CTLA4Ig inhibits effector T cells through regulatory T cells and TGF-β

Abstract

The CD28 costimulatory receptor is a critical regulator of T cell function, making it an attractive therapeutic target for the treatment of immune-mediated diseases. CTLA4Ig, now approved for use in humans, prevents naive T cell activation by binding to B7 proteins and blocking engagement of CD28. However, CTLA4Ig suppresses inflammation even if administered when disease is established, suggesting alternative mechanisms. We identified a novel, CD28-independent mechanism by which CTLA4Ig inhibits activated T cells. We show that in vitro, CTLA4Ig synergizes with NO from bone marrow-derived macrophages to inhibit T cell proliferation. Depletion of regulatory T cells (Tregs) or interference with TGF-β signaling abrogated the inhibitory effect of CTLA4Ig. Parallel in vivo experiments using an allergic airway inflammation model demonstrated that this novel mechanism required both macrophages and regulatory T cells. Furthermore, CTLA4Ig was ineffective in SMAD3-deficient mice, supporting a requirement for TGF-β signaling. Thus, in addition to preventing naive T cells from being fully activated, CTLA4Ig can turn off already activated effector T cells by an NO/regulatory T cell/TGF-β-dependent pathway. This mechanism is similar to cell-extrinsic effects of endogenous CTLA4 and may be particularly important in the ability of CTLA4Ig to treat chronic inflammatory disease.

Figure 1. CD28-independent inhibition of T cell proliferation by CTLA4Ig in vitro

CFSE labeled splenocytes were co-cultured with bone marrow derived macrophages and activated with PMA and ionomycin. Cells were harvested and stained with α-CD4 and proliferation determined by flow cytometry after 96 hours of culture. Where indicated, CTLA4Ig (10 μg/ml) and/or L-NMMA (100 μM) were included at the start of the culture. A) Representative CFSE dye dilution histograms of wild type splenocytes stimulated with PMA and ionomycin alone or in combination with BMDM and/or CTLA4Ig as indicated. The percentage of cells that have undergone at least one cell division is indicated. B) Splenocytes from wild type or CD28-deficient mice were co-cultured with wild type BMDM. Presented are the combined results from 6 independent experiments C) Splenocytes from wild type or NOS2 deficient mice were co-cultured with BMDM from wild type or NOS2 deficient mice. Presented are the combined results from 3 independent experiments. D) Splenocytes from wild type or CD80/CD86-double deficient mice were co-cultured with BMDM from wild type or CD80/CD86-double deficient mice. Presented are the combined results from 7 independent experiments. * = p<0.05, ** = p<0.01 by 2-tailed paired T-test

Figure 2. Inhibition of allergic airway inflammation by CTLA4Ig in vivo requires macrophages, and expression of CD80/CD86 and NOS2

A–C) C57Bl/6J mice were sensitized and challenged with OVA as described. Some mice received nothing, control liposomes or clodronate liposomes to deplete macrophages as indicated. Where indicated, CTLA4Ig was administered at the time of inhaled challenge only. D–F) Bone marrow chimeras established from C57Bl/6J or NOS2 deficient mice reconstituted with either C57Bl/6J or NOS2 deficient marrow mice. The mice were sensitized and challenged with OVA and treated with CTLA4Ig at the time of inhaled challenge. G–I) Bone marrow chimeras established with C57Bl/6J or CD80/CD86-double deficient mice reconstituted with bone marrow cells from C57Bl/6J mice were sensitized and challenged with OVA and treated with CTLA4Ig at the time of inhaled challenge. For all experiments, tissue was collected for analysis 72 hours after inhaled challenge. Each experimental group consisted of 5 mice and each experiment has been repeated 3 times. Shown is data from one representative experiment. * = p< 0.005 as compared to condition without CTLA4Ig as determined by Kruskall-Wallis with Dunns post test comparisons.

Figure 3. FoxP3+ cells are required for CTLA4Ig mediated inhibition of T cell proliferation and allergic airway inflammation

A–C) FoxP3-DTR mice were sensitized with OVA and challenged with either OVA or HEL. Mice were either untreated or administered DT to deplete FoxP3+ cells at the time of challenge. CTLA4Ig was administered to the indicated groups at the time of inhaled challenge. Specimens were collected 72 hours after inhaled challenge. Each experimental group consisted of 5 mice and has been repeated 3 times. Shown is data from one representative experiment. * = p< 0.05 as compared to condition without CTLA4Ig D) FoxP3-DTR mice were untreated or administered DT and splenocytes harvested 24 hours later, labeled with CFSE and co-cultured with BMDM alone, with CTLA4Ig or with L-NMMA. Proliferation was determined by flow cytometry after 96 hours. Presented are the combined results from 4 independent experiments. * = p<0.05 by 2-tailed paired T-test. E) splenocytes from OT-II/FoxP3-GFP/Rag1 KO mice (CD45.2+) were adoptively transferred into CD45.1+ C57Bl/6 recipients. The mice were then primed and challenged with OVA. One group received CTLA4Ig at the time of challenge. Spleen and lung draining lymph node harvested 48 hours later. Presented is the percentage of CD45.2+ cells that are FoxP3-GFP positive. Each experimental group consisted of 5 mice and has been performed 2 times. Shown is representative data from one independent experiment. *=p<.05, **= p< 0.01 as determined by Mann-Whitney U-test.

Figure 4. TGFβ is required for CTLA4Ig mediated inhibition of T cell proliferation

Splenocytes and BMDM were isolated from C57Bl/6J mice and co-cultured either alone or in the presence of A) α-TGFβ antibody or B) an inhibitor of TGFβR signaling. Proliferation was measured by flow cytometry after 96 hours. Presented are the combined results from 5 (panel A) and 3 (panel B) independent experiments. ** = p<0.01 by 2-tailed paired T-test

Figure 5. CTLA4Ig is ineffective in SMAD3-deficient mice

Wild type or SMAD3-deficient mice were sensitized with OVA and given an inhaled challenge with either OVA or HEL. CTLA4Ig was administered to groups of mice as indicated and samples were collected 72 hours after challenge. Each experimental group consisted of 5 mice and has been repeated 3 times. Shown is data from one representative experiment. * = p< 0.05 as compared to OVA primed and challenged without CTLA4Ig as determined by Kruskall-Wallis with Dunns post test comparisons.

Figure 6. Proposed model for CTLA4Ig

Binding of CTLA4Ig to CD80/CD86 expressed on an antigen presenting cell can prevent naïve T cell activation by preventing CD28 mediated costimulation (left panel). In addition, CTLA4Ig can suppress the proliferation of activated effector T cells (Teff) through a TGFβ and Treg dependent mechanism that is independent of CD28-costimulation (right panel). This pathway requires and is synergistic with macrophage derived NO.