# IgG subclass-specific N-glycosylation differentiates HRCT subtypes in idiopathic inflammatory myopathies-associated ILD

## TL;DR

This study shows that specific sugar patterns on IgG antibodies can help distinguish different lung disease types in patients with inflammatory muscle disorders.

## Contribution

The study identifies IgG subclass-specific N-glycosylation signatures that differentiate HRCT subtypes in IIM-associated ILD.

## Key findings

- Six IgG2-subclass glycopeptides were significantly altered in IIM patients with ILD compared to those without.
- Distinct glycoform signatures were found for cNSIP, fNSIP, and OP subtypes based on HRCT classifications.
- A model combining seven glycopeptides and clinical variables achieved 89% accuracy in classifying HRCT subtypes.

## Abstract

Idiopathic inflammatory myopathies (IIMs) frequently involve interstitial lung disease (ILD), a major contributor to morbidity and mortality. However, the immunological heterogeneity across radiologic ILD subtypes remains poorly defined. This study aimed to explore whether subclass-specific IgG N-glycopeptides could distinguish high-resolution computed tomography (HRCT)-based ILD patterns in IIM patients.

We analyzed plasma IgG subclass-specific N-glycopeptides in 145 IIM patients, including 98 with ILD (IIM-ILD), using intact glycopeptide mass spectrometry. Among IIM-ILD patients, 82 with available HRCT scans were classified into the three predominant subtypes: fibrotic nonspecific interstitial pneumonia (fNSIP, n=40), cellular NSIP (cNSIP, n=18), and organizing pneumonia (OP, n=24). A weighted multinomial logistic regression model was constructed using LASSO-selected glycopeptides and clinical variables.

Fifteen intact N-glycopeptides (IGPs) were quantified across all IIM patients. Compared to patients without ILD, IIM-ILD patients showed significant alterations in six IgG2-subclass IGPs. Crucially, distinct glycoform signatures were identified across the three HRCT subtypes: cNSIP was enriched in highly sialylated IgG2 forms; fNSIP showed elevated IgG1 and IgG3 glycoforms; and OP exhibited uniquely high IgG2-N5H4F1. These glycoforms correlated significantly with autoantibody profiles and clinical features. A multinomial logistic regression model integrating seven key IGPs and seven clinical variables achieved robust classification of the HRCT subtypes (macro-averaged AUC = 0.89).

Subclass-specific IgG N-glycosylation profiles reflect the immunological heterogeneity underlying IIM-ILD. Integrating these glycan signatures with routine clinical data creates a strong model for distinguishing HRCT-defined endotypes, supporting their potential to improve disease classification and guide future mechanistic research in autoimmune-related ILD.