Toll interacting protein (TOLLIP) serum concentration in patients with systemic sclerosis (SSc), with and without interstitial lung disease (ILD) Authors: Niels Schröder, Jitka Andrä, Diana Knuth-Rehr, Silke Leja, Nicolas Hunzelmann, Eda Börner, Kimberly Barbet, Francesco Bonella Published: 2026 ABSTRACT AbstractBackgroundSingle nucleotide polymorphisms (SNPs) within Toll interacting protein (TOLLIP) coding gene have been associated with progression, prognosis and treatment response in idiopathic pulmonary fibrosis (IPF). These SNPs seem to influence mRNA expression and serum levels of TOLLIP. The aim of this study was to investigate TOLLIP serum concentrations in a previously genotyped cohort of patients with systemic sclerosis (SSc) with and without interstitial lung disease (ILD).Patients and methods106 consecutive patients with SSc (77 female) with (n=53) and without ILD (n=53) and 60 healthy controls (HCs) were included. TOLLIP concentrations were measured using a commercially available ELISA kit (Cloud-Clone, USA) in serum samples from patients and HC previously genotyped for two SNPs within TOLLIP (rs3750920, rs5743890).ResultsPatients with SSc had tendentially lower TOLLIP serum concentrations than HC (0.111 (95% CI 0.101 to 0.123) vs 0.141 (95% CI 0.104 to 0.191) ng/µL, p=0.076). Patients with SSc-ILD showed significantly higher TOLLIP levels than patients with SSc without ILD (0.126 (95% CI 0.106 to 0.151) vs 0.098 (95% CI 0.091 to 0.105) ng/µL, p=0.010). At a cut-off value of ≥0.1135 ng/µL serum TOLLIP yielded a 71% sensitivity and 74% specificity for identifying ILD (AUC=0.74, 95% CI 0.593 to 0.889, p=0.007). Serum TOLLIP concentrations did not differ significantly across rs3750920 or rs5743890 genotypes. TOLLIP concentration inversely correlated with age (r=−0.256, p<0.001), whereas no significant associations with gender, body mass index, C-reactive protein, forced vital capacity or diffusing capacity of the lung for carbon monoxide at time of sampling were observed.ConclusionPatients with SSc show lower TOLLIP serum concentrations than HC, while patients with SSc-ILD exhibit higher TOLLIP levels than those without. TOLLIP might serve as a biomarker of ILD in patients with SSc.Keywords: Interstitial Fibrosis, Rare lung diseases, Patient Outcome Assessment, Innate Immunity WHAT IS ALREADY KNOWN ON THIS TOPIC Toll interacting protein (TOLLIP) regulates innate immune and fibrotic pathways. Toll interacting protein (TOLLIP) regulates innate immune and fibrotic pathways. Genetic variants in TOLLIP have been linked to disease susceptibility and outcome in idiopathic pulmonary fibrosis (IPF). Genetic variants in TOLLIP have been linked to disease susceptibility and outcome in idiopathic pulmonary fibrosis (IPF). In a previous study using the same patient cohort, we investigated TOLLIP genetic variants in systemic sclerosis (SSc), showing an association between genotype and progression of interstitial lung disease (ILD). In a previous study using the same patient cohort, we investigated TOLLIP genetic variants in systemic sclerosis (SSc), showing an association between genotype and progression of interstitial lung disease (ILD). Circulating TOLLIP protein levels in SSc are unknown. Circulating TOLLIP protein levels in SSc are unknown. WHAT THIS STUDY ADDS Serum TOLLIP levels are tendentially reduced in patients with SSc compared with healthy controls. Serum TOLLIP levels are tendentially reduced in patients with SSc compared with healthy controls. Within the SSc population, patients with SSc-associated interstitial lung disease (SSc-ILD) exhibit higher serum TOLLIP levels than those without ILD, suggesting a compensatory response to lung injury. Within the SSc population, patients with SSc-associated interstitial lung disease (SSc-ILD) exhibit higher serum TOLLIP levels than those without ILD, suggesting a compensatory response to lung injury. TOLLIP levels are further influenced by smoking status and immunosuppressive therapy. TOLLIP levels are further influenced by smoking status and immunosuppressive therapy. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY Serum TOLLIP may serve as a biomarker to improve risk stratification in patients with SSc with and without ILD. Serum TOLLIP may serve as a biomarker to improve risk stratification in patients with SSc with and without ILD. Introduction Systemic sclerosis (SSc) is a rare autoimmune connective tissue disease characterised by inflammation, vasculopathy and fibrosis. Pulmonary complications, namely, interstitial lung disease (ILD) and pulmonary hypertension,1 are still the leading causes of death.2 SSc-ILD is quite heterogeneous, with some patients experiencing a stable course over many years while others develop progressive lung function decline. The pathogenesis of fibrotic ILDs involves a complex interplay of innate and adaptive immune responses, with chronic inflammation driven by multiple immune cell subpopulations and molecular pathways that perpetuate tissue remodelling and fibrosis.3 There are numerous risk factors associated with disease progression and mortality, such as decline in forced vital capacity (FVC) and diffusing capacity of the lung for carbon monoxide (DLCO) over 1–2 years.4 5 Established predictors for ILD progression and mortality also include older age, male sex, diffuse cutaneous systemic sclerosis (dcSSc) and positivity for anti-topoisomerase I antibodies.6,8 Nevertheless, predicting ILD progression and mortality remains a challenge, emphasising the need for novel biomarkers to improve risk stratification.9 Toll interacting protein (TOLLIP) is a multifunctional protein and significant intracellular regulator of toll-like receptors (TLR) 2 and 4, modulating immune responses and preventing excessive inflammation.10 By inhibiting interleukin (IL)-1 receptor-associated kinase-1 (IRAK-1) activation and nuclear factor kappa B (NF-κB) signalling, it downregulates inflammatory cytokines such as IL-6, IL-13, tumour necrosis factor-alpha (TNF-α) and transforming growth factor-beta (TGF-β).11,13 Beyond its role in the innate immune regulation, TOLLIP is implicated in idiopathic pulmonary fibrosis (IPF) and a range of other conditions, including Parkinson’s, Alzheimer’s disease and inflammatory bowel disease.14,18 Emerging evidence from transcriptomic studies suggests that circulating monocyte subsets, particularly CD14+CD163−HLA-DRlow monocytes, are associated with disease progression and mortality in IPF, highlighting the importance of innate immune dysregulation in fibrotic lung diseases.19 Multiple genome-wide association studies have identified associations between TOLLIP single nucleotide polymorphisms (SNPs), located on chromosome 11, and different diseases like IPF and rheumatoid arthritis-associated interstitial lung disease (RA-ILD).20,23 These SNPs influence TOLLIP mRNA expression, leading to either an increase or decrease in TOLLIP serum concentrations. Additionally, certain non-synonymous SNPs (nsSNPs), for example, R28Q (rs149174945) and T40M (rs764227120), compromise TOLLIP protein stability and function independent of mRNA expression levels, through structural alterations that affect protein folding, binding affinity, and intracellular signalling capacity.24 The minor allele (C) of rs5743890 has been associated with a 20% reduction in TOLLIP mRNA levels in lung tissue of patients with IPF. Rs3750920 genotype has been associated with differential treatment response to N-acetylcysteine in IPF.17 20 25 In a previous study, we investigated the distribution and clinical relevance of two common TOLLIP SNPs (rs3750920 and rs5743890) in patients with SSc, demonstrating associations with ILD, disease progression and survival.26 The aim of this study was to investigate TOLLIP serum concentrations in patients with SSc with and without ILD, and to correlate these protein levels with TOLLIP genotype data previously obtained from the same patient cohort. This integrated genotype-phenotype approach allows for a comprehensive assessment of how genetic variants influence circulating TOLLIP protein levels. Methods Study subjects We retrospectively studied 106 patients with SSc (77 female) with (n=53) and without (n=53) ILD. The patients were recruited at the University Hospital Cologne from 2004 to 2022. Diagnosis of SSc was confirmed based on the 2013 American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) classification criteria.27 Presence of ILD was confirmed according to the most recent American Thoracic Society (ATS)/European Respiratory Society (ERS) guidelines.28 The 60 healthy subjects (35 female) were recruited at the University Hospital Essen. These healthy controls (HCs) between the ages of 30 and 70 years did not suffer from any lung disease and did not show any history of cancer, autoimmune diseases or immunodeficiency. The HC group included in the present analysis represents a subset of the HCs used in our previously published study of TOLLIP SNPs.26 The selection was based on the availability of serum samples. All included subjects provided written informed consent. Clinical definitions Disease progression of ILD was defined as either ≥10% relative decline in FVC% predicted or ≥5 to <10% decline in relative FVC% predicted and 15% relative decline in DLCO% predicted.29 Otherwise, the patients were defined as stable. We did not include death in the definition of progression since it was not possible to retrieve the cause of death (respiratory vs non-respiratory). Time to progression was calculated from the first available lung function test to the timepoint when progression criteria were fulfilled. Patients without at least two follow-up lung function assessments were excluded from progression analysis. Survival was defined as the time from date of first hospital admission to the date of death or last available follow-up. Disease duration was defined as timeframe between diagnosis of SSc and blood sampling. The follow-up time was calculated from the first registered hospital admission (baseline). A detailed description of DNA extraction, genotyping and pulmonary function measurement can be found in the previously published study from our group.26 TOLLIP serum concentrations Serum concentrations of TOLLIP were measured using a commercially available ELISA kit (SEC803Hu, Cloud-Clone, Houston, Texas, USA), following the manufacturer’s instructions. All samples showing detectable signals were analysed in duplicate, randomly selected samples were additionally measured for a third time to ensure accuracy and reproducibility of the results. Concentrations were determined by comparing the measured absorption with a standard curve provided with the kit. Statistical analysis Normal distribution for continuous variables was evaluated using the Kolmogorov-Smirnov test. Due to the skewed distribution of TOLLIP serum concentrations, data were log-transformed for statistical analysis and are presented as geometric mean with 95% CI. Parametric data are presented as mean±SD. Categorical variables are presented as either a percentage of the total or numerically. Continuous variables were analysed with the unpaired t-test when comparing two groups. Analysis of variance (ANOVA) was used to determine correlations when comparing more than two groups. To compare categorical variables, χ2 test and Fisher’s exact test were used. Spearman’s or Pearson’s correlation coefficient was calculated to assess correlations. Allele, genotype and haplotype frequencies were calculated by using SNPStats (Catalan Institute of Oncology, Spain). Pair-wise linkage disequilibrium between both SNPs was also calculated by using SNPStats. χ2 test and Fisher’s exact test were used to test for deviation from Hardy-Weinberg equilibrium (HWE). A cut-off value for serum TOLLIP was determined using receiver operating characteristic (ROC) curve analysis to evaluate the diagnostic accuracy for discriminating patients with SSc with ILD from those without ILD. The area under the ROC curve (AUROC) was calculated as a measure of overall discriminatory ability. The optimal cut-off value was determined using the Youden index, which identifies the threshold that maximises the sum of sensitivity and specificity without preferentially weighting either parameter. Sensitivity and specificity were calculated at the optimal cut-off point. This cut-off was subsequently used to stratify patients into TOLLIP-high and TOLLIP-low groups for analyses of disease progression and mortality. Dichotomisation of TOLLIP concentrations was performed to facilitate clinical interpretation and risk stratification. To assess the robustness of this approach, sensitivity analyses were conducted comparing Cox regression models using TOLLIP as a dichotomised variable versus a continuous variable. This was considered necessary because the narrow distribution of TOLLIP concentrations and the substantial proportion of imputed values below the limit of detection resulted in unstable estimates when TOLLIP was modelled continuously. Univariable and multivariable Cox proportional hazard regression were performed to examine the individual impact of age, gender, smoking history, medication, pulmonary hypertension, serum TOLLIP concentrations and pulmonary function tests on the disease course. Covariates were selected a priori based on established clinical predictors of survival and disease progression in SSc-ILD, including demographic factors (age, gender), smoking history, baseline lung function parameters (FVC% pred., DLCO% pred.), presence of pulmonary arterial hypertension and immunosuppressive therapy. A backward stepwise selection procedure was applied to identify independent predictors in the final multivariable model. Kaplan-Meier analysis with log-rank test was used to assess progression and overall survival based on the determined TOLLIP cut-off value and SNP genotypes. P values <0.05 were considered statistically significant. Survival was calculated from the time of first admission to the hospital. All statistical analyses were performed using SPSS V.29.0 (IBM, Armonk, New York, USA) provided by the University of Duisburg-Essen. Patient and public involvement No patients or members of the public were involved in the design, or conduct, or reporting or dissemination plans of our research, since the present study was designed as a retrospective analysis from a previously genotyped cohort, which limited the opportunity for prospective patient involvement. Results Characteristics of study subjects Demographics and clinical characteristics of the studied subjects are shown in table 1. At baseline, age (60±12.7 vs 46±11.6, p<0.01) and smoking history (40/66 vs 9/51, p<0.01) differed significantly between patients with SSc and HCs. Table 1Demographics and characteristics of the studied subjects at baselineVariablesAll SSc (n=106)SSc-non-ILD (n=53)SSc-ILD (n=53)HC(n=60)Age, years60±12.760±13.6*60±11.946±11.6Gender, female77 (73%)43 (81%)34 (64%)35 (58%)BMI, kg/m2†24 (21.6–26.4)23.6 (21.5–25.7)24.3 (22.1–27.5)25 (23.2–26.7)Smoking history, yes/no40/66*19/3421/329/51Pack years, N27±2232±2620±1017±6FVC, % pred.86±22100±17‡73±17–FEV1, % pred.84±2095±14‡72±18–DLCO, % pred.65±1972±17 ‡58±19–Antifibrotic treatment, yes21 (20%)0 (0%)‡21 (40%)–Immunosuppressive treatment, yes52 (50%)15 (28%)‡36 (69%)–CVD, yes14 (13%)6 (11%)8 (15%)–PAH, yes30 (28%)8 (15%)§22 (42%)–TOLLIP, ng/µL¶0.111 (0.101–0.123)0.098 (0.091–0.105)0.126 (0.106–0.151)0.141 (0.104–0.191)CRP, mg/dL†0.24 (0.1–0.6)*0.13 (0.08–0.29)§0.38 (0.2–0.94)0.10 (0.10–0.10)KL-6, U/mL†546 (346–1092)267 (251 – 346)‡843 (521–1392)–Time to progression, months†––92 (53–148)–Time to death or loss to follow-up, months†101 (53–157)84(26 – 149)130(71 – 176)–Otherwise indicated, values are expressed as mean±SD.*p<0.01 versus HC†Median, IQR‡p<0.001 versus SSc-ILD.§p<0.01 versus SSc-ILD.¶Geometric mean (95% CI).BMI, body mass index; CRP, C-reactive protein; CVD, cardiovascular diseases; DLCO, diffusion capacity of the lung for carbon monoxide; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; HC, healthy controls; ILD, interstitial lung disease; KL-6, Krebs von den Lungen-6; PAH, pulmonary arterial hypertension; SSc, systemic sclerosis; TOLLIP, Toll interacting protein. Demographics and characteristics of the studied subjects at baseline Otherwise indicated, values are expressed as mean±SD. p<0.01 versus HC Median, IQR p<0.001 versus SSc-ILD. p<0.01 versus SSc-ILD. Geometric mean (95% CI). BMI, body mass index; CRP, C-reactive protein; CVD, cardiovascular diseases; DLCO, diffusion capacity of the lung for carbon monoxide; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; HC, healthy controls; ILD, interstitial lung disease; KL-6, Krebs von den Lungen-6; PAH, pulmonary arterial hypertension; SSc, systemic sclerosis; TOLLIP, Toll interacting protein. The median follow-up time of patients with SSc was 101 (IQR 53–157) months. TOLLIP serum concentrations in the studied subjects The distribution of serum TOLLIP concentrations across all studied subgroups is shown in figure 1. From the 166 analysed samples, detectable signals enabling concentration determination were obtained for 61 samples. To ensure measurement accuracy and reproducibility, 50 of these samples were measured a second time and an additional 10 samples underwent a third measurement. For the remaining samples (n=105), whose absorption values fell below the assay detection limit, values were imputed by using the lowest detectable concentration above the detection threshold. Serum TOLLIP levels in patients with systemic sclerosis (SSc). Distribution of serum Toll interacting protein (TOLLIP) levels in all patients with SSc compared with healthy controls (A) and in patients with SSc, stratified by presence or absence of interstitial lung disease (ILD) (B). (C–D) Distribution of serum TOLLIP levels in all patients stratified by rs3750920 (C) and rs5743890 (D) genotype. Data are presented on a logarithmic scale. Dots are single measurements; orange crosses represent mean values and orange lines represent median values. Significance of the comparison is shown in the graphic. Groups were compared using unpaired t-test on log-transformed data (A–B) or one-way analysis of variance on log-transformed data (C–D). Patients with SSc tended to have lower TOLLIP serum concentrations than HCs (0.111 (95% CI 0.101 to 0.123) vs 0.141 (95% CI 0.104 to 0.191) ng/µL, p=0.076) (figure 1A). Patients with SSc-ILD showed significantly higher TOLLIP levels than patients with SSc without ILD (0.126 (95% CI 0.106 to 0.151) vs 0.098 (95% CI 0.091 to 0.105) ng/µL, p=0.010) (figure 1B). Influence of TOLLIP SNP genotype on TOLLIP serum concentration Data on genotype and allele frequencies, HWE, linkage disequilibrium and haplotype structure have been described in a previous publication from our group.26 The characteristics of patients with SSc stratified by TOLLIP rs3750920 and rs5743890 genotype are summarised in online supplemental tables S1 and S2. Figure 1 shows the serum TOLLIP distribution of all patients in relation to rs3750920 (figure 1C) and rs5743890 (figure 1D) genotype. No significant differences were observed. TOLLIP serum concentrations and correlation with clinical characteristics We did not observe any correlation with gender, body mass index (BMI), serum C-reactive protein (CRP), serum Krebs von den Lungen-6 (KL-6) or lung function tests. Neither Pearson nor Spearman correlation analyses showed a significant correlation between serum TOLLIP concentrations and baseline FVC % pred. or DLCO % pred. Serum TOLLIP levels weakly correlated with age (r=−0.256, p<0.001). Additionally, serum TOLLIP concentrations were significantly higher in patients under immunosuppressive treatment than in those without (0.125 (95% CI 0.105 to 0.149) vs 0.100 (95% CI 0.091 to 0.109) ng/µL, p=0.023) and tended to be higher in non-smokers compared with current or former smokers (0.128 (95% CI 0.108 to 0.151) vs 0.106 (95% CI 0.091 to 0.123) ng/µL, p=0.090). No significant correlation was observed between TOLLIP serum levels and the decline of FVC% pred. or DLCO% pred. over 1 year. When patients were stratified by median TOLLIP concentration (0.088 ng/µL), no significant differences in baseline FVC% pred. or DLCO % pred., nor in their annual decline, were observed between groups (online supplemental table S3). The lack of significant associations may in part be related to the limited sample size. Serum TOLLIP cut-off value for the identification of ILD We conducted a receiver operating characteristic (ROC) analysis to assess the association between serum TOLLIP and ILD. At a threshold of ≥0.1135 ng/µL, SSc with and without ILD could be separated with a sensitivity of 71% and specificity of 74% (AUC=0.74, 95% CI 0.593 to 0.889, p=0.007) (figure 2A–B). Using this cut-off, 17 of 22 patients (77%) in the TOLLIP-high group had ILD compared with 36 of 84 patients (43%) in the TOLLIP-low group (p=0.004), confirming the discriminatory capacity of the identified threshold. Diagnostic and prognostic performance of serum Toll interacting protein (TOLLIP). Receiver operating characteristic (ROC) curve for identifying interstitial lung disease in the entire study population (A) and corresponding performance metrics (B). (C) Kaplan-Meier survival curve for patients with systemic sclerosis (SSc) stratified by the cut-off value for serum TOLLIP concentrations comparing TOLLIP-low with TOLLIP-high group. Shaded areas represent 95% CIs and tick marks indicate censoring. SSc Patients were stratified into two groups based on the identified TOLLIP cut-off value: TOLLIP-low (< 0.1135 ng/µL, n=84) and TOLLIP-high (≥ 0.1135 ng/µL, n=22) group (table 2). We compared demographics and clinical characteristics of the two groups. Table 2Clinical characteristics of patients with SSc at baseline stratified by TOLLIP cut-off valueVariableTOLLIP-low (< 0.1135 ng/µL) (n=84)TOLLIP-high (≥ 0.1135 ng/µL) (n=22)P valueAge, years61±1357±120.146Gender, female59 (70%)18 (82%)0.278*BMI, kg/m224.6±3.923.8±4.60.417Smoking history, yes/no43/896/280.255*Pack years (n=21)29±2215±120.304FVC, % pred.89±2177±22 0.025 FEV1, % pred.86±2076±190.056DLCO, % pred.65±1963±190.653FVC decline, % pred./year−1.8±4.8−1.2±5.80.581FEV1 decline, % pred./year−1.5±4.2−1.1±4.20.710DLco decline, % pred./year−1.3±4.6−2.17±3.10.399PAH, yes23 (27%)7 (32%)0.681*Antifibrotic treatment, yes17 (20%)5 (23%)0.798*Immunosuppressive treatment, yes35 (42%)16 (73%) 0.009* CVD, yes13 (15%)1 (5%)0.178*ILD, yes36 (43%)17 (77%) 0.004* KL-6, U/mL521 (320–1002)713 (529–1266)0.151Otherwise indicated, values are expressed as mean±SD. Bold values indicate statistical significance (p<0.05).*Calculated by using χ2 test.†Median, IQR.BMI, body mass index; CVD, cardiovascular diseases; DLCO, diffusion capacity of the lung for carbon monoxide; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; ILD, interstitial lung disease; KL-6, Krebs von den Lungen-6; PAH, pulmonary arterial hypertension; SSc, systemic sclerosis. Clinical characteristics of patients with SSc at baseline stratified by TOLLIP cut-off value Otherwise indicated, values are expressed as mean±SD. Bold values indicate statistical significance (p<0.05). Calculated by using χ2 test. Median, IQR. BMI, body mass index; CVD, cardiovascular diseases; DLCO, diffusion capacity of the lung for carbon monoxide; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; ILD, interstitial lung disease; KL-6, Krebs von den Lungen-6; PAH, pulmonary arterial hypertension; SSc, systemic sclerosis. At baseline, patients in the TOLLIP-high group had significantly lower FVC% pred. (p=0.025). Furthermore, immunosuppressive treatment was more frequently administered in the TOLLIP-high group (p=0.009). No significant correlation was observed between groups regarding gender distribution, smoking history or the rate of pulmonary function decline over 1 year. The median serum KL-6 concentration tended to be higher in the TOLLIP-high group. As a sensitivity analysis, we stratified patients according to the median TOLLIP serum concentration (0.088 ng/µL) and compared clinical characteristics between groups (online supplemental table S3). Patients with detectable TOLLIP levels (≥0.088 ng/µL) showed no significant differences in age, gender, BMI, smoking status or lung function parameters compared with those with undetectable levels. A trend towards higher KL-6 levels and a slightly increased proportion of ILD was observed in the group with TOLLIP levels beyond the median. Survival and progression analysis Median survival of all patients with SSc was 101 (IQR 53–157) months, 130 (IQR 71–176) months in patients with SSc-ILD and 84 (IQR 26–149) months in those without ILD. Nine (9%) patients died until end of follow-up. There were no deaths in the group of patients with SSc-non-ILD and in the ILD group without ILD progression. Figure 2C shows the survival curve of all patients with SSc categorised into TOLLIP-low and TOLLIP-high groups. A trend toward shorter survival was observed in patients with SSc with high TOLLIP concentrations compared with those with low (158 vs 213 months, p=0.057). Disease progression occurred in 30 (57%) out of 53 patients with SSc-ILD at baseline. Median time to progression was 92 months (IQR 53–148). There were no differences in terms of demographics and clinical characteristics between patients who had progression or whose who remained stable at baseline (data not shown). Patients with low TOLLIP serum concentrations (<0.1135 ng/µL) showed a trend towards shorter progression-free survival with a higher rate of disease progression compared with the TOLLIP-high group (36% vs 19%, p=0.192, respectively) (online supplemental figure S1). Univariable and multivariable analysis for predictors of survival and progression We conducted univariable and multivariable Cox regression analysis to identify predictors of survival. After adjustment for smoking history, BMI, DLCO % pred. and pulmonary arterial hypertension, higher serum TOLLIP concentrations were associated with shorter survival time (HR 8.35, 95% CI 1.52 to 45.83, p=0.015) (table 3). In sensitivity analysis using TOLLIP as a continuous variable, this association remained consistent (HR 12.85, 95% CI 0.93 to 177.74, p=0.057) (online supplemental table S4). Table 3Univariable and multivariable Cox regression for predictors of survival in patients with SSc-ILDVariablesβHR95% CIP valueUnivariable analysisAge, years (continuous)0.1571.171.03 to 1.330.017Gender, (ref. male)−3.3580.040.001 to 1.210.064BMI, kg/m2 (continuous)−0.1550.860.57 to 1.280.452Smoking history (ref. yes)0.0211.020.12 to 9.110.985FVC, % pred. (continuous)0.0371.040.98 to 1.100.201DLCO, % pred. (continuous)−0.0270.970.91 to 1.040.424TOLLIP-high group (ref. yes)1.9276.871.06 to 44.690.044PAH (ref. yes)0.4471.560.31 to 7.830.586Immunosuppressive treatment (ref. yes)−1.4010.250.04 to 1.390.113Multivariable analysis*Age, years (continuous)0.1471.161.05 to 1.280.005Gender (ref. male)−3.3520.0350.003 to 0.430.009FVC, % pred. (continuous)0.0421.040.99 to 1.100.097TOLLIP-high group (ref. yes)2.1228.351.52 to 45.830.015Immunosuppressive treatment (ref. yes)−1.3870.250.05 to 1.280.127*Model obtained by using backward conditional stepwise regression (five steps), including BMI, smoking history, DLCO (% pred.) and PAH as covariates.BMI, body mass index; DLCO, diffusion capacity of the lung for carbon monoxide; FVC, forced vital capacity; PAH, pulmonary arterial hypertension; SSc-ILD, systemic sclerosis-associated interstitial lung disease; TOLLIP, Toll interacting protein. Univariable and multivariable Cox regression for predictors of survival in patients with SSc-ILD Model obtained by using backward conditional stepwise regression (five steps), including BMI, smoking history, DLCO (% pred.) and PAH as covariates. BMI, body mass index; DLCO, diffusion capacity of the lung for carbon monoxide; FVC, forced vital capacity; PAH, pulmonary arterial hypertension; SSc-ILD, systemic sclerosis-associated interstitial lung disease; TOLLIP, Toll interacting protein. For ILD progression, univariable and multivariable Cox regression analyses were performed in patients with SSc-ILD at baseline. We did not find any association between higher or lower TOLLIP levels and disease progression (online supplemental table S5). Discussion To the best of our knowledge, this is the first study investigating TOLLIP concentration in serum of patients with SSc. This analysis was performed as a follow-up investigation in the same patient cohort that had been previously genotyped for TOLLIP SNPs by our group, enabling an integrated genotype–phenotype assessment.26 We found that TOLLIP levels tended to be lower in patients with SSc compared with HCs, but elevated in patients with ILD compared with those without. Furthermore, serum TOLLIP concentrations did not differ significantly across rs3750920 or rs5743890 genotypes. A cut-off value of ≥0.1135 ng/µL showed moderate diagnostic accuracy for ILD, although no significant association with disease progression was observed. TOLLIP is predominantly expressed in immune cells and alveolar type I epithelial cells.30 In patients with IPF, an overall reduction of TOLLIP expression in lung tissue has been observed.20 Interestingly, despite this general downregulation, increased TOLLIP expression has been reported in atypical epithelial cells within fibrotic lungs, although the significance of this overexpression remains unclear.23 Given the increasing evidence that a tight homeostatic regulation of TOLLIP may be critical for controlling fibrogenesis and immune responses, we selected two SNPs within the TOLLIP coding gene that have previously been associated with IPF and investigated in the context of autoimmune diseases like RA.22 While most studies to date have focused on the mRNA expression in lung tissue, little is known about the actual circulating protein concentrations of TOLLIP. The reduced TOLLIP levels observed in patients with SSc resemble patterns seen in IPF, where decreased TOLLIP expression was detected in fibrotic lung tissue.20 Given the immunomodulatory and anti-fibrotic role of TOLLIP through inhibition of TLR 2/4 and NF-κB signalling,10,13 these findings may reflect a general loss of protective TOLLIP-mediated regulation in SSc. This is consistent with the emerging understanding that innate immunity plays a central role in the pathogenesis of fibrotic ILDs, with chronic inflammation driven by multiple immune cell subpopulations and molecular pathways that maintain tissue remodelling and fibrosis.3 Although serum TOLLIP concentrations tended to be lower in patients with SSc compared with HCs, the considerable overlap between groups indicates marked biological heterogeneity. This is not unexpected in a complex multisystem autoimmune disease such as SSc, where systemic biomarkers reflect dynamic immune regulation rather than a binary disease state. This suggests that TOLLIP may represent a marker of altered immune homeostasis rather than an isolated marker. The reduction in serum TOLLIP concentrations in patients with SSc may have significant downstream immunomodulatory effects. TOLLIP normally inhibits TLR2/4 and NF-κB signalling, thereby suppressing the production of proinflammatory and profibrotic cytokines including IL-6, IL-13, TNF-α and TGF-β.11,13 Loss of this protective TOLLIP-mediated regulation may therefore contribute to the chronic inflammation and progressive fibrosis characteristic of SSc.31 Although we did not measure these cytokines in our cohort, previous studies have demonstrated elevated levels of IL-6, TNF-α and TGF-β in patients with SSc, which correlate with disease activity and the extent of fibrosis.32 33 These findings underscore the broader relevance of innate immune dysregulation in fibrotic ILDs and support the hypothesis that reduced TOLLIP may contribute to aberrant monocyte and macrophage activation.19 Future studies incorporating simultaneous measurement of TOLLIP and downstream inflammatory markers would be valuable to confirm this mechanistic link. The finding that patients with SSc-ILD exhibited higher serum TOLLIP concentrations than those without ILD seems paradoxical at first. However, this observation may represent a compensatory mechanism in response to lung inflammation or epithelial injury to restore immune homeostasis. Previous studies in IPF have shown increased local TOLLIP expression in abnormal epithelial cells, despite overall downregulation at the tissue level.23 34 35 Another possible explanation could be that ILD-specific therapies may have influenced TOLLIP serum levels, as patients with higher serum TOLLIP concentrations were more often treated with immunosuppressive medications. Since these drugs influence cytokine signalling pathways and immune cell activation, it is possible that they could upregulate or maintain circulating TOLLIP. However, due to the retrospective nature of our study, we cannot determine whether therapy led to increased TOLLIP levels or if patients with higher TOLLIP were more likely to receive treatment. Whether this association is causal needs to be determined in future studies. In our cohort, serum TOLLIP concentrations did not differ significantly across rs3750920 or rs5743890 genotypes. While prior studies have linked TOLLIP variants to altered mRNA expression and treatment response in IPF, the translation of these genetic effects into measurable circulating protein levels may be modest and influenced by additional factors such as disease heterogeneity, treatment effects and the substantial proportion of values at the assay detection limit.17 20 Consequently, genotype–protein relationships in serum may require larger, prospectively collected cohorts and assays with greater sensitivity to be reliably detected. The weak inverse correlation between TOLLIP levels and age is consistent with evidence that ageing is associated with a shift towards a more proinflammatory immune phenotype and reduced expression of regulatory proteins, including TOLLIP.36 Similarly, the higher TOLLIP levels observed in never-smokers compared with current or former smokers may reflect reduced baseline immune activation. Experimental and clinical data show that cigarette smoke exposure downregulates TOLLIP expression in airway epithelial cells and lung tissue through epigenetic mechanisms, leading to impaired host defence and increased susceptibility to viral infections.37 Smoking is also associated with broad alterations in systemic immune and inflammatory markers, with current smokers exhibiting lower levels of several regulatory cytokines and higher levels of proinflammatory markers compared with never-smokers.38 These findings support the interpretation that lower TOLLIP levels in smokers reflect enhanced baseline immune activation and dysregulated immune responses. The observation that higher TOLLIP levels coincided with younger age and lower FVC may indicate that increased circulating TOLLIP reflects an early fibrotic or inflammatory disease phenotype, possibly related to epithelial activation during the onset of ILD. Despite these observations, serum TOLLIP concentrations did not correlate with ILD progression, while Cox regression analysis confirmed an association between higher TOLLIP levels and shorter survival. This association is consistent with recent meta-analyses and cohort studies, which have demonstrated that specific genetic variants of TOLLIP and elevated serum TOLLIP levels are predictors of worse survival and more rapid disease progression in IPF and SSc-ILD.21 25 39 Interestingly, in our previously published study on TOLLIP SNPs in patients with SSc-ILD, the rs5743890 minor allele (C) was significantly associated with a higher risk of ILD progression,26 whereas the current analysis did not show any predictive value of serum TOLLIP levels for either progression or survival. This discrepancy could be related to the fact that the levels of a circulating biomarker can be influenced by treatments and concomitant conditions, including fluctuations of disease activity. Further studies are needed to better understand the interaction between TOLLIP levels, TOLLIP SNPs and ILD development and progression. Notably, patients with SSc-ILD showed a longer median survival compared with those without ILD. After ruling out a possible role of comorbidities like cardiovascular disease and pulmonary hypertension, this paradox could be explained by earlier initial hospital presentation and longer follow-up period in patients with SSc with ILD, and by the use of a composite endpoint of death or loss to follow-up. Although this study presents novel findings, several limitations should be acknowledged. First, the monocentric design without a validation cohort limits the generalisability of our results. Second, the sample size of patients with SSc was too small to perform reliable subgroup analyses. Third, despite a clear trend towards lower serum TOLLIP concentrations in patients with SSc compared with HCs, considerable overlap between groups was observed, limiting its utility as a standalone diagnostic biomarker. Fourth, high-resolution CT scans were not systematically collected, precluding analysis of the association between radiological findings, TOLLIP SNPs and serum TOLLIP concentrations. Such analyses could provide important insights into whether TOLLIP levels reflect the extent or pattern of lung involvement. Fifth, serum levels of downstream inflammatory markers (eg, IL-6, IL-13, TNF-α, TGF-β) were not available, which would have allowed us to directly assess the functional consequences of reduced TOLLIP concentrations. Sixth, detailed information on immunosuppressive and antifibrotic treatment regimens, including combinations, dosage and duration was unavailable, limiting our ability to assess the impact of therapy on TOLLIP levels. In conclusion, this study provided a new insight into the role of TOLLIP serum concentrations in SSc, with and without ILD. TOLLIP concentrations are tendentially reduced in patients with SSc compared with HCs, but elevated in those with ILD compared with patients with SSc without lung involvement. While these levels differed by ILD presence, no significant differences were observed across TOLLIP genotype. Serum TOLLIP concentrations did not correlate with disease progression. However, elevated TOLLIP levels were independently associated with shorter survival in multivariable analysis. These findings suggest that circulating TOLLIP might reflect disease course, but influence of treatment cannot be excluded. Further validation with larger cohorts in longitudinal studies is mandatory. Supplementary material Footnotes Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Consent obtained directly from patient(s). Ethics approval: The study complies with the current laws in Germany and was approved by the University Duisburg-Essen Research Ethics Committee (21-9904-BO). Participants gave informed consent to participate in the study before taking part. Data availability free text: Requests should be directed to the corresponding author. Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research. Data availability statement Data are available upon reasonable request. References Associated Data Supplementary Materials Data Availability Statement Data are available upon reasonable request.