Homeostatic Expansion Occurs Independently of Costimulatory Signals

Mice

C57BL/6 (B6) background CD40^−/− mice were initially obtained from D. Parker (Portland, OR), and additional CD40^−/− and B6 background CD28^−/− mice were obtained from The Jackson Laboratory (Bar Harbor, ME). The 4-1BBL^−/− mice (19) were obtained in collaboration with J. Peschon (Immunex, Seattle, WA). BALB/c, BALB/c Rag^−/−, B6, and CD45.1 congenic B6 mice (carried under the strain name B6.Ly-5.2) were obtained from the National Cancer Institute, The Jackson Laboratory, or Taconic Farms. Normal and CD28^−/− DO11.10 TCR transgenic mice have been described previously (20, 21) and were maintained at the University of Minnesota. All mice were maintained under specific pathogen-free conditions and used at 6–12 wk of age. Recipient mice were maintained on antibiotic water (polymixin B sulfate and neomycin sulfate) throughout the course of the experiment.

Adoptive transfer

Cells were purified by negative selection utilizing magnetic cell sorting using MACS microbeads (Miltenyi Biotec, Auburn, CA). To purify CD44^low T cell, lymph node cells were depleted from adherent cells (90-min incubation on tissue culture-treated flasks at 37°C) and labeled with FITC-coupled Abs to B220 (clone RA3-6B2), I-A^b (AF6-120.1), and CD44 (IM-7) (0.0125 μg anti-B220 and anti-I-A^b per 1 × 10⁶ cells; 0.004 μg anti-CD44 per 1 × 10⁶ cells) (all from BD PharMingen, San Diego, CA). In some experiments, CD4 cells were also labeled with FITC anti-CD4 (GK1.5; BD PharMingen) (0.0125 μg per 1 × 10⁶ cells). Following staining, cells were subject to depletion using anti-FITC MACS microbeads. Flow through cells were >90% pure (established by flow cytometric analysis).

CD4⁺ cells from DO11.10 animals were purified utilizing a similar protocol using biotinylated (bio)³ anti-CD8 (53.6-7) plus bio-anti-B220, followed by streptavidin (SA) microbeads. In some cases, CD25⁺ cells were depleted at the same time using bio-anti-CD25 (7D4) (if applicable) (PharMingen). The percentage of CD4⁺CD25⁺ T cells before and after depletion was assessed by staining an aliquot of the cells with both CD25-PE Ab and SA-PE (to reveal cells bound by the depleting bio-CD25 Ab) and counterstaining for CD4 with CD4-APC or CD4-PerCP.

After purification, cells were stained with CFSE (Molecular Probes, Eugene, OR), as described previously (7, 22). Labeled donor cells were suspended in PBS, and 1–3 million cells were injected via the tail vein of recipient mice. In some cases, host mice were sublethally irradiated (700 cGy) 1 day before cell transfer. In some experiments, recipient mice were primed i.v. with 2 mg OVA (whole protein) plus 25 μg LPS (Sigma, St. Louis, MO) 1 day cell posttransfer. Cells were analyzed on day 3 postpriming.

Flow cytometry

Recipient mice were sacrificed at the time points indicated, and single cell suspensions were prepared separately from spleen and a pool of major lymph nodes. Lymph node and spleen cells were then stained with the fluorescently conjugated Abs to CD4, CD8, CD44, and CD45RB. Biotinylated Abs to CD25, CD45.1, CD45.2, CD28, and KJ1-26 (specific for the DO11.10 TCR) were used, staining being revealed using SA-TriColor, SA-PerCP, or SA-PE (all reagents from BD PharMingen). Cells were analyzed by using a FACSCalibur (BD Biosciences, San Jose, CA) and analyzed using FLOWJO (TreeStar, San Carlos, CA) software.

Experimental system

To test the role of CD28/B7 interactions in regulating homeostatic expansion, we studied proliferation of CFSE-labeled normal and CD28^−/− donor T cells after transfer into normal or lymphopenic B6 hosts. To minimize mouse to mouse variation, we employed a cotransfer system in which we premixed CD45.1⁺ wild-type and CD45.2⁺ CD28^−/− T cells before transfer and distinguished the two donor populations during analysis with CD45 allele-specific Abs. Similar transfers were performed using CD40^−/− or 4-1BBL^−/− hosts to determine the effect of disrupting CD40L/CD40 and 4-1BB/4-1BBL interactions (alone or in combination with CD28 deficiency) on homeostatic expansion.

No requirement for CD28, CD40, or 4-1BBL in T cell homeostatic expansion

Using the transfer scheme outlined above, we tested whether CD28 expression on T cells and CD40 or 4-1BBL expression on host APCs was required for T cell homeostatic expansion. As expected, neither wild-type nor CD28^−/− T cells proliferate after transfer into unmanipulated syngeneic B6 hosts (Fig. 1⇓, a–d). Also as expected, normal CD8 and CD4 T cells undergo one to four rounds of proliferation 14 days after transfer into irradiated syngeneic hosts (Fig. 1⇓, e and g, respectively). Interestingly, CD28^−/− T cells proliferated to the same extent as wild-type cells when transferred into irradiated wild-type hosts (Fig. 1⇓, f and h). This suggests that CD28 is not required for homeostatic expansion of either CD8 or CD4 subsets.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Role of CD28 and CD40 interactions in homeostatic expansion of cotransferred CD4 and CD8 T cells. One million CD44^low T cells from CD28^−/− and from CD45.1 congenic B6 animals were mixed at a 1:1 ratio, CFSE labeled, and injected i.v. into the indicated host mice. In some cases (irrad.), recipient mice were sublethally irradiated (700 cGy) 1 day before transfer. Recipient mice were sacrificed 14 days posttransfer, and lymph node cells were analyzed by flow cytometry using Abs to CD4, CD8, and CD45.1 to distinguish the different subsets of cells. The CFSE profile of CD4 and CD8 subsets and of CD28^−/− and wild-type populations is shown separately, but comes from the same mice in each row. Because the recipient mice are all CD45.1⁻, host cells appear in the analysis of the CD28^−/− subsets as a CFSE-negative population. Data are representative of two to three mice per group, and comparable results were obtained in at least three experiments.

It was possible that CD40L/CD40 or 4-1BB/4-1BBL costimulation could partially or completely compensate for CD28 deficiency. Indeed, 4-1BB and CD40L interactions can synergize with CD28 or even replace it in some T cell responses (17, 19, 23, 24). To test this, we transferred both wild-type and CD28^−/− T cells into CD40-deficient hosts (Fig. 1⇑, i–l) or into 4-1BBL-deficient hosts (Fig. 1⇑, m–p). Once again, homeostatic proliferation of both CD4 and CD8 T cell subsets was similar regardless of CD28 expression on the donor cells and CD40 or 4-1BBL expression in the host.

In all of these experiments, we observed similar donor T cell recoveries for both the normal and CD28^−/− donors after expansion in the various irradiated host strains (data not shown), which argues against a survival role for these costimulator molecules in homeostatic expansion, at least in the short term.

CD28^−/− DO11.10 TCR transgenic T cells undergo greater homeostatic expansion, but less Ag-driven proliferation than their CD28⁺ counterparts

The experiments above utilized polyclonal T cells and, although the overall degree of homeostatic expansion was similar with or without CD28, it was possible that different populations of T cells were responding, perhaps based on TCR specificity dictating CD28 dependence. Furthermore, we had to date looked exclusively in the C57BL/6 system, whereas there is some evidence that the impact of CD28/B7 deficiency is more extreme in certain BALB/c responses (25). Accordingly, we also studied homeostatic expansion of T cells from DO11.10 TCR transgenic mice, which are maintained on the BALB/c background. Data from several groups have shown that CD28 deficiency impairs the peptide/MHC (OVA/I-A^d) response of these T cells, both in vitro and in vivo (21, 26).

Thus, we transferred DO11.10 and DO11.10/CD28^−/− cells into irradiated BALB/c hosts and analyzed their homeostatic expansion. As shown in Fig. 2⇓, not only did we fail to see a requirement for CD28 in DO11.10 homeostatic expansion, there was evidence that CD28^−/− DO11.10 cells proliferate moderately better than their wild-type counterparts. The degree of this improved response was somewhat variable between individual animals, as exemplified in Fig. 2⇓. Thus, at a minimum we find that more CD28-deficient DO11 cells commit to homeostatic expansion than normal DO11 cells (see proportions of cells undergoing zero vs three divisions, Fig. 2⇓a), and that in some cases this was seen as an extra round of proliferation by the CD28^−/− DO11.10 cells (Fig. 2⇓b).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Homeostatic expansion of normal vs CD28^−/− DO11.10 cells. One million CFSE-labeled DO11.10 or CD28^−/− DO11.10 T cells were transferred into sublethally irradiated BALB/c hosts and analyzed 21 days after transfer. The CFSE profiles shown were gated on CD4⁺ KJ1-26⁺ cells. The number of rounds of proliferation (derived from comparison with CFSE levels in cells transferred into unirradiated BALB/c hosts) are indicated over the traces. Normal DO11.10 cells are represented by the shaded histogram; CD28^−/− DO11.10 cells are shown as the unfilled solid line. The data for a and b come from separate host mice in the same experiment and represent the range seen among three separate experiments involving three mice per group.

Given this surprising result, it was important to test that, in our hands, CD28 deficiency had the predicted effect of blunting the DO11.10 Ag-specific response, as reported by others (21, 26). Thus, we transferred CFSE-labeled wild-type or CD28^−/− DO11.10 T cells into irradiated or unirradiated BALB/c hosts. The unirradiated hosts were then immunized with OVA/LPS, and the proliferative response of the DO11.10 population was monitored. As expected, CD28 deficiency severely impaired the Ag-specific DO11.10 response (Fig. 3⇓, c vs b). However, as in Fig. 2⇑, homeostatic proliferation of the same CD28^−/− DO11.10 donor population was slightly greater than that of wild-type transgenic cells (Fig. 3⇓, g vs f).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Impact of CD28 expression and CD4⁺CD25⁺ T cell depletion (Depl.) on foreign Ag reactivity and homeostatic expansion of DO11.10 cells in irradiated hosts. Normal, CD28^−/−, and CD25 T cell-depleted DO11.10 T cells were transferred into normal (a–e) or irradiated (f–h) BALB/c hosts. The percentages of CD4⁺CD25⁺ cells in the normal, CD28^−/−, and CD25-depleted populations were 4.2%, 1.2%, and 0.5%, respectively, in the experiment shown. The unirradiated hosts were challenged with OVA/LPS 1 day after transfer and sacrificed 3 days later. Irradiated (f–h) and control unirradiated hosts (e) were not immunized and were analyzed 21 days after transfer. CFSE profiles shown are gated on CD4⁺ KJ1-26⁺ cells, except in a, which was gated on CD4⁺ KJ1-26⁻ cells as a control for Ag nonspecific proliferation. Data are representative of three mice in each group.

CD4⁺CD25⁺ cells do not contribute to the difference seen between normal and CD28^−/− DO11.10 T cells

Looking for an explanation for the slightly enhanced proliferation of DO11.10 CD28^−/− cells, we considered the contribution of CD4⁺CD25⁺ regulatory cells (27, 28). This subset, which can suppress autoimmune T cell responses, is underrepresented in CD28^−/− animals (29). This is also true in mice expressing the DO11.10 TCR transgene, in which CD28^−/− DO11.10 mice have fewer CD4⁺CD25⁺ cells compared with wild-type DO11.10 animals (3–5% vs 5–10%, respectively; data not shown). Thus, we tested the impact of depleting the donor DO11.10 cells of CD25⁺ cells. If the CD25⁺ subset were restraining normal DO11.10 homeostatic expansion, then removal of these cells should lead to even greater homeostatic expansion than seen for the CD28^−/− DO11.10 group. CD25⁺ regulatory cell depletion did not improve the robust Ag-specific T cell response in vivo, perhaps reflecting the fact that this response was already optimal (Fig. 3⇑d). Interestingly, however, CD25⁺ subset depletion failed to impact DO11.10 homeostatic expansion (Fig. 3⇑, compare h with f).

Since sublethal irradiation does not deplete all T cells in the host, it was possible that endogenous CD4⁺CD25⁺ cells could influence homeostatic expansion of the transferred cells. Hence, we also transferred these same DO11.10 populations into BALB/c Rag^−/− hosts. Similar to the results in the irradiated hosts, proliferation of CD25-depleted DO11.10 cells did not differ significantly from that of normal DO11.10 cells, whereas there was once again a subtle, but consistent increase in proliferation by the CD28^−/− DO11.10 population (Fig. 4⇓). Intriguingly, after expansion of DO11.10 cells transferred into Rag-deficient hosts, a sizeable proportion of the cells had proliferated sufficiently to dilute out the CFSE dye, which was not observed in irradiated hosts (compare Fig. 4⇓ with Figs. 2⇑ and 3⇑). This may relate to the fact that sublethally irradiated hosts recover endogenous T cell numbers during the course of the experiment (therefore limiting T cell space), whereas this will not occur in Rag^−/− hosts. We also noticed a slight increase in the total number of donor T cells recovered after transfer of CD25-depleted DO11.10 cells compared with wild-type or CD28^−/− cells, which was also not observed in irradiated hosts (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Impact of CD28 expression and CD4⁺CD25⁺ T cell depletion (Depl.) on homeostatic expansion in Rag^−/− hosts. Three million CFSE-labeled normal, CD28^−/−, or CD25 T cell-depleted DO11.10 cells were transferred into BALB/c Rag-2^−/− hosts. The percentages of CD4⁺CD25⁺ cells in the normal, CD28^−/−, and CD25-depleted populations were 4.1, 2.8, and 0.6%, respectively. Recipient mice were harvested 22 days after transfer, and the data shown are representative of three mice in each group.

Thus, these results do not support a role for the DO11.10 CD25⁺ subset in restraining homeostatic expansion of DO11.10 T cells.

Note added in proof.

While this manuscript was in press, Gudmundsdottir et al. (42) reported that CTLA4Ig blocked homeostatic expansion of a subset of DO11.10 Rag^−/− T cells in Rag^−/− recipients. Specifically, CTLA4Ig administration caused a selective loss of the rapidly proliferating subset and this inhibition was influenced by the number of donor T cells transferred. These data suggest B7 interactions might influence homeostatic proliferation of certain subsets under specific conditions, while our data argue that CD28 interactions are not obligatory for this response. The basis for these discordant results presumably lies in the different experimental approaches used.