A Systems Immunology Approach to Graft-Versus-Host Disease

Until recently, understanding of immune system organization and function was based on experiments and observations involving cellular responses within secondary lymphoid organs or peripheral blood. While these approaches have been useful in identifying molecules, cells and structures critical for generating productive immunity, their suitability in guiding research in the field of GVHD is less clear. For example, many of the elements important in promoting immunity against infections, including molecular sensors of microbial patterns, dendritic cells and organized lymphoid tissue are redundant in animal models of GVHD. Furthermore, it remains unexplained why only certain tissues are prone to GVHD injury despite widespread antigen expression. Our recent work has evaluated the concept that GVHD-target tissues are major participants in shaping tissue injury.

In order to test the concept that peripheral tissues provide extrinsic signals to effector T cells (T_eff) that directly regulate their capacity to induce injury, we have adopted a ‘systems immunology’ approach in which we tracked the evolution of the donor T cell immune response at multiple sites within the same host. In a clinically relevant BMT model, we transferred B6 donor T cells and bone marrow to lethally irradiated 129SV recipients that were MHC-matched but mismatched for multiple minor antigens. CD8⁺ T_eff were flow sorted from the lymphoid organs and peripheral tissues at day +7 and their gene expression was compared by hybridisation to the Affymetrix Genechip^® Mouse Gene 2.0 ST Array. We observed that T_eff infiltrating the skin and gut have transcriptional signatures very different to those of T_eff isolated from blood, bone marrow, spleen or lymph nodes. Furthermore, the gene expression profiles of T_eff within GVHD target organs were different from each other, even when comparing individual sub-compartments (e.g. dermis versus epidermis). To exclude the possibility that variations in T_eff resulted from differences in the repertoire of antigens presented in individual organs, we repeated these studies using a second MHC-matched, female into male BMT model involving the transfer of transgenic MataHari (Mh) CD8+ T cells recognizing a single, ubiquitously-expressed male antigen (UTY). Again, we observed similar distinctions between the profiles of T_eff derived from GVHD target organs versus those from secondary lymphoid tissues.

Within the skin, we found that migration of T_eff from the dermis to the epidermis was linked to increased transcription of several effector molecules (e.g. IFN-γ, TRAIL). To assess the role of tissue antigen-presenting cells in driving the epidermal T_eff transcriptosome, we exploited the Langerin-DTR system, in which epidermal Langerhans cells (LC) from the host can be selectively and locally depleted in vivo by administration of diphtheria toxin. We found that LCs were required in situ for epidermal T_eff function and the development of immunopathology in the skin. Transcriptional profiling showed LCs were required for up-regulation of programs for cell cycle commitment, mRNA synthesis and metabolic activity. We have compared our data set to publicly available data sets of immune cells from mice and humans and identified specific modules of co-regulated genes that are controlled by the presence of LC, e.g. the most differentially expressed genes are significantly associated with the Coarse Module 5 (CM5) defined by the Immunological Genome Project. Interestingly, the genes comprised by CM5 are related with T cell proliferation and changes in metabolic activity.

In summary, we have shown that T_eff in GVHD adopt tissue-specific transcriptional programs that are independent of the TCR repertoire. Identification of extrinsic factors that regulate peripheral organ-specific programs offers the potential to selectively block GVHD while maintaining GVL.