Abstract
Naive murine B cells are known to proliferate and differentiate in response to LPS or CpG, which bind to TLR4 and TLR9, respectively. However, the naive murine B cell compartment is heterogeneous and comprises four different B cell subsets: B-1a, B-1b, marginal zone (MZ), and follicular (FO) B cells. B-1a, B-1b, and MZ B cells are specialized in the response to thymus-independent Ag, and FO B cells are involved in the response to thymus-dependent Ag. This study was undertaken to compare those four naive B cell subsets for their responses to TLR agonists. Quantitative RT-PCR analysis revealed that expression of TLR transcripts differs quantitatively but not qualitatively from one subset to the other. All TLR agonists, with the exception of flagellin and poly(I:C), stimulate B cell proliferation whatever the subset considered. However, TLR ligation leads to massive differentiation of B-1 and MZ B cells into mature plasma cells (PC) but only marginally promotes PC differentiation of FO B cells. Moreover, TLR stimulation strongly up-regulates expression of Blimp-1 and XBP-1S, two transcription factors known to be instrumental in PC differentiation, in B-1 and MZ B cells but not in FO B cells. Altogether, our findings suggest that B-1 and MZ B cells are poised to PC differentiation in response to the microbial environment and that TLR agonists can be instrumental in stimulating Ab-mediated innate immune protection during microbial infections.
The mature naive B cell compartment in the mouse can be divided into four B cell subsets that belong either to the B-2 or the B-1 lineage. B-1 cells reside mainly in the peritoneal and pleural cavities and can be further subdivided into B-1a (B220lowIgMhighCD11b+CD5+) and B-1b cells (B220lowIgMhighCD11b+CD5−). B-2 cells are primarily located in secondary lymphoid organs and comprise two populations designated as follicular (FO)4 B cells (B220+CD23highCD21low) and marginal zone (MZ) B cells (B220+CD23lowCD21high), respectively (1). Due to their unique location near the marginal sinus, MZ B cells are considered critical determinants of host defense directed against encapsulated blood-borne bacterial Ags (2). Nevertheless, there is increasing evidence of a broader role for MZ B cells in both T-independent (TI) and T-dependent (TD) immune responses (3, 4). It has recently been documented that B-1b cells are mostly responsible for the adaptive immune response to TI Ag and exert a memory function (5, 6). Both B-1a and MZ B cells express polyreactive specificities with low affinities to a broad range of Ags and are thus responsible for the production of natural Abs, thereby contributing to innate immunity (1, 6). Conversely, FO B are primarily recruited by TD Ags. They lead to germinal center formation and subsequent production of somatically mutated PC and memory B cells with high-affinity Ag-binding capacities.
TLRs are pattern recognition receptors that play a central role as sensors of infection and inducers of innate and adaptive immune responses (7). Ten TLRs have been described in the mouse so far. Their most common natural or synthetic agonists include peptidoglycan and Pam3CSK4 (TLR1/2), bacterial lipoproteins and MALP2 (TLR2/6), dsRNA and poly(I:C) (TLR3), LPS (TLR4), flagellin (TLR5), ssRNA and imidazoquinolines (TLR7 and TLR8), unmethylated CpG oligodeoxynucleotides (TLR9), and profilin-like molecule (TLR11) (7, 8, 9, 10). It has long been known that certain TLR agonists can deliver B cell stimulatory signals (11, 12). Furthermore, Ruprecht and Lanzavecchia (13) demonstrated that TLR stimulation is required along with BCR triggering and T cell help to sustain optimal proliferation and differentiation of human naive B cells.
In humans, naive B cells express low to undetectable levels of TLRs while memory B cells constitutively express several TLRs that can promote their proliferation and PC differentiation (14, 15). This observation gave ground to the idea that microbial products could contribute to maintain a crucial function of the immune system. In the mouse, naive B cells proliferate and secrete Igs in response to LPS and CpG (11, 12), indicating that they express functional TLR4 and TLR9. However, little is known about the distribution of TLRs and the impact of their agonists on mouse naive B cell subsets. In particular, due to their strong contribution to the antibacterial B cell responses, we believed it was important to explore the susceptibility of B-1 and MZ B cells to TLR agonists.
In this study, we show that the four B cell subsets display a qualitatively similar but quantitatively different pattern of TLR transcript expression. Although all B cell subsets proliferate to a comparable extent to TLR agonists, only B-1 and MZ B cells differentiate into mature PC upon TLR ligation. In agreement with these findings, B-1 and MZ B cells but not FO B cells were found to strongly up-regulate Blimp-1 and XBP-1s, two transcription factors associated with PC differentiation, in response to TLR agonists.
Materials and Methods
Mice
Eight- to 14-wk-old C57BL/6 mice were purchased from Charles River Laboratories and maintained in pathogen-free conditions at the Plateau de Biologie Experimentale de la Souris (Lyon, France).
Flow cytometry and cell sorting
mAbs against CD21/CD35 (7G6), CD23 (B3B4), CD19 (1D3), B220 (RA3-6B2), CD5 (53-7.3), CD11b (M1/70), CD43 (S7), and CD138 (281-2) were purchased from BD Pharmingen. Flow cytometry was conducted using a FACSCalibur and CellQuest software (BD Biosciences). B-1 and B-2 cells were purified from spleens and from peritoneal lavages. For isolation of splenic subsets, single spleen suspensions were prepared by grinding spleens on 70-μm cell strainers in RPMI 1640 medium (Invitrogen Life Technologies). B cell-enriched populations were prepared by depleting T cells with an anti-Thy1.2 mAb (30-H12) and goat anti-rat MicroBeads (Miltenyi Biotec) on LD columns. For FO and MZ cell sorting, enriched B cells were stained with FITC-conjugated anti-CD21, PE-conjugated anti-CD23, and biotin-conjugated anti-CD19 followed by incubation with streptavidin-TriColor (Caltag Laboratories). FO and MZ were sorted by gating the CD19highCD23highCD21int and CD19highCD23lowCD21high cells, respectively, using a FACSVantage (BD Biosciences). For splenic B-1 cell sorting, enriched B cells were stained with FITC-conjugated anti-CD43, PE-conjugated anti-B220, PerCP/Cy5.5-conjugated anti-CD19, and allophycocyanin-conjugated anti-CD5 mAbs. Splenic B-1 cells were sorted by gating the CD19highB220lowCD43high and CD5high population. For isolation of peritoneal B-1 cells, total peritoneal cells were stained with FITC-conjugated anti-B220, PE-conjugated anti-CD11b, biotin-conjugated anti-CD5, and allophycocyanin-conjugated anti-CD19 mAbs, followed by incubation with streptavidin-TriColor. Peritoneal B-1 cells were sorted by gating the CD19highB220intCD11bhigh cells. B-1a and B-1b cells were further discriminated based on their differential expression of CD5. For isolation of peritoneal B2 cells, total peritoneal cells were stained with FITC-conjugated anti-B220, PE-conjugated anti-CD23, biotin-conjugated anti-CD11b mAbs, followed by incubation with streptavidin-TriColor. Peritoneal B-2 cells were sorted by gating the CD23highB220highCD11blow cells. The purity of the sorted populations was between 97 and 99%.
RNA isolation and quantitative real-time RT-PCR
Approximately 106 purified B cells were homogenized with 1 ml of TRIzol reagent (Invitrogen Life Technologies), and total RNA were isolated according to the manufacturer’s protocol. cDNA were synthesized by extension of a mix of oligo(dT) and random primers with SuperScript III reverse transcriptase (all from Invitrogen Life Technologies) in a mixture containing 1 μg of total RNA first digested by RNase-free DNase (2 U/μg RNA) for 15 min at 37°C. The reaction mix was diluted 1/10 and stored at −20°C until real-time PCR analysis. Specific primer sets for murine TLR1–9, MyD88, Toll-IL-1R domain-containing adaptor protein (TIRAP), Toll-IL-1R domain-containing adaptor-inducing IFN-β (TRIF), and Toll-IL-1R domain-containing adaptor-inducing IFN-β-related adaptor molecule (TRAM) adaptors molecules as well as Pax5, Bcl-6, Blimp-1, and XBP-1S transcription factors were designed using BEACON Designer software and were purchased from Invitrogen Life Technologies. All primer sets were intron spanning and the absence of amplicons corresponding to genomic sequence provided evidence for the lack of genomic DNA contamination in the analyzed cDNA samples (data not shown). The real-time PCR was performed on an Applied Biosystems PRISM 7000 using the Platinum SYBR Green qPCR Supermix UDG with a Rox Kit (Invitrogen Life Technologies) according to the manufacturer’s instructions in the Plateau Technique de PCR en temps réel (IFR128). The relative quantity of each transcript was normalized according to the mean of the expression of three different housekeeping genes: 18S RNA, hypoxanthine guanine phosphoribosyltransferase (HPRT), and ubiquitin and multiplied by 103.
The primer sequences (forward/reverse) used were: mTLR1, 5′-GGAAAAAGAAGACCCGCAATC-3′/3′-GACGGACACATCCAGAAGAAAAC-5′; mTLR2, 5′-CGCTCCAGGTCTTTCACCTC-3′/3′-AGGTCACCATGGCCAATGTA-5′; mTLR3, 5′-AAGACAGAGACTGGGTCTGGG-3′/3′-AAGGACGCCTGCTTCAAAGT-5′; mTLR4, 5′-ACTGTTCTTCTCCTGCCTGACA-3′/3′-GGACTTTGCTGAGTTTCTGATCC-5′; mTLR5, 5′-TTGGACTTGGGCCAAAGC-3′/3′-CTGGAGAGTCCACAGGAAAACA-5′; mTLR6, 5′-TCCGACAACTGGATCTGCTC-3′/3′-AAGACTTTCTGTTTCCCCGC-5′; mTLR7, 5′-ATACAGCTCAGCAAAAAGACAGTGT-3′/3′-TCCAGGAGCCTCTGATGAGA-5′; mTLR8, 5′-CTGACGTGCTTTTGTCTGCTG-3′/3′-AGGGAGTTGTGCCTTATCTCGT-5′; mTLR9, 5′-CTAGATGCTAACAGCCTCGCC-3′/3′-GTCACCTTCACCGCTCCTG-5′; mHPRT, 5′-TCATTATGCCGAGGATTTGGA-3′/3′-CAGAGGGCCACAATGTGATG-5′; mubiquitin, 5′-AAGAATTCAGATCGGATGACACACT-3′/3′-GCCACTTGGAGGTTGACACTTT-5′; m18S RNA, 5′-TTCCATCTCTCGCGCAATGG-3′/3′-GAGAGAAGACCACGCCAACG-5′; mBlimp-1, 5′-CTTGTGTGGTATTGTCGGGACTTTG-3′/3′-GTTGCTTTCCGTTTGTGTGAGATTTATC-5′; mPax5, 5′-CGGvACGCTGACAGGGATGG-3′/3′-ACCTCCAAGAATCATTGTAGGAAGAATAC-5′; mBcl-6, 5′-AGAGATGTGCCTCCATACTGCTG-3′/3′-CGTTGCAGAAGAAGGTCCCATTC-5′; and mXBP-1S, 5′-CTGAGTCCGCAGCAGGTGC-3′/3′-AGGGTCCAACTT GTCXAGAATGC-5′.
B cell culture and proliferation assay
Purified FO, MZ, and peritoneal B-1 B cells were plated in 96-well U-bottom plates (BD Biosciences) at 5 × 104 cells in 100 μl of RPMI 1640 containing 10% heat-inactivated FCS, 2 mM l-glutamine, 100 U/ml penicillin, 100 μM streptomycin, 2% HEPES, and freshly added 50 mM 2-ME (all from Invitrogen Life Technologies). B cells were stimulated with the following TLR agonists at the concentrations indicated in the text: CpG oligodeoxynucleotides with phosphorothioate linkage of the base (MWG Biotech) (CpG1668, 5′-TCCATGACGTTCCTGATGCT-3′; CpG1720, 5′- TCCATGAGCTTCCTGATGCT-3′); Pam3CSK4 (Calbiochem), MALP2 (Alexis), LPS from Escherichia coli 0111:B4, R848, poly(I:C), and flagellin from Salmonella typhimurium (InvivoGen). For proliferation assays, cells were cultured in triplicates for 48 h and [3H]thymidine (Amersham Biosciences) was added for the last 12 h (1 μCi/well). [3H]TdR uptake was measured using a TopCount (Packard Instrument) after harvesting.
Quantification of Ig secretion
Purified peritoneal B-1 and B-2 cells and splenic B-1, FO, and MZ B cells were cultured at 105 cells in 200 μl of RPMI 1640 medium as described above. Supernatants (SN) were harvested at day 7 and IgM, IgG, and IgA levels were measured by ELISA. Briefly, 96-well MaxiSorb plates (Nunc) were coated with goat anti-mouse Ig (H + L) Abs (Southern Biotechnology Associates) overnight at 4°C. The plates were washed and then blocked with 2% BSA in PBS for 90 min at 37°C. SN were diluted and added to the plates along with titrated purified mouse IgM, IgG, or IgA isotype control (Southern Biotechnology Associates) to establish a standard concentration curve. The plates were incubated for 90 min at 37°C, then washed, and goat anti-mouse IgM, IgG, or IgA conjugated to alkaline phosphatase (1/1000; Southern Biotechnology Associates) were added and incubated for 90 min. The ELISAs were developed by adding alkaline phosphatase substrate (1 mg/ml; Sigma-Aldrich). ODs were measured using a microplate reader (Molecular Devices) at 405/490 nm, and data were analyzed using the Softmax software (Molecular Devices).
Quantification of Ig-secreting cells (ISC)
Purified FO, MZ, and peritoneal B-1 B cells were cultured at 1 × 105 cells/200 μl of RPMI 1640 medium as described above. After 48 h of culture, cells were washed and distributed at various concentrations into Multiscreen HTS plates (Millipore) precoated with goat anti-mouse Igs (H + L) (Southern Biotechnology Associates) and then incubated for 24 h at 37°C in 5% CO2. Plates were washed, then treated with alkaline phosphatase-conjugated goat anti-mouse Igs (H + L) (Southern Biotechnology Associates) and developed with 5-bromo-4-chloro-3-indoyl phosphate/NBT (Sigma-Aldrich). ISC were enumerated using the KS ELISPOT (Zeiss).
Cytopreparation and immunohistological stainings
Purified FO, MZ, and peritoneal B-1 B cells were cultured at 1 × 105 cells per 200 μl of RPMI 1640 medium as described above. After 48 h of culture, cells were added in a cytocentrifuge chamber and centrifuged at 300 rpm for 5 min with low break. The slides were left to air dry before being fixed with methanol (for coloration) or ice-cold acetone (for enzymatic stainings) for 5 min. For May-Grünwald-Giemsa coloration, the cytospins were incubated with a 2/3 dilution of May-Grünwald solution (RAL Products) for 5 min, washed in distilled water, then incubated with 1/9 dilution of Giemsa (RAL Products) for 10 min. The cytospins were then washed under running water, air dried, and mounted. For cytoplasmic Ig staining, slides were saturated with TBS containing 10% FCS for 20 min at room temperature, then washed with 1× TBS and stained with a rat anti-mouse Ig (H + L) conjugated to biotin (1/50; DakoCytomation) for 1 h at room temperature. For calnexin staining, slides were saturated as described above and stained with a rabbit anti-mouse calnexin Ab (1/200; Sigma-Aldrich) for 1 h at room temperature, washed, and then incubated with a goat anti-rabbit Ig (H + L) Ab conjugated to biotin (1/500; Beckman Coulter) for 1 h at room temperature. The slides were washed, incubated with HRP-conjugated streptavidin (1/50; Sigma-Aldrich) for 1 h at room temperature and developed with 3-amino-9-ethylcarbazole tablets (Sigma-Aldrich). The slides were then counterstained with hematoxylin (DakoCytomation) for 1 min.
Statistical analysis
Statistics were calculated using GraphPad Prism software (version 4.0). Repeated measures one-way ANOVA test was applied where appropriate (∗, p < 0.05; ∗∗, p < 0.01).
Results
Distribution of TLRs in FO, MZ, and peritoneal B-1 B cell subsets
Peritoneal B-1a and B-1b cells and splenic FO and MZ B cells (B-1a, B-1b, FO, and MZ) were isolated by cell sorting to a purity ≥97%, and expression of the TLR transcripts was quantified by real-time RT-PCR. As shown in Fig. 1,A, TLR1, TLR7, and TLR9 were expressed predominantly whatever the B cell subset considered. TLR2, TLR3, TLR4, and TLR6 were expressed at intermediate levels, whereas TLR5 and TLR8 were almost undetectable as compared with the splenocyte controls. Quantitatively, no significant difference was found among the expression levels of TLR2, TLR4, TLR7, and TLR9 in the four B cell subsets (Fig. 1,B). Interestingly, three TLRs were preferentially expressed in B cell populations specialized in the response to TI Ag: TLR3 in MZ B cells, TLR6 in B-1a cells, and TLR8 in B-1a and B-1b cells (Fig. 1 B). In contrast, TLR1 was preferentially expressed in B-2 cells.
We also compared the expression of the four adaptor molecules involved in TLR signaling in B-1a, B-1b, MZ, and FO B cells. MyD88, TIRAP, and TRIF were expressed at similar levels in all B cell subsets, but TRAM, an adaptor of the TLR4 MyD88-independent signaling pathway, was preferentially expressed in B-1a cells (data not shown).
TLR agonists induce proliferation of FO, MZ, and peritoneal B-1 B cells
Because almost all TLRs were expressed in murine naive B cells, we next compared the impact of TLR ligation on the proliferative response of these populations. For this purpose, FO, MZ, and peritoneal B-1 B cells were stimulated for 60 h with graded concentrations of each TLR agonist and analyzed for [3H]TdR incorporation (Fig. 2). Two agonists, flagellin and poly(I:C), failed to stimulate B cell proliferation and thus are not included in Fig. 2. For flagellin, this result was expected due to the lack of TLR5 transcripts in mouse naive B cells. The result was more surprising for poly(I:C) because both the receptor (TLR3) and the adaptor (TRIF) are expressed in MZ B cells. The other TLR agonists can be classified into three categories depending on the differential susceptibility of B cell subsets to their growth-promoting effect: 1) those that induce a response of higher amplitude in B-2 cells than in B-1 cells (CpG); 2) those that induce a response of comparable amplitude in all B cell subsets (R848, Pam3CSK4, and MALP2); and 3) those that induce a response of higher amplitude in B-1 and MZ B cells (LPS). Altogether, these observations indicate that, with the exception of TLR3, the signaling pathways connected to TLRs are functional and promote a strong proliferative response in all B cell subsets.
MZ and peritoneal B-1 but not FO B cells secrete high levels of Ig in response to TLR agonists
We next examined the impact of TLR agonists on the differentiation of FO, MZ, and peritoneal B-1 B cells into ISC. The numbers of ISC were determined by ELISPOT at the end of a 48-h culture period with TLR agonists. This assay reveals all ISC, irrespective of the Ig isotype they produce. As illustrated in Fig. 3,A, CpG1668, LPS, R848, Pam3CSK4, and, to a lesser extent, MALP2 induce a massive differentiation of B-1 and MZ B cells into ISC. In contrast, these TLR agonists only marginally promote differentiation of FO B cells into ISC. Since it has been reported that MZ B cells generate effector cells more rapidly than FO B cells (16), a kinetics study of the generation of ISC in response to CpG1668 was performed. These experiments clearly demonstrated that at day 4 or day 6 after CpG1668 stimulation, FO B cells do not increase their numbers of ISC as compared with day 2 (data not shown). Measurement by ELISA of the levels of Igs secreted in 7-day culture supernatants (Fig. 3, B–D) confirmed the reduced capacity of FO B cells to differentiate into ISC in response to TLR agonists. IgM was by far the dominant Ig isotype produced by B-1 and MZ B cells in response to TLR agonists. IgG and IgA production was also stimulated to a lesser extent in those two populations, whereas it remained almost undetectable in FO B cells.
All B-1 cell populations are equally prone to PC differentiation in response to TLR agonists
It was recently reported that peritoneal B-2 cells show some phenotypical and functional characteristics of the B-1 cells (17). We thus compared 1) peritoneal B-1 and B-2 cells and 2) splenic and peritoneal B-1 cells for their PC differentiation capacity in response to TLR agonists. As shown in Fig. 4,A, splenic B-1 cells did not differ significantly from their peritoneal counterpart as measured by the IgM secretion in response to CpG1668 (Fig. 4,A). In contrast, peritoneal B-2 cells, which as FO B cells express high levels of CD23, only marginally secrete IgM in response to CpG1668 (Fig. 4 A), suggesting that they behave functionally as splenic FO B cells.
As we evoked previously, it has been documented that B-1a and B-1b cells fulfill distinct functions. B-1a cells are responsible for the secretion of natural Abs, whereas B-1b cells are mostly responsible for the adaptive immune response to TI Ag (6). To explore the response to TLR agonists of those two B-1 cell subsets, we purified peritoneal B-1a and B-1b cells and measured their Ig responses 7 days after TLR ligation. The levels of IgM secretion induced by CpG1668, LPS, R848, and Pam3CSK4 in B-1a and B-1b cells were comparable (Fig. 4,B). In contrast, the levels of IgG and IgA secretion were significantly higher in B-1b than in B-1a cells (Fig. 4, C and D), suggesting either that B-1b cells can switch in response to TLR agonists or that a fraction of this population has already undergone Ig isotype switching.
TLR agonists promote differentiation of B-1 and MZ but not of FO B cells into fully mature PC
The observation that TLR agonists induce a massive IgM secretion by B-1 and MZ but not by FO B cells suggested that the latter population could not differentiate beyond the early PC stage in response to TLR agonists. To address this question, we examined the cellular morphology, size of the endoplasmic reticulum/Golgi secretory apparatus, cytoplasmic Ig content, CD138 expression, and Ig secretion capacity of FO, MZ, and peritoneal B-1 B cells 48 h after treatment with CpG1668. As shown in Fig. 5, A–C, B-1 and MZ B cells stimulated by CpG1668 display the morphological features of bona fide PC, i.e., an abundant cyanophilic cytoplasm, a large archoplasm corresponding to the highly developed Golgi apparatus, a highly developed endoplasmic reticulum/Golgi apparatus, and a dense staining for cytoplasmic Igs. In contrast, CpG1668-stimulated FO B cells exhibit a high nuclear:cytoplasmic ratio, an intense basophilic staining of the cytoplasm characteristic of free ribosomes, and a weak staining for both calnexin and cytoplasmic Igs that appeared to be restricted to the perinuclear area (Fig. 5, A–C). In agreement with these findings, a large proportion of B-1 and MZ B cells express the plasma cell marker CD138 (syndecan-1) in response to CpG while FO B cells remain largely negative for CD138 expression (Fig. 5,D). Expression of CD138 in CpG-stimulated FO B cells did not increase at later time points (days 4 and 6), indicating that the lack of expression of these markers is not due to a delayed PC maturation (data not shown). Finally, the Ig secretion capacities of B-1 and MZ B cells following stimulation with CpG1668 were 4- to 5-fold higher, respectively, than that of FO B cells (Fig. 5 E). Interestingly, B-1 and MZ B cells but not FO B cells also display a constitutive Ig secretion capacity.
B-1 and MZ but not FO B cells express a PC-like transcription factor profile upon TLR ligation
Four main transcriptional regulators control PC differentiation: Pax5, Bcl-6, Blimp-1, and XBP-1s. Pax5 and Bcl-6 regulate a panel of genes involved in maintaining B cell identity and germinal center reaction, respectively, while Blimp-1 and XBP-1s are essential for PC differentiation. The reciprocal feedback loop between Bcl-6 and Blimp-1 ensures that activated B cells have two mutually exclusive fates: to enter the germinal center or the PC pathway (18). To further document the unique responsiveness of B-1 and MZ B cells to TLR agonists, we analyzed the expression pattern of these key transcriptional regulator transcripts by real-time quantitative RT-PCR before and after stimulation with CpG1668 for 48 h. As shown in Fig. 6,A, naive B-1, FO, and MZ B cells express similar levels of Pax5, which only declined substantially in B-1 cells after CpG1668 stimulation. Expression of Bcl-6 is constitutively high in FO B cells ex vivo and decreases after stimulation with CpG1668, whereas it remains low in B-1 and MZ B cells whatever their status of activation (Fig. 6,A). Although unstimulated B-1, FO, and MZ B express low levels of Blimp-1 and XBP-1S, stimulation with CpG1668 strongly up-regulates expression of these two genes in B-1 and MZ B cells but not in FO B cells (Fig. 6,A). Thus, the transcriptional profile of CpG-stimulated B-1 and MZ is compatible to that of fully mature PC and correlates with their high Ig secretion capacities. In contrast, the pattern of transcription factor in FO B cells correlates with the very low levels of IgM they secrete in response to TLR agonists. We next compared the expression levels of Bcl-6 and Blimp-1 transcripts in unstimulated FO, B-1, and MZ B cells to those of FO B cells after CpG1668 stimulation. As shown in Fig. 6 B, stimulated FO and unstimulated B-1 and MZ B cells express similar levels of Bcl-6 and Blimp-1 transcripts, suggesting that the PC differentiation capacity of B-1 and MZ B cells before TLR stimulation is equivalent to that of FO B cells after CpG1668 stimulation.
Discussion
The findings reported herein highlight profound functional differences between naive murine B cell subsets regarding their response to TLR agonists. We show that TLR triggering induces B-1 and MZ but not FO B cells to differentiate into fully mature PC. This assertion is supported by our observations that TLR signaling in B-1 and MZ B cells promotes 1) massive IgM secretion, 2) rapid up-regulation of the PC-associated transcription factors Blimp-1 and XBP-1S, 3) induction of CD138, and 4) morphological changes compatible with acquisition of a high secretory capacity. Several lines of evidence suggest that the poor capacity of FO B cells to secrete Igs in response to TLR agonists is not an intrinsic TLR defect. First, the distribution of TLR transcripts is qualitatively similar in all four naive B cell subsets. Second, on a quantitative basis, the levels of expression of TLR transcripts in FO B cells are either comparable (TLR2, TLR4, TLR7, and TLR9) or superior (TLR1) to those of B-1 and MZ B cells. Third, no qualitative nor quantitative differences were found between B cell subsets regarding the expression of three adaptor molecules involved in TLR signaling (MyD88, TIRAP, and TRIF). Finally, and most importantly, the TLR signaling pathway in FO B cells is fully functional as indicated by the strong growth response elicited by TLR agonists in this subset. In fact, two of them (CpG1668 and Pam3CSK4) induce a proliferative response of higher amplitude in FO B cells than in B-1 or MZ B cells. Therefore, as opposed to B-1 and MZ B cells, the signal supplied by TLR agonists is not sufficient to promote full differentiation of FO B cells into PC. Overall, these data suggest that to progress along the PC pathway, FO B cells require complementary signals such as those forwarded by engagement of the BCR or by cognate interaction with helper T cells. Indeed, we found that cross-linking the BCR on FO B cells with anti-μ Abs increases the numbers of PC generated in response to most TLR agonists (data not shown). These data are in agreement with previous data reporting that TLR agonists such as CpG or LPS enhance Ig secretion induced by anti-Ig-dextran (19, 20) or by the more recent observation of Ruprecht and Lanzavacchia (13) showing that optimal differentiation of naive human B cells into ISC can only be achieved when signals delivered by Ag, T cells, and TLR agonists are combined.
Altogether, our findings provide evidence that B-1 and MZ B cells are programmed for efficient maturation into ISC. This may be due, in part, to the fact that they possess an enhanced secretory apparatus (21). This could also be linked to the constitutively high levels of Blimp-1 and low level of Bcl-6 they express as compared with FO B cells. In this respect, our data extend previous observations by the group of Kearney (22) demonstrating that resting MZ B cells constitutively express higher levels of Blimp-1 than FO B cells. Interestingly, B-1 and MZ B cells can give rise to a few ISC when they are plated in the absence of exogenous stimuli (Fig. 3 A). Based on the ELISPOT data, we estimate that this Ig-secreting subcompartment constitutes ∼1% of the B-1 and MZ populations. At face value, the levels of spontaneous Ig secretion by B-1 and MZ B cells may be regarded as marginal but they are nonetheless comparable to the levels of Ig secretion induced from FO B cells upon stimulation with TLR agonists such as CpG oligonucleotides. In keeping with this, the levels of expression of the Blimp-1 transcript in unstimulated B-1 and MZ B cells are comparable to those reached in FO B cells upon stimulation with CpG1668. These results indicate that a fraction of the B-1 and MZ populations might be spontaneously secreting Igs in vivo. The group of Rothstein (23) has reached similar conclusions regarding B-1 cells but proposes that B-1 cells differentiate under a unique transcriptional program. Their study shows that both resting B-1 and B-2 fail to express Blimp-1 while, in accordance with recent data from the group of Calame (24), we find the spontaneous Ig secretion capacity of B-1 cells to be correlated with enhanced expression of the Blimp-1 transcript. One possible explanation for this discrepancy may lie in the B cell population used as comparative data for B-1 cells. We analyzed FO and MZ B cells separately, whereas in the study by Tumang et al. (23) B-1 cells have been compared with the whole B-2 cell compartment that includes both FO and MZ B cells. Because our present data show that MZ B cells and B-1 cells express a similar transcriptional profile, comparison of B-1 cells with B-2 cells could possibly mask the singularities of the PC-associated transcription factor pattern in B-1 cells.
It has recently been shown that peritoneal B-2 cells express some B-1b cell-like features (17), thus suggesting that the B cell functional attributes could also be shaped by the anatomical microenvironment. Our present data show that, despite these similarities, peritoneal B-2 cells behave like splenic FO B cells in response to TLR agonists. In the same way, splenic B-1 cells did not differ significantly from their peritoneal counterpart in terms of PC differentiation in response to CpG1668.
Under certain aspects, murine MZ and B-1 cells behave like human memory B cells with which they share the constitutive TLR expression and enhanced responsiveness to TLR agonists. The propensity of human memory B cells to differentiate into PC in response to TLR agonists has been proposed to provide a means whereby microorganisms sustain the titers of seric-protective Abs by continuously reactivating memory B cells in an Ag-independent fashion. For B-1 and MZ B cells, the physiological role fulfilled by this functional trait is unknown. B-1 and MZ B cells are involved in innate humoral functions such as production of gut IgA and natural Abs that help to prevent spreading of commensal or infectious microorganisms. It is thus tempting to speculate that microbial signals delivered through TLRs could be instrumental in these functions. The published literature does not support a crucial role for microbial signals and TLRs in establishment or maintenance of these innate B cell functions. For example, a recent report from Pasare and Medzhitov (25) has demonstrated that MyD88-deficient mice express normal serum IgA Abs (although IgA secretion in the gut had not been explored). Furthermore, contradictory results have been published regarding the reduction of serum IgM levels in naive MyD88-deficient mice or MyD88/TRIF double-deficient mice (25, 26), and natural Abs are still produced in germfree mice (27, 28). It cannot formerly be excluded that formation and replenishment of the pool of natural Abs involves compensatory mechanisms other than TLR signaling or that TLRs can be triggered by endogenous ligands. Nevertheless, available data rather support the notion that the unique ability of B-1 and MZ B cells to become PC upon TLR triggering is instrumental in stimulating Ab-mediated innate immune protection during microbial infection. It has been demonstrated that naive MyD88 knockout mice are more vulnerable to bacterial infection (29, 30) than naive wild-type mice. However, once the adaptive arm of the immune response has been mobilized by prior vaccination with heat-killed bacteria, MyD88 knockout mice and wild-type mice are equally protected from challenge with live bacteria. This indicates that disruption of the TLR signaling pathway primarily impairs innate immune protection mechanisms. The recent work from Tedder and colleagues (6) has revealed that B-1a cells are a crucial component of innate humoral responses. We thus propose that TLR stimulation in the course of natural infection allows a rapid raise of the production of protective natural Abs by B-1 and MZ B cells. This could contribute to reduce the microbial load below life-threatening levels until the adaptive arm of the immune response becomes fully operational.
Recent evidence has demonstrated that TLRs may play a pivotal role in B cell activation (25), suggesting that TLR agonists can be considered as potential adjuvants for conventional TD Ags. Our results demonstrating that B-1 and MZ B cells are poised to PC differentiation in response to TLR agonists suggest that the use of TLR agonists for adjuvantation of TI vaccines should also be considered.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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.
This work was supported by grants from the Institut National de la Santé et de la Recherche Médicale and Sanofi Pasteur.
Abbreviations used in this paper: FO, follicular; MZ, marginal zone; TI, thymus independent; TD, thymus dependent; PC, plasma cell; HPRT, hypoxanthine guanine phosphoribosyltransferase; ISC, Ig-secreting cell; SN, supernatant.