Back

Protein-level mapping reveals novel Immune State Types (ISTs) associated with Chlamydia trachomatis infection, inflammatory remodeling, and scarring in trachoma-hyperendemic Amhara Ethiopia

Shi, X. C.; Olagoke, O.; Titiheruw, B. R.; Chung, S.; Mengistie, H. D.; Asfaha, K.; Read, T. D.; Dean, D.

2026-10-06 microbiology
10.64898/2026.09.30.755767 bioRxiv
Show abstract

ABSTRACT Background: Trachoma remains the leading infectious cause of blindness. Over 100M people are at risk, and 1.9M already live with visual impairment or blindness. Ethiopia bears the greatest burden of disease with ~61M or 59% of the known affected global population. Although Chlamydia trachomatis (Ct) infection initiates disease, infection alone fails to explain why inflammation resolves in some but progresses to irreversible scarring and vision loss in others. Previous studies have largely examined gene expression or limited protein panels, leaving the conjunctival immune protein environment poorly characterized. We hypothesized that distinct conjunctival immune states capture these pathogenic processes and could reveal biomarkers for selecting treatment and surveillance strategies. Methods and Findings: We conducted a cross-sectional study of 667 participants residing in the hyperendemic Amhara region of Ethiopia at least one year after the most recent azithromycin mass drug administration to treat Ct. The upper tarsal conjunctivae was graded for TF and TI, trachomatous inflammation-follicular or intense, respectively; I-TS/TT, inflammation with scarring with or without trichiasis, defined as one or more eyelashes touching the globe of the eye; and TS and TT, scarring with or without trichiasis but no inflammation. T0 represented no trachoma. The protein concentrations of 41 immune mediators were quantified using Meso Scale Discovery electrochemiluminescence immunoassays. Associations with infection and grade were evaluated using covariate-adjusted models with false discovery rate correction. Bernoulli mixture modeling (BMM) revealed four distinct Immune State Types (ISTs). Both approaches supported a shared biological framework characterized by selective immune activation in Ct infection, broad inflammation in TF/TI, and chemokine remodeling versus immune quiescence in the scarred conjunctiva. Ct infection was associated with a focused IFN-{gamma}-centered response. IST1 captured this immune response where 51.9% of the participants had Ct; we therefore termed IST1 Quiescent Infection. In contrast, TF/TI was associated with higher concentrations of proinflammatory, Th1/Th17, chemokine, regulatory, and remodeling pathways mediators. This profile fit well within IST2, termed Broad Inflammation, where 72.4% of the participants had TF/TI. TS/TT retained considerable chemokine activity, whereas I-TS/TT showed a narrower, IL-17A-dominant inflammatory profile. BMM separated scarring into two states characterized by distinct profiles: IST3, Conjunctival Remodeling, with significantly stronger chemokine and T-cell maintenance signals; and IST4, Quiescent Scarring, with the majority of chemokines being significantly downregulated. This suggests that scarred conjunctiva can retain substantial immune activity or shift toward a more attenuated state. Conclusions: This study establishes the first IST framework for trachoma, revealing biologically distinct immune states with different potential clinical implications. IST1 indicates the need for antibiotic treatment, whereas IST2 suggests that novel immunotherapeutics might be effective in thwarting disease progression. The upregulated chemokines in IST3 may indicate risk of progression to TT necessitating prospective studies and potential novel antifibrotic interventions. Together, our findings provide a strong biological foundation for stratifying patients by immune state and matching them to preventive and therapeutic care. Prospective validation and streamlined biomarker panels will support the translation of this framework into field settings to help avert disease progression and blindness.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.