|Home | About | Journals | Submit | Contact Us | Français|
Dendritic cell (DC) maturation and migration are events critical for the initiation of immune responses. After encountering pathogens, DCs upregulate the expression of costimulatory molecules and subsequently migrate to secondary lymphoid organs. Calcium (Ca2+) entry governs the functions of many hematopoietic cell types, but the role of Ca2+ entry in DC biology remains unclear. Here we report that the Ca2+-activated nonselective cation channel TRPM4 was expressed in and controlled the Ca2+ homeostasis of mouse DCs. The absence of TRPM4, which elicited Ca2+ overload, did not influence DC maturation but did considerably impair chemokine-dependent DC migration. Our results establish TRPM4-regulated Ca2+ homeostasis as crucial for DC mobility but not maturation and emphasize that DC maturation and migration are independently regulated.
Dendritic cells (DCs) are crucial in the initiation of adaptive immune responses. These ‘professional’ antigen-presenting cells, which are present in almost all peripheral tissues, capture and process antigens in the periphery before migrating to secondary lymphoid organs, where they prime naive T lymphocytes1. DC-mediated modulation of immune responses depends on the DC maturation status. During maturation, DCs undergo several phenotypic changes (such as up-regulation of the expression of costimulatory and major histocompatibility complex (MHC) class II molecules) that result in more potent antigen-presenting cell function. DC pattern-recognition receptors, including Toll-like receptors, are necessary for the sensing of pathogen-associated molecular patterns during microbial aggression. Hence, DCs act as ‘sentinels’, trafficking between tissues and lymph nodes to deliver signs of ‘danger’ to appropriate immune effector cells2. In the absence of foreign antigens, DCs remain in an immature state but are able to present self antigen and thus are involved in peripheral tolerance3. DCs are heterogenous and have been classified according to criteria such as their hematopoietic origin, surface marker expression, developmental stage, mobility and tissue localization4. For example, migratory DCs such as Langerhans cells and dermal DCs migrate from the peripheral tissues through the lymphatics to the draining lymphoid organs, whereas circulating DCs such as plasmacytoid DCs circulate in the bloodstream.
An increase in the cytosolic free calcium concentration ([Ca2+]i)is a ubiquitous signal that influences a wide range of hematopoietic cell functions, including cell cycle progression, adhesion, maturation and mobility5. Cytoplasmic increases in Ca2+ concentration induced by antigen receptor engagement modulate the initiation and maintenance of many cell functions6,7. After bacterial or mechanical stimulation, Ca2+ fluxes are induced in DCs and between adjacent DCs by networks of nanotubes8. DC migration correlates with an influx of Ca2+ through Ca2+ release-activated Ca2+ channels9,10, and the engagement of chemokine receptors such as CCR7 triggers DC trafficking by considerably increasing Ca2+ influx11.
Thus, Ca2+ homeostasis seems to be a key determinant of DC activation, but the molecular mechanisms regulating the Ca2+ homeostasis of DCs are not well understood. The Ca2+-activated nonselective (CAN) channel TRPM4 (transient receptor potential melastatin 4; A003857) has been detected in hematopoietic cells at both molecular and functional levels12. The TRPM4 and TRPM5 (A003767) channels are the only CAN channels described so far in nonexcitable cells13,14. In physiological conditions, TRPM4 channels allow massive entry of Na+, thereby inducing membrane depolarization and decreasing the driving force for Ca2+ entry15. TRPM4 can regulate the amount of Ca2+ entry and related physiological responses of T lymphocytes and mast cells15,16.
To examine the function of Ca2+ homeostasis in DC biology, we generated and analyzed a TRPM4-deficient (Trpm4−/−) mouse model. We found that TRPM4 was expressed in and controlled the Ca2+ homeostasis of DCs. TRPM4, but not TRPM5, acted as a Ca2+ ‘gatekeeper’, preventing Ca2+ overload in DCs. Ca2+-dependent stimulation of Trpm4−/− DCs promoted downregulation of the expression of phospholipase C-β2 (PLC-β2), which resulted in less subsequent Ca2+ mobilization in Trpm4−/− cells. Furthermore, the migration of Trpm4−/− DCs was profoundly impaired, but their maturation was not affected. Thus, our results show that the maturation and migration of DCs are regulated differently.
To assess the function of Ca2+ in DC biology, we generated mice in which exons 3–6 of Trpm4 were removed by Cre recombinase–mediated excision. Trpm4−/− mice lacked the targeted exons and TRPM4 protein (Supplementary Fig. 1 online) and were normal and fertile. The sizes of the thymus, spleen and lymph nodes of Trpm4−/− mice were similar to those of their Trpm4−/− littermates. Likewise, the numbers of CD4+ T cells, CD8+ T cells and B cells were similar in Trpm4−/− and Trpm4−/− mice (data not shown).
However, we noted that the spleen of Trpm4−/− mice consistently had about 50% fewer CD11b+CD11c+ DCs (Fig. 1a). The percentage and absolute number of plasmacytoid DCs in Trpm4−/− mice was normal (about 0.3%; data not shown). Further phenotypic analysis showed that this effect was limited to specifically the CD8α− subset; absolute numbers of CD8α+ DCs were not affected (Supplementary Fig. 2a online). However, the steady-state maturation status of Trpm4−/− and Trpm4+/+ DCs was identical (Fig. 1b) independently of CD8α expression (Supplementary Fig. 2b), as assessed by the expression of costimulatory and MHC class II molecules. We then assessed the maturation of DCs in inflammatory conditions in Trpm4−/− mice in vivo. After subcutaneous injection of Escherichia coli, the expression of CD86 and MHC class II was similar on DCs from Trpm4−/− and littermate control mice (Fig. 1c). However, the absolute number of mature DCs in draining lymph nodes was lower in Trpm4−/− mice (Fig. 1d). Thus, we noted fewer splenic DCs in steady-state conditions and fewer migratory DCs in inflammatory conditions in TRPM4-deficient mice. However, the absence of TRPM4 did not affect DC maturation.
Cell mobility requires an increase in intracellular Ca2+. As TRPM4 is involved in Ca2+ homeostasis, and as the number of mature DCs in inflamed lymph nodes was lower in Trpm4−/− mice, we further analyzed the effect of Trpm4 deletion on the homing of DCs to lymph nodes. After sensitization of skin with a proinflammatory agent, immature DCs begin to mature and migrate to the draining lymph nodes17. Thus, we painted the back skin of Trpm4−/− mice and littermate control mice with a 1% solution of fluorescein isothiocyanate (FITC) 2 h before inoculating the mice with E. coli by scarification. We isolated cells from brachial and mesenteric lymph nodes at 12, 24 and 48 h after scarification and counted FITC+CD11c+ cells in lymph nodes by flow cytometry. We noted many fewer FITC+CD11c+ cells in draining (brachial) lymph nodes of TRPM4-deficient mice at all time points studied (Fig. 2a). We found no such cells in nondraining (mesenteric) lymph nodes. However, the frequency of CD11c+ cells was similar in the skin of Trpm4−/− and Trpm4+/+ mice in steady-state conditions (Fig. 2b).
To confirm the defect in DC migration in Trpm4−/− mice, we assessed the trafficking of adoptively transferred bone marrow–derived DCs (BMDCs) from Trpm4−/− and Trpm4+/+ mice. We labeled immature BMDCs (iDCs) from Trpm4−/− and Trpm4+/+ mice differentially with the cytosolic dye CFSE (Trpm4−/−, CFSElo; Trpm4+/+, CFSEhi) and injected equal amounts of the cells together into footpads of recipient mice with the same genetic background (Fig. 2c). We induced DC migration by inoculating mice with E. coli at 2 h after BMDC transfer, then counted CFSEhi and CFSElo DCs in popliteal lymph nodes 12, 24 and 48 h later. We detected significantly fewer Trpm4−/− DCs in the popliteal lymph nodes at all time points (Fig. 2c). As expected, DC mobility was much lower in mice not inoculated with E. coli (Fig. 2c). These results collectively suggest that the impaired DC migration noted in TRPM4-deficient mice was due to a DC-intrinsic defect.
The regulation of Ca2+ is essential for DC function; however, only a few studies have examined ionic channels in DCs10. Furthermore, of 28 members of transient receptor potential channels discovered, none have been studied in DCs18. Thus, we used the patch-clamp technique to measure TRPM4 channel activity in BMDCs.
We first compared the electrophysiological profiles of Trpm4−/− and Trpm4+/+ BMDCs using the whole-cell configuration of the patch-clamp technique. We used an [Ca2+]i of 1 μM to promote the activation of CAN channels. We detected an outwardly rectifying current in wild-type but not in TRPM4-deficient BMDCs; in the latter, we detected only a residual inward-rectifying current (Fig. 3a). By subtracting the mean residual current-voltage relationships of Trpm4−/− DC traces from those of Trpm4+/+ DCs, we identified a TRPM4-like current (Fig. 3a), as reported before for human embryonic kidney (HEK293) clones over-expressing TRPM4 (ref. 12).
We further characterized the TRPM4 current by studying cell-free excised membrane patches in the ‘inside-out’ configuration. We obtained typical single-channel recordings at several membrane potentials in symmetrical ionic conditions19 (Fig. 3b, inset). The corresponding current-voltage curve showed a linear conductance of 22.1 ± 0.6 pS (n = 6 cells; Fig. 3b). Changing the internal solutions showed that the channel did not discriminate between Na+ and K+ and did not conduct Cl− (Fig. 3c). Channel activity was increased by membrane depolarization (Fig. 3d) and by an increase in [Ca2+]i, with a half-maximum effector concentration of 14 μM (Fig. 3e,f), but was decreased by the internal application of ATP (Fig. 3g). Thus, our electrophysiology studies established that TRPM4 was expressed and functional in Trpm4+/+ BMDCs (20 of 42 ‘inside-out’ patches) but not Trpm4−/− BMDCs (0 of 58 ‘inside-out’ patches).
We next analyzed Ca2+ influx in Trpm4−/− and Trpm4+/+ iDCs by activation of the cells with an E. coli supernatant (ECS), as reported before8. We noted higher Ca2+ influx in Trpm4−/− iDCs than in Trpm4+/+ iDCs (Fig. 4a). To determine whether this was due to sustained Ca2+ entry or to the release of Ca2+ from internal stores, we incubated iDCs in Ca2+-free medium, stimulated them with ECS and added a solution of 2 mM Ca2+ 90 s later. We found no difference in Ca2+ release from internal stores (Fig. 4b), which suggested that the higher [Ca2+]i in Trpm4−/− DCs was due to enhanced Ca2+ entry.
We next sought to determine what type of receptors were involved in the E. coli-mediated Ca2+ response of DCs. We tested whether G protein–coupled receptors, which are involved in the stimulation of DCs by Mycobacterium tuberculosis20, were involved in DC activation by E. coli. We incubated BMDCs overnight with pertussis toxin, an inhibitor of transduction through G protein–coupled receptors, then activated them with ECS. Ca2+-mediated signals induced by E. coli were abolished in the presence of pertussis toxin (Fig. 4c), which confirmed the interaction between G protein-coupled receptors and E. coli. Among G protein–coupled receptors expressed on iDCs, chemokine receptors such as CCR5 are involved in Ca2+ signals20 and cell mobility21. Therefore, we assessed whether CCR5 was triggered by E. coli in DCs. Indeed, treatment with TAK-779, a potent antagonist of CCR5, resulted in much lower ECS-mediated Ca2+ signals in iDCs (Fig. 4d). E. coli can also directly bind to receptors for the Fc fragment of immunoglobulin G type III (FcRγIII)22; however, E. coli-induced Ca2+ signals in iDCs were not mediated by Fc receptors dependent on immunoreceptor tyrosine-based activation motifs, as iDCs from FcRγ-deficient mice had Ca2+ responses similar to those of control iDCs (Supplementary Fig. 3a online). These findings collectively indicate that Trpm4−/− iDCs have a prolonged Ca2+ signal after E. coli triggering of G protein-coupled receptors, in particular chemokine receptors.
TRPM5 is closely related to TRPM4 in terms of electrophysiological properties, and we did detect TRPM5 transcripts by RT-PCR in DCs (Fig. 4e). However, we detected no other CAN current in Trpm4−/− DCs. Quantitative RT-PCR showed that TRPM4 mRNA was almost 200 times more abundant than TRPM5 mRNA in wild-type DCs, but that TRPM5 expression was three times higher in BMDCs from Trpm4−/− mice than in those from littermate control mice (Fig. 4f). We therefore generated double-knockout Trpm4−/−Trpm5−/− mice by mating our Trpm4−/− mice with Trpm5−/− mice23. After being stimulated with ECS, iDCs from Trpm5−/− mice had Ca2+ responses identical to those of iDCs from wild-type mice, whereas Trpm4−/− Trpm5−/− iDCs had a significantly higher Ca2+ flux than did Trpm5−/− iDCs (Fig. 4g). Trpm4−/−Trpm5−/− mice also had fewer splenic DCs than did Trpm5−/− mice (Fig. 4h). These results show that TRPM4 is the main Ca2+-activated nonselective channel involved in Ca2+ homeostasis in mouse DCs and that TRPM5 cannot compensate for the lack of TRPM4.
As Trpm4−/− DCs had more Ca2+ influx after stimulation, we examined whether DC differentiation and maturation were also affected. We assessed the percentage of BMDCs (CD11c+CD11b+) generated by the addition of granulocyte-macrophage colony-stimulating factor in vitro every 2 d from day 4 to day 13. We noted similar differentiation of Trpm4−/− and Trpm4+/+ BMDCs (Fig. 5a).
To determine whether the higher [Ca2+]i permitted by TRPM4 deficiency affected DC maturation, we assessed maturation induced by a Ca2+-independent stimulus (LPS) and a Ca2+-dependent stimulus (paraformaldehyde-fixed E. coli). As reported before24, activation with bacteria allowed Ca2+ influx, whereas LPS did not (Supplementary Fig. 3b,c). Each stimulus upregulated the expression of maturation markers (CD86, CD80, MHC class II and CCR7) to a similar extent in Trpm4−/− and Trpm4+/+ BMDCs (Fig. 5b,c), which suggested that the enhanced Ca2+ influx due to Trpm4 deletion did not affect DC maturation. We also tested the rate of E. coli phagocytosis and DC survival after incubation with fixed bacteria. Each was similar in BMDCs from Trpm4−/− and Trpm4+/+ mice (Supplementary Fig. 4a,b online). Therefore, maturation, differentiation and phagocytosis were not affected by Ca2+ overload in Trpm4−/− DCs.
Given the known function of Ca2+ in promoting cell mobility, the observed lower migration of Trpm4−/− DCs, which had large amounts of Ca2+, was unexpected. To further characterize the Trpm4−/− DC mobility defect, we made time-lapse videos of Trpm4−/− and control BMDCs after activation. We stimulated iDCs with ECS and measured their trajectories in two-dimensional x-y projections every 5 min for 4 h. Immature BMDCs from Trpm4−/− mice were more mobile than were those from Trpm4+/+ control mice (Fig. 6a).
However, because we had found that DC migration was impaired in Trpm4−/− mice in vivo, we then assessed oriented migration in response to the CCR7 ligand CCL21, which is involved in the trafficking of DCs to lymph nodes25,26. In Transwell chambers, we compared chemotaxis of DCs matured for 24 h with fixed E. coli or with LPS. The chemotaxis of Trpm4−/− mature DCs (mDCs) was significantly lower than that of Trpm4+/+ DCs when we used E. coli for cell maturation (Fig. 6b), but it was significantly higher than that of Trpm4+/+ DCs when we used LPS for maturation (Fig. 6c). This lower chemotaxis of E. coli–matured Trpm4−/− DCs was conserved even at higher concentrations of CCL21 (Supplementary Fig. 5 online). In contrast, mDCs from Trpm4+/+ mice showed similar chemotaxis after stimulation with either E. coli or LPS (Fig. 6b,c).
Because the in vitro experiments showed greater mobility of Trpm4−/− DCs after LPS maturation, we tested the effect of LPS on the migration of Trpm4−/− iDCs in vivo by adoptive transfer assay. We differentially labeled Trpm4−/− and Trpm4+/+ iDCs with CFSE as described above, then injected the cells into the footpads of mice 2 h before LPS injection (50 ng per footpad) and counted migratory CFSE+ DCs 24 h after LPS injection. We detected significantly more Trpm4−/− DCs in the popliteal lymph nodes (Fig. 6d). These results collectively indicate that the alteration in Ca2+ homeostasis induced deregulation of DC migration but not DC maturation.
Cell migration to lymph nodes involves chemokine receptors coupled to G proteins. One important component of chemokine receptor signaling pathways is PLC27. PLC produces inositol-1,4,5-triphosphate, which, by binding to its receptor on the endoplasmic reticulum, causes rapid release of Ca2+ from endoplasmic reticulum stores. As reported above, maturation of Trpm4−/− DCs with E. coli resulted in less CCL21-mediated chemotaxis. Similarly, we found that Ca2+ entry after ECS stimulation was lower in mDCs from Trpm4−/− than in those from Trpm4+/+ mice (Fig. 7a). We also noted this lower Ca2+ mobilization after CCL21 stimulation (Fig. 7b). The use of Ca2+-free medium showed that the defect in ECS-induced Ca2+ flux in Trpm4−/− mDCs was due to impaired release of Ca2+ from internal stores (Fig. 7c). Furthermore, the release of Ca2+ stores induced by thapsigargin was identical in Trpm4−/− and Trpm4+/+ DCs (data not shown). These findings suggest that Trpm4−/− mDCs have defects in signaling pathways ‘upstream’ of the release of Ca2+ stores.
To determine if initial augmented Ca2+ entry during the maturation process was responsible for the lower Ca2+ flux noted during subsequent cell stimulations, we incubated iDCs with E. coli and 2.1 mM EGTA to decrease the Ca2+ concentration in the cell culture medium from 2 mM to 1 μM. Viability and maturation were similar for Trpm4−/− and Trpm4+/+ DCs after overnight treatment with 2.1 mM EGTA (data not shown). However, EGTA treatment resulted in identical Ca2+ responses induced by subsequent CCL21 stimulation of Trpm4−/− and Trpm4+/+ mDCs (Fig. 7d). In addition, when we matured DCs with LPS in presence of 2.1 mM EGTA, Trpm4−/− DCs still showed subsequent higher Ca2+ influx than that of Trpm4+/+ DCs (Supplementary Fig. 6a online). As expected, by decreasing the external Ca2+ concentration to 1 μM during Transwell assays, we abolished the CCL21-induced chemotaxis of LPS-matured DCs (Supplementary Fig. 6b). These data collectively emphasize the function of the initial Ca2+ influx in the alteration of subsequent CCL21-mediated responses (Supplementary Fig. 7 online).
We next assessed which ‘upstream’ signaling molecule were involved in this regulatory process. As PLC-γ2 regulates the migration of B cells28 and PLC-γ1 is involved in the unresponsiveness of T cells29,30, we tested various PLC isoforms for their involvement in the Ca2+-mediated regulation of DCs. We noted reproducibly lower PLC-β2 expression in E. coli–matured Trpm4−/− BMDCs (Fig. 7e). The proteasome inhibitor epoxomicyn had no effect on the lower PLC-β2 expression in E. coli-matured Trpm4−/− BMDCs (Supplementary Fig. 8 online). Moreover, overnight incubation with ionomycin, which mimics the Ca2+ overload noted in Trpm4−/− DCs, triggered similarly lower PLC-β2 expression in Trpm4+/+ DCs (Fig. 7e). This result suggests that Ca2+ overload is sufficient to induce PLC-β2 downregulation in mDCs independently of proteasome degradation. Our data are consistent with the consequences of Ca2+ overload induced by ionomycin in T cells29, in which PLC-γ1 degradation leads to unresponsiveness of T cells. However, the intensity of the bands corresponding to PLC-γ1 and PLC-γ2 was similar in immature and mature DCs from Trpm4−/− and Trpm4+/+ mice (Fig. 7f). Therefore, sustained entry of Ca2+ into Trpm4−/− DCs resulted in lower PLC-β2 expression and thus less Ca2+ mobilization during subsequent stimulations.
Ca2+ homeostasis is tightly regulated in hematopoietic cells, and the importance of Ca2+ influx in cell activation is widely acknowledged. DCs are central to the immune system, but little is known about the molecular mechanisms that control Ca2+ entry in DCs. To address the role of Ca2+ in DC function, we generated an animal model in which Ca2+ homeostasis was impaired by genetic ablation of Trpm4 channels. Trpm4 is a CAN channel that regulates Ca2+ influx by depolarizing the membrane and decreasing the Ca2+ driving force through store-operated channels15. In this study, we have shown that stimulation of DCs lacking Trpm4 led to intracellular Ca2+ overload with detrimental consequences for DC migration but not maturation.
Although we noted impaired migration of Trpm4-deficient DCs to lymph nodes, DC maturation was not altered in Trpm4-deficient mice; thus, DC functions are differentially sensitive to [Ca2+]i variations. Indeed, the maturation induced by a Ca2+-independent stimulus (LPS) and a Ca2+-dependent stimulus (E. coli) was identical for Trpm4−/− and Trpm4+/+ DCs. This indicated that the process of maturation in DCs is not directly affected by sustained Ca2+ entry.
Unlike maturation, the migration of DCs was highly dependent on variations in [Ca2+]i. In absence of Trpm4, DCs matured in a Ca2+-independent way (by LPS) had more chemotaxis and homing to the draining lymph nodes. In contrast, Ca2+-dependent (bacterial) maturation of Trpm4−/− DCs led to intracellular Ca2+ overload and impaired migration. Thus, in inflammatory conditions and during DC activation with pathogens, Trpm4 is needed to regulate initial Ca2+ influx. This ‘fine tuning’ of Ca2+ entry allows DCs to respond adequately to subsequent stimuli (such as chemokines). Therefore, in contrast to the common paradigm that an increase of Ca2+ influx is associated with an increase in cell activity, our results lead us to postulate that Ca2+ uptake must be tightly regulated to avoid DC unresponsiveness and that TRPM4 is essential to prevent such detrimental overactivation.
The relationship between [Ca2+]i and cell mobility has also been studied in T lymphocytes. T cell immobilization correlates with an increase in intracellular Ca2+ that seems to induce a ‘stop’ signal31. T cell mobility has been shown to be decreased by sustained Ca2+ influx. In a nonselective environment, immature thymocytes move rapidly while maintaining low intracellular Ca2+ concentrations. Conversely, an increase in [Ca2+]i is necessary and sufficient to abolish thymocyte mobility32. The Ca2+ ionophore ionomycin has been used in T cells to assess molecular pathways affected by Ca2+ overload30. Such results show that in these cells, subsequent steps of stimulation are responsible for the degradation of PLC-γ1, which drives them into an unresponsive state29. In B lymphocytes, another PLC isoform, PLC-γ2, is involved in the regulation of cell migration and homing. B cells from PLC-γ2-deficient mice show impaired migration controlled by CXC chemokines28. We found that neither PLC-γ1 nor PLC-γ2 was affected by Ca2+ overload in DCs. The β-isoforms of PLC are expected to be essential for DC migration, as chemokine receptors that guide DCs to secondary lymphoid organs33,34 involve Gα protein family members that activate the β-isoforms of PLC enzymes6. We suggest that in inflammatory conditions, the downregulation PLC-β2 induced by Ca2+ overload abrogates the homing of Trpm4−/− DCs to secondary lymphoid organs. Thus, specific PLC isoforms seem to be important in Ca2+-mediated regulation in a cell type-specific way.
TRPM5, the channel the most closely related to TRPM4, is also a CAN cation channel35. On the basis of our electrophysiological data, we estimate that the number of TRPM4 channels expressed at the cell surface is around 1,000 molecules per DC. Quantitative PCR showed that TRPM4 expression was almost 200 times higher than TRPM5 expression; thus, we estimate that each DC expresses only 5 or 6 TRPM5 molecules. Furthermore, we detected no CAN current in Trpm4−/− DCs. In addition, TRPM5 is not inhibited by the internal application of ATP35, but we found that ATP suppressed DC CAN activity in ‘inside-out’ patch-clamp experiments. To further rule out the possibility of a contribution of TRPM5 channels to the Ca2+ homeostasis of DCs, we generated Trpm4−/−Trpm5−/− double-knockout mice. Phenotypic and physiological analysis of Trpm4−/−Trpm5−/− DCs showed that they were not different from Trpm4−/−Trpm5+/+ DCs. Therefore, our data suggest that TRPM4 is the only CAN channel involved in the control of DC Ca2+ homeostasis.
The factors governing the migration of DCs from skin to lymph nodes have been studied extensively, but because of the difficulty in tracking subsets of migrating cells and DC precursors in vivo, less is known about DC trafficking in steady-state conditions36,37. Studies have shown that in addition to bone marrow, spleen is a lymphoid organ where DC precursors can seed and differentiate38,39. Also, findings indicate that mice lacking the receptor tyrosine kinase Flt3 have impaired DC development40. This defect in peripheral DC homeostasis seems to be dependent on both progenitor migration and in situ conventional DC division40. Our data have shown that TRPM4-deficient mice had constitutively fewer splenic CD8α− DCs. These lower numbers of DCs could be the result of impaired migration and/or development of Trpm4−/− DC precursors. Further detailed studies must be done to distinguish between these two possibilities.
Matured monocyte-derived DCs are increasingly being used as cancer immunotherapy treatments3. As successful priming of tumor-specific T cells necessitates that such mature DCs generated ex vivo have the ability to migrate to lymphoid organs and thus to respond to specific chemokines, our data suggest that care should be taken to avoid Ca2+ overload during ex vivo maturation procedures41,42. In summary, our results have shown that TRPM4 prevents Ca2+ overload and thus permits the homing of DCs to draining lymph nodes. In addition, our findings emphasize that DC migration and maturation are independently regulated.
Trpm4−/− mice were viable and fertile. Pups were born at the expected mendelian ratio. Double-knockout mice were generated by intercrossing of Trpm4−/− and Trpm5−/− mice (provided by C.S. Zuker). Double-knockout mice were viable and fertile and had no noteworthy abnormalities. In all experiments, mice were 6–14 weeks of age and controls were littermates. Mice were housed in specific pathogen-free conditions and were handled in accordance with French and European directives.
Bone marrow cells were extracted from femurs and tibias. Cells were cultured at 37 °C and 5% CO2 at a density of 0.5 × 106 cells per ml in Iscove's modified Dulbecco's medium (Sigma-Aldrich) containing 10% (vol/vol) FCS (PAA Laboratories), 1 mM glutamine (Invitrogen), 10 μM β-mercaptoethanol (Sigma-Aldrich) and 20 ng/ml of granulocyte-macrophage colony-stimulating factor from supernatants of J558 cells (mouse plasma-cytoma cell line of BALB/c origin transfected with the gene encoding granulocyte-macrophage colony-stimulating factor). DCs were matured for 24 h by the addition of LPS (10 ng/ml; Sigma-Aldrich) or paraformaldehydefixed E. coli (K12 strain; 1 × 106 bacteria per ml) to the culture medium. Where indicated, cells were matured in presence of 2.1 mM EGTA to chelate free Ca2+ in medium from 1.92 mM to 1 μM. E. coli fixation was done on bacteria growing in logarithmic phase with 1% (vol/vol) paraformaldehyde.
Cells were immunostained with fluorochrome-conjugated monoclonal antibodies to mouse CD11c (HL-3), CD11b (M1/70), CD8α (53-6.7), I-Ab (AF6-120.1), CD80 (16-10A1) and CD86 (GL1; all from Becton Dickinson). Cells were stained for CCR7 with antibody to CCR7 (anti-CCR7; 4B12; Serotec); as described43. Data were acquired with a FACSCalibur (Becton Dickinson) and analyzed with FlowJo 8.3.3 software (TreeStar).
Mice were subcutaneously injected in the footpads with 1 × 108 E. coli in PBS. At 24 h after injection, popliteal lymph nodes were extracted and were treated with DNAse and collagenase. DCs were purified with CD11c microbeads according to the manufacturer's protocol (Miltenyi Biotec) and were analyzed by flow cytometry.
Mice anesthetized with xylasin and ketamin (Sigma-Aldrich) were shaved and depilated. Mice were then painted with 150 μl of a 1% (wt/vol) solution of FITC (isomer I; Sigma-Aldrich) in acetone and dibutylphtalate (1:1). At 2 h after skin painting, mice were scarified with E. coli. Cells from draining and nondraining lymph nodes were analyzed by flow cytometry 12, 24 or 48 h after scarification.
Immature BMDCs from TRPM4-deficient or control mice were loaded with different concentrations of CFSE (carboxyfluorescein diacetate succinimidyl diester; Sigma-Aldrich). Cells in suspension (1 × 108 cells per ml) were labeled by the addition of either 500 μM (CFSEhi) or 50 μM (CFSElo) CFSE in PBS with 0.1% (wt/vol) BSA. Trpm4−/− and Trpm4+/+ BMDCs were then mixed in equivalent proportions, and 0.5 × 106 cells were injected in the footpads of recipient mice 2 h before injection of E. coli (5 × 106 bacteria per footpad) or LPS (50 ng per footpad). Single-cell suspensions from draining and nondraining lymph nodes were incubated with anti-CD11c and were analyzed by flow cytometry.
The expression of TRPM4 (from intestine) and PLC (from iDCs and mDCs after incubation for 1 h at 37 °C atadensityof 3 × 106 cells per ml, as reported29) was analyzed with 200 μg or 25 μg of total protein, respectively (concentration determined by Bio-Rad protein assay). Proteins were separated by 8% or 10% SDS-PAGE and were transferred to polyvinylidene difluoride membranes (Amersham). Protein expression was assessed by immunoblot analysis with the primary antibodies anti-PLC-γ1 (2822; Cell signaling), anti-PLC-γ2 (sc-407; Santa Cruz Biotechnology), anti-PLC-β2 (sc-206; Santa Cruz Biotechnology), anti-β-actin (ACTN05 (C4); AbCam) and polyclonal anti-TRPM4 (produced with Proteogenix). Horseradish peroxidase-conjugated antibodies were anti-rabbit (NA934; Amersham) and anti-mouse (115-035-068; Jackson Immunoresearch Laboratories).
Cells were loaded and acquired as described15 (Supplementary Methods online). For flow cytometry, cells were loaded for 45 min with pluronic acid and a probe of the acetoxymethyl ester of the fluorescent Ca2+ indicator Fluo-4 (1 μM; Invitrogen) in Kreb's solution (145 mM NaCl, 5 mM KCl, 1 mM Na2HPO4, 5 mM glucose, 1 mM CaCl2, 0.5 mM MgCl2, 10 mM HEPES and 0.1% (wt/vol) BSA, pH 7.4). Individual fluorescence values were then analyzed with FlowJo and Origin software to normalize the fluorescence with the first value according to the equation (F/F0)–1, where ‘F’ is the fluorescence at specific time point and ‘F0’ is the fluorescence at time 0. The area under the curve was calculated for each cell. Where specified, cells were treated for 18 h with pertussis toxin (1 μg/ml; Sigma-Aldrich) or for 5 min with TAK-779 (500 nM; National Institutes of Health AIDS Reagent Program).
Chemotaxis was assessed in 8-μm Transwell chambers (Corning). Matured BMDCs (1 × 105 cells) were loaded into the upper chamber previously coated for 1 h with Iscove's modified Dulbecco's medium and 1% (wt/vol) BSA, then CCL21 (25 ng/ml; R&D systems) was added to lower chamber (in Iscove's modified Dulbecco's medium and 1% (wt/vol) BSA), followed by incubation for 2 h at 37 °C and 5% CO2. The chemotaxis index was calculated, with normalization of values to those obtained in conditions without CCL21.
Statistical significance was calculated with the one-way analysis of variance test. P values of less than 0.05 were considered significant.
We thank S.Y. Lin (ABgenomics) and B. Koller (University of North Carolina) for help with generation of Trpm4−/−mice; C.S. Zuker (University of California, San Diego) for Trpm5−/− mice; E. Ferrary for help with the patch-clamp setup; M. Benhamou and U. Blank for advice and critical reading of the manuscript; and J. Bex, A. Bouhalfaïa and E. Couchi for help in animal care. Supported by the Association pour la Recherche sur le Cancer (G.B.), Institut National de la Santé et de la Recherche Médicale (M.D. and P.L.), Ministère de l'Enseignement Supérieur et de la Recherche, Université Paris 7 (N.S.), Fondation pour la Recherche Médicale (T.L. and P.L.) and Action Concertée Incitative Jeunes Chercheurs (P.L.).
Note: Supplementary information is available on the Nature Immunology website.
AUTHOR CONTRIBUTIONS G.B. and I.C.M. were responsible for all experiments involving in vivo analysis of Trpm4−/− and Trpm5−/− mice; M.D. did the electrophysiological experiments in the whole-cell configuration; R.G. did the electrophysiological experiments in the ‘inside-out’ configuration; G.B. and M.D. did Ca2+ imaging; G.B. did the Transwell assays; M.D. did the time-lapse experiments; T.L. contributed to biochemical experiments; N.S. was responsible for quantitative RT-PCR; G.B. and N.S. were responsible for genotyping of mice; J.-P.K. was responsible for the generation of Trpm4−/− mice, critical reading and comments; R.C.M. and F.V. provided experimental guidance; all authors critically reviewed and contributed to the manuscript; and P.L. directed and supervised all aspects of the study and the writing and editing of the manuscript.
Reprints and permissions information is available online at http://npg.nature.com/reprintsandpermissions/
UCSD-Nature Signaling Gateway (http://www.signaling-gateway.org): A003857 and A003767.