|Home | About | Journals | Submit | Contact Us | Français|
IL-10 plays a central role in restraining the vigor of inflammatory responses, but the critical cellular sources of this counter-regulatory cytokine remain speculative in many disease models. Using a novel IL-10 transcriptional reporter mouse, we found an unexpected predominance of B cells (including plasma cells) among IL-10 expressing cells in peripheral lymphoid tissues at baseline and during diverse models of in vivo immunological challenge. Use of a novel B cell-specific IL-10 knockout mouse revealed that B cell-derived IL-10 non-redundantly decreases virus-specific CD8+ T cell responses and plasma cell expansion during murine cytomegalovirus (MCMV4) infection and modestly restrains immune activation after challenge with foreign antibodies to IgD. In contrast, no role for B cell-derived IL-10 was evident during endotoxemia; however, while B cells dominated lymphoid tissue IL-10 production in this model, myeloid cells were dominant in blood and liver. These data suggest that B cells are an under-appreciated source of counter-regulatory IL-10 production in lymphoid tissues, provide a clear rationale for testing the biological role of B cell-derived IL-10 in infectious and inflammatory disease, and underscore the utility of cell type-specific knockouts for mechanistic limning of immune counter-regulation.
While the immune system is essential for host protection against pathogens, all immune responses themselves have potential for harming the responding host. It is thus not surprising that immune responses are tightly regulated in time, space and character. Despite this, dysregulated inflammation is central to the pathogenesis and expression of a broad range of diseases, both infectious and not. Although immunologists have traditionally focused on the mechanisms underlying immune response activation and class specification, control of immune response amplitude and resolution is just as important. Numerous, often overlapping and redundant mechanisms of immune counter-regulation have been identified in recent years, including counter-regulatory proteins and lipids, specialized regulatory cell types, inhibitory receptor/ligand pairs, and modulators of intracellular signaling pathways. Among these diverse mechanisms, IL-10 appears to play a uniquely important role.
Originally identified as a product of Th2-polarized CD4+ T cells that inhibited Th1 cell cytokine production (“cytokine synthesis inhibitory factor”), IL-10 has subsequently been shown to be produced by a plethora of adaptive and innate immune cell types and to have pleiotropic, largely counter-regulatory, activities across the immune system (1). Murine studies have underscored the central role played by IL-10 in restraining inflammatory responses: local and systemic, innate and adaptive, and polarized along diverse axes of response—Th1, Th2 and Th17) (1-3). During some experimental infections (e.g., toxoplasmosis), IL-10-mediated down-modulation of inflammatory responses is essential to prevent the development of lethal immunopathology (4). In other experimental infections (e.g., cutaneous leishmaniasis and infection with lymphocytic choriomeningitis virus), the counter-regulatory activity of IL-10 is required to induce latency or chronicity (5-7). IL-10 also restrains the expression of autoimmune and allergic disease (1), and is required to prevent mice from developing pathological gut inflammation in response to endogenous gut flora (8, 9).
Cells of the innate immune system (including monocytes, macrophages, dendritic cells, neutrophils, eosinophils, and innate lymphocyte populations), cells of the adaptive immune system (including CD4+ T cells, CD8+ T cells, gut intraepithelial T cells, and B cells), and even non-immune cells (e.g., keratinocytes and hepatocytes) can all express IL-10 (1, 10-14). This profusion of IL-10-producing cell types, along with difficulties in detecting IL-10 via intracellular staining without relatively non-physiological stimulation, have made it difficult to disentangle the roles played by specific IL-10-expressing cell types in regulating immune responses. The focus has largely been on CD4+ T cells in infectious models, where experimental attention has been paid to the regulatory role of IL-10 production by CD4+ T cell sub-populations (2).
In order to better define IL-10 production in vivo, we generated an IL-10 transcriptional reporter mouse. This reporter revealed an unexpected dominance of B cells among IL-10 expressing cells in lymphoid tissues in naïve mice and during in vivo immunological challenge. Of particular interest, a significant fraction of plasma cells and plasmablasts expressed IL-10 in both naïve and challenged mice. Use of a novel B cell-specific IL-10 knockout mouse revealed biologically important, non-redundant roles for B cell-derived IL-10 in immune counter-regulation during infection with MCMV, including restraint of the amplitude of virus-specific CD8+ T cell responses and down-modulation of plasma cell expansion. B cell-derived IL-10 also modestly restrained immune activation after challenge with foreign antibodies to IgD. B cell-derived IL-10 production was not required to restrain LPS-induced immune activation. However, despite B cell dominance in lymphoid organs, myeloid cells were the dominant IL-10-expressing cells in blood and liver after LPS challenge. These data suggest that B cells are an under-appreciated source of counter-regulatory IL-10 production in lymphoid tissues, and underscore the utility of cell type-specific knockouts for mechanistic analysis of immune counter-regulation.
A 20 kb genomic fragment containing Il10 along with flanking sequences was isolated via recombineering from BAC clone RP23-122P5 (BACPAC Resource Center, Children’s Hospital Oakland). A 6.8 kb EcoRI fragment containing the 5th exon, the endogenous polyA site and the 3′ UTR of Il10 was subsequently subcloned into pBluescript II KS (Stratagene). A floxed neomycin-IRES-eGFP cassette (15) was cloned into the HindIII site between the endogenous stop and polyA sites of il10, followed by subcloning of an HSV-TK cassette (16) into the SalI site of pBluescript II KS. After linearizing with NotI, the targeting vector was electroporated into a C57BL/6 ES cell line, and ES cell clones were selected with G418 and gancyclovir. Correctly targeted clones, screened initially by PCR followed by Southern blot confirmation (17) were injected into C57BL/6 albino blastocysts implanted into C57BL/6 females. Male chimeric mice were bred to C57BL/6 albino females to screen for germ line transmission. The neomycin cassette was floxed-out using C57BL/6 Zp3-Cre mice (Jackson laboratories, Bar Harbor), and correctly targeted heterozygous mice were interbred to generate homozygous Vert-X (Vert, fr. green; X, roman numeral 10) mice. Genotyping of Vert-X mice was performed by PCR using the following oligonucleotides: (a) il10 5′ ACCAAGGTGTCTACAAGGCCATGAATGAATT; (b) GFP 5′ GAGGAAATTGCATCGCATTGTCTGAGTAGGT; (c) il10 3′ CAAAGGCAGACAAACAATACACCATTCCCA.
Il10flox/flox mice, having loxP sites that flank exon 1 of IL10 (18, 19), were bred with CD19-Cre mice (20) in order to generate mice with B cell-specific inactivation of Il10. Southern analysis of cell-type specificity and efficiency of Cre-mediated deletion was performed as described (18): T and B cells were immunostained with anti-CD3 and anti-CD19 (both from BDPharMingen), respectively, and were separated with a MoFlo cell sorter (Cytomation); for isolation of macrophages, cell suspensions from peritoneal lavage fluid were stained with anti-F4/80 (Caltag Laboratories). Genotyping was performed by Southern blot analysis as described earlier (8, 18), or by PCR using the following oligonucleotides: IL-10 5′ CCA GCA TAG AGA GCT TGC ATT ACA, 5′ GAG TCG GTT AGC AGT ATG TTG TCC AG, CD19-Cre 5′ GCT GGG GCT CCC CTT TTC CTC T, 5′ CCC AGA AAT GCC AGA TTA CG. All mice were on the C57BL/6 background. Mice were housed under standard specific pathogen-free conditions. All animal experimentation was done in accordance with institutional guidelines.
TLR4-specific E. coli (K12) LPS was from Invivogen; E. coli (055:B5) and S. typhmurium LPS were from Sigma. CpG DNA (ODN 1826) was from Coley Pharmaceuticals or Metabion. Goat polyclonal anti-mouse IgD was generated as described (21). Parent virus stocks of MCMV strain K181+ (kindly provided by E. Mocarski) and MCMV strain Smith VR-1399 (obtained from ATCC) were prepared in NIH 3T3 cells (22). All virus stocks were stored at -70°C.
Single cell leukocyte suspensions were generated from mouse lymphoid organs as described (23). Peritoneal cells were harvested by lavage as described (24). Liver leukocytes were isolated as described (25), with minor modifications. Briefly, mouse livers were perfused with PBS via the hepatic portal vein, followed by removal of the gallbladder. Livers were mechanically dissociated with scissors, incubated with collagenase IV (Roche), and passed through a 100 μm nylon cell strainer (BD Falcon). After the non-leukocyte fraction was pelleted by low-speed centrifugation, the leukocyte fraction was pelleted and subjected to Percoll gradient purification. Erythrocyte lysis was performed with ACK lysis buffer (Biowhittaker). Salivary glands devoid of adjacent lymph nodes were excised, homogenized and incubated with collagenase IV (Roche). Digested samples were passed through a 40 μm nylon cell strainer (BD Falcon) and washed with PBS. Blood leukocytes were isolated from the inferior vena cava and transferred into EDTA-containing capillary blood collection microvette tubes (Sarstedt). Pelleted cells were stained, followed by red cell lysis as above.
After blockade of Fc receptors with CD16/32 blocking antibody (eBioscience), leukocytes were stained using monoclonal antibodies to the following cell surface markers: CD19-PE, CD19-PE-Cy7, CD1d-PE, B220-APC-Alexa 750, IgM-PE-Cy7, CD11b-Pacific Blue, CD11c-APC, CD11c-Alexa 700, TCRβ-APC, CD4-PE-Cy7, CD4-APC-Alexa 750, CD5-Percp-Cy5.5, CD8-Pacific Blue, CD25-PE, CD16/32 blocking antibody (all from eBioscience); CD8-PE-Cy7, CD138-APC, CD11b-Percp-Cy5.5, IgM-PE, CD19-FITC, (all from BDPharmingen); as well as isotype controls (eBioscience, BDPharmingen). For intra-cellular staining, cells were stained for cell surface markers, treated with BD Cytofix/Cytoperm and BD PermWash buffer, and then stained with IFN-γ-PE or -FITC (BD Pharmingen and eBioscience, respectively). Data were acquired using LSRII, FACSCanto and FACScan (Becton-Dickinson) flow cytometers, and analyzed using CellQuest and FloJo software. Splenic leukocyte subpopulations were purified using a FACSVantage SE FACS sorter (Becton-Dickinson).
Splenic leukocytes were restimulated with MCMV-specific peptides (M45: HGIRNASFI; m141: VIDAFSRL; M38: SSPPMFRV; M78: VDYSYPEV; M57: SCLEFWQRV; m139: TVYGFCLL; IE3: RALEYKNL and M36: GTVINLTSV; 1 μM; JPT Peptide Technologies) (26) for 6 h in the presence of Brefeldin A (Applichem).
FACS sorted splenic leukocytes were plated in 96-well PVDF membrane ELISPOT plates (Millipore) and analyzed for IL-10 protein expression using a mouse IL-10 ELISPOT kit (eBioscience). Spots were quantified on ELISPOT 3B analyzer (CTL, Cleveland, OH). Goat IgG-specific antibodies and systemic IL-10 was quantified from the harvested serum by ELISA assays as described previously (27, 28).
mRNA expression was quantified by qRT-PCR, using a LightCycler (Roche) as described (23) using the following primer sets: (a) IL-10: GAAGCATGGCCCAGAAATCA, TGCTCCACTGCCTTGCTCTT; (b) GFP: TGAGCAAAGACCCCAACGAG, TGGCTGGCAACTAGAGGCAC; β-actin: GGCCCAGAGCAAGAGAGGTA, GGTTGGCCTTAGGGTTCAGG.
In order to track IL-10-producing cells in vivo, a floxed neomycin-IRES-eGFP cassette (15) was inserted between the endogenous stop site and the polyA site of Il10 (Figure 1a). The neomycin resistance marker was excised by breeding with Zp3-Cre mice. Successful Cre-deletion was confirmed by Southern blot analysis (Figure 1b). All steps, from the BAC clone source of genomic DNA through Cre-mediated deletion were done with C57BL/6 mice to obviate the need for backcrossing. Homozygous, targeted mice, designated Vert-X (Fr. green, roman numeral 10), developed normally, were viable, fertile and without obvious phenotype; in particular, such mice lacked evidence of inflammatory bowel disease.
In order to address the fidelity of reporter expression, splenic leukocytes from Vert-X mice were stimulated in vitro with CpG DNA. 18 h after stimulation, CD19+GFP+ and CD19+GFP- cells were FACS sorted, and mRNA expression (GFP, IL-10) and protein secretion (IL-10) were quantified by RT-PCR and ELISPOT, respectively. As shown in Figure 2a and 2b, IL-10 mRNA expression and protein secretion segregated with GFP mRNA and protein expression. Similar fidelity of reporter expression was seen in CD11c+ cells (data not shown).
To address whether IL-10 is transcribed from both alleles, splenic leukocytes from Vert-X heterozygous and homozygous mice (along with wild type controls, to control for baseline auto-fluorescence) were stimulated in vitro with CpG DNA. 24, 48 and 72 h after stimulation, GFP expression was quantified by flow cytometry in CD19+B220+ B cells. While the percentage of GFP+ B cells was similar in heterozygotes and homozygotes (Figure 2c), the mean fluorescence intensity (MFI) of GFP expression in B cells from heterozygous mice was about half that of homozygous mice (Figure 2d). Similar results were seen following challenge with LPS and Pam3Cys (data not shown). These data suggest that IL-10/reporter expression is bi-allelic, as previously reported (29).
Approximately 1.3% of splenic leukocytes were found to express GFP in naïve or PBS-stimulated mice. Flow cytometric analysis revealed that the majority of splenic leukocytes exhibiting baseline GFP expression were B cells (CD19+B220+ and CD19+B220low/- cells [presumptive plasmablasts(30)]; CD138+ plasma cells) and CD4+ T cells (both CD25+ and CD25-) (Figure 3). Analyzed differently—by defining the percentage of cells in a given splenic leukocyte subpopulation that were GFP+, as opposed the percentage of GFP+ cells accounted for by a splenic leukocyte subpopulation—the highest proportion of baseline GFP expression was exhibited by CD19+B220low/- cells, at least 1/3 of which exhibited baseline GFP expression (Table I). Indeed, a right shift in the fluorocytogram of CD19+B220low/- cells from Vert-X mice, compared with wild type mice (data not shown), suggests that all such cells may express Il10, at least to some extent. Splenic marginal zone (CD19+B220+CD21hiCD23lo) and follicular (CD19+B220+CD21intCD23+) B cells also exhibited baseline GFP expression, albeit at lower percentages (6.2 ± 0.3% and 0.5 ± 0.06, respectively).
IL-10 plays a critical, non-redundant role in protecting mice from experimental endotoxemia (31-34). Increased GFP expression was observed in splenic leukocyte populations as early as 3 h after systemic challenge with LPS (data not shown). 48 h after challenge, approximately 10% of splenic leukocytes were GFP+ (Table I). Flow cytometric analysis revealed robust induced reporter expression in B cells, T cells, myeloid cells, dendritic cells and NK cells (Table I). B cells were the numerically dominant population of GFP+ splenic leukocytes (Table II). Further, quantification of mean fluorescence intensity (MFI) among GFP+ cells, revealed that no non-B cell population had a higher GFP MFI than B cells in this model (data not shown). Thus, B cells are the dominant splenic population of IL-10 expressing leukocytes after systemic LPS challenge. Among B cell subpopulations, the fraction of cells expressing GFP was highest in plasma cells (CD19+CD138+), followed by plasmablasts (CD19+B220low/-) and CD19+B220+ cells (Table I). Numerically, however, CD19+B220+ B cells represented the dominant population of GFP+ cells. While marginal zone and follicular B cells are not distinguishable by surface marker expression after LPS stimulation, a recent report has suggested particular enrichment of IL-10 expression by a regulatory population of CD19+CD1dhiCD5+ splenic B cells after in vitro challenge with LPS (35). We were unable to replicate this with in vivo LPS challenge: whereas 5.7 (± 0.8) × 106 cells were CD19+GFP+ 48 h after intravenous challenge with LPS, only 3.1 (± 0.67) × 104 of such cells were CD1dhiCD5+.
B cells were similarly dominant IL-10 producers in the peritoneal cavity, where 5.1 ± 0.3% of all leukocytes expressed GFP after intravenous LPS challenge. Flow cytometric analysis revealed that the majority of these cells (77.8 ± 4.3%) were CD19+CD5+ B1a cells, with CD19+CD5- B1b and B2 cells accounting for another 6.2 ± 0.9%. Notably, however, similar percentages of total peritoneal leukocytes and peritoneal B1a cells expressed GFP in naïve mice, suggesting that intravenous LPS challenge fails to induce IL-10 expression over baseline levels in peritoneal cells (data not shown).
Such B lymphocyte predominance of IL-10 competence after LPS stimulation was not the case in blood or liver leukocytes, however, 3.5 ± 0.6% and 4.3 ± 0.3% of which, respectively, were GFP+ 24 h after systemic LPS challenge (compared with ~0% at baseline). Under such conditions, B cells accounted for < 4% of the GFP+ cells in blood and liver. Myeloid (CD11b+) cells were the dominant such IL-10-expressing population in blood after LPS challenge (accounting for 87.3 ± 0.6% of GFP+ cells); myeloid (CD11b+) cells (49.1 ± 2.7%) and NK (NK1.1+TCR-) cells (28.5± 3.5%) dominated in liver.
In diverse other models of immunological challenge in which IL-10 has been shown to play an important counter-regulatory role, including systemic challenge with another TLR ligand (CpG DNA) and goat anti-mouse IgD antibody (36), and subcutaneous infection with MCMV, B cells also represented the dominant population of IL-10-expressing leukocytes in spleen (Table II). A similar dominance of B cell-derived reporter expression was seen in draining lymph nodes following subcutaneous infection with MCMV (data not shown).
The IL-10 competence of plasma cells was of particular interest. The above models of immune activation were associated with expansion of plasma cell numbers in lymphoid organs; such cells accounted for >15% of total splenic leukocytes after challenge with goat anti-mouse IgD (data not shown). Following MCMV infection and goat anti-IgD challenge, plasma cells accounted for 30% to > 50%, respectively, of GFP+ cells in lymphoid tissue (Figure 4).
In order to define the biology of such B cell IL-10 production, mice with a B cell-specific inactivation of the IL-10 gene were generated by breeding Il10flox/flox mice, which have loxP sites flanking exon 1 of Il10 (18, 19), with CD19-Cre mice (20). The B cell-specificity and efficiency of Cre-mediated deletion was verified by Southern blot analysis of DNA isolated from FACS-sorted leukocyte populations. As shown in Figure 5, complete deletion of the loxP-flanked fragment was observed in splenic B cells from Il10flox/floxCD19-Cre mice, whereas no or only insignificant deletion was observed in CD3+ splenic T cells, peritoneal macrophages or tail biopsies from the same animals. Thus, inactivation of the IL-10 gene in Il10flox/floxCD19-Cre mice is efficient in B cells and is not detected in other cell types by Southern blot analysis.
We have previously shown that rectal prolapse is a sensitive indicator of inflammatory bowel disease in IL-10 mutant mice (18). As previously reported, IL-10-/- and Il10flox/floxCD4-Cre mice exhibited rectal prolapse of variable penetrance by the age of six months (18). In contrast, 0 of 68 Il10flox/floxCD19-Cre mice aged 6 months or older developed prolapse. The absence of intestinal inflammation was confirmed by histological analysis of H&E-stained paraffin sections of small intestine and colon from Il10flox/floxCD19-Cre mice (data not shown). Thus, B cell-specific IL-10 production does not appear to have a critical, non-redundant function in protection from pathological inflammation of the gut.
IL-10 protects against host morbidity during experimental MCMV infection: IL-10-deficient mice exhibit increased disease in the face of more efficient control of virus replication (37). IL-10 also facilitates the establishment of chronic infection in salivary glands, restraining CD4+ T cell-mediated viral clearance and facilitating horizontal transmission (38, 39). Quantification of antigen-specific CD8+ T cell responses by flow cytometric analysis of MCMV peptide-driven IFN-γ production revealed a significant increase in the percentage of MCMV-specific CD8+ cells in Il10flox/floxCD19-Cre mice, compared to wild type mice (Figure 6a) 7 d after infection. Thus, B cell-derived IL-10 plays a non-redundant role in restraining the acute MCMV-specific CD8+ T cell response. B cell-derived IL-10 also decreased plasma cell numbers during acute MCMV infection (Figure 6b).
Injection of mice with a foreign antibody to mouse IgD activates a B cell-dependent, Th2-polarized immune response against antigenic determinants derived from the foreign antibody (36). B cells activated by membrane IgD cross-linking facilitate subsequent CD4+ T cell activation by processing and presentation of the anti-IgD antibody. In turn, activated CD4+ T cells secrete cytokines and provide contact-dependent help that leads to further B cell proliferation and differentiation (36). Endogenously-produced IL-10 regulates this model, as shown by the dramatic increase in nearly all splenic leukocyte populations in anti-IgD-inoculated IL-10 knockout mice, compared to wild type mice (Figure 7a-d). Il10flox/floxCD19-Cre mice produced significantly less IL-10 in this model (Figure 7e). Despite the dramatic increase in plasma cell numbers and IL-10-expressing plasma cells observed after challenge of wild type mice with foreign antibodies to IgD, and the heightened increase in plasmablast and plasma cell expansion observed in IL-10 knockout mice, however, Il10flox/floxCD19-Cre mice exhibited no increases above those seen in wild type mice (Figure 7b); findings that were mirrored by their marginally significant increase in goat IgG-specific antibody production (Figure 7f). That said, Il10flox/floxCD19-Cre mice exhibited significant increases in splenic CD4+CD25- T cell, CD19+B220+ B cell and CD11b+ myeloid cell numbers after anti-IgD challenge (Figure 7b-d), indicating that B cell-derived IL-10 modestly restrains immune activation in this model.
IL-10-deficient mice are exquisitely sensitive to systemic challenge with LPS and CpG DNA (19, 31-34). We have previously shown that mice with selective deletion of IL-10 in T cells have unaltered sensitivity to endotoxic shock (18), whereas mice with selective deletion of IL-10 in macrophages and neutrophils exhibit sensitivity to LPS (but not CpG) intermediate between wild type and IL-10-deficient mice (19). To define whether B cell-derived IL-10 plays a non-redundant role in counter-regulation of innate responses to LPS and CpG, Il10flox/floxCD19-Cre and control mice were challenged i.p. with LPS or CpG DNA. In contrast to IL-10-deficient mice, mice with selective deletion of IL-10 in B cells did not exhibit increased sensitivity to LPS (Figure 8) or CpG (data not shown). Mirroring this, analysis of plasma proinflammatory cytokine production by luminex assays revealed that IL-10-deficient, but not Il10flox/floxCD19-Cre mice had significantly greater proinflammatory cytokine and chemokine production after TLR challenge than wild type mice (data not shown). Thus, B cell-derived IL-10 does not appear to play a non-redundant counter-regulatory role during systemic TLR ligand challenge.
Despite the importance of IL-10 to immune homeostasis and counter-regulation, the critical cellular sources of immunoregulatory IL-10 in most infectious, autoimmune and allergic models, have generally remained speculative. The wide range of cells capable of producing IL-10, difficulties with direct in vivo or ex vivo identification of IL-10 producing cells, and experimental problems with unequivocally defining the role of a particular cell type in IL-10-mediated counter-regulation in vivo (it being harder to show that a particular cell type actually does, as opposed to can, play this role) have all hindered mechanistic insight. Transgenic approaches offer a clear path through these experimental difficulties. Using IL-10 transcriptional reporter mice, we found an unexpected predominance of B cells among IL-10 expressing cells in peripheral lymphoid tissues in both naïve immune system-activated mice. Employing B cell-specific IL-10 knockout mice, we demonstrated that B cell-derived IL-10 has a major, non-redundant role in immune counter-regulation during MCMV infection, a more modest role in restraining immune activation after challenge with foreign antibodies to IgD, and no evident non-redundant role during endotoxemia. B cell-specific IL-10 knockout mice also fail to develop inflammatory bowel disease, indicating that B cell-derived IL-10 is not required for protection from pathological inflammation of the gut.
The reporter fidelity and robustness exhibited by Vert-X mice suggest that they will provide a useful tool for further mechanistic studies of IL-10-mediated counter-regulation. In particular, the Vert-X mouse appears to report IL-10 expression with greater sensitivity than the previously-described tiger mice, whose T cells appear to exhibit a gene-dose dependent decrease in IL-10 protein production (29). Vert-X mice provide information complementary to that provided by a recently published translational IL-10 reporter mouse (40), an experimental system in which cells that have expressed IL-10 protein remain marked for sustained periods. An advantage of transcriptional reporter mice is their ability to unveil the in vivo kinetics and localization of the initiation of gene expression. Subsequent events in the translation and release of cytokine protein are not addressed. There may be qualitative differences in GFP and IL-10 translation; the IRES sequence allows for cap-independent translation of GFP while that of IL-10 is likely to be cap-dependent. Further, IL-10 is thought to be secreted rapidly, while the GFP reporter remains intracellular. Finally, IL-10 has been reported to undergo post-transcriptional regulation due to mRNA destabilizing sequences in the 3′-untranslated region (41, 42); it is possible that this process is modified by the inserted construct. Indeed, use of the IVCCA assay (28) revealed somewhat increased systemic IL-10 production in Vert-X than in wild-type mice, both in the basal state and during cutaneous infection with L. major (data not shown). Despite this and the exquisite sensitivity of L. major infection to IL-10, the course of infection proceeded identically in Vert-X and wild type mice (data not shown). Therefore, while the data provided by all such reporters must be interpreted with caution, these mice enable unique access to in vivo information about IL-10 gene expression.
One example of novel, important information provided by our study of these mice is the relative importance of IL-10 produced by B cells versus other cell types for regulation of immune activation induced by different stimuli. Our results reveal a non-redundant role for B cell-derived IL-10 during MCMV infection, a more modest role after anti-IgD challenge, and no evident role after TLR challenge, even though B cells are a major source of IL-10 in all three models. Possible reasons for this variation include differences in the location, timing and overall vigor of the resulting immune responses in these models. Although B cells are the dominant IL-10-expressing cell type in peripheral lymph organs after LPS challenge, myeloid cells dominate in blood and liver—and Il10flox/floxLysM-Cre mice exhibit increased sensitivity to LPS challenge (19). Additionally, LPS-induced toxicity appears to be largely dependent on myeloid cell secretion of inflammatory cytokines; secretion that may only be suppressed efficiently by autocrine production of IL-10. Early MCMV infection is associated with local viral replication and local immune responses in lymphoid tissue, whereas the dramatic systemic immune activation of multiple IL-10 producing cell types following TLR challenge may lead to functional redundancy among cellular sources of this cytokine. Finally, as noted above, IL-10 expression may not always directly reflect IL-10 protein secretion. However, these data do suggest the likelihood that B cells play an underappreciated role in immune counter-regulation and provide a clear rationale for testing the biological role of B cell-derived IL-10 in these and other models. Previous attempts to address such issues have been beset by technical problems. B cell reconstitution of μMT or RAG-deficient mice leads to a reversed ratio of B cells to T cells, and a reversed ratio of marginal zone to follicular B cells, in spleen (data not shown). Bone marrow chimera studies involving μMT/IL-10-/- and B cell-sufficient (IL-10 wild type) mice, while often compelling (43), can lead to variable proportions of non-B cells expressing IL-10. Adoptive transfer of in vitro activated IL-10 wild type vs knockout B cells (44) may provide clear evidence that B cells can provide IL-10-mediated counter-regulation, without necessarily demonstrating that they normally do this in vivo. Future such studies should be facilitated by Il10flox/floxCD19-Cre and other cell type-specific IL-10 knockout mice.
Increasing percentages of mature B cells, plasmablasts, and plasma cells express IL-10, both in naïve and immunologically challenged mice. This may represent an instance of what may be a general phenomenon: the more differentiated or polarized a cell type is, the more likely it is to express IL-10—arguably why IL-10 was first discovered in Th2 clones, given the increased cellular divisions needed for Th2 vs. Th1 differentiation (45, 46). However, as IL-10 can act as an in vitro differentiation factor for human plasma cells (47-49), such IL-10 production by B cells might act in cis, facilitating plasma cell development. Naïve IL-10 knockout mice exhibit no deficit in plasma cell numbers in spleen or bone marrow, however (data not shown), and such mice exhibit increased plasma cell numbers after anti-IgD challenge. Further, a lack of B cell-derived IL-10 led to increased plasma cell numbers during MCMV infection. Thus, B cell-derived IL-10 either inhibits plasma cell expansion in cis in mice in vivo or, more likely, acts in trans to provide counter-regulatory activity for the overall immune response at the level of antigen presentation.
In addition to well-recognized immunopathogenic roles, particularly in autoimmune disease, B cells have also long been recognized to be able to play immunoregulatory roles. The paradigmatic B cell-derived effector molecule, immunoglobulin, can dampen immune responses through engagement of ITIM-bearing Fc receptors (50), something that has been exploited therapeutically (51). Obviously, immunoglobulins can also regulate the amplitude of immune responses through neutralization of pathogens and pathogen-derived molecules. The current studies add to previous evidence that B cell-derived IL-10 can play an immunoregulatory role in autoimmune disease. Both experimental allergic encephalomyelitis (EAE) induced by active immunization and inflammatory bowel disease (IBD) developing in TCRα-deficient mice are exacerbated by the absence of mature B cells (52, 53). In EAE, B cell deficiency was associated with a lack of recovery from what is normally a monophasic disease (52); bone marrow chimera systems provided strong evidence that the critical missing counter-regulatory factor in μMT mice was B cell-derived IL-10 (43). In the case of IBD, TCRα-/-Igμ-/- double deficient mice developed more severe disease than TCRα-deficient mice, which was ameliorated by the transfer of wild type, but not IL-10-deficient, B cells (53, 54). B cells have also been shown to be able to provide protection against arthritis in the collagen-induced arthritis model through an IL-10-dependent mechanism (44). These models, as well as models implicating IL-10 from neonatal B cells in the impaired activation of neonatal dendritic cells by TLR ligands (55, 56), will likely benefit from reexamination with currently available experimental approaches. More broadly, careful use of reporter and cell type-specific knockout mice should facilitate mechanistic analysis of the role played by IL-10 and other central counter-regulatory mediators in immune homeostasis and disease.
We thank Leah Flick, Sara Wojciechowski, and Tatyana Orekov for technical assistance, and E. Mocarski and J.M. Reddehase for MCMV strains.
1Supported by grants from the National Institute of Allergic and Infectious Diseases (AI057992 and AI052099) to C.L.K. and F.D.F., respectively; and the Deutsche Forschungsgemeinschaft (SFB 621 and SFB 589) to W.M. and A.R., respectively.
4Abbreviations used in this paper: