The Effect of Antigen Stimulation on the Migration of Mature T Cells from the Peripheral Lymphoid Tissues to the Thymus

Although the maturation and export of T cells from the thymus has been extensively studied, the movement of cells in the opposite direction has been less well documented. In particular, the question of whether T cells which have been activated by antigen in the periphery are more likely to return to the thymus had been raised but not clearly answered. We examined this issue by activating T cells present in the periphery with their cognate antigen, and assessing migration to the thymus. TCR-transgenic cells from OT-I mice (Thy1.2+), which recognise the ovalbumin peptide OVA257–264 in the context of H-2Kb, were transferred into otherwise unmanipulated Thy1.1+ C57BL/6 mice. Recipient mice were injected i.v. with 5 μg peptide (SIINFEKL) approximately 24 hours later. The numbers of donor-derived (Thy1.2+) cells in the thymus and peripheral lymphoid tissue were determined. The results clearly show increased numbers of transgenic cells in the thymus 3 days after antigenic stimulation. However, since numbers of transgenic cells increased in the spleen and LN in about the same proportion, the data do not support the notion that there is highly increased selective migration of activated T cells to the thymus. Rather, they suggest that a sample of peripheral cells enters the thymus each day, and that the mature immigrants detected in the thymus merely reflect the contents of the peripheral T cell pool.


INTRODUCTION
thymus migrate to the spleen and lymph nodes (LN), especially early in life, and in the absence of a thymus The thymus is the primary source of T cells for the these organs remain essentially devoid ofT cells. This peripheral lymphoid organs. T cells produced in the pathway has been extensively quantitated in a number * This work was supported by grants from the National Health and Medical Research Council of Australia.
of species, and in young adult animals represents an export rate of about 1% of thymocytes per day Scollay and Shortman, 1985). The reverse pathway, of mature T cells migrating from the periphery back into the thymus is less often considered, although several studies have shown that this is not a major pathway in normal animals (Dumont et al., 1984;Michie et al., 1988;Hirokawa et al., 1989;Agus et al., 1991). Our own unpublished data suggest that about 1% of the mature phenotype cells in the thymus may be mature T cell immigrants (Coward and Scollay, in preparation). Some studies have suggested that this pathway might preferentially be used by activated T cells. An early report indicated that activated antigen-specific T cells migrated to the thymus, and persisted there, whereas migration of non-activated cells to the thymus was not detected (Naparstek et al., 1982). Similarly, Fink et al. (1984) demonstrated that peripheral T cells activated in vivo provide a greater contribution to the thymic memory CTL response than non-activated cells. In that study, it was not possible to determine the proportion of thymocytes which were donor derived. More recently, it was shown that blasts (generated in parent F combinations) homed to the thymus, and represented approximately 0.4% of mature T thymocytes, although the rate of homing was greatly increased by irradiation of hosts (Agus et al., 1991). Others have shown that, compared to 'naive' CD4 + T cells, there is a preferential accumulation of antigen-experienced T cells in the rat thymus (Bell et al., 1995;Westerman et al., 1996). As is the case for 'normal' peripheral T cells (Hirokawa et al., 1989), accumulation of activated cells within the thymus is largely restricted to the medulla (Pabst and Binns, 1989;Agus et al., 1991;Westermann et al., 1996).
One shortcoming of the above studies is that thymic immigration of in vivo-activated clonal T cell populations has not been analysed. As part of a longer study on migration of mature cells into the thymus, we have addressed this question in a model where TCR transgenic T cells present only in the periphery could be stimulated with their cognate peptide, and their migration to the thymus assessed. OT-I mice (Hogquist et al., 1994) carry a transgene for a MHC class I-restricted TCR with Vc2 and V135 variable regions. The TCR recognises the ovalbumin peptide OVA257_264 presented by H-2Kb, the MHC haplotype of C57BL/6 mice. In these mice, the great majority of peripheral T cells are CD8/CD4-, and express the transgenic TCR (Koniaris et al., 1997). These cells are readily activated in vivo upon administration of the OVA peptide (Koniaris et al., 1997).
The data described in this paper clearly demonstrate increased numbers of transgenic cell in the thymus after peripheral antigenic stimulation, but since cell numbers are also increased in the periphery, the results do not support the concept that activated cells show selective migration to the thymus.

Phenotype of Donor Cell Inoculum
Phenotyping of pooled donor spleen and LN cells prior to transfer indicated that CD8 + cells were far more prevalent than CD4 / cells (CD4:CD8 ratio of approximately 1:9), and that the majority (97%) of CD8 / cells expressed Vt2 chains in their TCR, and were therefore peptide specific ( Figure 1A&B).

Quantitation of Donor-Derived Cells in Peripheral
Lymphoid Organs Recipient animals were anaesthefised, and perfused with PBS and FCS. Spleen and Mesenteric LN (MLN) cell suspensions were prepared, and stained to quantitate and phenotype donor-derived cells. In control mice which had been injected with donor cells followed by PBS instead of peptide three days previously, a population of donor-derived cells could be clearly seen, which comprised approximately 2% of total splenocytes, or 10 to 21% of CD8 + splenocytes ( Figure 1C; Table I). Similarly, in the LN, donor-derived cells represented 3 to 5% of total LN cells, or 9 to 18% of CD8 / cells ( Figure 1E; Table I). Three days post-peptide injection, donor-derived cells represented 12 to 17% of total splenocytes, and  Figure 1D&F; Table I). Thus, there was a large increase in donor-derived Ag-specific cells in peripheral lymphoid tissue three days after peptide injection.
By day six after peptide, the proportion of total lymphocytes and CD8 + lymphocytes which were donor-derived was similar to the control PBS-injected group (Table I).
When total cell numbers were considered, there was a 2 to 3-fold expansion in the number of splenocytes 3 days after peptide injection. This difference was not apparent 6 days after peptide, nor was it seen when LN cell numbers were compared. When absolute donor (Thyl.2/) cell numbers were compared, there was a 15 to 20-fold expansion in spleens, and a 6 to 8-fold expansion in MLN 3 days after peptide injection, leading to an approximate 15 to 20-fold increase in the number of donor-derived cells in the total peripheral pool (based on the assumption that the total cell pool in all LNs is small relative to cell numbers in the enlarged spleen). Six days post-peptide there were negligible increases in absolute Thyl.2 / cell numbers (Table I).  (Table II).  Table II). In terms of total cell numbers, the thymuses of peptide-injected mice had slightly fewer cells than those of PBS-injected mice (Table II). However, compared to PBS-injected mice, the absolute number of donor-derived (Thyl.2/) cells detected in the thymus was increased 16 to 20-fold 3 days after peptide administration (Table II). This difference was reduced to 3-fold at the latter time-point (Table II).
Proportions of Transgene-Expressing (Vt2+) Donor-Derived CD8 + Cells In order to ensure that donor-derived cells were indeed potentially antigen responsive, we assessed the proportion of donor CD8 / cells which were Vet2+.
Apart from slightly lower values in the thymus at day 6 after peptide administration, essentially all (approximately 95%) CD8 / cells expressed the transgenic Vct2 chain in their TCR (Table III), and were therefore antigen-specific.

DISCUSSION
Although thymic maturation and rates of thymic export of mature T cells have been extensively studied Scollay and Shortman, 1985), the reverse pathway, of immigration of mature T cells into the thymus, has been less well analyzed. The current study aimed to determine whether T cells activated in the periphery show increased selective migration to the thymus. By activating peripheral T cells with specific antigen in vivo we could avoid some of the pitfalls of previous studies, in which activated cells were adoptively transferred (Naparstek et al., 1982;Fink et al., 1984;Agus et al., 1991;Bell et al., 1995;Westermann et al., 1996). While the cells involved had been transferred from congenic donors, and had suffered the rigors of being prepared as cell suspensions, they were not manipulated in any other way. This model is therefore relatively physiological, as there was no organ transplantation or recipient manipulation (apart from the i.v. injection), and provides a situation in which there is a population of T cells present in the periphery but absent from the thymus unless there has been immigration. Our data confirm experiments in other systems (Dumont et al., 1984;Michie et al., 1988;Hirokawa et al., 1989; Agus et al., 1991) which show that migration back to the thymus in normal, unstimulated animals is very low, and also confirm our own experiments using carboxy fluorescein (diacetate) succinimidyl ester (CFSE)-labeled cells from normal, non-transgenic donors which show the same thing (Coward and Scollay, in preparation). Those CFSE experiments also showed long-term survival and apparently normal migration of transferred spleen and LN cells in this experimental set up. Other studies show similar long-term survival and/or tissue-specific migration of fluorescein isothiocyanate-labeled cells Chin and Cahill, 1984;Abernethy et al., 1990;Witherden et al., 1990). For the experiments in this paper, the transferred T cells were allowed approximately 24 hours for 'recovery' before antigen stimulation in order to minimize any effects of the earlier ex vivo manipulation.
It was clear that the transferred transgenic T cells were stimulated to proliferate in vivo with the specific peptide, as evidenced by the loss of CFSE staining intensity (not shown) and the large increase in cell numbers. This was true three days after stimulation, although by 6 days, most of the donor cells had apparently disappeared, presumably due to antigen induced apoptosis, as shown in similar systems by several investigators (Webb et al., 1990;Kyburz et al., 1993;Moskophidis et al., 1993;Koniaris et al., 1997). Three days after antigen administration, there were significant changes in the sizes of the lymphoid organs. The cellularity of the thymus was reduced, presumably due to the well characterised stress effects often associated with experimental manipulation, including delivery of antigen. The reduced thymic cellularity may have been due to deletion of cortical thymocytes, possibly mediated by cortico-steroids or TNF produced by peripherally activated T cells (Martin and Bevan, 1997). In contrast, the cellularity of the peripheral organs was increased, due to expansion of the transgenic T cells and cytokine-induced "bystander" proliferation (Tough et al., 1996;Ehl et al., 1997). For this reason we have focussed this discussion on absolute T cell numbers, as it is difficult to interpret the relevance of changes in proportion in the face of varying organ sizes.
There Was clearly an increase in the number of cells which migrated into the thymus following antigen-specific T cell activation in the periphery, confirming the findings of earlier studies (Naparstek et al., 1982;Fink et al., 1984;Agus et al., 1991;Bell et al., 1995;Westermann et al., 1996). This increase amounted to about 16 to 20-fold 3 days after injection with peptide. However, at this same time point there had also been an increase in the number of cells in the periphery, by about 15 to 20-fold in the spleen and 6 to 8-fold in the MLN. In these animals with enlarged spleens, this must represent an increase in the total peripheral pool only slightly less than the 15 to 20-fold value seen in the spleen. Thus, the increase in the thymus (16 to 20-fold) was very close to that seen in the periphery (approximately 15 to 20-fold). Therefore, the results do not support the notion that there is preferential migration of activated peripheral T cells back to the thymus. Rather they support the idea that there may be a steady level peripheral cells to the thymus which is proportional to their numbers in the periphery. While it is clear that our data do not prove such a model, they are consistent with it, but not with a preferential migration model. There is one additional point which should be made. We cannot exclude the possibility that the increased numbers of transgene-positive donor cells in the thymus result from proliferation inside this organ, rather than increased migration. Our data do not allow us to distinguish between these alternatives. Previous investigations of immigration of mature T cells to the thymus were limited by the fact that activated cells were defined and purified on the basis of cell-surface marker expression (Bell et al., 1995;Westermann et al., 1996), or involved adoptive transfer of previously activated cells (Fink et al., 1984;Agus et al., 1991). Additionally, measurement of immigrating cells has frequently relied upon relatively non-quantitative radioactive, immunohistochemical or in vitro cell culture techniques (Naparstek et al., 1982;Fink et al., 1984;Westermann et al., 1996). By transferring large numbers of clonal T cell populations, administering specific peptide several days later, and quantitating thymic immigrants by flow cytometry we have been able to circumvent some of these shortcomings. Our results are in broad agreement with the previous studies, showing an enhanced accumulation of activated cells within the thymus, but for the first time allowing a comparison 130 of changes in the pool from which the activated cells have come.
The functional significance of peripheral T cells in the thymus, especially previously activated T cells, is unknown. It has been suggested that, since the thymic medulla contains both epithelial and bone marrow-derived cells which express both MHC and co-stimulatory molecules (Picker and Siegelman, 1999), it is a potential site for immune reactions (Michie et al., 1988), although there is no clear evidence to support such a case. However, it is possible that the presence of activated cells in the thymus might play a role in intrathymic selection and maturation events, as suggested by several investigators (Naparstek et al., 1982;Michie et al., 1988;Agus et al., 1991). Such a mechanism would presumably require delivery to, and recognition of, peripherally-derived peptides within the medulla, which may selectively alter production of thymic emmigrants. The possibility that such peptide/self-antigen plays an important role in medullary proliferation or emigration remains to be demonstrated (see Scollay and Godfrey, 1995 for discussion).