Association of Plasma IL-6 with Indoor Radon Exposure in Children with Non-Allergic Asthma Authors: Saleh Alsulami, Youn Soo Jung, Kari Nadeau, Perdita Permaul, Longxiang Li, Petros Koutrakis, Jonathan M. Gaffin, Wanda Phipatanakul, Tina M. Banzon Published: 2026 ABSTRACT Background/Objectives: Radon exposure has recently been associated with asthma morbidity, including increased airway inflammation and school absenteeism in children, though limited data on underlying biological mechanisms exist. Interleukin-6 (IL-6), a pleiotropic cytokine implicated in both Type 2-low airway inflammation and radon-related lung carcinogenesis, may represent a key mechanistic link between radon exposure and asthma morbidity. We aimed to evaluate the association between indoor radon exposure and plasma IL-6 levels in children with asthma and whether this relationship differs by allergic sensitization status. Methods: We analyzed baseline data from the School Inner-City Asthma Study, a prospective cohort of children aged 4–13 years with persistent asthma. Monthly indoor radon concentrations at each participant’s residential ZIP Code Tabulation Area were estimated using a validated spatiotemporal prediction model. Plasma IL-6 was measured from baseline blood samples. Multivariable linear mixed-effects models with random intercepts for school were used to assess the association between radon exposure and IL-6, adjusting for demographic, clinical, and socioeconomic covariates. Effect modification by allergic sensitization was evaluated using an interaction term. Results: Among 144 participants, 62.5% were allergen-sensitized. The median home radon concentration was 46.6 Bq/m3 (range 30.7–99.9), and the mean plasma IL-6 was 0.22 pg/mL (SD 0.41). A significant interaction was observed between radon exposure and allergic sensitization status (β-interaction = −0.012; p = 0.014), indicating differential effects by phenotype. Among non-sensitized children, higher radon exposure was associated with increased IL-6 levels (β = 0.0088; p = 0.044), corresponding to a 0.32 pg/mL rise in IL-6 per 37 Bq/m3 increase in radon. No significant association was observed among sensitized children. Conclusions: Indoor radon exposure is associated with higher plasma IL-6 levels in non-sensitized children with asthma, suggesting a potential IL-6–mediated pathway linking radon exposure to asthma morbidity in the Type 2-low phenotype. These findings highlight heterogeneity in environmental asthma responses and support further investigation into radon mitigation as a modifiable factor to improve asthma outcomes. IL-6 may serve as a biomarker to identify children most susceptible to radon-related airway inflammation, guiding personalized mitigation strategies and targeted interventions to improve asthma outcomes. Future studies should incorporate direct home radon measurements, comprehensive endotyping panels, and longitudinal biomarker sampling to validate these findings and elucidate whether IL-6 trans-signaling pathways mediate radon-induced airway injury in non-allergic asthma. 1. Introduction Asthma is a heterogeneous chronic inflammatory airway disease characterized by variable airflow obstruction, airway hyperresponsiveness, and complex immune dysregulation [1]. Advances in immune profiling of patients with asthma have led to the recognition of distinct asthma endotypes broadly categorized into Type 2-high and Type 2-low endotypes [2]. Type 2-high asthma is characterized by allergen-driven type 2 inflammation with elevated interleukin-4 (IL-4), IL-5, and IL-13, whereas Type 2-low asthma involves alternative inflammatory pathways, including IL-6, IL-17, and tumor necrosis factor-α (TNF-α), often with neutrophilic or pauci-granulocytic patterns and reduced corticosteroid responsiveness [3]. Elevated plasma IL-6 has been associated with greater asthma severity, lower lung function, and metabolic dysregulation, suggesting IL-6 as a potential biomarker of a higher-risk, Type 2-low-skewed phenotype [4]. Mechanistically, IL-6 exerts its effects through two distinct pathways: classical signaling via membrane-bound IL-6 receptor (IL-6R) expressed on neutrophils, macrophages, and T cells, and trans-signaling via soluble IL-6R (sIL-6R), which enables IL-6 to activate cells that lack membrane IL-6R, such as bronchial epithelial cells and airway smooth muscle cells [5]. Epithelial IL-6 trans-signaling has been shown to define a distinct asthma subset characterized by reduced epithelial barrier integrity, increased airway remodeling, and frequent exacerbations in the absence of Type 2 inflammation [6]. Furthermore, IL-6, in combination with TGF-β and IL-1β, promotes the differentiation of Th17 cells, which produce IL-17 and drive neutrophilic airway inflammation—a hallmark of severe, corticosteroid-resistant asthma [7,8]. A positive feedback loop between IL-6 and IL-17 in the airways may perpetuate this non-Type 2 inflammatory cycle [5]. This dual role of IL-6 in both promoting Th17-driven neutrophilic inflammation and inhibiting regulatory T cell differentiation underscores its importance as a central mediator in Type 2-low asthma pathophysiology. Understanding these distinct endotypes is central to the emerging paradigm of personalized medicine in asthma, wherein biomarker-guided phenotyping may inform individualized therapeutic and environmental management strategies. Environmental exposures are increasingly recognized as important drivers of airway inflammation and asthma morbidity, yet their role in Type 2-low asthma remains incompletely understood. One exposure of growing interest is radon, an odorless radioactive gas produced from the natural decay of uranium in soil that can accumulate in indoor environments [9]. The heaviest noble gas, radon (222Rn), originates from the natural decay series of uranium-238, a radioactive mineral with a half-life of 4.5 × 109 years naturally present in earth’s soils, crust, and rocks. Radon may enter a house or building through fissures and cracks [10,11], with its levels affected by various geologic factors, building characteristics, the presence of radon mitigation, and energy efficiency, which may affect generalizability of exposure and account for the heterogeneity of radon levels [12]. Radon is well established as the leading cause of lung cancer among non-smokers [11]. Moreover, long-term radon exposure can induce cellular and lung tissue injury via oxidative stress, mitochondrial dysfunction, and mitophagy [13]. Additionally, IL-6 may play a key mechanistic role in radon-mediated lung injury. Genetic studies of uranium miners with high occupational radon exposure have shown that IL6 promoter variants associated with increased IL-6 expression confer a higher risk of lung cancer—particularly squamous cell carcinoma—underscoring the relevance of IL-6–driven inflammation in radon-related lung disease [14]. Recently, our group demonstrated that radon exposure is linked to airway inflammation [15], asthma symptoms, and higher rates of school absenteeism in school-aged children with asthma [16]. Additionally, these findings are corroborated by evidence from other groups showing that chronic home radon exposure is associated with higher inflammatory biomarker concentrations in children and adolescents [17] and that residential radon exposure can induce DNA damage in peripheral blood lymphocytes [18], reinforcing the biological plausibility of radon’s deleterious effects on airway inflammation. Thus, we sought to determine whether indoor radon exposure is associated with plasma IL-6 levels in childhood asthma and assess whether this relationship differs between allergic and non-allergic asthma phenotypes to identify potential mechanisms for this exposure–outcome relationship. 2. Materials and Methods This analysis utilizes data from the School Inner-City Asthma Study (SICAS), a prospective cohort study of children aged 4–13 years with asthma attending urban elementary schools in the northeastern United States (ClinicalTrials.gov NCT 02291302; PMID: 34547084). Eligible participants had a physician diagnosis of asthma at least one year prior to enrollment and evidence of active disease, defined by the presence of at least one of the following in the preceding 12 months: wheezing symptoms, the use of daily controller medications, an asthma-related emergency department or urgent care visit, an asthma-related hospitalization, or at least one course of systemic corticosteroids. Baseline data were collected during a research visit prior to the school year and included demographics, clinical history, aeroallergen sensitization, and blood samples. Detailed study methodology and community-based recruitment procedures have been described previously [19]. This study was approved by the Boston Children’s Hospital Institutional Review Board, and written informed consent/assent was obtained from all participants and their parent or guardian, respectively. Patient consent was not required because this work is a secondary analysis of a published article, and informed consent was obtained from all participants involved in the original study. Aeroallergen sensitization status was determined as a surrogate marker for Type 2-high asthma using skin prick testing (MultiTest device, Lincoln Diagnostics, Decatur, IL) or specific Immunoglobulin E (IgE) levels (ImmunoCAP, PhadiaAB, Uppsala, Sweden), with “any sensitization” defined as positivity to at least one allergen (≥3 mm wheal or IgE ≥ 0.35 kU/L). Aeroallergens included for testing were house dust mite (Dermatophagoides pteronyssinus and Dermatophagoides farinae), German cockroach (Blatella germanica), cat, dog, Alternaria tenuis, mouse pelt, rat pelt, Aspergillus, Cladosporium, Penicillium, oak tree pollen, timothy grass pollen, and ragweed mix (Greer, Lenoir, North Carolina; ALK-Abelló). The plasma IL-6 level was measured using the Meso Scale Discovery V-PLEX multiplex assay platform. The primary outcome was plasma IL-6 levels (pg/mL). Residential radon exposure was estimated using a validated spatiotemporal prediction model predicting monthly indoor radon levels at the ZIP Code Tabulation Area (ZCTA) level across the Greater Boston area based on home zip code [12,20]. Briefly, the model was developed using over 363,783 short-term basement radon measurements collected between 2005 and 2018 via certified passive detectors. A two-stage prediction framework was employed: in stage one, twelve base learners spanning six statistical methods—random forest, gradient boosting machine, neural network, generalized linear model, generalized additive model, and robust linear regression—were trained on a set of 68 covariates encompassing geological, building, socioeconomic, meteorological, and spatiotemporal predictors. In stage two, predictions from the base learners were aggregated using a non-negative geographically and temporally weighted regression (NN-GTWR) ensemble, which allows the relative contribution of each base model to vary locally according to its spatiotemporal prediction performance. The ensemble model achieved a cross-validated R2 of 0.63 and a root mean square error of 22.6 Bq/m3, outperforming all individual base models. Each participant was linked to their ZCTA-level radon estimate based on residential address, and the primary exposure was defined as the average indoor radon concentration (Bq/m3) corresponding to the month of the blood draw visit. Descriptive statistics were used to summarize participants’ baseline characteristics. The analysis evaluated continuous models of indoor radon exposure. To assess the association between indoor radon exposure and IL-6 level, we applied a multivariable linear mixed-effects model with a random intercept for school levels. Household income (above vs. below USD 25,000/year) had 43 missing values. We imputed these using k-nearest neighbors (k = 5), using age, sex, race/ethnicity, private insurance, head-of-household education, marital status, school, and home ZIP code to define distances; the binary income indicator was assigned by neighbor majority. We tested effect modification by allergic sensitization status (“any sensitization”) using an interaction term between radon exposure and sensitization status in the multivariable model. This approach was based on the known relationship of season and allergic sensitization to airway inflammation and asthma outcomes. The model adjusted for potential confounders, including age, sex, recent cold infection, ICS controller medication use (any vs. none), race (Black, White, and Other), annual household income (above vs. below USD 25,000 per year), BMI, and seasonality. A p-value < 0.05 was considered statistically significant. All analyses were conducted using R version [4.5.1]. 3. Results In 144 participants, 36.1% reported income