Introduction: Rejection and infection are two risk factors for chronic lung allograft dysfunction (CLAD), all together limiting success of lung transplantation (LuTx). Traditional histology-based diagnosis of acute cellular rejection (ACR) is limited by low sensitivity and specificity, including difficulties to distinguish ACR from infection-mediated inflammation. Many previous biomarker and mechanistic studies of rejection after human LuTx do not distinguish donor cells from recipient cells, leading to difficulties in interpreting their clinical significance. Despite the importance of T cells in driving alloresponses and immune defense, their dynamic repopulation, clonal distribution, alloreactivity and anti-microbial reactivity after LuTx are largely unknown.
Method: We utilized a combination of flow cytometry for chimerism determination and T cell phenotyping, and high-throughput sequencing of T cells from post-Tx bronchoalveolar lavage (BAL) and PBMCs. We set up mixed lymphocyte reactions with pre-Tx cells and then identified alloreactive T cell receptors (TCRs) in graft-vs-host (GvH) and host-vs-graft (HvG) directions. We defined and tracked pathogen-reactive TCRs by integrating publicly available databases.
Results: Patient (Pt) 2 and Pt5 were excluded from our cohort (n=13) due to early death. Remaining patients had a low rate of ACR (3/11), but high rates of infection (11/11) and either definite (4/11) or probable (6/11) CLAD. Recipient CD14+ cells showed rapid infiltration in BAL post-Tx, suggesting a role in priming GvH response. Much higher levels of HvG compared to GvH clones were associated with faster recipient T cell repopulation in the BAL. We observed high levels of HvG-reactive TCRs among recipient mappable repertoires in BAL, but not paired PBMCs, of patients with early ACR. In many other patients, high peak levels of HvG-reactive TCRs were identified in BAL, but not PBMCs. However, pathology results were read as negative for early ACR in these cases, despite multiple complications occurring later, including de novo DSA, late ACR, multiple infections driven by different types of pathogens, and definite or probable CLAD, consistent with potentially missed diagnose of ACR by histology and the need of refined immunosuppression. Recipient graft-infiltrating CD8 T cells gradually acquired tissue-resident memory features, and elevated CD28 expression on these cells correlated with early ACR. In patients who were infected with EBV and CMV post-Tx, we observed an increase in EBV- or CMV-reactive TCRs in the BAL that are enriched for non-alloreactive repertoires.
Conclusions: Our data suggest that cellular and clonal changes may have already happened in lung allografts when histology is not diagnostic of rejection. Developing an immunological toolset at the chimeric, alloreactive clonal and phenotypic levels is expected to facilitate rejection diagnosis and mechanistic understanding of rejection, infection and CLAD after human LuTx.