Discovery and Validation of Graft-Versus-Host Disease Biomarkers

Abstract SCI-50

Graft-versus-host disease (GVHD) following allogeneic hematopoietic cell transplantation (HCT) is the major cause of non-relapse mortality. A major barrier to GVHD research and treatment is that the diagnosis and prognosis relies almost entirely on the presence or absence of clinical symptoms, which can only be confirmed by biopsy of one of three target organs: the skin, the gastrointestinal (GI) tract, or the liver. Currently, no laboratory tests exist to predict the risk of developing GVHD, the responsiveness to treatment, or patient survival. I will first describe what the large-scale identification of proteins in the plasma of GVHD patients is revealing about candidate biomarkers. Using state-of-the-art antibody arrays, tandem mass-spectrometry as a discovery engine, and ELISA for validation in thousands of samples, colleagues from the University of Michigan and Fred Hutchinson Cancer Research Center found that biomarkers of disease-related systemic (1) and tissue-specific changes (2) can be detected in the plasma of patients. Second, I will describe how we provided the first demonstration showing that these biomarkers are associated with GVHD clinical outcomes and prognosis, including long-term survival, independent of clinical stages. We showed that plasma elafin levels have significant diagnostic and prognostic power as a biomarker of skin GVHD (2). Using the same proteomics strategy, we discovered regenerating-islet-derived-3-α (REG3α) as a biomarker of lower GI GVHD and subsequently validated it in two independent sets totaling 1014 patients from three different centers. This marker provides important prognostic information, including response to GVHD treatment and survival. Third, I will describe how the ability to identify patients at high risk for GVHD early in their transplant and treatment course has important therapeutic consequences, including more stringent monitoring and/or preventative care. To determine if the diagnostic biomarkers could predict GVHD before the appearance of clinical symptoms, we evaluated the four most informative biomarkers (IL2R-α, TNFR1, elafin, and REG3α) in samples from 513 unrelated HCT patients. The combination of predicted probabilities at days 7, 14, 21, and 28 post-HCT correctly predicted the absence of grade II-IV GVHD in 80% of patients without GVHD (i.e., specificity), and the development of grade II-IV GVHD in 59% of patients who had not already developed GVHD at the time of the test sample (i.e., sensitivity). Our next step will be to design a preemptive clinical trial using these results and with additional refinements. Fourth, I will describe the shortcomings in the prediction of the response to GVHD therapy. Thus, we further analyzed biomarkers of resistance to GVHD therapy and found that they were different from the biomarkers discovered at GVHD onset. We subsequently identified novel pathways, such as drug resistance, glucocorticoids receptor signaling, glycolysis, catabolic processes, the oxidative stress response, the IL-1 family, and T helper cell signaling. We are currently validating these candidate biomarkers for response to treatment in approximately 1200 samples. The candidates include seven that were discovered with in-depth proteomic analysis along with the six biomarkers that we previously identified as diagnostic markers. Interestingly, several of these novel biomarkers offer important insights into the biology of GVHD and are potential therapeutic targets.