Abstract
RANKL–RANK signaling regulates numerous physiologic processes such as bone remodeling, lymph node organogenesis, central thermoregulation, and formation of a lactating mammary gland in pregnancy. Recently, a receptor activator of NF-κB ligand (RANKL)-blocking Ab has been approved for human use in potentially millions of osteoporosis and cancer patients. However, germline deficiencies in RANKL or receptor activator of NF-κB (RANK) also lead to strong B cell defects in mice and human patients, suggesting that RANKL–RANK inhibition could interfere with B cell physiology and thereby trigger immunologic side-effects. To address this key question—that is, whether RANKL–RANK signaling affects B cell physiology directly or the observed defects are secondary because of the severe osteopetrosis—we generated B cell-specific RANK knockout mice. We show that B cells deficient for RANK undergo normal development and do not show any obvious defects in Ab secretion, class switch recombination, or somatic hypermutation. Our data indicate that ablation of the RANKL–RANK pathway has no direct adverse effect on B cell physiology.
Receptor activator of NF-κB ligand (RANKL, also known as ODF, TRANCE, OPGL, and TNFSF11) is the activating ligand of the receptor RANK (receptor activator of NF-κB, also known as TRANCE-R and TNFRSF11A), whereas osteoprotegerin (OPG; also known as OCIF, TNFRSF11B, TR1, FDRC1) acts as a natural decoy receptor for RANKL and, therefore, as a negative regulator of RANKL–RANK signaling (1). The RANKL/RANK/OPG system was shown to regulate a number of physiologic processes. RANKL–RANK signaling is essential for bone remodeling by regulating osteoclast development and function (1). Moreover, RANKL and RANK are indispensable for lymph node organogenesis (2, 3), for the development of a lactating mammary gland (4), the formation of AIRE+ medullary thymic epithelial cells (5), and were shown to regulate the fever response in the CNS (6). In addition, RANKL–RANK can promote progestin driven breast cancer (7), can act as a soil factor for cancer cells metastasizing to bone (8), or drive local invasion of experimental breast and prostate tumors (9).
Interestingly, it was also reported that B lymphocytes express RANK (10). The analysis of mice carrying a germline deletion of either RANK or RANKL revealed dramatic defects in B cell development, resulting in reduced numbers of peripheral B cells (2, 3). Importantly, human patients with a mutation in RANK also exhibit B cell defects such as reduced serum Ig levels, hypogammaglobulinemia, or impaired Ab responses to Ags (11). Denosumab is a fully human mAb against RANKL that has recently been approved as a treatment against osteoporosis of postmenopausal women, bone loss in men undergoing hormone ablation therapy against prostate cancer, and the effects of bone metastases from solid tumors (12–19). It has therefore been proposed that blocking RANKL might directly interfere with B cell function. However, both mice and human patients lacking functional RANK or RANKL develop osteopetrosis because of the absence of osteoclasts (2, 3, 11, 20). Osteopetrotic bones are denser than normal bones and exhibit a lack of bone marrow cavities, which are the regular environment for B cell development (1). Therefore, it is also conceivable that B cell defects in RANK deficient mice and human patients are secondary to the absence of the natural site of B cell development—the bone marrow cavities.
To address this question, we generated mice that lack RANK specifically in B cells. In this study, we report that mice that lack RANK in B cells generate regular primary and secondary lymphoid organs, and RANK-deleted B cells undergo development normally. Furthermore, we demonstrate that basic B cell functions involved in humoral immunity such as Ab secretion, Ig class switch recombination (CSR), or somatic hypermutation (SHM) are undisturbed in the absence of RANK signaling. These data indicate that RANKL/RANK have no direct involvement in B cell physiology.
Materials and Methods
Mice
Mice carrying the conditional RANKFlox (RANKF) allele and the RANKΔ allele have been described previously (21). The mutant RANK alleles were backcrossed to C57BL/6 more than ten times. MB1-cre mice (22) were obtained on a C57BL/6 background. We crossed RANKF/F females with mb1-cre+ RANKΔ/+ males to obtain mb1-cre+ RANKΔ/F experimental mice and RANKF/+ control animals. All mice were maintained at the animal colony of Institute of Molecular Biotechnology according to institutional guidelines.
Fluorescence activated cell sorting
hiCD28+Lin− (CD4−CD8α−CD11b−CD19−) where lineage depletion was performed by MACS sorting (Miltenyi Biotec). FACS analysis was performed on a FACSCalibur (BD Biosciences), a FACSCanto (BD Biosciences), and on an LSRFortessa (BD Biosciences) apparatus. Cell sorts were performed using a FACSAria (BD Biosciences) apparatus.
Absolute splenocyte numbers were determined by counting total splenocytes after RBC lysis with a CASY1 counter and subsequent calculation of T cell and B cell numbers based on ratios from FACS experiments.
Histology
Cryosections from spleens were fixed with acetone and blocked in TBS, 0.1% BSA. Samples were incubated with biotin-coupled TCR-β (H57-597) or FITC-coupled anti-B220 (RA3-6B2) or IgD (11-26c.2a) Abs (BD) or with biotinylated PNA, followed by streptavidin-HRP (BD) and anti Fluorescein alkaline phosphatase Fab Fragments (Roche). Fast Diaminobenzidine tablets (Sigma) and Fast Blue RR salt (Sigma) were used for detection according to the manufacturers’ instructions.
Immunizations
Fresh sheep RBCs (Innovative Research) were washed in PBS three times; 1 × 108 cells per mouse were injected i.p. Analysis was done after 12 d.
Determination of Ig levels
Somatic hypermutation assay
Efficiency of SHM was assesses as described previously (23). B220+PNAhiCD95+ splenic germinal center B cells were sorted from sheep RBC (SRBC)-immunized mice. Upon DNA extraction, the VHJ558–JH4 intragenic region was PCR amplified with Pfu DNA polymerase (Stratagene), cloned into Zero Blunt-TOPO vector (Invitrogen), sequenced, and analyzed in comparison with nonimmunized splenocytes.
Class switch recombination
CD43− B cells were isolated from spleens by MACS (Miltenyi Biotec) and stimulated for 5 d with LPS (20 μg/ml) or IL-4 (50 ng/ml) plus αCD40 (1 μg/ml) to induce switching to IgG3 or IgG1, respectively. Percentages of switched B lymphocytes were assessed by flow cytometry (24).
PCR analyses
RANK wt (256 bp), floxed (390 bp), and Δ (566 bp) alleles were identified using PCR. The following primers and amplification conditions were used: 5′-GGCAGAACTCGGATGCACAGATTGG-3′, 5′-AGTGTGCCTGGCATGTGCAGACCTT-3′, 5′-CTGGTGGTTGTTCTCCTGGTGTCAT-3′ with 35 cycles of 95°C for 30 min, 60°C for 30 min, and 72°C for 30 min. For detection of RANK mRNA expression, RNA was isolated with Trizol (Invitrogen), followed by reverse transcription with SuperscriptII (Invitrogen) and random hexamers (Roche). Real time PCR was performed using the oligonucleotides 5′-CCAGGAGAGGCATTATGAG-3′ and 5′-CATTCCAGGTGTCCAAGTA-3′ (40 cycles of 95°C for 10 min, 60°C for 60 min).
Results
Deletion of RANK in B cells
To analyze a direct effect of RANK deficiency on B cells, we used the conditional RANK allele in which exons 2 and 3 are flanked by loxP sites. Cre-mediated deletion of exons 2 and 3 results in a frame shift and a RANK null allele (Supplemental Fig. 1A) (21). For a specific deletion in B cells, we crossed the conditional RANK allele to mb1-cre mice, which efficiently delete in the complete B cell lineage from the earliest pro B cell stage onwards (22). Efficiency of Cre-mediated deletion of RANK exons 2 and 3 and consequent loss of RANK expression in B cells was confirmed by PCR for the deleted and floxed alleles (Supplemental Fig. 1B) and the absence of RANK mRNA expression (Supplemental Fig. 1C). In all the experiments described below, we analyzed mb1-cre+ RANKΔ/F mice (hereafter termed RANKΔB) that lack RANK in B cells, and, as controls, RANKF/+ animals. Both experimental mice and controls were obtained as littermates by crossing RANKF/F females to mb1-cre+ RANKΔ/+ males.
Normal B cell development
It has been reported that absence of RANKL–RANK signaling in mice can lead to a block in B cell development (2, 3). We therefore analyzed B lymphocytes in bone marrow, spleen, and lymph nodes by FACS analysis. B cells develop in the bone marrow from hematopoietic stem cells to pro-B cells (IgM−CD19+c-kit+), pre-B cells (IgM−CD19+CD25+), and immature B cells (IgM+B220int). Mice lacking Rank in B cells display similar frequencies of pro-B cells, pre-B cells, and immature B cells in bone marrow when compared with control RANKF/+ animals (Fig. 1A). In spleens, we analyzed population frequencies of immature (B220+CD93+) and mature B cells (B220+CD93−). B220+CD93+ immature B cells can be subdivided into T1 (IgM+CD23−), T2 (IgM+CD23+), and T3 (IgMloCD23+) transitional B cell subsets. Furthermore, we analyzed the mature B cell subsets of follicular B cells (B220+CD21+CD23+) and marginal zone B cells (CD21hiCD23lo). All of those B cell populations were present at comparable ratios in RANKΔB and control mice (Fig. 1B). Moreover, ratios of Igκ and Igλ expressing B cells in bone marrow and spleen were similar in RANKΔB and control mice (Supplemental Fig. 2). We finally analyzed peritoneal B-1 B cells (B220+CD11b+IgM+CD5+; Fig. 1C) and plasma cells (Lin−CD28+CD138+) in bone marrow and spleen (Fig. 1D). Again, all those B cell populations showed similar frequencies in the presence and absence of RANK. Finally, assessment of total splenocyte numbers, as well as CD3ε+ splenic T cell, CD19+, or IgM+ splenic B cell numbers resulted in similar values (two-tailed t test) in RANKΔB and control animals (Table I). Therefore, we conclude that RANK deletion in B cells does not affect B cell development in bone marrow, spleen, and the peritoneal cavity.
B cell development. A, FACS analysis of pro-B cells (IgM−CD19+c-kit+), pre-B cells (IgM−CD19+CD25+), and immature B cells (IgM+B220int) in bone marrow. B, FACS analysis of immature (B220+CD93+) and mature B cells (B220+CD93−; left panels); T1 (IgM+CD23−), T2 (IgM+CD23+), and T3 (IgMloCD23+) transitional B cells (middle panels), and follicular B cells (B220+CD21+CD23+) and marginal zone B cells (CD21hiCD23lo; right panels) in spleen. C, FACS analysis of B-1B cells (B220+CD11b+IgM+CD5+) isolated from the peritoneal cavity. D, FACS analysis of plasma cells (Lin−CD28+CD138+) in bone marrow (left panels) and spleen (right panels). FACS blots are representative of at least three independent experiments and at least three mice per group, 6–12 wk of age.
Formation of secondary lymphoid organs
In RANKL-deficient mice, it was observed that B cell zones in spleens were dramatically reduced in size (3) (Fig. 2A). However, it was not clear whether this phenotype could be directly attributed to the lack of RANKL–RANK signaling in B cells or whether it was secondary for example to the osteopetrosis phenotype and a resulting defect in lymphocyte development. To answer this question directly, we first analyzed spleen sections of RANKΔB mice. Histologic stainings of B and T lymphocytes revealed normal B and T cell zones in splenic white pulp in RANKΔB mice (Fig. 2B). Moreover, we observed normal development of splenic germinal centers upon immunization with SRBCs (Fig. 2C). This result was corroborated by quantification of B220+PNAhi germinal center B cells by FACS analysis (Supplemental Fig. 3). Therefore, splenic architecture and organization of B cell and T cell zones do not depend on RANK signaling in B lymphocytes.
Histologic analysis of RANKL−/− and RANKΔB spleen sections. A, White pulp in spleens from RANKL−/− and wild type mice stained for IgD (brown). Original magnification ×50. B, White pulp in spleens from RANKΔB and control mice stained with anti-B220 (blue) and anti–TCR-β (brown) Abs to identify B cell and T cell zones. Original magnification ×20. C, White pulp in spleens from SRBC immunized RANKΔB and control mice stained with an anti-IgD (blue) Ab and PNA (brown) to identify B cell follicles and germinal centers, respectively. Original magnification ×20. Representative results are shown from at least three independent experiments and at least four mice per group, 8–12 wk of age, with similar results.
It was reported that mice with a germline deletion of RANK or RANKL do not develop any lymph nodes, but do develop Peyer's patches at reduced numbers and size (2, 3). Moreover, human patients with germline mutations in RANK do not show palpable lymph nodes. Therefore, we examined axillary, inguinal, mesenteric, and popliteal lymph nodes as well as Peyer's patches but did not find an obvious defect in morphology or in numbers of these organs in RANKΔB mice (not shown). Ratios of B cells and T cells in Peyer's patches and lymph nodes were comparable in RANKΔB and control animals (Supplemental Fig. 4). Therefore, we conclude that RANKL–RANK signaling in B cells is dispensable for the generation of lymph nodes.
Intact B cell functions in RANKΔB mice
It was reported that RANK mutations in human patients can lead to reduced Ig levels in serum and to impaired Ab responses to Ags. To assess whether RANK deficiency in B cells can cause such phenotypes, we measured serum Ig levels in nonimmunized and in SRBC-immunized mice. We determined levels of serum IgM, IgG1, IgG2a, IgG2b, IgG3, IgA, and IgE in RANKΔB and control mice, but did not detect a significant difference in any of the Ig isotypes, neither in nonimmunized nor in SRBC-immunized mice (Fig. 3). Another way to investigate efficiency of Ig CSR is to stimulate splenic B cells in vitro. Stimulation with LPS leads to switching to IgG2b and IgG3, whereas stimulation with IL-4 and αCD40 results in switching to IgG1 and IgE. Efficiency of CSR was similar in B cells from RANKΔB and from control mice (Fig. 4). Finally, we determined SHM in RANKΔB mice. To this end, we immunized mice with SRBCs and FACS sorted B220+PNAhiCD95+ activated B cells that undergo SHM (23). We did not observe an SHM defect in B cells lacking RANK, as the mutation rates were comparable in RANKΔB and control animals (Table II). These data indicate that RANK signaling in B cells is dispensable for Ab production, efficient Ig CSR, and SHM.
Ig levels in serum. Serum Ig levels of all Ig isotypes from nonimmunized (A) or SRBC-immunized (B) mice were assessed. Values obtained from RANKΔB animals are shown as triangles, values obtained from control animals are shown as squares and are shown for each Ig isotype for both nonimmunized (A) and SRBC-immunized (B) mice. Triangles and squares represent values from individual animals. Data are shown as mean ± SEM. Data are shown from three independent experiments. Three to 11 mice, 6–14 wk of age, were used per group.
Class switch recombination. FACS analysis of splenic B cells from RANKΔB and control mice stimulated with IL-4/anti-CD40 (A) or with LPS (B) inducing class switch recombination to IgG1 (A) or IgG3 (B), respectively. FACS blots are representative of four independent experiments. Four mice per group, 7–12 wk of age, were used with similar results.
Discussion
We show that B cell intrinsic expression of RANK is not required for normal development or homeostasis of B lymphocytes in bone marrow and in the peripheral lymphoid organs. RANKΔB mice exhibit normal populations of pro-B cells, pre-B cells, immature or transitional B cells, mature B cells such as marginal zone B cells and follicular B cells, and plasma cells in bone marrow and spleen as well as normal numbers of B-1 B cells in the peritoneum. Our data differ from previously published work on mice carrying a germline mutation for either RANK or RANKL (2, 3). In these mice, a dramatic reduction of peripheral B cells was observed, combined with a partial developmental block from the pro- to the pre-B cell stage in RANKL-deficient mice (3). Because such whole body RANKL and RANK mutant mice exhibit severe osteopetrosis, it has been difficult to ascertain whether RANKL–RANK act directly in B cells. Osteopetrotic bones of RANKL and RANK knockout mice are denser than normal bones and, importantly, exhibit an almost complete lack of bone marrow cavities. The bone marrow is the natural environment for B cell development as well as the residence of hematopoietic stem cells from which all immune cell types are derived (25). Therefore, one feasible theory has always been that the observed B cell defects are secondary to the absence of bone marrow cavities in RANKL- and RANK-deficient mice rather than being a direct effect of absent RANKL–RANK signaling on B lymphocytes. Our data strongly support an argument against a direct effect of RANKL–RANK signaling on B cell development.
In RANKΔB mice, we found Peyer's patches and lymph nodes of normal morphology with normal B cell and T cell numbers. In contrast, germline mutations of RANKL or RANK result in the complete absence of lymph nodes in mice (2, 3). Similarly, human patients carrying inactivating RANKL and RANK mutations lack palpable lymph nodes (11, 20). Our data demonstrate that RANKL–RANK signaling in B cells is not a requirement for the development of lymph nodes. The exact cellular and molecular mechanisms by which RANKL and RANK control lymph node organogenesis during embryogenesis remain largely elusive (1, 26).
It was reported that human patients carrying RANK mutations can exhibit disturbed B cell functions resulting in conditions such as hypogammaglobulinemia, a lack of Ab response to Ag, or in the presence of reduced serum Ig levels (11). Because we observed high RANK expression levels in IL-4/anti-CD40 activated B cells (Supplemental Fig. 1C), we investigated various functions of activated B cells, including CSR, SHM, and Ab secretion to serum with or without immunization. Again, RANKΔB mice did not display any obvious defects, arguing against an essential role for RANKL–RANK signaling in B lymphocyte functions.
Our work examines various aspects of B cell biology in the absence of RANKL–RANK signaling, without identifying an essential function of RANK in B lymphocytes. That leaves us with the question of the actual biologic function of the observed expression of RANK in B cells. One possible explanation is a functional redundancy of RANK signaling in B cells. Alternatively, although we performed comprehensive analyses of B cell development and basic B cell functions, the potential role of RANK signaling in B cells could be hidden in more complex functions of B cells that were not addressed in this study. Overall, RANK signaling does not seem to have a major role in B cell physiology.
Disclosures
J.M.P. owns stock of Amgen. The other author has no financial conflicts of interest.
Acknowledgments
We thank Dr. Shane Cronin, Dr. Ulrich Elling, Dr. Reiko Hanada, Dr. Toshikatsu Hanada, Dr. Gregory Neeley, Dr. Daniel Schramek, and Dr. Gerald Wirnsberger for helpful comments and discussions, Qiong Sun for help with histologic stainings, and the Institute of Molecular Biotechnology service departments, especially Gerald Schmauss and Harald Scheuch, for technical support.
Footnotes
J.M.P. is supported by grants from the Institute of Molecular Biotechnology of the Austrian Academy of Sciences, the Austrian Ministry of Sciences, the Austrian Academy of Sciences, Genome Research in Austria (AustroMouse), and a European Union European Research Council Advanced Grant. T.P. is supported by a Marie Curie International Incoming Fellowship. The research leading to these results has received funding from the European Community’s Seventh Framework Programme (FP7/2007-2013) under Grant Agreement 252210.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- CSR
- class switch recombination
- OPG
- osteoprotegerin
- PNA
- peanut agglutinin
- RANK
- receptor activator of NF-κB
- RANKL
- receptor activator of NF-κB ligand
- SHM
- somatic hypermutation
- SRBC
- sheep RBC.
- Received July 15, 2011.
- Accepted November 28, 2011.
- Copyright © 2012 by The American Association of Immunologists, Inc.