Plasma Proteome Analyses Identify Predictors of Graft Rejection Prior to Stem Cell Infusion and Support Interferon-Mediated Ferroptosis As a Novel Mechanism of Graft Rejection

Introduction No established biomarkers or effective interventions exist for graft rejection. We previously reported that CXCL9, a downstream marker of interferon, differentiates graft rejection from other complications but is only useful at the time of rejection. We hypothesized that biological differences in the host proteome exist prior to HSCT and that interferon-mediated pathways drive host elimination of donor cells during graft rejection. Objectives We sought to identify novel graft rejection mechanisms, biomarkers and therapies. Methods We studied the plasma proteome of 21 HSCT recipients prior to stem cell infusion (day -2 to day 0) and at the time of rejection (or timepoint matched days for controls) using an aptamer-based proteomics analysis (SomaScan). Twelve patients suffered graft rejection and associated fever and 9 (matched for age and diagnosis) developed fevers after HSCT but not graft rejection (febrile controls). Results A total of 723 proteins were significantly different in graft rejection patients at the time of rejection vs controls (Fig. 1A). CXCL11 was the most differentially expressed protein during rejection (p=5.2e-8, log2 fold change= 12.9) but did not differ from controls at baseline (Fig. 1B). In contrast, LAG3, an interferon regulated protein involved in activated T-cell expansion, was different in rejection patients both prior to stem cell infusion (p=0.009) and during rejection (p=0.007, Fig.1C). A total of 158 proteins were different prior to stem cell infusion in patients with later rejection vs controls.Ingenuity pathway analysis (IPA) of proteins differentially expressed prior to stem cell infusion found differences in interferon-γ pathways that promote leukocyte recruitment (Fig.2A), suggesting some recipients are primed for rejection by interferons before donor stem cells are infused. IPA of differentially expressed proteins at time of rejection shows a combination of interferon-γ, α, and β pathways resulting in cell death and leukocyte activation (Fig. 2B). Ferroptosis signaling was also significantly increased in rejection patients at the time of rejection vs controls (p=1e-3). The ferroptosis database (FerrDB) identified 32 genes corresponding to proteins differentially expressed in rejection patients (Table 1). Our data suggest interferon activation during rejection induces ferroptosis-mediated cell death of donor cells. Conclusion This is the first study to establish highly significant differences in biologically relevant plasma proteins prior to stem cell infusion. Our findings suggest early intervention with targeted therapies may prevent impending rejection and identifies measurable proteins for graft rejection surveillance. Our findings strongly validate the integral role of interferons in graft rejection and identify a novel mechanism of interferon-mediated cell death in graft rejection, ferroptosis.