IL-7–Dependent B Lymphocytes Are Essential for the Anti-polysaccharide Response and Protective Immunity to Streptococcus pneumoniae

T cell-independent (TI) Ags can be classified into two types. Bacterial LPS, a TI type 1 (TI-1) Ag, induces polyclonal activation of B cells at high concentrations, primarily because of their mitogenic activity. TI-2 Ags, however, are characterized by the presence of repetitive epitopes that can efficiently cross-link BCRs and induce activation of Ag-specific B cells (1). Capsular polysaccharides on a variety of bacterial pathogens, such as Streptococcus pneumoniae, Neisseria meningitidis, and Bacillus anthrasis, are classic examples of TI-2 Ags (2, 3). Numerous studies have also been performed using the model TI-2 Ag 4-hydroxy-(3-nitrophenyl)-acetyl (NP) hapten conjugated to Ficoll, a synthetic polysaccharide (4–6).

TI-2 Ags are capable of inducing robust Ab responses in adults; however, they are poorly immunogenic in young children. Similar to young children, young mice under the age of 4 wk are impaired in TI-2 responses (7). Possible factors that may account for the impaired TI responses in the young include defects in B cell activation and the lack of appropriate B cell subsets (8–16).

Mature murine B cells are developmentally, phenotypically, and functionally heterogeneous. They are divided into follicular (FO), marginal zone (MZ), B1a, and B1b subsets. FO B cells, also referred to as B2 cells, mount Ab responses to T cell-dependent (TD) Ags, whereas MZ B, B1a, and B1b cells contribute to TI Ab responses (14, 17–19). Upon immunization with NP-Ficoll, pyk-2^−/− mice, deficient in MZ B cells but not B1 cells, generate IgM and IgG3 Ab responses >2 orders of magnitude higher than preimmune levels (20), indicating that a robust TI response can occur in the absence of MZ B cells. CD19^−/− mice, sufficient in the B1b but not the B1a subset, can generate protective Ab responses to pneumococcal polysaccharide (PPS), suggesting that B1b cells play an important role in responding to TI Ags (21). Indeed, reconstitution of immunodeficient Rag1^−/− mice with various B cell subsets demonstrated that B1b cells generate the majority of the IgM and IgG3 responses to PPS type 3 (PPS3) and NP-Ficoll (5, 21).

Although a variety of Ags that drive the TI B1b cell response have been identified to date (5, 19, 21–23), the developmental origin of Ag-specific B1b cells is unknown. The ontogeny of B cells early and late in life originates from fetal liver and bone marrow lymphopoeisis, respectively (24). Analysis of knockout mice revealed that IL-7 and Flt3 ligand (Flt3L, also known as Flk2L) contribute to B lymphopoeisis (25, 26). IL-7 signaling is mediated by a heterodimeric receptor composed of an IL-7Rα-chain and the common cytokine receptor γ_c chain. IL-7 is essential for B cell development in the bone marrow of adult mice (27–29). The presence of large numbers of B1 and MZ B cells in IL-7^−/− mice indicate that these B cell subsets can be generated by IL-7–independent B lymphopoiesis. This IL-7–independent B cell development is likely to be mediated by the action of Flt3L on primitive B cell progenitors in fetal liver (30), because IL-7–deficient mice have an arrest in adult bone marrow B lymphopoiesis and mice lacking both Flt3L and IL-7 have a complete deficiency of B lymphopoiesis (26). Although B cells and Abs are present in humans with mutations in IL-7Rα, the ability of those B cells to respond to specific Ags is not known (31).

The unique, age-specific roles of Flt3L and IL-7 in B lymphopoiesis and the impaired TI Ab response of young mice despite the presence of B1b, B1a, and MZ B cells led us to hypothesize that IL-7–dependent B cells are required for the response to PPS in adults. In the current study, we show that B lymphopoiesis driven by IL-7 is essential for TI immunity.

The Institutional Animal Care and Use Committee have approved these studies. Mice were housed in microisolator cages with free access to food and water and were maintained in a specific pathogen-free facility. C57BL6/J (wild-type) and B6.129S7-IL-7r^tm1Imx/J (IL-7Rα^−/−) were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice deficient in IL-7 (IL-7^−/−) (27) on the C57BL/6 background were obtained from Dr. C. Mackall’s colony at the National Cancer Institute, National Institutes of Health (Bethesda, MD). IL-7 transgenic (IL-7Tg) mice on the C57BL6 background (32) were originally generated by Dr. R. Ceredig (Institut National de la Santé et de la Recherche Médicale, Strasbourg, France) and obtained from Dr. A. Feeney (The Scripps Research Institute, La Jolla, CA). Mice 3 wk of age and mice 8–14 wk old were considered young and adult mice, respectively.

For pneumococcal infections, 5 × 10³ CFU S. pneumoniae WU2, a type 3 strain (33, 34), were injected i.p., and survival of immunized mice was monitored.

Ten micrograms of 23-valent PPS vaccine (Pneumovax 23; Merck & Co., Whitehouse Station, NJ) (35) or 50 μg NP-Ficoll (NP⁴⁰-aminoethyl carboxymethyl-Ficoll; Biosearch Technologies, Novato, CA) dissolved in 100 μl Dulbecco’s PBS (Mediatech, Herndon, VA) were used to immunize mice i.p. Blood samples were obtained 0, 7, and 14 d following immunization.

Total IgM or IgG levels in naive mice were measured, according to the manufacturer’s instructions (Bethyl Laboratories, Montgomery, TX). The hapten NP-specific response was measured by coating plates with NP-conjugated BSA (NP²³-BSA; Biosearch Technologies). Pneumovax23, PPS3, PPS14-specific IgM, and IgG3 were measured by coating 96-well plates with either 50 μl Pneumovax 23 (5 μg/ml), PPS3, or PPS14 (5 μg/ml; American Type Culture Collection, Manassas, VA). All plates were washed and blocked with 2% BSA in PBS (pH 7.2) for 2 h at room temperature. Blood samples of immunized mice were diluted 1/25, 1/100, or 1/500, samples were centrifuged (16,000 × g for 10 min), and supernatant was used. Bound IgM or IgG3 was measured using HRP-conjugated goat anti-mouse IgM or IgG3. Specific Ab levels were interpreted as nanograms per microliter equivalents using IgM or IgG3 standards.

To determine the frequency of B1a and B1b cells, peritoneal cavity cells were harvested from individual mice and the cell concentration adjusted to 2.5 × 10⁷/ml in staining medium (deficient RPMI 1640 medium [Irvine Scientific, Santa Ana, CA] with 3% new calf serum, 1 mM EDTA). After blocking FcRs with 2.4G2 Ab (1 μg/10⁶ cells), an aliquot of 25 μl peritoneal cavity cells was incubated in a microtiter plate with appropriately diluted Ab. To determine the frequency of FO and MZ B cells, 25 μl spleen cells was stained with appropriate Abs. The Abs anti–IgM-FITC (clone 1B4B1), anti–Mac1-allophycocyanin (clone M1/70), and anti–CD5-PerCP (clone 53-7.3) were purchased from eBioscience (San Diego, CA); anti–CD23-PE (clone B3B4) and anti–CD21-FITC (clone 7G6) were from BD Pharmingen (San Diego, CA); and with anti–B220-TC (clone RA3-6B2) was purchased from Caltag Laboratories (Burlingame, CA). NP²³-PE was purchased from Biosearch Technologies. After staining, cells were washed twice with staining medium and the preparations were analyzed on a FACSCalibur (BD Biosciences, San Jose, CA) using the CellQuest software (BD Biosciences). Data were analyzed using the FlowJo software program (Tree Star, San Carlos, CA).

Statistics were performed using the Prizm 5 software program (GraphPad, La Jolla, CA). To determine statistically significant differences, the Student’s unpaired t test was performed and two-tailed p values were given.

Impaired responses in neonates (9–12 d old) to TI-2 Ags can be attributed to the lack of MZ B cells (8, 14). However, 3 to 4-wk-old mice are still unable to respond to TI Ags (7), despite having a developed MZ B cell compartment (Table I). Because B1b cells mount anti-PPS and anti–NP-Ficoll Ab responses (5, 21), we examined the peritoneal cavity of young (3 wk old) and adult (8–14 wk old) mice for B1b cells. Young C57BL6 mice had all the mature B cell subsets, including B1b cells, involved in TI responses, although the numbers were reduced compared with adult mice (Table I). The reduction in B1b, B1a, and MZ B cell numbers in young mice can be attributed to the physically smaller coelomic cavity and spleen, respectively.

B1b cells generated early in life are derived from IL-7–independent fetal liver B lymphopoiesis (28, 36, 37). B1b cells are also generated efficiently from adult bone marrow B lymphopoiesis (36, 37) that is presumably driven by IL-7 (28, 29). To examine the role of IL-7 signaling in the ontogeny of all B cell subsets involved in TI responses, we analyzed mice deficient either in IL-7 or in IL-7Rα. IL-7^−/− mice generated B1b cells with a higher frequency than wild-type mice (Fig. 1), although the absolute numbers were ~2-fold less than wild-type levels (Table I). These results demonstrate that efficient B1b lymphopoiesis occurs in the absence of IL-7. The frequency as well as absolute numbers of B1b cells in adult IL-7Rα^−/− mice was comparable to that of IL-7^−/− mice (Fig. 1, Table I). Because mice deficient in both Flt3L and IL-7Rα lack all mature B cell subsets, the B1b cells generated in the absence of IL-7Rα–mediated signaling are likely to be derived from Flt3L-driven fetal liver B lymphopoiesis (25, 26). IL-7^−/− mice have B1a cells and MZ B cells in abundance (Fig. 1, Table I) as reported earlier (28). We have also confirmed that MZ B and FO B cell subsets in IL-7Rα^−/− mice are reduced severely (38, 39). Despite a reduction in B1a, MZ B, and FO B cells, total serum IgM and IgG levels in IL-7^−/− or IL-7Rα^−/− mice were significantly higher than in wild-type mice (Supplemental Table I) as reported previously (28).

IgM is the dominant and the earliest Ig isotype produced during TI responses (17), and IgM memory is critical for controlling S. pneumoniae in humans (13). Therefore, we measured specific IgM responses either to Pneumovax23, a polyvalent vaccine composed of polysaccharides from 23 serotypes of S. pneumoniae, or to PPS3 or PPS14, two of the PPS serotypes present in the vaccine. We found that young but not adult wild-type mice are impaired in responding to Pneumovax23, PPS3, or PPS14 (Fig. 2A) and die upon challenge with S. pneumoniae strain WU2 (PPS3 serotype) (Fig. 2B). To examine whether developmentally distinct B cells contribute to TI responses in adults, we immunized adult mice deficient in IL-7 or IL-7Rα signaling with Pneumovax23. Despite the presence of all TI B cell subsets including B1b cells (Fig. 1, Table I), IL-7– or IL-7Rα–deficient mice were incapable of mounting specific responses to Pneumovax23, PPS3, or PPS14 (Fig. 2A). These results demonstrate that IL-7–dependent B cells are crucial for generating an anti-PPS response.

IL-7Rα–deficient mice had higher preimmune levels of Pneumovax 23- and PPS3-reactive IgM than the wild-type mice, but the increase in specific Abs upon immunization in these mice was not statistically significant (Fig. 2A). Furthermore, the increase in the PPS-specific IgM response in IL-7^−/− mice upon immunization was minimal. IgG3 is also a major Ig isotype generated during TI responses in mice (40). Immunized adult wild-type mice but not adult IL-7^−/− or IL-7Rα^−/− or young wild-type mice generated a robust anti–Pneumovax 23-specific IgG3 response (Fig. 2A). To test whether this increase and/or the higher baseline reactivity (wild-type versus IL-7Rα^−/−) (Fig. 2A) was optimal for protection or not, we challenged mice with S. pneumoniae 4 wk postimmunization. All the immunized adult wild-type mice that were able to generate a response survived the challenge (Fig. 2B). However, a majority of the immunized adult IL-7^−/− or IL-7Rα^−/− mice succumbed to infection (Fig. 2B), suggesting that a functional and PPS-specific IgM and IgG3 dependent on IL-7–mediated B lymphopoiesis was critical for protective immunity.

To test whether IL-7–dependent B cells also mediate a specific response to another TI Ag, we immunized IL-7^−/− or IL-7Rα^−/− mice with NP-Ficoll. The binding to NP by preimmune IgM of adult IL-7Rα^−/− mice was significantly higher than that of wild-type mice (p > 0.016) (Fig. 3). Because of the variable preimmune baseline-binding values, the postimmunization data of mutant mice were interpreted as the fold increase compared with their preimmune levels. In adult wild-type mice, NP-specific IgM levels were more than an order of magnitude higher than their preimmune IgM levels (Fig. 3). In contrast, immunized adult IL-7^−/− or IL-7Rα^−/− mice generated only a 2-fold increase in Ag-specific IgM responses, suggesting that IL-7–dependent B cells play a major role in generating NP-specific Abs.

The above results demonstrated that IL-7–driven B lymphopoiesis is critical for TI responses. Although at present it is not clear how IL-7–driven B lymphopoiesis is contributing to TI responses, it is known that IL-7Rα–mediated signaling is involved in generating BCR diversity, proliferation, and differentiation during B cell development (41, 42). B cells in the young are known to have limited BCR diversity (43–45) because they are presumably generated in an IL-7–independent mechanism and therefore may have a restricted BCR repertoire that could limit the ability of young mice to recognize a wide range of Ags. To examine whether expression of IL-7 can restore the impaired responses to TI Ags in the young, we immunized young IL-7Tg mice with Pneumovax23. We found that young IL-7Tg mice generate robust IgM responses to Pneumovax23 that were comparable to adult wild-type mice (Fig. 4A). Moreover, unlike young wild-type mice (Fig. 2), Pneumovax23-immunized young IL-7Tg mice survived S. pneumoniae challenge (Fig. 4B). These data demonstrate that enforced expression of IL-7 can restore TI responses and protective immunity to S. pneumoniae in the young. Because the frequency of MZ B (46), B1a, and B1b cells (Table II) in young IL-7Tg were comparable to young wild-type mice, it is possible that overexpression of IL-7 may have induced qualitative changes in the B cells of young mice.

Rag1^−/− mice reconstituted with adult naive B1b cells survive if immunized with PPS3 prior to S. pneumoniae challenge (21). Protection of Pneumovax23-immunized adult wild-type mice upon S. pneumoniae challenge can be attributed to an expansion of Ag-specific B cells. The NP hapten system is widely used for studying Ag-specific B cells in a variety of murine models (47–49). Adoptive transfer of purified B cells into Rag1^−/− mice implicated that peritoneal cell preparations containing B1b cells mount an anti-NP Ab response upon NP-Ficoll immunization (5). To directly identify which B cell subset contributes to the majority of the TI response, wild-type adult mice were immunized with NP-Ficoll and analyzed for NP binding B cells by flow cytometry using NP-conjugated PE (NP-PE). In adult wild-type mice, ~65% of the B cells recognizing NP were of the B1b phenotype, whereas the remaining ~20 and ~15% were of B2 and B1a phenotype, respectively (Fig. 5). This trend was observed in naive as well as in immunized mice (Fig. 5). The binding of NP-PE by B cells of NP-immunized mice was specific because NP conjugated to BSA (NP-BSA) competitively inhibited this binding (Fig. 5). These data support the notion that the resistance of immunized but not naive mice to S. pneumoniae infection is likely to be attributable to an expansion of Ag-specific B cell subsets, predominantly including B1b. The ability of purified naive B1b cells to generate anti–Pneumovax23-specific IgM upon immunization indicates that Ag-specific B1b cells do not remain quiescent but expand and differentiate into Ab-secreting cells (Supplemental Fig. 1).

TI responses are an integral part of humoral immune defense against a variety of pathogenic microorganisms (1, 2). TI responses are impaired in the young but not adults. B lymphopoiesis late but not early in life is IL-7 dependent. In the current study, we show that the Ab responses to TI Ags are dependent on B cells generated by IL-7–dependent B lymphopoiesis. Importantly, we also show that TI responses in the young can be restored by IL-7.

Recent studies have revealed that B1b cells play a critical role in TI responses (14). Despite the presence of B1b cells, TI responses in the young are impaired. B1b cells are generated efficiently from precursors in fetal liver as well as in adult bone marrow (50). Because B1b cells generated during the perinatal period persist throughout life by self-renewal (36, 37), the B1b cell pool in adults is developmentally heterogeneous. B1b cells generated perinatally from fetal liver progenitors share a number of characteristics with B1a cells, including self-replenishment and feedback regulation of development (36, 37). The Abs produced by B1a and B1b cells generated early in life have limited repertoire of antigenic specificities. Such Abs are referred as natural Abs because they are generated spontaneously in naive mice and in normal individuals in the absence of apparent Ag stimulation (51). Although natural Abs are generally of low affinity to self or evolutionarily conserved Ags (52), they have been shown to play a role in the early control of pathogen burden in bacterial and viral infection models (51, 53–55). However, natural Abs do not play a critical role in clearance of Borrelia hermsii, a bacterial pathogen, whose elimination requires B1b cells (22). Moreover, natural Abs do not recognize FhbA, a B. hermsii Ag target for B1b cells (22), suggesting that B1b cells produced early in life do not contribute significantly to the anti-B. hermsii response (18, 56). Natural Abs alone are not sufficient to clear S. pneumoniae (21). Instead, a division of labor between the two B1 cell subsets is required to control S. pneumoniae (21, 57); natural Abs produced by B1a cells limit infection of S. pneumoniae by virtue of recognizing evolutionarily conserved Ags, such as phosphorylcholine (17, 58), whereas the anti–PPS-specific response generated by B1b cells are critical for protection against serotype-specific pneumococcal infection (21, 57).

The strength of BCR signaling is known to be critical for the development of B1a and B1b cells (59). Xid mice are defective in BCR signaling and are therefore deficient in B1a and B1b cell subsets and have reduced levels of serum IgM (60). Defective BCR signaling is also responsible for the impaired TI responses in xid mice (61). The fact that IL-7^−/− or IL-7Rα^−/− mice generate B1a and B1b cell subsets (Fig. 1, Table I) suggests that the impaired TI responses in these mice are not because of a defect in BCR signaling strength. Moreover, the B cells generated in the absence of IL-7Rα signaling are not deficient in generating spontaneous IgM or IgG. In fact, the basal levels of serum IgM and IgG Abs in naive IL-7^−/− or IL-7Rα^−/− mice are significantly higher than in wild-type mice (Supplemental Table 1). These results suggest that the impaired responses to specific TI Ags (e.g., PPS or NP-Ficoll) are not because of a defective BCR signaling.

The reason(s) for the lack of specific Ab responses to TI Ags in mice deficient in IL-7 signaling is currently under investigation. B1 (Table II) and MZ B (46) cell numbers in young IL-7Tg mice are comparable to young wild-type control mice, yet they generated a significantly more Ab response to PPS (Fig. 4). Although the B1b cells in IL-7^−/− mice were reduced only by 2-fold (Table I), they did not mount anti-PPS response even after 14 d postimmunization (Fig. 2). Therefore, the inability of IL-7^−/− adult mice to mount a TI response is likely to be attributable to a qualitative rather than quantitative difference in the B cells. Failure to express an extensive and suitable BCR repertoire by B cells could be a potential reason for the lack of Ag-specific Ab response not only in IL-7^−/− mice but also in young wild-type mice. The B cell repertoire in adult wild-type mice represents all VH genes, even more so those, that are distal to the DJ segments (43–45). In contrast, B cells of young wild-type mice preferentially express VH genes proximal to DJ but not the major families of VH genes that are distal to DJ and known to be critical for extensive BCR diversity (43–45). Similarly, in the absence of IL-7–mediated signaling, the usage of several large families of VH genes distal to DJ in the IgH locus does not occur because of a constraint in the accessibility of distal VH genes for VDJ recombination events (41, 62). It is known that TdT, an enzyme necessary for the addition of nontemplate nucleotides between V-D and D-J junctions during VDJ recombination, is not expressed early in life (63–65), and the B1a and B1b cells generated at this time of life have limited junctional diversity in their BCRs (63–65). IL-7 also appears to play a role in the expression of TdT (66). Thus, IL-7 signaling plays a critical role in generating a wider range of BCR diversity. It is possible that the B cells generated early in life or in the absence of IL-7Rα signaling might have a “hole in the BCR repertoire” that can account for the lack of recognition to a wide range of TI Ags, such as NP-Ficoll and PPS. Besides a role in generating BCR diversity, IL-7Rα is also known to transmit distinct signals for proliferation and differentiation during B lymphopoiesis (42). Therefore, IL-7–dependent B cells that develop in the bone marrow could be qualitatively distinct from the B cells generated independent of IL-7 early in life.

The TNF family of receptors, namely B lymphocyte-activating factor receptor, transmembrane activator and calcium modulator and cyclophilin ligand interactor, and B cell maturation Ag play an important role in TI humoral responses (67–69). It has been reported that decreased expression of these receptors on B cells is associated with the lack of humoral responses in neonates (15, 16). TLR stimulation can upregulate the expression of TACI on neonatal B cells. TLR agonists, such as CpG oligonucleotides, have been shown to restore anti–NP-Ficoll response in the young (15, 70). These findings suggest that in addition to IL-7–dependent B lymphopoiesis other immunostimulatory pathways, such as TLR and TACI signaling, play an important role in augmenting TI responses.

Analysis of TI responses in humans and mice revealed striking similarities in both species (7). An identical reactivity of serum from B. hermsii-infected individuals (71) and murine B1b cells (22) to FhbA indicates that humans have functional equivalents of B1b cells. The PPS vaccine generates a protective response in humans, and in the murine system, the B1b subset generates the bulk of the protective Ab response to PPS (21, 57). It is possible that the young might lack sufficient protection from a developmentally distinct B1b cell repertoire driven by IL-7–driven B lymphopoiesis.

We thank Dr. David Briles and Janice King for providing S. pneumoniae strain WU2; Drs. Crystal Mackall for IL-7^−/− mice and Rod Ceredig and Ann Feeney for IL-7Tg mice, respectively; Dr. Mustafa Akkoyunlu for discussion; and Dr. Tim Manser for critical review of the manuscript.

Disclosures The authors have no financial conflicts of interest.