TCR Repertoires in Graft-Versus-Host-Disease (GVHD)-Target Tissues Reveals Tissue Specificity of the Alloimmune Response

Introduction: Graft-versus-host-disease (GVHD) arises from the inflammatory cascade triggered by alloreactive T cell following allogeneic hematopoietic cell transplantation (allo-HCT). The tissues most frequently involved by GVHD include the small and large intestines, skin, and liver. Prior studies of T cell receptor (TCR) sequencing in diagnostic biopsies and blood have identified dominant T cell clones in GVHD-affected tissues that were not abundant in circulation or shared across multiple patients, but this has not been studied in sites poorly accessible to biopsy nor in lymphoid tissues in humans. We hypothesized that the GVHD-affected tissues have distinct TCR repertoires reflecting the differential expression of allo-antigens at different anatomic sites.

Methods: We performed rapid autopsies on patients whose post allo-HCT course was complicated by GVHD to profile the TCR repertoire in tissues inaccessible to biopsy. Tissues were obtained from seven patients (HLA-identical allografts with a variety of graft sources and GVHD-prophylaxis regimens), all with active GVHD and/or on immunosuppression for GVHD control. Spleen, liver, skin, and multiple sites along the gastrointestinal (GI) tract were sampled, when possible, from the esophagus to the rectum. When available, blood and bone marrow mononuclear cells were also viably preserved. T cell receptors were sequenced from 38 different snap-frozen tissues from five patients via genomic DNA based next-generation sequencing of the TCR-beta CDR3 (ImmunoSeq, Adaptive Biotechnologies). Four out of five subjects sequenced had skin and GI GVHD, one only GI. In parallel, TCRs were sequenced from GVHD-affected tissues from a major and minor mouse model of GVHD: BALB/c 7-14 days after HCT with C57BL/6 T cells and C57BL/6 mice > 30 days after HCT with LP/J T cells.

Results: Sequences representing 264,678 productive TCRs were recovered from the human autopsy samples with over 100 unique clones for nearly all tissues (mean 1539; standard deviation 1626). We found virtually no TCR clones defined by nucleotide sequence shared across patients, even in the context of shared HLA haplotypes (4/5 patients shared HLA-A*02-01). This is consistent with prior studies of diagnostic biopsies describing a paucity of clones shared across patients. We observed greater repertoire overlap between sites within the GI tract compared to other tissues in a given patient, as measured by the Jensen Shannon Diversity (JSD) index, especially compared to the skin, another GVHD-affected tissue. Despite differences in global repertoires across tissues, some clones were shared across all samples for a given patient, with at least one clone present among the top twenty clones by frequency for each tissue with sufficient sampling. A strikingly similar pattern of repertoire sharing across tissues was observed in the TCR repertoires of a major and minor mouse model of GVHD. While tissues within the GI tract and mesenteric lymph nodes shared the most clones in the mice, there were also dominant clones shared across tissues outside the GI tract. Individual mice with GVHD had highly divergent repertoires from each other in spite of the markedly controlled setting including inbred mice with the same donor T cell pool for each transplant, consistent with the notion that T cells that mediate GVHD may be present at very low frequency in the pool of naïve T cells in the allograft.

Conclusion: This work characterizes T cell repertoire in GVHD in human tissues that have not been previously available, including lymphoid tissues and small intestine samples. Combined with the mouse data, this study reveals the relationships between TCR repertoires across different tissues, highlighting the striking diversity of clones driving the GVHD process across tissues and individuals. Although some even dominant clones are shared across all tissues within an individual, each sampled site had a tissue-specific clonal composition. While GVHD-directed therapy focuses primarily on the inhibition of the T cells themselves, additional attention must be devoted to understanding the pattern of target tissue-specific antigen expression in order to identify the key drivers of GVHD.