We investigated the mechanism by which α/β T cells expand upon transfer to T cell–deficient host mice by injecting carboxyfluorescein diacetate succinimidyl ester–labeled T cells into mice depleted of T cells by sublethal irradiation. We found that CD4+ T cells divided when transferred to irradiated hosts and that the division of more than half of these cells required class II expression. However, division of transferred CD4+ T cells did not occur in irradiated hosts that expressed class II molecules occupied solely by the peptide responsible for thymic selection, indicating that peptides distinct from those involved in thymic selection cause the division of CD4+ T cells in irradiated mice. These data establish that class II–bound peptides control the expansion of CD4+ T cells transferred to T cell–deficient hosts and suggest that the same peptides contribute to the maintenance of T cell numbers in normal mice.
The number of mature T cells in an adult mouse is controlled by the rates of T cell production and division, and by the factors that influence T cell survival. Each of these processes involves interactions between TCR and MHC molecules. The production of new T cells occurs in the thymus 1, where contact between newly expressed α/β TCRs and self-peptides presented by MHC molecules selects thymocytes to mature 2,3,4,5,6,7. TCR–MHC interactions also affect mature T cell survival. Studies in which T cells were transferred to mice that lack MHC expression have shown that a normal CD4+ or CD8+ T cell life span depends on peripheral expression of the selecting MHC molecule 8,9,10,11,12. Whether the peptides bound to MHC proteins also affect T cell survival has not been addressed, but at least one study suggests that the selecting peptide–MHC ligand may be sufficient to promote a normal T cell life span 9.
The size of the T cell pool is also affected by the division of mature T cells 13,14,15,16. In normal mice, most T cells divide infrequently, such that approximately half the T cells in a mouse divide within a period of 3–8 wk 17,18. However, if a T cell deficit exists, the rate of T cell division increases for unknown reasons. For example, when T cells are transferred to nude, irradiated, SCID or TCR-α knockout (KO)1 mice, the transferred cells expand rapidly 19,20,21,22,23. Whether such expansion requires antigen presence and/or MHC expression has been the subject of several reports. Studies have shown that naive transgenic CD8+ T cells specific for the male HY antigen expand in normal mice only in the presence of antigen 12,24. A different study demonstrated that 2B4 transgenic CD4+ T cells specific for pigeon cytochrome c in the context of Ek expand in irradiated, bone marrow–reconstituted mice only in the presence of antigen 23. This study also established that expansion of naturally occurring CD4+ T cells after transfer to irradiated mice is diminished when class II blocking antibody is administered to the host. The finding that CD4+ T cell expansion in T cell–deficient hosts requires class II expression has been confirmed by a recent report showing that CD4+ T cells could not expand in TCR-αKO mice lacking class II expression 22. Whether different peptides presented by class II molecules influence this process is unknown. Because TCR transgenic T cells require antigen to expand in T cell–deficient mice, peptides presented by class II molecules are presumably required for naturally occurring CD4+ T cells to divide when transferred to T cell–deficient mice. However, the involvement of class II–bound peptides has not been confirmed for normal T cells.
The purpose of our study was to determine whether class II–bound peptides influence the division of CD4+ T cells transferred to T cell–deficient mice. Because we sought a direct measure of the extent of division of individual T cells after transfer to each host, T cells were labeled with the dye carboxyfluorescein diacetate succinimidyl ester (CFSE) before transfer to various sublethally irradiated hosts. Due to the approximately twofold decrease in CFSE fluorescence after each cell division, CFSE allowed us to assess the extent of division of individual transferred T cells 25,26. We found that a significant fraction of CD4+ T cells divided when placed into irradiated normal mice. Most of the CD4+ T cell division that occurred in irradiated normal mice failed to occur in irradiated MHC class II KO (MHCIIKO) mice, confirming that CD4+ T cell division in T cell–deficient mice requires MHC class II expression. When the peptide presented by class II in the irradiated recipient was restricted to the peptide responsible for positive and negative selection of the T cells, division of the transferred CD4+ T cells did not occur. These experiments establish that peptides distinct from those involved in thymic selection control CD4+ T cell expansion in T cell–deficient mice and suggest that these same peptides may influence the T cell pool in normal mice by causing mature CD4+ T cells to divide.
Materials And Methods
Reagents and Mice.
B6.PL-Thy1a/Cy (B6.PL) and BALB/cByJ (BALB/c) mice were purchased from The Jackson Laboratory. C57BL/6NTacfBR (B6) and C57BL/6TacfBR-[KO]Abβ (MHCIIKO) animals were purchased from Taconic Farms. Ab EpTKO (β2-microglobulin KO, wtAbKO, IiKO, AbEp+), TKO (β2-microglobulin KO, wtAbKO, IiKO), wtAb+DKO (β2-microglobulin KO, wtAb+, IiKO), and DO TCR+ RAG2KO BALB/c mice were bred under specific pathogen-free conditions in the Biological Resource Center (BRC) at National Jewish Medical and Research Center. CFSE was purchased from Molecular Probes Inc. PE-labeled mAbs against Thy1.1 and CD4, cy-chrome (CyC)-labeled anti-CD8, allophycoerythrin-labeled anti-CD4, and CyC-labeled streptavidin were purchased from PharMingen. FITC-labeled and biotinylated anti–TCR-β (H597) and anti-DO 11.10 idiotype (KJ1) were prepared in our laboratory. Cells were stained and analyzed as previously described 5,27.
T Cell Transfer.
T cells were purified to >95% using Cellect Mouse T columns (Biotex Labs.). They were stained with CFSE by incubation in 1.5 μM CFSE in balanced salt solution for 15 min at 37°C at 107 cells/ml, washed once, and transferred intravenously. Mice were irradiated immediately before transfer with 450 rads (137Cs source; AEC). Similar results were obtained if transfer was delayed until 36 h after irradiation.
T Cells Divide When Transferred to T Cell–deficient Mice.
The effect of sublethal irradiation of host mice on the rate of division of T cells transferred to the host was determined. T cells were purified from the lymph nodes of B6.PL (Thy1.1+) mice, labeled with CFSE, and injected into B6 (Thy1.2+) recipients that either had or had not been sublethally irradiated. CD4+ and CD8+ T cells that were transferred to normal, nonirradiated B6 mice showed little evidence of division 7 d after transfer (Fig. 1). However, CD4+ and CD8+ T cells that were transferred to irradiated B6 mice divided substantially during the same period. In irradiated B6 hosts, CD8+ T cells divided more quickly than did CD4+ T cells, although a small subpopulation of CD4+ T cells divided so many times that it had almost lost its CFSE fluorescence (Fig. 1, histogram region A). These data showed, as expected from previous studies, that transferred T cells divided more quickly in irradiated host mice than in nonirradiated host mice.
Because whole-body irradiation affects many tissues, it is possible that the T cell division observed in irradiated mice did not result from T cell deficiency, but instead from additional effects of irradiation. To be certain that a T cell deficiency results in the increased division of transferred T cells, CFSE-labeled T cells were transferred to mice that lacked T cells due to induced genetic mutations. CFSE-labeled T cells from BALB/c mice were transferred to RAG2KO (α/β-TCR T cell−, γ/δ-TCR T cell−, B cell−) mice, and CFSE-labeled T cells from B6.PL mice were transferred to TCR-βKO (α/β-TCR T cell−, γ/δ-TCR T cell+, B cell+) mice. Both RAG2KO and TCR-βKO mice lack endogenous α/β T cells. 7 d after transfer, CD4+ and CD8+ T cells had divided in both RAG2KO and TCR-βKO recipient mice (Fig. 2). However, the rate of CD4+ T cell division was much faster in the RAG2KO and TCR-βKO hosts than it was in irradiated hosts. Because the rates of T cell division in irradiated and genetically T cell–deficient hosts differed, it is possible that the mechanism responsible for division in each host also differed. However, together these experiments indicate that the rate of α/β-TCR+ CD4+ and CD8+ T cell division increased in response to an α/β T cell deficit.
Most CD4+ T Cell Division in Irradiated Mice Requires MHC Class II Expression.
To determine whether division of T cells that were transferred to irradiated mice required TCR–MHC interactions, we tested whether CD4+ T cells divided in hosts that did not express MHC class II molecules. Purified B6.PL lymph node T cells were transferred to nonirradiated or irradiated B6 or MHCIIKO mice. Analyses 7 d after transfer showed that transferred CD4+ T cells did not divide in nonirradiated B6 or MHCIIKO hosts (Table). However, transferred CD4+ T cells did divide in both irradiated B6 and irradiated MHCIIKO hosts, although more division occurred in irradiated B6 hosts than in irradiated MHCIIKO hosts (Fig. 3, summarized in Table). 55% of transferred CD4+ T cells divided in irradiated B6 hosts, whereas only 22% of transferred CD4+ T cells divided in irradiated MHCIIKO hosts. As shown in Table, the difference in the percentage of transferred CD4+ T cells that divided in irradiated B6 and irradiated MHCIIKO mice was consistent over many experiments. These data indicate that most of the division that occurred among CD4+ T cells transferred to irradiated B6 mice required class II expression, which confirmed previous studies of CD4+ T cell expansion in T cell–deficient mice 22,23.
Although most of the CD4+ T cells that were transferred to irradiated hosts required class II expression to divide, some CD4+ T cells divided in the absence of class II expression. Among the transferred CD4+ T cells that divided in irradiated MHCIIKO mice were two groups of cells. The first group of CD4+ T cells that divided in irradiated MHCIIKO mice went through a single round of cell division. This single round of division may represent a response by the T cells to the cytokine environment generated in the host by irradiation 28,29,30. The second group of CD4+ T cells divided many times in MHCIIKO mice (see Fig. 1, histogram region A). During and after division these cells expressed high levels of CD44 (data not shown), which suggests that antigen, possibly presented by class I or nonclassical MHC molecules, caused these cells to divide. The rapidly dividing cells were excluded from the calculations presented here because they appeared to represent a distinct phenomenon.
After the transfer of T cells to irradiated mice, the total number of transferred T cells recovered from the lymph nodes and spleen of an irradiated recipient was typically 10–20% of the total number of cells that were initially transferred. To exclude the possibility that transferred T cells trafficked to and divided within tissues other than spleen and lymph node, lymphocytes were prepared from the liver, lung, and small intestine of irradiated B6 or irradiated MHCIIKO mice that had been injected with CFSE-labeled B6.PL T cells 7 d earlier. The total number of transferred T cells obtained from these tissues represented ∼4% of the number initially transferred, but the T cells obtained from these tissues had divided less than had the T cells in the lymph nodes and spleen from the same mice (data not shown). Thus, trafficking of transferred CD4+ T cells to tissues other than the spleen and lymph node did not influence the division of CD4+ T cells in irradiated B6 or MHCIIKO mice. This finding provided additional support for the conclusion that most CD4+ T cells required class II expression to divide in irradiated hosts.
The Peptide–MHC Ligand Responsible for Thymic Selection Does Not Cause Division of CD4+ T Cells in Irradiated Mice.
To assess whether peptides influenced MHC class II–dependent CD4+ T cell division, we exploited mice that express a single class II–peptide combination, Ab covalently linked to a peptide from Eα (AbEp), and no other class II or class I proteins, and no invariant chain (Ii) (AbEpTKO). A recent report involving mice constructed to express a single peptide bound to class II molecules by a different method suggested that such mice contained many class II–bound peptides 31. Experiments similar to those performed in this report indicate that this is not the case in AbEpTKO animals (Marrack, P., unpublished data). All tests of Ab EpTKO mice indicate that positive and negative selection of all CD4+ T cells in AbEpTKO mice occurs only on the AbEp molecule, and that the APCs from these mice are unable to present any other self- or foreign peptide in association with class II molecules due to 100% occupancy of Ab by Eα 5,32. Because a normal array of self-peptides is not presented by class II during negative selection of CD4+ T cells in AbEpTKO mice, mature CD4+ T cells in these mice are not tolerant to the self-peptides presented by class II in a wild-type Ab+ mouse. Consequently, a large percentage of AbEpTKO CD4+ T cells react strongly to the normal self-peptides presented by Ab 5. As discussed below, we used this reactivity as a control to show that AbEpTKO T cells were capable of division.
To determine whether interactions between TCR and the class II–peptide ligand responsible for thymic selection could induce division of CD4+ T cells in irradiated mice, AbEpTKO T cells were labeled with CFSE and transferred to irradiated hosts of three genotypes: (i) mice that expressed no MHC class I, class II, or Ii (TKO); (ii) mice that expressed only the selecting ligand (AbEpTKO mice); and (iii) mice that expressed wild-type Ab, but no class I or Ii (wtAb+DKO). 7 d after transfer, host lymph node and spleen cells were stained for TCR and CD4 and analyzed. In each of these hosts a population of cells stained positively for TCR and CD4 but were CFSE negative. These cells were derived from the host, not the donor, and they appear in Fig. 4 because there was no surface marker that could be used to distinguish between host and donor CD4+ T cells.
As expected, few CD4+ AbEpTKO T cells divided when transferred to irradiated TKO hosts, which do not express class II molecules (Fig. 4, summarized in Table). CD4+ AbEpTKO T cells also failed to divide when transferred to irradiated AbEpTKO hosts, which express the class II–peptide ligand responsible for thymic selection. In fact, on average, more of the AbEpTKO T cells divided in irradiated TKO recipients than in irradiated AbEpTKO recipients (Table). Thus, the selecting ligand had no effect on the rate of division of AbEpTKO T cells transferred to irradiated recipients. The lack of division by the transferred AbEpTKO T cells was not due to a defect in the AbEpTKO T cells themselves; AbEpTKO T cells divided extensively in irradiated wtAb+DKO hosts, which express wild-type Ab bound to peptides to which AbEpTKO T cells are not tolerant (Fig. 4). These data showed that interactions between TCR and the specific peptide–MHC ligand responsible for thymic selection did not cause division of T cells in irradiated hosts.
A recent study found that although naive transgenic CD8+ T cells required antigen to divide in irradiated hosts, memory transgenic CD8+ T cells divided in irradiated mice in the absence of specific antigen 12. To address whether the peptide–MHC ligand responsible for thymic selection is capable of causing the division of memory CD4+ T cells, Vβ8+ T cells from AbEpTKO mice were activated before transfer by injecting the mice with 150 μg staphylococcal enterotoxin B (SEB) and 7 μg LPS. Injection of SEB and LPS resulted in the activation of Vβ8+ T cells, as assessed by the increased number of Vβ8+ cells observed 2 d after SEB injection (data not shown). 1 wk after the SEB injection, T cells from these AbEpTKO mice were purified, labeled with CFSE, and transferred to irradiated AbEpTKO, TKO, and wtAb+DKO recipients. 1 wk later, the amount of division that had occurred among transferred Vβ8+ CD4+ cells was determined (Table). The data indicate that little division of previously activated Vβ8+ CD4+ T cells occurred in irradiated mice that lacked MHC expression or expressed only the selecting MHC–peptide ligand. As expected, the transferred Vβ8+ CD4+ cells divided rapidly in wtAb+DKO mice, indicating that the lack of division of these cells in AbEpTKO and TKO mice was not due to anergy induced by SEB priming (data not shown). Thus, unlike memory CD8+ T cells, memory CD4+ T cells do not appear to divide in irradiated mice that express only the selecting MHC molecule.
Transgenic DO T Cells Do Not Divide in Irradiated Hosts That Support Positive Selection.
To provide additional evidence that the peptide–MHC ligand responsible for thymic selection could not induce division of CD4+ T cells in irradiated mice, CFSE-labeled transgenic CD4+ DO T cells, which bear TCRs specific for OVA peptide in the context of Ad and are positively selected in BALB/c mice 33, were transferred to irradiated BALB/c mice. 7 d after transfer, analysis showed that few DO T cells divided in irradiated BALB/c hosts (Fig. 5, summarized in Table). Because the peptides found in the periphery of BALB/c mice are similar to the peptides that support selection of the DO TCR in the thymus 34, this experiment indicated that ligands involved in thymic selection could not induce division of DO T cells in an irradiated recipient.
To demonstrate that DO T cells were able to divide in response to antigen, irradiated BALB/c mice that had previously received CFSE-labeled DO T cells were injected with OVA peptide, for which DO T cells are specific. 7 d after the injection of 0.1, 1.0, or 10 μg of OVA peptide, the extent of division among DO T cells was determined (Fig. 5). Analysis showed that the amount of division of DO T cells reflected the amount of antigen administered to the irradiated BALB/c recipient. Although injection of 0.1 μg OVA peptide into an irradiated BALB/c host induced little division, injection of 1.0 or 10 μg OVA peptide caused a progressive increase in the extent of division of DO T cells. This result confirmed the previous finding that transgenic CD4+ T cells required antigen to divide when transferred to T cell–depleted mice 23 and indicated that the extent of division among DO T cells in irradiated BALB/c mice reflected the amount of peptide antigen administered. These findings suggest that the one or two rounds of division seen among normal B6.PL CD4+ T cells that were transferred to irradiated B6 hosts may have occurred in response to TCR stimulation by relatively low levels of peptides.
Previous reports have demonstrated that depletion of T cells in mice creates an environment in which T cells divide when transferred into these mice 19,20,21,22,23. Past experiments have also shown that the T cell division in such mice depends on TCR contact with MHC. For example, division of CD4+ T cells transferred to T cell–deficient mice depends on the expression of class II molecules 22,23. The findings presented here confirm that CD4+ T cell division in T cell–deficient mice requires class II expression and establish that class II–bound peptides control this division.
As expected, a smaller percentage of CD4+ T cells from normal mice divided when transferred to irradiated MHCIIKO mice than divided when transferred to irradiated B6 mice. To determine whether the peptides presented by class II were involved in this class II–dependent division, CD4+ T cells from mice expressing only one class II–peptide combination were transferred to irradiated hosts. AbEpTKO CD4+ T cells, which were selected in the thymus exclusively by Ep bound to Ab, did not divide when transferred to irradiated AbEpTKO hosts, in which all Ab molecules were occupied by the selecting Ep peptide. Because AbEpTKO hosts expressed class II molecules, the lack of division among transferred AbEpTKO T cells indicated that division of CD4+ T cells in irradiated mice depends on peptide presentation by class II molecules and not on other effects class II expression may have on irradiated hosts. Therefore, the class II–dependent division of naturally occurring B6.PL CD4+ T cells after transfer to irradiated B6 hosts was caused by peptides presented by class II molecules.
The lack of AbEpTKO T cell division in irradiated AbEpTKO hosts also established that the peptides that controlled division of CD4+ T cells in irradiated mice were distinct from those responsible for thymic selection. These peptides have not been specifically identified, but we refer to them as unfamiliar because they are distinct from those involved in thymic selection. These unfamiliar peptides were expressed extrathymically after irradiation and may be derived either from intestinal flora or from the mouse itself.
The requirement for unfamiliar peptides suggests that TCR specificity determined whether a particular CD4+ T cell divided when transferred to an irradiated host. This suggestion is supported by findings described here and in a previous report 23 that transgenic CD4+ T cells did not divide in T cell–deficient hosts in the absence of antigen. Thus, CD4+ T cells divided when transferred to irradiated mice in response to specific interactions between TCRs and unfamiliar peptides presented by class II molecules. The reason these specific interactions occurred in irradiated mice but did not occur in nonirradiated mice is unknown. There are at least two explanations for the observation that CD4+ T cells experienced greater exposure to unfamiliar peptides when transferred to irradiated mice than when transferred to normal mice. First, irradiation may have changed the distribution of peptides presented by class II molecules by altering the representation of specific self- or foreign peptides. This could have resulted from the phagocytosis of large numbers of apoptotic mouse cells by APCs or from transient infection of irradiated mice by normal flora due to the breakdown of mucosal barriers. Second, by causing apoptosis of lymphocytes, irradiation could have decreased the competition among T cells for interaction with peptides normally presented by class II molecules. The explanation that decreased competition led to increased exposure of CD4+ T cells to peptides is particularly appealing because it could also explain the rapid rate of division of T cells that were transferred to TCR-βKO and RAG2KO hosts, both of which lack endogenous T cells to compete with transferred CD4+ T cells for interactions with APCs. Further work will be required to determine whether increased contact between transferred CD4+ T cells and unfamiliar peptides in irradiated mice occurred because of decreased competition among T cells or due to a change in the distribution of peptides presented by class II molecules in the host.
The data presented here demonstrate that CD4+ T cells from normal mice divide in response to interactions with unfamiliar class II–bound peptides when transferred to irradiated hosts. Because the unfamiliar peptides that induce division in irradiated mice may also be presented, perhaps at lower levels, by class II molecules in normal mice, it is possible that the same peptides that cause division of CD4+ T cells in irradiated mice are also responsible for causing extrathymic expansion 13 of T cells in normal mice. Thus, the 20% of the CD4+ T cell repertoire that required class II expression to divide in irradiated hosts may have been created in donor mice by extrathymic expansion in response to the same peptides that induced division in irradiated mice. If so, then the peptides presented by class II molecules in a normal, healthy mouse may significantly influence the T cell repertoire, as has been suggested in an earlier study of peripheral T cell selection 24. Additional work will have to be done to determine whether unfamiliar peptides influence the T cell repertoire in normal mice by causing CD4+ T cell division and, if so, whether this division occurs only in the context of an active immune response.
The authors thank Ella Kushnir for breeding the AbEpTKO and TKO mice and Terry Potter for evaluating the manuscript.
This work was supported by United States Public Health Service grants AI17134, AI18785, and AI22295.
1used in this paper: CFSE, carboxyfluorescein diacetate succinimidyl ester; CyC, cychrome; Ii, invariant chain; KO, knockout; SEB, staphylococcal enterotoxin B