Notch controls the magnitude of T helper cell responses by promoting cellular longevity

Abstract

Generation of effective immune responses requires expansion of rare antigen-specific CD4(+) T cells. The magnitude of the responding population is ultimately determined by proliferation and survival. Both processes are tightly controlled to limit responses to innocuous antigens. Sustained expansion occurs only when innate immune sensors are activated by microbial stimuli or by adjuvants, which has important implications for vaccination. The molecular identity of the signals controlling sustained T-cell responses is not fully clear. Here, we describe a prominent role for the Notch pathway in this process. Coactivation of Notch allows accumulation of far greater numbers of activated CD4(+) T cells than stimulation via T-cell receptor and classic costimulation alone. Notch does not overtly affect cell cycle entry or progression of CD4(+) T cells. Instead, Notch protects activated CD4(+) T cells against apoptosis after an initial phase of clonal expansion. Notch induces a broad antiapoptotic gene expression program that protects against intrinsic, as well as extrinsic, apoptosis pathways. Both Notch1 and Notch2 receptors and the canonical effector RBPJ (recombination signal binding protein for immunoglobulin kappa J region) are involved in this process. Correspondingly, CD4(+) T-cell responses to immunization with protein antigen are strongly reduced in mice lacking these components of the Notch pathway. Our findings, therefore, show that Notch controls the magnitude of CD4(+) T-cell responses by promoting cellular longevity.

Fig. 1.

Notch ligands increase recovery of activated CD4⁺ T cells. (A) Naive CD4⁺ T cells were stimulated with A20 cells expressing DLL4 (black triangles) or empty vector (open circles) in the presence of 1 μg/mL anti-CD28 and different concentrations of anti-CD3. Shown is the number of viable cells recovered at day 4. (B) Naive CD4⁺ T cells were stimulated with 10 μg/mL plate-bound anti-CD3, 1 μg/mL soluble anti-CD28, and 5 μg/mL plate-bound control-Ig or DLL4-Fc, and the number of viable cells was determined at day 5. (C) Naive CD4⁺ T cells were stimulated and analyzed as in B with different concentrations of control-Ig, DLL1-Fc (Left), or DLL4-Fc (Right). Black triangles/bars represent DLL1/4-Fc and open circles/bars control-Ig samples in B and C. Results are representative of 3 (A), more than 10 (B), and 2 experiments (C).

Fig. 2.

DLL4-Fc promotes survival of activated CD4⁺ T cells. Naive CD4⁺ T cells were stimulated as in Fig. 1B. (A) Viability (7AAD) and proliferation (CFSE) were analyzed on different days. (B–E) Black bars represent DLL4-Fc and open bars control-Ig samples. (B) Cell cycle distribution on day 2 (Upper) and 3 (Lower). Shown is the mean of three independent experiments + SEM. (C) Recoveries of total (Left) and viable cells (Right) on different days after activation. (D) Cells were stimulated as in Fig. 1B in the presence or absence of 20 μM QVD and analyzed at day 5. Percentages of viable cells are shown. Mean of three independent experiments + SEM is shown. (E) Naïve CD4⁺ T cells from Fas^lpr and control mice were activated and analyzed at day 5 as in Fig. 1B. A–D are representative of three experiments, and E is representative of two experiments.

Fig. 3.

DLL4-Fc-induced survival depends on Notch1, Notch2, and RBPJ. Naive CD4⁺ T cells were stimulated and analyzed at day 5 as in Fig. 1B. Black bars represent DLL4-Fc samples and open bars control-Ig samples. (A) Cumulative results using CD4⁺ T cells from three individually tested Notch1^fl/flCD4-Cre⁺ (N1ko) and five Notch1^fl/flCD4-Cre⁻ (WT) mice. (B) Cumulative results using CD4⁺ T cells from four individually tested Notch2^fl/flCD4-Cre⁺ (N2ko) and five Notch2^fl/flCD4-Cre⁻ (WT) mice. (C) CD4⁺ T cells from Notch1^fl/flNotch2^fl/flCD4-Cre⁻ (WT) or Notch1^fl/flNotch2^fl/flCD4-Cre⁺ (N1/2ko) mice. (D) CD4⁺ T cells from Notch3^+/−Notch4^+/− (WT) or Notch3^−/−Notch4^−/− (N3/4ko) mice. (E) CD4⁺ T cells from RBPJ^fl/flCD4-Cre⁻ (WT) or RBPJ^fl/flCD4-Cre⁺ (RBPJko) mice. A and B show the mean + SEM. ***P < 0.0001 (WT); ***P < 0.0005 (N1ko); **P < 0.004 (N1ko); *P < 0.0118 (N2ko) (unpaired, two-tailed t test). Results shown in C–E are individual mice representative of three experiments.

Fig. 4.

DLL4-Fc induces a broad survival program. Naive CD4⁺ T cells (Thy1.2⁺) were stimulated as in Fig. 1B for 3 d. (A) Surface expression of markers. Filled gray histograms represent staining controls; dark gray lines, specific staining control-Ig stimulated cells; black lines, specific staining DLL4-Fc stimulated cells. (B) Relative mRNA abundance for Bcl-2 (Left) and Bcl-3 (Right) was measured by quantitative RT-PCR and normalized against β-actin. Black bars represent DLL4-Fc samples and open bars control-Ig samples. Individual data points from six independent experiments are shown. Paired, two-tailed t test: **P = 0.0018 (Bcl-2) and **P = 0.003 (Bcl-3). (C) Western blot for Bcl-2 (Lower) or β-actin (Upper) using lysates from cells generated as in A. (D) Activated cells were transduced with retroviral expression vectors encoding NICD1 (Left), Bcl-2 (Center), Bcl-3 (Right) (all black bars) or control (open bars) linked to Thy1.1 via an IRES (internal ribosome entry site) sequence. The percentages of viable Thy1.1⁺ cells were determined after different periods of culture. Days indicate the time from the start of the culture. (E) Relative expression of selected genes as determined by Illumina mouse gene chip. All genes were significantly induced by DLL4-Fc (P < 0.01). (F) Surface expression of ICOS as in A.

Fig. 5.

Notch signaling is required for optimal CD4⁺ T-cell responses in vitro. Naïve OTII transgenic CD4⁺ T cells were stimulated with ovalbumin protein presented by splenic CD11c⁺ DCs for 5 d. Viability was analyzed as in Fig. 1A. Black bars represent WT (RBPJ^fl/flCD4-Cre⁻), and open bars represent RBPJ knock out (RBPJ^fl/flCD4-Cre⁺) CD4⁺ T cells from littermate mice. Results are representative of three experiments.

Fig. 6.

Notch signaling is required for optimal CD4⁺ T-cell responses in vivo. Mice were immunized s.c. with KLH in alum (A), KLH in SEA (B), or SEA alone (C and D). After 8 d, CD4⁺ T cells from draining lymph nodes were restimulated in vitro with splenic APCs loaded with KLH or SEA. [³H]thymidine incorporation during the last 12 of 72 h culture was measured and displayed as cpm. Mouse genotypes are as in Fig. 3. (A and B) WT represented by closed symbols; N1/2 KO, open symbols. (C) WT, closed circles and squares; N1 KO, open circles; N2 KO: open squares. (D) WT, closed circles; RBP KO, open circles. Each symbol represents a single mouse. Results are representative of at least three experiments.