Characterization of immune functions in TRAF4-deficient mice

Abstract

Tumour necrosis factor receptor associated factor 4 (TRAF4) is a member of the TRAF family of proteins which are cytoplasmic adaptor molecules strongly implicated in multiple immune functions. A previous investigation of TRAF4 biological functions by gene targeting in mice has shown a role for TRAF4 in embryonic development and neurulation in vivo. However, unlike other TRAF family members, the role of TRAF4 in the immune system is still unknown. To address this question, we performed an extensive characterization of the immune development and immune functions of TRAF4-deficient mice. Our analyses did not reveal any defects in development of T and B lymphocytes, granulocytes, macrophages and dendritic cells, and no defects in reactive oxygen species production and phagocytosis by neutrophils. Cellular and humoral responses against T-cell-dependent antigens were normal, as was dendritic cell maturation in response to microbial components and antigen uptake by dendritic cells. However, we demonstrated that dendritic cells from TRAF4-deficient mice exhibited reduced migration both in transwell experiments and in vivo. These results suggest that TRAF4 is not strictly required for immune development and functions but could participate in immune functions by facilitating immune cell migration.

Introduction

Tumour necrosis factor receptor-associated factor 4 (TRAF4), a member of the TRAF family of proteins, was originally cloned from metastatic breast cancer, where it was overexpressed.^1–3 Recently, we reported that TRAF4 is commonly overexpressed in human carcinomas.⁴ Six TRAFs have been described to date which exert a wide range of biological functions, such as in adaptive and innate immunity, embryonic development, stress responses and bone metabolism.^5–8 They serve as adaptive proteins and mediate signals from the tumour necrosis factor receptor (TNFR) and/or interleukin-1/Toll-like receptor (IL-1R/TLR) families. Reports suggest that TRAF4 could be recruited upon signalling triggered via the p75-NGFR (nerve growth factor receptor),⁹ LT-βR (lymphotoxin-β receptor),¹⁰ GITR (glucocorticoid-induced TNFR-related protein)¹¹ and TLR,¹² and in mitogen-activated protein kinase pathways.^13–15

Studies of TRAF mutants or with TRAF knockout mice clearly demonstrate that most TRAF proteins are strongly implicated in multiple immune functions. In TRAF1-deficient mice, an enhanced proliferation of T cells to T-cell receptor and TNFR stimulation is observed, suggesting that TRAF1 is a negative regulator of signalling in these cells.¹⁶ TRAF2 deficiency leads to splenomegaly, lymphadenopathy, enhanced TNF-induced cell death and increased B-cell proliferation in response to various signals.^17,18 In contrast, the CD40-mediated activation of B cells is impaired in TRAF5-deficient¹⁹ and TRAF3-deficient mice, defective T-dependent immune responses are observed and lymphoid organs are smaller.²⁰ Finally, a series of lines of evidence demonstrate that TRAF6 plays a key role in B-cell proliferation and dendritic cell (DC) maturation and is required for lymphoid organogenesis.^21–23

TRAF4 is a widely expressed protein, detected at a basal level in most tissues and organs as well as during embryonic development. It is also highly expressed in murine common lymphoid progenitors.²⁴ In most cells of the immune system, a basal expression of TRAF4 is detected after birth but it increases upon activation of T and B cells both at the messenger RNA^12,25–29 and protein levels (I. Cremer and C. H. Régnier, our unpublished observations). These findings together with the prominent immunological roles of all the other TRAF members prompted us to clarify the biological significance of TRAF4 in the immune system. For this purpose we performed an extensive characterization of the immune development and immune functions in TRAF4-deficient mice.⁷ We show that the absence of TRAF4 neither alters immune cell development nor in vitro B-cell and T-cell proliferative responses or maturation of dendritic cells. Furthermore, immune responses to T-cell-dependent antigens in vivo appear normal. However, we observed reduced DC migration in vitro and in vivo in the absence of TRAF4, suggesting its implication in immune cell migration.

Materials and methods

Mice

TRAF4-deficient mice (129/SvJ background), described previously,⁷ were bred in a pathogen-free facility at the Charles River Laboratories (L’Arbresle, France). Animals were used for experiments between 8 and 12 weeks of age. All animal experiments were performed in accordance with the approved institutional animal welfare guidelines.

Reagents

Lipopolysaccharide (LPS) from Escherichia coli (055:B5), lipoteichoic acid (LTA), poly (I:C), and phytohaemagglutinin A (PHA) were all from Sigma (La Verpilliere, France). CpG oligodeoxynucleotide 2084 (TCC TGA CGT TGA AGT) was purchased from MWG Biotech (Roissy, France) and recombinant TNF-α, IL-4, IL-2, CCL19 and CCL21 were from R&D Systems (Lille, France). Anti-CD40 (clone HM40-3), anti-CD3 (clone 145-2C11) and anti-CD28 (clone 37·51) were obtained from BD Biosciences (Le Pont-De-Claix, France) and F(ab′)₂ goat anti-mouse immunoglobulin M (IgM) was purchased from Jackson Immunoresearch (Suffolk, UK). The B-unit of Shiga toxin coupled to ovalbumin (STxB-OVA) was obtained by chemical coupling³⁰ and 2′,7′-dichloro-fluorescein diacetate (DCFDA), Alexa Fluor 488 E. coli and 5-(and 6)-carboxyfluorescein diacetate succinimidyl ester (CFSE) were supplied by Molecular Probes (Cergy Pontoise, France). Ovalbumin–fluorescein isothiocyanate (OVA-FITC) was from Sigma.

TRAF4 expression

Thymus, spleen, lymph nodes, T and B lymphocytes and DCs were analysed for TRAF4 expression by Western blotting. B cells were negatively sorted from spleen using the B-cell isolation kit (with a purity of 98% as judged by flow cytometric analysis of cells stained with anti-B220 antibody), and T cells were negatively sorted from lymph nodes using the T-cell isolation kit (Miltenyi Biotec, Paris, France) (with a purity of 99% as judged by flow cytometric analysis of cells stained with anti-CD3 antibody). DCs were generated from mouse bone marrow as described previously.³¹ Cell lysates were prepared in 50 mm Tris–HCl, pH 7·5, 150 mm NaCl, 2 mm ethylenediaminetetraacetic acid, 0·5% Triton X-100, 2 mm Na₃VO₄, 10 mm NaF, 10 mm sodium pyrophosphate, supplemented with a complete protease inhibitor cocktail (Boehringer, Paris, France). Proteins in cell lysates were quantified using the quick start Bradford protein assay kit (BioRad, Marnes-La-Coquette, France) to ensure that all the samples contained similar amounts of protein. Proteins (10 μg) were resolved in 4–12% gradient Tris–glycine polyacrylamide gels (Invitrogen, Cergy Pontoise, France) and were transferred to nitrocellulose membranes. Blots were probed with antibodies against TRAF4 and actin (Santa Cruz, Le Perray en Yvelines, France).

In vitro DC culture

Dendritic cells were generated from mouse bone marrow as described.³¹ Bone marrow was flushed from the femurs and the tibias, and red blood cells were lysed by incubation in lysis buffer containing 0·9 mm NH₄HCO₃ and 130 mm NH₄Cl for 1 min. Cells were plated at a density of 1 × 10⁶ cells and were cultured for 6 days in RPMI-1640 medium containing 10% fetal calf serum (FCS) and 100 units/ml of penicillin and streptomycin, supplemented with l-glutamax and 10 μg/ml granulocyte–macrophage colony-stimulating factor (GM-CSF; from J558L-conditioned medium). For the experiments, DCs (> 80% CD11c⁺), comprising immature DCs were used. Alternatively, DCs were positively sorted from spleen using the CD11c⁺ cell isolation kit (Miltenyi Biotec).

For phenotypic analysis of DC maturation, day 6 DCs were stimulated with 100 ng/ml LPS, 10 μg/ml poly (I:C), 1 μg/ml LTA, 10 μm CpG, 1 μg/ml TNF-α, or with medium alone for 48 hr, harvested and analysed with anti-CD11c and anti-CD86 antibodies.

Antigen uptake by DCs

Day 6 bone-marrow-derived DCs (4 × 10⁵) were incubated with 2 μg/ml OVA-FITC. After 10 min at 4° or 37°, the cells were washed and the phagocytosis of OVA-FITC was analysed using a FACScalibur cytometer (BD Biosciences). Flow cytometry data were analysed using cellquest software (BD Biosciences).

Bone marrow neutrophils: purification and functions

Neutrophils were purified from bone marrow as described.³² Cell purity, determined by fluorescence-activated cell sorter (FACS) staining with anti-Ly-6G antibody (BD Biosciences), was > 99%. For the phagocytosis assay, bone marrow neutrophils (10⁶ cells) were incubated with AlexaFluor 488 E. coli, at a multiplicity of infection of 5, for 30 min, at 37°. To determine reactive oxygen species production, bone marrow neutrophils were preincubated for 6 hr with DCFDA following the recommendations of the manufacturer. The cells were then treated with H₂O₂ for 30 min and analysed using the FACScalibur cytometer.

Immunization and enzyme-linked immunosorbent assay (ELISA)

STxB-OVA was obtained by chemically coupling the B subunit of Shiga toxin to ovalbumin as previously described.³⁰ Mice were immunized intraperitoneally with STxB-OVA (5 μg/ml) and αGalCer (α-galactoside ceramide) (2 μg/ml) and boosted with STxB-OVA alone by the same route at day 21. Seven days after the second immunization, CD8⁺ T cells were enriched from the spleen using the CD8 T-lymphocyte enrichment set (BD Biosciences), and their specificity was analysed by tetramer staining (see above). Serum samples were collected 28 days after the initial immunization. Levels of OVA-specific immunoglobulins IgG1, IgG2a and IgG2b were determined by ELISA as previously described)³⁰. As standard, IgG1, IgG2a, IgG2b mouse anti-OVA (Southern Biotechnologies, Montrouge, France) were used.

Immunofluorescence staining, cell sorting and flow cytometry

Single-cell suspensions prepared from thymus, spleen, bone marrow and lymph nodes were surface stained with monoclonal antibodies conjugated to biotin, FITC, phycoerythrin (PE), cychrome or allophycocyanin. The monoclonal antibodies used were anti-CD3 (145-2C11), anti-CD4 (RM4-5), anti-CD8α (53-6.7), anti-B220 (RA3-6B2) anti-IgD (11-26c.2a), anti-IgM (II/41), anti-CD43 (S7), anti-CD11b (M1/70), anti Gr-1 (RB6-8C5), purchased from BD Biosciences, and anti-CD11c (HL3) and anti-CD86 (GL1), purchased from eBioscience (Montrouge, France). For the detection of regulatory T cells, the monoclonal antibodies used were PE Texas Red-conjugated anti-CD4 (RM4-5; Caltag, Cergy Pontoise, France), PE-conjugated CD25 (3C7; BD Biosciences) and Alexa-647-conjugated Foxp3 (150D; Biolegend, Saint Quentin en Yvelines, France).

To detect OVA_257–264/K^b-specific CD8⁺ T cells, the cells were stained with OVA_257–264/K^b tetramer reagents according to the manufacturer’s recommendations (Beckman-Coulter Immunomics, Marseille, France). Briefly, cells were incubated with PE-labelled tetramer (for 60 min at 4°). After incubation and washes, labelled monoclonal antibodies were used to phenotype the tetramer-positive CD8⁺ T cells. An irrelevant tetramer recognizing a vesicular stomatitis virus-derived peptide in the context of K^b molecules (TTK^bVSV) was used. After staining with antibodies, fluorescent signals on the cell surfaces were analysed with a FACScalibur cytometer, and flow cytometry data were analysed using cellquest software (BD Biosciences).

Proliferation assays

For proliferation assays, B cells were negatively sorted from the spleen using the B-cell isolation kit, and T cells were negatively sorted from lymph nodes using the T-cell isolation kit (Miltenyi Biotec). Purified T lymphocytes from lymph nodes (99% pure as judged by flow cytometric analysis of cells stained with anti-CD3 antibody) were cultured in RPMI-1640 supplemented with 10% FCS and 50 mm 2-mercaptoethanol. Proliferation assays were performed by plating out 5 × 10⁴ T cells per well in triplicate for each condition in 96-well plates containing 200 μl medium. Cells were cultured for 48 hr in culture medium alone or in the presence of plate-bound anti-CD3 antibody (12 μg/ml) with or without anti-CD28 (10 μg/ml), with IL-2 (20 U/ml), with anti-CD3 antibody plus IL-2, or with PHA (10 ng/ml) with or without IL-2. After 48 hr, 1 μCi [methyl-³H]thymidine (TdR; Amersham, Buckinghamshire, UK) was added to each well, and [³H]TdR incorporation was quantified by scintillation counter 18 hr later.

Purified B lymphocytes from spleen (98% pure as judged by flow cytometric analysis of cells stained with anti-B220 antibody) were cultured in RPMI-1640 supplemented with 10% FCS, 50 μm 2-mercaptoethanol and 1 mm sodium pyruvate (Gibco, Cergy Pontoise, France). Proliferation assays were performed by plating out 10⁵ B cells per well in triplicate for each condition in 96-well plates containing 200 μl medium. Cells were stimulated with 10 μg/ml LPS, 10 μm CpG DNA, or with 10 μg/ml F(ab′)₂ goat anti-mouse IgM, or 10 μg/ml anti-CD40 antibody, with or without 10 ng/ml recombinant IL-4. After 24, 48 or 72 hr, 1 μCi [³H]TdR (Amersham) was added to each well, and [³H]TdR incorporation was quantified by scintillation counter 18 hr later (Molecular Devices, Saint-Gregoire, France).

Cytokine production

T lymphocytes were purified from lymph nodes, and were cultured for 48 hr in culture medium alone or in the presence of plate-bound anti-CD3 antibody (12 μg/ml) with anti-CD28 (10 μg/ml). The supernatants were then collected and kept at −80° until used. The cytokines IL-2, IL-4, IL-6, IL-10, IL-12, TNF-α and interferon-γ (IFN-γ) were detected in the supernatants using a cytometry bead array (BD Biosciences), and the IL-17 was detected by ELISA (R&D Systems).

Transwell migration assays

In vitro chemotaxis assays were performed using Costar Transwell inserts in 24-well plates. Cells were washed three times and resuspended in serum-free medium containing 1 mm HEPES. Mature bone-marrow-derived DCs or DCs isolated from spleen (1 × 10⁶) were put in the upper well in a volume of 100 μl, using 5 μm pore-size Transwell inserts. The lower well contained 600 μl serum-free Dulbecco’s modified Eagle’s medium supplemented with various concentrations of the CCL19 or CCL21 chemokines. The plates were incubated at 37° for 90 min before harvesting the migrated cells on the lower chamber. Harvested migrated cells were counted using light microscopy. The level expression of the chemokine receptor for CCL19 and CCL21 on mature DCs was carried out using the anti-CCR7 antibody (R&D Systems).

In vivo dendritic cell migration

Mature DCs were labelled with CFSE (Molecular Probes) as follows. Cells were washed twice in phosphate-buffered saline containing 5% bovine serum albumin (PBS-BSA 0·5%) then resuspended in PBS-BSA 0·1% at 10⁸ cells/ml. CFSE (10 μm) was added and left at 37° for 8 min, then extensively washed with PBS-BSA 0·5%. Cells were resuspended in PBS alone at 10⁸ cells/ml and injected into the footpad of wild-type mice (30 μl). Cervical and popliteal lymph nodes, and spleens were collected 48 hr later, treated with 1 mg/ml collagenase D (Roche, Roches-Roussillon, France) for 40 min at 37°, and mashed. The percentage of migrating cells was measured by flow cytometry using a Facscalibur (BD Biosciences).

Results

TRAF4 expression by organs and cells of the immune system

To confirm TRAF4 expression in organs and cells of the immune system in 129SvJ mice, we performed Western blot experiments on thymus, lymph nodes, spleen, purified T and B lymphocytes and DCs. In all the organs and cells of the immune system tested, a basal expression of TRAF4 was detected, with a higher expression in thymus and spleen, and in B cells and DCs (Fig. 1). These findings, together with the prominent immunological roles of all the other TRAF members, prompted us to clarify the biological significance of TRAF4 in the immune system.

Figure 1.

Western blot analysis of TRAF4. Total thymus, spleen, lymph nodes, T and B lymphocytes and dendritic cell lysates from wild-type mice were fractionated on a 4–12% denatured polyacrylamide gel, and the Western blot was probed with monoclonal antibody against TRAF4, or monoclonal antibody against actin as a control of protein charge.

Normal development of T and B lymphocytes, granulocytes, macrophages and DCs in the absence of TRAF4

Since TRAF4 expression can be detected in thymus, bone marrow and spleen, we first examined if haematopoietic cells develop normally in the absence of TRAF4. Mice that were TRAF4^−/− have thymus, spleen and lymph nodes of normal size, and their cell composition was comparable to those of wild-type littermates (data not shown). Detailed FACS analysis revealed no difference in the percentages of CD4⁺ CD8⁺, CD4⁻ CD8⁻, CD4⁺ CD8⁻ and CD4⁻CD8⁺ cells in the thymus between TRAF4^−/− and TRAF4^+/+ mice (Fig. 2a). Bone marrow from TRAF4^−/− mice had normal percentages of IgM⁺ B cells with normal expression of B220 and CD43 (Fig. 2b). The percentages of immature (IgM^+/− B220⁺) and mature B cells (IgM^hi B220⁺) were comparable between TRAF4^−/− mice and their wild-type littermates. Similarly, the percentages of pre-B and pro-B cells were similar. Taken together, these results indicated that TRAF4 expression was not necessary for T- and B-lymphocyte development.

Figure 2.

Lymphocytes, granulocytes, macrophages and dendritic cells develop normally in the absence of TRAF4. Flow cytometric analysis of cells in the thymus, lymph nodes, bone marrow, and spleen of 8-week-old wild-type (TRAF4^+/+) and TRAF4-deficient (TRAF4^−/−) mice. Thymocytes and cells from lymph nodes were stained for CD4 and CD8 (a). Bone marrow cells were stained for B220, immunoglobulin M (IgM), and CD43 (b). Splenocytes and cells from lymph nodes were stained for CD25 and Foxp3 expression (c). Splenocytes were stained for B220, IgD, and IgM (d), and for Gr-1, CD11b and CD11c (e). Numbers indicate the percentage of cells within each quadrant or region. Dot plots are representative of six mice of each genotype.

Lymph nodes and spleens from TRAF4^−/− mice had normal CD3, CD4 and CD8 T-cell surface markers (Fig. 2a and data not shown). The percentages of regulatory T cells (CD25⁺ Foxp3⁺) in the spleen and in lymph nodes were similar between TRAF4^−/− mice and wild-type littermates (Fig. 2c). In addition, type 1 transitional/marginal zone (MZ/T1, IgM^high/IgD^low), type 2 transitional (T2, IgM^high/IgD^high) and follicular (FO, IgM^low/IgD^high) B220⁺ B cells in the spleens were present in similar proportions in both TRAF4^−/− and TRAF4^+/+ mice (Fig. 2d). These results extend and confirm that normal B-cell development and differentiation occur in the absence of TRAF4. Moreover, analysis of Gr-1, CD11b and CD11c cell surface markers of spleen cells revealed no differences in the percentages of granulocytes (Gr-1+) and myeloid populations, including macrophages (CD11b⁺ CD11c^+/−) and dendritic cells (CD11c⁺), between TRAF4^−/− mice and their wild-type littermates (Fig. 2e). The percentages of DCs (CD11c⁺) in the lymph nodes and the percentages of Langerhans cells in the epidermis, determined by immunohistochemistry labelling of major histocompatibility complex class II-expressing cells in the ear epidermis of mice, were also similar between TRAF4^−/− mice and their wild-type littermates (data not shown). These results indicated that TRAF4 was not essential for myeloid cell and Langerhans cell development.

Since members of the TRAF family are recruited in the signal transduction pathways of TNFR, we investigated whether TRAF4 might regulate cell survival, as has been observed for TRAF2.¹⁷ The detection of spontaneous or TNF-α-induced apoptosis revealed no difference between TRAF4^−/− and TRAF4^+/+ thymocytes (data not shown).

Normal neutrophil functions in the absence of TRAF4

To directly assess the role of TRAF4 in acute reactive oxygen species production and in phagocytosis, we studied bone marrow neutrophils isolated from wild-type and TRAF4^−/− mice. DCFDA fluorescence was used to detect in situ reactive oxygen species generation. A 6-hr treatment by H₂O₂ increased DCFDA fluorescence intensity at the same level in TRAF4^−/− and wild-type neutrophils (Fig. 3a). Similarly, the uptake of AlexaFluor 488 E. coli was not affected by the absence of TRAF4 (Fig. 3b). These data indicate that neutrophil functions were normal in the absence of TRAF4.

Figure 3.

TRAF4 is not essential for polynuclear cell functions. Neutrophils were purified from bone marrow of wild-type (TRAF4^+/+) and TRAF4-deficient (TRAF4^−/−) mice. The cells were preincubated with DCFDA and stimulated with H₂O₂ for 6 hr, to determine reactive oxygen species activity (a) or were incubated for 30 min at 37°, with AlexaFluor 488 Escherichia coli (b). The cells were then analysed using a FACScalibur cytometer. Results are representative of two independent experiments, each performed on three mice of each genotype.

Normal B- and T-cell proliferation in TRAF4-deficient mice

To determine whether TRAF4 deficiency results in T-cell dysfunction, T cells from peripheral lymph nodes of TRAF4^−/− mice were tested for their ability to proliferate in response to anti-CD3-mediated antigen receptor stimulation. TRAF4-deficient T lymphocytes proliferated as well as their wild-type littermates (Fig. 4a) in response to anti-CD3. Similar results were obtained with costimulation by the anti-CD28 antibody or exogenously added IL-2, or in response to the polyclonal mitogen PHA.

Figure 4.

TRAF4 is not essential for B- and T-lymphocyte proliferation. (a) Purified lymph node T cells from TRAF4^+/+ and TRAF4^−/− mice were cultured for 48 hr in medium alone or in the presence of stimulatory anti-CD3, anti-CD3 plus anti-CD28 antibody, or phytohaemagglutinin (PHA). Stimulation was performed in the absence or presence of interleukin-2 (IL-2) as indicated. Proliferation was measured by [³H]thymidine incorporation uptake for 18 hr. Data-points represent the mean ± SD of triplicate wells. Results are representative of three independent experiments. (b) Purified spleen B cells from wild-type (TRAF4^+/+) and TRAF4^−/− mice were cultured in the presence of lipopolysaccharide (LPS), CpG DNA, anti-immunoglobulin M (IgM) antibody alone or with IL-4, or in the presence of anti-CD40 antibody alone or with IL-4. Proliferation was measured by [³H]thymidine incorporation uptake for 18 hr on the days indicated. Data-points represent the mean ± SD of triplicates wells.

To determine whether the stimulation of B cells through TLRs involves TRAF4, we measured TRAF4^−/− B-lymphocyte proliferation in response to the pathogen-associated molecular patterns LPS and CpG DNA, which activate TLR4 and TLR9, respectively. Resting B cells were purified from spleens of TRAF4^−/− and TRAF4^+/+ mice and cultured for 3 days in presence of LPS or CpG DNA. The results showed that the proliferative response of TRAF4^−/− B cells, as measured by [³H]thymidine incorporation, was comparable to that of B cells purified from wild-type littermates (Fig. 4b). Several TRAF members have been reported to be recruited in B cells after CD40 activation.³³ Furthermore, stimulation of human primary B cells after B-cell receptor or CD40 cross-linking induces TRAF4 expression (I. Cremer, C.H. Regnier unpublished data). We therefore examined the proliferative response of B cells from TRAF4^−/− mice to IgM or CD40 activation alone or in combination with IL-4. Figure 4(b) shows that B cells from TRAF4^−/− mice have normal proliferation to anti-IgM or anti-CD40 antibodies alone or with IL-4. Taken together, these results indicate that TRAF4 is not required for TLR-, IgM- or CD40-mediated activation of B cells.

Normal cytokine production by activated T cells in TRAF4-deficient mice

To determine whether TRAF4 deficiency results in a defect in cytokine secretion by T lymphocytes, T cells from peripheral lymph nodes of TRAF4^−/− mice were tested for their ability to produce IL-2, IL-4, IL-6, IL-10, IL-12, IL-17, TNF-α and IFN-γ cytokines after anti-CD3 and anti-CD28 stimulation (Table 1). The main cytokines produced by T cells are IL-2 and IFN-γ, showing a T helper type 1 (Th1) cytokine profile, which has been previously described in a 129SvJ genetic background. A production of IL-17 was also detected, indicating that T cells are able to differentiate into Th17 cells. Table 1 shows that the production of IL-2, IFN-γ and IL-17 by T cells purified from TRAF4^−/− mice was slightly lower than the production by T cells from wild-type mice. However, the significance of the differences observed was low (Table 1). The other cytokines tested were not detected, except TNF-α and IL-6, which are produced at a very low and comparable level between TRAF4^−/− mice and their wild-type littermates. Together, these data indicate that the potential of Th1, Th2 and Th17 T-cell differentiation is not affected in the absence of TRAF4.

Table 1.

Cytokine production by activated lymph node T cells

Cytokine (pg/ml)	IL-2*	IL-4	IL-6	IL-10	IL-12	IL-17**	TNF-α	IFN-γ***
TRAF4+/+	422 ± 15	nd	1·47	nd	nd	29 ± 8	88·7	1347 ± 54
TRAF4−/−	339 ± 30	nd	1·49	nd	nd	20·3 ± 5·4	87·6	895 ± 247

nd, not detected.

^*P =0·02,

^**P =0·19,

^***P = 0·06 (Student’s t-test).

IFN-γ, interferon-γ; IL-2, interleukin-2; TNF-α, tumour necrosis factor-α; TRAF4, tumour necrosis factor receptor-associated factor.

TRAF4 is dispensable for T-cell-dependent antigen response in vivo

To determine the role of TRAF4 protein in T-cell-dependent immune responses and assess B- and T-cell function in vivo, we immunized TRAF4^−/− mice and wild-type littermates with the T-cell-dependent antigen OVA linked to STxB by chemical coupling.³⁰ Vaccination of mice with STxB-OVA has been shown to prime a specific anti-OVA cytotoxic T-lymphocyte response.³⁴ TRAF4^−/− and TRAF4^+/+ mice were immunized intraperitoneally with STxB-OVA and αGalCer, and were boosted 21 days later with the same dose of STxB-OVA alone. The sera and spleens of immunized mice were collected on day 28. CD8⁺ T cells were enriched from spleens and anti-ovalbumin-specific CD8⁺ T cells were detected using OVA_257–264/K^b tetramer labelling. Figure 5 shows that labelling with the irrelevant tetramer TTK^bVSV, specific for VSV protein, revealed no anti-VSV-specific CD8⁺ T cells, whereas labelling with TTK^bSL8, specific for the peptide SL8 of OVA, allowed the detection of similar percentages of anti-OVA-specific CD8⁺ T cells between TRAF4^−/− and wild-type littermate mice (respectively 1·78% and 2·05%, Fig. 5a) with a mean of 1·67 ± 0·6% and 1·46 ± 0·62% for TRAF4^−/− and wild-type littermate mice, respectively. These results suggest that TRAF4 is not required for the induction of a specific anti-OVA CD8⁺ T-cell-dependent response in vivo.

Figure 5.

Normal response to the T-dependent antigen ovalbumin (OVA) in the absence of TRAF4. Wild-type and TRAF4^−/− littermate mice were intraperitoneally immunized with 1 nmol Shiga toxin B-subunit (STxB) -OVA at day 0 and boosted with the same dose at day 28. (a) Eight days after the second immunization, the number of OVA-specific CD8⁺ T cells in the spleen that was determined by staining with with OVA_257–264/K^b tetramer (TTKb SL8). As a control, cells were also stained with irrelevant tetramer (TTK^bVSV). The values represent the percentages of gated CD8 T cells, which stain positively with the tetramer. Representative data from four mice of each genotype analysed are shown. (b) Eight days after the second immunization, twofold serial dilutions of serum – 1 : 1000 initially for immunoglobulin G1 (IgG1) and 1 : 100 for IgG2a and IgG2b – were analysed by enzyme-linked immunosorbent assay for anti-OVA specific immunoglobulins. Data are the means of four mice of each genotype analysed.

Basal levels of IgM, IgG1, IgG2a, IgG2b, IgG3 and IgA in the serum of TRAF4^−/− mice were similar to those of wild-type littermates (data not shown). Figure 5(b) shows that TRAF4^−/− mice produced normal IgG1, IgG2a and IgG2b anti-OVA responses 28 days after immunization with STxB-OVA (Fig. 5b), suggesting intact T-cell help and intact immunoglobulin isotype switching in TRAF4^−/− B cells. These results suggest that antigen-specific antibody responses are not dependent on TRAF4 and they indicate that T-cell-dependent in vivo responses do not require TRAF4.

Having observed no defect in T-cell and B-cell function in TRAF4^−/− mice, we therefore addressed the question of whether antigen-presenting cell function was impaired or not.

Normal DC maturation and antigen uptake in the absence of TRAF4

To examine possible function of TRAF4 in DC maturation and antigen uptake, we first examined the antigen uptake by immature DCs (Fig. 6). The DCs were incubated with ovalbumin coupled to FITC at 4° or at 37°, and the antigen uptake was determined 10 min later by FACS analysis. Figure 6(a) shows that OVA-FITC was efficiently phagocytosed at 37°, and that the antigen uptake was as efficient in wild-type DCs as in TRAF4^−/− DCs (49% of phagocytosed OVA-FITC by TRAF4^+/+ DCs and 48% by TRAF4^−/− DCs). These results indicate that TRAF4 is not required for antigen capture by DCs. We then examined the phenotypic maturation of cultured DCs, obtained in vitro from bone marrow cells cultured in the presence of GM-CSF for 6 days. The DCs from TRAF4^−/− mice and from age-matched control littermates were stimulated with LPS (for TLR4), poly (I:C) (for TLR3), LTA (for TLR2), CpG DNA (for TLR9), or TNF-α for 48 hr, and surface expression of CD11c and B7.2 (CD86) was assessed by flow cytometry. No significant difference was observed in the surface expression nor in the percentages of B7.2-expressing cells between untreated immature CD11c-positive TRAF4^−/− and control DCs (Fig. 6b). Consistent with previous reports,^35,36 B7.2 was strongly upregulated on wild-type DCs in response to all TLR stimuli tested and TNF-α. Up-regulation of B7.2 was similar on TRAF4^−/− DCs compared to wild-type DCs in response to all the stimuli tested (Fig. 6b). These results indicate that optimal DC maturation in vitro in response to TLR ligands or to TNF-α was independent of TRAF4. Moreover, IL-6, TNF-α, IL-12 and IFN-β cytokine production was induced at the same levels in TLR-stimulated DCs from TRAF4^−/− and from TRAF4^+/+ mice (data not shown). These results indicate that microbial components stimulate cytokine production from DCs in a TRAF4-independent manner.

Figure 6.

Normal dendritic cell (DC) antigen uptake and maturation in the absence of TRAF4. DCs were generated from bone marrow cells of TRAF4^−/− mice or control littermates in vitro. (a) DCs were incubated with ovalbumin-fluorescein isothiocyanate (OVA-FITC), at 4° or at 37° and the antigen uptake was determined after 10 min by flow cytometry. (b) DC were left untreated (medium) or stimulated with lipopolysaccharide (LPS), poly (I:C), lipoteichoic acid (LTA), CpG DNA or tumour necrosis factor-α (TNF-α) as indicated for 48 hr, and CD86 levels were analysed by flow cytometry on CD11c-expressing cells. Results are representative of at least three independent experiments.

Reduced DC migration in vitro in the absence of TRAF4

A recent report suggested a role for TRAF4 in endothelial cell migration,³⁷ we thus tested the migratory capacities of TRAF4-deficient immune cells. We chose to study the migration of mature DCs because of the important migratory capacities of these cells. Mature DCs are known to migrate in response to the chemokine CCL19 or CCL21.^38,39 We thus used a CCL19- or CCL21-dependent Transwell migration assay to assess the chemotactic response of DCs. Addition of CCL19 induced a dose-dependent increase of cell migration, with a peak of migration at the dose of 100 ng/ml, both in TRAF4^−/− mice and in the wild-type littermates (Fig. 7a). However, in the absence of TRAF4, the total number of migrated DCs was significantly reduced, compared to wild-type littermate-derived DCs in the absence (34% reduced) or the presence (32% reduced) of 100 ng/ml CCL19 (Fig. 7a). Similar results were obtained in response to CCL21 (Fig. 7b). We next investigated whether the migration of DCs isolated from spleens of TRAF4^−/− mice was also reduced. CD11c-expressing DCs, purified from wild-type and TRAF4^−/− mice, were subjected to an in vitro migration assay in response to CCL21 (Fig. 7c). Again, the results show that the capacities of TRAF4-deficient DCs are reduced compared to their wild-type littermates. To determine whether the reduced DC migration was the result of reduced cell surface expression of the CCL19 and CCL21 receptor, we examined CCR7 expression by CD11c⁺ DCs. The results showed that similar percentages of dendritic cells expressed CCR7 in TRAF4^−/− (43·3%) and in wild-type mice (41·6%) (Fig. 7d). These results clearly demonstrate that the absence of TRAF4 reduces DC migration, even in the absence of chemokine, and that this phenomenon is not the result of a defect in chemokine receptor expression.

Figure 7.

Reduced in vitro migration of TRAF4^–/– dendritic cells. Mature bone marrow-derived wild-type and TRAF4^−/− dendritic cells (a, b) or mature spleen dendritic cells purified from wild-type and TRAF4^−/− mice (c) were assayed for their ability to migrate across transwell plates in the absence or presence of various concentrations of CCL19 (a), or in the presence of 100 ng/ml CCL21 (b, c). The data indicate the total number of migrated cells in duplicate wells ± SD in three independent experiments (three mice of each genotype/per experiment). Mann–Witney’s U-test was used for statistical analysis. The level of expression of chemokine receptors for CCL19 and CCL21 on mature bone marrow-derived wild-type and TRAF4^−/− dendritic cells was measured using the anti-CCR7 antibody (d).

In vivo reduced DC migration in the absence of TRAF4

Having observed a reduced migration of DCs in vitro, we analysed the in vivo migration of these cells. Mature bone-marrow-derived DCs from TRAF4^−/− or from wild-type mice were labelled with CFSE and injected into the footpads of wild-type recipient mice. Two days later, the percentages of cells that had migrated to the cervical lymph nodes, the draining popliteal lymph nodes and the spleen were determined (Fig. 8). The results showed that, as expected, no fluorescent cells were recovered from the cervical lymph nodes, which do not drain the footpad. However, CFSE-labelled cells were detected in the popliteal lymph nodes and in the spleen. Figure 8 shows that the percentages of CFSE-positive cells in the popliteal lymph nodes and in the spleen were fivefold and twofold higher, respectively, when wild-type DCs were injected, compared with TRAF4^−/− DCs. These results confirm that in the absence of TRAF4, the DCs have a reduced migration capacity.

Figure 8.

Reduced in vivo migration of TRAF4^−/− dendritic cells. Mature bone-marrow-derived wild-type and TRAF4^−/− dendritic cells were stained with CFSE, and were injected into the footpad of wild-type recipient mice. The cervical and popliteal lymph nodes and the spleen were recovered 48 hr later, and the percentages of CFSE positive cells were determined by flow cytometry. Dot plots represent the expression of CFSE relative to SSC. Results are representative of two independent experiments, each performed on three mice of each genotype.

Discussion

In this study, we demonstrate that DCs from TRAF4^−/− mice exhibit reduced migration in transwell experiments. In contrast, none of the other immune functions that we investigated were altered in the absence of TRAF4, and we could not observe any significant defect in development and organogenesis of the immune system in TRAF4-deficient mice.

Altogether these results suggest that TRAF4 does not play a major role in the immune system compared with other TRAF members, at least in the 129/SvJ genetic background. The absence of phenotype in TRAF4^−/− mice might also indicate that TRAF4 and other protein(s) have overlapping, redundant roles in the immune system. Since TRAF family members present strong sequence and structure homologies, we cannot exclude that another TRAF protein could have a redundant role with TRAF4, and could compensate the absence of TRAF4 in the immune system during mouse development. However, none of the five other TRAF proteins were overexpressed in the thymus and spleen of TRAF^−/− mice (C.H. Régnier, unpublished data).

Taken together, our observations show that TRAF4 is a particular member of the TRAF protein family because no alteration in the immune system was observed in TRAF4^−/− mice, in contrast to other TRAF^−/− mice. The role of TRAF4 seems to be predominantly exerted during neural development. In a previous study, we showed that TRAF4^−/− mice exhibit embryonic lethality, tracheal malformation, and defects in the development of the axial skeleton and closure of the neural tube during embryogenesis.⁷

Among the TRAF protein family, TRAF3^40–42 and TRAF6⁴³ have been shown to be adaptor molecules linking upstream TLR signals to downstream gene activation. TRAF3 is required for TLR7/9-dependent IFN-α production by Flt3-ligand-derived DCs,⁴² and is required for TLR-induced type I IFN and IL-10 production by DCs.⁴⁰ TRAF6 is crucial for MyD88-dependent mitogen-activated protein kinase and nuclear factor-κB activation.^44,45

A recent study suggested that TRAF4 could also be involved in the TLR signal transduction pathway, being a negative regulator in the TLR-signalling cascade. It has been observed that in HEK293 cells, TRAF4 forms a complex with TRAF6, TRIF (Toll–IL-1 receptor domain-containing adaptor-inducing IFN-β) and IRAK1 (IL-1 receptor-associated kinase) in addition to p47^phox, and counteracts the signal transduction mediated by TRAF6 and TRIF.¹² In the present work, the levels of TRAF6 and TRIF signals were not modified in the absence of TRAF4 (data not shown), and the maturation and cytokine production of DC induced by TLR stimulation were not significantly altered in the absence of TRAF4. In addition, in vivo experiments revealed that TRAF4^−/− mice are not hyperresponsive to LPS (data not shown).

A recent study suggests that TRAF4 is implicated in the migration of endothelial cells.³⁷ Cell migration is regulated by many signalling pathways, and it has been recently reported that ERK1/2 signalling can modulate the migration of cells by controlling the turnover of focal adhesions and the dynamics of actin polymerization.^46,47 NADPH oxidase was also shown to mediate endothelial cell migration in a Rac1-dependent manner.^48,49

Numerous studies have demonstrated that focal complex formation is critical to cell migration, and TRAF4 has been observed in focal complexes, and implicated in cell migration. In endothelial cells, endogenous TRAF4 and p47^phox,a subunit of the NADPH oxidase, target focal complexes in association with Hic-5, because a colocalization of endogenous Hic-5 and TRAF4 has been observed within vinculin aggregates, and initiate Rho GTPase signalling. Small interfering RNA-mediated knock-down of TRAF4 diminishes human umbilical vein endothelial cell migration across fibronectin-coated filters in response to a vascular endothelial growth factor gradient, demonstrating a role for TRAF4 in endothelial cell migration after association with p47^phox and formation of focal adhesion complexes.³⁷ There are no indications in the literature regarding the role of TRAF4 in the migration of immune cells, and particularly DCs, for which migration is an important process for initiating immune responses. Immature DC, localized in peripheral tissues, migrate to draining lymph nodes after pathogen recognition to present processed antigen to T lymphocytes. The directed migration of these cells through peripheral tissues is crucial for the activation of the immune response. We show herein that DCs from TRAF4^−/− mice exhibit reduced migration in transwell experiments, and reduced in vivo migration from the footpad to the draining lymph nodes. These results demonstrate for the first time a possible role of TRAF4 in immune cell migration. However, despite this reduced DC migration observed in vitro and in vivo, the numbers of resident DCs in lymph nodes and in spleen, and the numbers of Langerhans cells in the epidermis, are similar between TRAF4^−/− mice and their wild-type littermates. We have not observed an impaired immune response in vivo, that would be attributed to a deficiency in DC migration and in antigen presentation.

It seems that the mechanism for the reduced DC migration is probably the result of an intrinsic inability of the cells to migrate, because the reduced migration is observed even in the absence of chemokine. In leucocytes, the podosome dynamic structures are required for normal migration. Podosomes are composed of mainly the same structural and regulatory proteins as are seen in the more commonly found focal adhesion, but they are unique in their requirement for Wiskott–Aldrich syndrome protein (WASP).⁵⁰ It has been demonstrated that WASP is essential for podosome formation in DCs.⁵¹ In murine WASP-null DC cultures, the number of podosomes is reduced, and leucocyte migration in vitro and in vivo is defective.⁵² However, WASP-deficient mice are capable of mounting a normal immune response,⁵³ similar to TRAF4^−/− mice.

The lower migration observed in the absence of TRAF4 could be the result of a defect in actin polymerization and podosome formation, and further investigation into the molecular mechanisms involved, and particularly into the formation of podosomes, will help to address this issue.

Abbreviations

BSA

bovine serum albumin

CFSE

carboxyfluorescein diacetate succinimidyl ester

DC

dendritic cell

DCFDA

2′,7′-dichloro-fluorescein diacetate

ELISA

enzyme-linked immunosorbent assay

FCS

fetal calf serum

FACS

fluorescence-activated cell sorting

FITC

fluorescein isothiocyanate

GM-CSF

granulocyte–macrophage colony-stimulating factor

IgM

immunoglobulin M

IL-1

interleukin-1

LPS

lipopolysaccharide

LTA

lipoteichoic acid

OVA

ovalbumin

PBS

phosphate-buffered saline

PHA

phytohaemagglutinin

STxB

B-subunit of Shiga toxin

ST-xB-OVA

STxB chemically coupled to OVA

TLR

Toll-like receptor

TNF

tumour necrosis factor

TRAF

TNF receptor associated factor

Conflict of interest

The authors declare no competing financial interests.