PROJECT 3 ABSTRACT Gastrointestinal (GI) graft-vs-host disease (GVHD) remains the major cause of non-relapse mortality (NRM) after hematopoietic cell transplant but GVHD symptom severity at onset predicts outcomes too poorly to guide treatment and thus all patients are treated with prolonged courses of high dose systemic steroids, under treating some and over treating others. Steroid treatment for GVHD is itself a major cause of morbidity and complications include infections, reduced physical function, and poor quality of life (QOL). In the current cycle, we developed and validated the MAGIC algorithm probability (MAP), which combines two serum GI GVHD biomarkers into a single value that measures the extent of GI crypt damage. The MAP predicts clinical outcomes such as response to treatment and NRM more accurately than clinical severity alone and thresholds categorize patients into high or low risk groups. Patients with high risk GVHD (35% of GVHD) account for 75% of NRM. In contrast, our preliminary data show that a subset of patients with low risk GVHD who exhibit both clinical and biomarker responses during the first two weeks of treatment represent an ultra-low risk population with 91% response to steroids and 2% NRM. Ultra-low risk patients comprise 40% of all GVHD. Our central hypothesis is that biomarker-guided, risk adapted treatment of GVHD will improve outcomes. We will test this hypothesis in two clinical trials. Specific Aim 1 leverages the discovery in Project 2 that activation of RIPK1 causes destruction of intestinal stem cells and its inhibition can both prevent and reverse GI GVHD in preclinical models. In this Aim, we will conduct a multicenter Phase 2 clinical trial to test whether the addition of a RIPK1 inhibitor to steroid treatment in patients with high risk GVHD increases treatment response rates and decreases NRM. In Specific Aim 2 we will conduct a Phase 2 multicenter clinical trial that tests whether ultra-low risk GVHD can be successfully treated with a brief exposure to steroids. We will use real-time clinical and biomarker monitoring to reduce steroid exposure by more than 50%. We will quantify the efficacy of this approach by comparing the frequency of severe infections in trial patients to a well matched contemporaneous control cohort. In a subaim, we will test whether our approach improves QOL as measured by patient reported outcomes in cases compared to controls. These two trials thus use innovative and complementary, risk-adapted treatment strategies to treat both high and low risk GVHD. If the trials are successful, we will have devised new treatments for the large majority of patients who develop GVHD and will have addressed the under-treatment of high risk GVHD as well as the over- treatment of low risk GVHD. .