Homeostatic proliferation is a barrier to transplantation tolerance

Abstract

Despite the ease of inhibiting immune responses by blockade of T-cell costimulation in naive rodent models, it is difficult to suppress those responses in animals with memory cells. Studies demonstrating the importance of alloreactive T-cell deletion during tolerance induction have promoted use of peritransplant T-cell-depleting therapies in clinical trials. But potentially complicating wide-scale, nonspecific T-cell depletion is the finding that extensive T-cell proliferation can occur under conditions of lymphopenia. This process, termed homeostatic proliferation, may induce acquisition of functional memory T cells. Here, using clinically relevant mouse models of peripheral T-cell depletion, we show that residual nondepleted T cells undergo substantial homeostatic expansion. In this setting, costimulatory blockade neither significantly suppresses homeostatic proliferation nor prevents allograft rejection. In addition, T cells that have completed homeostatic proliferation show dominant resistance to tolerance when adoptively transferred into wild-type recipients, consistent with known properties of memory cells in vivo. These findings establish the importance of homeostatic proliferation in clinically relevant settings, demonstrate the barrier that homeostatic proliferation can present to the induction of transplantation tolerance, and have important implications for transplantation protocols that use partial or complete peripheral T-cell depletion.

Figure 1

Homeostatic proliferation after partial T-cell depletion. (a) BrdU labeling of live CD3⁺ cells from B6 mice treated with depleting antibodies to CD4 and CD8 (mAb) and/or and thymic irradiation. Six days after the last antibody dose, mice were fed BrdU for six days until sacrifice. (b) B6 mice underwent thymectomy and were treated with depleting antibodies to CD4 and CD8. Five days after the last antibody dose, 20 million CFSE-labeled syngeneic T cells were adoptively transferred. (c) Same protocol as in b, except congenic B6.Thy1.1 CFSE-labeled T cells were used for transfer into B6.Thy1.2 mice. Data are representative of three experiments.

Figure 2

Homeostatic proliferation induces resistance to tolerance induction by costimulatory blockade. (a) B6 (WT) or B6-scid (SCID) mice received BALB/c cardiac allografts. B6-scid mice were injected intravenously with the indicated numbers of B6 lymph node cells on the day of transplantation (n = 5–10 mice per group). (b) B6-scid mice received BALB/c cardiac allografts plus 2 million B6 lymph node cells. Mice were also treated with DST, CTLA4Ig, control Ig or antibody to CD154 (200 μg/dose). All mice received irradiated DST on the day of transplantation (n = 5–10 mice per group). (c) B6-scid Thy1.2⁺ mice were injected with 2 million CFSE-labeled B6.Thy1.1⁺ lymph node cells and treated as indicated. Mice were killed on day 5 (left panels) or day 10 (right panels). Data are for live Thy1.1⁺ cells (representative of four experiments). (d) Euthymic B6 mice were given depleting antibodies to CD4 and CD8, received a BALB/c cardiac allograft 9 d later, and were treated with the indicated agents. Control mice (Δ) did not undergo prior T-cell depletion. P = 0.0038 for T-cell-depleted mice treated with control Ig and DST, compared with non-depleted mice treated with DST and CTLA4Ig (n = 6); P = 0.0255 for T-cell-depleted mice treated with CTLA4Ig and DST, compared with non-depleted mice treated with DST and CTLA4Ig (n = 12).

Figure 3

Kinetics of homeostatic proliferation and long-term effect on function and tolerance induction. (a) B6-scid mice were injected intravenously with 50 million B6 lymph node cells and followed for the indicated time points. Six days before sacrifice, BrdU was added to the drinking water. Normal B6 mice were fed BrdU for six days prior to sacrifice as a control. Data for live T cells are displayed (representative of six experiments). (b) B6-scid mice were injected intravenously with 50 million B6 lymph node cells 12 weeks before receiving a BALB/c heterotopic vascularized cardiac allograft. Mice were also treated with DST and control Ig or CTLA4Ig (n = 6 mice per group). (c) B6-scid mice were injected with 50 million B6 lymph node cells and killed after 12 weeks. Purified CD25⁺or CD25⁻ cells were tested for suppressor function. CD25⁺ cells from wild-type, unmanipulated B6 mice were used as controls. Histograms show proliferation, as CFSE dye dilution, of the responder population. R_f, calculated responder frequency (percentage of responder T cells that divided). Data shown are representative of two experiments.

Figure 4

Tolerance resistance in wild-type mice. (a) T cells that underwent homeostatic proliferation in scid mice dominantly transferred tolerance resistance. B6-scid mice were adoptively transferred with 50 million B6 lymph node cells. Mice were killed after 12 weeks, and post-homeostatic-proliferation T cells were purified. Unmanipulated B6 mice, B6 mice given 10 million post-homeostatic-proliferation T cells, and B6 mice given 10 million T cells from naive syngeneic mice received BALB/c cardiac allografts and treatment with DST and CTLA4Ig (n = 5–6 mice per group). P < 0.0001 for mice receiving post-homeostatic-proliferation cells, compared with unmanipulated mice or mice that received control naive cells. (b) T cells that underwent homeostatic proliferation in thymectomized wild-type recipients acquired tolerance resistance. Wild-type B6 mice were thymectomized, treated with 600 rad of irradiation and depleting antibodies to CD4 and CD8, and adoptively transferred with 50 million B6 lymph node cells. Thirteen weeks later, the animals received BALB/c cardiac allografts and treatment with DST and either control Ig (n = 5) or CTLA4Ig (n = 7). P = 0.7507.