It has been known for decades that neonates are susceptible to transplant tolerance, but the immunological mechanisms involved remain to be fully elucidated. Recent evidence indicates that the maturation state of DCs responding to an allograft may have a profound impact on whether immunity or tolerance ensues. Given that TLR activation is a key process leading to DC maturation, we hypothesized that DCs from neonates have defective TLR immune responses. Contrary to our hypothesis, we found that murine neonatal DCs demonstrated enhanced TLR responses in comparison to adult counterparts in vitro. However, we found that neonatal B cells possess unique immunoregulatory functions as they impaired DC responses to TLR activation in an IL-10-dependent fashion. Functionally, we demonstrated that TLR-activated neonatal, but not adult, B cells impaired Th1, but not Th2, T cell alloimmune responses in vitro and in vivo, in models of alloimmune priming and allotransplantation. We conclude that neonatal B cells possess unique immunoregulatory properties that inhibit DC function and modulate alloimmunity in our murine experimental systems.
Neonates demonstrate increased susceptibility to infection, impaired responses to vaccination, and a bias toward Th2 vs Th1 T cell responses (1, 2, 3, 4). Furthermore, the first description of actively acquired immune tolerance was in the neonate; Billingham, Brent, and Medawar (5) demonstrated that fetal mice injected with allogeneic lymphoid cells at day 17 of gestation are able to accept a skin graft from the donor strain as adults. Subsequent studies demonstrated that tolerance could be induced by injecting neonates shortly after birth, rather than fetal mice (6, 7). Later reports indicated that neonates are not intrinsically unable to mount immune responses; immunity can be induced by manipulating the Ag dose or strength of signal (8, 9, 10).
Recent reports demonstrate that human infants also show a privileged response following transplantation; 80% of infants given ABO-incompatible heart transplants accept those grafts (11, 12, 13, 14), whereas this usually leads to a lethal outcome in adult patients. Transplant acceptance in these infants is associated with B cell tolerance (15). Understanding the mechanisms that lead to neonatal transplant tolerance will give us insight into the pathways that modulate alloimmune responses and may aid in the development of protocols to induce transplant tolerance in adults.
In murine models of neonatal transplant tolerance, the immunological mechanisms associated with graft acceptance vary according to the strain combination used; development of hemopoietic chimerism with donor cells, Th2-skewing of the immune response (associated with increased IL-4 and IL-10 levels), T cell deletion, anergy, and regulatory mechanisms may all play a role (16, 17, 18, 19, 20). Neonatal T cells preferentially mount Th2 responses in vivo, although in the right circumstances, these T cells are capable of mounting adult-like Th1 responses (4, 8, 9, 10, 21). Hence, the immunological properties of the neonate that lead to this tolerant state remain unclear.
There is a growing appreciation of the role of innate TLR immune responses in modulating adaptive alloimmunity. We, and others, have shown that TLR responses are important during the recognition of allografts and have a profound effect on the acquisition of transplant tolerance (22, 23, 24, 25). Thus, impaired TLR responses might be one of the components contributing to tolerance induction in the neonatal transplant model.
The characteristics of the neonatal innate immune system are unclear, with conflicting reports suggesting that responses are similar to the adult or the impaired. Several reports describe altered production of proinflammatory cytokines following stimulation of human neonatal cord blood cells with innate immune agonists: the secretion of some cytokines is reduced relative to adult counterparts (e.g., TNF-α), while that of others is preserved (e.g., IL-6) or even enhanced (e.g., IL-8); altered up-regulation of costimulatory molecules and impaired T cell priming have also been described (26, 27, 28, 29, 30, 31). There is disparate data regarding whether dendritic cells (DCs)3 and monocytes purified from neonatal mice demonstrate intact or impaired cytokine production and whether they possess an equal ability to prime T cells, relative to their adult counterparts (32, 33, 34, 35). Importantly, some of these studies were performed before the appreciation of the role of TLRs in innate immunity. Furthermore, many of these studies do not distinguish between intrinsic impairments in cell function and the effects of soluble factors in neonatal serum, such as adenosine, which have been implicated as modulators of neonatal innate immune function (36, 37, 38).
We hypothesized that neonates have defective TLR innate immune responses, leading to impaired alloimmune responses. Contrary to our hypothesis, we found that purified murine neonatal DCs demonstrated enhanced secretion of proinflammatory cytokines and up-regulation of costimulatory molecules in response to TLR agonists, relative to their adult counterparts. In contrast, we found that neonatal B cells demonstrated immunoregulatory functions in our in vitro and in vivo model systems. Neonatal B cells produced substantial amounts of the immunosuppressive cytokine IL-10 in response to a range of TLR agonists and suppressed production of IL-12p40 and up-regulation of costimulatory molecules by DCs in response to TLR agonists, whereas adult B cells did not. This suppression was IL-10 dependent. Neonatal B cells also demonstrated quantitative and qualitative differences relative to their adult counterparts, notably, an increased abundance of the CD5+ B-1a subpopulation. Finally, neonatal, but not adult, B cells suppressed Th1, but not Th2 alloimmune T cell responses, effectively skewing the response toward Th2, both during in vitro MLRs, and in vivo, following their injection into adult mice in models of alloimmune priming and allotransplantation. Thus, neonatal B cells may play an active role in modulating alloimmune responses.
Materials and Methods
C57BL/6 (H2b), BALB/c (H2d), and CAF1 (A × BALB/c F1) mice and C57BL/6 IL-10−/− mice (39) were obtained from The Jackson Laboratory. The Yale University Institutional Animal Care and Use Committee approved the use of animals in this study. C57BL/6 TLR-2−/− and TLR-4−/− mice were provided by S. Akira (Osaka University, Japan) and were crossed together to generate TLR-2/4−/− mice.
Purification of DCs, B cells, and CD4+ T cells from neonatal and adult mice
Neonatal mice, aged 1–7 days, and adult mice, aged 6–8 wk, were euthanized and their spleens were harvested. The tissue was dissociated into a cell suspension by mastication between frosted slides. The cells were spun down, resuspended in 5 ml RBC lysis buffer (0.8% NH4Cl, 0.01% KHCO3 by volume in water plus 100 μM EDTA), and incubated on ice for 5 min to lyse RBCs. The cells were then washed with PBS and passed through a nylon filter to remove the capsid and debris. EasySep magnetic selection kits (StemCell Technologies) were used to isolate DCs by positive selection (>80%, EasySep Mouse CD11c Positive Selection kit) and B cells (>90% EasySep Mouse B Cell Enrichment kit), CD3+ T cells (>90%, EasySep Mouse T Cell Enrichment kit) and CD4+ T cells (>90%, EasySep Mouse CD4+ T Cell Enrichment kit) by negative selection, according to the manufacturer’s protocol.
Generation of bone marrow-derived DCs (BMDCs)
BMDCs were generated as previously described (40). In brief, femur bones were flushed with complete RPMI 1640 medium (Invitrogen Life Technologies, containing 5% FCS plus 50 mM 2-ME plus 20 μg/ml gentamicin plus 1% l-glutamine) to collect bone marrow cells. RBC were lysed by resuspending the cells in a hypertonic ammonium chloride solution. T cells and B cells were depleted by incubating the cellular suspension for 1 h at 37°C with GK 1.5, TIB120, TIB211, and B220 mAbs (provided by I. Mellman, Yale University, New Haven, CT) in the presence of rabbit complement. The remaining cells were resuspended in complete RPMI 1640 medium containing 5 ng/ml GM-CSF. Nonadherent cells were removed on day 2, and the medium was replaced on days 2 and 4. BMDCs were collected on day 5 (purity >80% was confirmed by staining with CD11c Ab, coupled with FACS analysis).
DCs, B cells, and unfractionated spleen cells were cultured in RPMI 1640 containing l-glutamine plus 10% FCS plus 1% penicillin/streptomycin in round-bottom 96-well plates, or ELISPOT plates coated with the appropriate Ab, in 200 μl final volume, at 37°C, 5% CO2. TLR agonists were added at the following concentrations, unless otherwise indicated: LPS 5 μg/ml (Sigma-Aldrich), peptidoglycan 1 μg/ml (Sigma-Aldrich), poly(I:C) 0.2 μg/ml (Sigma-Aldrich), CpG oligonucleotide 10 μg/ml (WM Keck Oligonucleotide Synthesis Facility, Yale University, New Haven, CT, sequence: tccatgacgttcctgacgtt), ssRNA40, 5 μg/ml (InvivoGen, complexed with LyoVecu2122 cationic liposome, to enhance stability and facilitate cell uptake), and flagellin 1 μg/ml (InvivoGen).
Unless otherwise indicated, 1 × 105 DCs were cultured ± 2 × 106 B cells/well. For the IL-10 ELISPOT and ELISA, 1 × 105 B cells were added per well. For the unfractionated spleen cell experiments, 4 × 106 cells were added per well. After an 18-h culture period, medium was collected for ELISA analysis, and ELISPOT plates were developed (except the IL-10 ELISPOT, which was developed after a 42-h culture period to facilitate detection).
Mixed leukocyte reaction
For the primary in vitro MLRs, 1 × 105 irradiated (28Gy) spleen cells (stimulators) were mixed with 1 × 105 CD4+ allogeneic T cells (responders) in a round-bottom plate, in complete Bruff’s medium (Invitrogen Life Technologies, containing 1 g/L glucose, 10% FCS, 1% penicillin/streptomycin, l-Glutamine, and 55 μM 2-ME). After 4 days’ culture at 37°C 5% CO2, the cells were resuspended in RPMI 1640 (containing 10% FCS plus 1% penicillin/streptomycin) and transferred to ELISPOT plates coated with IL-2, IFN-γ, or IL-4. For the recall MLRs, 2 × 103 to 3 × 105 irradiated (28Gy) spleen cells from the donor strain (stimulators) were mixed with 3 × 105 spleen cells or 2 × 105 CD3+ T cells (responders) isolated from recipients of allogeneic cells or skin grafts, as indicated. These cells were seeded directly into ELISPOT plates, as described above. ELISPOT plates were developed after an 18-h culture period, according to the manufacturer’s instructions.
For the experiments where medium from cocultures was added to the MLR, 1 × 106 DCs were incubated ± 2 × 106 B cells for 18 h in the presence of 5 μg/ml LPS and then the medium containing soluble factors was collected and transferred to an MLR, as described above. This higher number of DCs was used to facilitate the study of the effect of cytokines on T cell responses (we confirmed that IL-12p40 production by this higher number of DCs was effectively suppressed by the neonatal B cells in this experiment, data not shown). These cultures were set up in complete Bruff’s medium to facilitate the MLR. Furthermore, “mock” wells were seeded containing LPS but no cells. This medium was used as a control to assess the direct effects of LPS on the MLR.
ELISA and ELISPOT
IL-12p40 ELISA was performed using the BD OptEIA mouse IL-12p40 ELISA set (BD Biosciences), according to the manufacturer’s protocol. IL-10 ELISA was performed with the same protocol using purified and biotinylated IL-10 Abs and recombinant murine IL-10 (eBioscience).
ELISPOT was performed using purified and biotin-conjugated TNF-α, IL-6, IL-10, IFN-γ, IL-2, and IL-4 Abs (eBioscience), according to the manufacturer’s protocol.
Flow cytometry and isolation of CD19+CD5+ and CD19+CD5− cell subsets
Cells were suspended at a concentration of 2 × 107/ml in PBS buffer containing 1% FCS/BSA. The relevant primary Ab (eBioscience) was added and the mixture was incubated on ice for 30 min. The cells were then washed in buffer and resuspended at the original concentration. Streptavidin-PerCP secondary Ab was then added, and the mixture was incubated on ice for 15 min. The cells were washed with PBS and then fixed in 1% paraformaldehyde on ice for 15 min. A final PBS wash was performed, and then the cells were resuspended in PBS at a concentration of 5 × 106/ml. Fluorescence data was acquired with a FACSCalibur flow cytometer (BD Biosciences), and then analyzed with FlowJo software (Tree Star).
For isolation of CD19+CD5+ and CD19+CD5− subsets, B cells were first purified from neonatal and adult mice, as described in the previous section, and then stained as described above. After incubation with the secondary Ab, the cells were not fixed, but immediately washed and resuspended in PBS, 0.5% FCS at 2 × 107/ml; live CD19+CD5+ and CD19+CD5− populations were purified using a FACSAria flow cytometer (BD Biosciences).
Peritoneal injection of BALB/c mice with CAF1 BMDCs ± neonatal B cells
CAF1 BMDCs (1 × 106) were cultured ± 2 × 106 neonatal or adult BALB/c B cells for 18 h in the presence of 5 μg/ml LPS. The cells were washed with PBS and then resuspended in a volume of 200 μl. These cells were injected into the peritoneum of adult BALB/c mice (n = 3/group). Fourteen days later, the mice were sacrificed and their spleens were harvested. Spleen cells were harvested and used in a recall MLR, as described above.
Skin transplantation in conjunction with injection of neonatal B cells
BALB/c neonatal or adult B cells were cultured for 18 h in the presence of 5 μg/ml LPS, washed extensively, and then 1 × 107 cells were injected into the tail vein of adult BALB/c mice (n = 3/group). On the same day, these mice (and noninjected controls, n = 3) were transplanted with a C57BL/6 trunk skin graft, as previously described (22). In brief, trunk skin was harvested from a donor mouse. The recipient was anesthetized, and then the thorax was shaved and cleaned with betadine and then ethanol. A patch of skin (∼1 cm2) was cut away from the thorax, and then a piece of donor skin of the same size was situated in the graft bed and attached with surgical staples. Antibiotic ointment and bandages were applied to the graft. Fluids were s.c. administered postoperatively and carprofen analgesic was provided in the drinking water for several days. Twenty-one days later, the mice were sacrificed, and their spleens were harvested. CD3+ T cells were purified and used in a recall MLR, as described above.
All experiments were performed with three or more replicates; all data shown is representative of at least two independent experiments. Data was analyzed using GraphPad Prism Software. A two-tailed Student’s t test, assuming nonequal variance between groups, was used to determine the significance of observed pair-wise differences between groups. A two-way repeated measures ANOVA was used to measure the statistical significance of observed differences with two independent variables. A p value <0.05 was considered statistically significant.
Neonatal DCs demonstrated enhanced TLR innate immune responses
To examine whether TLR responses are intact in neonates, we purified DCs from the spleens of BALB/c neonatal mice (aged 1 to 7 days) and adult mice (aged 6 to 8 wk) and cultured them for 18 h with a range of TLR agonists, including peptidoglycan (which activates TLR-2 and TLR-6), poly(I:C) (TLR-3), LPS (TLR-4), flagellin (TLR-5), ssRNA40 (a ssRNA oligonucleotide containing a GU-rich sequence; TLR-7), and CpG (a ssDNA oligonucleotide containing stimulatory CpG motifs; TLR-9). With every agonist tested, neonatal DCs demonstrated increased secretion of the proinflammatory cytokines IL-6 and TNF-α, relative to adult DCs (Fig. 1⇓, a and c, data shown for LPS, peptidoglycan and poly(I:C)). In addition, neonatal DCs produced a greater amount of the IL-12p40 subunit than adult DCs in response to LPS (Fig. 1⇓b). Neonatal DCs also up-regulated the costimulatory molecule CD40 to a greater degree than adult counterparts, following exposure to LPS (Fig. 2⇓a). A similar result was obtained with CD86 (data not shown).
We found similar results in a heterogeneous population of splenic APCs and in purified splenic macrophages, and these differences were maintained in the C57BL/6 strain (data not shown). Finally, we found that irradiated neonatal and adult C57BL/6 DCs were equally able to prime allogeneic BALB/c CD4+ T cells in a one-way primary MLR (Fig. 1⇑d). Thus, in contrast to our hypothesis, murine neonatal DCs demonstrated enhanced TLR responses and similar T cell priming ability relative to adult DCs.
A cell population within the spleen impaired the up-regulation of costimulatory molecules on DCs in response to LPS
In addition to stimulating purified DCs, we treated unfractionated spleen cells with LPS, and then examined up-regulation of costimulatory molecules in the gated CD11c+ DC population. In contrast to our results with purified DCs, we found that neonatal DCs within the total cultured spleen cells demonstrated a lower baseline level of CD40 at rest and reduced up-regulation of CD40 in response to LPS vs adult counterparts (Fig. 2⇑b). We obtained similar results with CD86 (data not shown). This suggested that another cell population within the neonatal spleen might be modulating DC responses to TLR agonists.
Neonatal B cells suppressed IL-12p40 production and up-regulation of costimulatory molecules by DCs in response to LPS
Several prior reports have shown that adult and neonatal B cells exhibit immunosuppressive functions (41, 42, 43, 44, 45, 46, 47). To test whether neonatal B cells altered DC responses to TLR agonists, we isolated neonatal and adult splenic B cells by magnetic selection, and cultured them together with adult splenic DCs, in the presence of LPS. Eighteen hours later, medium was harvested and IL-12p40 levels were assessed by ELISA. Neonatal B cells suppressed IL-12p40 production from adult DCs in a dose-dependent fashion, whereas adult B cells failed to do so (Fig. 3⇓a; note that neonatal and adult B cells alone produced no IL-12p40 in the presence or absence LPS, data not shown). Neonatal B cells also suppressed IL-12p40 production from neonatal DCs (Fig. 3⇓b). Due to greater availability of material, all additional experiments were done with adult DCs. We found that neonatal B cells also reduced DC up-regulation of costimulatory molecules CD40 (Fig. 3⇓c) and CD86 (data not shown) in response to LPS, in a dose-dependent fashion. Both neonatal and adult B cells produced TNF-α and IL-6 upon TLR activation (data not shown). These data indicate that neonatal B cells suppress specific aspects of the DC response to TLR agonists in vitro.
TLR activated neonatal B cells produced a secreted factor which suppressed IL-12p40 production by TLR-activated DCs
To determine whether a secreted factor is responsible for the neonatal B cell suppression of LPS-induced IL-12p40 secretion by DCs, we mixed either 1) neonatal B cells or, 2) the medium from LPS-activated neonatal B cells, with adult DCs, and cultured them for 18 h in the presence of LPS. We found that medium transferred from LPS-activated neonatal B cells suppressed IL-12p40 production from DCs equally well to neonatal B cells present in the culture (Fig. 3⇑d), indicating that a secreted factor is sufficient for suppression.
Neonatal B cells produced IL-10 following activation with specific TLR agonists
IL-10 is known to suppress IL-12 production (48). Previous studies suggest that adult marginal zone and B-1 B cells produce IL-10 in response to TLR agonists (42, 43, 44, 45). Furthermore, a prior report suggests that neonatal splenic B cells produce substantial amounts of IL-10 in response to the CpG TLR agonist, and therefore suppress the innate immune response elicited by CpG in neonatal DCs (47). To assess whether other TLR agonists elicit a similar effect, we cultured B cells from neonatal or adult mice in the presence of various TLR agonists for 2 days and measured IL-10 production by ELISPOT. We found that a substantial number of neonatal B cells produced IL-10 following stimulation with LPS, peptidoglycan, and CpG (Fig. 4⇓a). There was little IL-10 production with poly(I:C), RNA40, or flagellin (data not shown). In comparison, adult B cells produced relatively little IL-10 in response to TLR stimulation. Furthermore, when neonatal B cells were exposed to LPS, the levels of IL-10 in the medium increased over a 24-h time course; this was not true when neonatal B cells were cultured at rest or when adult counterparts were cultured in the presence of LPS (Fig. 4⇓b).
In TLR-2/4−/− neonatal mice, IL-10 was secreted in response to CpG (TLR-9 agonist), but not LPS (TLR-4 agonist) or peptidoglycan (TLR-2 agonist; Fig. 4⇑c). This indicates that IL-10 production by neonatal B cells required stimulation of specific TLRs by the appropriate agonist.
IL-10 production and TLR signaling in the neonatal B cell were critical for suppression of DC IL-12p40 production
To determine whether IL-10 is critical to the aforementioned suppressive effect, we isolated B cells from IL-10−/− neonatal mice (39) and cultured these cells together with wild-type (WT) adult DCs. We found suppression of DC IL-12p40 production when WT, but not IL10−/−, neonatal B cells were added to the culture, suggesting that IL-10 is necessary for this suppressive effect (Fig. 5⇓a). Furthermore, addition of recombinant murine IL-10 at the same concentration found in medium from LPS-activated WT neonatal B cells (717.5 pg/ml, measured by ELISA), mimicked the suppressive effects of this medium (Fig. 5⇓b), demonstrating that IL-10 is also sufficient for the effect.
To clarify the role of TLR signaling in this suppressive effect, we isolated B cells from TLR-4−/− or WT C57BL/6 neonatal mice, and cultured them with WT adult DCs for 18 h in the presence or absence of LPS or CpG. We found that TLR-4−/− neonatal B cells failed to suppress IL-12p40 production from DCs in the presence of LPS (a TLR-4 agonist; Fig. 5⇑c). However, TLR-4−/− neonatal B cells suppressed IL-12p40 production from DCs in the presence of CpG (a TLR-9 agonist) in a similar fashion to WT neonatal B cells (Fig. 5⇑d). This result indicates that neonatal B cells must be TLR activated to suppress DC IL-12p40 production.
Neonatal B cells were phenotypically different from adult counterparts
We examined whether the different properties of neonatal and adult B cells were associated with different maturation status or the presence of different B cell subsets in these two populations. To investigate this, we stained neonatal and adult spleen cells with fluorescently labeled Abs to CD19 (a B cell marker), B220 (a B cell marker which is down-regulated upon plasma cell activation), CD69 (an early activation marker), MHC class II (MHCII) (I-A/I-E), CD44 and CD86 (plasma cell markers), and CD5 and surface IgM (B-1a cell markers). We found that the percentage of CD19+B220+ B cells increased with age (Fig. 6⇓a). Within the CD19+B220+ B cell population, B220 levels were similar regardless of age (Fig. 6⇓b). The percentage of CD69high B cells increased slightly between day 22 and adulthood (Fig. 6⇓c) and the percentage of MHCII (I-A/I-E)high B cells increased gradually up to day 18, and then sharply between day 18 and adulthood (6 wk; Fig. 6⇓d), which suggested that the neonatal B cells were less mature relative to adult counterparts. Conversely, the percentage of CD44high B cells decreased from day 10 to day 12 (Fig. 6⇓e) and the percentage of CD86high cell was slightly higher in neonates up to 18 days of age vs adults (Fig. 6⇓f), which suggested a more mature phenotype in neonates. CD5+ B-1a B cells were much more abundant in neonates and decreased with age until day 18 (Fig. 6⇓g), consistent with a prior report (49). In line with this observation, the percentage of IgMhigh cells was higher in day 3 neonates (Fig. 6⇓h). In sum, our data demonstrate that the suppressive properties of neonatal B cells were associated with a larger B-1a cell subset and a difference in the expression of several maturation markers vs adult counterparts.
Both neonatal and adult B cells displayed a slight increase in maturation following their magnetic purification, as indicated by the markers described above. However, we do not think our results were explained by a transient, artificial maturation of B cells because only TLR-activated neonatal B cells were able to produce IL-10 and exert the suppressive effect; purified neonatal B cells that were cultured in the absence of TLR agonists (or that lacked the appropriate TLR receptor) did not exhibit these effects (Figs. 4⇑ and 5⇑). Furthermore, resting the cells for 1 day after magnetic purification did not affect their ability to produce IL-10 in response to LPS or their suppressive properties (data not shown).
Suppression by neonatal B cells was primarily due to the CD19+CD5+ subset
Prior work suggests that CD19+CD5+ neonatal and adult B-1a cells secrete IL-10 following CpG stimulation (43, 47). Because this population is more abundant in the neonatal spleen, it is possible that the increased IL-10 production and suppressive function of neonatal B cells may be merely due to increased numbers of CD5+ B-1a cells. To test this, we isolated neonatal and adult splenic B cells, then purified CD19+CD5+ and CD19+CD5− cells by FACS. These cells were then treated with TLR agonists and IL-10 production was measured. A large number of CD19+CD5+ neonatal B cells produced IL-10 in response to LPS, whereas few CD19+CD5− neonatal B cells, CD19+CD5+ or CD19+CD5− adult B cells produced IL-10 following LPS stimulation (Fig. 7⇓a).
Furthermore, the neonatal and adult CD19+CD5+ and CD19+CD5− populations were added to adult DCs and cultured for 18 h in the presence of LPS, and then IL-12p40 production was measured. The greatest degree of IL-12p40 suppression was observed with CD19+CD5+ neonatal B cells. A lesser degree of suppression was observed with CD19+CD5− neonatal B cells and CD19+CD5+ and CD19+CD5− adult B cells (Fig. 7⇑b). Interestingly, B cells isolated from neonatal blood did not produce IL-10 in response to TLR stimulation; however, we did not detect CD5+ B-1a cells in the neonatal blood (data not shown). These data demonstrated that the neonatal B-1a subset was predominantly responsible for IL-10 secretion and suppression of DC IL-12p40 production in response to LPS. Adult B-1a cells did not recapitulate these effects on a cell by cell basis.
Neonatal B cells, but not adult B cells, suppressed an allogeneic primary MLR, and altered the Th1 vs Th2 balance
We used an in vitro MLR to examine the functional consequences of TLR-activated neonatal B cells on T cell alloimmune priming. We cultured C57BL/6 CD4+ T cells with irradiated allogeneic BALB/c spleen cells, in the presence or absence of LPS-activated syngeneic (C57BL/6) neonatal or adult B cells. We found that neonatal B cells suppressed the T cell response to the allogeneic spleen cells, as measured by IL-2 production, whereas adult B cells failed to do so (Fig. 8⇓a).
IL-12 has been implicated in skewing the immune response toward Th1 during T cell priming (50). As neonatal B cells suppress DC production of IL-12p40, we next examined whether neonatal B cells therefore alter the ability of DCs to skew the T cell response toward Th1 through the production of soluble factors. For this purpose, we stimulated DCs with LPS in the presence or absence of neonatal or adult B cells for 18 h. We then isolated the medium from this culture (containing soluble factors produced by the DCs and/or B cells) and added this medium to an MLR containing C57BL/6 CD4+ T cells and irradiated BALB/c spleen cells. Production of the Th1 cytokine IFN-γ and the Th2 cytokine IL-4 were measured by ELISPOT. We found that the medium derived from 1) DCs plus LPS or, 2) DCs plus LPS plus adult B cells induced a Th1-biased T cell response, as indicated by a high IFN-γ/IL-4 ratio. In contrast, the medium from 3) DCs plus LPS plus neonatal B cells induced a Th2-biased T cell response, as indicated by a low IFN-γ/IL-4 ratio (Fig. 8⇑b). This indicates that TLR activated neonatal B cells acted in concert with DCs to alter the balance between Th1 and Th2 alloimmunity in vitro, via the production of soluble factors.
Neonatal B cells reduced T cell priming and skewed the response toward Th2 vs Th1 in adult mice following coinjection with allogeneic cells
Next, we tested the effects of neonatal B cells on T cell priming in an in vivo model. For this purpose, we chose the CAF1 (A × BALB/c F1) donor to BALB/c recipient strain combination because this combination has been used in neonatal tolerance induction models (19, 20, 51), and therefore may be susceptible to in vivo modulation of alloimmune responses by neonatal B cells.
We cultured CAF1 BMDCs with neonatal or adult BALB/c B cells for 18 h in the presence of LPS. BMDCs were used in this experiment instead of splenic DCs because it is possible to generate large numbers of these cells (as with splenic DCs, we found that BMDCs produced IL-12p40 upon LPS stimulation, and this was suppressed by neonatal B cells, data not shown). The cells were then harvested and injected into the peritoneum of adult BALB/c mice. In this experiment, the CAF1 BMDCs were allogeneic to the immunized recipient and therefore primed the recipient T cells, whereas the B cells were syngeneic and gender-matched to the recipient, and therefore merely participated in modulating the character of the response. Two weeks after the injection, spleen cells were purified from the mice, and the effector T cell response to irradiated CAF1 stimulator spleen cells was measured in a recall MLR.
Spleen cells isolated from mice injected with 1) LPS-treated CAF1 BMDCs in the absence of syngeneic B cells produced IL-2, IFN-γ, and IL-4 in response to ex vivo activation with irradiated CAF1 stimulators, in a dose-dependent fashion (Fig. 9⇓, a–c). In comparison, spleen cells from mice injected with 2) LPS-treated CAF1 BMDCs plus neonatal B cells demonstrated reduced production of IFN-γ (Fig. 9⇓a) and IL-2 (Fig. 9⇓b) with each dose of CAF1 stimulators, however, production of IL-4 was similar in these two groups. Thus, neonatal B cells effectively skewed the response toward Th2, as indicated by a low IFNy/IL-4 ratio (Fig. 9⇓d). In contrast, spleen cells from mice injected with 3) LPS-treated CAF1 BMDCs plus adult B cells demonstrated similar/increased IFN-γ, IL-2, and IL-4 production to the group without B cells. Because the B cells were syngeneic to the host in this experiment, this effect is likely to be the result of the neonatal B cells indirectly modulating the character of the immune response to the allogeneic CAF1 BMDCs during the initial priming event. In sum, neonatal, but not adult, B cells suppressed Th1, but not Th2, alloimmune responses in this in vivo model of T cell priming.
Neonatal B cells reduced the Th1 response to an allograft, following their injection into adult skin transplant recipients
Finally, we tested whether neonatal B cells could modulate alloimmune responses in the setting of allograft transplantation. This time, we chose the more stringent (fully MHC mismatched) C57BL/6 to BALB/c allogeneic combination. We activated neonatal or adult BALB/c B cells with LPS for 18 h, then injected them i.v. into gender-matched BALB/c adults (we found that the i.v. tail vein route accomplished more effective B cell trafficking to the spleen vs the i.p. route, data not shown). The same day, the recipients were transplanted with a C57BL/6 skin graft. Twenty-one days later, the mice were sacrificed and their spleens were harvested. CD3+ T cells were prepared by magnetic purification and restimulated with irradiated C57BL/6 spleen cells in a recall MLR, and IFN-γ and IL-2 production were measured by ELISPOT.
We found that T cells from allograft recipients who had received LPS-activated neonatal B cells demonstrated significantly reduced IFN-γ production across a dose range of stimulators in comparison to T cells from allograft recipients that had received adult B cells or no B cells (Fig. 10⇓a). We also observed a slight decrease in IL-2 production in the group that received neonatal B cells, vs adult B cells or no B cells, with the highest dose of stimulators, but this effect was not statistically significant (Fig. 10⇓b). T cells from the transplant recipient groups demonstrated similar IL-4 production to T cells from naive mice (data not shown). Furthermore, allograft rejection occurred at a similar tempo between the experimental groups (data not shown). These data indicate that the LPS-activated neonatal B cells suppressed adult Th1 alloimmune responses, but this effect was not sufficient to prolong allograft survival in this murine transplant model system.
It has been known for many years that neonates demonstrate impaired immune responses and susceptibility to transplant tolerance. Although key advances have been made in our understanding of this process, most studies were performed before the appreciation of the innate immune system. A full understanding of the mechanisms which contribute to impaired alloimmune responses in the neonate will aid in our understanding of the processes that regulate alloimmunity, and may help facilitate the development of clinical protocols to induce transplant tolerance in adults. Furthermore, the identification of defective or inhibitory neonatal immune pathways may help us develop strategies to combat infections and achieve effective vaccination in human infants. TLR activation can have a profound effect on immune regulation (22, 23, 24, 52, 53). Hence, we conducted the current study to examine whether TLR immune responses are defective in neonatal hosts.
Contrary to our hypothesis, we found that murine neonatal splenic DCs demonstrate enhanced TLR responses, compared with adult counterparts (Figs. 1⇑ and 2⇑). However, in unfractionated neonatal splenocytes, some aspects of the DC TLR immune response are down-regulated by an alternative cell population (Fig. 2⇑), which we have shown to be neonatal B cells.
Neonatal B cells secrete the immunoregulatory cytokine IL-10 and suppress DC production of IL-12p40 and up-regulation of costimulatory molecules in response to TLR agonists (Figs. 3⇑ and 4⇑). This suppressive effect is achieved through the production of a secreted factor (Fig. 3⇑), and requires TLR-activation of the B cell (Fig. 5⇑). We have demonstrated that IL-10 is the suppressive factor as it is both necessary and sufficient for the suppressive effect (Fig. 5⇑). We found that neonatal B cells are quantitatively and phenotypically different from their adult counterparts (Fig. 6⇑). The CD5+ B-1a subset is more prevalent in neonates; furthermore, IL-10 production and the suppressive phenotype are predominantly a feature of this B-1a subset (Figs. 6⇑ and 7⇑). Interestingly, LPS-activated neonatal B cells suppress Th1 but not Th2 alloimmune responses both in vitro, as indicated in a primary MLR (Fig. 8⇑), and in vivo following their injection into adult mice along with allogeneic lymphocytes (Fig. 9⇑). Finally, LPS-activated neonatal B cells reduce Th1 T cell responses in an in vivo allograft transplantation model (Fig. 10⇑). Adult B cells do not exhibit similar effects.
TLR activation is essential to mount an effective immune response against certain invading pathogens. However, this process must be tightly controlled as inappropriate or excessive TLR activation may lead to harmful effects including autoimmunity and septic shock. Suppression of immunity may be especially important during gestation to prevent immune reaction to maternal tissues, or excessive inflammatory reactions to microbes, which could result in spontaneous abortion (54). Negative regulators of the TLR activation pathway include soluble decoy TLRs, intracellular regulators (such as IL-1-receptor-associated kinase M, SOCS, NOD2, Toll-interacting protein, and phosphoinositide-3 kinase) and transmembrane protein regulators (such as ST2, single Ig IL-1-related receptor, and TNF-related apoptosis-inducing ligand) (55).
Although TLR activation generally promotes an immune response, under certain circumstances, TLR activation is immunosuppressive. For example, activation of TLRs directly on Tregs may enhance their regulatory function (56, 57, 58, 59). Furthermore, DCs may produce small amounts of IL-10 upon TLR-2 or TLR-4 activation, resulting in autocrine regulatory activity, and cross-talk between different TLRs (48, 60). Furthermore, TLRs may be down-regulated, or cells may undergo apoptosis following excessive TLR activation, causing a negative feedback loop (55).
Previous reports have documented the existence of “regulatory B cells” with immunosuppressive functions, mediated by the production of IL-10 and TGF-β (41, 42, 43, 44, 45). These reports show that adult marginal zone B cells and B-1 cells produce IL-10 in response to TLR agonists. In our experimental system, our results indicate that neonatal B cells manifest a superior capacity to produce IL-10 in comparison to adult counterparts.
We demonstrate that IL-10 production and suppression of DC IL-12p40 production are predominantly a feature of the CD19+CD5+ neonatal B cell population, although a more minor effect was achieved with CD19+CD5− neonatal B cells. The latter effect could be due to the fact that, in addition to CD5− B-2 cells, the neonatal spleen might contain B-1b cells that are CD5− but play a similar role to B-1a cells.
IL-12p40 is known to promote the Th1 pathway (41). In keeping with this paradigm, medium derived from DCs treated with LPS, which contained high levels of IL-12p40, promoted a Th1-biased response during an in vitro MLR. Furthermore, medium derived from DCs that were treated with LPS in the presence of neonatal B cells, which had reduced levels of IL-12p40, favored a Th2 response in this assay. Recent reports indicate that IL-23, which shares the IL-12p40 subunit, is involved in the Th17 pathway (61). In future studies, it would be interesting to determine the levels of IL-23 in medium derived from DCs that are LPS activated in the presence or absence of neonatal B cells, and the effect of this medium on the development of Th17 responses during T cell priming.
Previous studies have shown that that tolerance induction in neonates is associated with increased production of IL-10 and IL-4 (20) and a Th2-skewed T cell response (62). Furthermore, injection of IL-12 along with semiallogeneic spleen cells into the peritoneum of neonatal mice induces the development of a Th1 response, and leads to rejection of the donor strain skin allograft in adulthood (51). In an alternative model of neonatal skin tolerance, where neonatal skin is grafted onto an adult recipient in conjunction with short-term immunosupression, grafts from IL-10-deficient neonates were rejected faster than those from WT neonates (63). Thus, the levels of IL-10 and IL-12 may be critical to modulating the balance between immunoregulation and alloimmunity in the neonate.
Although neonatal B cells were sufficient to suppress Th1 IFN-γ T cell responses to an allograft, this was not sufficient to modulate the tempo of allograft rejection in our experimental system. However, this does not rule out a role for neonatal B cells in this process. It is important to consider that there are several differences in the immune systems of neonates and adults, including quantitative and phenotypic differences in immune cell composition and serum components, such as adenosine (26, 37, 38), which modulate the immune response. These factors are likely to interact in complex and synergistic ways, resulting in the impaired immune responses and improved allograft survival observed in neonates. In the future, it would be interesting to test whether neonatal B cells synergize with other tolerance protocols, such as costimulatory blockade, to improve allograft survival in adult mice, and whether B cells are critical to murine neonatal tolerance models. It would also be interesting to examine whether B cells in human infants produce IL-10 during infection or following transplantation of ABO incompatible grafts.
In conclusion, we have shown that murine neonatal B cells suppress certain aspects of the DC response to TLR agonists, via the production of IL-10, in vitro. Furthermore, neonatal B cells modulate T cell alloimmunity, suppressing Th1 but not Th2 T cell alloimmune responses in our in vitro and in vivo experimental models of alloimmunity and allotransplantation. This mechanism may be one of the factors which contribute to impaired immunity and the development of transplant tolerance in neonates, in concert with other factors, such as altered immune cell frequency and phenotype, and immunoregulatory serum components. Future investigation into the basis of neonatal transplant tolerance will yield vital insights into the mechanisms that modulate alloimmunity and may allow us to develop clinical protocols that are relevant to the adult. Furthermore, this information may allow us to develop more effective strategies to combat infection and achieve vaccination in infants.
We thank Bethany Tesar, Xin Yang, Hua Shen, Steven Kerfoot, and Christian Arroyo for their technical assistance and Mark Shlomchik, Yale University, for his critical reading of the manuscript.
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 the Roche Organ Transplant Research Foundation (29991650 and 86166890) and in part from National Institutes of Health Grant R01 AI064660 awarded to D.R.G.
↵2 Address correspondence and reprint requests to Dr. Daniel R. Goldstein, The Anlyan Center S469, 333 Cedar Street, 3FMP, New Haven, CT, 06520. E-mail address:
↵3 Abbreviations used in this paper: DC, dendritic cell; BMDC, bone marrow-derived DC; WT, wild type; MHCII, MHC class II.
- Received October 4, 2006.
- Accepted May 29, 2007.
- Copyright © 2007 by The American Association of Immunologists