Oral Presentation International Celiac Disease Symposium 2026

Serum proteomic profiling of celiac disease: Extending inflammatory signatures with the novel Protein XI multiplex platform (147038)

Jostein H. Ibsen 1 2 , Stine R. Lund 1 2 , Knut E. A. Lundin 1 2 , Louise Fremgaard Risnes 1 2
  1. Norwegian Coeliac Disease Research Centre, Oslo, Norway
  2. Oslo University Hospital, Oslo, NORWAY, Norway

Celiac disease (CeD) is a gluten-driven enteropathy with wide-ranging systemic effects. Using the Olink Target 96 Inflammation panel (Uppsala, Sweden, www.olink.com), we showed that intestinal immune activity in CeD leaves a detectable serum footprint, with markers (CXCL9, TNFRSF9 and IL-17A) tracking mucosal damage and treatment response (Ibsen et al, Gastroenterology, in press). Building on this, we applied ProteinXI, a new high-multiplex quantitative  immunoassay platform from Mesoscale (Rockville, MD, USA, www.mesoscale.com), to capture a substantially wider proteomic signature and further link circulating proteins to serology and mucosal pathology.

We reanalyzed 100 serum samples (57 individuals) from this previously described secondary care cohort . Forty-three CeD patients provided paired samples at diagnosis (UCeD) and after one year on a gluten-free diet (TCeD), alongside 14 controls. The platform quantified 391 proteins in absolute concentration, with 87.3% data completeness after quality control. Log2-transformed values were analyzed in R (FDR < 0.05).

Principal component analysis separated samples into three clusters, with treatment status as the dominant axis of variation and disease status (CeD versus controls) as the secondary. Pathway enrichment (GO, Reactome) revealed heightened leukocyte and lymphocyte differentiation and hematopoiesis in UCeD relative to controls, whereas TCeD showed enrichment for VEGF signaling and macrophage colony-stimulating factor. 

Several proteins tracked closely with mucosal damage severity, demonstrating strong cross-platform concordance by confirming previously identified Olink markers, while revealing novel markers with even stronger signals. An exploratory three-protein machine-learning model predicted mucosal healing with an AUC of 0.934, rising to 0.953 when combined with IgA-TG2, exceeding IgA-TG2 alone (AUC=0.914).

These findings reinforce that gut-level inflammation in CeD is systemically detectable and that gluten-free treatment leaves residual immunological signals after one year. Targeted protein panels derived from this signature offer a promising, non-invasive route to monitor disease activity and mucosal recovery.