Technical Comments

Comment on “Thymic Origin of Intestinal αβ T Cells Revealed by Fate Mapping of RORγt+ Cells”

Science  10 Jun 2005:
Vol. 308, Issue 5728, pp. 1553a
DOI: 10.1126/science.1107363

Intraepithelial T lymphocytes of the murine gut bearing CD8αα homodimers (CD8αα-IEL) have been proposed to originate locally, through a differentiation process initiated in lineage negative (lin) c-kit+IL7Rα+ gut precursors (13). In a recent study, Eberl and Littman (4) claimed that all CD8αα-IEL bearing the αβ T cell receptor (TCR) originated in the thymus and are progeny of CD4+CD8+ [double-positive (DP)] thymocytes. In our opinion, the experiments are not conclusive and the authors' interpretation is not compatible with previously published data.

In their first set of experiments, a Rorct)-EGFP (enhanced green fluorescent protein) construct was used to delete RORγt-expressing cells and their progeny. In Rorct)-deficient mice, c-kit+IL-7Rα+ precursors were absent and TCR-αβ CD8αα-IEL considerably reduced. Expression of a Rorct)-Bcl-xL construct had no effect on linc-kit+IL7Rα+ cells but restored the normal cell cycle and survival of CD8αα TCRαβ IELs. From these experiments, Eberl and Littman concluded that CD8αα TCRαβ IELs could be generated in the absence of gut precursors (4).

In our view, the major caveat of these experiments is the methodology used to identify gut precursors. CD8αα IELs (5, 6) and the vast majority of gut precursors (3) express B200 and CD11c. The authors used antibodies to these markers to deplete Lin+ cells, thus excluding the most abundant local T cell–committed precursors from their studies [Supporting Online Material (SOM), note 1]. This makes it difficult to evaluate whether gut precursors in Rorct)-deficient or Rorct)-Bcl-xL mice may be simply reduced in number or even modified/enriched in particular phenotypes. This discussion of “reduction versus absence” is of fundamental importance in the gut because CD8αα TCRαβ IEL populations represent major expansions of a relatively small number of mature T cell clones (7). Reduction in precursor numbers does not necessarily lead to a reduction in the number of CD8αα IELs, because a low number of precursors can be amply compensated by an increased expansion of CD8αα mature T cells (SOM, note 2).

The second set of experiments in (4) used genetic cell fate mapping. The Cre recombinase was expressed under the control of a short synthetic CD4 promoter, inducing the excision of a STOP cassette and GFP expression in cells (and their progeny) in which the promoter was activated (4). Linc-kit+IL7Rα+ gut cells and CD4CD8CD3 [triple negative (TN)] thymocytes were found to be GFP. Because both DP thymocytes and CD8αα TCRαβ IELs expressed GFP, Eberl and Littman concluded that CD8αα TCRαβ IELs must originate from DP-GFP+ thymocytes (4). This conclusion relies on the assumptions that GFP expression is a reliable marker for Cre-recombinase activity and that this enzyme is first induced at the DP stage. We believe that the authors' data demonstrate otherwise.

Previous data measuring CD4-Cre activity clearly and directly demonstrated that CD4-Cre is fully functional well before the DP stage, because it induced Notch deletion in 60% of TN thymocytes (8). Because only DP cells were GFP+ in (4), we must conclude that GFP does not label all cells in which the CD4-Cre construct is active. This dissociation (Cre activity versus reporter gene detection) represents a major caveat of all genetic fate experiments and has also been described in other deletion-GFP-induction systems (9). Expression of the reporter genes is not sufficient to identify “start cells” (9), because reporter genes may only become detectable well after Cre-recombinase induction. Indeed, for a reporter gene to be detected, the Cre recombinase mRNA and protein have to be previously induced to reach efficient levels to fully delete STOP cassettes. Only then can reporter gene transcription be activated, the reporter mRNA expression induced, and the reporter protein synthesized. Consequently, this lag time may provide ample opportunity for a Cre+ GFP cell to migrate and originate a GFP+ progeny elsewhere. Therefore, conclusions on precursor-progeny experiments can never rely solely on the identification of reporters but rather require direct determination of Cre mRNA expression and Cre activity (SOM, note 3). Because the experiments reported in (4) lacked these controls, it is not possible to establish whether CD8αα TCRαβ IELs are GFP labeled because they derive from DP thymocytes or from Cre+ GFP precursors localized in the gut, bone marrow, or elsewhere. GFP labeling of CD8αα TCRαβ IELs could even be due to the direct expression of Cre recombinase in this mature population (SOM, note 4).

We also suggest that the authors' interpretation is not compatible with previously published data. It was demonstrated that linc-kit+IL7R+ possess TCRβ D-J rearrangements and express GATA-3, Rag-1, and pre-Tα (3, 10). These cells have been shown to generate abundant and mature precursor types in the gut (SOM, note 5), where TCR-β and TCR-γ rearrangements, and subsequently TCR-α rearrangements, could be detected (3). All of these precursor types are present in normal mice and therefore are not an idiosyncrasy of lymphopenic conditions. TCR-α–deficient mice show a severe block in the transition between Lin gut subpopulations, which indicates that TCR-α rearrangements must occur locally (3). c-kit+IL7R+ precursors fully reconstitute both TCR-αβ and TCR-γδ CD8αα T-IEL after adoptive transfer (2, 3). Finally, selective expression of IL-7 in the gut epithelium selectively reconstitutes this γδ compartment, formally demonstrating that TCR-γδ precursors must be present in the gut because TCR-γδ rearrangements occur in situ (11). However, the authors report that CD8αα TCR-γδ IELs were normal in Rorct)-deficient mice, and this, we suggest, contradicts the authors' claim that these mice lack all gut precursors. Thus, clarification of potential relationships between precursor cells of different origins and the gut differentiation processes remains an open question.

Supporting Online Material

www.sciencemag.org/cgi/content/full/308/5728/1553a/DC1

SOM Text

References

References and Notes

Related Content

Navigate This Article