Independent and Opposing Roles For Btk and Lyn in B and Myeloid Signaling Pathways

Btk^−/− (10) and lyn^−/− mice (14), each on a mixed C57B/6 × 129/Sv genetic background, were crossed to generate Btk^+/−lyn^+/− F1 progeny. These F1 animals were mated resulting in wild-type (wt), Btk-deficient, Lyn-deficient, and Btk/Lyn-deficient progeny. Genotypes were determined by Southern blot (10) or PCR (14) analysis of tail biopsy DNA as described. For the analysis of limiting Btk dosage in Fig. 5, crosses were performed as above starting with Btk^−/− mice carrying an Ig heavy chain enhancer/ promoter–driven Btk transgene expressing ∼25% of endogenous Btk levels in B cells (34). The resulting progeny were on a mixed C57B/6 × 129/Sv × Balb/c background. The presence of the Btk transgene was determined by Southern blot as previously described (34).

Single cell suspensions from spleen, bone marrow, or peritoneal wash were depleted of red blood cells and stained with combinations of the following antibodies (from PharMingen, San Diego, CA, unless otherwise noted): PE-conjugated (PE) anti-B220 (RA2-6B2); FITC-conjugated (FITC) anti-IgM (R6-60.2); anti-IgD^a (AMS 9.1) PE; anti-IgD^b (217-170) PE; anti-CD5 (53-7.3) FITC; anti–Thy-1.2 (53-2.1) FITC; anti-CD4 (H129.19) PE; anti-CD8 (53-6.7) FITC; anti-Ter119 (TER-119) PE; anti-Gr1 (RB6-8C5) PE; and anti-Mac1 (M1/ 70HL) FITC (Boehringer Mannheim Corp., Indianapolis, IN). Data was acquired on a FACScan^® (Becton Dickinson, San Jose, CA) and analyzed using LYSIS II software. Live cells were gated based on forward and side scatter. A gate characteristic of lymphoid cells based on forward and side scatter was used for acquisition of peritoneal cells.

Mice were fed 0.25 mg/ml bromodeoxyuridine (BrdU) and 2.5% glucose in their drinking water continuously for up to 15 d. Single cell suspensions of spleens and peripheral blood were depleted of red blood cells, stained with anti-BrdU FITC (Becton Dickinson) and anti-B220 (RA3-6B2) PE (PharMingen) as previously described (36), and analyzed as above.

Total splenocytes were depleted of red blood cells and plated in RPMI with 10% heat-inactivated FCS at 10⁶/ml. Where indicated, goat anti–mouse IgM F(ab′)₂ fragments (Jackson ImmunoResearch Labs., West Grove, PA) were added at either 2 or 20 μg/ml. At 24 h, BrdU (Sigma Chemical Co., St. Louis, MO) was added to a final concentration of 10 μM. Cells were harvested at 48 h and FACS^® analysis was performed as above.

B220⁺ spleen cells were isolated using the Minimacs magnetic bead system (Miltenyi Biotec, Inc., Auburn, CA) according to the manufacturer's instructions. Single cell suspensions were depleted of red blood cells before incubation with magnetic beads. B cell–enriched populations were >90% B220⁺ by FACS^® analysis. B220⁺ splenic B cells were seeded into 96-well plates at 5 × 10⁵/ml in RPMI with 10% heat-inactivated FCS. Where indicated, cells were incubated for 60 h with 2 or 20 μg/ml goat anti–mouse IgM F(ab′)₂ fragments (Jackson ImmunoResearch Labs.). 1 μCi [³H]thymidine (NEN Life Science Products, Boston, MA) was added per well for the final 12-18 h. Cells were harvested and counted on a scintillation counter.

Plates were coated with 2 μg/ml goat anti–mouse Ig (Southern Biotechnology Associates, Huntington, AL) and blocked with 1% BSA in borate-buffered saline (BBS). Serum or Ig standards (mouse IgM, IgG₁, IgG_2a, IgG_2b, IgG₃, and IgA; Sigma Chemical Co.) were diluted serially into BBS and added to wells in duplicate. Plates were washed, incubated with secondary antibody (goat anti–mouse IgM, IgG₁, IgG_2a, IgG_2b, IgG₃, or IgA-alkaline phosphatase [ALPH], Southern Biotechnology Associates) diluted 1:500 in BBS/0.05% Tween 20/1% BSA and developed with an ALPH substrate kit (Bio-Rad, Hercules, CA). OD₄₀₅ was read on a Vmax kinetic microplate reader (Molecular Devices Corp., Sunnyvale, CA).

Mice were immunized intraperitoneally with 100 μg of keyhole limpet hemocyanine (KLH) (Sigma Chemical Co.) in incomplete Freund's adjuvant (GIBCO BRL, Gaithersburg, MD), boosted on day 21 with 50 μg of KLH in PBS, and bled on day 28. ELISAs were performed as above with the following modifications: plates were coated with 8 μg/ ml of KLH, and secondary antibodies were goat anti–mouse IgM-ALPH and goat anti–mouse IgG₁-ALPH.

Mice were immunized intraperitoneally with 10 μg of 2,4,6-trinitrophenyl (TNP)-Ficoll (a gift of Dr. John Inman, National Institutes of Health, Bethesda, MD) and bled 6 d later. ELISA was performed as above with the following modifications: plates were coated with 25 μg/ml of TNP-BSA in PBS, and serum and secondary antibody (goat anti–mouse IgM-ALPH) were diluted into PBS/0.1% BSA, and 0.05% Tween 20.

Unimmunized mice were bled at 16–20 wk of age. Serum antibodies to dsDNA were measured in triplicate by ELISA as previously described (14).

Unimmunized mice were bled at 16–20 wk of age. Serum antibodies to nuclear antigens were measured by immunofluorescence as previously described (14).

We bred Btk^+/−lyn^+/− mice to generate wild-type, Btk^−/−, lyn^−/−, and Btk^−/−lyn^−/− progeny in order to examine the interaction of these kinases in B cell development and function. All genotypes occurred in the predicted Mendelian frequencies, indicating that embryonic development occurs normally in the absence of both Btk and Lyn. Btk^−/−lyn^−/− mice were healthy and fertile. lyn^+/+ and lyn^+/− animals were indistinguishable in the assays described below and were therefore used interchangeably. All mice were on a mixed C57B/6 × 129/Sv background. Although the phenotype of Btk deficiency has been described to differ according to genetic background (10, 37, 38), little variation was observed within each group of mice analyzed here.

To understand whether Lyn and Btk are components of a common signaling pathway or function independently in B cells, we examined B cell development in the absence of Btk alone, Lyn alone, or both Btk and Lyn. Both the frequency and number of splenic B cells in 8-wk-old Btk^−/−lyn^−/− mice was reduced two- to fourfold relative to either Btk^−/− or lyn^−/− animals and four- to sixfold compared with wild-type controls (Fig. 1,A and Table 1). The remaining B cells had an immature IgM^hiIgD^lo phenotype similar to Btk^−/− B cells (Fig. 1,B). This decrease was specific to the B lineage. No significant difference in myeloid cell numbers was observed in the spleens of young mice, and T cell numbers were reduced less than twofold compared with wild-type controls (Table 1). The frequency of conventional (B220⁺CD5⁻) B cells in the peritoneum was also diminished in Btk^−/−lyn^−/− mice compared with single knockouts alone (Fig. 1,D, Table 1).

Either a block in the bone marrow phase of B cell development or reduced half life of peripheral B cells could be responsible for the small number of splenic B cells in Btk^−/− lyn^−/− mice. Both the B220⁺IgM⁻ population, consisting of proB and preB cells, and B220^intIgM⁺ immature B cells were present in similar frequencies in bone marrow from wild-type, Btk^−/−, lyn^−/−, and Btk^−/−lyn^−/− mice (Fig. 1,C and Table 1). Therefore, the antigen-independent phase of B cell development proceeds normally in the absence of both Btk and Lyn. B220^hiIgM⁺ recirculating B cells were three- to fourfold less frequent in both lyn^−/− (as previously reported in references 12 and 13) and Btk^−/−lyn^−/− mice than in wild-type controls (Fig. 1,C and Table 1). This is consistent with the significant reduction in mature B cell numbers in the periphery.

The turnover rate of the peripheral B cell population was assessed with in vivo BrdU labeling since an early developmental block did not account for the reduced number of splenic B cells in Btk^−/−lyn^−/− mice. Short-lived B cells in the periphery are replaced by newly generated cells that have incorporated BrdU, whereas long-lived resting cells remain unlabeled. Both splenic and peripheral blood B cells from Btk^−/−lyn^−/− mice turned over faster than wild-type, Btk^−/−, or lyn^−/− B cells (Fig. 2). Greater than 70% of Btk^−/− lyn^−/− B220⁺ cells were labeled with BrdU after 8 d, compared with ∼35% in wild-type mice and 50% in mice lacking either Btk or Lyn alone (Fig. 2). These observations indicate that the reduced number of B cells in mice lacking both Btk and Lyn is due to poor survival in the periphery rather than decreased production of B cells in the bone marrow.

To determine whether the reduced number of peripheral B cells in Btk^−/−lyn^−/− mice was accompanied by alterations in serum Ig levels, the amount of each Ig isotype present in the serum of 6–8-wk-old wild-type, Btk^−/−, lyn^−/−, and Btk^−/−lyn^−/− mice was measured (Fig. 3). Btk^−/− mice had a >10-fold reduction in IgM and IgG₃ levels, and a 2–4-fold decrease in IgG_2a and IgG_2b. IgM and IgA levels were slightly elevated in lyn^−/− mice compared with wild-type controls, although the 5–10-fold increase in these isotypes described in lyn^−/− mice in two previous studies (12, 13) was not observed. This may be due to differences in the age of the animals at the time of analysis, since serum IgM levels in lyn^−/− mice increase with time (12, 13). Btk^−/−lyn^−/− mice resembled Btk^−/− mice with respect to IgM, IgG₃, and IgG_2a levels, and had increased IgA similar to lyn^−/− animals. The amount of IgG_2b and IgG₁ was reduced in Btk^−/−lyn^−/− mice compared with all other genotypes, consistent with the low numbers of peripheral B cells in these animals.

The dramatic effect of double Btk and Lyn deficiency on B cell numbers suggested that B cell function may also be more severely inhibited in the absence of both kinases. The ability of Btk^−/−lyn^−/− mice to mount an immune response to the T cell–independent antigen TNP-Ficoll and the T cell–dependent antigen KLH was assessed. Btk^−/−lyn^−/− mice failed to produce antibodies against TNP-Ficoll (Fig. 4,A), as expected, since this response requires Btk. Mice lacking either Btk or Lyn alone have relatively normal secondary responses to T cell–dependent antigens such as KLH (Fig. 4, B and C). Strikingly, anti-KLH IgM was not detectable in Btk^−/−lyn^−/− mice (Fig. 4,B). IgG₁ titers against KLH were measurable but lower on average than in mice of all other genotypes (Fig. 4 C). Although the impaired IgG₁ response could be attributed to low B cell numbers, the reduction in anti-KLH IgM was significant even when corrected for cell number. This indicates that Btk and Lyn are redundant for the production of IgM against T-dependent antigens.

Several aspects of B cell function differ significantly between lyn^−/− and Btk^−/− mice despite the similar reduction in conventional B cell numbers. Most strikingly, lyn^−/− B cells are hyperresponsive to BCR cross-linking (14, 20), whereas Btk^−/− cells fail to proliferate in response to anti-IgM (10). A transgenic model, in which varying doses of a Btk transgene driven by the Ig heavy chain enhancer and promoter are expressed on an xid or Btk^−/− background, has been used to demonstrate that Btk dosage is limiting for response to BCR cross-linking (34). B cells expressing low levels of transgenic Btk and no endogenous Btk (Btk^lo) develop normally but have reduced sensitivity to BCR cross-linking (34). This system was used to determine whether the net effect of Lyn is to activate or inhibit Btk function. Low doses of Btk should be less effective in mediating mitogenic response to BCR engagement on a lyn^−/− background if Lyn is an essential upstream activator of Btk during signal initiation. Alternatively, the ability of limiting amounts of Btk to transmit BCR signals should be enhanced in the absence of Lyn if the negative role of Lyn predominates.

A Btk transgene expressing ∼25% of endogenous Btk levels in splenic B cells (34) was crossed onto both Btk^−/− (Btk^lo) and Btk^−/−lyn^−/− (Btk^lo lyn^−/−) backgrounds. The frequency and absolute number of B220⁺ and IgM^loIgD^hi cells were increased two- to threefold in Btk^lo spleens relative to Btk^−/− spleens in both the presence and absence of Lyn (reference 34 and data not shown), indicating that the transgene restored Btk-dependent signals for maintenance of B cell numbers. Btk^lo B cells were less sensitive to BCR engagement than were wild-type B cells (reference 34 and Fig. 5,A). In contrast, the BCR-induced proliferative response of Btk^lo lyn^−/− B cells was indistinguishable from that of lyn^−/− B cells even at low doses of anti-IgM (Fig. 5,A). Btk^−/−lyn^−/− B cells failed to proliferate upon anti-IgM stimulation (Fig. 5,A). This was not due to altered T or myeloid cell function, as the same result was obtained with purified B cells (Fig. 5,B) and total splenocytes (Fig. 5 A). Lyn deficiency therefore enhances Btk-dependent signaling by the BCR but cannot bypass a requirement for Btk. These observations suggest that Lyn has a net inhibitory effect on Btk-dependent BCR signaling pathways.

The development of autoimmunity in aged lyn^−/− mice is another feature that distinguishes them from Btk^−/− and xid mice (12–14). The xid mutation has been shown to prevent autoantibody production in both NZB×NZW mice (39, 40) and me mice (41). Although six out of six lyn^−/− mice older than 16 wk developed IgM and IgG antibodies against both dsDNA and nuclear antigens, no Btk^−/−lyn^−/− animals of similar age displayed signs of autoimmunity (Table 2). Peritoneal B1 cells are believed to be a major source of anti-self antibodies in autoimmune strains of mice (42, 43). Btk^−/−lyn^−/− mice, like Btk^−/− mice, have a reduced frequency of B220⁺CD5⁺ cells in the peritoneum relative to wild-type or lyn^−/− animals (Fig. 1,D, Table 1).

Splenomegaly resulting from extramedullary hematopoiesis occurs after 14 wk of age in the absence of Lyn (12–14). No myeloid phenotype has been described in xid or Btk^−/− mice or in XLA patients, although a partial defect in FcεRI-mediated cytokine induction has been reported in xid and Btk^−/− mast cells (44). Consistent with these observations, the increased frequency of myeloid and erythroid cells characteristic of lyn^−/− spleens was also observed in old Btk^−/−lyn^−/− mice (Fig. 6,B). Surprisingly, splenomegaly did not occur in Btk^−/−lyn^−/− mice. Spleens of 10–11-mo-old Btk^−/−lyn^−/− mice were four- to fivefold smaller by both weight (0.208 ± 0.042 g vs. 1.1 ± 0.4 g, n = 2) and cell count (3.74 × 10⁷ ± 2.5 × 10⁷, n = 5, vs. 1.6 × 10⁸ ± 0.72 × 10⁸, n = 6, nucleated cells) than those of age-matched lyn^−/− mice (Fig. 6 A). These results suggest that although extramedullary hematopoiesis does not require Btk, the subsequent expansion of myeloid and erythroid elements in lyn^−/− mice is Btk dependent.

A large body of evidence derived from biochemical and tissue culture–based experiments demonstrates that activation of Btk requires Src family kinases (22–28). However, the increased severity of the B cell phenotype in Btk^−/−lyn^−/− mice compared with lyn^−/− mice alone indicates that Btk can still transmit signals in the absence of Lyn. In fact, the ability of limiting doses of Btk to signal is enhanced in lyn^−/− B cells. This suggests either that Lyn is redundant with other Src family kinases for the activation of Btk in mature B cells or that Btk can be activated by Src family–independent mechanisms. Although activation of Btk is critically dependent on Src kinases in a fibroblast model (22, 23, 26), additional regulatory mechanisms cannot be ruled out in the B lineage.

Although Lyn may contribute to the activation of Btk in concert with other Src kinases, it also plays a Btk-independent role in the maintenance of the peripheral B cell population. lyn^−/− mice have fewer B cells than did wild-type littermates (12–14). This has been suggested to result from increased negative selection of B cells that are hypersensitive to BCR cross-linking (14). Diminished B cell numbers in Btk^−/−lyn^−/− mice could be explained by the additive effect of reduced positive selection in the absence of Btk and increased negative selection in the absence of Lyn. However, this is unlikely since lyn^−/− B cells do not respond to BCR engagement when they also lack Btk. Lyn is therefore likely to transmit a positive signal for B cell survival that is independent of Btk.

B cells from both Btk^−/− and lyn^−/− mice have an increased turnover rate relative to wild-type B cells. The number of long-lived B cells is even further reduced in the absence of both Btk and Lyn. This could be secondary to a block in maturation as the long-lived B cell pool consists predominantly of mature B cells (45). However, this possibility is unlikely as Btk^−/− and Btk^−/−lyn^−/− mice have a similar developmental block at the IgM^hiIgD^hi to IgM^loIgD^hi transition. The reduced half-life of Btk^−/−lyn^−/− B cells could also be a result of impaired positive BCR signaling due to the combined effects of Lyn deficiency on signal initiation and Btk deficiency on signal transmission. The BCR is required for survival of B cells in the periphery (46), and deletion of the cytoplasmic tail of Igα in mice results in a B cell phenotype (4) similar to that of Btk^−/−lyn^−/− mice. Independent roles for Btk and Lyn in BCR signaling are also supported by the recent demonstration that the Btk/Tec family kinase Itk and the Src family kinase Fyn have independent functions in TCR signaling (47). Alternatively, Btk and Lyn may be redundant for CD40 signaling. Btk^−/− lyn^−/− mice resemble mice deficient in both Btk and CD40 (48, 49). The impaired response to T cell–dependent antigens would also be explained by failure to transmit CD40 signals (50–52). Finally, defects in homing of Btk^−/− lyn^−/− B cells to the proper compartments in secondary lymphoid organs could contribute to their poor survival (53).

Mice lacking the three Src family kinases Lyn, Blk, and Fyn have a block in development at the proB to preB transition (Tarakhovsky, A., personal communication) similar to that observed in Ig heavy chain– (5) or surrogate light chain–deficient (6) mice. These combined results imply that Src family kinases are redundant for the transmission of preB receptor signals. B lymphopoiesis is also blocked at the preB stage in XLA patients (33), suggesting that Btk is an essential substrate of Src family kinases in human preB receptor signaling. In contrast, the antigen-independent phase of B cell development is normal in Btk^−/− mice even in the absence of Lyn. The critical target of Src family kinases in murine preB cells is probably Syk rather than Btk since the proB to preB transition is impaired in syk^−/− mice (7, 8).

Hypersensitivity to BCR cross-linking in lyn^−/− B cells indicates that Lyn plays a critical role in the negative regulation of BCR signaling. Both the failure of Btk^−/− lyn^−/− B cells to proliferate in response to anti-IgM and the observation that Btk is no longer limiting for response to BCR engagement in the absence of Lyn suggest that Lyn downregulates Btk-dependent signaling pathways. The ability of Btk to promote depletion of intracellular calcium stores in response to BCR cross-linking is prevented by FcγRIIb signaling (27). This inhibition may be mediated by Lyn since FcγRIIb function is partially impaired in lyn^−/− B cells (14). The negative regulatory role of Lyn is not limited to Btk-dependent pathways, as BCR-induced activation of the classical mitogen-activated protein kinase (MAPK) pathway does not require Btk (54, 55) but is enhanced in lyn^−/− B cells (14).

The transgenic mice expressing low levels of Btk (34) are shown here (Fig. 5 A) to be a useful sensitized system with which to identify negative regulatory components of Btk signaling pathways. Similarly, molecules that contribute positively to Btk signaling could be defined by mutations that further impair BCR signaling in Btk^lo mice. It will be interesting to determine the effect of mutations in the remaining Src family kinases, BCR signal threshold modulators, and other B cell signaling molecules on the transmission of BCR signals by limiting amounts of Btk.

A distinguishing characteristic of older lyn^−/− mice is the development of splenomegaly due to extramedullary hematopoiesis (12– 14). Surprisingly, although splenomegaly did not occur in old Btk^−/−lyn^−/− mice, these animals had a similar increase in the frequency of myeloid and erythroid cells as lyn^−/− mice. This suggests that the splenomegaly in lyn^−/− mice is caused by two separate defects. The first, in which the frequency of splenic myeloid and erythroid cells is increased, is independent of Btk. This phase could result from either a shift in the site of hematopoiesis to the spleen or simply “space filling” (56) secondary to a reduction in the number of lymphoid cells. Myeloid and erythroid elements that are present in lyn^−/− spleens then expand in a Btk-dependent manner. Lyn deficiency may render myeloid cells hypersensitive to cytokines, analogous to the reduction of BCR signaling thresholds in B cells. This enhanced response may be attenuated in the absence of Btk. Btk has been implicated as a component of the IL-5 (25, 57, 58), IL-6 (59), and IL-10 (60) cytokine pathways in B cells. However, no alterations in myeloid or erythroid cell development have been reported in xid mice, Btk^−/− mice, or XLA patients except for some defects in FcεRI signaling in mast cells (44).

Btk may serve in a general capacity to regulate mitogenic responses and cell survival. These functions would normally be observed only in B cells because of redundant signaling pathways in other lineages. Variation in genetic context or the mutation of other signaling molecules on a Btk^−/− background may reveal additional roles for Btk in the development and function of hematopoietic cells.

We thank Kim Palmer, Fiona Willis, and Prim Kanchanastit for excellent technical assistance and Vivien Chan for technical advice. We are grateful to Dr. John Inman for his gift of TNP-Ficoll. We also thank Julia Shimaoka and Jamie White for assistance with manuscript preparation.

ALPH

alkaline phosphatase

BBS

borate-buffered saline

BCR

B cell antigen receptor

BrdU

bromodeoxyuridine

Btk

Bruton's tyrosine kinase

Btk^lo

mice lacking the endogenous Btk gene and expressing 25% of endogenous levels of Btk in B cells from an Ig heavy chain enhancer/promoter–driven transgene

KLH

keyhole limpet hemocyanin

TNP

2,4,6-trinitrophenyl

me

motheaten

wt

wild-type

xid

X-linked immunodeficiency

XLA

X-linked agammaglobulinemia