Summary
Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is potentially curative for many blood disorders
but is often complicated by graft-versus-host disease (GvHD). Despite major advances in therapeutic strategies,
GvHD-related morbidity and mortality remain high, and better tools to predict the GvHD risk are urgently needed.
The immunologic basis of GvHD is the recognition by donor T cells of minor histocompatibility antigens (mHAgs),
arising from nonsynonymous single nucleotide polymorphism (SNP) differences between donor and recipient.
However, only some 50 mHAgs have been reported thus far. Current advances in the characterization of
germline and somatic events within coding exomes through next-generation sequencing, coupled with
improvements in HLA class I and II epitope prediction, offer an unprecedented opportunity to systematically
identify candidate mHAgs on an individualized basis. We hypothesize that identification of tissue-specific mHAgs,
predicted through baseline analysis of patient and donor DNA, can enable a personalized genomics-informed
risk assessment of GvHD, with the potential to help refine current donor selection algorithms and to enable
personalized tailoring of post-transplant immunosuppression. Our aims are thus to: (1) Build an epitope
prediction pipeline to systematically identify mHAgs, incorporating state-of-the-art tools for SNP discovery
from WES, expression filters designed through comprehensive analysis of single-cell RNA-seq and proteomic
profiling of a set of commonly targeted GvHD tissues (i.e. skin, liver, colon and lung), inflammation signatures,
and Y-chromosome specific expression (in the case of female¿ male transplants). This will be then coupled with
HLA class I and II prediction. Epitope predictions will be confirmed through direct detection of HLA-bound
peptides by immunopeptidome-based evaluations of GvHD-affected target tissues. (2) Define and track the T
cell responses to predicted mHAgs in patients with GvHD. To link prediction of mHAg targets with functional
T cell responses, we will establish proof-of-concept by testing for mHAg-specific T cell responses across three
clinically relevant settings: (i) patients experiencing organ-specific GvHD, immunoproteomically characterized in
Aim 1; (ii) patients who received sequential allo-HSCT from 2 distinct HLA-identical donors and experienced
GvHD following one but not the other transplant; and (iii) patients who received post-transplant
cyclophosphamide after matched related donor (MRD) HSCT as part of GvHD prophylaxis. (3) Test the
feasibility of mHAg prediction to generate a prognostic GvHD risk score. To determine if individual mHAg
burden relates to clinically observed GvHD, we will focus on patients with myeloid neoplasia, and perform WES
analysis on a discovery set of DNA from ~200 MRD recipient-donor pairs from patients treated at DFCI, and then
on an extension set of ~200 pairs from URD HSCT, from whom information regarding GvHD occurrence is
available. Through a planned validation analysis of ~600 other pairs (also myeloid neoplasia) from CIBMTR, we
will assess if this approach is applicable cross-institutionally, and for patients across ethnicities.