Differential Expression of Immunity Related Genes and Early Prediction of Severe Graft Versus Host Disease after Allogeneic Hematopoietic Cell Transplantation

Background:

Allogeneic hematopoietic cell transplantation (HCT) is curative for hematological malignancies. However, Graft versus host disease (GVHD) is the most common and debilitating complication of HCT as 10-15% HCT recipients die and 25% suffer long-term due to GVHD. Effective but toxic prophylaxes for GVHD (e.g., anti-thymocyte globulin) exist that can be given preemptively to patients with high risk of GVHD. However, an early and accurate identification of patients who are at a high risk of developing severe GVHD is the prerequisite for any preemptive therapy. The 'onset' of symptoms associated with GVHD is the final manifestation of a series of cellular and molecular perturbations that take place when host epithelial cells are attacked by the donor immune cells. With the aim of identifying potential functional biomarkers for early prediction of severe GVHD, a comprehensive panel of 594 genes shared among 24 immunology-related gene networks were retrospectively analyzed at different posttransplant time points prior to the onset of GVHD .

Methods:

Two hundred and two RNA specimens from 117 first allogeneic HCT recipients with (n=49) or without (n=68) clinically significant GVHD were analyzed. RNA was extracted from cryopreserved peripheral blood mononuclear cells (PBMNCs) collected at three early post-transplant time points: one week (n=32), one month (n=85) and two months (n=80) after transplantation. Messenger RNA counts for 594 human genes with immunity related functions and 15 internal reference genes were obtained using Nanostring based gene expression CodeSet profiling. Empirical Bayes statistics for differential gene expression, receiver operating characteristic (ROC) and principle component analyses were performed using R statistical packages to identify the gene sets with highest sensitivity and specificity for early prediction of clinically significant GVHD.

Results:

A 'GVHD transcript panel' comprising of highly upregulated 6 genes with T cell related pro-inflammatory functions was identified at one month post transplantation in patients with clinically significant GVHD (grade II-IV acute GVHD (median day of onset = 34 days) and/or chronic GVHD requiring systemic therapy, median day of onset = 108 days). The panel consists of CD27 (AUC=0.93; p=3.4X10^-8); B- and T-lymphocyte attenuator, BTLA ((AUC=0.92; p=3.4X10^-8),Inducible T-cell Co-stimulator, ICOS (AUC=0.90; p=3.5X10^-7), CD5 (AUC=0.87; p=8.5X10^-7), CD3D (AUC=0.88; p=8.8X10^-7) and CD28 (AUC=0.91; p=9.9X10^-7). The GVHD transcript panel predicted clinically significant GVHD with sensitivity of 85% and specificity of 95%.

Conclusions:

The findings strongly favor the early prediction of clinically significant GVHD made possible by differential expression profile of genes associated with T cell related pro-inflammatory functions The identified 'GVHD transcript panel' is highly sensitive and specific in identifying patients at a high risk of developing GVHD and can pave the way to precision HCT medicine.