TCR Tg cells commonly are used in the assessment of the properties of T cells. They are easy to manipulate and to visualize. Their defined TCR expression allows one to monitor the same clone throughout the immune response (39
). This characteristic is fundamental to determine if the changes in population properties throughout the response are due to the selection of particular clones of antigen-specific cells. However, it was suggested recently that when TCR Tg cells are present at high precursor frequencies, their intraclonal competition for antigen leads to their suboptimal activation and abnormal differentiation (2
). Several reasons prompted us to reexamine this claim in greater detail. First, several studies showed that a short-term contact with an antigen is sufficient to trigger a complete CD8 differentiation program (18
), and that extensive CD8 expansion is not a prerequisite for efficient memory generation (3
). Second, in other studies effector CD8 numbers seem to hit a similar ceiling regardless of initial variability in precursor numbers or specificity (22
), arguing for stimulation-tailored rather than T-cell-intrinsic differentiation pathways. Finally, other data suggested alternative explanations to the different behavior of high- and low-density TCR Tg cell transfers. The adoptive transfers of >105
TCR Tg cells have been shown to alter the kinetics of pathogen clearance and the timing of peak CD8+
T-cell expansion (10
). Since high- and low-dose transferred populations were studied systematically in different recipients where antigen loads and antigen clearance are known to be different, alterations in the course of infection could account for the different population properties in mice that received different numbers of TCR Tg cells.
Competition for antigen and clonal competition also occur in normal immune responses and contribute to the immunodominance hierarchy observed. A partial or complete compensation for a loss of a particular epitope by other specificities has been known to occur (1
). In some circumstances it has been suggested that the cytokine-mediated active suppression of dominant clones over subdominant ones (immunodomination) occurs (48
), but the existence of such active immunosuppression still is disputed (24
). Besides, although immunodominance has been studied widely in many infectious models, it still is unclear whether dominant and subdominant populations diverge in their functional capacities and protection capabilities (6
). To address these issues, in addition to conventional tests, we performed a powerful single-cell multigene expression study of several antigen-specific populations during the course of LCMV infection in mice. When studied in the same infectious context in the same mice at the same time point of the response, the T-cell populations of different specificities and present at different frequencies showed remarkably similar features. Thus, except at the earliest stages of infection (days 4 to 5) when GP276-specific cells expressed less granzyme A mRNA than NP396- and GP33-specific cells, dominant and subdominant cell effector and memory had remarkably similar cytokine (Ifng
, and Tgfb1
) and cytotoxic gene expression (Prf1
, and Fasl
) and coexpression profiles. Previous comparisons of cytokine profiles after in vitro stimulation also failed to reveal major differences (49
), and we found that cell surface markers’ expression most frequently was overlapping. As an exception, the subdominant GP276-specific population showed some delay in CD27 upregulation and KLRG1 downregulation at day 8, but these differences disappeared in the memory phase, when these cells’ phenotypes were equivalent to those found in GP33-specific cells. Conversely, the NP396-specific memory cohort usually had a larger fraction of CCR7−
cells than cell populations with other peptide specificities, but otherwise they expressed the same KLRG1 and IL-7R labeling, and it was reported previously that this cell type eventually also upregulates the expression of both of these ligands. Overall, these data directly argue against the hypothesis that dominant and subdominant populations follow disparate differentiation pathways. These findings were confirmed even when major differences in clonal abundance were introduced artificially by the adoptive transfer of Tg cells.
The differentiation profiles of monoclonal T-cell populations recently have fallen under scrutiny, since several reports suggested that the artificial introduction of TCR Tg CD8 cells in numbers exceeding those of endogenous cells of similar epitope specificity (5
) resulted in the inadequate differentiation of TCR Tg cells (2
). These reports, however, focused mainly on CD62L and IL-7R expression analysis, and functional assays were performed only at a single time point of the infection. These studies also did not take into consideration possible differences in response kinetics that could result from the introduction of a large cohort of naïve Tg cells. Indeed, abundant and rare clone behavior always was studied in different mice, where Tg cells could be submitted to different antigen loads and abundant and rare clone accumulation peaked at different time points (2
). Supporting the notion that previously reported differences between high- and low-dose transfers can be explained by a different response kinetics, adoptive transfers of >105
precursors were shown to accelerate the kinetics of pathogen clearance and CD8 expansion (10
Contrary to those studies, we compared TCR Tg and endogenous cells of the same epitope specificity from the same animals, where both faced exactly the same antigen exposure and showed similar response kinetics. Moreover, when Tg cells are present, the endogenous GP33-specific population expands very little, which should prevent any TCR downregulation early in the response. We found that under these conditions, Tg and endogenous GP33-specific CD8 cells retrieved from the same mice always were remarkably similar. They not only had initiated IL-7R and CD62L upregulation precociously at day 8 but also showed similar phenotypes and gene expression profiles at the response peak. The analysis of CD8 T cells with other specificities in these transferred mice also supported the notion that high-dose transfers only accelerate response kinetics. Indeed, we found that in P14-injected mice both NP396- and GP276-specific populations also had initiated IL-7R and CD62L upregulation at day 8 after infection. Surprisingly, these cells appeared to be otherwise unaffected by the presence of high frequencies of TCR Tg cells. Their frequency and their capacity to produce IFN-γ was similar in mice left untreated or receiving P14 Tg cells. These results demonstrate that high-frequency adoptive transfers do not inhibit overall endogenous responses but only influence the expansion of T-cell populations with the same TCR specificity.
Our results also do not support the notion that high-frequency transfers induce major modifications in the properties of memory cells. We demonstrated that memory cells on day 90 that arose from 5 × 103
and 5 × 105
P14 cells did not differ in functional capacities such as stimulation-induced cytokine secretion. We did not find evidence for the predominant generation of CD62L+
Tg memory cells in high-frequency transfers. In our hands, the GP33-specific endogenous memory cells frequently expressed more CD62L than the Tg memory cells. Differences between the present and previously published results (2
) could be due to mouse-to-mouse variability, as we found in our experiments, or to the fact that we always evaluated endogenous and Tg cells present in the same mouse. Our results suggest that studies showing a preferential expression of CD62L in high-frequency cells were not exhaustive, and that the conclusion that these cells only generate CD62L+
(a major argument to suggest abnormal differentiation) is unreliable.
Overall, these data suggests that high-frequency adoptive transfers just accelerate response kinetics, and that Tg cells only compete with the endogenous cells that share the same TCR specificity. It is likely that such competition is greatly influenced by the relative avidity/cross-reactivity of the TCR Tg cells with respect to the average avidity/cross-reactivity of the endogenous antigen-specific cells. Different Tg CD8s populations were classified according to these parameters in the hierarchy OT1 > P14 > anti-HY (13
), which appears to correlate directly with their inhibitory effect on endogenous responses. Indeed, the transfer of the high-avidity/cross-reactive OT-1 clone virtually abrogates endogenous responses, while P14 transfers have a smaller effect (2
). In contrast, in high-frequency anti-HY Tg transfers to normal mice, the endogenous cells partially outcompete the Tg population. Both Tg and endogenous responses show reduced amplitude and become similarly represented in the overall anti-HY response (47
TCR downregulation is a rapid and dose-dependent corollary of T-cell activation in vitro (43
) but is rather transitory, lasting for about 24 h. TCR downregulation also was detected in acute infections in vivo (8
), but due to the lack of other markers to identify antigen-specific cells, these previous studies could not evaluate fully the extent of this phenomenon. Here, we established that Tg cells identified by an allogeneic marker, even when present at physiologic frequencies, downregulated TCR expression, and a major fraction fully lost TCR cell surface expression and failed to bind tetramers. This behavior is likely a common feature of CD8 immune responses, since we also found it in other infectious models and in other TCR Tg cells (P14 or OT-1 cells immunized with Listeria
-expressing GP33 [LM-GP33] or LM-OVA, respectively; unpublished data).
Several aspects of this phenomenon must be emphasized. In contrast to the transient loss of TCR after in vitro activation, in vivo responding populations could remain TCR negative throughout a long time period during the expansion phase; activation status and tetramer binding were inversely correlated, allowing for the possibility that more activated cells could be rendered completely invisible by prominent TCR downregulation.
To summarize, the detailed analysis of CD8 T cells responding to different LCMV epitopes in the same infectious environment showed that relative clone abundance or TCR specificity did not alter substantially the properties of effector and memory cells. From this perspective, the current notion that high-frequency transfers of naïve Tg cells induce abnormal T-cell differentiation must be toned down. We found that differences in Tg behavior can be explained by a different response kinetics, that abundant Tg and rare endogenous cells with the same peptide specificity had overlapping properties, and that Tg cells did not affect the amplitude or the quality of the endogenous response to other LCMV peptides. It also was demonstrated recently that high-frequency transfers did not affect the quality of the memory responses (53
). In contrast, the use of TCR Tg cells that can be recognized by allotype markers revealed that during acute infection, when high viral loads are present, a substantial fraction of responding cells downregulate their TCR and fail to bind MHC tetramers, and that tetpos
cells have different properties. Therefore, TCR-Tg mice may be fundamental for the evaluation of the entirety of the early immune response.
Finally, the important and long-lasting loss of TCR expression we found to occur during the expansion phase has major implications for our capacity to study early events in the vast majority of acute infections in the mouse (when Tg cells are not available) and, more importantly, in humans. Studies based on the tetramer binding identification and/or magnetic bead purification of antigen-specific cells likely are incomplete and biased (14
), since they select subpopulations with peculiar properties that do not represent the overall characteristics of the responding peptide-specific cohort. Moreover, it is at present unclear if any of the methods currently used to identify responding cells will be able to do so and in which circumstances. The failure to bind tetramers is due to TCR downregulation. It therefore is possible that the vast majority of tetneg
cells also are undetected through cytokine expression after in vitro stimulation with specific peptides, since these responses depend on the cell surface expression of the peptide-specific TCR. Moreover, we found that TCR downregulation increased when higher doses of virus were injected, suggesting that the higher the viremia the more incomplete will be our assessment of the acute response. This important pitfall must be taken into consideration: we may fail to detect a major cohort of responding cells when high virus loads are present.