Discovery of the transcription factor ZNF683/hobit as a central regulator of pathogenic T cell proliferation in acute gvhd

Abstract Background: Severe acute GVHD (aGVHD) remains a major cause of mortality after transplant, especially with CNI/MTX prophylaxis. However, the molecular mechanisms driving breakthrough aGVHD with CNI/MTX remain largely unknown. The ABA2 trial led to FDA approval of abatacept, a CD28:CD80/86 costimulation blocker, as an adjunct to CNI/MTX that significantly prevented aGVHD. Using ABA2 samples, we combined multiparameter flow cytometry with RNA-seq to identify the cells and pathways that drive aGVHD under CNI/MTX, and how abatacept controls them. We have discovered the transcription factor, ZNF683 (HOBIT), classically defined as regulating cytotoxicity and tissue resident memory programming, but recently identified as mediating T cell responses to Immune Checkpoint Inhibitors (ICI), as a central regulator of breakthrough alloproliferation during aGVHD, which can be controlled with CD28:CD80/86 blockade. Methods: PBMCs from ABA2's 8/8 HLA-matched arm [n=69 in CNI/MTX/Placebo (PBO) n=73 in CNI/MTX/abatacept (ABA)] were studied. Multiparameter flow tracked 308 immune populations, including 227 naïve/memory T cell subtypes/cell states. Flow-sorted Day+21 to +28 CD4 and CD8 T cells from 88 patients underwent bulk RNA-Seq, and Day+28 T cells from 36 patients (18/arm) underwent scRNA-seq; those developing aGVHD before sample collection (median aGVHD onset Day+36) were excluded. The role of HOBIT in T cell proliferation was interrogated with CellTrace-labeled mixed lymphocyte reactions (MLR) and CD3/CD28 stimulation, comparing non-modified T cells, lentivirus-mediated HOBIT overexpression (HOBIT-OE), and transduction controls. The Division Index (‘DI’, the average # divisions a cell has undergone) was calculated with FlowJo. Results: In PBO, the major immunologic signature of aGVHD was T cell proliferation, with multiplexed flow analysis identifying Ki67+ CD4 Tcm and Tem as significantly associated with Grade (Gr) 2-4 aGVHD. Gene Set Enrichment Analysis (GSEA) of CD4+ and CD8+ bulk RNA-Seq from Gr 2-4 vs Gr 0-1 aGVHD revealed that, of the top 20 C2cp GSEA signatures (ranked by Normalized Enrichment Score, 'NES'), 18/20 and 19/20, respectively, were associated with proliferation/cell cycle. scRNA-Seq of PBO proliferating T cells further identified type-I interferon signaling enriched in Gr 2-4 aGVHD (Reactome 'Interferon alpha/beta Signaling': NES=2.6, p <0.002). Pseudo-bulk differential expression (DE) analysis (Gr 2-4 vs Gr 0-1 aGVHD, fold-change > 2; adjusted p < 0.05) identified candidate drivers of inflammatory proliferation. Five genes were DE in proliferating CD4+ cells (Up:IL2RA, PELI1, BCL2L11; Down:HOBIT, HLA-DR5,) and two in CD8+ cells (Down:HOBIT, MXRA7). The transcription factor HOBIT was the only DE gene in both, downregulated 5.9x in CD4+ and 4.5x in CD8+ Gr 2-4 aGVHD proliferating T cells (p=0.038, p=0.008). Importantly, abatacept controlled both the inflammatory proliferation of T cells, and normalized HOBIT expression in Gr 2-4 aGVHD. A dynamic role for HOBIT in T cell proliferation was validated using bead-based stimulation assays and MLRs. With CD3/CD28 stimulation, HOBIT-OE significantly increased the proportion of T cells entering division vs controls (Day 3 DI for CD4+: 1.73 vs 1.21; CD8+: 1.79 vs 1.23, n=6, p=0.03 for each). However, when non-gene-modified T cells (capable of modulating HOBIT expression) were assessed in MLRs, significant transcriptional downregulation of HOBIT occurred as alloproliferation progressed, with >20-fold HOBIT downregulation in highly proliferating vs non-proliferating T cells (n=8 biologic replicates, p=0.008). These experiments reveal complex transcriptional control of HOBIT during T cell activation, with its expression enhancing the probability that T cells enter division, followed by HOBIT downregulation as alloproliferation proceeds. Conclusions: We have discovered a prominent role for the transcription factor HOBIT in regulating inflammatory T cell alloproliferation that drives CNI/MTX breakthrough aGVHD, and its control with abatacept. These results are the first to identify HOBIT as a key regulator of alloreactivity, uncovering a new link between the mechanisms that harness the nascent T cell activation potential unleashed by ICI, and those that drive aGVHD. They nominate HOBIT as a critical regulator of multiple facets of T cell activation, and underscore the central role that CD28:CD80/86 signaling plays in modulating HOBIT expression, and its downstream effects.