AFFILIATIONS
1 City of Hope Hospital, Duarte, California, United States
2 City of Hope Hospital, Duarte, California, United States
3 City of Hope Hospital, Duarte, California, United States
4 City of Hope Hospital, Duarte, California, United States
5 City of Hope Hospital, Duarte, California, United States
6 City of Hope Hospital, Duarte, California, United States
7 City of Hope Hospital, Duarte, California, United States
8 City of Hope Hospital, Duarte, California, United States
9 City of Hope Hospital, Duarte, California, United States
10 City of Hope Hospital, Duarte, California, United States, Fred Hutch Cancer Center, Seattle, Washington, United States
11 City of Hope Hospital, Duarte, California, United States
12 Memorial Sloan-Kettering Center, New York, New York, United States
13 City of Hope Hospital, Duarte, California, United States
14 City of Hope Hospital, Duarte, California, United States
15 City of Hope Hospital, Duarte, California, United States
16 City of Hope Hospital, Duarte, California, United States, Karolinska Institute, Solna, Sweden, Karolinska Universitetssjukhuset, Solna, Sweden
ABSTRACT
Failure of thymic regeneration after cytoreductive therapy and allogeneic hematopoietic cell transplantation (allo-HCT) produces prolonged T-cell lymphopenia. While no regenerative therapies are currently approved, prior work implicates thymic regulatory T (Treg) cell expansion in immune reconstitution. Treg cells support regeneration in many tissues, yet whether they deploy tissue-specific effector programs after injury is unknown.
We applied paired single-cell RNA and TCR sequencing of Treg cells from target organs of transplant-related toxicity: matched thymus, spleen, bone marrow, liver, lung and skin at day 4 after sublethal total body irradiation (SL-TBI, 550 cGy). Two conserved states resolved Tregs across all tissues: quiescent (Sell, Foxp1, Bach2) and regenerative (Pdcd1, Icos, Tigit). Clonal expansion was confined to the regenerative compartment (p=0.02), and shared αβ-TCR clonotypes connected all tissues, indicating a single recirculating Treg pool across organs. We then asked whether effector output was shared or tissue specific. Supervised factor analysis resolved distinct tissue-restricted regenerative programs, with Trefoil factor 1 (Tff1), a small, secreted trefoil-domain peptide, as the top thymus-restricted factor. Intrathymic TFF1 protein rose after SL-TBI, coinciding with local Treg expansion, and a Tff1GFP reporter confirmed the highest TFF1+ frequency in thymic Treg cells. Among IL-1 family receptors on thymic Tregs, Il1rl1 (ST2) correlated most with Tff1. Thymic fibroblasts were the dominant Il33 source, IL-33 but not IL-18 induced TFF1-GFP in thymic Treg cells in vitro (p<0.01), and fibroblast-restricted Il33 deletion (PdgfraCreIl33fl/fl) abrogated thymic TFF1 after injury.
To separate local stimulation of Tregs from clonal identity, we compared clonotypes shared between thymus and at least one other tissue. Tff1 expression was higher in the thymic member of the same clone (p=1.3e-10). Furthermore, Tff1CreERTR26LSLtdTomato fate mapping supported a transient, environment-induced state, since after injury, cells with a history of Tff1 expression distributed equally across all tissues.
We next generated germline Tff1−/− and conditional Foxp3CreYFPTff1Fl/Fl mice. Both mice had normal homeostatic thymic cellularity but reduced thymic cellularity at day 7 after SL-TBI (p<0.01, n=14–17), with reduced DN, DP and SP8 and preserved SP4 thymocytes, while retaining normal post-injury Treg numbers. In allo-HCT (LP/J into B6 or Tff1−/−), loss of Tff1 led to reduced thymic cellularity and DP thymocytes at day 14 (p<0.05, n=9–10). Thymic epithelial cell (TEC) numbers fell in both models (SL-TBI p<0.01; allo-HCT p<0.05). Day 4 single-cell RNA-seq of TECs showed a selective loss of cortical TECs with loss of cell-cycle programs and downregulation of Pax1, Pax9, Foxn1 in Tff1−/− mice, confirmed functionally by BrdU incorporation in vivo (p<0.01).
Conversely, recombinant TFF1 induced proliferative gene programs in primary TEC organoids. Tff1-overexpressing organoids showed increased phospho-ERK1/2 and phospho-p38. ATAC-seq in these organoids revealed gain of PAX-family motif accessibility with reduced accessibility at stress-response genes and FOS and JUN motifs. CUT&RUN confirmed increased occupancy of these transcription factors at cell-death loci in non-Tff1-treated TECs.
To test therapeutic delivery, Tff1 was overexpressed in T cell precursors and adoptively transferred 3 hours after injury. At day 7, recipients had increased thymic cellularity compared with empty-vector controls and increased TEC numbers (p<0.001), with equivalent numbers of transduced cells recovered, indicating that the benefit was not attributable to homing, survival or expansion of the transferred cells.
In human thymus subjected to ex vivo injury, TFF1 protein increased in culture supernatant, and recombinant human TFF1 induced MAPK activation with upregulated proliferative and downregulated stress associated transcriptional programs in thymic stromal cells.
Recirculating Treg cells thus share a conserved regenerative identity after injury but deploy distinct, tissue-restricted effector programs. In the thymus this program converges on TFF1, which acts directly on thymic epithelium to sustain proliferation, and PAX-associated TEC stemness. TFF1 accelerates thymic recovery, providing a strategy to shorten lymphopenia after cytoreductive therapy and allo-HCT.
DISCLOSURES
Jonah Pierce: No relevant conflicts of interest to declare.