Research

The human intestine is exposed to a vast array of molecules from the microbiome, diet, and pathogens. However, it remains unknown how the intestinal immune system maintains tolerance to beneficial microbes and nutrients while generating an inflammatory response to pathogens. This gap in knowledge has hindered the rational design of therapeutics for diseases that arise from the disruption of the gut immune system - including IBD, food allergy, and immunotherapy-resistant cancer. Our research has two branches, both of which attempt to understand the gut immune system at the level of molecular mechanisms.

​

(1) Building a model system of the gut microbiome

Many previous studies characterized the function of gut bacteria under the artificial condition of mono-colonization using germ-free mice. However, it is unknown how each bacterial strain affects the immune system in the context of a bacterial community.  To address this question, we are developing a synthetic model of the gut microbiome with the following features:

i) The strains are defined.

ii) The bacterial community is complex enough (>100 strain scale) and taxonomically resembles the human microbiome.

iii) The community mimics the immune-modulatory function of the human microbiome.

​

(2) Exploring T cell response at the repertoire level.

T cells orchestrate the immune responses by recognizing a molecular cue, an antigen, from bacteria and diet using the T cell receptor (TCR). However, our current understanding of T-cell responses in the gut is based on a small subset of bacteria and dietary antigens, like SFB and OVA. Mapping of TCRs to antigens has never been achieved in the physiological condition of a bacterial community and complex diet. We are exploring this point with two approaches.

​

In vivo: We raise mice under the condition of a defined bacterial community and/or a defined diet so that we can test each factor of the gut ecosystem experimentally. We profile the immune response against each individual bacterial strain or ingredient in a diet at the single TCR level. We develop experimental tools for high-throughput antigen discovery. We use our T cell profiling results to design therapeutic interventions – adding or dropping out bacteria or dietary components from the gut ecosystem - to skew immune reactions toward tolerance or inflammation.

​

Computational: TCRs are extremely diverse, making it difficult to predict an antigen from primary TCR sequences. However, if we identify many TCRs that target the same antigen, we find a certain pattern, a "specificity group", in the TCR sequences. We are collecting many examples of TCR specificity groups for antigens from the microbiome and diet. This extensive dataset allows us to build a computation pipeline that is capable of inferring T cell antigens from TCR sequences.