Macrophage IL-4 polarization is restricted by soluble CD40 agonists and antigen-induced CD154 but not by constitutive CD154 expressed by CD4⁺ T cells

Abstract

Stimulation of macrophages via CD40 promotes their classical activation. Therefore, CD154 (CD40 ligand) can be expected to oppose macrophage polarization and proliferation induced by IL-4. However, there are limited experimental data to support this, which is additionally complicated by the possibility of differential effects of CD40 agonists in different formats/contexts. Whereas canonically CD4⁺ T cells upregulate CD154 strongly following exposure to cognate antigen, naïve CD4⁺ cells constitutively express significant levels of CD154, which could be a tonic signal. Soluble CD154 and agonistic CD40 antibodies also trigger CD40 signaling. We explored these questions in a reductionist model of IL-4 delivery to mouse peritoneal cavity cells in vitro and in vivo. Soluble CD40 agonists inhibited M(IL-4) polarization, with a stronger effect on RELM-α than on Ym1 (Chil3), as well as inhibiting IL-4-induced proliferation. CD154 provided by CD4⁺ cells in the context of an antigen-specific interaction blunted macrophage RELM-α expression but did not affect Ym1. Macrophages negatively regulated, via CD40, constitutive cell-surface CD154 on naïve CD4⁺ cells, both in vitro and in vivo. The large peritoneal macrophages of CD40 KO mice showed a moderately enhanced RELM-α response to IL-4, but this was not a cell-autonomous effect. No differences between WT and CD40 KO mice were detected in IL-4-induced macrophage proliferation. We conclude that strong CD40 stimulation, including stimulation by CD154 expressed by antigen-specific CD4⁺ cells, blunts selected macrophage responses to IL-4, and that constitutive CD4⁺-cell CD154, in spite of interacting with CD40 on macrophages, does not directly influence macrophage responses to IL-4.

Constitutive CD40 ligand expressed by CD4 cells interacts with macrophage CD40 but does not affect macrophage responses to IL-4, whereas stronger CD40 agonists inhibit these responses.

1. Introduction

CD154 (CD40L) is a TNFα-family protein that is expressed on the surface of activated CD4⁺ T cells but can also be expressed on other cell types.^1–4 The major receptor for CD154 is CD40, a TNFα receptor family cell-surface protein that is expressed by antigen-presenting cells (APC) and various other cell types. The CD154-CD40 interaction participates in CD4⁺ T cell priming by dendritic cells (DCs), CD4⁺ T cell help to B-cells, classical monocyte-macrophage activation by CD4⁺ T cells, and further immune processes.^2–14

CD154 protein is present in intracellular stores in antigen-experienced CD4⁺ T cells, and it is translocated to the cell surface in response to TCR stimulation.^15,16 TCR stimulation also strongly stimulates transcription of the CD154 gene, presumably in part to fill or refill the intracellular stores.^17–21 In addition, naïve CD4⁺ T cells constitutively express CD154 mRNA and protein and the latter can reach the cell surface.^{16  ,22  ,23} The exposure of CD154 at the cell CD4⁺ T cell surface is strongly inhibited by interactions with CD40-expressing cells, which cause internalization and lysosomal degradation of CD154.^17,18,24 This form of regulation affects both CD154 externalized after TCR triggering and constitutive CD154, and it is exerted by CD40-expressing cells engaged or not in cognate interaction with the CD4⁺ cell.^{16–18,23,24} The in vivo availability of constitutive CD154 as measured in lymphoid organs is regulated both by B and non-B cells.²³

CD154 can have effects on CD40-expressing cells in the absence of cognate interactions, as exemplified by the biological effects of soluble CD154, CD40 agonist antibodies and cells with forced expression of CD154.^{6,7,12,13,18,24–27} This opens the possibility that constitutive CD154 expressed by naïve CD4⁺ cells may impact CD40-expressing cells. Constitutive CD154 has been reported to influence monocyte activation in vitro²¹ and the survival of autoreactive B cells in an autoimmunity model.²³ Thus, constitutive cell-surface CD154 could represent a tonic signal that is kept at low levels (in WT animals) as a result of its consumption through interactions with CD40-expressing cells.^16,23

The role of the CD154-CD40 interaction in type 2-biased immune responses is not well understood.¹⁰ In particular, there is a paucity of information on the role of this interaction in macrophage responses in type 2 contexts, in which macrophages undergo M (IL-4) polarization and proliferation in response to IL-4 and IL-13.^28–30 Given that soluble CD40 agonists and CD154 expressed by antigen-specific CD4⁺ T cells enhance classical macrophage activation,^5–9 it seems likely that CD40 signaling antagonizes M(IL-4) polarization. Indeed, the limited information available indicates that agonistic CD40 antibody inhibits expression of M(IL-4) markers in tumor-associated macrophages in vivo.⁵ However, the outcome of CD40 signaling can vary substantially depending on the strength of stimulation and the specific phenotype of the myeloid cell that receives the signal^31,32 Accordingly, endogenous CD154 (both from lymphoid and non-lymphoid cells) is generally required for the full priming of Th2 responses but exogenous CD40 agonists instead polarize towards Th1 responses.^32–39

No information is available on the impact of CD154 on macrophage proliferation induced by IL-4. In B cells, the combined stimulation through CD40 and the IL-4 receptor promotes proliferation and polyclonal IgE production.^{25–27,40–46} If CD154, including constitutive CD154, was necessary for IL-4-induced macrophage proliferation, this may provide an explanation to a standing question in this area: why IL-4 as sole exogenous stimulus efficiently and directly induces macrophage proliferation in vivo but hardly does so in vitro.^29,47 Given its subtle regulation, it is conceivable that the availability of constitutive cell-surface CD154 is very different in cell cultures from what it is seen in vivo.²³

In this work, we addressed the impacts of exogenous CD40 agonists, CD154 provided by specific CD4⁺ T cells and constitutively expressed CD154 on the responses of peritoneal cavity macrophages to IL-4. We provide evidence that the first 2 classes of CD40 agonists have negative impacts on some macrophage responses to IL-4 whereas constitutive CD154 does not significantly influence such responses in spite of being regulated by macrophage CD40. In this context, we show that macrophages and B cells exposed to IL-4 and CD40 agonists respond in starkly different ways in terms of cell proliferation, so that poor CD154 availability certainly does not explain the absence of IL-4-induced macrophage proliferation in cell culture.

2. Materials and methods

2.1. Mice

WT C57BL/6 mice were purchased from DILAVE (Ministry of Agriculture and Fisheries of Uruguay). B6.129P2-Cd40tm1Kik/J (CD40 gene-deficient, CD40 KO) mice originally imported from Jackson Laboratories, as well as their WT controls, were provided by URBE (Facultad de Medicina, Universidad de la República, Uruguay). CD40 KO mice were co-housed with the WT controls for at least 3 wk in order to equilibrate their microbiota. B6.Cg-Tg(TcraTcrb)425Cbn/J (OT-II) mice imported from Jackson Laboratories were provided by UBAL (Institut Pasteur of Montevideo). Mice were kept at the Instituto de Higiene animal house. All mice were used between 8 and 14 wk of age. Mice were euthanized using isoflurane. The procedures used were approved by the Comisión Honoraria de Experimentación Animal (CHEA, Universidad de la República, Uruguay: protocols 101900-000491-18, 101900-000520-18, 101900-001483-19, 070151-000031-22, 070151-000039-22, 070151-000040-22, 070151-000042-22, 070151-000019-24, 070151-000020-24, 070151-000006-24 and 070151-000030-24).

2.2. Biological reagents

Recombinant mouse IL-4 was purchased from PeproTech. The anti-IL-4 antibody 11B11 was from BioXCell. The CD40-agonistic 1C10 antibody was purchased from Biolegend, in “Ultra-LEAF^TM” format and the GL117 antibody from PolPharma was used as control. The CD154-blocking MR1 antibody and its control (monoclonal Armenian hamster IgG) were purchased from Biolegend, in “Ultra-LEAF^TM” format. The CpG oligonucleotide was from Enzo.

2.3. Flow cytometry

Cells were plated in 96-well V-bottom plates, stained with viability probe, blocked using 25 ng per well of anti-CD16/32 (Biolegend) in 10% v/v normal rat serum (URBE, Facultad de Medicina, Universidad de la República, Uruguay) in FACS buffer (PBS, 2 mM EDTA, 0.1% w/v BSA), and surface-stained. Cells were then either fixed in 1% w/v PFA, washed, and permeabilized overnight with permeabilization buffer (Thermo) for staining RELM-α, Ym1 (Chil3) and/or intracellular CD154, or they were incubated overnight in eBioscience™ Foxp3/Transcription Factor Fixation/Permeabilization (Thermo) for staining Ki-67, BrdU and/or EdU. For BrdU staining, cells were treated with DNAse (Sigma; 50 μL per well of preparation containing 690 units/mL) after permeabilization and before intracellular staining. EdU was detected with Alexa Fluor 647-azide using Click-iT™ Plus EdU Flow Cytometry Assay Kit assay kit (Thermo) according to the manufacturer's instructions. Finally, cells were washed, spun at 350 g for 5 min, and re-suspended in FACS buffer. The antibodies and other probes used are summarized in Table S1. The data were acquired in a FACS Canto II Cytometer (BD) or in a BD LSRFortessa™ X-20 and then analyzed using the FlowJo software package. Absolute cell numbers informed were obtained by multiplying total cell numbers obtained with the help of cell counters (Nexcelom Cellometer K2 or ImmunoSpot CTL, using the AOPI stain for cell viability) by the frequency of the cell type in the sample determined by flow cytometry. Thresholds of expression were set with the help fluorescence minus one (FMO) controls; for CD40, this threshold coincided with the threshold set using CD40 KO cells. For BrdU and EdU, “no DNAse” and “no Alexafluor 647 picolyl azide” controls were used respectively, instead of FMO controls.

2.4. Quantitation of cytokines by ELISA

RELM-α and Ym1 in cell supernatants of peritoneal cavity lavage fluid were measured with the help of purified and biotinylated antibodies from PeproTech followed by streptavidin-peroxidase (Thermo) and the Mouse Ym1/Chitinase 3-like 3 DuoSet ELISA (RyD), respectively.

2.5. Cell culture assays with peritoneal cavity cells or isolated peritoneal macrophages

Three hundred thousand unfractionated peritoneal cavity cells from mice either untreated or injected 3 d prior with 800 μL of aged 4% m/v thioglycollate medium (Sigma) were plated per well in 48-well plates, as source of unstimulated or recently recruited macrophages, respectively. Alternatively, unstimulated macrophages were negatively purified from peritoneal cavity cells of untreated mice using biotinylated antibodies against CD19, SiglecF, CD3, Gr1, TCRβ, and Ly6C (Biolegend) and Dynabeads biotin binder (Thermo) according to the beads manufacturer's instructions, and 150,000 macrophages were plated per well. In all cases, the cells were plated in RPMI medium with 10% v/v fetal bovine serum (FBS) and left overnight before addition of the stimuli and further culture for 48 h. IL-4 was used at 80 ng/mL, sCD154 at 1 to 2 μg/mL, the CD40 agonist antibody 1C10 or corresponding control antibody at 4 μg/mL, and M-CSF at 40 ng/mL. When used, BrdU (10 μM) was added 24 h before end of culture. In experiments involving carboxyfluorescein succinimidyl ester (CFSE) dilution, cells were stained previous to their culture with the Vybrant™ CFDA SE Cell Tracer Kit (Thermo) following the manufacturer's instructions.

2.6. Assays involving co-culture of peritoneal macrophages with spleen cells

The cell preparations used as sources of WT or CD40 KO peritoneal cavity macrophages were plated first, in 48-well plates, and in RPMI with 10% v/v FBS, then 600,000 spleen cells per well (as source of CD4⁺ T cells) were added in the mentioned medium containing 55 μM beta-mercaptoethanol (Gibco), and the cells co-cultured in the absence or presence of IL-4, for 3 h for studying CD154 regulation or 48 h for measuring macrophage polarization and/or CD4⁺ T cell proliferation. For studying the interaction between unstimulated macrophages and specific CD4⁺ T cells, peritoneal cavity cells from untreated mice were depleted of B cells, as described above for macrophage isolation but using only anti-CD19; then 200,000 macrophage-enriched cells were plated per well, followed by spleen cells from “OT-II” mice. For studying the interaction between recently recruited macrophages and specific CD4⁺ T cells, 300,000 unfractionated peritoneal cavity cells obtained from mice injected with thioglycollate medium were plated per well, followed by spleen cells from “OT-II” mice. In both these experimental systems, the macrophage-containing cell preparations were incubated for 2 h in serum-free medium with ovalbumin MHCII peptide (1 μg/mL; Genscript) or with medium only, and washed before addition of the spleen cells. For studying the interaction between unstimulated macrophages and CD4⁺ cells expressing constitutive cell-surface CD154, 150,000 macrophages isolated from peritoneal cavity cells were plated per well, followed by spleen cells from CD40 KO mice.

2.7. In vivo experiments in mice

IL-4 (1 μg) was injected i.p. or s.c. (as indicated for each experiment) in complex with 5 μg of anti-IL-4 antibody (clone 11B11); this complex is termed IL-4c.²⁹ Effects of IL-4c were always assessed 48 h post-injection. The CD40 agonist antibody 1C10 or its control was injected i.p. in a dose of 150 ng per g of body mass, at the time of injection of IL-4c or vehicle. The CD154 blocking antibody MR1 or its control were injected i.p. in a dose of 100 μg per mouse, at the time of injection of IL-4c or vehicle.

BrdU (1 mg per mouse) or EdU (0.5 mg per mouse) was injected s.c. 16 h before euthanasia. For experiments involving transfer of peritoneal macrophages, donor macrophages were labelled in vivo with the PKH26-PCL probe (Sigma), by injecting i.p. 175 pmol of probe per mouse in the diluent provided by the manufacturer, either 2 or 24 h before euthanasia. Labelled macrophages were isolated as previously described from peritoneal cavity cells pooled from either 5 WT or 5 CD40 KO mice, and 200,000 macrophages in 250 μL of PBS were transferred to each recipient mouse. Recipient mice were injected with IL-4c one week after cell transfer. In experiments for verifying that the PKH26-PCL does not alter macrophage responses to IL-4, mice were administered IL-4c 24 h after injection of the probe or vehicle. In experiments involving the transfer of both peritoneal macrophages and CD4⁺ T cells, first 150,000 CD4⁺ cells, purified (∼90% purity) from the spleen of CD40 KO mice using the Dynabeads® Untouched^TM Mouse CD4 cells kit (Thermo), were injected i.p. in 250 μL volumes into recipient CD40 KO mice. One day later, 200,000 unlabeled peritoneal cavity macrophages isolated from WT or CD40 KO donors as previously described (or vehicle only) were injected, and mice were euthanized 48 h later.

2.8. Statistical methods

Data arising from 2 or more independent experiments were analyzed by 2-way parametric statistics. Specifically, the robustlmm R software package was used for robust estimation of linear mixed-effect models, and then the estimated marginal means were obtained with the help of the emmeans R package.^48,49 In the cases in which a single experiment was available, one-way ANOVA or t-test was used. The Benjamini and Hochberg correction for minimizing the false discovery rate⁵⁰ was used in all analyses involving comparisons between more than two experimental groups. Paired data, corresponding to donor and recipient cells within each mouse, were analyzed by paired t-test. When necessary for homogenizing variances of different groups, data were log-transformed, as indicated in each case; in such cases, geometric means (instead of arithmetic means) are shown in the corresponding graphs.

3. Results

3.1. Exogenous soluble CD40 agonists inhibit responses to IL-4 in peritoneal cavity macrophages

Before assessing the impact of CD40 agonist on IL-4 mediated macrophage activation, we assessed CD40 expression in our experimental system. Under basal conditions, peritoneal cavity macrophages comprise the numerically dominant large peritoneal macrophages (LPM) and the less abundant small peritoneal macrophages (SPM)⁵¹ (see Supplementary Fig. S1a–d for peritoneal cell populations and their analysis by flow cytometry). As expected, LPM, SPM, and B cells were found to express CD40, expression being higher in LPM than in SPM and B cells (Supplementary Fig. S1e). IL-4 delivery to the peritoneal cavity caused slight decreases in CD40 expression in LPM and SPM, and had the opposite effect on B cells (Supplementary Fig. S1e). In 48 h cultures of unstimulated peritoneal cavity cells, CD40 in macrophages was enhanced by IL-4 (see Supplementary Fig. S2a–d for these observations and flow cytometry of peritoneal cell cultures). In cell culture experiments, we did not attempt to discriminate LPM from SPM, as LPM cultured in the absence of retinoic acid undergo partial de-differentiation,^52,53 so that they can no longer be distinguished from SPM on the basis of F4/80 and MHCII.

In unfractionated peritoneal cavity cell cultures, two CD40 agonists, namely soluble CD154 (sCD154) and the agonistic antibody 1C10, strongly inhibited the induction of the M(IL-4) product RELM-α, whereas they had marginal or weak inhibitory effects on Ym1 and CD206 (mannose receptor) (Fig. 1a). The blunting of RELM-α induction was also observed with isolated macrophages, suggesting that it is due to a direct signal on macrophages, whereas Ym1 was not significantly inhibited in this system (Fig. 1b; see Supplementary Fig. S3A and B for macrophage isolation). In CD40 KO cells, the inhibition of RELM-α expression by sCD154, even if weaker and less consistent than in WT cells, was significant (Fig. 1c), suggesting that receptors additional to CD40 contribute to the inhibition observed in WT cells. Ym1 was not inhibited by sCD154 in CD40 KO cells (data not shown). As expected, the agonistic CD40 antibody was not inhibitory in CD40 KO cells (Supplementary Fig. S2E).

Fig. 1.

Soluble exogenous CD40 agonists inhibit the polarization and proliferation of basal peritoneal cavity macrophages induced by IL-4. Unfractionated peritoneal cavity cells from untreated WT mice (a) or macrophages isolated thereof (b) were cultured for 48 h in the presence of IL-4 and/or soluble CD154 (sCD154) or the CD40-agonistic antibody 1C10; then RELM-α, Ym1, and CD206 measured by flow cytometry as indicators of M (IL-4) polarization. RELM-α was also measured in macrophages present in unfractionated peritoneal cavity cells from CD40 KO (or as a control WT) mice treated with sCD154 in the absence or presence of IL-4 (c). The CD40 KO vs WT cell results are presented after normalization over the response of cells from each genotype to IL-4 as sole stimulus, and each point represents an independent experiment; statistical comparisons between genotypes were carried out on the normalized data, while comparisons within each genotype including comparisons with the response to IL-4 alone (vertical arrows) were carried out on the crude data. Also, in the unfractionated WT cell cultures, Ki-67 expression, CFSE dilution and BrdU incorporation were also measured as indications of proliferation, in macrophages (d) and B cells (e). M-CSF was included as a positive control condition in (d). Also, WT mice were injected i.p. with the 1C10 antibody or control antibody GL117 together with IL-4c or its vehicle control, and 2 days later euthanized for the analysis by flow cytometry of RELM-α and Ym1 in Tim4⁺ and Tim4⁻ macrophages (f) as well as Ki-67 expression and BrdU incorporation by Tim4⁺ and Tim4⁻ macrophages (g) and B cells (h). Except in part (c), data from different independent experiments are shown with different symbols. Asterisks without connecting lines indicate differences between the conditions that include IL-4 stimulation and the corresponding conditions without IL-4.

As expected,⁴⁷ IL-4 did not induce the proliferation of macrophages in cell cultures (Fig. 1d). sCD154 also did not elicit macrophage proliferation, by itself or in the presence of IL-4, in contrast to its effect on peritoneal cavity B cells (Fig. 1e). Although macrophages in unfractionated cell cultures did incorporate BrdU after stimulation with sCD154 plus IL-4, this was not accompanied by expression of Ki-67 or dilution of CFSE, both of which were readily detected after stimulation with the macrophage mitogen M-CSF (Fig. 1d). This suggested that macrophages in unfractionated cell cultures exposed to stimuli known to cause B cell proliferation (ie CD40 agonist plus IL-4) may take up BrdU from B cells, probably through phagocytosis.

We also tested the effects of CD40 agonists on inflammatory macrophages (IMacs) recruited to the peritoneal cavity in response to thioglycolate medium. The term “inflammatory” in this context does not imply that these cells are classically activated but rather to the fact that they originate from monocytes recruited recently and in response to a stimulus. IMacs also expressed CD40 although at lower levels than LPM (Supplementary Fig. S4A–D). Similar to isolated unstimulated macrophages (Fig. 1b), in IMacs, sCD154 inhibited RELM-α expression but had only a marginal effect on Ym1 at the dose tested (Supplementary Fig. S4E). Also similar to the results in unstimulated macrophages (Fig. 1d), sCD154 did not bring about proliferation of IMacs in the absence or presence of IL-4, whereas it did cause proliferation of B cells in the same cultures (Supplementary Fig. S4F and G). Secondary incorporation of BrdU into macrophages was not detected, possibly because B cells make only a small fraction of the cells in this preparation (Supplementary Fig. S4B).

Similar to what was observed in the cell culture experiments, in vivo the CD40 agonistic 1C10 antibody blunted IL-4-induced RELM-α expression without significantly inhibiting Ym1 expression (Fig. 1f). This was observed both in Tim4⁺ macrophages (the majority of LPM) and Tim4⁻ macrophages (a minority of LPM and all SPM)⁵⁴ (see Supplementary Fig. S5). The definitions of LPM and SPM used in other experiments could not be used when injecting the 1C10 antibody, which altered F4/80 and MHCII expression in LPM, without altering the fraction of Tim4⁺ cells within all cavity macrophages (data not shown). Proliferation of both Tim4⁺ and Tim4⁻ macrophages induced by IL-4 was strongly inhibited by the CD40 agonist whereas the opposite effect was observed in the case of B cells (Fig. 1g and h). Secondary acquisition of BrdU by macrophages was not observed in vivo.

In summary, soluble exogenous CD40 agonists antagonize the effects of IL-4 on peritoneal cavity macrophages in terms of RELM-α expression and of proliferation, and they inhibit Ym1 expression only weakly.

3.2. The effects on macrophage M(IL-4) polarization of CD154 delivered by specific CD4⁺ T cells are similar to those of soluble CD40 agonists

Next we set up an in vitro system to find out if soluble CD40 agonists mimic the effects of CD154 provided by a CD4⁺ T cell engaged in antigen-specific interaction with a macrophage. Unstimulated B-cell depleted peritoneal cavity cells as a source of macrophages were pulsed with ovalbumin peptide and cultured with spleen cells of “OT-II” mice as a source of antigen-specific CD4⁺ T cells, in the absence or presence of IL-4 (Fig. 2a; see Supplementary Fig. S3C and D for B cell depletion). Antigen presentation in the culture caused some CD4⁺ T cells to express cell-surface CD154, as measured 3 h after commencement of co-culture (Fig. 2b and c); cell-surface CD154 was partly repressed when CD40 was present in the macrophages (compare WT vs CD40KO) but was not altered by IL-4. Antigen presentation also caused proliferation of CD4⁺ cells, which in our system was independent of whether macrophages expressed CD40 or not (Fig. 2d). In the presence of an antigen-specific interaction with CD4⁺ T cells, the up-regulation of RELM-α induced by IL-4 reached higher levels in CD40 KO than in WT macrophages, whereas no analogous difference was observed for Ym1 (Fig. 2e and f). The same conclusion was obtained when the data were normalized over the responses in the absence of antigen presentation, as to eliminate any effect of differences between the WT and CD40 KO macrophage preparations in their intrinsic responses to IL-4 (data not shown). This suggests that the effects on M(IL-4) polarization of CD154 provided by CD4⁺ T cells in specific interaction with macrophages are similar to those of soluble CD40 agonists. The antigen-specific interaction with CD4⁺ T cells inhibited RELM-α and also weakly Ym1 expression even in CD40 KO macrophages. Both CD40-independent effects of antigen-induced CD4⁺-cell CD154 (similar to those observed for soluble CD154; Fig. 1c) and IFN-γ produced by antigen-stimulated CD4⁺ T cells even in the absence of macrophage CD40 (data not shown) could contribute to these observations.

Fig. 2.

Cell-surface CD154 expressed by CD4⁺ T cells in specific interaction with basal peritoneal macrophages in cell culture negatively regulates IL-4-induced RELM-α. B-cell depleted peritoneal cavity cells from WT or CD40 KO mice as a source of macrophages were loaded with ovalbumin MHC class II peptide or not, washed and co-cultured with spleen cells from “OT-II” mice as a source of antigen-specific CD4⁺ T cells, in the absence or presence of IL-4, as summarized in (a). Cell-surface CD154 in CD4⁺ T cells was measured after 3 h, as shown in (b) through representative dot-plots and in (c) with % positive cell data for individual culture wells. After 48 h, Ki-67 was measured in CD4⁺ T cells as an indication of proliferation (d), and RELM-α (e) and Ym1 (f) were measured in macrophages. In experiments not shown it was verified that the spleen cell preparation used did not contain significant numbers of macrophages. Data from different independent experiments are shown with different symbols. Asterisks without connecting lines indicate differences between the conditions that include IL-4 stimulation and the corresponding conditions without IL-4.

In an analogous experiment setup, the M(IL-4) polarization of IMacs was not detectably affected by CD154 (Supplementary Fig. S6A–F), in possible agreement with the observations that IMacs express lower levels of CD40 than LPM (Supplementary Fig. S4C).

In sum, CD154 expressed by CD4⁺ T cells in antigen-specific cognate interaction with unstimulated macrophages can antagonize the effects of IL-4 in terms of RELM-α expression, similar to what was observed for soluble CD40 agonists.

3.3. CD154 is constitutively available in the peritoneal cavity as a result of its cell-surface expression by naïve CD4 + cells

With the purpose of analyzing if constitutive CD154 may be a tonic signal for macrophages in tissues, we first analyzed the availability of CD154 in the basal peritoneal cavity. In WT mice, CD154 was only detected clearly in CD4⁺ T cells, mostly intracellularly but also at the cell surface (Fig. 3a). Discrimination of CD4⁺ T cells into naïve, central memory (CM) and effector memory (EM) phenotypes on the basis of expression of CD62L and CD44 showed that cell-surface CD154 was present in WT animals mostly in naïve cells (Fig. 3b). In contrast, total (surface plus intracellular) CD154 levels were considerably higher in EM than in CM and naïve CD4⁺ T cells (Fig. 3b). The levels of cell-surface as well as total CD154 were much enhanced in CD40 KO animals in comparison to WT, especially in naïve CD4⁺ T cells. Of note, cell-surface CD154 levels in CD40 KO mice can be taken as an indication of “potential” CD154 in WT animals, ie the total amount of signal that would be available if CD40-expressing cells did not “consume” it. Our results for CD154 availability in the peritoneal cavity are largely coincident with published data for spleen and lymph nodes.^16,23 However, our data show that naïve CD4⁺ T cells in the peritoneal cavity of WT mice have significant levels of cell-surface CD154, in contrast to naïve CD4⁺ T cells from lymph nodes or spleen (Fig. 3b and data not shown).^16,23 IL-4 injection caused slight decreases in total and/or cell-surface CD154, significant in WT but not in CD40 KO mice (Fig. 3b). We interpret these results in terms of IL-4 likely decreasing CD154 gene expression,⁵⁵ and this decrease impacting on total CD154 protein more strongly in the presence of CD40-dependent CD154 internalization and degradation than in its absence.²⁴

Fig. 3.

CD154 is available in the basal peritoneal cavity as a result of its expression by naïve CD4⁺ T cells. Cell-surface and total (cell-surface plus intracellular) CD154 was measured in different peritoneal cavity cell types of untreated WT mice (a), and in naïve, CM and EM CD4⁺ cells (b) of WT and CD40 KO mice either untreated or injected with IL-4c 48 h earlier. Data from different independent experiments are shown with different symbols. Data for cell-surface CD154 in (b) were log-transformed for statistical analysis. Representative histograms depicting CD154 expression in CD4⁺ and CD8⁺ T cells (a) and in the subtypes of CD4⁺ T cells in the absence of IL-4 (b) are shown in addition to the % positive cell data for individual mice. Asterisks without connecting lines indicate differences between the conditions that include IL-4 stimulation and the corresponding conditions without IL-4. The differences shown in (b) between WT and CD40 KO mice in cell-surface CD154 expression in CD4⁺ T cells were also significant in terms of geometric mean fluorescence intensity (P < 0.01 for CM CD4⁺ T cells in the absence of IL-4, and P < 0.001 for the remaining comparisons).

Three days after injection of thioglycolate medium into WT mice, cell-surface and total CD154 in bulk peritoneal cavity CD4⁺ T cells were detected at levels similar to those in the basal cavity (Supplementary Fig. S7).

Summarizing, CD154 is available in the peritoneal cavity, mostly as the result of its cell-surface expression by naïve CD4⁺ T cells, although in WT mice it goes mostly unnoticed due to its “consumption” through interactions with CD40. Even in the presence of systemically available IL-4, naïve CD4⁺ T cells are a ready source of this “potential” CD154.

3.4. Constitutive CD154 interacts with CD40 on peritoneal cavity macrophages

The CD40-dependent regulation of cell-surface CD154 in naïve CD4⁺ T cells was previously found to depend on both B and non-B cells when measured in spleen, and almost exclusively on non-B cells when measured in lymph nodes and blood.²³ Given these precedents and the high CD40 levels detected in LPM in comparison to peritoneal B cells (Supplementary Fig. S1E), macrophages are a candidate cell type to regulate cell-surface CD154 in the CD4⁺ T cells found in the peritoneal cavity. The existence of such regulation would imply that constitutive CD154 on CD4⁺ T cells does interact with macrophage CD40, thus making it possible that constitutive CD154 additionally acts as a tonic signal for macrophages. In cell culture, CD40 expressed by peritoneal cavity macrophages induced the down-regulation of cell-surface CD154 in bulk spleen CD4⁺ T cells (Fig. 4a and b), of which most are naïve cells.²³ Macrophage CD40 additionally caused total CD154 in CD4⁺ T cells to diminish (Fig. 4b), presumably because CD154 is internalized and undergoes lysosomal degradation.²⁴ The presence of IL-4 had no effect in this system, except for causing a slight increase in cell-surface CD154 (without change in total CD154) in the absence of CD40 (Fig. 4b).

Fig. 4.

CD40 expressed by peritoneal macrophages can control the expression of cell-surface CD154 by CD4⁺ T cells. Isolated peritoneal cavity macrophages from WT or CD40 KO mice were cultured for 3 h with spleen cells from CD40 KO mice as a source of CD4⁺ T cells expressing constitutive cell-surface CD154, in the absence or presence of IL-4, as summarized in (a). The levels of cell-surface and total CD154 were measured in the CD4⁺ cells from spleen before culture and after being cultured with WT or CD40 KO macrophages in the absence or presence of IL-4 (b). In addition, CD40 KO mice received CD4⁺ cells isolated from the spleens of CD40 KO mice i.p., and 24 h later received isolated WT or CD40 KO macrophages or no macrophages i.p. (c). Forty-eight hours after the last cell transfer, these mice were euthanized for the measurement of cell-surface CD154 in naïve, CM and EM CD4⁺ cell found in the peritoneal cavity (D). In (d), only data from the two experimental groups that received macrophages were taken into account for the statistical analysis. In (b) and (d), data from different independent experiments are shown with different symbols. Representative histograms depicting CD154 expression in CD4⁺ T cells for each condition (except conditions in the presence of IL-4) are shown in addition to the % positive cell data for individual mice. Asterisks without connecting lines indicate differences between conditions with IL-4 and the corresponding conditions without IL-4. The data for cell-surface CD154 expressed in terms of geometric mean fluorescence intensity showed significant differences between spleen WT and CD40 KO CD4⁺ T cells before culture (P < 0.05), between CD40 KO CD4⁺ T cells cultured with WT and CD40 KO macrophages (P < 0.001, both in the absence or presence of IL-4), and between mice transferred WT and CD40 KO macrophages in the case of naive CD4⁺ T cells only (P < 0.05).

In order to test if macrophages can regulate cell-surface CD154 in naïve CD4⁺ T cells in the peritoneal cavity of live mice, an obvious experiment would be to transfer purified WT or CD40 KO macrophages into the peritoneal cavity of CD40 KO mice. However, this approach was made difficult by the unexpectedly small number of naïve CD4⁺ T cells found in the peritoneal cavity of CD40 KO mice (Supplementary Fig. S8), since naïve CD4⁺ T cells allow the most sensitive detection of regulation of cell-surface CD154 (Fig. 3b). We therefore transferred WT or CD40 KO peritoneal cavity macrophages (or no macrophages) into the peritoneal cavity of CD40 KO mice that previously received purified spleen CD4⁺ T cells from CD40 KO mice (Fig. 4c). Forty-eight hours after macrophage transfer, naïve CD4⁺ cells present in the peritoneal cavity of mice that received WT macrophages expressed significantly less cell-surface CD154 than naïve CD4⁺ T cells in mice that received CD40 KO macrophages (Fig. 4d). Transfer of WT macrophages also impacted negatively on the fraction of EM CD4⁺ T cells that expressed cell-surface as well as total CD154. It is noteworthy that in our experimental setup only ∼ 3.5% of peritoneal cavity macrophages in the CD40 KO mice that received WT macrophages were CD40^+/+, and yet this was enough to bring about significant drops in cell-surface CD154.

Thus, CD40 expressed on peritoneal cavity macrophages interacts with, and negatively regulates, constitutive cell-surface CD154 on CD4⁺ T cells.

3.5. Constitutive CD154 has only a weak and indirect impact on the M(IL-4) polarization and no detectable effects on the proliferation of macrophages induced by IL-4

As mentioned, the outcome of CD40 signaling in macrophages can vary depending on agonist strength.³¹ The M(IL-4) polarization of peritoneal macrophages tended to be stronger in unfractionated cell cultures than in purified macrophage cultures (Fig. 1a and b), opening the possibility that polarization may be enhanced by a non-macrophage cell type, potentially naïve CD4⁺ T cells. Blocking CD154 with an antibody did not affect M(IL-4) polarization in unfractionated cell cultures (Supplementary Fig. S9A). The lack of effect of CD154 blocking in this system may merely reflect low levels of cell-surface CD154, as these were observed to drop over time in culture, even in the absence of CD40 and more markedly in the presence of IL-4 (Supplementary Fig. S9B). Blocking constitutive CD154 also did not significantly affect peritoneal macrophage responses to IL-4 in vivo (Supplementary Fig. S9C and D).

As the short-term antibody blocking strategy may have limitations, we compared the effects of IL-4 in WT vs CD40 KO mice. LPM of CD40 KO mice showed modestly enhanced responses to IL-4 in terms of RELM-α, and more weakly also in terms of Ym1 (Fig. 5a). In addition, the known basal expression of RELM-α by SPM⁵¹ (but not its IL-4-potentiated expression) was significantly higher in the absence of CD40 (Fig. 5a). In contrast, the proliferation responses to IL-4 of both LPM and SPM were normal in CD40 KO mice (Fig. 5b).

Fig. 5.

RELM-α expression induced by exogenous IL-4 in LPM is enhanced in CD40 KO mice, but this is not a cell-autonomous effect. WT or CD40 KO mice were injected with IL-4 (s.c.) and RELM-α and Ym1 (a) and proliferation (b) responses to this stimulus measured in LPM and SPM 48 h later. In addition, peritoneal cavity macrophages previously labelled in vivo with PKH26-PCL were isolated from WT or CD40 KO mice and transferred i.p. into WT recipients, and 1 wk later recipient mice were injected (s.c.) with IL-4c, as summarized in (c). Forty-eight hours later, RELM-α expression was measured in donor and recipient LPM (d). Data from different independent experiments are shown with different symbols. Asterisks without connecting lines indicate differences between conditions with IL-4 and the corresponding conditions without IL-4. The asterisk over the lines connecting paired data in (d) indicates the only difference between donor and recipient LPM within each individual mouse, identified by paired statistics.

We additionally tested the responses of isolated WT or CD40 KO peritoneal cavity macrophages transferred into the peritoneal cavity of WT mice (Fig. 5c). Transferred macrophages were pre-labelled with the PKH26-PCL dye, which in separate experiments was verified to minimally affect LPM or SPM responses to IL-4 (Supplementary Fig. S10A and B). However, the response of labelled transferred WT LPM to IL-4 was significantly weaker than that of the (also WT) recipient macrophages in terms of Ym1 expression and proliferation (Supplementary Fig. S11A and B). The same effects were present after transfer of unfractionated peritoneal cavity cells (data not shown), indicating that it is cell transfer and not purification that causes the problem. SPM also displayed differences in responses to IL-4 between WT donor and WT recipient cells, in this case also for RELM-α expression (Supplementary Fig. S11A and B). Hence, we could only obtain reliable data for the RELM-α response of LPM, and these did not show difference between CD40 KO donor and WT recipient cells within each mouse (Fig. 5d; compare paired red/pink and black/grey symbols). A difference observed between CD40 KO donor and WT donor LPM was mirrored by a similar difference between the LPM of the respective WT recipients (grey/black symbols in Fig. 5d), indicating that it does not reflect a cell-autonomous effect of CD40. Thus, the macrophage transfer experiments suggest that the negative effect of CD40 on LPM RELM-α expression previously observed in fully WT or CD40 KO animals is not cell-autonomous.

Finally, we tested if macrophage responses to IL-4 may be affected by CD40-dependent detection of constitutive CD154 in an optimized cell culture system. To counterbalance the fall in CD154 expression by CD4⁺ T cells in culture (Supplementary Fig. S9B), we used spleen cells from CD40 KO mice as a source of CD4⁺ cells initially expressing constitutive cell-surface CD154 at its maximum levels. These spleen cells were co-cultured with isolated unstimulated peritoneal cavity macrophages (as summarized in Fig. 4a, but using 48-h cultures). In this system, no differences in M(IL-4) polarization between WT and CD40 KO cells were observed (Supplementary Fig. S12), showing that constitutive CD154 is not capable of directly impacting on peritoneal macrophage M(IL-4) polarization. In addition, the availability of constitutive CD154 did not alter the fact that macrophages did not proliferate in vitro in response to IL-4 (data not shown).

In summary, in the absence of antigen-induced CD154 or exogenous CD40 agonists, CD40 expression has only weak and probably indirect impacts on peritoneal cavity macrophage M(IL-4) polarization, and no effect on their IL-4-driven proliferation.

4. Discussion

In the initial part of this work, we showed that exogenous CD40 agonists blunt macrophage responses to IL-4 (Fig. 1), consistent with a report in which M(IL-4) markers including RELM-α and Ym1 were measured in tumour-associated macrophages of mice treated with CD40 agonist antibody.⁵ Our results extend the conclusions to peritoneal cavity macrophages and provide evidence in isolated macrophages, which suggests that at least part of the inhibitory effects are due to direct signalling on macrophages. Our results also show that sCD40L can inhibit RELM-α expression independently of CD40 (Fig. 1c). The CD11b/CD18 integrin, an alternative CD154 receptor highly expressed on macrophages,¹ was previously observed to negatively regulate the expression of M(IL-4) marker arginase 1 in adipose tissue macrophages from obese mice.⁵⁶

Importantly, we show that CD40 agonists also inhibit IL-4-driven macrophage proliferation. It is notable that macrophages and B cells coexisting in a culture well or in the peritoneal cavity respond to CD40 agonist plus IL-4 in qualitatively different ways in terms of proliferation (Fig. 1d, e, g, and h). A related conclusion is that shortage of CD40 agonists is not the reason why macrophages in culture fail to proliferate in response to IL-4. One of the exogenous agonists that we used (sCD154) is to all practical purposes identical to CD154 that is proteolytically shed from cells and that is a marker of inflammation in vivo.^18,20,27,57 Thus, under inflammatory conditions that include a type 2 component, the effects of IL-4 and probably IL-13 on macrophages may be antagonized by shed CD154.

Exogenous CD40 agonists inhibited more strongly those responses to IL-4 known to require the PI3K/Akt/mTORC1 pathway (RELM-α, proliferation) than those that do not require this pathway (Ym1, CD206) (Fig. 1a, b, f, g and Supplementary Fig. S4E).^58–62 The mechanism underlying this contrast is not obvious, as CD40 agonists are thought to activate the PI3K/Akt pathway in myeloid cells.^63–66

CD154 provided by CD4⁺ T cells in cognate interaction with unstimulated peritoneal macrophages had a negative impact on the IL-4-induced up-regulation of RELM-α but not on Ym1 (Fig. 2e and f). The experimental system that we used to study the effects of CD154 provided by specific CD4⁺ T cells is not ideal: only a minor fraction of the TCR-transgenic CD4⁺ cells attained detectable levels of cell-surface CD154 in the presence of antigen, probably because the CD4⁺ T cells were not primed. In addition, the specific interaction in the culture well was inhibitory towards M(IL-4) markers independently of CD40. Even with these limitations, the gradation from a weak effect of CD40 expression on RELM-α to no effect on Ym1 was reminiscent of the effects of soluble CD40 agonists in vitro and in vivo. Hence the soluble CD40 agonists probably do imitate the effect of CD154 provided in antigen-specific contexts, and we envisage that both forms of CD154 antagonize those effects of IL-4 on macrophages that are PI3K/Akt-dependent. Our results also suggest that recruited macrophages are less sensitive than non-elicited ones (mostly LPM) to the mentioned effects of CD154 expressed by antigen-specific CD4⁺ T cells (compare Fig. 2e with Supplementary Fig. S6E). Thus, during ongoing adaptive immune responses, recruited macrophages may more easily acquire M(IL-4) phenotypes than LPM and possibly other resident macrophages.

We also show that CD40 expressed by peritoneal cavity macrophages can control cell-surface CD154 after its up-regulation triggered by TCR engagement (Fig. 2c). Previously human Langerhans’ cells and B cells had been shown to exert this control,^17,18,24 with no data available for mouse macrophages.

In the second part of our work, we explored constitutive endogenous CD154. In the peritoneal cavity under basal conditions, essentially only naïve CD4⁺ T cells were found to express cell-surface CD154, and this expression was strongly enhanced by the absence of CD40 (Fig. 3a). These results extend to a non-lymphoid site the observations made by Cyster's group for lymphoid organs.²³ In contrast to those previous observations and our own results for lymphoid organs (data not shown), naïve CD4⁺ T cells in the peritoneal cavity of WT mice have appreciable levels of cell-surface CD154 (Fig. 3b). This contrast may reflect a lower frequency of encounters with CD40-expressing cells in the peritoneal cavity in comparison to the lymphoid organs. As mentioned, non-B cells are important for the CD40-dependent control of constitutive cell-surface CD154 in lymphoid organs.²³ Put together with our observations that peritoneal cavity macrophages can regulate constitutive cell-surface CD154 (Fig. 4b and d), the above suggests that macrophages (possibly together with DC) may play a major part in regulating constitutive CD154 exposure both in lymphoid and non-lymphoid sites. Our in vivo results (Fig. 4d) additionally provide evidence of encounters between macrophages and naïve CD4⁺ cells in the peritoneal cavity under homeostatic conditions.

Although constitutive CD154 in the peritoneal cavity interacts with macrophages, our results indicate that it does not directly affect macrophage responses to IL-4 (Fig. 5d; Supplementary Figs. S9C, D, and S12). This lack of effect could be explained by total CD154 expression being somewhat lowered by IL-4 (Fig. 3B and Supplementary Fig. S9B). However, constitutive CD154 does not seem to impact on peritoneal macrophages in the absence of IL-4 either, as hinted by the absence of effects of CD154 blocking or CD40 gene deficiency in terms of basal MHCII expression in LPM (data not shown). The contrast with the effects of CD154 provided by antigen-specific CD4⁺ T cells probably reflects the potentiation of CD40 signalling afforded by the clustering of CD154 to the immunological synapse.^67,68 In addition, the mentioned contrast likely reflects different levels of expression of cell-surface CD154, which was clearly higher in antigen-stimulated CD4⁺ T cells than in bulk CD4⁺ T cells from CD40 KO mice (Supplementary Fig. S6B).

The reported effect of constitutive CD154 on the survival of autoreactive B cells is not cell-intrinsic,²³ and in our experiments we did not observe effects on B cells of CD40 or CD154 under basal conditions (data not shown). Thus, the cell type(s) that are target(s) of constitutive CD154 are probably different from both macrophages and B cells. Instead, macrophages (and possibly also B cells) may act as a buffer for this constitutive CD154 signal. The function, if any, of constitutive CD154 expressed by naïve CD4⁺ T cells is unknown. Our observation that naïve CD4⁺ T cells are essentially missing in the peritoneal cavity of CD40 KO mice (Supplementary Fig. S8) might provide a clue about this function, provided that the impact of CD40 deficiency reflects the lack of an interaction with CD154 expressed by these cells.

In conclusion, soluble CD40 agonists and antigen-induced CD154 blunt peritoneal cavity macrophage responses to IL-4. The inhibition is stronger for those effects of IL-4 that are known to require PI3K/Akt signalling (RELM-α, proliferation) than for Ym1 (which is PI3K/Akt independent) and are stronger for non-elicited than for recently recruited macrophages. In contrast to soluble CD40 agonists and antigen-induced CD154, constitutive CD154 expressed by naïve CD4⁺ T cells does not affect macrophage responses to IL-4, in spite of it being negatively regulated by (and thus interacting with) macrophage CD40.

Author contributions

M.S.-M.: Investigation, Writing—review, editing, and revision; I.G.-A.: Investigation, Data curation—formal analysis, Writing—review, editing, and revision; C.C.: Investigation, Writing—review, editing, and revision; L.B.: Contributing biological reagents, Writing—review, editing, and revision; P.H.P.: Investigation. Writing—review, editing, and revision; C.M.F.: Investigation, Conceptualization, Writing—review, editing, and revision; S.J.J.: Conceptualization, Writing—review, editing, and revision; J.E.A.: Conceptualization Project Administration/oversight, Writing—review, editing, and revision; A.D.: Conceptualization, Project Administration/oversight. Writing—original draft.

Supplementary material

Supplementary material is available at Journal of Leukocyte Biology online.

Funding

This work was funded by the International Centre for Genetic Engineering and Biotechnology (ICGEB) grant CRP/URY18-02 (to A.D.), Universidad de la República, Uruguay through CSIC I + D grants number 10 (year 2017) and 22520220100058UD (to A.D.), and by Wellcome Trust grants 106898/A/15/Z and MRC-UK MR/V011235/1 (to J.E.A.). In addition, it was supported by the Wolfson-Bob Sim Uruguay Initiative through travel grants to M.S.-M. and S.J.J., the Company of Biologists, ANII (Government of Uruguay) and PEDECIBA through travel grants to M.S.-M., by ANII through postgraduate scholarship to M.S.-M., and by Universidad de la República through CAP scholarships to I.G.-A. and M.S.-M.