Purpose: Solid organ transplant is lifesaving for children with end-stage heart, liver, or kidney disease. Protection of the transplanted organ from rejection requires long-standing immunosuppression. However, despite our arsenal of immunosuppressive agents, 25% of pediatric transplant recipients have a rejection episode in the first-year post-transplant. The goal of our study was to identify novel and robust surrogate endpoints of allograft status.
Methods: Blood was obtained from pediatric recipients of liver, heart, kidney, or intestinal grafts enrolled at seven sites in the NIAID-sponsored Clinical Trials of Organ Transplantation in Children (CTOTC)-06, a prospective multi-institutional study. PBMCs from 52 children were analyzed, both from patients who had stable graft function (n=24) in the 12-month period before and after sample collection and patients who had an episode of biopsy-proven rejection (n=28) within 30 days after sample collection. These cohorts were well-balanced with respect to clinical characteristics. PBMCs were thawed, barcoded, pooled, and stained with 37 metal-conjugated antibodies against cell surface and intracellular proteins expressed in immune cell lineages and analyzed by single-cell mass cytometry (cytometry by time-of-flight; CyTOF). To identify differences in population frequency between stable and rejection cohorts, we computed cell-type proportions for each population by calculating the number of cells of each population divided by the total number of cells in the respective lineage. 
Results: Principal component analysis revealed that allograft type is an important driver of differences in post-transplant immune composition. Liver and kidney recipients were more like each other and distinct from heart recipients. Heart recipients had higher proportions of CD4 EM, CD4 effector T cells, plasmablasts, and naïve B cells compared to liver or kidney recipients. We observed the same phenomenon using linear discriminant analysis, an independent method of analyzing differences between cohorts.

When controlling for differences in allograft type, we observed an increase in the proportion of CD8 naive and CD8 central memory cells in patients with graft rejection. In this same cohort, we also observed a decrease in the proportion of a novel T cell population (CD45⁺CD3⁺CD19^-CD4⁺CD8^-CD25^loCD5⁺CD38^-FoxP3^loCD45RA^-).

Conclusions: Single-cell CyTOF analysis identified distinct differences in immune response based on allograft status and identified a population whose abundance is associated with increased graft stability. Additional studies will determine if this T cell population can be harnessed to predict or prevent allograft rejection.