*p < 0.05), (I) IL-2(pg/mL) (*p < 0.05), (J) IL-17(pg/mL) in blood plasma of patients with FMF, SpA and FMF/SpA with healthy control groups

Synovial fluid was collected from SpA and FMF/SpA patients for treatment with intra-articular steroids, and the screening of pro-inflammatory cytokines in synovial fluids (Fig. 4) revealed higher concentration of IP-10 by the SpA patient groups compared to the FMF/SpA patient group (Fig. 4C, *p < 0.05). Similarly, IL-17 levels produced by the SpA patient group showed an increase compared to FMF/SpA (Fig. 4K, *p < 0.05). IL-1β, TNF-α, IFN-ɣ, IL-6, IL-8, IFN-α2, GM-CSF, IL-22, and IL-2 did not show any significant differences between two groups (Fig. 4A, B, D, E, F, G, H, I, J, L, p > 0.05). Although the cellular source of the measured cytokines is unclear, the sample size is small, this result suggested increased group 3 ILC3 (IL-22, IL-17), and perhaps ILC2 (IL-6) activity [29, 30–32].

Fig. 4.

Proinflammatory cytokine production in the synovial fluid of SpA and FMF/SpA patients Proinflammatory cytokine production in the synovial fluid of SpA(n:3) and FMF/SpA(n:4) patients. Comparison of (A) IL-1β (pg/mL), (B) TNF-α (pg/mL), (C) IP-10 (pg/mL) (*p < 0.05), (D) IL-8 (pg/mL), (E) IFN- α2 (pg/mL), (F) GM-CSF (pg/mL), (G) IFN-β (pg/mL), (H) IFN-ɣ (pg/mL), (I) IL-6 (pg/mL) (J) IL-22 (pg/mL), K) IL-17 (pg/mL), L) IL-2 (pg/mL) in synovial fluid of patients with SpA and FMF/SpA

The gene expression of ILC-activating cytokines in PBMCs from FMF, SpA, FMF/SpA patients, and healthy controls

To better assess the state of ILC activity across patient groups, gene expression of ILC activating cytokines in total PBMCs from FMF, SpA and FMF/SpA patients and healthy controls were compared. The results revealed higher thymic stromal lymphopoietin (TSLP) gene expression in the SpA patient group compared to healthy controls and FMF patients (Fig. 5E, **p ≤ 0.01). The expression of IL17, IL23p19, IL1β, IL33, IL6, TGFβ genes did not show any significant difference between the groups (Fig. 5A, B, C, D, F, G, p > 0.05).

Fig. 5.

ILC-activating cytokine gene expression by PBMCs ILC-activating cytokine gene expression by PBMCs: Elevated TSLP expression. Comparison of (A) IL17, (B) IL23p19, (C) IL33, (D) TSLP (**p ≤ 0.01), (E) IL6, (F) TGFβ gene expression of pro-inflammatory cytokines in PBMCs from patients with FMF, SpA and FMF/SpA with healthy control group. Each dot represents a pooled PBMC sample from two distinct donors. Healthy controls(n: 5–6 pooled samples from 12 donors), FMF(n: 5–6 pooled samples from 12 donors), SpA(n: 5–6 pooled samples from 11 donors), FMF/SpA(n: 5–6 pooled samples from 12 donors) patients

Cytokine gene expression by sorted total ILCs across FMF, SpA, FMF/SpA patients, and healthy controls

Total ILCs (Lineage-CD161 + CD127+, Fig. 1A) were sorted from PBMCs of FMF, SpA and FMF/SpA patients and healthy controls via FACSAria III and gene expression of indicated cytokines in total ILCs from the patient and control groups was analyzed (Fig. 6). The results revealed an increase in gene expression of IL4 in the FMF/SpA group compared to the FMF group (Fig. 6A, *p < 0.05). An increase in the gene expression of IL-5 was observed in the FMF/SpA group compared to the control and FMF groups (Fig. 6B, *p < 0.05). GMCSF gene expression also showed increase in the FMF/SpA group compared to all groups (Fig. 6F, ***p < 0.001, ****p < 0.0001). Gene expression of IL13, IFNɣ, TNFα, IL22 did not show any significances between the groups (Fig. 6A, C, D, E, G, p > 0.05) Collectively, these results supported activated ILC2 signature in FMF/SpA group.

Fig. 6.

Increased gene expression of type 2 immunity associated effector cytokines by sorted peripheral blood total ILCs in FMF/SpA. (A) Comparison of IL4 (*p < 0.05), (B) IL5 (*p < 0.05), (C) IFNG, (D) TNFA, (E) GMCSF (***p < 0.001, ****p < 0.0001), (F) IL22, gene expression of total ILCs signature cytokines derived from PBMCs of FMF, SpA and FMF/SpA patients(*p < 0.05) Each dot represent a pool of two different donor derived sorted total ILC. Healthy controls(n: 5–7 pooled samples from 14 donors), FMF(n: 5–7 pooled samples from 14 donors), SpA(n: 3–4 pooled samples from 8 donors), FMF/SpA(n: 3–4 pooled samples from 8 donors) patients

Discussion

In this study, we performed a combined phenotypic and functional assessment of innate lymphoid cells (ILCs) in pediatric FMF, SpA, and overlap FMF/SpA during clinically active disease. Three main findings emerged. First, the frequency and absolute number of circulating total ILCs and ILC1/2/3 subsets were comparable among FMF, SpA, FMF/SpA, and healthy controls, and ILC subset distributions in synovial fluid were also similar between SpA and FMF/SpA. Second, despite stable ILC numbers, FMF/SpA was characterized by an enhanced type 2–skewed ILC transcriptional program, with increased IL4 and IL5 expression in sorted peripheral blood ILCs, together with increased GMCSF expression. Third, SpA patients exhibited higher PBMC TSLP expression and a synovial compartment enriched for IP-10 and IL-17, suggesting that local stromal/chemokine cues and type 3 inflammatory circuits may be particularly prominent in SpA compared with FMF/SpA.

Altered ILC composition and function have been described across immune-mediated inflammatory diseases, including systemic sclerosis, systemic lupus erythematosus, and rheumatoid arthritis, although the direction of change varies by tissue and disease stage [28, 30–36]. In SpA specifically, several reports emphasize tissue-enriched innate-like effector programs, including ILC3 expansion and IL-22/GM-CSF production in inflamed joints and mucosal sites [19–21, 39–41]. In contrast, our peripheral blood phenotyping did not reveal differences in total ILC or subset abundance among groups. This is consistent with the concept that ILC perturbations in inflammatory arthritis may be more apparent at the tissue level or at the level of activation state rather than absolute numbers, particularly when patients are sampled during treatment and at variable points along the inflammatory trajectory [19–21, 37, 38]. The limited number of synovial fluid samples however supports interpreting these observations with appropriate caution.

A key observation in our study is that FMF/SpA exhibited increased expression of IL4 and IL5 in sorted peripheral blood ILCs, consistent with enhanced ILC2-like effector programming despite unchanged ILC2 numbers. ILC2s can be activated by epithelial/stromal “alarmins” and cytokines (including IL-33, IL-25, and TSLP) and rapidly produce type 2 cytokines that shape local inflammation and crosstalk with regulatory pathways [13–16]. In experimental arthritis, increasing ILC2 activity can attenuate disease and protect from bone destruction, and IL-9–producing ILC2s have been implicated in resolution programs [17, 18, 42–44]. While these studies support a potential anti-inflammatory or tissue-protective role for type 2 innate programs in some settings, type 2 cytokines are context-dependent and should not be interpreted as uniformly anti-inflammatory. In particular, the concurrent increase in GMCSF expression in FMF/SpA ILCs is notable, as GM-CSF has been linked to myeloid activation and inflammatory amplification in arthritides, and ILC3-like GM-CSF programs have been reported in SpA joints [19]. Together, our data suggest that overlap FMF/SpA may reflect a distinct immune activation state in which ILC effector programming shifts toward type-2 cytokines (IL-4/IL-5) while also displaying features that could support inflammatory myeloid circuits (GM-CSF), rather than simply representing an additive combination of FMF and SpA pathways.

We also observed a compartmentalized cytokine/chemokine pattern that differed between SpA and FMF/SpA. In plasma, FMF/SpA showed increased IL-6, IL-8, and IP-10 compared with other groups, suggesting heightened systemic inflammatory signaling. In synovial fluid, SpA patients displayed higher IP-10 and IL-17 compared with FMF/SpA, although the results require confirmation with larger cohort size. IP-10/CXCL10 is a chemokine induced downstream of interferon-driven inflammation and can promote recruitment of CXCR3-expressing lymphocytes into inflamed tissues [45]. Consistent with this concept, synovial CXCR3/CXCL10 axis activity has been demonstrated in pediatric inflammatory arthritis [22]. The coexistence of increased synovial IP-10 and IL-17 in SpA is also compatible with a local inflammatory circuit in which chemokine gradients support immune cell accumulation alongside type 3 effector programs that are central to SpA pathogenesis [9–12]. Although our study does not identify the cellular source of synovial IP-10 or IL-17, these findings raise the possibility that chemokine-driven recruitment and IL-17–dominated effector pathways may be more pronounced in SpA, whereas FMF/SpA may display a different balance between systemic cytokines and joint-local inflammation.

Although plasma IL-6, IL-8, IP-10, and IL-22 differed significantly in the complete cohort, these differences were no longer statistically significant in the sensitivity analysis following ROUT-identified outlier exclusion. We therefore interpret these findings as reflecting biological heterogeneity rather than uniform disease-wide cytokine signatures. Notably, the ROUT algorithm identified up to three patients in some groups, corresponding to approximately 30% of the study population. In relatively small cohorts of rare inflammatory diseases, exclusion of such a substantial proportion of patients may inadvertently remove biologically meaningful heterogeneity rather than true experimental outliers, particularly in the absence of technical evidence supporting measurement error. Increasing evidence indicates that many immune-mediated diseases comprise distinct molecular endotypes characterized by differential cytokine expression, including type 2-high asthma, interferon-high systemic lupus erythematosus, and cytokine-defined subsets of rheumatoid arthritis [46–48]. Consequently, markedly elevated cytokine concentrations observed in a subset of patients should not automatically be considered statistical artifacts in the absence of technical evidence supporting measurement error, but may instead represent biologically distinct inflammatory subgroups. Nevertheless, given the limited sample size of our study, these findings should be interpreted with caution and require validation in larger, independent cohorts to determine whether these high cytokine-producing patients represent reproducible disease endotypes.

Another important consideration when interpreting the plasma cytokine findings is the potential influence of ongoing immunomodulatory therapy. All patients in the SpA and FMF/SpA groups received NSAIDs, while subsets were also treated with corticosteroids, methotrexate, TNF inhibitors (adalimumab or etanercept), or IL-1 inhibitors for FMF. These therapies are known to modulate systemic inflammatory cytokines and may also influence innate lymphoid cell function. Consequently, circulating cytokine measurements, particularly IL-6, IL-8, IP-10, and IL-22, are likely to be the findings most susceptible to treatment-related effects. Likewise, the observed alterations in TSLP and ILC-associated gene expression may have been influenced by immunomodulatory therapy. In contrast, the relatively modest differences observed in circulating ILC subset frequencies are less likely to be explained solely by treatment. Because subgroup analyses according to treatment were not feasible owing to the limited sample size and heterogeneous treatment regimens, residual treatment-related confounding cannot be excluded and should be considered when interpreting these findings.

Finally, PBMC gene expression analyses showed increased TSLP expression in SpA. TSLP is a stromal/epithelial-derived cytokine that can support type 2 polarization and influence innate lymphocyte activation [13, 16]. In our cohort, increased PBMC TSLP expression did not coincide with clear numerical expansion of ILC2s in blood or synovial fluid, but it may reflect broader activation of alarmin-responsive pathways that could shape the inflammatory milieu. Future studies that directly measure TSLP protein levels in plasma/synovial fluid, assess TSLP receptor expression on ILC subsets, and integrate tissue-level sampling will be needed to determine whether TSLP is mechanistically linked to ILC activation or represents a biomarker of stromal activation in pediatric SpA.

This study has limitations. Patients were sampled during flares, but many were receiving anti-inflammatory and/or biologic therapies, which may modulate cytokine profiles and ILC activation states. Synovial fluid sample size was limited, reducing power for joint-compartment comparisons. Additionally, some transcriptional analyses used pooled samples, which reduced the ability to resolve inter-individual heterogeneity. Lastly, the due to limited sample size, further analyzing patients with high and low plasma cytokine expression could not be done.

Conclusion

Despite these constraints, our data support the conclusion that pediatric FMF/SpA is associated with altered ILC activation (type 2 cytokine signature) without overt numerical changes, and that SpA exhibits joint-local enrichment of IP-10 and IL-17 along with increased PBMC TSLP expression. Larger, longitudinal cohorts with deeper tissue-informed profiling will be important to define whether these immune signatures track with disease activity, treatment response, and clinical outcomes, and whether the IP-10/CXCL10 axis or innate lymphoid effector programs represent actionable therapeutic nodes in pediatric SpA and FMF/SpA.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by [Ayşenur Paç Kısaarslan], [Zehra Büşra Azizoğlu], [ Büşra Şeniz Demir], [Sümeyra Özdemir Çiçek], [Hakan Poyrazoğlu], and [Ahmet Eken]. The first draft of the manuscript was written by [Ayşenur Paç Kısaarslan] and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This study was supported in part by the Erciyes University BAP TDK-2022-11916 and TUBA GEBIP-2021, TUBITAK TESVIK 2024 Funds to AE.

Data availability

The data underlying this article are available in the article and in its online supplementary material. All the raw data generated are available upon reasonable request from the corresponding authors Aysenur Pac Kisarslan and Ahmet Eken authors.

Tables

Competing interests

The authors declare no competing interests.

Ethics approval

The study was approved by the Erciyes University Clinical Research Ethics Committee (Decision No: 2022/91) and adhered to the Declaration of Helsinki.

Consent to participate

Written informed consent was obtained from the parents and patients.

Consent to publish

The authors affirm that human research participants provided informed consent for publication.