Regulation of the Germinal Center Response by MicroRNA-155

See allHide authors and affiliations

Science  27 Apr 2007:
Vol. 316, Issue 5824, pp. 604-608
DOI: 10.1126/science.1141229


MicroRNAs are small RNA species involved in biological control at multiple levels. Using genetic deletion and transgenic approaches, we show that the evolutionarily conserved microRNA-155 (miR-155) has an important role in the mammalian immune system, specifically in regulating T helper cell differentiation and the germinal center reaction to produce an optimal T cell–dependent antibody response. miR-155 exerts this control, at least in part, by regulating cytokine production. These results also suggest that individual microRNAs can exert critical control over mammalian differentiation processes in vivo.

MicroRNAs are emerging as key players in the control of biological processes, and the stage-specific expression of certain microRNAs in the immune system suggests that they may participate in immune regulation. One such microRNA is miR-155, produced from the non–protein-coding transcript of the bic gene. bic was discovered as a recurrent integration site of avian leukosis virus in chicken lymphoma cells (1). The hairpin from which miR-155 is processed represents the only evolutionarily conserved sequence of the bic gene, indicating that miR-155 mediates bic function (24). bic/miR-155 has been shown to be highly expressed in a variety of human B cell lymphomas, including the Hodgkin-Reed-Sternberg cells in Hodgkin's disease, and miR-155 transgenic mice develop B cell lymphomas (58). In humans, bic/miR-155 expression was detected in activated mature B and T lymphocytes (7, 9), including germinal center (GC) B cells (3, 7), as well as in activated monocytes (10). Germinal centers represent sites of antibody affinity maturation and memory B cell generation in T cell–dependent antibody responses (11).

To obtain insights into the physiological function of bic/miR-155, we generated two mutant mouse strains. In the first, a major portion of the bic second exon, including miR-155, was replaced by a β-galactosidase (lacZ) reporter (12), generating a loss-of-function allele designated bic/miR-155–/–. The reporter allows one to study the bic/miR-155 expression pattern through lacZ expression (fig. S1) (13). Northern blots showed that miR-155 expression was completely ablated in activated bic/miR-155–/– B cells (fig. S1C). To generate the second mutant strain, we used a previously established knock-in strategy (14), to conditionally express miR-155 and an enhanced green fluorescent protein (EGFP) reporter in mature B cells, in a Cre-dependent manner (fig. S2A) (15). For simplicity, mice carrying the miR-155 knock-in and the CD21-cre alleles will be referred to as B cellmiR-155 mice.

The gut-associated lymphoid tissue (GALT), including Peyer's patches (PPs) and mesenteric lymph nodes (mLNs), contains both B and Tcells and activated, proliferating B cells undergoing GC reactions in response to chronic stimulation by gut-derived microbes. We found increased fractions of GC B cells in both PPs and mLNs of B cellmiR-155 mice (Fig. 1A and figs. S3 and S4A), and most of these cells, as well as the non-GC B cells, expressed the EGFP reporter. In contrast, in bic/miR-155–/– mice, the fraction of GC B cells, determined by fluorescence activated cell sorting (FACS) and immunohistochemistry, was significantly reduced in PPs and mLNs (Fig. 1A and figs. S3 and S4, A and B). In bic/miR-155+/– mice, the vast majority of the non-GC B cells were negative for lacZ, whereas ∼60% of GC B cells expressed the lacZ reporter (Fig. 1B and fig. S1B). Because the detection of β-galactosidase activity depends on the sensitivity of the assay as well as the persistence of the enzyme in dividing cells, we conclude that many or perhaps all GC B cells express bic/miR-155 in the course of the GC response.

Fig. 1.

bic/miR-155 regulates the GC response and is induced upon activation. (A) The percentage of PP GC B cells was determined by FACS in bic/miR-155–/– mice (n = 13) and controls (n = 15), and in B cellmiR-155 mice (n = 16) and controls (n = 17). (B) (Left panels) bic promoter activity was measured by LacZ staining in GC B cells with the use of FACS. (Right panels) In B cellmiR-155 mice, bic/miR-155 expression in mature B cells is reported by EGFP expression. (C) RT-PCR was used to detect bic in progenitor, resting B cells and anti–IgM-(Fab′)2–stimulated mature spleen Bcells (10 μg/ml). The smaller transcript represents the spliced form of bic. (D) miR-155 expression was detected by Northern blots in the same samples as in (C). LPS, lipopolysaccharide.

To further characterize miR-155 expression, we isolated spleen B cells from wild-type mice stimulated through the B cell receptor (BCR), CD40, or with mitogens that bind Toll-like receptors (TLRs). Although little bic/miR-155 expression was seen in cells before activation, strong up-regulation was detected under each of these activation conditions (Fig. 1, C and D). The signaling requirements were different for BCR versus TLR/CD40-mediated bic/miR-155 induction. The former appeared to depend on the calcineurin/NFAT (nuclear factor of activated T cells) pathway, but not NEMO, an essential component of the nuclear factor κB (NF-κB) signaling pathway. The latter required both MyD88 and NEMO (fig. S5) (16). A kinetic analysis upon BCR cross-linking showed that both bic and miR-155 up-regulation was transient, with a maximum induction of the former at 3 hours and the latter at 24 hours, consistent with a precursor-product relationship (Fig. 1, C and D). Thus, in the GC response, B cells may up-regulate bic/miR-155 at its initiation or recurrently during proliferation and selection by antigens.

We have also observed that bic/miR-155 expression was absent in nonlymphoid organs (lungs, kidney, brain, liver, and heart) as well as in resting, naïve CD4+ T cells, but strong up-regulation occurs upon activation of these cells by T cell antigen receptor cross-linking (fig. S6), in accord with earlier work in the human (3, 9).

The reduced fraction of GC B cells in the GALT of bic/miR-155–/– mice, together with its increase in mutants overexpressing miR-155 in B cells, suggests that miR-155 may indeed mediate bic function and may also be involved in the control of the GC reaction. To determine whether bic/miR-155 is also involved in induced GC responses in the spleen, we immunized mice with alum-precipitated 3-hyroxy-4-nitro-phenylacetyl (NP) coupled to chicken gamma globulin (CGG), which normally initiates a GC response accompanied by the production of antigen-specific antibodies detectable at day 7 after immunization and reaching a peak 1 week later (17). Antigen-specific immunoglobulin G1 (IgG1) antibody titers and the fractions of GC B cells were compared between immunized mutant and wild-type mice. In mice overexpressing miR-155, the antibody response was marginally enhanced at both time points (Fig. 2A). In contrast, the bic/miR-155–/– mice produced about one-fifth as much NP-specific antibody titers as their littermate controls (Fig. 2A). The percentages and numbers of spleen GC B cells were higher in the miR-155–overexpressing mice, but reduced in the knockouts, compared to controls, most notably on day 14 after immunization (figs. 2B and S7). Furthermore, bic/miR-155–/– spleens displayed reduced numbers of GCs that appeared smaller than those of controls and B cellmiR-155 mice (Fig. 2, C and D). Together, these results complement those obtained for GC formation in the GALT and indicate that miR-155 plays a specific role in the control of the GC reaction in the context of a T cell–dependent antibody response.

Fig. 2.

bic/miR-155–/– mice show impaired T cell–dependent antibody responses. (A and B) Mice were immunized intraperitoneally with NP-CGG/Alum and analyzed on days 7 and 14 after immunization. Open and closed symbols: experiments 1 and 2; triangles: controls; diamonds: B cellmiR-155; squares: bic/miR-155–/–. (A) NP-specific IgG1 levels were measured by enzyme-linked immunosorbent assay. (B) The percentage of spleen CD38loFashi GC B cells was determined by FACS. Un: unimmunized. (C) Immunohistochemistry was performed on day 14 NP-immunized spleen sections from wild-type (a), B cellmiR-155 (b), and knockout mice (c) to detect GCs (brown, PNA+; blue, hemotoxylin). High-magnification image is shown in (d). Images are representative of threemiceper group. (D) Number of GCs (±SEM) was determined from sections in (C); n = 3 mice per group. (E) The frequency of W33L replacement was determined by sequence analyses with spleen GC B cells 12 or 14 days after NP-CGG immunization.

We next investigated a possible molecular basis for the effects of miR-155. With no obviously relevant predicted miR-155 targets in hand, we focused on basic features of the GC response; namely, B cell proliferation, the generation of somatic antibody mutants, and selection of mutated B cells that bound antigen with high affinity. We also examined the production of tumor necrosis factor (TNF) and lymphotoxin-α (LT-α) and LT-β by mutant and control B cells, because it is known that TNF and LT-α, produced by B cells, are critical for GC formation (1820). When bic/miR-155–/– B cells were induced to proliferate in vitro by a variety of stimuli, their proliferation profile, determined by dilution of the cell-bound carboxyfluorescein diacetate succinimidyl ester (CFSE) label, was indistinguishable from that of control cells (fig. S8). There was thus no indication from these experiments that miR-155 expression is directly involved in the control of B cell proliferation. The anti-NP response is characterized by the preferential usage of the VH186.2 gene segment of the IgHb haplotype. Furthermore, high-affinity anti-NP antibodies acquire a tryptophan-to-leucine mutation at position 33 (W33L) (17). GC B cells were thus isolated from bic/miR-155+/– and bic/miR-155–/– mice on day 12 or 14 after immunization with NP-CGG, and rearranged VH186.2 gene segments were sequenced (17). Although there were no notable differences in overall mutation frequency between control and mutant cells, the selection for the W33L mutation was compromised in bic/miR-155 knockout cells (Fig. 2E). Therefore, although miR-155 is not required for somatic hypermutation of antibody genes in GC B cells, it contributes to an optimal selection of cells acquiring high-affinity antibodies. A possible clue to an understanding of the defective GC reaction in bic/miR-155 knockout mice came from the analysis of cytokine production by activated B cells from knockout and control mice. Two days after in vitro activation by BCR cross-linking, TNF production by bic/miR-155–/– B cells was noticeably reduced when compared with that of the controls (Fig. 3A). Consistent with this, the concentration of TNF in culture supernatants of the mutant B cells was about one-third of that in control supernatants (Fig. 3B). The differences in TNF production between knockout and wild-type cells were also apparent at the level of gene expression, as demonstrated by reverse transcription–polymerase chain reaction (RT-PCR) analysis of TNF-specific transcripts (Fig. 3C). We further showed, by RT-PCR, that lt-a but not lt-b expression is also compromised in the mutant cells. These defects were also observed in ex vivo sorted GC and non-GC B cells from mLNs of the knockout mice, where B cells may be chronically activated by exposure to bacterial antigens (Fig. 3D). Together, these data suggest that miR-155 controls the GC response at least in part at the level of cytokine production. Although this control may follow pathways of posttranscriptional gene silencing, we note that fragile-X mental retardation–related protein 1 (FXR1) and Argonaute-2, an RNA-binding protein involved in the microRNA pathway, were shown to associate with an AU-rich element in the 3′ untranslated region (UTR) of the TNF mRNA during translation activation (21). A conserved miR-155 binding site (AGCGUUA) downstream of this element could contribute to the targeting of Argonaute-2 and FXR1 to the TNF 3′ UTR.

Fig. 3.

bic/miR-155–/– B cells are deficient in TNF and LT-α production. All experiments were done with CD19+ mature spleen B cells. (A) TNF expression determined by FACS was analyzed before and after stimulation with anti-IgM-F(ab′)2 antibodies for 2 days (gated on blasted cells). Numbers in panels represent the percentage of cells. Histograms display the amount of TNF expressed at the indicated time points after stimulation. (B) TNF production was measured with the Beadlyte mouse cytokine kit in supernatants from (A). (C) mRNAs were detected by RT-PCR in cells from (A). Lanes: WT (lane 1), bic/miR-155+/– (lane 2) and bic/miR-155–/– mice (lanes 3 and 4), and no cDNA input (lane 5). Data are representative of five independent experiments. (D) mLNs non-GC and GC B cells were sorted, and RT-PCR was performed as in (C).

Because bic/miR-155 is also expressed in T cells upon activation and differential cytokine production is a hallmark of T cell differentiation into T helper cell 1 (TH1) and TH2 effector cells, we tested TH1 and TH2 differentiation of knockout and control T cells in vitro (22, 23). We found that T cell differentiation proceeded normally in both cases (Fig. 4A). However, when T cells were cultured under conditions that promote neither differentiation pathway or suboptimally promote TH2 differentiation, bic/miR-155–/– cells produced more interleukin-4 (IL-4) and less interferon-γ (IFN-γ), suggesting that they were more prone to TH2 differentiation than controls (Fig. 4 and fig. S9). In addition, mutant T cell cultures generated more cells producing IL-10, a cytokine known to dampen immune responses (24, 25).

Fig. 4.

bic/miR-155–/– T cells show a TH2 cytokine bias accompanied by a higher fraction of cells producing IL-10. (A) TH1 (IL-12 + anti–IL-4), TH2 (IL-4 + anti–IFN-γ + anti–IL-12), nonpolarizing (THN; no addition of exogenous cytokines or blocking antibodies) conditions were used to study T cell differentiation with purified CD4+ T cells from peripheral lymph nodes. On day 5, intracellular IL-4 and IFN-γ production was measured by FACS (mean ± SD; five knockout and four wild-type mice from three independent experiments). (B) Cells prepared as in (A) were differentiated under the influence of a limited quantity of IL-4 (12.5 U/ml) (mean ± SD, four knockout and three wild-type mice, from two independent experiments). Numbers in panels represent the percentage of cells.

Although it remains to be seen whether these observations relate to the impaired GC response in the mutants, the present experiments establish, through a combined genetic loss- and gain-of-function approach, that miR-155 is critically involved in the in vivo control of specific differentiation processes in the immune response and that it exerts its functions at least partly at the level of control of cytokine production.

Supporting Online Material

Materials and Methods

Figs. S1 to S9


References and Notes

View Abstract

Navigate This Article