Identification of Microbiota-Reactive CD4+ T Cells in the Post-Transplant Setting

Background: The development of acute graft-vs.-host disease (GVHD) is primarily initiated by alloreactive donor T cells. However, we have recently shown using established murine transplant models that clonal expansion of donor CD4+ T cells post-transplant is also dictated by reactivity against the recipient microbiota which can act as a source of antigen to exacerbate GVHD (Yeh AC et al., Immunity 2024). To date, microbiota-targeting CD4+ T cells have been identified and cloned in the context of murine B6 background, which carries a single major histocompatibility antigen class II (MHCII) allele, I-A(b). However, there have been no microbiota-targeting CD4+ T cells identified in the patient setting thus far. Unlike in murine models, a major challenge in screening for reactive donor T cells in the clinic is due to the diversity of MHCII alleles across patients. Here, we develop a screening methodology that enables identification of microbiota-reactive donor T cells on the native MHCII background using both donor and recipient PBMC samples through detection of CD40L, a surface marker of T cell receptor (TCR) activation that allows rapid isolation of viable T cells for downstream characterization. Methods: Our platform includes three major components. First, given a limited number of cells in PBMC samples collected very early post-transplantation before the administration of post-transplantation cyclophosphamide (PTCy) and calcineurin inhibitors, we utilize an established rapid expansion protocol to cryopreserve a large and redundant pool of T cells (REP T cells) harvested from recipient PBMC post-transplantation. We next generate fresh monocyte-derived dendritic cells (moDCs) using a 5-day fast DC expansion protocol from donor PBMCs. We establish a co-culture incorporating a 10:1 ratio of REP T cells and moDCs on a 96-well plate along with bacterial lysate of interest. While REP T cells upregulate CD40L compared to naïve T cells, 24-hour lead-in co-culture with moDCs downregulates CD40L to baseline levels and enabled subsequent CD40L detection as an assay for TCR activation. On day 2, anti-CD40L antibody was added in co-culture along with monensin for optimal CD40L detection within 12-24 hours. Using this method, activated REP T cells are detected by flow cytometry by day 3 of co-culture. Results: As proof of concept, we utilized cryopreserved PBMCs isolated on day 2 from a patient undergoing a matched unrelated donor transplant prior to PTCy (NCT03970096). We subjected each cryopreserved sample to 1 round of REP to generate a pool of over 80M cells (>10x fold expansion). Upon co-culture of REP T cells with donor-derived moDCs, we demonstrated that CD40L was more sensitive at detecting TCR activation compared to IFNg, TNF, IL2, and CD107a using Staphylococcal enterotoxin B (SEB) (15-25% vs. 1-10%) while also preserving cell viability. We next conducted a co-culture comparing 3 different sterile bacterial lysate pools including oral strep (3 strains), enterococcus (7 strains), and commensal anaerobes (10 strains) at a lysate concentration of 10mg/mL REP T cells significantly upregulated CD40L within 24 hours when exposed to the commensal anaerobe pool but not to the other two pools when compared to buffered saline control. CD40L positive CD4+ T cells against the commensal anaerobe pool were subsequently flow-sorted and clonally expanded from a 96-well plate with 2 additional REP cycles. Subsequent rescreening for activity against the commensal anaerobe pool identified 1 out of 8 clones that were confirmed positive upon repeat assay. Conclusions: We developed a novel method that enables the detection of microbiota-reactive T cells post-transplantation in the context of native MHCII genotype. This tool can be used isolate T cell clonotypes of interest for downstream characterization and epitope validation.