Cutting edge commentary: origins of B-1 cells.

The origin of B-1a cells, a minority population of B cells that express CD5, are abundant in coelomic cavities, and often produce autoantibodies, has been the subject of study for many years. Accumulating evidence demonstrates that the hypothesis that only B cells arising in fetal or neonatal tissues have the potential to become B-1a cells cannot be true. Rather, B cell receptor-mediated signaling initiated by ligation of autoantigen has now been shown to be required for induction of the B-1a phenotype. Furthermore, cells with a functional B-1a phenotype can be induced from adult precursors by appropriate Ag. At the same time, microenvironment-specific events may determine the likelihood that a given B cell, either adult or fetal derived, enters this pathway. CD5 expression and possibly localization to the peritoneum appear to provide some protection to autoreactive cells otherwise slated for elimination.

Bruton's tyrosine kinase links the B cell receptor to nuclear factor kappaB activation.

The recognition of antigen by membrane immunoglobulin M (mIgM) results in a complex series of signaling events in the cytoplasm leading to gene activation. Bruton's tyrosine kinase (BTK), a member of the Tec family of tyrosine kinases, is essential for the full repertoire of IgM signals to be transduced. We examined the ability of BTK to regulate the nuclear factor (NF)-kappaB/Rel family of transcription factors, as the activation of these factors is required for a B cell response to mIgM. We found greatly diminished IgM- but not CD40-mediated NF-kappaB/Rel nuclear translocation and DNA binding in B cells from X-linked immunodeficient (xid) mice that harbor an R28C mutation in btk, a mutation that produces a functionally inactive kinase. The defect was due, in part, to a failure to fully degrade the inhibitory protein of NF-kappaB, IkappaBalpha. Using a BTK-deficient variant of DT40 chicken B cells, we found that expression of wild-type or gain-of-function mutant BTK, but not the R28C mutant, reconstituted NF-kappaB activity. Thus, BTK is essential for activation of NF-kappaB via the B cell receptor.

The ability of CD40L, but not lipopolysaccharide, to initiate immunoglobulin switching to immunoglobulin G1 is explained by differential induction of NF-kappaB/Rel proteins.

Antibodies of the immunoglobulin G1 class are induced in mice by T-cell-dependent antigens but not by lipopolysaccharide (LPS). CD40 engagement contributes to this preferential isotype production by activating NF-kappaB/Rel to induce germ line gamma1 transcripts, which are essential for class switch recombination. Although LPS also activates NF-kappaB, it poorly induces germ line gamma1 transcripts. Western blot analyses show that CD40 ligand (CD40L) induces all NF-kappaB/Rel proteins, whereas LPS activates predominantly p50 and c-Rel. Electrophoretic mobility shift assays show that in CD40L-treated cells, p50-RelA and p50-RelB dimers are the major NF-kappaB complexes binding to the germ line gamma1 promoter, whereas in LPS-treated cells, p50-c-Rel and p50-p50 dimers are the major binding complexes. Transfection of expression plasmids for NF-kappaB/Rel fusion proteins (forced dimers) indicates that p50-RelA and p50-RelB dimers activate the germ line gamma1 promoter and that p50-c-Rel and p50-p50 dimers inhibit this activation by competitively binding to the promoter without activating the promoter. Therefore, germ line gamma1 transcription depends on the composition of NF-kappaB/Rel proteins.

An NFAT-dependent enhancer is necessary for anti-IgM-mediated induction of murine CD5 expression in primary splenic B cells.

CD5 is a 67-kDa membrane glycoprotein the expression of which in murine splenic B cells is induced by surface IgM cross-linking. To analyze this induction, we transiently transfected primary splenic B cells with luciferase reporter constructs driven by various wild-type and mutated CD5 5'-flanking sequences. The transfected cells were subsequently cultured in medium with or without F(ab')2 anti-IgM (anti-IgM), and luciferase expression was assayed. Using this approach, we identified a 122-bp enhancer element necessary for anti-IgM-mediated induction of the CD5 promoter. Electrophoretic mobility shift assays indicated that four inducible and four constitutive complexes form on the enhancer fragment in nuclear extracts of primary B cells. Supershift assays revealed that two of the inducible complexes contained NFATc. Point mutations that abolished NFAT binding severely impaired enhancer function. Thus, CD5 is a target of NFAT in B cells. A third inducible complex required an intact H4TF-1 site. One of several constitutive complexes required an intact Ebox site while a second required an intact putative ets binding site. Mutation of the H4TF-1, Ebox, and Ets sites, in the presence of wild-type NFAT sites, significantly reduced the activity of the enhancer. Therefore, the induction of B cell CD5 expression requires NFAT binding and binding to at least one of three additional sites in the CD5 enhancer.

Protein-tyrosine phosphatase SHP-1 is dispensable for FcgammaRIIB-mediated inhibition of B cell antigen receptor activation.

The inhibitory Fc receptor, FcgammaRIIB, provides a signal that aborts B cell antigen receptor activation, blocking extracellular calcium influx. Because the protein-tyrosine phosphatase SHP-1 binds tyrosyl phosphorylated FcgammaRIIB and FcgammaRIIB-mediated inhibition is defective in motheaten (me/me) mice, which do not express SHP-1, it was proposed that SHP-1 mediates FcgammaRIIB signaling in B cells (D'Ambrosio, D., Hippen, K. L., Minskoff, S. A., Mellman, I., Pani, G., Siminovitch, K. A., and Cambier, J. C. (1995) Science 268, 293-297). However, SHP-1 is dispensable for FcgammaRIIB-mediated inhibition of FcepsilonRI signaling in mast cells (Ono, M., Bolland, S., Tempst, P., and Ravetch, J. V. (1996) Nature 383, 263-266), prompting us to re-examine the role of SHP-1 in FcgammaRIIB signaling in B cells. We generated immortalized sIgM+, FcgammaRIIB+ cell lines from me/me mice and normal littermates. Co-ligation of FcgammaRIIB and the sIgM antigen receptor inhibits calcium influx in both cell lines. Inhibition is reversed by preincubation with anti-FcgammaRIIB antibodies, indicating that it is mediated by FcgammaRIIB. The inositol 5' phosphatase SHIP is recruited to tyrosyl-phosphorylated FcgammaRIIB in both cell lines. FcgammaRIIB-mediated CD19 dephosphorylation also occurs in the presence or the absence of SHP-1. Our results establish that SHP-1 is dispensable for FcgammaRIIB-mediated inhibition of sIgM antigen receptor signaling.

An essential role for Bruton's [corrected] tyrosine kinase in the regulation of B-cell apoptosis.

Mutations of the Bruton's tyrosine kinase (btk) gene cause X-linked agammaglobulinemia (XLA) in humans and X-linked immune deficiency (Xid) in mice. To establish the BTK role in B-cell activation we examined the responses of wild-type and Xid B cells to stimulation through surface IgM and CD40, the transducers of thymus independent-type 2 and thymus-dependent activation, respectively. Wild-type BTK was necessary for proliferation induced by soluble anti-IgM (a prototype for thymus independent-type 2 antigen), but not for responses to soluble CD40 ligand (CD40L, the B-cell activating ligand expressed on T-helper cells). In the absence of wild-type BTK, B cells underwent apoptotic death after stimulation with anti-IgM. In the presence of wild-type but not mutated BTK, anti-IgM stimulation reduced apoptotic cell death. In contrast, CD40L increased viability of both wild-type and Xid B cells. Importantly, viability after stimulation correlated with the induced expression of bcl-XL. In fresh ex vivo small resting B cells from wild-type mice there was only barely detectable bcl-XL protein, but there was more in the larger, low-density ("activated") splenic B cells and peritoneal B cells. In vitro Bcl-XL induction following ligation of sIgM-required BTK, was cyclosporin A (CsA)-sensitive and dependent on extracellular Ca2+. CD40-mediated induction of bcl-x required neither wild-type BTK nor extracellular Ca2+ and was insensitive to CsA. These results indicate that BTK lies upstream of bcl-XL in the sIgM but not the CD40 activation pathway. bcl-XL is the first induced protein to be placed downstream of BTK.

Adult bone marrow contains precursors for CD5+ B cells.

To determine directly whether B cell precursors of adult origin are capable of generating CD5+ B cells, we reconstituted neonatal C3H. SCID mice with adult C57BL/6 bone marrow and analyzed splenic B cells 10 months later. Surface staining and flow cytometry revealed that the B cells were of donor origin and that 30% were CD5+. This confirms that in vivo generated CD5+ B cells can be adult derived. After anti-IgM (but not lipopolysaccharide) stimulation in vitro, virtually all of the B cells from the bone marrow-reconstituted mice expressed surface CD5. Sequence analysis of expressed VHDJH genes from the CD5+ B cells present after anti-IgM stimulation revealed a high frequency of N nucleotide addition in CDR3 regions. The presence of N nucleotides indicates that these sequences were derived from CD5+ B cells of adult origin rather than from long-lived fetal precursor B cells present in either the adult bone marrow at the time of transfer or adult spleen. These experiments demonstrate conclusively that adult bone marrow contains precursors for CD5+ B cells and that unlike fetal liver-derived precursors these express terminal deoxynucleotidyl transferase.

B-cell activation by crosslinking of surface IgM or ligation of CD40 involves alternative signal pathways and results in different B-cell phenotypes.

Treatment of small resting B cells with soluble F(ab')2 fragments of anti-IgM, an analogue of T-independent type 2 antigens, induced activation characterized by proliferation and the expression of surface CD5. In contrast, B cells induced to proliferate in response to thymus-dependent inductive signals provided by either fixed activated T-helper 2 cells or soluble CD40 ligand-CD8 (CD40L) recombinant protein displayed elevated levels of CD23 (Fc epsilon II receptor) and no surface CD5. Treatment with anti-IgM and CD40L induced higher levels of proliferation and generated a single population of B cells coexpressing minimal amounts of CD5 and only a slight elevation of CD23. Anti-IgM- but not CD40L-mediated activation was highly sensitive to inhibition by cyclosporin A and FK520. Sp-cAMPS, an analogue of cAMP, augmented CD40L and suppressed surface IgM-mediated activation. Taken together these results are interpreted to mean that there is a single population of small resting B cells that can respond to either T-independent type 2 (surface IgM)- or T-dependent (CD40)-mediated activation. In response to different intracellular signals these cells are induced to enter alternative differentiation pathways.

Activation of B-cells by sIgM cross-linking induces accumulation of CD5 mRNA.

The surface membrane molecule CD5 is expressed on mature T cells and on the B-1a subpopulation of B cells. These CD5 positive B cells express an antibody repertoire with a relatively high frequency of self-reactivity. There is uncertainty about the origins of CD5 B cells and the reasons for this are reviewed. Recent reports which relate to the lineage/selection debate are discussed. For instance, an increase in the frequency of CD5 B cells is a feature of several genetically determined polysystem autoimmune syndromes. In the case of motheaten (me, mev) the pathogenesis of this increase in CD5 B cells is not yet understood, even though the mutation has been mapped to the Hematopoietic cell protein-tyrosine phosphatase (Hcph) gene. Another mutation which affects B cell development, X-linked immunodeficiency (xid), encodes a point mutation in a B cell cytoplasmic tyrosine kinase. Expression of xid in otherwise normal mice causes a lack of CD5 B cells and a shift in the antibody repertoire. Interestingly, expression of both xid and motheaten results in an amelioration of autoantibody production. Evidence is presented that in B cells regulation of expression of CD5 can occur at the level of mRNA and that cross-linking of sIgM can induce the accumulation of CD5 mRNA. The overall concept advanced is that cells expressing natural autoantibodies are triggered via sIgM ligation to become CD5 B cells.

Function of bone marrow stromal cell lines derived from nude mice.

We previously reported that in double deficient nude.xid mice B cells failed to develop and their bone marrow did not produce mature B cells in vitro. However, when progenitors from nude.xid bone marrow were placed on a preestablished normal stromal cell line (AC6) they differentiated into surface IgM+ cells. This raised the possibility of a deficiency of nude and nude.xid stromal cells such that they were incapable of supporting the maturation of X-linked immune deficiency (xid) B cells. Here we ask whether bone marrow stromal cells from nude and nude.xid mice have the ability to support xid B cell lymphopoiesis. A primary stromal cell layer derived from nude mice supported xid B cell differentiation in vitro. We derived panels of stromal cell lines by transfection of primary stromal cell layers with a retrovirus encoding SV40 large T Ag. Several bone marrow stromal cell lines derived from nude and nude.xid mice supported xid B cell differentiation from CD43+/CD45 (B220-) to CD45 (B220+) and from CD45 (B220+)/surface IgM- to surface IgM+. Supporting cell lines expressed both IL-7 and insulin-like growth factor I. The frequencies of bone marrow stromal cells capable of supporting xid B cell differentiation were similar in normal, xid, nude, and nude.xid mice. These results demonstrate that nude and nude.xid mice have bone marrow stromal cells with normal abilities to support B cell maturation.

ABL oncogenes directly stimulate two distinct target cells in bone marrow from 5-fluorouracil-treated mice.

Mice reconstituted with BCR/ABL-infected 5-fluorouracil-treated bone marrow are considered a model system for human chronic myelogenous leukemia, a malignancy that arises in hematopoietic stem cells. These animals develop multiple types of hematopoietic tumors, which could arise either from undifferentiated cells that mature during tumor development or from progenitors committed to different lineages. To examine the BCR/ABL-sensitive target cells present in the marrow of mice treated with 5-fluorouracil, we used a single-step in vitro assay. These experiments revealed that both the P210 and P185 BCR/ABL proteins and the related v-abl protein induce lymphoid and myeloid colonies, colony types that mimic two of the prominent types of tumors found in the reconstitution model. The lymphoid colonies were similar to lymphoid colonies found following infection of normal bone marrow with respect to differentiation state and tumorigenicity. The cells in the myeloid colonies were differentiated and non-tumorigenic. Fluorescence-activated cell sorting revealed that most of the lymphoid and myeloid colonies arose from distinct precursors and that the lymphoid colonies arose from B-lineage-committed cells. These data suggest that most of the lymphomas observed in the reconstitution model arise from committed progenitors that are distinct from those involved in the myeloid disease.

B cell deficiency progresses with lineage maturation in nude.X-linked immunodeficient mice B cell deficiency progresses with lineage maturation.

Previously we showed that unlike normal, nude, or X-linked immune deficient (xid) mice, nude.xid mice are deficient in bone marrow pre-B cell targets for Abelson murine leukemia virus transformation. We show that nude.xid bone marrow is deficient in both CD45(B220)+ and CD45(B220)- surface (s)IgM- progenitors that give rise to B cell colonies in Whitlock-Witte cultures. CD45(B220)+ precursors had normal differentiation potential in vitro. CD45(B220)- precursors differentiated into CD45(B220)+ cells at the same rate as normal controls, but acquired sIgM at a much slower rate. These results correlated with the observation that in nude.xid mice the severity of B lineage defects correlates with maturity: a profound (ninefold) deficit of sIgM+, CD45(B220)+ mature B cells, a fivefold deficit in the sIgM-, CD45(B220)+ precursors of short term B cell colonies (colonies forming within 4-5 days in Whitlock-Witte cultures), and a moderate (twofold) decrease in the frequency of sIgM-, CD45(B220)- (less mature) precursors of long term B cell colonies (colonies forming after 14 days of Whitlock-Witte culture. Thus the combination of the nude and xid mutations produces a deficiency in early B cell progenitors and the deficiency becomes more profound with further maturation. Therefore the lack of mature B cells is the result of a cascade effect. Inasmuch as bone marrow progenitors are affected, and these are the source of the vast majority of B cells, most B cells are affected by the xid mutation and the xid defect cannot be attributed to a loss of a fetal lineage of B cells. These results suggest that xid affected cells lack the capacity to progress efficiently through differentiation in the absence of an exogenous factor(s) that is dependent on the product of a normal allele at the nude locus. This product might be supplied in vivo by a T cell or T cell-dependent source and/or epithelial elements such as bone marrow stromal cells all of which are known to be affected by the nude mutation.

Production of 17.2.25 mu transgenic and endogenous immunoglobulin in X-linked immune deficient mice.

In M54 mice transgenic for a completely rearranged mu(a) heavy chain there is a decrease in total B cells and the rearrangement of endogenous immunoglobulin genes is partially inhibited. Surprisingly, however, endogenous immunoglobulin gene rearrangement and significant heavy chain polypeptide production does occur. We tested the hypothesis that only CD5+ B cells produce endogenous immunoglobulin by taking advantage of the fact that X-linked immune deficient (xid) mice normally are deficient in CD5+ B cells. We found that the frequency of CD5+ splenic B cells was similar in XxidY transgenic and non-transgenic F1 males, and in XxidX transgenic and non-transgenic F1 females. In both XxidX and XxidY transgenic F1 mice some, but not all, splenic B cells are CD11b+. There was a striking deficit of splenic B cells expressing endogenous immunoglobulin in XxidY transgenic mice, although this was not true for peritoneal cells. Thus, the introduction of the 17.2.25 mu transgene does not prevent the development of CD5- B cells nor does it limit endogenous immunoglobulin gene arrangement and expression solely to CD5+ B cells. However, in mice capable of expressing B cell surface CD5 or CD11 this transgene can lead to expansion of the fraction of B cells positive for these molecules. We conclude that while the introduction of the 17.2.25 mu transgene alters the frequencies of B cell populations, maturation is not limited to one subpopulation.

Loss of CD23 is a consequence of B-cell activation. Implications for the analysis of B-cell lineages.

When splenic CD5- B cells are stimulated with antiimmunoglobulin they become CD5+ and have a prolonged in vitro life. Further treatment with IL-6 induces a loss of surface CD23 and IgD; that is, they resemble freshly isolated peritoneal CD5+ cells. These data suggest that the CD5 phenotype is induced after sIg-mediated B-cell activation. Additional support for this view arises from the observation that the loss of CD23 and IgD can be induced by another activation inducer, LPS, although in this case CD5 is not expressed. Thus, activation by anti-Ig plus IL-6 or by LPS induces CD23 loss. Consistent with the hypothesis that the loss of CD23 is a consequence of activation, we now report that the surface expression of CD23 varies inversely with the amount of total cellular RNA. We also find both CD23 positive and negative B cells among freshly isolated splenic CD5- B cells. In young mice a proportion of small splenic CD5+ B cells are CD23+, providing additional evidence that CD23 is present on all B cells prior to activation. A comparison of the features of CD5+ B cells and the antibody responses to thymus-dependent and thymus-independent antigens leads us to hypothesize that the CD5 phenotype arises as a consequence of thymus-independent type 2 (TI-2) stimulation. The relationship of CD5 expression to B-cell lineage (fetal vs. adult bone marrow) is discussed.

The Ig VH repertoire of fetal liver-derived pre-B cells is influenced by the expression of a gene linked to X-linked immune deficiency.

Ig VH repertoire differences between normal and x-linked immune deficiency- (xid) expressing mice are well established. To test the hypothesis that such differences might exist as early as the pre-B stage of ontogeny we generated panels of xid fetal liver derived Abelson murine leukemia virus transformants with H chain Ig VDJ rearrangements. Cells from CBA/Tufts.xid mice used VH genes from many families, with no demonstrable preference for 3' genes. Analysis of cells derived from (CBA/Tufts.xid X CBA/Tufts)F1 mice showed preferential usage of 3' family genes in the phenotypically normal females, even though V to DJ joins were made in vivo. The defective male mice did not show this marked preferential usage. A similar, but less marked, effect on VH gene usage was seen in mice with X-linked immune deficiency and a BALB/c background. Taken together, these results show that either X-linked immune deficiency, or a closely linked gene, affects fetal pre-B cells such that the usual pattern of predominant usage of 3' family genes is altered.

Surface markers, heavy chain sequences and B cell lineages.

A unifying theory of B cell development and lineage commitment is presented. There are two firmly established B lineages: cells which normally arise only from fetal sources and lack N insertions in their rearranged heavy chains; and N-containing cells which arise from adult bone marrow precursors (and perhaps from late fetal sources). Commitment to the expression of CD5 and the capacity for long-life (or self-renewal) are induced as a consequence of sIg cross-linking, typically by a repeating epitope, thymus independent type two antigen. Alternatively, activation resulting from cognate interaction with a helper T cell does not induce CD5 but results in lower expression of J11d. In this case activation occurs in the absence of sIg cross-linking. It is further proposed that differences in the Ig repertoire make it highly likely that fetal/neonatal, but not adult derived B cells will be induced to express CD5. The model offers a plausible explanation for the correlation of CD5 expression and natural autoantibody production by neonatal B cells. Possible sources of pathogenic autoantibody are discussed in the context of this model.

Treatment of murine CD5- B cells with anti-Ig, but not LPS, induces surface CD5: two B-cell activation pathways.

Anti-Ig stimulated murine B cells express high levels of surface CD5 (ly-1) and increased CD44 while maintaining surface IgD, CD23 and J11d. Sorting of CD5- and CD5+ cells demonstrates that anti-Ig induces CD5 expression rather than the selective expansion of CD5+ cells. Anti Ig plus interleukin-6 (IL-6) induces the CD23, IgD, low ly-5 (B220) (CD45low), J11dhigh phenotype of typical CD5+ peritoneal B cells. In contrast, lipopolysaccharide (LPS)-stimulated B cells have high levels of CD44 but decreased surface IgD, CD23 and J11d and no CD5. Thus LPS and anti-Ig generate activated cells with differing phenotypes. Induced CD5+ cells have increased viability, even in the absence of added exogenous factors, while the viability of CD5- B cells is dependent on factors such as IL-4. We conclude that conventional CD5- B cells can be activated by either of two pathways: one generating CD5+ B cells; the other yielding conventional activated cells. We hypothesize that the first path requires slg cross-linking and corresponds to T-independent (type 2) stimulation, while cognate interaction with helper T cells in the absence of slg cross-linking induces B cells to enter the second path.

CD4-, CD8- gamma/delta cells from normal mice respond to a syngeneic B cell lymphoma and can induce its differentiation.

Lymph nodes and spleens from normal unimmunized mice contain small numbers of CD3+, CD4-, CD8- (double negative, DN) T cells. Of these, approximately one-third express the marker Ly-5(B220) in a form previously seen only on normal B cells and a population of DN T cells found in mice genetically prone to develop autoimmunity. DN T cells proliferate when co-cultured with a syngeneic surface Ig+ lymphoma, CH12. After one cycle of stimulation with CH12 almost all of the responding CD3+ DN cells express Ly-5(B220), suggesting that it is an activation marker for some DN T cells. The CH12 responding population also contains cells with two other phenotypes, Thy-1+, CD4-, CD8-, Ly-5(B220)+, sIgM-, CD3- and Thy-1+, CD4+, CD8-, Ly-5(B220)-, sIgM-, CD3+. The Ly-5(B220)+, CD3- population is no longer found after repeated stimulation. While the relationship between these three populations is unknown, DN T cells can proliferate in the absence of CD4+ or CD8+ cells and therefore their proliferation is not dependent on the presence of other T cells or lymphokines produced by CD4+ or CD8+ T cells. Anti-CD3 immunoprecipitation of CH12-responding cells reveals at least seven different receptor proteins of which five can also be precipitated with an anti-(C gamma 1/C gamma 2) monoclonal antibody. Thus at least three different TCR gamma-delta heterodimers are expressed by CH12-responding T cells. The Thy-1+, CD4-, CD8-, Ly-5(B220)+ cells can provide help to CH12 cells for Ig secretion even in the absence of the nominal antigen for the B lymphoma cells.(ABSTRACT TRUNCATED AT 250 WORDS)