TCR-based lineage tracing: no evidence for conversion of conventional into regulatory T cells in response to a natural self-antigen in pancreatic islets

Abstract

Foxp3-expressing regulatory T (T reg) cells derive primarily from selection in the thymus. Yet conversion of mature conventional CD4(+) T (T conv) cell lymphocytes can be achieved in several conditions, such as transforming growth factor beta treatment, homeostatic expansion, or chronic exposure to low-dose antigen. Such conversion might provide a means to generate peripheral tolerance by "converting" potentially damaging T cells that react to self-antigens. We tested this hypothesis in mice transgenic for the BDC2.5 T cell receptor (TCR), which is representative of a diabetogenic specificity that is naturally present in NOD mice and reactive against a pancreatic self-antigen. In the thymus, before any exposure to antigen, clonotype-positive T reg and T conv cells express a second TCRalpha chain derived from endogenous loci. High-throughput single-cell sequencing of secondary TCRs of the Valpha2 family showed their joining CDR3alpha regions to be very different in T reg and T conv cell thymocytes. These specific CDR3alpha motifs, thus, provided a "tag" with which to test the actual impact of T conv to T reg cell conversion in response to peripheral self-antigen; should the autoreactive clonotypic TCR induce T conv to T reg cell conversion upon encounter of cognate antigen in the pancreas or draining lymph node, one would expect to detect tag CDR3alpha motifs from T conv cells in the T reg cell populations. Sequencing large numbers of peripheral BDC(+)Valpha2(+) cells showed that little to no conversion occurs in response to this pancreatic autoantigen.

Figure 1.

Concomitant expression of BDC2.5 and Vα2 TCR chains on both T reg and T conv cells. (A) Flow-cytometric analysis and sorting of CD4SP thymocytes from BDC2.5/TCRα^+/− mice (n = 3). Vα2-expressing T reg and T conv cells were sorted according to the indicated gates. The CD25 versus FoxP3 intracellular staining depicts the frequency of FoxP3 positivity in each gate. (B) Cells were singly sorted into 96-well PCR plates, leaving every third well empty as a control. cDNA made from single cells were split into two PCR reactions for Vα2 and Foxp3 transcripts, as depicted (bottom).

Figure 2.

High-frequency Vα2 TCR sequences from T reg and T conv cell thymocytes. (A) Frequency of individual TCR sequences. The bar histogram depicts the frequency in T conv and T reg cell repertoires (shaded bars above the midline and open bars below, respectively). Bracket represents the contribution of sequences #16–34 to the total T reg cell repertoire. (B) Tabulation of TCR sequences indicating Vα usage, Jα usage, CDR3 length (as defined by the amino acids that are bordered by the invariant C residue in TCRVα genes and the F/W-G-X-G Jα motif), and the percentage of TCR sequences within the T reg and T conv cell populations. Total sequences sampled are noted at the bottom.

Figure 3.

Paucity of T conv to T reg cell conversion in disease-relevant localities. (A) Bar histograms representing TCR sequences from LN, PLN, and pancreas (shaded bars above the midline and open bars below, respectively). Bracket represents the contribution of sequences #16–34 to the total T reg cell repertoire. (B) Table of TCR sequences indicating their frequency (as a percentage) of the indicated population and organ. Total sequences sampled are listed below.

Figure 4.

Mouse-to-mouse and organ-to-organ constancy of the T conv cell repertoire. Heat map of MH indices for each two-way comparison. Color scale indicated above. N represents the number of TCR sequences in the corresponding sample. ACE values estimate the number of unique sequences within the combined population sampled.