The role of TLR4 in mature B cell activation is well characterized. However, little is known about TLR4 role in B cell development. Here, we analyzed the effects of TLR4 and TLR2 agonists on B cell development using an in vitro model of B cell maturation. Highly purified B220+IgM− B cell precursors from normal C57BL/6 mouse were cultured for 72 h, and B cell maturation in the presence of the TLR agonists was evaluated by expression of IgM, IgD, CD23, and AA4. The addition of LPS or lipid A resulted in a marked increase in the percentage of CD23+ B cells, while Pam3Cys had no effect alone, but inhibited the increase of CD23+ B cell population induced by lipid A or LPS. The TLR4-induced expression of CD23 is not accompanied by full activation of the lymphocyte, as suggested by the absence of activation Ag CD69. Experiments with TLR2-knockout mice confirmed that the inhibitory effects of Pam3Cys depend on the expression of TLR2. We studied the effects of TLR-agonists on early steps of B cell differentiation by analyzing IL-7 responsiveness and phenotype of early B cell precursors: we found that both lipid A and Pam3Cys impaired IL-7-dependent proliferation; however, while lipid A up-regulates B220 surface marker, consistent with a more mature phenotype of the IgM− precursors, Pam3Cys keeps the precursors on a more immature stage. Taken together, our results suggest that TLR4 signaling favors B lymphocyte maturation, whereas TLR2 arrests/retards that process, ascribing new roles for TLRs in B cell physiology.
In the adult mouse, B lymphopoiesis occurs in the bone marrow through the commitment and differentiation of hemopoietic stem cells (1, 2). B lineage cells are characterized by cell surface expression of CD19 and B220, and the earliest B cell precursors express several well-described surface markers such as AA4, c-kit, CD43, BP-1, and heat stable antigen (3, 4, 5, 6). At a later stage characterized as pre-B-II stage (6), or fraction D (4), c-kit and CD43 are down-modulated, while IL-2R is expressed on cell surface and maintained until the immature B lymphocyte stage (6). Along B cell development, V-D-J rearrangements lead to formation of the clonotypic BCR and surface expression of IgM molecules (7). Newly formed surface IgM+ (sIgM+) B lymphocytes are still immature cells bearing AA4, which is progressively lost as B lymphocytes mature (3), and must pass through transitional stages before the full maturation (8, 9, 10). During the late differentiation steps, transitional B lymphocytes acquire CD23 and IgD, which are considered as markers of B cell maturity (8, 9, 10, 11).
Mature B cells can be activated to proliferation and Ig secretion, in a BCR-independent manner, by microbial products such as LPS, lipoproteins, and CpG-enriched DNA (12, 13, 14). Recently, receptors for such microbial ligands and their molecular signaling pathways have been characterized. These receptors belong to a recently described family of germline-encoded and highly conserved molecules, the TLR family, which includes at least 10 members in the mouse species (15). TLRs are present in the cell types involved in innate immunity and are capable of recognizing several pathogen-associated molecular patterns, playing a key role in the activation of monocytes and granulocytes. TLR4 recognizes the LPS present in cell walls of Gram-negative bacteria (16, 17), while TLR2 recognizes bacterial lipoproteins and is important for triggering innate immune response against Gram-positive bacteria (18). The activation of mature B lymphocytes to proliferation and Ig secretion by LPS and bacterial CpG-DNA has been confirmed to be mediated by TLR4 and TLR9, respectively (17, 19). Triggering of murine B lymphocytes by bacterial cell wall lipoproteins is very likely mediated by TLR2 (12, 18).
The final differentiation of the maturing B cell is thought to be dependent on the positive selection of the lymphocyte upon engagement of the BCR with self-Ags, suggesting a role for cellular activation in B cell maturation (20, 21, 22). Although TLRs are well characterized as functionally triggering receptors for mature B lymphocytes, little is known about their expression and function in B cell development. TLR and BCR signaling can exhibit synergy in activation of mature B cell (23) and it is not known whether triggering through TLRs could act in B cell development as well. Here, using an in vitro model of B cell differentiation, we have studied the role of TLR4 and TLR2 agonists, lipid A/LPS, and Pam3Cys respectively, in B cell maturation. Our results show that signaling through TLR4 and TLR2 modulates B cell differentiation from early precursor stage. We found that, while TLR4 promotes maturation, TLR2 engagement arrests/retards B cell maturation. Simultaneous addition of TLR2 and TLR4 ligands revealed a previously unknown antagonism between those stimuli, suggesting an inhibitory cross-talk between the respective signaling pathways in maturing B lymphocytes. Experiments with TLR2-knockout (KO) mice showed that Pam3Cys effects depend on the expression of TLR2. This study reveals a new role for TLRs in B cell maturation, raising questions about their physiological role in B lymphopoiesis and repertoire selection in vivo.
Materials and Methods
Mice and cells
C57BL/6 and C57BL/10ScCr mice were from Fluminense Federal University and from the Oswaldo Cruz Foundation. TLR2-KO mice were obtained from Centro de Pesquisas René Rachou from breeding stock originally provided by Dr. Akira (Osaka University). Bone marrow (BM)34Cl lysis buffer and BM and spleen cells were washed with OptiMEM plus 10% FCS.
For depletion of IgM+ B lymphocytes, BM cells were incubated in MACS buffer (PBS containing 2 mM EDTA, 5% FCS) with anti-mouse IgM MicroBeads (Miltenyi Biotec) for 20 min on ice, washed, and resuspended in MACS buffer, according to specifications of the manufacturer. Cells were applied onto the magnetic column (Miltenyi Biotec) and the effluent was collected, and washed in OptiMEM plus 10% FCS. For purification of B220+IgM− B precursors, IgM+-depleted cell preparations were incubated for 20 min on ice with anti-B220 MicroBeads (Miltenyi Biotec) and processed as described above. Cells were applied to the magnetic column (Miltenyi Biotec), and after washing, retained cells were collected as the B220+ fraction. Viable cells were scored by trypan blue dye, using a Neubauer hemacytometer, and the purity of the cell preparations were verified by FACS analyses. IgM+ cells corresponded to <1% of B lineage cells after depletion, and B220+ cells corresponded to >99% of recovered cells after positive selection.
Cell staining and FACS analysis
Fresh BM cells, MACS-purified cells, and cultured cells were monitored by flow cytometry. The Abs used for staining were as follows: PE anti-CD23, PE anti-c-kit, FITC anti-B220, biotin anti-CD69 (BD Pharmingen), allophycocyanin goat anti-mouse IgM (Caltag Laboratories), biotin anti-IgD, FITC goat anti-mouse anti-IgM (Southern Biotechnology Associates), and biotin anti-AA4 (493 hybridoma provided by Dr. A. Rolink, Basel Institute, Basel, Switzerland; the mAb was purified and biotin-conjugated according to standard protocols). Biotinylated mAbs were revealed with Alexa-Fluor 680-R-PE streptavidin (Molecular Probes) or with allophycocyanin streptavidin (Caltag Laboratories). Cells were incubated with the mAbs in FACS buffer (PBS, 1% FCS, 0.05% sodium-azide) for 20 min on ice and washed twice with FACS buffer. When biotinylated mAbs were used, another step of incubation with Alexa 680-PE streptavidin or allophycocyanin streptavidin were performed under the same conditions as described above. Except for the four-color-stained cell samples, propidium iodide was added to the samples immediately before data acquisition at 0.5 μg/ml, for dead cell exclusion. Data were acquired on a FACSCalibur (BD Biosciences) and analyzed using CellQuest software (BD Biosciences).
Cultures for B cell differentiation
Purified B220+IgM− cells were cultured in 96-well flat-bottom plates (Corning) at 2 × 105 cells per well in OptiMEM plus 10% FCS, 5 × 10−52 at 37°C. Stimuli added to cultures included highly purified, lipoprotein-free LPS preparations (provided by Dr. D. Golenbock, University of Massachusetts Medical School, Worcester, MA and Dr. P. Tobias, Scripps Research Institute, La Jolla, CA) at 10 μg/ml, lipid A from Salmonella minnesota Re595 (Calbiochem) at 1 μg/ml, and Pam3Cys-Ser-(Lys)4 (EMC Microcollections) at 1 μg/ml. At 24, 48, or 72 h of culture, cells were harvested and counted; living cells were determined excluding dead cells with trypan blue dye and processed for FACS analyses. Statistical analysis of the data (one-sample analysis, Student’s t test) was performed using StatView software (Abacus Concepts); results were considered statistically significant if p < 0.05.
Cultures for early B cell precursors expansion
In experiments performed to study early stages of the B cell precursors, BM cells were cultured at 5 × 105 cells/ml in OptiMEM with the supplements described above plus rIL-7 for 5 days to obtain pro-B cell-enriched population. The source of rIL-7 was the culture supernatant of J558 cells transfected with IL-7 cDNA, added to the culture at a final concentration of 10%. Harvested cells were submitted to MACS for the depletion of IgM+ cells. The recovered B220+IgM− B cell precursors at the end of the fifth day represented ∼85% of the total cells after elimination of IgM+ cells. Cells were recultured in 96-well flat-bottom plates at 2 × 105 cells per well with rIL-7 for 48 h, without or with either Pam3Cys (1 μg/ml), lipid A (1 μg/ml), or both at 1 μg/ml. Living cells were counted excluding dead cells with trypan blue dye and processed for FACS analyses.
Spleen cells from C57BL/6 or TLR2-KO mice were cultured with different concentrations of Pam3Cys or lipid A at 2 × 105 cells per well in 96-well flat-bottom plates in 200 μl/well of OptiMEM supplemented with 10% FCS in a humidified atmosphere of 7% CO2 at 37°C. After 48 h, cultures were pulsed with 1 μCi of [3H]thymidine and incubated for further 24 h. Cells were harvested, and the incorporation of [3H]thymidine was measured by scintillation counting.
In vitro generation of immature and transitional B lymphocytes
In vitro experimental systems for studying B cell differentiation have long been established as useful models for dissecting B cell maturation process (4, 24, 25, 26). To study the effects of TLR ligands on B cell development, we used a system of B cell differentiation in vitro from highly purified B220+IgM− B cell precursors (>99% purity). This culture system allowed B220+IgM− B cell precursors to differentiate into newly formed sIgM+ B cells, as previously established (4, 26). The first IgM+ B lymphocytes arise at 24 h, and continue to accumulate during cell culture, representing 40–50% of B220+ cells after 72 h (Fig. 1⇓A). Beyond the 72-h culture period, the absolute number of recovered IgM+ B lymphocytes decreased while cell mortality increased (data not shown). The percentage of IgM+CD23+ cells, which represent a more mature developmental stage of the B lymphocyte, augments during the culture, up to 10% of the total IgM+ B cells at 72 h (Fig. 1⇓A).
In adult BM, three major subsets of B lymphocytes have been characterized according to B220 and IgM expression patterns (8): the B220lowIgMlow immature B cells, which correspond to the earliest stage of the IgM+ B lymphocyte, the B220dull/highIgMhigh transitional B cells, a more advanced stage of B cell maturation, and the B220highIgM+ follicular recirculating mature B lymphocytes. Fig. 1⇑B shows the IgM/B220 flow cytometry plot representative of fresh BM cells from young mice (8–12 wk of age), indicating the gates we used to define the three different B cell subsets: I, immature; T, transitional; and M, mature. The gates were defined according to the original description of Carsetti et al. (8). For each of these subsets we studied in more detail the phenotype of the B cell generated in vivo and in vitro. We analyzed the up-regulation of maturation markers acquired as the B cell differentiates, CD23 and IgD, and the down-modulation of AA4, an immature B cell marker that is progressively lost during differentiation, from the transitional B cell population to the mature stage. Down-regulation of AA4 is considered as the hallmark discriminating immature and transitional from mature B cell (3, 9). As can be seen in Fig. 1⇑B, in fresh BM cells, while immature B cells are mostly CD23−IgD−, a significant part of the transitional B cells are already CD23+ and IgD+ and mature B cells, instead, are all AA4−CD23+IgD+. AA4− cells, which represented ∼10 and 30% of immature and transitional subsets, respectively, were probably mature B cells that could not be completely separated from these populations with the gating used in the B220/IgM plot. If AA4− cells are excluded from immature and transitional gates, then the percentage of CD23+ cells drops to ∼10% in transitional B subset, and 0% in immature B.
B lymphocytes generated in vitro bear essentially a typical immature/transitional phenotype (Fig. 1⇑C), mostly expressing AA4. Immature and transitional B cell subsets obtained in culture express levels and percentages of CD23 and IgD similar to the respective B cell subsets generated in vivo. In vitro-generated transitional B cells display higher levels of AA4+ compared with their in vivo equivalent, but the level of expression of this Ag is diminished when compared with immature B cells, suggesting that this population obtained in vitro indeed represents a more advanced stage of B cell maturation. Overall, the in vitro maturation culture generated immature and transitional B lymphocytes with a phenotype very similar to the lymphocytes differentiated in vivo. A very small percentage of phenotypically mature B cells could be generated in vitro, confirming that cell culture systems of differentiation do not generate mature B lymphocytes (27).
TLR4 and TLR2 ligands have distinct effects on B lymphocyte differentiation
The addition of TLR4 agonists, lipid A, or highly purified LPS, to the cultures of B220+IgM− B cell precursors promoted an increase in the absolute number and percentage of IgM+CD23+ B lymphocytes, resulting in a 2- to 4-fold increase in the percentage of CD23+ B lymphocytes (Fig. 2⇓A and Table I⇓). The percentage of IgM+CD23+ cells augments during the 72-h culture period, which could suggest the proliferation of a small number of IgM+CD23+ cells. However, this seems unlikely because the addition of LPS in the last 24 h of culture had the same result as the addition of LPS in the beginning of the culture (Table II⇓). Indeed, the effect of LPS is already observed even if the addition occurred 12 h before the end of the culture (data not shown). Absolute cell counts are also shown in Tables I⇓ and II⇓, allowing the conclusion that the percent increase of IgM+CD23+ B cells cannot be accounted by cell death or proliferation, and represents the transition of a subset of B cells in culture, from CD23− to CD23+. It is interesting to note that only IgMhigh B cells express CD23, both in control and LPS-treated cultures (Fig. 2⇓B). The effect of LPS on B cell maturation was not observed in cultures with B cell precursors obtained from C57BL/10ScCr mice, a natural null-mutant for TLR4 (data not shown).
Expression of CD23 induced by LPS could depend on the activation of the B lymphocyte. We have addressed this point comparing the expression of CD69 and CD23 following the addition of LPS to cultures of mature splenic B cells or B cells generated in vitro (Fig. 3⇓A). In the presence of LPS, B cells generated in vitro showed a limited up-regulation of CD69 Ag in 24 h of culture and there was not a significant increase of the marker even at 72 h, in contrast to the massive and sustained expression of CD69 in mature B lymphocytes (Fig. 3⇓A). Moreover, four-color staining for B220, IgM, CD69, and CD23 revealed that the limited expression of CD69 following the addition of LPS bears no correlation with the expression of CD23 (Fig. 3⇓B).
We next investigated if the agonist specific for the TLR2, Pam3Cys, would have similar effects as TLR4 agonists in B cell maturation, inasmuch as those receptors share most of their signaling transduction pathways (15). However, the addition of Pam3Cys to the cultures of B220+IgM− B cell precursors had no effects on the percentage of CD23+ B cells (Fig. 4⇓A). Interestingly, when Pam3Cys and lipid A were simultaneously added to the culture, the increase in the percentage of CD23+ cells promoted by LPS/lipid A was completely abolished by Pam3Cys (Fig. 4⇓A). Pam3Cys alone inhibited the expected down-modulation of AA4 expression on transitional B cells (Fig. 4⇓B), when compared with control. In contrast, the presence of LPS/lipid A did not interfere in the down-modulation of AA4 expression (Fig. 4⇓B). The expression of IgD was also inhibited in the presence of Pam3Cys alone (Fig. 4⇓C). Altogether, these results suggest that, besides inhibiting TLR4 induction of CD23, engagement of TLR2 alone can interfere and arrest the differentiation in a late step of B cell maturation.
The effect of Pam3Cys is abrogated in TLR2-KO
Pam3Cys has been characterized as a specific TLR2 ligand; however, to rule out the possibility that Pam3Cys would be acting through an alternative receptor, we also studied the effects of Pam3Cys in the TLR2-KO mice to confirm its dependence on TLR2. As shown in Fig. 5⇓A, splenocytes from normal C57BL/6 mice responded with intense proliferation in response both to Pam3Cys and lipid A. By contrast, TLR2-KO mouse splenocytes did not proliferate with Pam3Cys, while the response to lipid A was not affected, showing that the proliferative response of mature B lymphocytes to Pam3Cys is TLR2-dependent. We then tested the effects of TLR2 and TLR4 ligands in cultures of purified B220+IgM− cells from TLR2-KO mice (Fig. 5⇓B). We verified that lipid A and highly purified lipoprotein-free LPS, induced a 2-fold increase in the percentage of IgM+CD23+ transitional B cells relative to control. Pam3Cys addition alone did not show any effect. When Pam3Cys was added simultaneously with LPS/lipid A no inhibition of their activity was observed, demonstrating that the effect of Pam3Cys is directly dependent upon the activation via TLR2.
TLR2 and TLR4 ligand effects on early B cell precursors
The experiments described above characterize the effects of TLR4 and TLR2 in late steps of B cell maturation. We then investigated the activity of lipid A and synthetic Pam3Cys on early B cell precursors. For that purpose, B cell precursors were assayed for the effects of Pam3Cys and lipid A on the expression of B220, c-kit, IgM, and IL-7 responsiveness (Fig. 6⇓). BM cells were cultured for 5 days with IL-7 to obtain a B cell precursor population highly enriched for IL-7 responsiveness, and then recultured in the presence of IL-7 without or with Pam3Cys and/or lipid A for further 48 h. While in cultures with IL-7 alone, the number of B cell precursors recovered was 2-fold the input number, in cultures with Pam3Cys, lipid A or both, no changes in the B cell precursor numbers occurred, showing that the presence of the TLR ligands impaired the proliferative expansion of B cell precursors in response to IL-7 (Fig. 6⇓A). The addition of lipid A resulted in higher expression of B220 on B cell precursors when compared with control and Pam3Cys-treated cultures (Fig. 6⇓B). Interestingly, we found that upon the addition of both lipid A and Pam3Cys, the effect of lipid A prevailed over the Pam3Cys (Fig. 6⇓B), in contrast to the results found in the study of CD23+ population, where TLR2 predominated, suggesting a developmental control on dominance of the TLR4/TLR2 signaling.
There is little information about the direct effects of LPS, as well as of other TLR agonists, on B cell development. Indeed, most studies investigate the response to LPS of the maturing B lymphocyte without addressing the effect of LPS in the maturation process itself (3, 9, 24, 25). In vitro studies with transformed 70Z/3 pre-B cell line showed that LPS promotes expression of IgM on cell surface, indicating that LPS can provide stimulus for differentiation (28). To study direct effects of TLR2 and TLR4 agonists in B cell differentiation, we used an in vitro culture system of B cell maturation from normal, highly purified B220+IgM− cells. In vitro culture of B220+IgM− cells allows for the differentiation of a significant number of sIgM+ B lymphocytes that make up near 50% of B220+ cells after 72 h (Fig. 1⇑A and Table I⇑); the percent enrichment for B cells is true differentiation and cannot be attributed to selective loss of viability of B cell precursors (Tables I⇑ and II⇑). Comparison of the phenotype of in vitro with in vivo maturation supports the conclusion that differentiation in vitro generates mostly immature and transitional B lymphocytes, with a phenotype very similar to their homologous populations generated in vivo (Fig. 1⇑, B and C). In line with previously described results, few mature B lymphocytes could be generated in vitro (26, 27). Our culture systems consisted of purified B220+IgM− cells, and support B cell maturation until the T1 (IgMhighAA4+CD23−) and T2 (IgMhighAA4+CD23+) transitional stages, while T3 (IgMlowAA4+CD23+) is poorly represented (9).
Addition of TLR4 agonists—lipoprotein-free LPS or lipid A—to purified B220+IgM− cell cultures promoted a 2- to 4-fold increase in the generation of CD23+ B cells (Fig. 2⇑ and Table I⇑). Again, the percent enrichment for CD23+ B cells is true differentiation and cannot be attributed to selective loss of viability of CD23− cells or proliferation of CD23+ cells (Tables I⇑ and II⇑). This is particularly evident in cultures where LPS was added in the last 24 h, resulting in the same increase in IgM+CD23+ population as for cells cultured in the continuous presence of LPS for 72 h (Table II⇑). The percent increase of CD23+ B cells induced by LPS is dependent on the expression of TLR4, as indicated by results with cultures of B cell precursors from C57BL10.ScCr mice, a natural TLR4-deficient mutant (data not shown). CD23 is a low-affinity receptor for Fcε (FcεRII) highly expressed on mature follicular B lymphocytes (29). It is considered as a maturation marker due to the close correlation of arising of this marker on B lymphocyte surface with progression in maturation process, as has been characterized for the transitional B cell subsets (9). The data obtained with cultures of purified B220+IgM− cells in the presence of lipid A or LPS showed that TLR4 can stimulate B cell maturation as suggested by the increase in the more mature CD23+ B cell population. It is interesting to note that in control cultures, the expression of CD23 only begins in B lymphocytes with transitional phenotype B220dull/highIgMhigh (Fig. 2⇑B), which have down-modulated the expression of AA4 relative to immature B lymphocytes, as expected accordingly with the in vivo maturation process (Fig. 1⇑C). We found that the CD23+ B lymphocytes generated in the presence of TLR4 agonists also arise at transitional stage expressing low levels of AA4, just like the control cultures, thus suggesting that CD23 expression induced by TLR4 engagement obeys the differentiation program, and cannot be expressed out of its timing in maturation (Figs. 2⇑B and 4⇑B). Little is known about the stimuli necessary to promote the expression of CD23 in maturing B cells. A recent study has shown that incubation of immature B cells with BAFF, characterized as an important factor for survival and maturation of transitional B lymphocytes (30, 31), is able to induce expression of CD23 (32). Our results indicate that TLR4 signaling can also provide stimuli for directing the expression of CD23 on late steps of B lymphocyte maturation.
The final maturation stages of B lymphocyte, in which immature B cells progress along the transitional stages, are not associated with cell proliferation (9, 10). However, evidences for negative or positive selection of B cell clones depending on the engagement of the clonotypic BCR (20, 21, 22) suggest that intracellular signaling linked to B cell activation by surface Ig receptor is directing these maturation steps. LPS is a potent mitogen for mature B cells, promoting extensive proliferation and differentiation of the lymphocytes into Ig-secreting plasma cells. It is interesting to speculate whether the action of LPS in B cell differentiation in vitro would depend on the full activation of the B lymphocyte, or would follow a pattern of partial activation suggested by the positive/negative selection through the BCR. We have investigated this point analyzing the expression of the very early activation Ag CD69. CD69 is readily and massively expressed by B and T cells a few hours following activation through the clonotypic receptors or mitogens (33). CD69 is expressed in conditions of suboptimal stimulation, being a very sensitive indicator of lymphocyte activation. Strikingly, the addition of LPS to B cells generated from purified B220+IgM− cells in vitro resulted in little expression of CD69 (Fig. 3⇑A). In contrast, mature splenic B cells show a massive and sustained expression of CD69 upon stimulation with LPS (Fig. 3⇑A). Moreover, there is no correlation between the expression of CD23 and CD69, as assessed by four-color flow cytometry analysis (Fig. 3⇑B). These results are in agreement with the interpretation that stimuli inducing the expression of CD23 upon triggering of TLR4 do not promote a full cell activation, analogous to BCR-derived intracelullar signaling involved in B cell maturation (34). Further evidence of the differential activation of LPS in maturing B cells was obtained analyzing the expression of AA4 Ag, which is up-regulated in mature lymphocytes upon LPS activation (data not shown and Ref. 35), but is down-regulated in CD23+ B cells generated in vitro (Fig. 1⇑C). Altogether these data show that signaling through TLR4 can promote B cell maturation, a finding that has not been previously described.
Pam3Cys inhibited the increased generation of CD23+ cells in the presence of lipid A (Fig. 4⇑), suggesting that engagement of TLR2 blocked TLR4-induced maturation. We sought further evidences that engagement of TLR2 alone could be interpreted as arrest in B cell maturation. Interestingly, while transitional B cells (B220dull/highIgMhigh) down-modulate AA4 in the presence of lipid A, the equivalent population in the presence of Pam3Cys kept a high expression of AA4 (Fig. 4⇑B). This result confirms that TLR2 and TLR4 agonists have opposite effects on B cell maturation. The finding that Pam3Cys completely inhibits lipid A-mediated increase of the generation of CD23+ cells suggests a new cross-talk interaction between those TLR signaling pathways. Experiments with TLR2 KO mice (Fig. 5⇑) showed that the inhibition mediated by Pam3Cys cannot be attributed to competition for the TLR4, with Pam3Cys behaving as a hypothetical antagonist for that receptor; the presence of TLR2 is necessary for the inhibition to occur. These experiments also rule out putative molecular interactions between Pam3Cys and lipid A. In most cell culture systems studied in vitro, TLR2 and TLR4 share many of their biological activity promoting the synthesis of a common set of cytokines, although TLR2 seems to be less comprehensive than TLR4, but both stimuli act to promote the synthesis of inflammatory mediators (15). It has also been demonstrated that simultaneous addition of agonists for different TLRs can result in a synergistic effect on transcription factors and cytokine production (36). Surprisingly, our results show a clear antagonism between TLR4 and TLR2 engagement in B cell maturation. To our knowledge, this is the first report of a direct biological antagonism between these TLRs. Cross-tolerance between TLR2 and TLR4 has been shown to require previous stimulation of the first receptor to inhibit the other (36, 37), and it has been explained by inhibition of downstream signaling pathways or down-regulation of receptors upon preincubation with heterologous TLR agonists. However, here the ligands were added simultaneously, and our results indicate a competition between two concurrent activation pathways each one leading to distinct cell fates, rather than a desensitization phenomenon. TLR4 signal transduction involves both MyD88-dependent and independent pathways, whereas TLR2 signals only through MyD88 (38), suggesting a possible explanation for the different effects of TLR4 and TLR2 agonists. Studies with others TLRs restricted to the MyD88-dependent (TLR7/9) or MyD88-independent (TLR3) signal transduction pathways are necessary to test this hypothesis. Alternatively, LPS may also act through RP105 besides TLR4, involving different signaling cascades (39).
We also investigated the effects of TLR4 and TLR2 agonists in early stages of B cell development analyzing cultures of IL-7-responsive B220+IgM− cells. In the presence of Pam3Cys or lipid A, the expansion of B cell precursors in response to IL-7 is almost totally inhibited, showing that these early B cell precursors are also direct targets of those TLR agonists (Fig. 6⇑A). The increased B220 expression in cultures that received lipid A is again suggestive of a maturation effect of TLR4 signaling for early B cell precursors (Fig. 6⇑B). In contrast, B cell precursors in Pam3Cys-added cultures did not show up-regulation of B220. Therefore, despite the similar effect in blocking IL-7-mediated pro-B cell expansion, TLR2 seems unable to trigger the differentiation process, while TLR4 would push it into further maturation. Interestingly, dominance between TLR2 or TLR4 effects in B lineage cells seems to be dependent on the state of differentiation: in IL-7-responsive pro-B/pre-B-I cells the TLR4-mediated effect prevailed over TLR2 (Fig. 6⇑B), whereas the TLR2-mediated inhibition prevails over TLR4-mediated increase in the generation of CD23+ B cells.
The results discussed above raise the question of a role for TLR4 and TLR2 in B cell development in vivo. Although TLRs are considered as receptors for pathogen-associated molecular patterns, endogenous ligands for TLR4 have been described (38), including extracellular matrix molecules, which are known to play important roles in hemopoietic differentiation. We have analyzed by flow cytometry the B cell lineage in BM cells from TLR2- and TLR4-deficient mice, but no significant abnormality in the pre-B, immature B or transitional B cell populations was noted (data not shown). This observation has to be taken cautiously because cell numbers in these compartments are under homeostatic control. The positive/negative selection of B cell clones depends on signaling through the BCR; however, many examples of cross-talk between TLR- and BCR-mediated activation have been reported and can contribute to the overall activation of the lymphocyte (23, 40). An effect of TLR4/TLR2 in B cell development may eventually become apparent in the expression of B lymphocyte repertoire.
It is known that signal transduction triggered by BCR engagement differs between immature and mature B cells and correlates with negative/positive selection in B cell maturation (34, 41). It is interesting to speculate about an analogous picture for TLR signaling. It has been reported that immature B cells are less responsive to activation by LPS to proliferate and secrete Igs, although this notion is still controversial (3, 9). It is possible that the different results reported in the literature were caused by variable amounts of impurities in LPS, mainly lipoproteins (TLR2 agonists) (42, 43). The antagonism between TLR4 and TLR2 ligands in the generation of CD23+ transitional B cells may suggest that each TLR signaling pathway would undergo a differential regulation in B cell development. Recently, evidence for a differential activation through TLR4 in immature B compared with mature B cells was obtained, suggesting a role for lipid rafts in the quality of TLR signaling (44). It is interesting to observe that behavior of the BCR complex in the membrane also changes during B cell maturation, particularly upon engagement of the receptor, which colocalizes with the distribution of lipid rafts in mature B cells, but not in immature (45). These results may suggest a molecular basis for the simultaneous alteration of BCR and TLR4 signaling in immature B, as TLR4 is normally associated with lipid rafts. Further studies on the functional maturation of BCR signaling vs TLR signaling would be necessary to understand how those receptor pathways are integrated in the same cell type and how they affect the generation of the emerging B cell repertoire.
We thank Dr. G. Dos Reis, A. Coutinho, M. Bozza, and M. Bellio for suggestions and critical reading of the manuscript, Dr. Golenbock and Dr. P. Tobias for providing highly purified LPS preparations, Dr. A. Rolink for providing mAb 493 hibridoma, and Dr. L. Zingali and A. L. de Oliveira-Carvalho for purifying 493 mAb.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported by grants from Conselho Nacional de Pesquisas, Fundação de Amparo á Pesquisa do Estado do Rio de Janeiro, and Fundação Universitária José Bonifácio.
↵2 Address correspondence and reprint requests to Dr. Alberto Nobrega, Department of Immunology, Institute of Microbiology, Federal University of Rio de Janeiro, Rio de Janeiro 21941-590, RJ, Brazil. E-mail address:
↵3 Abbreviations used in this paper: BM, bone marrow; sIgM, surface IgM; KO, knockout.
- Received September 7, 2004.
- Accepted March 14, 2005.
- Copyright © 2005 by The American Association of Immunologists