B cell chronic lymphocytic leukemia
IL-5 transgene-congenic (NZB×NZW)F1
IL-5 transgene-negative (NZB×NZW)F1
IL-5 preferentially activates B1 cells to produce natural antibodies cross-reactive to self antigens. To determine the role of IL-5 in antibody-mediated autoimmune disease, we generated systemic lupus erythematosus (SLE)-prone (NZB×NZW)F1 mice congenic for IL-5 transgene (TG-F1). The transgene unexpectedly reduced the incidence of lupus nephritis. Anti-DNA antibodies in sera and those produced by splenic B cells in vitro were markedly decreased in TG-F1 mice, while total polyclonal Ig levels were comparable to those in IL-5 transgene-negative (NZB×NZW)F1 (non-TG-F1) littermates. Flow cytometry-sorted splenic B1 cells showed a significant reduction of anti-DNA antibody synthesis in response to IL-5, while proliferative responses to IL-5 did not significantly differ between TG-F1 and non-TG-F1 mice. As TG-F1 mice aged, frequencies of peripheral B1 cells progressively increased, and the mice frequently developed B cell chronic lymphocytic leukemia (B-CLL). Our results suggest that dysregulated, continuous high expression of IL-5 in SLE-prone mice may directly or indirectly mediate a skewed signaling of proliferation/differentiation of self-antigen-activated B1 cells, leading to suppression of autoimmune disease, but instead to aberrant expansion of B1 cells, giving rise to B-CLL. Thus, this model may provide a clue to the pathogenesis of both SLE and B-CLL.
Currently, at least two subpopulations of mouse B cells, B1 and B2 cells, can be divided on the basis of differences in phenotype, physiology and antibody repertoire. The majority of B1 cells express CD5 on their cell surface (B1a cells), while there is a minor subset of B1 cells (B1b cells) that lack the CD5 phenotype 1, 2. Functionally, B1 cells mainly participate in innate immunity, in contrast to conventional B2 cells, which participate in acquired immunity. B1 cells produce IgM natural antibodies of a low-avidity nature, which are polyreactive and cross-react with a variety of self antigens 3, 4.
IL-5 is a growth and differentiation factor of eosinophils and B cells 5. Takatsu and collaborators 6, 7 found that the IL-5 receptor is preferentially expressed on CD5+ B1 cells 6, and that IL-5 selectively maintains B1 cells in an in vitro bone marrow culture 7. Furthermore, IL-5-transgenic C3H mice showed a massive eosinophilia and an expansion of B1 cells associated with increased production of IgM antibodies to DNA 8. Thus, even in normal healthy mice, B1 cells continuously exposed to IL-5 have the potential to produce large amounts of IgM autoantibodies. However, IL-5-transgenic C3H mice do not develop autoimmune disease, probably because B1 cells in IL-5-transgenic mice on a C3H genetic background do not show IgM-to-IgG class switching and affinity maturation.
(NZB×NZW)F1 mice spontaneously develop a severe autoimmune disease closely resembling human SLE 9. Young F1 mice before the onset of SLE show IgM hypergammaglobulinemia, including anti-DNA autoantibodies. We earlier found that CD5+ B1 cells are virtually entirely responsible for these events 10, 11. Unlike the case in normal healthy strains of mice, B1 cells in (NZB×NZW)F1 mice are hyper-responsive to IL-5, and produce large amounts of IgM in vitro in the presence of IL-5 11–13. Of note was the finding that in response to IL-5, these B1 cells change their phenotype with the Mott cell, i. e. a pathological state of plasma cells containing large numbers of intracellular inclusions of Ig 14. Whereas Mott cells are rarely found in healthy individuals, they do appear in large numbers in lymphoid organs and other inflammatory tissues in autoimmune diseases including SLE 15, 16. Thus, it is suggested that IL-5 contributes to aberrant B1 cell activation and differentiation found in early life of (NZB×NZW)F1 mice, which mediate IgM hypergammaglobulinemia and a high frequency of Mott cell formation.
Development of SLE in (NZB×NZW)F1 mice, mainly characterized by lupus nephritis, is significantly associated with IgM-to-IgG class switching and affinity maturation of anti-DNA antibodies, which occurs when the animals are 5–6 months of age 9. However, characterization of cell surface phenotypes of anti-DNA antibody-producing precursors showed that while anti-DNA IgM was produced in vitro by CD5+ B1 cells, anti-DNA IgG was mainly produced by cells in the pre-plasma stage with the CD5– phenotype. Using hybridoma technology and Ig VH repertoire analysis, however, we 17 and others 18 showed evidence that precursors of affinity-selected pathogenic IgG anti-DNA clones in aged F1 mice with overt SLE originate from expanded low-affinity IgM anti-DNA clones. Thus, we postulated that IL-5-mediated hyper-activation of B1 cells in young (NZB×NZW)F1 mice may be a prerequisite for later production of pathogenic high-affinity IgG autoantibodies, leading to an early onset of severe SLE. To validate this hypothesis, we established IL-5 transgene-congenic (NZB×NZW) F1 (TG-F1) mice, and examined effects of IL-5 on SLE.
2.1 Establishment of IL-5 transgene-congenic (NZB×NZW) F1 mice
We established IL-5 transgene-congenic NZB and NZW strains by backcrossing (NZB×IL-5-transgenic C3H) F1 and (NZW×IL-5-transgenic C3H) F1 mice to NZB and NZW, respectively. Mice with over 15% of peripheral eosinophils were selected as transgene-positive. These congenic NZB and NZW mice were crossed to obtain (NZB×NZW) F1 hybrids, and the final screening for IL-5 transgene in F1 hybrids was done using Southern blot analysis (Fig. 1). The frequency of peripheral eosinophils per total leukocytes (mean and SE) was higher in TG-F1 (24.6+2.0%) than found in transgene-negative F1 (non-TG-F1) littermates (2.4+0.4%) at 2 months of age (p<0.001). While serum IL-5 was undetectable in non-TG-F1 mice, significant amounts were detected in TG-F1 mice (94.3+19.4 U/ml).
2.2 Effect of IL-5 transgene on SLE features
We first examined serum levels of IgM and IgG anti-DNA antibodies and total IgM and IgG. Compared to non-TG-F1 mice, TG-F1 mice unexpectedly showed a significant reduction in both IgM and IgG anti-DNA antibodies at 5 and 10 months of age (Fig. 2A), while amounts of total serum IgM and IgG did not significantly differ between the two groups of F1 mice (Fig. 2B).
The development of lupus nephritis was in parallel with the age-associated increase in titers of IgG anti-DNA antibodies, and the cumulative incidence of proteinuria in TG-F1 mice was much lower than that found in non-TG-F1 mice (Fig. 2C). While all of the 12 non-TG-F1 mice died of lupus nephritis by 15 months of age, 12 of 15 (80%) TG-F1 mice were still alive at this age (Fig. 2C). Histological and immunofluorescence examinations of renal glomeruli showed that glomeruli in non-TG-F1 mice were markedly enlarged with massive deposition of immune complexes positive for periodic acid-Schiff (PAS) stain, with IgG and C3 both in mesangial areas and along capillary walls. In contrast, glomeruli in the TG-F1 mice were of normal size and limited amounts of IgG and C3 deposits were observed only in mesangial areas (Fig. 3).
2.3 Effect of IL-5 transgene on B1 cell frequencies
We then examined age-associated changes in frequencies of CD5+ B1 cells per total B220+ B cells in peripheral blood, using flow cytometry (FCM) (Fig. 4A). Frequencies of B1 cells progressively increased with aging in TG-F1 mice and were significantly higher than those found in non-TG-F1 mice at 5 months of age onward (Fig. 4B). Absolute B1 cell counts/μl (mean and SE) in the peripheral blood of TG-F1 and non-TG-F1 mice aged 12 months were 10,299+2,019 and 2,605+306, respectively, and the difference was statistically significant (p<0.02). There was no significant difference in absolute numbers of T cells, B2 cells and a subset of CD5–B220dull B1 cells (B1b cells) 1, 2 between TG-F1 and non-TG-F1 mice (data not shown).
In TG-F1 mice, 6 out of 15 mice (40%) showed B1 cell frequencies over 80% at 13 months of age. Using criteria that mice with total peripheral leukocyte counts/μl over 25,000 (mean + 2SD in non-TG-F1 mice over 12 months of age) are diagnosed as having B cell chronic lymphocytic leukemia (B-CLL), these 6 TG-F1 mice eventually developed B-CLL by the age of 22 months. Fig. 5A shows a representative FCM profile of peripheral blood lymphocytes in these mice, in which CD5dullB220dull B-CLL cells are markedly expanded. Blood film presented numerous B-CLL cells having a basophilic cytoplasm and an atypical large cleaved nucleus with a course granular chromatin (Fig. 5B). These B-CLL cells show massive infiltration into interstitial tissues of the liver, kidney and lung (Fig. 5C–E). Fig. 5F shows a representative Southern blot analysis of genomic DNA from T cell-depleted spleen cells for clonal Ig VH region gene rearrangements, using an Ig H chain joining region (JH) probe. While DNA from a 5-month-old TG-F1 mouse showed smear pattern, a single rearranged band was observed with DNA from a 22-month-old TG-F1 mouse with B-CLL.
2.4 Changes in cell surface phenotypes of splenic B cells
Fig. 6 shows expression levels of cell surface molecules that can potentially affect proliferation and differentiation of B cells. Compared to B2 cells, B1 cells preferentially expressed IL-5Rα and CD80 on their cell surfaces; however, there was no difference between TG-F1 and non-TG-F1 mice. MHC class II I-A levels were almost identical between B1 and B2 cells, and there was no difference between TG-F1 and non-TG-F1 mice. Of note was the finding that IL-2Rα levels were markedly up-regulated on the majority of B1 but not B2 cells in TG-F1 mice, compared to those found in non-TG-F1 mice.
2.5 In vitro potential of splenic B cells to produce antibodies
Table 1 shows in vitro potential of T cell-depleted spleen cells from 7-month-old TG-F1 and non-TG-F1 mice to produce anti-DNA antibodies. We earlier found that splenic B cells obtained from aged (NZB×NZW) F1 mice spontaneously produce significant amounts of IgM and IgG anti-DNA antibodies in vitro, and that the antibody syntheses are up-regulated in the presence of IL-5 and IL-6, respectively 13. In keeping with these findings, spleen cells from non-TG-F1 mice spontaneously produced significant amounts of IgM and IgG anti-DNA antibodies, and the levels were significantly up-regulated in the presence of IL-5 and IL-6, respectively. In the same experiments using spleen cells from TG-F1 mice, both spontaneous and IL-5-induced productions of anti-DNA IgM were significantly lower than those found in non-TG-F1 mice. Levels of spontaneous anti-DNA IgG production in TG-F1 mice were much lower than in non-TG-F1 mice, and IL-6 did not significantly promote the antibody synthesis. In contrast, when we measured amounts of polyclonal IgM and IgG spontaneously produced in the same cultures, there were no significant differences between TG-F1 and non-TG-F1 mice. IL-5 up-regulated polyclonal IgM and IL-6 up-regulated IgG syntheses; however, amounts produced were almost identical between TG-F1 and non-TG-F1 mice (Table 1).
|Mice||IgM class||IgG class|
|Anti-DNA antibodies (U/ml)|
|Total Ig (μg/ml)|
We then examined in vitro potential of FCM-sorted splenic CD5+B220dull B1 cells to produce IgM antibodies. Irrespective of the presence or absence of IL-5, amounts of anti-DNA IgM produced in TG-F1 mice were much smaller than those in non-TG-F1 mice, while there were no significant changes in total amounts of IgM in supernatants between TG-F1 and non-TG-F1 mice (Table 2).
|IgM anti-DNA antibodies (×10–1 U/ml)|
|Total IgM (μg/ml)|
2.6 In vitro proliferative response of B cells
Table 3 shows in vitro proliferative responses of splenic CD5+B220dull B1 cells to IL-5. In keeping with earlier findings 12, B1 cells from non-TG-F1 mice were hyper-responsive to IL-5 and showed high proliferative responses. B1 cells from TG-F1 mice also showed comparable levels of IL-5 hyper-responsiveness, indicating that this property of B1 cells in (NZB×NZW) F1 mice remains unchanged in TG-F1 mice. B2 cells from both TG-F1 and non-TG-F1 mice were negative for proliferative response to IL-5 (data not shown). Consistent with the finding that IL-2Rα expression was markedly up-regulated on the majority of B1 cells from TG-F1 mice (Fig. 6), proliferative responses to IL-2 of TG-F1 B1 cells were significantly higher than those of non-TG-F1 B1 cells.
|Mice||[3H]Thymidine incorporation (×103 cpm)|
The transgene-mediated hyper-expression of IL-5 unexpectedly suppressed autoimmune disease in TG-F1 mice. The transgene, in some cases, is known to reduce mature B cells and to suppress antibody responses of B cells 19, 20. Here, however, this may not be the case, because there is evidence that the original IL-5-transgenic C3H mice, a donor of the transgene, showed a significant expansion of B1 cells in association with production of large amounts of IgM anti-DNA antibodies 8. As is the case in IL-5-transgenic C3H mice, our TG-F1 mice showed an eosinophilia and a marked expansion of B1 cells; nevertheless, autoantibody production was strikingly down-regulated. It is also unlikely that certain suppressive genetic elements for autoantibody synthesis were introduced from the C3H strain, because B1 cells from the non-TG-F1 littermates responded well to IL-5 in vitro, proliferated even in the absence of prior activation signals and produced large amounts of IgM anti-DNA antibodies, as was evident in wild-type (NZB×NZW) F1 mice 13.
Our data clearly showed that the reduction of anti-DNA antibody synthesis is not due to the suppressive effect of IL-5 transgene on polyclonal B cells, because total amounts of IgM and IgG in sera or those produced in vitro in B cell cultures were almost identical between TG-F1 and non-TG-F1 mice. There are several possibilities to account for these findings. TG-F1 mice may have skewed B1 cell repertoires. As IL-5 transduces signals for proliferation to polyclonal B1 cells in an antigen-independent manner, relative frequencies of anti-DNA clones in TG-F1 mice may become low in expanded polyclonal B1 cell repertoires, resulting in low serum levels of both IgM and IgG anti-DNA antibodies even after self antigen stimulation. Alternatively, self-reactive B1 cells in wild-type (NZB×NZW) F1 mice are relatively resistant to Fas-mediated apoptosis 21. Dysregulated, continuous stimulation by IL-5 may render such activated anti-DNA clones susceptible for Fas-mediated elimination or unresponsive to signals for antibody synthesis in TG-F1 mice. To our knowledge, however, there have been no reports suggesting the involvement of IL-5 in the induction of immunological tolerance.
Another possibility may relate to the observed aberrant up-regulation of IL-2Rα expression on a majority of B1 cells in TG-F1 mice (Fig. 6). B1 cells in (NZB×NZW) F1 mice are intrinsically hyper-proliferative in response to IL-2, although the IL-2 signal does not induce their differentiation 22, 23. One can speculate that in TG-F1 mice, up-regulated IL-2R expression induces skewed signaling for abnormal propagation of these anti-DNA clones, which eventually dysregulates their differentiation and affinity maturation; thus leading to reduction not only in anti-DNA IgM of low-avidity nature but also in affinity-selected anti-DNA IgG.
In addition, a continuous stimulation of self-antigen-activated B1 cells with large amounts of IL-5 may cause a skewed IL-5 receptor signaling, i.e. positive for proliferation and negative for differentiation and maturation. Several distinct signaling cascades are involved in cell proliferation and maturation in the IL-5 receptor-signaling pathway. IL-5 activates a number of kinases, including Bruton-type tyrosine kinase (Btk), Janus kinase (Jak2), Lyn and Raf-1, as well as the phosphatase SHP2 24–28. Among these, Btk and Jak2 are essential for proliferation of B–lineage cell lines 24, 25. Stimulation of B cells with IL-5 together with CD38 ligation induces expression of the Blimp-1 gene, capable of driving the terminal differentiation of B cells into IgM-secreting cells 29, and induces an μ–γ1 switch recombination to IgG1 secretion in a STAT5-dependent manner 30. Whether there is any skewed signaling of the transgenic IL-5 in B1 cells of TG-F1 mice is now under investigation in our laboratory.
Although both IL-5-transgenic C3H mice and TG-F1 mice showed aberrant expansion of B1 cells, the extent was much greater in the latter. As (NZB×NZW) F1 mice have several susceptibility loci for abnormal expansion of B1 cells 31–35, IL-5 appears to augment this property of B1 cells in these mice. Lack of such genetic elements in C3H mice may account for the failure of the development of both SLE and B-CLL in IL-5-transgenic C3H mice 8. Considering that the IL-2R expression is highly up-regulated on B1 cells in TG-F1 mice, aberrant, progressive proliferation of B1 cells in TG-F1 mice can be attributed to combined effects of both IL-5 and IL-2 cytokines and the heretofore unidentified genetic elements. Together with the long-lived capacity 1, 2, such B1 cells are likely to be highly susceptible for accumulation of genetic alterations, giving rise to malignant transformation.
Our present model suggests that certain regulatory abnormalities for proliferation and/or differentiation of B1 cells are a crossroad between B-CLL and autoimmune disease. Patients with B-CLL frequently develop intercurrent systemic autoimmune diseases, and genetic factors may predispose to this event 36. In this context, we earlier found that while (NZB×NZW) F1 mice, carrying H-2d from NZB and H-2z from NZW, develop severe SLE 37–39, homozygous H-2z/z-congenic (NZB×NZW) F1 mice do not develop SLE, but instead show abnormal proliferation of B1 cells, giving rise to B-CLL 31, 32, the same phenotypic change of the disease observed in TG-F1 mice. Thus, it may be that both genetic and environmental factors can induce a certain shared dysregulation of B1 cells. Taken collectively, further studies of our TG-F1 mouse model may provide clues to molecular mechanisms regulating proliferation and differentiation of self-reactive B1 cells, as related to autoimmune disease and B-CLL.
4 Materials and methods
To obtain IL-5 transgene-congenic NZB and NZW mice, NZB or NZW mice were crossed with murine IL-5 cDNA-transgenic C3H/HeN mice 8, and F1 mice were selectively backcrossed with NZB or NZW mice for eight generations. IL-5-transgenic C3H/HeN mice constitutively show increased frequencies of peripheral eosinophils 8. Thus, mice with peripheral eosinophils of over 15% were regarded as positive for the IL-5 transgene. Percentages of eosinophils per total leukocytes were calculated under light microscopy with a blood smear stained with Giemsa solution. Selected mice were crossed to produce (NZB×NZW) F1 mice, and presence of the transgene was confirmed by Southern blot hybridization, using BamHI-digested tail DNA 8 (Fig. 1). The IL-5 probe used was the 461-bp SacI-AccI fragment of pSP6 K-mTRF23 40.
4.2 Measurement of IL-5
Anti-IL-5 mAb NC17 (100 μg/ml) was coated onto ELISA plate, and nonspecific binding sites were blocked with 2% BSA in PBS. Samples were then applied to wells and incubated overnight at room temperature. After washing wells with PBS containing 0.05% Tween-20, polyclonal rabbit anti-mouse IL-5 antibodies were added, followed by a 4-h incubation at room temperature. Then, peroxidase-conjugated goat anti-rabbit Ig and 2,2′-azino-di(3-ethyl)-benzthiazoline sulfonic acid were used as a coloring system. Optical density was read at 405 nm. A standard curve was obtained using purified rIL-5.
4.3 Cell counts and cytological examination
Peripheral leukocyte counts were made using a MEK-6158 Automatic Blood Cell Counter (Nihon Koden, Tokyo, Japan). Leukocyte-rich populations were separated from 40 μl of heparinized blood using a density gradient lymphocyte separator M-SMF (Japan Immuno Research Laboratories Co. Ltd., Takasaki, Japan). Leukocyte film for cytological examination was prepared using the Cytospin 3 Cell Preparation System (Shandon Scientific Ltd., UK), then stained with Giemsa.
4.4 FCM analysis
Frequency of peripheral CD5+ B1 cells per total B cells was examined using FACStarPLUS (Becton Dickinson, San Jose, CA). Peripheral blood was treated with ammonium chloride to remove red blood cells. Aliquots of 5×105–10×105 cells in 20 μl of PBS containing 0.2% BSA/0.05% NaN3 were incubated with FITC-labeled anti-mouse CD45R (B220) (clone RA3–6B2) and PE-labeled anti-mouse CD5 (clone 53–7.3) mAb at 4°C for 30 min. Percent frequency of CD5+ B1 cells per total B220+ B cells was calculated using the FITC/PE filter system.
Expression of IL-5Rα, IL-2Rα (CD25), CD80 and MHC class II (I-A) on B1 and B2 cells was examined by three-color FCM analysis. Spleen cell suspension was first incubated with FITC-labeled anti-mouse B220 (6B2) and PE-labeled anti-mouse CD5 mAb. Stained cells were further incubated with biotinylated mAb to IL-5Rα (H7) 41, CD25, CD80 (RM80) or I-A (m52) (K24–199), followed by allophycocyanin-labeled streptavidin. All incubations were done at 4°C for 30 min. Cells in the compartment with CD5+B220dull+ B1 and CD5–B220+ B2 phenotypes were gated and examined.
4.5 Cell sorting and culture
T cell-depleted spleen cells were obtained by treatment of cells with a mixture of mAb to CD4, CD8 and Thy1.2 plus rabbit complement at 37°C for 45 min. Contamination of T cells was under 2%. CD5+B220dull B1 cells were FCM-sorted after staining cells with FITC-conjugated anti-B220 and PE-conjugated anti-CD5 mAb. Purity of sorted B1 cells always exceeded 95%.
For in vitro antibody assay, aliquots of 2×105 cells in 200 μl of RPMI 1640 consisting 5×10–5 M 2-mercapoethanol, penicillin (100 U/ml), streptomycin (100 mg/ml) and 10% FCS were cultured in 96-well flat-bottomed plates in the absence or the presence of mIL-5 (50 U/ml) or mIL-6 (500 U/ml) at 37°C for 7 days. For the proliferation assay, 2×105 cells per well in 96-well round-bottomed plates were cultured in the absence or the presence of mIL-5 (50 U/ml) or hIL-2 (5,000 U/ml) at 37°C for 3 days with addition of 0.5 μCi of [3H]thymidine during the final 24 h and were then subjected to measurements of radioactivity.
4.6 Measurement of serum levels of Ig and class-specific anti-DNA antibodies
Isotype-specific Ig and anti-DNA autoantibodies were measured by ELISA with isotype-specific antibodies. The DNA-binding activities were expressed in U, referring to a standard curve obtained by serial dilution of a standard serum pool from 4- to 7-month-old NZB mice for IgM class and from over 9-month-old (NZB×NZW) F1 mice for IgG class, both containing 1,000 U activities/ml 39. Double-stranded DNA was obtained from calf thymus DNA (Sigma Chemical Co., St. Louis, MO).
4.7 Histopathology and immunofluorescence staining
Organs were removed from necropsied animals, fixed in 4% paraformaldehyde, embedded in paraffin, and stained with hematoxylin and eosin or PAS solution. For immunofluorescence staining, frozen sections were incubated for 60 min at room temperature with FITC-labeled goat antibodies to IgG or to C3 (ICN Pharmaceuticals, Inc., Aurora, OH).
4.8 Southern blot analysis for B cell clonality
Approximately 10 μg of genomic DNA extracted from liver or T cell-depleted spleen cells was digested with EcoRI, fractionated on 0.7% agarose gels, transferred to nitrocellulose filters, and probed with radiolabeled JH probe, a 2.0-kb BamHI/EcoRI fragment that induces JH3 and JH4 segments 42. Hybridizations were done in 6× SSPE/5× Denhardt's solution/0.5% SDS/100 μg/ml of salmon sperm DNA at 65°C and washed in 2× SSPE/0.1% SDS and 0.2× SSPE/0.1% SDS at 65°C (1× SSPE = 0.18 M NaCl/10 mM sodium phosphate, pH 7.4/1 mM EDTA; 1× Denhardt's solution = 0.02% BSA/0.02% Ficoll/0.02% polyvinylpyrrolidone).
The onset of renal disease was monitored by biweekly measurement of proteinuria, as described 37. A proteinuria of 111 mg/100 ml or more was regarded as being positive.
4.10 Statistical analysis
Statistical analysis was made using Student's t-test and a chi-square test. p values of <5% were considered to have a statistical significance.
We thank M. Ohara for language assistance. This work was supported in part by Grant-in-Aid for Scientific Research on Priority Areas and for COE Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and grant from The Organization for Pharmaceutical Safety and Research, Japan.