Regulation of murine IgE production in SJA/9 and nude mice. Potentiation of IgE production by recombinant interleukin 4

Serum IgE levels were determined in different strains of mice with enzyme-linked immunosorbent assay by using rat monoclonal anti-murine IgE antibodies in normal and in Nippostrongylus brasiliensis-infected mice. After infection, serum IgE levels were high in BALB/c and CB-20, low in SJL/J and SJA/20 mice, and not detected at all in SJA/9 and nude mice. Surface IgE-positive cells were greatly increased in BALB/c and SJL/J mice after infection, but not in SJA/9 and nude mice. Most surface IgE-positive spleen cells were also surface IgM- and surface IgD-positive. When spleen cells from SJA/9 or nude mice were stimulated in vitro with lipopolysaccharide and recombinant interleukin 4 (formerly B cell-stimulating factor 1), IgE was produced and detected in the supernatants of these cultures. In addition, surface IgE-positive cells could be detected in these cultures. Most of the surface IgE-positive cells were surface IgM- and surface IgD-negative, unlike those seen in the spleens of Nippostrongylus-infected BALB/c and SJL/J mice. These observations show that SJA/9 and nude mice have IgE-producing precursor B cells, and after appropriate stimulation interleukin 4 can induce them to secrete IgE.     

Studies on IgE production in mice

IgE levels after Nippostrongylus brasiliensis infestation were high in all strains of mice examined except in the SJA/9 strain. However, surface IgE levels on B cells were not different from those observed in the SJL strain. Most surface IgE-bearing cells also had surface IgM molecules. We suggest that the surface IgE-bearing cells are precursors of IgE-producing cells. In adoptive transfer, hapten-specific IgE antibody was produced when hapten-primed cells from SJA/9 mice were transferred together with cells from SJL donors infested with N. brasiliensis but not when the cells were from SJA/9 donors. It is suggested that the lack of collaboration between hapten-primed B cells and cells from infested mice is a genetic trait of the SJA/9 strain.     

Attenuation of allergen-induced responses in CCR6-/- mice is dependent upon altered pulmonary T lymphocyte activation

We have established a defect in CCR6-/- mice in response to a cockroach allergen airway challenge characterized by decreased IL-5 production, reduced CD4+ T and B cells as well as decreased eosinophil accumulation. To determine the nature of the defect in CCR6-/- mice T lymphocyte populations from allergen-sensitized wild-type mice were transferred into sensitized CCR6-/- mice. The reconstituted response was characterized by an increase in IL-5 levels, eosinophil accumulation, and serum IgE levels in recipient CCR6-/- mice. Analysis of lymphocytes from draining lymph nodes of CCR6+/+ and CCR6-/- sensitized or challenged mice demonstrated a significant decrease in IL-5 and IL-13 production in CCR6-/- mice. In contrast, the systemic response in allergen-rechallenged spleen cells demonstrated no significant alteration in allergen-induced cytokine production. Transfer of isolated splenic T lymphocytes from sensitized CCR6+/+ mice induced airway hyperresponsiveness in wild-type but not CCR6-/- naive mice, suggesting that T cells alone were not sufficient to induce airway hyperresponsiveness in CCR6-/- mice. Additional analysis demonstrated decreased CD11c+, CD11b+ and CD11c, and B220 subsets of dendritic cells in the lungs of CCR6-/- mice after allergen challenge. Using in vitro cell mixing studies with isolated pulmonary CD4+ T cells and CD11c+ cells from CCR6+/+ or CCR6-/- mice, we demonstrate alterations in both CCR6-/- T cells and CCR6-/- pulmonary APCs to elicit IL-5 responses. Altogether, the defect in CCR6-/- mice appears to be primarily due to an alteration in T cell activation, but also appears to include local pulmonary APC defects.     

The mechanism of a defective IFN-gamma response to bacterial toxins in an atopic dermatitis model, NC/Nga mice, and the therapeutic effect of IFN-gamma, IL-12, or IL-18 on dermatitis

NC/Nga (NC) mice raised under conventional conditions (Conv. NC mice) spontaneously develop dermatitis similar to human atopic dermatitis, whereas NC mice raised under the specific pathogen-free conditions do not develop dermatitis. In the present study, we show that the representative Th1 cytokine, IFN-gamma levels in the sera of NC mice, injected with either staphylococcal enterotoxin B or endotoxin (LPS), to be severalfold lower than those of normal mice. The low IFN-gamma response to staphylococcal enterotoxin B was correlated to the lack of regular Vbeta8(+) T cells and Vbeta8(+) NK T cells, and the low IFN-gamma response to LPS was correlated to an impaired IL-18 production of macrophages. The CD3-stimulated IL-4 production from liver and spleen T cells from Conv. NC mice in vitro was greatly augmented. The serum IL-4 levels of untreated Conv. NC mice also were higher than those of normal mice and specific pathogen-free NC mice. Treatment of Conv. NC mice either with IFN-gamma, IL-12, or IL-18 twice a week from 4 wk of age substantially inhibited the elevation of the serum IgE levels, serum IL-4 levels, and dermatitis, and IL-12 or IL-18 treatment also reduced the in vitro IL-4 production from CD3-stimulated liver T cells. The systemic deficiency in the Th1 response to bacterial stimulation thus leads to a Th2-dominant state and may induce an abnormal cellular immune response in the skin accompanied with an overproduction of IgE and a susceptibility to dermatitis in NC mice.     

IL-3 promotes production of IL-4 by splenic non-B, non-T cells in response to Fc receptor cross-linkage

A spleen cell population that lacks CD3, CD4, CD8, Thy-1, B220, and class II major histocompatibility complex cell-surface markers (non-B, non-T cells) produces IL-4 when cultured in wells coated with IgE. Their production of IL-4 in response to plate-bound (PB)-IgE is strikingly enhanced by IL-3, and in the presence of IL-3, these cells also produce IL-4 in response to PB-IgG2a. The effect of IL-3 is not mimicked by IL-1, IL-2, IL-5, IL-6, IL-7, granulocyte-macrophage CSF (GM-CSF) or IFN-gamma. Non-B, non-T cells cultured with IL-3 for 12 h acquire the capacity to produce enhanced amounts of IL-4 in response to subsequent culture with PB-Ig even if IL-3 is omitted from the second culture. Irradiated cells also respond to IL-3 with enhanced capacity to produce IL-4 to PB-Ig, indicating that cell proliferation is not required for the effect of IL-3. The IL-3 effect can be obtained in vivo; treatment of mice with a total dose 90,000 U of synthetic IL-3 over a 3-day period results in the presence of splenic and peritoneal cavity non-B, non-T cells that produce enhanced amounts of IL-4 in response to PB-Ig. The FcR that mediates the response to PB-IgE appears to be Fc epsilon RI because cells can be sensitized with IgE anti-DNP mAb, washed, cultured for 15 h at 37 degrees C, washed again, and stimulated to produce IL-4 with 0.1 to 1 ng/ml of TNP10-OVA. IL-3 does not appear to mediate its function by increasing the number of Fc epsilon RI because it can exert its effect when cultured with non-B, non-T cells after they have been sensitized with IgE anti-DNP. However, IL-3 pretreatment does affect the signaling process in that non-B, non-T cells sensitized with IgE anti-DNP show strikingly reduced production of IL-4 to concentrations of TNP10-OVA of 100 ng/ml or more whereas cells pretreated with IL-3 show little or no diminution in IL-4 production at concentrations of TNP10-OVA up to 1 microgram/ml.     

Cutting edge: IL-18-transgenic mice: in vivo evidence of a broad role for IL-18 in modulating immune function

IL-18 has been shown to be a strong cofactor for Th1 T cell development. However, we previously demonstrated that when IL-18 was combined with IL-2, there was a synergistic induction of a Th2 cytokine, IL-13, in both T and NK cells. More recently, we and other groups have reported that IL-18 can potentially induce IgE, IgG1, and Th2 cytokine production in murine experimental models. Here, we report on the generation of IL-18-transgenic (Tg) mice in which mature mouse IL-18 cDNA was expressed. CD8+CD44high T cells and macrophages were increased, but B cells were decreased in these mice while serum IgE, IgG1, IL-4, and IFN-gamma levels were significantly increased. Splenic T cells in IL-18 Tg mice produced higher levels of IFN-gamma, IL-4, IL-5, and IL-13 than control wild-type mice. Thus, aberrant expression of IL-18 in vivo results in the increased production of both Th1 and Th2 cytokines.     

A heritable defect in IL-12 signaling in B10.Q/J mice. I. In vitro analysis

B10.Q mice are normally susceptible to the induction of collagen-induced arthritis. We noted that one subline of B10.Q mice, B10.Q/J, was completely resistant to disease induction when immunized with collagen in CFA. B10.Q/J mice have a global defect in the generation of Th1 responses, and Ag-specific T cells derived from this strain failed to produce IFN-gamma. Because T cells from these mice could produce normal amounts of IFN-gamma when activated by IL-12/IL-18-independent stimuli, the defect appeared to be a failure to respond to IL-12. This defect extended to NK cells, which also failed to produce IFN-gamma when stimulated by IL-12. The capacity of NK cells, but not activated T cells, to produce IFN-gamma in response to IL-12 could be partially restored by IL-18. The expression of the IL-12R beta1- and beta2-chains on T cells and NK cells from B10.Q/J mice was normal. However, activated T cells from B10.Q/J mice did not signal normally through the IL-12R and manifested a defect in their capacity to phosphorylate Stat4. This defect was partial in that it could be overcome by increasing both the concentration of IL-12 and the incubation times in the Stat4 phosphorylation assays. Because Stat4 function is apparently intact in B10.Q/J mice, the defect in IL-12 signaling can be localized between the IL-12R complex and Stat4. This subtle abnormality in IL-12 responsiveness results in a profound defect in the generation of Th1 cells and the development of autoimmune disease.     

Basophils amplify type 2 immune responses, but do not serve a protective role, during chronic infection of mice with the filarial nematode Litomosoides sigmodontis

Chronic helminth infections induce a type 2 immune response characterized by eosinophilia, high levels of IgE, and increased T cell production of type 2 cytokines. Because basophils have been shown to be substantial contributors of IL-4 in helminth infections, and because basophils are capable of inducing Th2 differentiation of CD4(+) T cells and IgE isotype switching in B cells, we hypothesized that basophils function to amplify type 2 immune responses in chronic helminth infection. To test this, we evaluated basophil function using the Litomosoides sigmodontis filaria model of chronic helminth infection in BALB/c mice. Time-course studies showed that eosinophilia, parasite Ag-specific CD4(+) T cell production of IL-4 and IL-5 and basophil activation and IL-4 production in response to parasite Ag all peak late (6-8 wk) in the course of L. sigmodontis infection, after parasite-specific IgE has become detectable. Mixed-gender and single-sex worm implantation experiments demonstrated that the relatively late peak of these responses was not dependent on the appearance of circulating microfilariae, but may be due to initial low levels of parasite Ag load and/or habitation of the developing worms in the pleural space. Depletion of basophils throughout the course of L. sigmodontis infection caused significant decreases in total and parasite-specific IgE, eosinophilia, and parasite Ag-driven CD4(+) T cell proliferation and IL-4 production, but did not alter total worm numbers. These results demonstrate that basophils amplify type 2 immune responses, but do not serve a protective role, in chronic infection of mice with the filarial nematode L. sigmodontis.     

IL-4 production by T cells from naive donors. IL-2 is required for IL-4 production

Utilizing a sensitive and selective assay for IL-4, it was shown that lymph node T cells from naive mice could produce small amounts of this lymphokine in response to anti-CD3 antibodies adsorbed to culture dishes. The capacity of these cells to produce IL-4 in response to plate-bound anti-CD3 was substantially enhanced by the addition of IL-2 to the culture and was strikingly inhibited by monoclonal anti-IL-2 antibody. Thus, IL-2 appears to be essential for IL-4 production by anti-CD3 antibody-stimulated T cells from naive mice. The effect of IL-2 was not mediated either by preferential proliferation or survival of precursors of IL-4 producing cells, indicating that IL-2 regulates T cell production of IL-4. IL-4 producing capacity of T cells from naive mice was found mainly among CD4+ T cells. Large T cells produced much more IL-4, on a per cell basis, than did small T cells. In contrast, small T cells appeared to be equal or superior to large T cells in producing IL-2. The superiority of large T cells in IL-4-producing capacity was not accounted for by a lack of an accessory cell population from the small T cells as addition of large spleen cells depleted of both B and T cells did not enhance IL-4 production by small lymph node T cells. These results suggest that the bulk of IL-4 production by T cell populations, from normal mice, in response to anti-CD3 depends upon cells that are already activated and that IL-2 is required for such production.     

Collagen-induced arthritis in mice. Relationship of collagen-specific and total IgE synthesis to disease.

The relationship between production of IgE and collagen-induced arthritis in mice was examined. Collagen-specific IgE was produced as a consequence of immunization of DBA/1 mice with chicken type II collagen emulsified in CFA. We observed a rise in collagen-specific IgE antibody levels at the onset of CIA clinical and histologic signs in DBA/1 mice. This rise in IgE paralleled that of IgG2a anticollagen antibodies, an isotype implicated in the pathogenesis of CIA by other laboratories. The collagen-specific IgE contained in the plasma of mice with CIA could arm basophils for Ag- (collagen) dependent degranulation. Collagen-specific IgE may thus contribute to CIA by promoting mast cell degranulation in the synovia of susceptible mice immunized with chick type II collagen; but, further work is required to establish such a role for IgE in CIA. However, genetic differences in disease susceptibility could not be accounted for by quantitative differences in collagen-specific IgE production. Further, comparable levels of IgE anticollagen antibodies were observed in animals with active CIA and after spontaneous remission, thereby confirming that the presence of such antibodies is insufficient for disease. Total IgE levels peaked just before spontaneous remission indicating active production of IL-4. IL-4 was administered to animals with CIA to determine if this lymphokine could be involved in the remission process. IL-4 facilitated remission of CIA. Enhanced total IgE production may thus be a marker for activation of Th2 cells that produce lymphokines such as IL-4 and IL-10, factors that may be involved in the spontaneous remission process.     

CD40 stimulation provides an IFN-gamma-independent and IL-4-dependent differentiation signal directly to human B cells for IgE production.

IgE induction from human cells has generally been considered to be T cell dependent and to require at least two signals: IL-4 stimulation and T cell/B cell interaction. In the present study we report a human system of T cell-independent IgE production from highly purified B cells. When human cells were co-stimulated with a mAb directed against CD40 (mAb G28-5), there was induction of IgE secretion from purified blood and tonsil B cells as well as unfractionated lymphocytes. Anti-CD40 alone failed to induce IgE from blood mononuclear cells or purified B cells. The effect of the combination of anti-CD40 and IL-4 on IgE production was very IgE isotype specific as IgG, IgM, and IgA were not increased. Furthermore, anti-CD40 with IL-5 or PWM did not co-stimulate IgG, IgM, or IgA and in fact strongly inhibited PWM-stimulated IgG, IgM and IgA production from blood or tonsil cells. IgE synthesis induced by anti-CD40 plus IL-4 was IFN-gamma independent as is the in vivo production of IgE in humans; the doses of IFN-gamma that profoundly suppressed IgG synthesis induced by IL-4, or IL-4 plus IL-6, had no inhibitory effect on anti-CD40-induced IgE production. Anti-CD23 and anti-IL-6 also could not block anti-CD40 plus IL-4-induced IgE production, but anti-IL-4 totally blocked their effect. IgE production via CD40 was not due to IL-5, IL-6 or nerve growth factor as none of these synergized with IL-4 to induce IgE synthesis by purified B cells. Finally, we observed that CD40 stimulation alone could enhance IgE production from in vivo-driven IgE-producing cells from patients with very high IgE levels; cells that did not increase IgE production in response to IL-4. Taken together, our data suggest that the signals delivered for IgE production by IL-4 and CD40 stimulation may mimic the pathway for IgE production seen in vivo in human allergic disease.     

Parallel regulation of IL-4 and IL-5 in human helminth infections.

To investigate the relationship between cytokine production and the increased levels of serum IgE and peripheral eosinophilia commonly accompanying human helminth infections, we studied the ability of PBMC of normal (N1) (n = 18) and eosinophilic individuals with helminth infections (H1) (n = 9) to produce IL-3, IL-4, IL-5, granulocyte-macrophage-CSF, and IFN-gamma in vitro after stimulation with PMA (50 ng/ml) and ionomycin (1 microgram/ml). The two groups differed in both the levels of serum IgE and eosinophilia. For mitogen-induced production of granulocyte-macrophage-CSF and IFN-gamma, there was no difference in cytokine production between the two groups. In marked contrast, supernatants from PBMC of infected individuals had significantly higher levels of IL-4 (mean = 213 pg/ml for N1 and 944 pg/ml for H1, p less than 0.02), IL-5 (mean = 180 pg/ml for N1 and 1118 pg/ml for HL, p less than 0.001), and IL-3 (mean = 13900 pg/ml for N1, 28029 pg/ml for H1, p less than 0.05). In addition, helminth-infected patients had approximately 5-fold greater numbers of T cells capable of producing IL-5 and 2.5-fold greater frequency of IL-4-secreting cells than did normal individuals; GM-CSF- and IFN-gamma-producing T cell numbers were not significantly different in the two groups. IL-3-producing cell frequencies could not be evaluated by this method. There was a direct correlation between IL-4 production and IL-5 production at the level of both protein production and frequency of T cells capable of producing these cytokines. These data indicate that individuals with reactive eosinophilia and elevated serum IgE have an expanded population of lymphocytes producing IL-4 and IL-5 and the association of the two suggests that the regulation of IL-4 and IL-5 may be linked.     

IL-4 suppression of in vivo T cell activation and antibody production

Injection of mice with a foreign anti-IgD Ab stimulates B and T cell activation that results in large cytokine and Ab responses. Because most anti-IgD-activated B cells die before they can be stimulated by activated T cells, and because IL-4 prolongs the survival of B cells cultured with anti-Ig, we hypothesized that treatment with IL-4 at the time of anti-IgD Ab injection would decrease B cell death and enhance anti-IgD-induced Ab responses. Instead, IL-4 treatment before or along with anti-IgD Ab suppressed IgE and IgG1 responses, whereas IL-4 injected after anti-IgD enhanced IgE responses. The suppressive effect of early IL-4 treatment on the Ab response to anti-IgD was associated with a rapid, short-lived increase in IFN-gamma gene expression but decreased CD4+ T cell activation and decreased or delayed T cell production of other cytokines. We examined the possibilities that IL-4 stimulation of IFN-gamma production, suppression of IL-1 or IL-2 production, or induction of TNF-alpha or Fas-mediated apoptosis could account for IL-4's suppressive effect. The suppressive effect of IL-4 was not reversed by IL-1, IL-2, or anti-TNF-alpha or anti-IFN-gamma mAb treatment, or mimicked by treatment with anti-IL-2Ralpha (CD25) and anti-IL-2Rbeta (CD122) mAbs. Early IL-4 treatment failed to inhibit anti-IgD-induced Ab production in Fas-defective lpr mice; however, the poor responsiveness of lpr mice to anti-IgD made this result difficult to interpret. These observations indicate that exposure to IL-4, while T cells are first being activated by Ag presentation, can inhibit T cells activation or promote deletion of responding CD4+ T cells.     

Impaired protection against Trichinella spiralis in mice with high levels of IgE

Helminth infection induces production of a large amount of immunoglobulin E (IgE) to nonhelminth antigens. Although such “irrelevant” IgE is a major proportion of total IgE in the host, its biological significance remains unclear. Therefore, I examined protective activity against Trichinella spiralis in mice with high levels of IgE by repeated injections of anti-dansyl IgE monoclonal antibody or Nippostrongylus brasiliensis infection. Injected anti-dansyl IgE occupied IgE receptors on mast cells in naive mice. Protective activity against T. spiralis, determined with number of muscle larvae 5 weeks after infection, was impaired in mice treated with anti-dansyl IgE. The impaired protection was found in mice treated with anti-dansy IgE 7 and 14 days after infection, but not 21 and 28 days after infection, indicating that IgE-dependent protection operates at an early stage after infection. In the next experiments, mice were infected with N. brasiliensis 4 weeks before T. spiralis infection to obtain high levels of IgE. The protective activity against T. spiralis was decreased by N. brasiliensis infection. On the other hand, protection against T. spiralis was comparable in IgE-deficient SJA/9 mice and in anti-IgE-treated BALB/c mice with or without N. brasiliensis infection, suggesting that impairment of protection is dependent on IgE. These results indicate that the high levels of irrelevant IgE are beneficial for helminths and, alternatively, that anti-helminth IgE antibodies are protective for hosts. In addition, the impaired protection was found in IgE high-responder mice but not in low-responder mice, suggesting that protection against T. spiralis is controlled by IgE responsiveness in the host.     

Both naive and memory T cells can provide help for human IgE production, but with different cytokine requirements.

Most in vitro systems for the induction of IgE production by human B cells require both IL-4 and the presence of T cells. Little is known about the mechanism of T cell help or the ability of different T cell subsets to provide this helper activity. In the present study we demonstrate that, in the presence of exogenous IL-4, anti-CD3 stimulated naive T cells (CD4+CD45RA+) are potent helper cells for human IgE production. In their presence, as little as 750 autologous B cells can produce up to 100 ng/ml IgE. This response was found over a broad range of anti-CD3 concentrations. IgE helper activity by naive T cells was inhibited by IL-2. Under all conditions tested, naive T cells were unable to provide help for IgM production. This is in contrast to activated memory T cells (CD4+CD45RO+), which are very efficient helper cells for IgM or IgE production, provided that IL-2 or IL-2 plus IL-4 are present respectively.     

Differential production of Th1- and Th2-derived cytokines does not determine the genetically controlled or vaccine-induced rate of cure in murine visceral leishmaniasis

Recent studies with models of cutaneous leishmaniasis have provoked much interest in the role of CD4+ T cell subsets in determining the outcome of infectious disease. In Leishmania major infections, cure vs progressive disease correlates with the expansion of Th1-like or Th2-like CD4+ populations, respectively. We have investigated whether similar responses are associated with the differential patterns of infection seen in models of visceral leishmaniasis, caused by L. donovani. Splenic lymphocytes from infected Lsh congenic C57BL/10 (Lshs;H-2b) and B10.L-Lshr (Lshr;H-2b) mice and MHC congenic non-curing B10.D2/n (Lshs;H-2d) mice were examined for the production of cytokines representative of these CD4+ populations (IL-2, IL-3, IL-4, IL-5, and IFN-gamma). In all three strains examined, there was no evidence for the production of Th2-restricted cytokines. In addition, levels of serum IgE were depressed during the early phase of infection, indicative of in vivo IFN-gamma production. In the non-curing B10.D2/n strain, late phase of infection was associated with the decreased ability to produce cytokines in response to Ag and not with the production of IL-4 or IL-5 in response to Ag or mitogen. Serum IgE levels were also not raised above levels seen in uninfected controls. C57BL/10 mice were vaccinated with SDS-PAGE fractionated amastigote Ag bound to nitrocellulose and cytokine levels determined at various times after infection. The protocol used for vaccination was able to induce significant modulation of the course of infection in this strain and it was clear that IFN-gamma production in vitro provided an excellent correlate of rate of cure. Occasional individuals produced low levels of IL-5 in culture in response to parasite Ag, but this did not correlate with disease progression. Together, these data suggest that over-expansion of Th2-type cells and production of their specific cytokines (IL-4 and IL-5) is not a contributing factor to the variable long term course of L. donovani infection in these strains of mice.     

The role of OX40 ligand interactions in the development of the Th2 response to the gastrointestinal nematode parasite Heligmosomoides polygyrus

In these studies, we examined the effects of OX40 ligand (OX40L) deficiency on the development of Th2 cells during the Th2 immune response to the intestinal nematode parasite Heligmosomoides polygyrus. Elevations in IL-4 production and total and Ag-specific serum IgE levels were partially inhibited during both the primary and memory immune responses to H. polygyrus in OX40L(-/-) mice. The host-protective memory response was compromised in OX40L(-/-) mice, as decreased worm expulsion and increased egg production were observed compared with H. polygyrus-inoculated OX40L(+/+) mice. To further examine the nature of the IL-4 defect during priming, adoptively transferred DO11.10 T cells were analyzed in the context of the H. polygyrus response. Although Ag-specific T cell IL-4 production was reduced in the OX40L(-/-) mice following immunization with OVA peptide plus H. polygyrus, Ag-specific T cell expansion, cell cycle progression, CXCR5 expression, and migration were comparable between OX40L(+/+) and OX40L(-/-) mice inoculated with OVA and H. polygyrus. These studies suggest an important role for OX40/OX40L interactions in specifically promoting IL-4 production, as well as associated IgE elevations, in Th2 responses to H. polygyrus. However, OX40L interactions were not required for serum IgG1 elevations, increases in germinal center formation, and Ag-specific Th2 cell expansion and migration to the B cell zone.