To determine the necessity for any individual BAFF receptor in the development of systemic lupus erythematosus (SLE).
To determine the necessity for any individual BAFF receptor in the development of systemic lupus erythematosus (SLE).
Bcma-, Taci-, and Br3-null mutations were introgressed into NZM 2328 mice. NZM.Bcma−/−, NZM.Taci−/−, and NZM.Br3−/− mice were evaluated for lymphocyte phenotype and BAFF receptor expression by flow cytometry; for B cell responsiveness to BAFF by in vitro culture; for serum levels of BAFF and total IgG and IgG anti–double-stranded DNA (anti-dsDNA) by enzyme-linked immunosorbent assay; for renal immunopathology by immunofluorescence and histopathology; and for clinical disease.
BCMA, TACI, and B lymphocyte stimulator receptor 3 (BR3) were not surface-expressed in NZM.Bcma−/−, NZM.Taci−/−, and NZM.Br3−/− mice, respectively. Transitional and follicular B cells from NZM.Br3−/− mice were much less responsive to BAFF than were the corresponding cells from wild-type, NZM.Bcma−/−, or NZM.Taci−/− mice. In comparison with wild-type mice, NZM.Bcma−/− and NZM.Taci−/− mice harbored an increased number of spleen B cells, T cells, and plasma cells, whereas serum levels of total IgG and IgG anti-dsDNA were similar to those in wild-type mice. Despite their paucity of B cells, NZM.Br3−/− mice had an increased number of T cells, and the numbers of plasma cells and levels of IgG anti-dsDNA were similar to those in wild-type mice. Serum levels of BAFF were increased in NZM.Taci−/− and NZM.Br3−/− mice but were decreased in NZM.Bcma−/− mice. Despite their phenotypic differences, NZM.Bcma−/−, NZM.Taci−/−, and NZM.Br3−/− mice had renal immunopathology and clinical disease that were at least as severe as that in wild-type mice.
Any single BAFF receptor, including BR3, is dispensable for the development of SLE in NZM mice. Development of disease in NZM.Br3−/− mice demonstrates that BAFF–BCMA and/or BAFF–TACI interactions contribute to SLE, and that a profound, life-long reduction in the numbers of B cells does not guarantee protection against SLE.
One of the characteristic features of systemic lupus erythematosus (SLE) is B cell hyperactivity. Accordingly, any factor that positively affects B cells has an a priori likelihood of playing a pathogenetic role in SLE. One such factor is BAFF (BLyS), a 285–amino acid type II transmembrane protein member of the tumor necrosis factor ligand superfamily (1, 2). In vitro and in vivo studies have demonstrated that BAFF is a vital B cell survival factor (3–5), and that BAFF promotes differentiation of immature B cells to mature B cells (6) and immunoglobulin class switching and production (7). Indeed, BAFF-deficient mice display marked global reductions in mature B cell numbers and in baseline and antigen-driven serum levels of immunoglobulins (8, 9).
The connection between BAFF and SLE is very strong. Constitutive overexpression of BAFF in non-autoimmune mice leads to SLE-like features, including elevated titers of multiple circulating autoantibodies and immune complex–mediated glomerulonephritis (10, 11). In human SLE, circulating BAFF levels are elevated in as many as 50% of patients (12–14), and BAFF expression correlates with disease activity (15, 16).
Importantly, elimination/neutralization of BAFF leads to prevention/amelioration of SLE. A genetic deficiency of BAFF protects SLE-prone NZM 2328 (NZM) mice from clinical disease (17), and both (NZB × NZW)F1 and MRL.lpr mice manifest enhanced survival in response to BAFF antagonists (11, 18, 19). Two phase III clinical trials of the anti-BAFF monoclonal antibody (mAb) belimumab in human SLE documented its efficacy and safety (20, 21), prompting the Food and Drug Administration to approve belimumab for the treatment of SLE (for review, see ref.22). In the phase III trials, a clinical response to belimumab was greatest among patients who were anti–double-stranded DNA (anti-dsDNA) positive and harbored low complement levels at baseline (23), suggesting that therapeutic benefit arising from BAFF neutralization is substantially mediated by inhibiting pathogenic autoreactive B cells and production of pathogenic autoantibodies.
BAFF has 3 receptors, BCMA, TACI, and B lymphocyte stimulator receptor 3 (BR3; also known as BAFF receptor), but it is not known which BAFF receptor(s) is required for the SLE-promoting effects of BAFF. Of note, single deficiency of the individual BAFF receptors in non-autoimmune mice yields markedly divergent phenotypes.
BCMA-deficient mice display a near-normal phenotype. They harbor normal numbers of lymphocytes and lymphocyte subsets, the in vitro functions of these cells are normal, and the mice manifest no obvious immunodeficiency (9, 24). However, immunized BCMA-deficient mice do not maintain as many antigen-specific long-lived immunoglobulin-secreting plasma cells in their bone marrow as do corresponding wild-type mice (25). In principle, BCMA deficiency in the context of SLE might reduce the numbers of pathogenic autoreactive long-lived plasma cells and, thereby, attenuate SLE.
Mice deficient in BR3 display a phenotype similar to that of BAFF-deficient mice, including reduced numbers of spleen B cells and mature recirculating B cells in the bone marrow and markedly reduced baseline and antigen-induced serum levels of immunoglobulins (26, 27). In bone marrow–chimeric mice harboring both wild-type B cells and B cells that bear a BR3 mutant, the B cells bearing the mutant BR3 manifest decreased survival (28). Collectively, these observations point to BAFF–BR3 interactions as essential for the prosurvival effects of BAFF on peripheral B cells. Given the central role for B cells in SLE, one could anticipate that BR3 deficiency would markedly attenuate SLE.
The phenotype of TACI-deficient mice dramatically differs from that of BR3- or BCMA-deficient mice. The B cell numbers are increased in TACI-deficient mice (29, 30) and, as these mice age, they experience elevated circulating titers of autoantibodies, immune complex–mediated glomerulonephritis, and premature death (31). Nevertheless, TACI deficiency could have a net down-modulatory effect in an SLE-prone host. TACI-deficient mice generate impaired immunoglobulin responses to T cell–independent antigens (29, 30), and expression of TACI may control the ability of B cells to produce immunoglobulins in response to BAFF (32). Although TACI deficiency leads to the expansion of follicular helper T cells, it concurrently impairs plasma cell survival and leads to a decreased number of long-lived plasma cells in the bone marrow (33, 34). Thus, TACI deficiency in an SLE-prone host may attenuate, rather than aggravate, SLE.
To directly assess the necessity for BCMA, TACI, or BR3 in the development of SLE, we generated NZM.Bcma−/−, NZM.Taci−/−, and NZM.Br3−/− mice. Despite considerable divergence among these mice in their lymphocyte and serologic phenotypes, all developed renal immunopathology and clinical disease to at least the same extent as did wild-type NZM mice. Thus, any single BAFF receptor, including BR3, is dispensable for the development of full-blown clinical SLE in NZM mice, and lymphocyte and serologic profiles can be dissociated from the development of severe pathologic and clinical disease. These observations may have important ramifications for B cell–directed therapies, and NZM.Br3−/− mice offer a model system in which full-blown clinical SLE develops in the face of profound long-term B cell depletion.
Female mice from 5 congenic NZM strains were studied: wild-type NZM, NZM.Baff−/−, NZM.Bcma−/−, NZM.Br3−/−, and NZM.Taci−/−. NZM.Baff−/− mice have previously been described (17). To generate NZM.Bcma−/−, NZM.Br3−/−, and NZM.Taci−/− mice, the Bcma−/−, Br3−/−, and Taci−/− genotypes from Bcma−/−, Br3−/−, and Taci−/− mice, respectively (9, 27, 29), were individually introgressed into NZM mice through a marker-assisted selection protocol, using microsatellite markers spanning the entire genome and including markers to ensure that all known SLE susceptibility loci remained intact (17, 35). Results are presented for mice from the N7 backcross generation or later, at which time all mice bore the NZM genotype at all tested markers. All mice were maintained in specific pathogen–free quarters at the University of Southern California (USC), and the experiments were approved by the USC Institutional Animal Care and Use Committee.
To detect the disrupted Bcma, Br3, or Taci gene fragments, genomic DNA extracted from mouse tail clippings was amplified by polymerase chain reaction (PCR) for 30 cycles at 94°C for 60 seconds, 55°C (62°C for Taci) for 60 seconds, and 72°C for 60 seconds. The primer sequences used are as follows: for Bcma, primer 1 was 5′-TCACTGTGGAAACACTGTTGCGCCATG-3′, primer 2 was 5′-GATATCCTGATCATCGGTCTTCAGATGC-3′, and primer 3 was 5′-CTGTCCACATTGCAACTGTTACCTGGG-3′; for Br3, primer 1 was 5′-CGCGGTTTCATTCTAGACTACAGGG-3′, primer 2 was 5′-ACACGCAGTTTCTCACCAGAGGGTCGAAGC-3′, and primer 3 was 5′-ATCCTGATCATCGGTCTTCAGATGC-3′; for Taci, primer 1 was 5′-CCTCAGGCCAGGAGCTTTTAGGGAGAA-3′, primer 2 was 5′-CCAGCATCCCCTCTGCTCTGGTTTTAT-3′, and primer 3 was 5′-CCTGGGTGGAGAGGCTTTTTGCTTCCT-3′. All 3 primers for each gene fragment were added to a single reaction mixture. The band sizes for the amplified PCR products are as follows: for intact Bcma, 325 bp; for disrupted Bcma, 425 bp; for intact Br3, 180 bp; for disrupted Br3, 379 bp; for intact Taci, 440 bp; for disrupted Taci, 300 bp.
To determine T cell and B cell subsets, spleen mononuclear cells were stained with combinations of fluorochrome-conjugated mAb specific for CD3, CD4, CD8, CD44, CD62L, B220, CD19, CD21, CD23, or CD69 (BD PharMingen or eBioscience) and analyzed by flow cytometry (36). For plasma cell determination, spleen or bone marrow cells were stained with a combination of fluorochrome-conjugated mAb specific for CD4, CD8, Gr-1, F4/80, B220, IgD, and CD138 (BD PharMingen or eBioscience). Plasma cells were taken as CD4−CD8−Gr-1−F4/80−IgD−B220−CD138+ cells. For BAFF receptor expression, B cells or plasma cells were additionally stained with fluorochrome-conjugated mAb specific for BCMA, TACI (both from R&D Systems), or BR3 (eBioscience).
Spleen mononuclear cells were stained with a combination of biotinylated mAb specific for CD4, CD8, F4/80, Gr-1, and TER-119 (BD PharMingen or eBioscience), and untouched B cells were isolated by magnetic sorting with a MiniMACS column and streptavidin microbeads (Miltenyi Biotec). The B cells were then stained with a combination of fluorochrome-conjugated mAb specific for CD19, B220, CD93, CD23, and CD21/35 (BD PharMingen, eBioscience, or BioLegend) and sorted. Transitional B cells were taken as CD19+B220+CD93+CD21/35− cells, and follicular B cells were taken as CD19+B220+CD93−CD23+CD21/35− cells. Sorted B cells (1 × 105/0.2 ml) were cultured in triplicate in 96-well flat-bottomed plates in RPMI 1640 medium plus 10% fetal calf serum, with or without exogenous recombinant human BAFF (rhBAFF; 100 ng/ml) (Human Genome Sciences). The viability of the cultured cells was determined at 48 hours by staining with DAPI (Invitrogen).
Serum levels of total IgG and anti-dsDNA antibodies were determined by enzyme-linked immunosorbent assay (ELISA) (37). Autoantibody optical density (OD) values were normalized to the mean OD of serum from 5-month-old MRL.lpr mice, with the latter arbitrarily assigned a value of 100 units/ml. BAFF expression in serum was measured by ELISA, as previously described (38).
Sections of formalin-fixed kidneys were stained with hematoxylin and eosin, Masson's trichrome, and Jones' silver methenamine and assessed by light microscopy for glomerular activity (hypercellularity, necrotizing lesions, karyorrhexis, cellular crescents, hyaline deposits), tubulointerstitial activity (interstitial cellular infiltration, tubular cell necrosis), chronic glomerular pathology (glomerulosclerosis, fibrous crescents), and chronic tubulointerstitial pathology (tubular atrophy, interstitial fibrosis). Each category was subjectively scored on a 0–3-point scale, for a maximum composite score of 12 (39).
Sections of snap-frozen kidneys were stained for IgG or C3 deposition using the fluorescein isothiocyanate–conjugated goat F(ab′)2 fragment of anti-mouse IgG or C3 antibodies (MP Biomedicals) (37).
Albustix reagent strips for urinary protein (Bayer) were dipped in mouse urine and assigned a score (range 0–4) by visual color comparison with the supplied standard color key. Severe proteinuria was defined as a score of ≥3 on 2 consecutive examinations.
All analyses were performed using SigmaStat software (SPSS). Parametric testing between 2 groups was performed by the unpaired t-test, and parametric testing among 3 or more groups was performed by one-way analysis of variance (ANOVA). When the data were not normally distributed or the equal variance test was not satisfied, nonparametric testing was performed by the Mann-Whitney rank sum test between 2 groups and by Kruskal-Wallis one-way ANOVA among 3 or more groups.
To confirm at the protein level that the specific BAFF receptor was not expressed, spleen B (CD19+) cells from wild-type NZM, NZM.Bcma−/−, NZM.Taci−/−, and NZM.Br3−/− mice were evaluated for surface expression of the individual BAFF receptors (Figure 1A). TACI was surface-expressed in all mice except NZM.Taci−/− mice, and BR3 was surface-expressed in all mice except NZM.Br3−/− mice. Surface BCMA was not detected in any of the mice, which is consistent with BCMA being undetectable on non–terminally differentiated mature B cells in non-autoimmune mice (40, 41). Therefore, we analyzed spleen and bone marrow plasma cells and documented that BCMA was surface-expressed in wild-type NZM mice but was specifically not expressed in NZM.Bcma−/− mice (Figure 1B).
The loss of individual BAFF receptors led to highly divergent serum levels of BAFF. Whereas serum levels of BAFF were much higher in NZM.Taci−/− and NZM.Br3−/− mice than in wild-type NZM mice, they were decreased in NZM.Bcma−/− mice (Figure 1C). As expected, BAFF was not detected in sera from NZM.Baff−/− mice.
The differences in serum levels of BAFF among NZM.Bcma−/−, NZM.Taci−/−, and NZM.Br3−/− mice were associated with only limited effects on B cell surface expression of the reciprocal 2 BAFF receptors. BR3, which was absent in NZM.Br3−/− mice, was highly and equally surface-expressed by the B cells of NZM.Bcma−/−, NZM.Taci−/−, and wild-type NZM mice. TACI, which was absent in NZM.Taci−/− mice, was highly expressed in the other mice, albeit expression was somewhat lower in NZM.Br3−/− mice than in NZM.Bcma−/− or wild-type NZM mice (Figure 1A). Surface expression of any BAFF receptor on T cells was minimal in all of the mice and was not detectably up-regulated or down-regulated (data not shown).
The clinical courses in NZM.Bcma−/−, NZM.Taci−/−, and, most surprisingly and unexpectedly, NZM.Br3−/− mice were indistinguishable from that in wild-type NZM mice. No differences among the mouse cohorts were detected in development of either severe proteinuria or mortality. Severe proteinuria began to develop in each of the cohorts at 4–5 months of age, and ∼90% of the mice in each cohort were affected by 12 months of age. Mortality in each of the cohorts was noted as early as 6–7 months of age, with ∼90% of the mice in each cohort being dead by 12 months of age (Figure 2A).
Development of renal immunopathology in NZM mice deficient in one BAFF receptor, including NZM.Br3−/− mice, was consonant with their clinical courses. Glomerular deposition of IgG and C3 was as robust in NZM.Br3−/− mice as in wild-type NZM mice, with substantial IgG and C3 deposition in the glomeruli of NZM.Bcma−/− and NZM.Taci−/− mice as well (Figure 2B). Moreover, renal pathology, including glomerular hypercellularity, glomerular crescents, mesangial matrix deposition, interstitial inflammation and fibrosis, tubular atrophy, and perivascular leukocyte infiltration, was at least as severe in all the BAFF receptor–deficient NZM mice (including NZM.Br3−/− mice) as in wild-type NZM mice (Figure 2C). Indeed, renal histology scores were actually higher in NZM.Taci−/− and NZM.Br3−/− mice than in wild-type NZM mice at 5 months of age and were still higher in NZM.Taci−/− mice than in wild-type NZM mice at 8 months of age (Figure 2D).
Notwithstanding the similar clinical courses and renal immunopathologic findings in NZM.Br3−/−, NZM.Taci−/−, and NZM.Bcma−/− mice, their B cell phenotypes were very different from each other and from that of wild-type NZM mice. In comparison with the numbers in wild-type NZM mice, the numbers of total, follicular, and marginal zone B cells were greatly reduced in NZM.Br3−/− mice (Figure 3A, parts a–c). Of note, the numbers of activated (CD69+) B cells and B cells expressing the costimulatory CD80 molecule were similar in NZM.Br3−/− and wild-type NZM mice (Figure 3A, parts d and e), whereas the difference between NZM.Br3−/− and wild-type NZM mice in the number of B cells expressing the costimulatory CD86 molecule was significant (Figure 3A, part f). Although indisputably abnormal, the B cell phenotype of NZM.Br3−/− mice was not as severe as that of NZM.Baff−/− mice; the numbers of total B cells and every B cell subset tested (other than marginal zone B cells) were significantly greater in NZM.Br3−/− mice than in NZM.Baff−/− mice (Figure 4, parts a–f).
In contrast to the reductions in the number of total B cells and several B cell subsets in NZM.Br3−/− mice, total B cells and individual B cell subsets were expanded in NZM.Taci−/− and NZM.Bcma−/− mice (Figure 3A). The numbers of total, follicular, and marginal zone B cells were all greatly increased in NZM.Taci−/− mice, as were CD69+, CD80+, and CD86+ B cells. Total B cells were also expanded in NZM.Bcma−/− mice, albeit not to the extent as in NZM.Taci−/− mice. Whereas the number of marginal zone B cells in NZM.Bcma−/− mice was not different from that in wild-type NZM mice, the expansion of CD69+, CD80+, and CD86+ B cells in NZM.Bcma−/− mice was very similar to that in NZM.Taci−/− mice.
The T cell profiles among NZM.Br3−/−, NZM.Taci−/−, and NZM.Bcma−/− mice were less divergent than the B cell profiles. The numbers of total, CD4+, CD8+, and CD4+ naive cells were increased in NZM mice deficient in any one BAFF receptor relative to those in wild-type NZM mice (Figure 3B, parts a–d). Moreover, the number of CD4+ memory cells was increased in NZM.Bcma−/− mice but not in NZM.Taci−/− or NZM.Br3−/− mice (Figure 3B, part e). The expansion of CD4+ memory cells in NZM.Bcma−/− mice was proportionate to the expansion of CD4+ naive cells, resulting in CD4+ naive:memory cell ratios in these mice consistent with those in wild-type NZM mice, whereas the expansion of CD4+ naive cells in NZM.Br3−/− mice, and to a lesser extent in NZM.Taci−/− mice, was out of proportion to the expansion of CD4+ memory cells, leading to increased CD4+ naive:memory cell ratios (Figure 3B, part f). In striking contrast to the T cell profile in NZM.Br3−/− mice, the numbers of total T cells and T cell subsets were markedly reduced in NZM.Baff−/− mice (Figure 4, parts g–k). This does not appear to reflect a difference in T cell activation between these mice, because no difference in their CD4+ naive:memory ratios was observed (Figure 4, part l).
The increased circulating BAFF levels in NZM.Br3−/− mice coupled to a marked reduction in the number of B cells strongly suggested that such B cells were highly unresponsive to BAFF. To confirm this impression, transitional B cells and follicular B cells from NZM.Bcma−/−, NZM.Taci−/−, NZM.Br3−/−, and wild-type NZM mice were cultured with or without exogenous rhBAFF and assessed for viability after 48 hours. Whereas rhBAFF enhanced the in vitro survival of transitional and follicular B cells from NZM.Bcma−/−, NZM.Taci−/−, and wild-type NZM mice, rhBAFF had no discernible effect on these cells from NZM.Br3−/− mice (Figure 5). Of interest is the observation that in vitro survival of B cells (especially follicular B cells) from NZM.Bcma−/− or NZM.Taci−/− mice in either the absence or presence of rhBAFF was greater than that of B cells from wild-type NZM mice; this finding is consistent with the greater numbers of spleen B cells in NZM.Bcma−/− or NZM.Taci−/− mice compared with wild-type NZM mice (Figure 3A).
In NZM.Baff−/− mice, the reduced number of B cells is paralleled by reduced numbers of spleen and bone marrow plasma cells (17, 39). Strikingly, no diminution in the numbers of either spleen or bone marrow plasma cells was observed in NZM.Br3−/− mice, despite their low numbers of B cells (Figures 6A and B). In NZM.Bcma−/− and NZM.Taci−/− mice, the number of spleen plasma cells was greater than that in wild-type NZM mice, although bone marrow plasma cell numbers were similar across all of the mouse cohorts.
Neither serum levels of total IgG nor serum levels of IgG anti-dsDNA immutably paralleled plasma cell numbers. Despite the differences in spleen plasma cell numbers, the serum levels of total IgG and IgG anti-dsDNA in NZM.Bcma−/− and NZM.Taci−/− mice did not differ from those in wild-type NZM mice, at any age tested (Figures 6C and D). In contrast, serum levels of total IgG and IgG anti-dsDNA were significantly reduced in NZM.Br3−/− mice compared with WT mice at 4–6 months of age, in spite of the fact that the number of plasma cells in these mice was similar to the number in WT mice. Of note, by 7–9 months of age, serum levels of IgG anti-dsDNA in NZM.Br3−/− mice were no longer lower than those in age-matched wild-type NZM mice, while serum levels of total IgG in NZM.Br3−/− mice remained substantially reduced.
Whereas the contribution of BAFF to SLE has been firmly established (11, 17–21), the contribution of any individual BAFF receptor to SLE had not previously been investigated. To that end, we generated and investigated SLE-prone NZM mice deficient in BCMA, TACI, or BR3, and several remarkable conclusions can be drawn from our findings.
First, the B cell phenotypes are highly divergent. Whereas B cells are expanded in NZM.Bcma−/− and NZM.Taci−/− mice, they are dramatically reduced in NZM.Br3−/− mice. These observations are congruent with the B cell expansion observed in TACI-deficient non-autoimmune mice (29, 30), the B cell expansion in BCMA-deficient B6.lpr and B6.Nba2 mice (42), and the greatly reduced number of B cells in BR3-deficient non-autoimmune mice (26, 27). Because BCMA and TACI each bind APRIL (43) and, therefore, could normally serve as a “sink” for APRIL, increased circulating APRIL levels in NZM.Bcma−/− and NZM.Taci−/− mice may have contributed to the B cell expansion in these mice. Because, unfortunately, repeated attempts at developing a quantitative ELISA for murine APRIL were unsuccessful, we were not able to measure serum levels of APRIL.
Of note, the reduction in the number of B cells in NZM.Br3−/− mice is not as great as that in NZM.Baff−/− mice. Although engagement of BR3 by BAFF indisputably plays a vital role in B cell survival (26–28), our observations suggest a modest yet discernible contribution to B cell survival from engagement of TACI and/or BCMA as well. Because the B cell population is expanded instead of contracted in NZM.April−/− mice (39), it is likely that BAFF–TACI and/or BAFF–BCMA interactions, rather than APRIL–TACI and/or APRIL–BCMA interactions, are the relevant interactions.
Second, there is a divergence in the numbers of spleen and bone marrow plasma cells and in serum levels of total IgG and IgG anti-dsDNA among the BAFF receptor–deficient NZM mice. Although spleen plasma cells are modestly increased in NZM.Bcma−/− and NZM.Taci−/− mice, bone marrow plasma cells are not. This dichotomy between spleen plasma cells and bone marrow plasma cells suggests that regulation of spleen plasma cells is more dependent on BCMA and TACI than is regulation of bone marrow plasma cells. Additional experimentation is needed to evaluate this possibility.
The finding that the numbers of bone marrow plasma cells in NZM.Bcma−/− and NZM.Taci−/− mice were similar to those in WT mice was unexpected, because BCMA- or TACI-sufficient mice have more long-lived antigen-specific bone marrow plasma cells compared with BCMA- and TACI-deficient non-autoimmune mice (25, 33, 34). The nature of the antigen inducing the long-lived plasma cell response may be important. BCMA and TACI may play vital roles in long-lived plasma cell responses to foreign proteins (predominantly T cell–dependent responses), whereas they may play a small role in long-lived plasma cell responses to self polynucleotides (predominantly T cell–independent responses). Studies are currently under way to delineate what, if any, differential role BCMA and TACI have in T cell–dependent versus T cell–independent plasma cell responses.
In any case, serum levels of total IgG and IgG anti-dsDNA in NZM.Bcma−/− and NZM.Taci−/− mice are no different from those in wild-type NZM mice. Our findings contrast with the increased circulating levels of autoantibodies in BCMA-deficient B6.lpr or B6.Nba2 mice (42) and with the profound serologic autoimmunity that develops in TACI-deficient non-autoimmune mice (31). Our results highlight the importance of in vivo models of bona fide SLE when investigating a given molecule or biologic pathway. Results obtained in non-autoimmune models or “incomplete SLE” models may not accurately predict outcomes in bona fide SLE.
Most surprisingly, NZM.Br3−/− mice have numbers of spleen and bone marrow plasma cells similar to those observed in WT mice. Whereas the number of plasma cells in NZM.Baff−/− mice is substantially reduced in parallel with the marked reduction in the number of B cells (17, 39), the reduction in B cells in NZM.Br3−/− mice is not accompanied by reductions in spleen or bone marrow plasma cells. Because expression of BR3 declines and expression of BCMA increases as B cells differentiate into plasmablasts/plasma cells in BAFF receptor–intact hosts (25, 44), the engagement of BCMA (and/or TACI) by the increased levels of BAFF in NZM.Br3−/− mice appears capable of supporting plasma cell responses similar to those in WT mice. Although transitional B cells and follicular B cells in NZM.Br3−/− mice are much less responsive to BAFF in terms of survival than are the corresponding cells in NZM.Bcma−/−, NZM.Taci−/−, or wild-type NZM mice, BAFF continues to promote the differentiation of the small fraction of B cells that do survive in the absence of BR3.
Of interest, the serum levels of total IgG in NZM.Br3−/− mice at 4–6 months of age are lower than those in age-matched wild-type NZM mice, and this difference is even more apparent at 7–9 months of age. Importantly, whereas serum levels of IgG anti-dsDNA are initially lower in NZM.Br3−/− mice than in wild- type NZM mice, they “catch up” by 7–9 months of age. One possibility is that autoreactive B cells are more dependent on BAFF than are non-autoreactive B cells (45, 46). The elevated levels of circulating BAFF in NZM.Br3−/− mice may preferentially support the autoreactive B cells within the contracted B cell pool, so that the increases in autoantibody titers are out of proportion to increases in total IgG. Alternatively, there may be a BAFF-independent mechanism that drives autoantibody production similar to that in NZM.Baff−/− mice, which develop robust serologic autoimmunity over time (17). APRIL may be one such contributor, given that serum IgG autoantibody levels are lower in NZM.Baff−/−.April−/− mice than in NZM.Baff−/− mice (39).
Third, T cell expansion is a feature of NZM mice deficient in any single BAFF receptor. This contrasts with the normal numbers of T cells in the corresponding non-autoimmune mice (9, 24, 26, 27, 29, 30) and the decreased number of T cells in NZM.Baff−/− mice (17, 39). In light of the substantial reduction in the number of total B cells in both NZM.Baff−/− mice and NZM.Br3−/− mice, the divergent T cell profiles are especially noteworthy. It may be that the preservation of activated B cells in NZM.Br3−/− mice is crucial to T cell expansion. Regardless, the T cell response in NZM.Bcma−/−, NZM.Taci−/−, and NZM.Br3−/− mice is not monolithic. CD4+ naive cells expanded similarly in all of these mice, whereas CD4+ memory cells significantly expanded only in NZM.Bcma−/− mice. Expansions of both CD4+ naive and memory cells were also observed in BCMA-deficient B6.lpr mice (42).
Fourth, serum levels of BAFF among the BAFF receptor–deficient NZM mice are highly divergent. In NZM.Taci−/− and NZM.Br3−/− mice, serum levels of BAFF are increased. This is expected, because the loss of BAFF receptor “sinks” for circulating BAFF should lead to decreased removal of BAFF from the circulation. The ensuing increased BAFF levels likely contribute to the B cell and T cell expansions observed in NZM.Taci−/− mice and to the T cell expansion in NZM.Br3−/− mice, because BAFF not only is a potent B cell survival factor but also has agonist effects on T cells (47, 48).
In contrast, serum levels of BAFF in NZM.Bcma−/− mice are lower than those in wild-type NZM mice. BCMA expression on mature B cells is very low (40, 41), so BCMA is much less of a “sink” for BAFF than is either TACI or BR3. As a result, BAFF levels in NZM.Bcma−/− mice might be expected to be equal to those in wild-type NZM mice. Why BAFF levels are actually lower in NZM.Bcma−/− mice remains unknown, and further investigation is warranted.
Regardless, the reduced serum levels of BAFF in NZM.Bcma−/− mice sharply contrast with the increased serum levels of BAFF in BCMA-deficient B6.lpr mice (42). BCMA deficiency in combination with the dysregulated Fas pathway in lpr mice may promote BAFF production, whereas BCMA deficiency in the context of an intact Fas pathway may not. Because B cells and T cells are expanded in both NZM.Bcma−/− and B6.lpr mice despite the marked disparity in serum levels of BAFF, B cell and/or T cell expansion in a BCMA-deficient environment may be relatively BAFF independent. Again, further investigation is warranted.
Fifth, the cellular and serologic differences among the BAFF receptor–deficient NZM mice notwithstanding, development of renal immunopathology in these mice was at least as severe as that in wild-type NZM mice and, most strikingly, the time courses of clinical disease among all the mice were indistinguishable. That is, even though NZM.Br3−/− mice (which bear a disrupted Br3 gene) do not express BR3 on their B cells, harbor vastly reduced numbers of mature B cells, and have transitional and follicular B cells that are markedly less responsive to BAFF in terms of survival than are the corresponding cells from NZM.Bcma−/−, NZM.Taci−/−, or wild-type NZM mice, they develop clinical and pathologic SLE that is at least as extensive as that in wild-type NZM mice. This unequivocally demonstrates that any single BAFF receptor, including BR3, is dispensable for the development of clinical SLE, and that the presence of any 2 BAFF receptors is sufficient for full development of clinical and pathologic disease. NZM mice expressing only BCMA, TACI, or BR3 (i.e., deficient in 2 BAFF receptors) are currently being generated to determine whether any single BAFF receptor is sufficient for development of SLE.
Our findings in NZM.Br3−/− mice may explain an apparent clinical paradox and, thereby, have important therapeutic ramifications for human SLE. Treatment with belimumab, which results in only modest depletion of B cells, was successful in 2 independent phase III SLE trials (20, 21), whereas treatment with rituximab, which results in considerable B cell depletion, failed to achieve its end points in 2 independent phase II/III SLE trials (49, 50). Because circulating BAFF levels in patients with SLE increase following rituximab treatment (51), these increased BAFF levels may preferentially drive the remaining autoreactive B cells and promote ongoing disease activity. Therapeutic approaches that inactivate B cells rather than depleting them may circumvent this problem by not triggering a rise in BAFF levels. Such agents are in development (52), and clinical trials with them should be highly informative.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Drs. C. O. Jacob and Stohl had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. C. O. Jacob, Quinn, Cancro, Stohl.
Acquisition of data. C. O. Jacob, Yu, Guo, N. Jacob, Quinn, Sindhava, Goilav, Putterman, Migone.
Analysis and interpretation of data. C. O. Jacob, Guo, N. Jacob, Quinn, Sindhava, Cancro, Stohl.
Author Mignone is an employee of Human Genome Sciences.