Graft-versus-host disease prophylaxis shapes T cell biology and immune reconstitution after hematopoietic cell transplant

Abstract

Successful hematopoietic cell transplant requires immunosuppression to prevent graft-versus-host disease (GVHD), a lethal, T-cell-mediated post-transplant complication. The phase 3 BMT CTN 1703 trial demonstrated superior GVHD-free/relapse-free survival for post-transplant cyclophosphamide (PT-Cy)-based GVHD prophylaxis versus tacrolimus/methotrexate (Tac/MTX), but did not improve overall survival. To compare T-cell biology between GVHD prophylaxis regimens, 324 patients were co-enrolled onto BMT CTN 1801 (NCT03959241). We quantified T-cell immune reconstitution using multi-modal analysis, including T-cell receptor (TCR) sequencing of 2,359 longitudinal samples (180,432,350 T-cells). Compared to Tac/MTX, PT-Cy was associated with an early, substantial reduction in TCR diversity that was sustained for 2 years. PT-Cy led to a T-cell reconstitution bottleneck, including reduced thymic output and virus-associated TCRs. Decreased D+14 TCR diversity predicted prevention of chronic GVHD, but also correlated with increased moderate-to-severe infections. This study reveals how distinct immunosuppression strategies have significant effects on the global immune repertoire, underpinning post-transplant clinical outcomes.

Figure 1:. Overview of the BMT CTN 1801 study.

Patients were co-enrolled with the BMT CTN 1703 study that randomized GVHD prophylaxis to Tac/MTX or PT-Cy. Peripheral blood samples were obtained longitudinally and TCR-sequencing used to assess the diversity of T cell receptors present after transplant.

Figure 2:. PT-Cy impairs the size and diversity of the T-cell repertoire.

a, Global TCR repertoire diversity as measured by inverse Simpson diversity and Shannon entropy was calculated for graft samples (N=101). Comparisons between treatment arms were made using Mann-Whitney U tests. b, Global TCR repertoire diversity as measured by inverse Simpson diversity (left) and Shannon entropy (right) was calculated for each patient at each timepoint (N=324 patients). Log-normalized values were compared between arms using a mixed effects model for repeated measures testing, followed by Fisher’s least significant difference posttests at each timepoint. Asterisks indicate P values for these comparisons, * P<0.05, ** P<0.01, *** P<0.001.

Figure 3:. TCR diversity correlates with post-transplant clinical outcomes.

In a (chronic GVHD requiring systemic immunosuppression) and b (moderate to severe infection), all patients with Day 14 samples available (N=281 patients) were dichotomized into high and low diversity groups using the median inverse Simpson diversity index value. The cumulative incidence of chronic GVHD and infection were plotted through 1 year after transplant. The 1-year estimates of the cumulative incidences are provided by group along with 95% confidence intervals. Gray’s test was used for comparisons between high and low diversity groups. Univariate Fine-Gray regression models were also fit for each outcome and set of diversity groups and used to present a subdistribution hazard ratio for the lower diversity group relative to the higher diversity group. The competing risk for chronic GVHD and infection was death in the absence of the applicable outcome. c, Survival without moderate or severe infection was also plotted through 1 year after transplant using Kaplan-Meier curves. Log-rank test was used for group comparisons and a univariate Cox regression model fit to present a subdistribution hazard ratio. d, Overall survival was assessed for patients dichotomized into high and low diversity groups (N=281 patients) based on median day 14 inverse Simpson diversity and compared using the Kaplan-Meier estimator with a log rank test. All endpoints were censored at 1-year post-transplant or last follow-up, whichever occurred first.

Figure 4:. GVHD prophylaxis shapes the composition of immune reconstitution post-transplant.

The number of T cells (a), naive CD4 (b) and CD8 T cells (c) and CD4 regulatory T cells (d) were measured by flow cytometry (N=40 patients). Data are presented as mean±SEM, and values were compared between arms at each timepoint using mixed effects models for repeated measures testing with Fisher’s least significant difference posttests at each timepoint. Asterisks indicate P values for these comparisons, * P<0.05, ** P<0.01, *** P<0.001. e, UMAP visualization of single cell RNA sequencing of day 28 and day 365 (N=8 for PT-Cy and N=5 for Tac/MTX). T cells colored by T cell subsets (CD4 vs CD8, naïve vs memory) (left). UMAP of T cells colored by the size of the TCR clone detected (log scale) (right). TCR clones were identified as being from the stem cell graft based on annotation of sequences from the leftover graft bag/syringe. f, The frequency of graft-derived clones at the indicated timepoints was calculated for each patient who had a graft sample available for comparison (N=101 patients). g, Schematic of the T cell bottleneck effect following treatment with PT-Cy.

Figure 5:. Origin of T cells building the post-transplant immune compartment.

a, Intracellular cytokine staining followed by flow cytometry was used to assess Ki67 expression in T cells from each subset (N=40 patients). The frequency of proliferating CD4+ naïve T cells is displayed as mean±SEM at each timepoint. Asterisks indicate P values for the comparison between arms at each timepoint using mixed effects models for repeated measures testing with Fisher’s least significant difference posttests at each timepoint, * P<0.05, ** P<0.01, *** P<0.001. b, The Morisita Index, a measure of repertoire stability/turnover, was calculated for each pair of timepoints listed; higher values indicate a more stable TCR repertoire (N=324 patients). TCR clones present at each of Day 28 (Panel c) and Day 63 (Panel d) were plotted at all timepoints (N=264 and 238 patients, respectively). e, Thymic output measured by quantification of T cell receptor excision circles (TREC) from genomic DNA is shown as box and whisker plots stratified by study arm (N=291 patients). Values were compared between arms at each timepoint using multiple Mann-Whitney U tests. f, The number of singleton TCRs associated with cytomegalovirus, Epstein-Barr virus, SARS-CoV-2, herpes simplex virus-1, herpes simplex virus-2 and parvovirus are displayed (N=324 patients). Data are represented as mean ± SEM and comparison between groups made by Mann-Whitney U test. Asterisks indicate P values for these comparisons, * P<0.05, ** P<0.01, *** P<0.001.