Antigen-specific T-T interactions regulate CD4 T-cell expansion

Abstract

The regulation of CD4 T-cell numbers during an immune response should take account of the amount of antigen (Ag), the initial frequency of Ag-specific T cells, the mix of naive versus experienced cells, and (ideally) the diversity of the repertoire. Here we describe a novel mechanism of T-cell regulation that potentially deals with all of these parameters. We found that CD4 T cells establish a negative feedback loop by capturing their cognate major histocompatibility class (MHC)/peptide complexes from Ag-presenting cells and presenting them to Ag-experienced CD4 T cells, thereby inhibiting their recruitment into the response while allowing recruitment of naive T cells. The inhibition is Ag specific, begins at day 2 (long before Ag disappearance), and cannot be overcome by providing new Ag-loaded dendritic cells. In this way, CD4 T-cell proliferation is regulated in a functional relationship to the amount of Ag, while allowing naive T cells to generate repertoire variety.

Figure 1

Ag-experienced T-cell proliferation is preferentially inhibited during on ongoing immune response. (A) Two-cohort experimental design. A first cohort of 10⁶ naive CD45.2 CFSE-labeled Marilyn cells was injected intravenously into CD45.2 B6 mice that were immunized by injection of 10⁶ H-Y peptide–loaded LPS-matured DCs into the footpad the next day. After a variable (n) number of days, a second cohort of either 10⁶ naive, 5 × 10⁶ effector, or 2 × 10⁶ memory CD45.1 CFSE-labeled Marilyn cells was injected. CFSE profiles of gated TCR⁺CD45.1⁺ cells from the popliteal and inguinal lymph nodes draining the immunization site (DLN) were assessed 6 days after transfer of the second cohort. (B) Kinetics of the response of transferred Marilyn TCR Tg T cells. Naive (10⁶) or memory (2 × 10⁶) CD45.1 CFSE-labeled Marilyn cells were injected intravenously into B6 mice that were primed the next day by injection of 10⁶ H-Y peptide–loaded LPS-matured DCs into the footpad. The percentage of transferred T cells (CD45.1⁺TCR⁺) in the draining lymph node was measured at the indicated time points and compared with the percentage measured at day 1 in the absence of antigen. Each dot represents one mouse. Pooled data from 2 independent experiments. (C) Stronger inhibition of Ag-experienced versus naive T-cell proliferation. Naive (left panels), effector (middle panels), or memory (right panels) Marilyn T cells were injected into primed B6 hosts in the absence (top panels) or the presence (bottom panels) of a previous cohort of responding T cells injected 6 days earlier, as in panel A. Dot plots of gated TCR⁺CD45.1⁺ cells from the DLN are representative of at least 3 experiments with 2 mice each per group. The percentage of undivided cells, among total Tg Marilyn T cells, is indicated. (D) Antigen persists in vivo at least 18 days. H-Y peptide–loaded LPS-matured DCs were injected into the footpad of female B6 mice, previously injected (right panels) or not (left panels) with naive CD45.2 CFSE-labeled Marilyn LN cells. At the indicated time points, naive CD45.1 CFSE-labeled Marilyn LN cells were transferred intravenously. Dot plots of gated TCR⁺CD45.1⁺ cells from the DLN injected at the indicated time after DC priming and analyzed 6 days later. Representative of 2 independent experiments with 2 mice per group for each one. (E) Functional avidity measurement in vivo. Naive (10⁶) or memory (2 × 10⁶) CD45.1 CFSE-labeled Marilyn cells were injected intravenously into B6 mice, which were then primed by injection into the footpad of 10⁶ LPS-matured DCs loaded with the indicated amount of H-Y peptide. Five days later, the number of naive and memory T cells recovered in the DLN was measured.

Figure 2

The preferential inhibition of memory T cells is observed under physiological conditions. (A) Physiologic numbers of responding endogenous polyclonal T cells preferentially inhibit memory T-cell proliferation. Naive (10⁶; top panels) or memory (2 × 10⁶; bottom panels) CD45.1 CFSE-labeled Marilyn cells were injected intravenously into B6 (i-vi) or OT-II (vii,viii) mice immunized by injection of 10⁶ H-Y peptide–loaded LPS-matured DCs into the footpad. (i,ii) Low initial frequencies of endogenous responding T cells do not inhibit naive or memory Marilyn T cells; naive or memory CD45.1 Marilyn T cells were injected on the same day as the H-Y peptide–loaded DCs. (iii-vi) Higher frequencies occurring later in the response inhibit proliferation of newly entering cells. Six days after the initiation of the primary endogenous response by injection of peptide-pulsed DCs into naive B6 hosts (iii,iv), or after a boost of mice that had been immunized 30 days previously (v,vi), new naive or memory CD45.1 CFSE-labeled Marilyn cells were injected, and proliferation was analyzed 6 days later. (vii,viii) Extremely low initial frequencies do not expand to inhibitory numbers in 6 days. Marilyn T cells were injected and analyzed as in panels iii,iv, but into OT-II hosts. Plots are of gated TCR⁺CD45.1⁺ cells from the DLN, and are representative of at least 2 experiments with at least 2 mice each per group. (B) Small numbers of responding T cells strongly inhibit small numbers of memory T cells. Five thousand naive (top panels) or memory (bottom panels) CD45.1 Marilyn T cells were injected into primed B6 hosts in the absence (left panels) or the presence (right panels) of a previous cohort of 5000 CD45.2 Marilyn T cells injected 6 days earlier, as in Figure 1A. Pooled draining lymph nodes of 8 to 20 mice were studied. Dot plots of gated TCR⁺CD45.1⁺ cells with the percentage of undivided cells are shown. (C) An immune response to endogenous DCs that have processed Ag from a peripheral solid tumor preferentially inhibits memory T-cell proliferation. CD45.2 B6 mice received a transplant of an MCA101 tumor cell line, transfected with the DBY protein. Two days later, they were given a first cohort of CD45.2⁺ Marilyn T cells (right panels) or not (left panels). Naive (top panels) or memory (bottom panels) CD45.1⁺ Marilyn T cells were injected 5 days later. Dot plots of gated TCR⁺CD45.1⁺ cells from the immunization DLN are representative of at least 3 experiments with 2 mice each per group.

Figure 3

The inhibition is Ag specific and requires the presence of the first cohort of T cells. (A) The inhibition depends on the presence of the first cohort of T cells. Naive (10⁶) or memory (2 × 10⁶) Marilyn T cells were injected into OT-II hosts that had been primed 6 days earlier with H-Y peptide–loaded mature DCs, in the presence (iii-vi) or not (i,ii) of a first cohort of Marilyn-Lat-DTR T cells. The day before the injection of the second cohort, Marilyn-Lat-DTR cells were deleted (v,vi) or not (iii,iv) by the injection of DT. Dot plots of gated CD4⁺CD45.1⁺ cells from the DLN studied 6 days later (day 6 + 6) are representative of at least 2 experiments with 2 mice each per group. (B) The inhibition is Ag specific. CD45.1 B6 mice were injected or not (i-iv) with a first cohort of naive CFSE-labeled Marilyn (v-viii) or OT-II (ix-xii) Tg CD4 T cells and were then immunized with LPS-matured DCs pulsed with both OVA and H-Y peptides. Six days later, CFSE-labeled naive (top panels) or memory (bottom panels) OT-II and Marilyn cells were coinjected in the same mouse. Proliferation was measured on the second cohorts (CD45.2⁺, and Vβ6⁺ for Marilyn, Vα2^hi for OT-II) in the same mouse 6 days later. Representative of 2 independent experiments with 2 mice each per group.

Figure 4

Activated T cells acquire MHC-II molecule during the ongoing immune response in vivo. (A) Acquisition of MHC-II molecules by activated T cells in vivo. Naive Tg CD45.1 Marilyn cells (10⁶) were injected intravenously into B6 mice that were immunized by injection of 10⁶ H-Y peptide–loaded A^b-GFP knockin DCs into the footpad. Two days later, CD69, A^b, and GFP expression were analyzed on the transferred T cells in the draining LN and the nondraining LN (CD45.1⁺TCR⁺) by 5-color FACS with doublet exclusion. (B) Proliferating T cells acquire MHC-II in vivo. Naive CFSE-labeled Tg CD45.1 Marilyn cells (10⁶) were injected intravenously into B6 mice and immunized as in panel A. Two days later, CD69 and A^b expression was analyzed in the draining LN on the transferred T cells (CD45.1⁺TCR⁺) by 5-color FACS with doublet exclusion. (C) The inhibition starts at day 1 and plateaus by day 2. H-Y peptide–loaded DCs (10⁶) were injected into the footpad of female B6 mice that contained (or not) a first cohort of 10⁶ naive CD45.2 CFSE-labeled Marilyn LN cells. At the indicated time points, a second cohort of 10⁶ CD45.1 CFSE-labeled naive Marilyn LN or 2 × 10⁶ memory Marilyn T cells was injected, and the CFSE profile assessed 6 days later. The percentage of divided TCR⁺CD45.1⁺ Marilyn cells in the second cohort is shown for cells in the presence of a first cohort relative to the percentage seen in its absence. Each point is an individual mouse.

Figure 5

Marilyn CD4 T cells capture MHC-II/H-Y in vitro, and induce death of memory but not naive T cells. (A) Purification of MHC-II–bearing T cells. Lymph node CD45.2 Marilyn cells were incubated with H-Y peptide–loaded (100 nM), CFSE-labeled, LPS-matured DCs. Twenty hours later, the T cells were FACS sorted according to CD11c and CFSE expression with exclusion of the doublets. (i) Postsort analysis of the CFSE^neg/CD11c^neg T-cell fraction. (ii,iii) Expression of A^b by the sorted T cells. (B) T cells that have captured MHC-II/H-Y complexes present them to naive and memory Marilyn T cells. (i,ii) CD45.2 fixed (as described in “In vitro experiments”) naive LN Marilyn cells (0.5 × 10⁶), (iii,iv) 0.5 × 10⁶ purified CD45.2 fixed T cells that had captured MHC-II/H-Y complexes, or (v,vi) 5000 purified H-Y–bearing fresh DCs were used to stimulate 2 × 10⁴ CFSE-labeled CD45.1 naive Marilyn LN cells or negatively purified memory Marilyn T cells. CD69 expression was measured the following day. Representative of 3 experiments. (C) T cells that have captured MHC-II/H-Y complexes suppress memory but not naive Marilyn T-cell proliferation. Purified CD45.2 fixed naive LN Marilyn cells (0.5 × 10⁶), or 0.5 × 10⁶ purified CD45.2 fixed T cells that had captured MHC-II/H-Y complexes (H-Y T cells), or not (unpulsed T cells), and/or various numbers of H-Y–bearing fresh DCs were used to stimulate 2 × 10⁴ CD45.1 naive or memory Marilyn T cells that had been negatively purified by FACS and then CFSE labeled. Proliferation was measured at day 3 by CFSE dilution. Percentage of undivided cells is indicated in each panel, and numbers of recovered cells/well are indicated at the bottom of each panel. Representative of 4 experiments. (D) Fresh DCs, able to stimulate naive cells, cannot relieve inhibition of memory T cells. B6 mice were injected with CD45.2 CFSE-labeled Marilyn cells and were immunized the next day with H-Y peptide–loaded LPS-matured DCs in one hind footpad. Naive or memory CD45.1 CFSE-labeled Marilyn cells were transferred 5 days later. Four hours later, some of the mice (bottom panels) received a second immunization with H-Y peptide–loaded DCs into both the original and the contralateral footpad. The CFSE profile is shown on the second cohort (TCR⁺ CD45.1⁺ cells) in the original DLN (left panels) or in the contralateral DLN (right panels) 6 days after they were injected. Representative of 2 experiments with 2 mice per group in each.

Figure 6

Model of CD4 T-cell autoregulation of an immune response. (1) At the onset of the immune response, naive (and also memory in the case of a secondary response, which is not represented here) CD4 T cells are recruited and stimulated by activated APCs displaying antigen. (2) During activation, these CD4 T cells capture the specific MHC-peptide complexes bound by their TCRs. (3) At later time points, the number of activated CD4 T cells that have captured MHC-peptide complexes increases. As these activated Ag-bearing T cells begin to outnumber the APCs, there is increasing probability that newly arriving T cells (and any T cells that have just been activated and need a second hit to continue to divide) will encounter them before encountering a proper professional APC. (4) This T-T interaction leads to inhibition of the Ag-experienced T cells, whereas the normal interaction of naive cells with the professional APCs leads to proliferation.

Figure 7

Ag-presenting Marilyn T cells directly inhibit memory T-cell proliferation in vivo. (A,B) Naive (10⁶) or memory (2 × 10⁶) CD45.1 CFSE-labeled Marilyn T cells were injected intravenously into CD45.2 female hosts. In parallel, CD45.1/2 Marilyn T cells were incubated in vitro overnight with CFSE-labeled H-Y peptide–loaded LPS-matured DCs or with DCs and an anti-TCR antibody. The next day, the 2 sets of in vitro–activated Marilyn T cells were FACS sorted according to CFSE and CD11c expression and injected into the footpad. Six days later, H-Y–loaded DCs were injected into the footpad to stimulate the T cells. CFSE patterns were studied in the DLN 6 days later. Dot plots on gated CD45.1⁺/CD45.2⁻/Vb6⁺/CD4⁺ cells are representative of 2 independent experiments with 2 mice in each.