Longitudinal tear cytokine biomarkers: an analysis from the Close Assessment and Testing for Chronic Graft-vs.-Host disease (CATCH) protocol: Predictive tear cytokine biomarkers for ocular GVHD.

Abstract

Background: Ocular graft-versus-host disease (oGVHD) is one of the most common initial manifestations of chronic GVHD (cGVHD) leading to significant morbidity and reduced quality of life. Early detection of oGVHD using susceptibility/risk biomarkers is urgently needed to enable preemptive therapy.

Objectives: In this subset analysis of patients enrolled on the CATCH Study (NCT04188912), we tested whether changes in tear film cytokines or ocular symptoms, as assessed by the Lee symptom scale (LSS) eye subscale, can predict oGVHD onset.

Study design: LSS eye subscores, Inflammadry (MMP9) and conjunctival washing samples were collected before hematopoietic cell transplantation (HCT) and every 2 months (mos) until 12 mos. A custom-designed 13-plex human cytokine magnetic bead panel was used to measure: IL-10, IL-17A, IL-1Ra, IL-1α, ELA2, IL-1β, LIGHT/TNFSF14, NGAL, OSM, IL-8, IP-10, TNF-α, and VEGF-A. Cytokine levels at the pre-HCT visit were compared across the groups using the Kruskal-Wallis test. Fold change (FC) of the cytokines, defined as post-HCT value divided by pre-HCT value, was calculated and FC≥2 was used in further analyses. oGVHD diagnosis was based on the NIH diagnostic criteria and having an eye score ≥1. Cox regression models were used to examine the longitudinal relationships between potential predictors and oGVHD development.

Results: Of the 44 patients included, 18 developed oGVHD, 11 had cGVHD without oGVHD, and 15 did not have any cGVHD. Median age was 64.5 years, median time from HCT to cGVHD was 6.4 mos and to oGVHD was 8.3 mos. There were no significant differences in baseline cytokine levels among groups. None of the tear cytokines or the InflammaDry MMP9 test predicted oGVHD onset. Clinically meaningful change in LSS eye score was associated with subsequent oGVHD development when compared to cGVHD without eye involvement (HR 2.5, 95% CI 1.2-5.1, p=0.01); and when compared to controls (HR 3.0, 95%CI 1.4-6.0, p=0.004) but the PPV of LSS change ≥15 points was low (27.6%), with a higher NPV (89.4%).

Conclusions: This is the first prospective longitudinal study of tear cytokines and symptoms in a cohort of patients observed closely through HCT for development of cGVHD. We were not able to identify any biological susceptibility/risk markers for oGVHD. Patient-reported symptoms as measured by the LSS are associated with oGVHD development but the low PPV and overlap with diagnostic criteria limit its usefulness as a biomarker to guide preemptive treatment studies.