The key roles of thrombin and fibrinogen in human infant and mice atopic dermatitis models

Atopic dermatitis (AD) affects between 15% and 24% of children in the United States1 and up to 20% of children worldwide.2 Adults with AD have increased cardiovascular risk including thromboembolic events compared to adults without AD.3 Indeed, associations between altered cutaneous and systemic clotting and atopy in adults have been observed,4-6 including a recent study which showed that adults with AD have a significant increase in cardiovascular/atherosclerosis-related proteins compared with controls.7 However, whether children with AD have altered hemostatic function remains unclear. Moreover, despite these associations in adults, the causal relationship and directionality of the association between AD and clotting has not been investigated. Herein, we investigate clotting's role in AD using the Mechanisms of Progression of Atopic Dermatitis to Asthma in Children (MPAACH) cohort8 and an AD mouse model. Briefly, the MPAACH cohort is a longitudinal cohort of children with AD. Children with no history of AD or other allergic disease are recruited in parallel. The MPAACH cohort biorepository includes extensive demographic, clinical, and biomarker data including skin AD severity and transepidermal water loss (TEWL) measurements from lesional (active or historical) and never-lesional (skin never affected by AD and ≥ 10 cm away from a site of active or historical lesions), along with extensive banked longitudinal biospecimens. A full, detailed description of MPAACH and the methods used in this study8 are found within the supplementary materials. We analyzed plasma samples that had not undergone multiple freeze–thaw cycles from 99 MPAACH children. Except for age, the subset and full MPAACH cohort were similar (Table S1). Control participants without AD were recruited in parallel with the MPAACH cohort and were also included in this substudy (Table S2A). The controls were similar to the cases in terms of sex and race, but they were younger than the cases. Due to the noted age difference between the cases and controls, we restricted the analysis to children under 1 year of age including additional demographic comparisons in the <1 year of age group (Table S2B). In this age-controlled analysis, we found that children with AD had no difference in time to clot initiation (Figure 1A p = .98), but children with AD had increased peak thrombin (Figure 1B p = 8.5e-07), decreased time to peak thrombin (Figure 1C p = .04), increased rate of thrombin generation/velocity index (Figure 1D p = 8e-04), and increased total thrombin generated (Figure 1E p = .002) compared to children without AD. Insufficient children without AD that are over the age of 1 as part of MPAACH, therefore a similar analysis is not presented for this subset. Moreover, compared with controls, in the entire cohort, children with AD had shorter clot initiation phase time (p < .001, Figure S1A), higher peak thrombin (p = .0015, Figure S1B), shorter peak thrombin time (p < .001, Figure S1C), and faster thrombin generation rate (p = .003, Figure S1D). Total thrombin generation did not differ between groups (Figure S1E). Children with moderate-to-severe AD had increased total thrombin generation compared to those with mild disease (p = .043, Figure S1A). Lastly, in all children with AD, increased total plasma thrombin was associated with lesional TEWL (β = .30, p = .001, Figure S1F) and never-lesional TEWL (β = .17, p = .02, Figure S1G).These differences reflect a significant increase in the thrombin amplification phase of thrombin generation in children with AD under 1 year of age. Thrombin cleaves fibrinogen to initiate fibrin clot formation.9 Total plasma fibrinogen was inversely associated with lesional TEWL (β = −.1, p = .04). Plasma fibrinogen was not associated with other AD biomarkers (Table S3). Platelet count was not associated with lesional or never-lesional TEWL (Table S3). Lastly, two common prothrombotic variants, factor V Leiden (rs6025) and prothrombin G20210A (rs1799963) frequencies in the children with AD were similar to the general population (Table S4). To determine if thrombin is mechanistically coupled to AD pathogenesis, mice were given chow with dabigatran (7.5 mg/kg of chow), a direct thrombin inhibitor,10 or control diet. Mice (n = 40) underwent three repeated cutaneous allergen patches to induce an AD-like phenotype.11 After three patches, mice receiving dabigatran had decreased TEWL (p = .03) compared with control (Figure 2A) and a substantial decrease in symptom score (p = .07, Figure 2B). To determine whether fibrinogen impacted AD development we performed immunohistochemistry and ELISA-based quantification of fibrinogen on mouse skin. Dabigatran administration resulted in decreased fibrinogen staining (Figure 2C) and significantly lower fibrinogen (p = .0025, Figure 2D) in the Asp group. To determine fibrinogen's direct impact on AD pathogenesis, we used mice with complete (Fib−/−) and partial (Fib+/−) fibrinogen deficiency and wildtype controls (FibWT) in our AD model (n = 30). After three patches, Fib−/− and Fib+/− mice had decreased disease development evidenced by decreased TEWL (p = .0079 and p = .016, respectively; Figure 2E) and median disease severity (p = .0092 and p = .014, respectively; Figure 2F) compared with WT. Notably, Fib+/− mice carry fibrinogen levels that are half of normal, which does not impair hemostasis, but resulted in significant attenuation in AD severity. Our studies are the first to establish associations of pediatric AD with thrombin and contributions of thrombin and fibrinogen to AD, demonstrating that hemostatic system regulation in childhood AD is pathologically altered. Based on our findings we suggest a mechanistic model whereby increased thrombin generation leads to increased conversion of fibrinogen to fibrin, which drives AD pathogenesis and is associated with skin barrier dysfunction and disease severity in pediatric AD. In addition thrombin may act in a synergic role with other skin barrier damage mediators, thus blocking thrombin contributes to decreased barrier damage. Our findings are novel and provide insights into the link between clotting and AD in adults.3 Previously, directionality of associations between clotting and AD remained speculative, our findings provide initial support that thrombin and fibrinogen may contribute to AD pathogenesis. Our results suggest that thrombin and fibrinogen could be used as novel targets for therapy in in patients with AD. Future investigations should elucidate the role thrombin mediated conversion of fibrinogen to fibrin as well as clotting mechanisms impact AD pathogenesis. This work is funded by NIH U19AI070235 (GKKH, JBM, LJM) and 5K12HD28827–27 (MGS) and the American Academy of Allergy, Asthma, and Immunology Foundation (Faculty Development Award, MGS). Alyssa Filuta: Investigation; Peter Amezcua: Investigation; Brandy Ruff: Investigation; Jocelyn M. Biagini, PhD: Resources, Writing–Review and Editing; John Kroner: Resources, Data Curation; Hua He: Data curation, Writing–Original Draft, Visualization; Eric Brandt: Investigation; Lisa J. Martin: Validation; Data Curation, Visualization, Writing–Review and Editing; Joseph S. Palumbo: Conceptualization, Resources, Writing–Review and Editing; Gurjit K. Khurana Hershey: Conceptualization, Resources, Writing–Review and Editing, Funding Acquisition; Michael G. Sherenian: Conceptualization, Supervision, Project administration; Writing–Original Draft, Review, and Editing, Resources, Methodology, Funding Acquisition. We would like to acknowledge and thank the participants and their families who volunteered to be a part of the MPAACH study. No authors have competing interests to disclose. All data may be made available upon request, there are no restrictions on materials or data for the experiments conducted in this study. Data S1: Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.