DA (RT1^a), Piebald Virol Glaxo rat strain (PVG) (RT1^c), and Lewis (RT-1^l) rats were bred and maintained in the animal house, Liverpool Hospital. All animals were fed standard chow and given water ad libitum. The study was carried out in accordance of the “Australian Code for the Care and Use of Animals for Scientific Purposes (NHMRC)” and Animal Ethics Committee of the University of New South Wales (UNSW), Australia. Animal experimental protocols were approved by the Animal Ethics Committee of the UNSW Australia.

DA rats weighing 180–230 g were anesthetized with either ether or isoflurane and heterotopically grafted with adult PVG heart (14). Graft rejection was assessed as cessation of palpable beat (21). Tolerance was induced by intraperitoneal injection of 7 mgm/kg of an anti-CD3 mAb (G4.18), as described (13, 14, 19). Hosts with good functioning grafts for >150 days were considered tolerant. The cells studied were from spleen and lymph nodes of tolerant animals, >150 days after transplantation (14, 19).

Anti-rat mAb used were G4.18 (CD3), Ox35 (CD4), MRCOx8 (CD8), MRCOx39 (CD25, IL-2R alpha chain), LECAM-1 (CD62L, l-selectin), and MRCOx33 (CD45RA, B cells) (BD-Pharmingen, San Diego, CA, USA). Anti-mouse/rat FOXP3 (FJK-16s) (eBioscience, San Diego, CA, USA) was used as per the manufacturer’s instructions. Immunostained lymphocytes were analyzed on a FACScan (Becton Dickenson, San Jose, CA, USA) using CellQuest software (Becton Dickenson).

Recombinant (r) IL-2, rIL-4, rIL-5, rIL-10, rIL-12p70, rIL-13, rIFN-γ, and rTGF-β cytokines were produced and quantified, as described (22). Each cytokine was added to cultures at ≥200 U/ml. The IL-4-transfected cell line (28) was a gift of Dr. Barclay (Pathology, Oxford, UK).

Single cell suspensions from spleen and lymph node were prepared and RBC lysed as described (29). An indirect panning technique to deplete CD8⁺ T and B cells, followed by CD25 enrichment using PE conjugated MRCOx39 mAb and anti-PE microbeads (Miltenyi Biotech Australia, Macquarie Park, NSW, Australia) as described (21, 25, 29). Enriched CD4⁺CD25⁻ T cells were >96% CD4⁺ with <3% CD25^hi cells. Enriched CD4⁺CD25⁺ T cells were 85–95% CD25⁺ with greatest enrichment for CD25^hi cells and had 70–80% FOXP3⁺cells.

The methods were as previously described (22, 25, 29). Briefly, stimulator cells were irradiated cells from thymus of naïve rats, which do not produce T cell cytokines (13). Stimulator cells from thymus are as effective or more effective than irradiated spleen cells as stimulator cells (29). Stimulation in MLC with specific-donor PVG cells was compared to stimulation to self-DA and third party Lewis.

Micro-cultures in U-bottom microtiter plates (Greiner) contained 2 × 10⁴ stimulator cells and 2 × 10⁵ responder cells in a total volume of 200 µl. Proliferation was assessed by ³H thymidine incorporation, as described (22, 29). Each experimental sample had 3–6 replicates and results were expressed as a mean ± SD.

As naïve CD4⁺CD25⁺ T cells are poor at proliferation in MLC without IL-2, the methods were refined to eliminate non-specific background proliferation (22, 29). These modifications included use of 20% Lewis rat serum rather than fetal calf serum and irradiated thymic stimulator cells.

In time course assays, proliferation was assessed at days 2, 3, 4, 5, and 6. For limiting dilution assays, serial twofold dilutions of T cells subsets starting at 1 or 2 × 10⁵ cells/well were cultured with 2 × 10⁴ stimulators cells/well and were assayed for proliferation at day 4. Stimulation index was calculated as proliferation of test MLC with cytokine added divided by proliferation of the control MLC with no cytokine.

mRNA from cells was extracted and reverse transcribed as described (22, 25). Primers for il-5ra, ifngr, il5, ifng, foxp3, and methods of RT-PCR were as described (22, 25).

Parametric data were expressed as mean ± SD and were compared with Student’s t-test using Statview for Mac (Abacus Concepts, Berkley, CA, USA). Wilcoxon Signed Rank test was used for non-parametric results. Statistical significance was set at p < 0.05.

Tolerance was induced by treatment of DA rats with anti-CD3 mAb (G4.18) at the time of transplantation with a heterotopic PVG heart graft, as described (13, 14, 19). Spleen and lymph node cells from tolerant hosts were from rats with good functioning heart grafts for over 150 days that were receiving no ongoing immunosuppressive therapy.

Figure 1A shows representative immunostaining and FACS analysis for CD4, CD8, CD25, and FOXP3. The proportion of lymphocyte subsets in lymph node and spleen of tolerant and naïve DA rats was similar for CD4⁺ T cells, CD8⁺ T cells, or B cells (Figure 1B). There was no increase in the proportion of CD4⁺CD25⁺ T cells in tolerant hosts, which remained at <20% of peripheral lymphocytes (Figure 1B, n = 16). 70–80% of CD4⁺CD25⁺ T cells expressed FOXP3 in both naïve and tolerant hosts.

The unfractionated, CD4⁺, CD4⁺CD25⁺, and CD4⁺CD25⁻ populations were prepared as described in the Section “Materials and Methods.” Figure 2 shows a representative FACS analysis of enriched T cell subsets from tolerant hosts. CD4⁺CD25⁻ T cells had <1% CD8⁺, <2% CD25^hi, and <2% FOXP3^hi. CD4⁺CD25⁺ T cells were >85% CD25⁺ and 70–80% FOXP3⁺ (Figure 2). This subpopulation had 5–10% of CD8^loT cells that was also CD4⁺. Preparations had 1–5% B cells, but these cells do not proliferate in rat MLC (30).

The time course of proliferation of naïve DA cells to self-DA and fully allogeneic PVG is illustrated in Figure 3A. Naïve unfractionated and enriched CD4⁺ T cells had a similar time course with proliferation to PVG peaking at day 4 and 5 and waning by day 6. The response to self-DA was delayed and peaked at day 5.

With naïve CD4⁺CD25⁻ T cells, the proliferation to PVG and to DA was similar at the peak on day 5. Prior to that, there was a slightly greater proliferation to PVG compared to self-DA. In subsequent experiments with CD4⁺ and CD4⁺CD25⁻ T cells, proliferation was assayed at day 4, when the differences between response to PVG and DA were greatest.

Naïve CD4⁺CD25⁺ T cell proliferation was much less than unfractionated lymphoid cells, CD4⁺ T cells, or CD4⁺CD25⁻ T cells (Figures 3A,B). Proliferation to PVG was significantly greater than to self-DA at day 3 and 4, after which it waned. Counts were small and never more than two thousand, often in the hundreds. In subsequent experiments, CD4⁺CD25⁺ T cell proliferation was assayed at day 3 or 4.

CD4⁺ T cells from tolerant (Figure 4A) or naïve rats (Figure 3B) had similar proliferation to specific-donor PVG and third-party Lewis, and the response to self-DA was less than to either alloantigen. The similarity of the response of unfractionated and CD4⁺ T cells from tolerant hosts to PVG and third party was confirmed in limiting dilution assays (Figure 4B). This showed there is not full clonal deletion of CD4⁺ T cells reactive to PVG, as previously described in this model of tolerance (14).

Removal of CD4⁺CD25⁺ T cells from naïve CD4⁺ T cells resulted in enhanced proliferation of the remaining CD4⁺CD25⁻ T cell population in MLC to PVG, Lewis, and self-DA, compared to that of unfractionated lymphoid cells (Figure 3A) and CD4⁺ T cells (Figures 3A,B). This suppression is not antigen-specific as the response to PVG, Lewis, as well as to self-DA was enhanced by removal of CD4⁺CD25⁺ T cells (Figures 3A,B) (29). The removal of naïve CD4⁺CD25⁺ T cells unmasked a significant autologous response in MLC, which was delayed compared to that to alloantigen, but at its peak was nearly as great as to alloantigen. This autologous response of naïve CD4⁺CD25⁻ T cells has been described previously, for example (29). The enhanced response by removal of naïve CD4⁺CD25⁺ T cells was confirmed in a serial dilution MLC, where the stimulator cells numbers were constant, and the responder cells were serially diluted twofold starting at 2 × 10⁵ cells per well, and out to 6.25 × 10⁴ per well (Figure 3C). There was significantly enhanced proliferation of naïve CD4⁺CD25⁻ T cells compared to unfractionated CD4⁺ T cells at several dilutions. Thus, with naïve hosts, the minority CD4⁺CD25⁺ T cells (10%) suppressed proliferation of the majority CD4⁺CD25⁻ T cells (90%) population within CD4⁺ T cells.

In contrast, CD4⁺CD25⁻ T cells from tolerant hosts had no increase in proliferation to specific-donor PVG compared to CD4⁺ T cells from tolerant hosts but had significantly greater proliferation to the third-party Lewis (Figure 4B). The reduced response of CD4⁺CD25⁻ T cells from tolerant hosts to PVG may be due to clonal pruning. This finding also suggested that the CD4⁺CD25⁺ T cells in unfractionated CD4⁺ T cells from tolerant hosts did not suppress proliferation of CD4⁺CD25⁻ T cells from tolerant hosts to specific-donor but inhibited these cells’ proliferation to third-party or self.

There was a small and similar response of naïve CD4⁺CD25⁺ T cells to PVG or Lewis alloantigen that was greater than to self-DA (Figure 3B). This response to alloantigen peaked at day 3 and waned after day 4 (Figure 3A). Thus, we report day 3 proliferation.

The response of CD4⁺CD25⁺ T cells from tolerant animals was different. CD4⁺CD25⁺ T cells from tolerant hosts had significantly lower proliferation to PVG than to third party Lewis (Figure 4A). No increase in proliferation of CD4⁺CD25⁺ T cells from tolerant hosts to PVG alloantigen was observed on days 2 through to 5 (data not shown). Figure 4A, in the bottom row, shows results from four other separate experiments demonstrating low proliferation of CD4⁺CD25⁺ T cells from tolerant hosts to PVG that was not significantly different to the response to self-DA in most experiments. CD4⁺CD25⁺ T cells from tolerant hosts retained MLC responsiveness to third-party Lewis stimulator cells, similar to that of naïve CD4⁺CD25⁺ T cells.

This was a paradoxical finding, as it was assumed that donor alloantigen-specific CD4⁺CD25⁺ T cells would be increased in tolerant hosts.

We have shown that the survival in vitro of tolerance transferring CD4⁺ T cells requires both stimulation with specific-donor alloantigen and cytokines from activated lymphocytes (16, 18, 23, 24). Thus, we examined which T cell cytokines supported proliferation of CD4⁺CD25⁺ T cells from tolerant hosts to specific-donor antigen but not to third-party antigen or self-DA.

Proliferation of naïve CD4⁺CD25⁺ T cells to all stimulator cells is enhanced by addition of rIL-2 or rIL-4 as previously described (22, 25, 26) and replicated in Figure 5A. rIL-2 and rIL-4 also induced proliferation of CD4⁺CD25⁺ T cells from tolerant hosts to self- or PVG and Lewis stimulator cells (Figure 5B). This polyclonal expansion by rIL-2 or rIL-4 was observed in four separate experiments. Neither rIL-2 nor rIL-4 selectively expanded CD4⁺CD25⁺ T cells from tolerant hosts to specific-donor PVG. The increased proliferation induced by rIL-2 or IL-4 to PVG and to third-party Lewis varied. Although there is a difference in the experiment in Figure 5B, this was not consistent, as stimulation index showed no difference in response to specific donor PVG and third party Lewis (Figure 5C).

Other Th1 cytokines, rIFN-γ, rIL-12p70, and rIL-15 did not enhance proliferation of naïve CD4⁺CD25⁺ T cells PVG, Lewis, or self-DA. In some cultures, rIFN-γ significantly suppressed proliferation (Figure 5A). The Th2 cytokines, rIL-5, rIL-10, and rIL-13 also did not enhance proliferation of naïve CD4⁺CD25⁺ Treg (Figure 5A). Thus, naïve CD4⁺CD25⁺ T cells were not activated by the other Th1 and Th2 cytokines we tested.

With CD4⁺CD25⁺ T cells from tolerant hosts, rIL-5 or rIFN-γ enhanced the response to specific-donor PVG but not to third-party Lewis or self-DA stimulator cells (Figure 5B).

With rIFN-γ, proliferation of CD4⁺CD25⁺ T cells from tolerant hosts was significantly enhanced to donor PVG compared to the response with rIFN-γ and DA stimulator cells, in 6 of 7 separate experiments (Figure 5C). The response to PVG with rIFN-γ was also greater than the response of CD4⁺CD25⁺ T cells from tolerant hosts to PVG where there was no cytokine or supernatant from non-transfected CHO-s cells (p < 0.01). rIFN-γ did not enhance proliferation of CD4⁺CD25⁺ T cells from tolerant hosts to third-party Lewis (p < 0.01) (Figure 5C) in any of five separate experiments. With rIFN-γ, the response to PVG was greater than the response to Lewis in all five separate experiments where there was direct comparison (p < 0.05–0.001). The reproducibility of this selective enhancement of proliferation by rIFN-γ to specific donor alloantigen is summarized as stimulation index in Figure 5C. With rIFN-γ, this response to specific donor is significantly different to proliferation to third-party, which was not enhanced by rIFN-γ (p < 0.01) (Figure 5C).

rIL-5 was the only Th2 cytokine that enhanced proliferation of CD4⁺CD25⁺ T cells from tolerant hosts to specific-donor PVG but not to self-DA or third party Lewis. This increased stimulation induced to PVG by rIL-5 was replicated in all six experiments (p < 0.05–0.01). No increased response to third party Lewis was observed in four experiments (p < 0.01). Comparing the effects of rIL-5 on proliferation expressed as stimulation indexes, the proliferation to specific-donor PVG was significantly greater than its effect on proliferation to third-party Lewis (Figure 5C).

rIL-12p70 enhanced proliferation of CD4⁺CD25⁺ T cells from tolerant hosts (Figure 4B), but this was not consistently observed (Figure 5C). 4 of 6 replicate experiments showed rIL-12 enhanced proliferation to PVG, and 1 of 4 showed enhanced proliferation to third party Lewis (Figure 5C). Comparison of stimulation indexes with rIL-12 showed no significant differences in responses to PVG and Lewis.

No other cytokine including rIL-15, rIL-12p40 (data not shown), rIL-10, rIL-13, nor rTGF-β (data not shown) enhanced proliferation of CD4⁺CD25⁺ T cells from tolerant hosts to either PVG or third-party Lewis alloantigen (Figure 5B).

CD4⁺CD25⁺ T cells from tolerant hosts expressed ifngr and il5ra whereas naïve CD4⁺CD25⁺ T cells did not express il5ra and had low expression of ifngr (Figure 6A). Further, CD4⁺CD25⁺ T cells from tolerant hosts also expressed il5 and ifng. These findings were consistent with our previous report of two pathways of alloactivation of naïve tTreg (22, 26), summarized in Figure 7.

CD4⁺CD25⁺ T cells from tolerant hosts had greater foxp3 expression compared to those from naïve animals, consistent with activated Treg.

To confirm the functional requirement of receptors for IL-5 and IFN-γ, we tested their effect on MLC of CD4⁺ T cells from tolerant hosts to PVG. As described above, CD4⁺ T cells from tolerant hosts had no enhancement in proliferation when the CD4⁺CD25⁺ T cells were removed, suggesting the CD4⁺CD25⁺ T cells from hosts did not suppress CD4⁺CD25⁻ T cells from tolerant hosts. We examined if this failure to suppress was due to insufficient IFN-γ or IL-5 in MLC. Both cytokines reduced proliferation of CD4⁺ T cells from tolerant hosts to specific-donor alloantigen in MLC compared to MLC without cytokines (Figure 6B).

However, rIFN-γ, but not rIL-5, also suppressed MLC of naïve CD4⁺ T cells (Figure 6). This effect of rIFN-γ is consistent with the known effect of IFN-γ on T cell proliferation (29). These results suggested IL-5 and possibly IFN-γ preserved the suppressor function of the antigen-specific CD4⁺ Treg in CD4⁺ T cells from tolerant hosts during stimulation in MLC.