B Cell Depletion with Anti-CD79 mAbs Ameliorates Autoimmune Disease in MRL/lpr Mice1

MRL/lpr mice develop a spontaneous systemic lupus erythematosus-like autoimmune syndrome due to a dysfunctional Fas receptor, with contributions from other less well-defined genetic loci. The removal of B cells by genetic manipulation not only prevents autoantibody formation, but it also results in substantially reduced T cell activation and kidney inflammation. To determine whether B cell depletion by administration of Abs is effective in lupus mice with an intact immune system and established disease, we screened several B cell-specific mAbs and found that a combination of anti-CD79α and anti-CD79β Abs was most effective at depleting B cells in vivo. Anti-CD79 therapy started at 4–5 mo of age in MRL/lpr mice significantly decreased B cells (B220+CD19+) in peripheral blood, bone marrow, and spleens. Treated mice also had a significant increase in the number of both double-negative T cells and naive CD4+ T cells, and a decreased relative abundance of CD4+ memory cells. Serum anti-chromatin IgG levels were significantly decreased compared with controls, whereas serum anti-dsDNA IgG, total IgG, or total IgM were unaffected. Overall, survival was improved with lower mean skin scores and significantly fewer focal inflammatory infiltrates in submandibular salivary glands and kidneys. Anti-CD79 mAbs show promise as a potential treatment for systemic lupus erythematosus and as a model for B cell depletion in vivo.


B cells and T cells are both important in the pathogenesis of lupus. Recent work has emphasized the complexity of the functions of B cells in the genesis of autoimmunity.
Beside their essential role as producers of Abs, it is now apparent that B cells can serve as regulatory cells, APCs (1), producers of cytokines (2), and as essential elements in lymphogenesis (3,4). Regulation may occur by diverse mechanisms, which include enhancing or suppressing effects of Abs or other soluble factors on other B cells or on other key regulatory elements of the immune system (5)(6)(7).
Accordingly, experiments using B cell-deficient MRL/lpr mice suggest that B cells may be of key importance in autoimmunity through functions extending beyond autoantibody production. MRL/lpr mice lacking B cells (J H D Ϫ MRL/lpr) showed substantial reduction of T cell inflammatory renal infiltrates and prolonged survival as compared with MRL/lpr mice (8). Additionally, depletion of lpr B cells in mixed chimeras reduced both autoantibody production and T cell proliferation, despite the presence of ϩ/ϩ B cells (9). In MRL/lpr mice engineered to have B cells expressing surface-bound (sIg-MRL/lpr) but not secretory Ig, however, much of this inflammatory component of disease was restored, notably T cell infiltrative lesions (10). B cells or their products other than Ig thus appear to regulate disease aspects of lupus in the MRL/lpr mouse model. B cell depletion therapy is being tried in a number of human autoimmune diseases, including systemic lupus erythematosus (SLE), 3 and has been approved for the treatment of rheumatoid arthritis (11). It is disappointing that such approaches in murine models have been difficult, both because of limited options for depleting Abs and because of the demonstrated resistance of SLE mice to depletion (12).
We have therefore screened a variety of B cell-specific mAbs in C57BL/6 and MRL/lpr mice for B cell depletion. We found that anti-CD79␣ and CD79␤ mAbs were particularly effective in depleting B cells. In this study, we describe the in vivo effects of anti-CD79␣ and anti-CD79␤ in MRL/lpr mice.

Materials and Methods
Mice MRL/lpr, MRL/lpr-Thy1.1, MRL/lpr-IgH b , and C57BL/6 (B6) mice were maintained in our barrier animal facility. These congenic strains of MRL/ lpr mice were used for experiments as they became available. Breeders of MRL/lpr were from The Jackson Laboratory. Breeders of MRL/lpr-Thy1.1 were from Dr. S. Ikehara (Osaka, Japan). MRL/lpr-IgH b lpr were developed in our facility (13).
For initial screening of Abs for their ability to deplete B cells, B6, B6/lpr, and MRL/lpr mice were given 1 mg of mAb or an equal volume of PBS by i.p. injection. Eight days after mAb treatment, mice were sacrificed and splenic B cells were enumerated.
Sera were collected every 2-3 wk and stored at Ϫ20°C for analyses of autoantibody levels, total serum IgG, and IgM by ELISA, as described below. At the end of 6 wk of treatment, kidney and salivary glands were explanted, fixed in PBS-buffered formalin, processed for H&E staining, and examined histologically.

ELISA for autoantibodies against chromatin and dsDNA
Polyvinylchloride microtiter plates (Dynatech Laboratories) were coated with 100 l borate-buffered saline (pH 8.5) containing 3 g/ml chicken chromatin purified from chicken erythrocyte nuclei (9) or 3 g/ml dsDNA purified from calf thymus (Sigma-Aldrich). Coated plates were allowed to incubate overnight at 4°C. Plates were washed five times in BBS and then incubated with 200 l coating buffer (borate-buffered saline, 0.4% Tween 80, 0.5% BSA, filtered) for 1 h at 4°C. Coating buffer was removed, and 100 l diluted sera was added and incubated at 4°C overnight. Serum samples were diluted 1/500 for anti-chromatin and 1/1000 for anti-dsDNA. A reference serum, which was pooled from older MRL/lpr mice with high titers of autoantibodies, was diluted serially 2-fold from 1/500 to 1/1,024,000. Plates were washed, and 100 l of 0.2 g/ml biotinylated-F(abЈ) 2 goat anti-mouse IgG (Fc chain-specific, Jackson ImmunoResearch Laboratories) was added. After a 1-h incubation at room temperature, plates were washed and incubated with avidin-alkaline phosphatase (Zymed Laboratories) for an additional 1 h at room temperature. Plates were washed and paranitrophenyl phosphate substrate (Sigma-Aldrich) in 0.01 M diethanolamine (pH 9.8) was added. Plates were read at various time points with an automated ELISA reader (Molecular Devices).
Autoantibody results from individual ELISAs were standardized against the reference serum, and the results were expressed as the logarithm of equivalent dilution factor (EDF), where EDF ϭ (the dilution of the reference MRL/lpr serum that gives an OD equal to that of the sample serum) ϫ 10 6 (9).
ELISA for mouse anti-hamster IgG, total IgG, and total IgM ELISA was performed as described above with the following exceptions: purified Syrian hamster IgG (MP Biomedicals) at 2 g/ml, goat anti-mouse IgG at 3 g/ml, or donkey anti-mouse IgM at 4 g/ml (both from Jackson ImmunoResearch Laboratories) were used as coating Ags for mouse antihamster IgG, total IgG, and total IgM, respectively. Serum samples were diluted 1/600 for mouse anti-hamster IgG, 1/2,000,000 for total IgG, and 1/200,000 for total IgM. The biotinylated detecting Abs were goat antimouse pFcЈ for IgG and donkey anti-mouse IgM F(abЈ) 2 for total IgM (both from Jackson ImmunoResearch Laboratories). Anti-hamster IgG titer was compared based on raw OD of sera from different treatment groups. Total IgG and IgM were determined based on standard curves developed using mouse IgG (clone HB63) and mouse IgM (clone TEPC 183) in these assays.

Evaluation of nephritis and sialadenitis
Mouse kidneys and salivary glands were fixed in formalin, embedded in paraffin, sectioned, and stained with H&E. Focal inflammatory infiltrates in kidneys and submandibular salivary glands were expressed as number of foci per 4 ϫ 4 mm area of tissue. Additionally, renal pathology was quantified as previously described (15). Briefly, for light microscopy, the se-  (Table II, Expt. I). Shown are means and SEM of the percentages of peripheral blood lymphocytes that were (A) B220 ϩ CD19 ϩ cells; (B) B220 ϩ cells with surface-bound hamster Ab (B220 ϩ hIgG ϩ ), as detected by anti-hamster Ab; and (C) B220 ϩ cells that bound ex vivo added FITC-anti-CD79␤ (B220 ϩ CD79b ϩ ). PBS (n ϭ 3-5 mice); hamster IgG group (n ϭ 3-6); AB (n ϭ 6 -8). Numbers of mice were less at later time points either because mice were deceased or sacrificed to examine kidney histology. Except for week 0, the differences between AB and control groups were significant at p Ͻ 0.00001 for each time point.  (5) a Age at the start of treatment. b AB group received 0.5 mg each of F11-172 (anti-CD79␣) and HM79-16 (anti-CD79␤). verity of glomerular, interstitial, and vascular lesions was determined independently by blinded grading by one of us (M.M.) on a scale of 0 to 4ϩ. Multiple sections at a minimum of two different levels were examined. Each section typically involved evaluation of Ͼ50 glomeruli, Ͼ25 blood vessels, and the interstitium contained within two to three longitudinal sections of kidney.

Evaluation of dermatitis
Skin lesions were monitored weekly and were scored on a scale of 0 to 2ϩ. Mice without notable hair loss or skin lesions were scored as 0; mice with hair loss only were scored as 1; and mice with hair loss and inflammatory skin lesions were scored as 2.

Statistical and data analyses
Two-tailed Student's t test or ANOVA were performed on numerical data. Two-tailed Fisher's exact test was performed on kidney scores and skin lesion scores. 2 test was performed on survival data. For serum autoantibody levels, analyses were conducted using mixed effects models that took into account the correlation between repeated measurements on the same animal. Tests of significance were based on likelihood ratio or Wald tests (16). In brief, we determined whether there was evidence first of changes over time, second whether there were differences between groups, and last whether differences among groups persisted over time. We were able to fit a linear model to the data; that is, we assumed that any changes over time would follow an approximately straight line. Other types of models (e.g., where we looked for differences at specific times and did not assume the linear effect over time) did not yield substantially different results compared with the mixed effect model. Analysis was conducted only on animals that completed the study, and mice that did not survive the full course of experiment or those sacrificed before the end of experiment were excluded. This analysis is most conservative, as we excluded mice that died during the experiment and presumably were the most affected.

Anti-CD79␣ and anti-CD79␤ mAbs, but not other anti-B cell Abs, deplete MRL/lpr B cells
We screened several anti-B cell Abs in B6, B6/lpr, and MRL/lpr mice for in vivo B cell depletion by measuring splenic B220 ϩ sIgM ϩ B cells 1 wk after an injection of 1 mg anti-B cell Ab. Some anti-mouse B cell mAbs (anti-CD22 or anti-CD23) were entirely nondepleting even in B6 mice, whereas others (anti-CD19 and anti-CD45R/B220) depleted only in B6 and B6/lpr but not in MRL/lpr mice (Table I). Anti-CD19 and anti-CD21 given alone (0.5 mg each, three weekly injections) or together did not affect MRL/lpr splenic B cell numbers (not shown). Daunomycin-conjugated Abs (anti-CD45R/B220, anti-CD23, or anti-CD24) injected at 0.3 mg per mouse weekly for 3 wk also failed to deplete B cells in MRL/lpr mice (Table I and data not shown). Despite the fact that some of these Abs were ineffective at depleting B cells, all of the mAbs could be demonstrated by FACS to have coated B cells in vivo (data not shown). Only anti-CD79␤ (HM79-16) was found to deplete splenic B cells in B6 and B6/lpr mice and in MRL/lpr mice 8 days after an injection of 1 mg Ab (Table I). Additionally, B6 mice treated with increasing doses of anti-CD79␤ (0.1, 0.3, 0.5, or 1 mg) showed increasing amounts of B cell depletion (ϳ40% depletion for 0.1 mg and 60% for 1 mg), with maximum depletion level achieved by 0.5 mg anti-CD79␤ (data not shown).
Subsequently, we also tested anti-CD79␣ (F11-172) and found that one injection of 0.5 mg anti-CD79␣ Ab reduced peripheral blood B cells in MRL/lpr mice to a similar extent as did 0.5 mg anti-CD79␤ alone or both Abs used in combination (data not shown), despite that these two Abs partially blocked each other in FACS staining (data not shown) and might compete for binding to B cells.
Thus, MRL/lpr mice were resistant to B cell-specific mAb-mediated B cell depletion with the exception of anti-CD79␣ and anti-CD79␤ Abs.

Peripheral blood B cell depletion and recovery in MRL/lpr after a single dose of anti-CD79 Abs
To determine the best time for re-treatment, we measured peripheral blood B cell counts in MRL/lpr-IgH b mice following one injection of the combined anti-CD79 Abs (0.5 mg each of anti-CD79␣ and anti-CD79␤), first every day (days 1-3) and then at 3-day intervals, for up to 24 days postinjection. Peripheral blood B cells were ϳ35% before treatment, declined starting day 1 postinjection, and reached a minimum of 10% on day 6 ( Fig. 1A), representing a maximum depletion of 75% B cells. B cell levels were then unchanged between days 6 and 9, but started to increase from day 9 and returned to pretreatment levels by 13 days postinjection (Fig. 1A). Thus, weekly injections of these Abs might be required to maintain substantial B cell depletion in MRL/lpr mice.
To determine whether B cells that remained in circulation were fully coated by anti-CD79 Abs in vivo, we collected and stained cells with either FITC-anti-hamster Ab (hIgG), which revealed in vivo-acquired anti-CD79 Abs, or with FITC-anti-CD79␤ (CD79␤), which revealed available free CD79 Ag on the cell surface. Taken together, these analyses would indicate whether resistance of B cells to depletion was due to insufficient surface coating of anti-CD79 Abs.
We found that on day 9 postinjection, the number of circulating B cells (B220 ϩ CD19 ϩ ) were the same as those with surface hamster Ab (B220 ϩ hIgG ϩ ), but very few cells at this time point picked up ex vivo anti-CD79␤ Ab (Fig. 1B). Thus, B cells found in circulation on day 9 postinjection were still fully coated with anti-CD79 Abs, suggesting that their lack of clearance at this point was not due to insufficient surface coating by the injected Abs.
After B cells had returned to preinjection levels (day 20), however, the circulating B cell profile was different from those on day 9. On day 20, few circulating B cells carried in vivo-bound anti-CD79 (B220 ϩ hIgG ϩ ), which was expected with the injected Ab concentration decreasing over time. Consistent with this, most remaining B cells bound ex vivo FITC-anti-CD79␤ Ab. However, their fluorescent intensity of ex vivo bound anti-CD79␤ was lower than cells from PBS-treated mice (Fig. 1, C and D). A similar pattern was seen with peripheral blood B cells collected on day 24 postinjection (Fig. 1D), suggesting that after anti-CD79 treatment, CD79␤ low expression B cells were the first subset of cells to reenter the circulation.

Repeated anti-CD79 treatment depletes B cells from peripheral blood, spleen, lymph node, and bone marrow
To determine whether anti-CD79 Abs can maintain B cell depletion, MRL/lpr-Thy1.1 mice were treated with weekly injections of 0.5 mg each of anti-CD79␣ and anti-CD79␤ Abs for 17 wk (Table  II, Expt. 1). Control groups received 1 mg control hamster IgG or PBS. Peripheral blood B cells (B220 ϩ CD19 ϩ ) remained ϳ75% depleted throughout the 17 wk of treatment ( Fig. 2A). Additionally, all remaining B cells showed in vivo-acquired anti-CD79␣/␤ Abs (B220 ϩ hIgG ϩ ) but no binding to ex vivo anti-CD79␤ (B220 ϩ CD79b ϩ ) (Fig. 2, B and C). Serum anti-CD79␤ Ab levels were ϳ10 g/ml just before each weekly injection (data not shown). Similar persistent B cell depletion and complete saturation of peripheral blood B cells by in vivo anti-CD79 Abs were found in additional studies where anti-CD79␣ and anti-CD79␤ Abs were administered weekly for a total of 9 or 11 wk (Table II, Expt. II and Expt. III; data not shown).
Lymph nodes of mice after 11 weekly anti-CD79 treatments showed Ͼ50% lower weight, total lymphocyte counts, ϳ60% reduction of B220 ϩ CD19 ϩ B cells, B220 ϩ sIgM ϩ mature B cells, and a significant reduction of B220 ϩ CD24 ϩ immature B cells (Fig. 4, A-C). Unlike splenic B cells, the remaining B cells in lymph nodes of anti-CD79-treated mice not only carried surfacebound hamster Ab (B220 ϩ hIgG ϩ ), but also picked up ex vivo anti-CD79␤ (Fig. 4D). Indeed, the numbers of lymph node B220 ϩ hIgG ϩ and B220ϩCD79b ϩ cells were similar (Fig. 4C), indicating that the resistant lymph node B cells were only partially saturated by in vivo anti-CD79 Abs.
Bone marrow of MRL/lpr-IgH b mice after 11 weekly anti-CD79 treatments also showed significant reduction in B220 ϩ CD19 ϩ B cells, B220 ϩ sIgM ϩ mature B cells, and a significant reduction of B220 ϩ CD24 ϩ immature B cells (Fig. 5A). Similar to lymph node B cells and unlike splenic B cells, the remaining B cells in bone marrow were characterized by partial saturation by in vivo anti-CD79 Abs (Fig. 5).
In a separate study, we analyzed susceptibility of follicular and marginal zone B cell subsets to anti-CD79-mediated depletion in MRL/lpr-Thy1.1 mice given six weekly injections of anti-CD79 Abs (Table II, Expt. IV). Similar to the longer treatment, total splenic B220 ϩ CD19 ϩ B cells decreased ϳ60% in absolute numbers (Fig. 6A), suggesting that there was no further depletion with longer treatment. Both follicular (FO) B cell numbers (CD19 ϩ CD23 high CD21 int ) and marginal zone (MZ) B cells (CD19 ϩ CD23 Ϫ CD21 high ) were depleted to a similar extent (FO, 78 Ϯ 10% depletion; MZ, 90 Ϯ 4% depletion) (Fig. 6B). CD19 ϩ IgD ϩ B cells were also significantly depleted (Fig. 6C). With most of the FO and MZ cells removed from spleen by anti-CD79, most (75%) of the remaining splenic CD19 ϩ B cells were CD21 low and CD23 low newly formed (NF) B cells (NF, CD19 ϩ CD23 low CD21 low ) (Fig. 6B, upper panel). However, there was no true expansion of the NF B cell subset, as their absolute numbers in spleens of anti-CD79-treated mice were the same as those in controls (Fig. 6B, lower panel). Thus, NF B cells in MRL/ lpr mice were apparently resistant to depletion by anti-CD79. We found that such resistance correlates with a lower expression level of CD79 molecules. In MRL/lpr controls, staining with anti-CD79␤ Ab along with anti-CD19, CD21, and CD23 revealed that NF B cell subsets showed lower fluorescence intensity of anti-CD79␤ binding than did FO and MZ B cells (Fig. 6D). Consistent with this pattern of CD79␤ expression, NF B cells from anti-CD79-treated mice had less surface-bound hamster Ab (anti-CD79␣ and anti-CD79␤) than did FO and MZ B cells (data not shown).

Changes in T cell subsets in MRL/lpr mice
Since congenitally B cell-deficient MRL/lpr mice showed decreased T cell-mediated inflammation in the kidney and decreased T cell activation, we asked whether anti-CD79 treatment is associated with reduced T cell activation. Thus, we enumerated T cell subsets from the spleen of MRL/lpr-Thy1.1 mice after 6 wk of anti-CD79 treatment (Table II, Expt. IV). We found no significant change in splenic CD4 or CD8 T cell numbers but an unexpected expansion of double-negative T cells in both frequency and absolute number (Fig. 7, A and B), which had been reported to have regulatory functions.
Activated, memory, and naive subsets of CD4 ϩ and CD8 ϩ T cells were analyzed based on their CD62L and CD44 expression (Fig. 7C). Naive (CD62L high CD44 low ) CD4 ϩ T cells in anti-CD79-treated mice rose significantly in relative frequency and expanded 2.5-fold in absolute cell number (Fig. 7, D and E). Most CD4 T cells in spleen of MRL/lpr mice were of memory phenotype (CD62L low CD44 high ) and were unaffected by anti-CD79 treatment, as the absolute numbers of memory CD4 ϩ T cells were unchanged in anti-CD79-treated mice (data not shown). The ap-parent decrease in percentage of memory CD4 ϩ T cells in spleen of anti-CD79-treated mice was a result of the true expansion of naive CD4 ϩ T cell (Fig. 7D). Activated CD4 ϩ T cells (CD62L high CD44 high ) were the second largest CD4 T cell subset in the spleen of MRL/lpr mice and also did not show change by anti-CD79 treatment in either relative frequency (Fig. 7D) or absolute number (data not shown). Splenic CD8 ϩ T cells showed no changes in the frequency of activated, memory, and naive T cell subsets (data not shown).
Thus, anti-CD79 treatment in MRL/lpr mice appeared to alter T cell phenotypes, particularly affecting CD4 ϩ naive subsets and double-negative T cells without changing the frequency or absolute number of activated or memory CD4 ϩ T cells.

Repeated anti-CD79 treatment reduced anti-chromatin IgG, prolonged survival, and reduced tissue inflammation
MRL/lpr mice have a shortened lifespan due to systemic autoimmunity, with manifestations including high serum autoantibody levels, hyperimmunoglobulinemia, and inflammation in various organs (17).
MRL/lpr mice develop high titers of autoantibodies, such as anti-chromatin and anti-dsDNA IgG. We found in three separate studies that the serum levels of anti-chromatin IgG were either suppressed or decreased over time by anti-CD79 treatment, whereas they rose steadily in control-treated mice (Fig. 8). However, serum levels of anti-dsDNA IgG were unaffected by anti-CD79 treatment (data not shown). Total serum IgG was also unaffected, although all mice exhibited hyperimmunoglobulinemia at the initiation of treatment (data not shown). Total serum IgM was also unaffected (data not shown).
Mouse anti-hamster IgG (MAHA) was significantly lower in anti-CD79␣/␤ hamster IgG-treated MRL/lpr mice than in control hamster IgG-treated mice (data not shown). This finding suggests a suppressive effect of anti-CD79 Abs on the production of mouse Ab against foreign proteins.
Anti-CD79 treatment also improved survival rate and reduced the incidence of skin lesion development. The survival rates at week 17 for anti-CD79-treated mice (Table II, Expt. I), control hamster IgG, and PBS groups were 83%, 20%, and 25%, respectively ( p Ͻ 0.05 by 2 test starting at week 14) (Fig. 9A). All mice were free of skin lesions at the time of initiation of treatment. By week 11, 100% of the control mice had developed skin lesions, but only 38% of the anti-CD79-treated mice had skin lesions (Fig. 9B) ( p Ͻ 0.05 by Fisher's exact test, for week 11 and week 14). Since mortality was higher in control groups and may affect the evaluation of the incidence of skin lesions, we compared skin lesion development in survivors only and found significantly lower occurrence of skin lesions in the anti-CD79-treated group after 8 wk of treatment (Fig. 9C). However, mice that developed skin lesions on anti-CD79 treatment did not become lesion-free with more treatment. Similar protective effects of anti-CD79 were observed in MRL/lpr-IgH b mice given 11 weekly anti-CD79 Abs (Table II, Expt. II) (data not shown). Improvement of survival might be due to not only to a systemic reduction of anti-chromatin autoantibodies and reduction of B cells, but also to the reduction in kidney injuries. The effect of anti-CD79 treatment on the kidney was evaluated in mice that underwent short-term (6-wk) treatment (Table II, Expt. IV), during which time there was no significant mortality in the control groups. Kidneys of anti-CD79-treated mice had strikingly smaller (Fig. 10,  A and B) and less focal inflammatory infiltrates compared with the PBS control group (Fig. 10E). Furthermore, anti-CD79 treatment significantly reduced interstitial nephritis, as fewer mice on anti-CD79 than in the control group scored Ͼ2 (Fig. 10F, p ϭ 0.0152, Fisher's exact test). Glomerulonephritis was less severe in two of five (40%) mice in the anti-CD79 group, while all mice in PBS groups had scores Ͼ2 (Fig. 10F). However, the difference was not significant.
Salivary glands of MRL/lpr mice are known to develop extensive focal inflammatory infiltrates around blood vessels and ducts. MRL/lpr females treated for 6 wk with anti-CD79 Abs showed substantially smaller lesions and significantly lower density of FIGURE 7. Splenic T cell changes in anti-CD79-treated MRL/lpr. Data were from same experiment as described in Fig. 6. A, Relative frequency of T cell subsets. B, Absolute number of T cell subsets. C, Gating of activated, memory, and naive subsets based on CD62L and CD44 expression. Shown are representative data from a pair of mice. D, Relative frequency of CD4 ϩ T cell subset. E, Absolute number of naive CD4 ϩ T cells. Shown are means and SEM. Significant differences ‫,ء(‬ p Ͻ 0.05; ‫,ءء‬ p Ͻ 0.01) were between AB and control-treated group.

Discussion
The idea of using anti-CD79 Abs as immunosuppressive agents was raised as early as 1993 by Nakamura et al. (18), mainly because they discovered that crosslinking human CD79␤ in vitro activates pre-B cells but blocks terminal differentiation of all plasma cell isotypes (18). We demonstrate in a series of in vivo studies that anti-CD79 can have a different immunomoluatory effect; that is, by mediating B cell depletion. We showed in MRL/lpr mice that a combination of CD79␣ and CD79␤ can effectively deplete B cells from peripheral blood, spleen, bone marrow, and lymph nodes; decrease tissue inflammation in skin, kidney, and salivary glands; decrease anti-chromatin IgG; and improve survival.
MRL/lpr, a spontaneous model of lupus, and the experimentally induced model of lupus are known to be very resistant to B cell depletion by anti-B cell Abs (12) (R. A. Eisenberg, unpublished observations). This feature is likely due to the autoimmune state of these mice and possibly to B cell activation. It is therefore very encouraging to have found that anti-CD79␣/␤ not only can deplete MRL/lpr B cells but also worked at doses that were more than 10 times lower than those that have been reported to be required of anti-CD20 in MRL/lpr mice (12). The clones of anti-CD79␣ and anti-CD79␤ Abs used in our studies partially compete with each other for binding. Still, both Abs given at the same time depleted B cells as effectively as did anti-CD79␤ alone and better than anti-CD79␣ alone, suggesting that they were used at levels exceeding the minimal doses. The in vivo effects of each Ab given alone at 0.5 mg or at either higher or lower doses remain to be determined.
We used both Abs to treat MRL/lpr mice, as we postulated that both are required to maximize depletion and to interfere with B cell signaling. Unlike CD20, CD79␣ and CD79␤ have important signaling functions for B cells as they couple to MHC II and BCR (19) and initiate downstream tyrosine phosphorylation and intracellular release of Ca 2ϩ . It is appreciated that CD79␣ (Ig␣) and CD79␤ (Ig␤) participate in both Ag-dependent and Ag-independent signaling (20). Thus, unlike anti-CD20, anti-CD79 Abs might have dual functions in vivo. Our study raises the question as to whether B cell-targeting Abs can exert immunomodulatory effects through mechanisms in addition to B cell removal. We found that most of the remaining B cells in peripheral blood and spleen were saturated in vivo with anti-CD79 Abs, whereas B cells remaining in bone marrow and lymph nodes were partially coated by anti-CD79␣ and anti-CD79␤. These surface-bound anti-CD79 Abs may modulate B cell signaling. While more specific studies are required to address the issue of whether anti-CD79 Abs have a direct immunosuppressive effect, our study suggests that they may. Indeed, mice treated with anti-CD79␣/␤ developed almost no mouse anti-hamster Abs, whereas mice treated with control hamster Abs developed high levels of the potentially neutralizing Abs. This apparent immunosuppressive effect of the anti-CD79 hamster Abs may be due to deletion of B cells, but given the incomplete degree of this effect, we suggest that additional mechanisms may be operative.
Our findings show that some subsets of B cells are effectively depleted by anti-CD79 Abs. These were follicular (78 Ϯ 10%), marginal zone (90 Ϯ 4%), and IgD ϩ B cells (80 Ϯ 10%) of spleen,    that other B cell types would be resistant to anti-CD79-mediated in vivo effects. Cells unaffected would be those expressing a lower level of CD79␣/␤ (as in newly formed B cells) or IgG1-expressing B cells, as suggested by a recent study showing that survival of IgG1-expressing B cells is less dependent on CD79␣ signaling than on IgM/IgD-expressing B cells (21). In contrast, IgG2a-expressing cells would be susceptible to anti-CD79-mediated in vivo effect, as IgG2a receptor signaling requires CD79␣ and CD79␤ (22). We have yet to test surface IgG1 and IgG2 expression on the remaining B cells in anti-CD79-treated MRL/lpr mice, or the serum levels of IgG isotypes.
Serum anti-chromatin IgG appeared to be more sensitive than anti-DNA or total serum IgG or IgM to anti-CD79 treatment, suggesting either that B cell clones producing anti-chromatin IgG have a relatively rapid turnover rate compared with other B cell clones or that they still express CD79␣ and CD79␤. In human SLE patients, anti-nucleosome and anti-dsDNA levels were also lower following anti-CD20 treatment (23). Therefore, some autoimmune B cell clones in both human or mouse models may have a shorter turnover rate than other B cell clones producing different Abs. Although anti-CD79 Abs and other B cell-targeting Abs cannot directly target some Ab-secreting cells, such as plasma cells, that have lost surface expression of CD20 or CD79␣/CD79␤, these Abs can nevertheless affect fast turnover Ab-forming cells by removing their precursors.
Anti-CD79␣/␤ can maintain peripheral B cells at maximum depletion of 75% for 9 days. This duration of B cell depletion is shorter than that of anti-mCD20 Abs in B6 mice, which can keep peripheral blood B cells depleted for as long as 20 days . The difference in the time course of B cell recovery may be due to the autoimmune nature of MRL/lpr mice, the difference in the Ags, and/or the difference in Abs targeting them. As Fc␥Rmediated removal by macrophages is an important mechanism for Ab-mediated B cell depletion (24), the nature of the Abs (hamster anti-mouse Abs) used in our study may contribute to the shorter duration of B cell depletion. The hamster Fc portion of anti-CD79␣ (clone F11-172) and anti-CD79␤ (clone HM79-16) may not be a perfect match for FcR of mice and thus can potentially lower the efficiency of FcR-mediated clearance of B cells. Thus, chimeric Abs with the variable regions of these two clones fused to mouse Fc regions might be useful in determining the full potential of B cell depletion and the impact on clinical disease in MRL/lpr mice by anti-CD79 Abs. Since the IgG2a isotype of B cell-targeting Ab mediates the greatest B cell depletion, the Fc region would preferably be that of mouse IgG2a. Thus, the efficacy of anti-CD79␣ and anti-CD79␤ Abs may be further improved.
T cell phenotypes were altered in several ways by anti-CD79 treatment. In spleen of anti-CD79-treated MRL/lpr mice, naive CD4 ϩ T cells were expanded in absolute numbers, whereas the numbers of activated or memory CD4 ϩ T cells were not affected by treatment. Although there was no reduction in the number of activated CD4 ϩ T cells in mice that received six weekly injections of Ab, tissue inflammation in both the salivary glands and kidneys was significantly lower than in control groups. Thus, reduction in tissue inflammation did not appear to require reduction of either the frequency or absolute number of activated T cells. Another unexpected finding of this study was the significant increase of double-negative T cells in spleens of MRL/lpr mice treated with six weekly injections of anti-CD79 Abs. The effect was specific, as neither CD4 ϩ nor CD8 ϩ T cells changed in absolute numbers. Since B cell depletion in B6 mice with anti-CD20 does not appear to affect T cell profiles, our findings that double-negative T cells increase in MRL/lpr mice treated with anti-CD79 Abs may be specific to the MRL/lpr strain and/or the effect of anti-CD79␣/␤. Double-negative T cells from MRL/lpr mice and other mouse strains have been found both in vivo and in vitro to exhibit regulatory T cell function (25). In particular, double-negative regulatory T cells can kill CD8 ϩ T cells and CD4 ϩ T cells in an Agspecific manner that requires functional FAS and FAS receptor (26). Based on this, however, we do not think the expanded double-negative T cell subset in anti-CD79-treated MRL/lpr mice plays a role in down-modulation of the inflammatory state, as these mice have a mutated FAS receptor. Taken together, the lack of evidence of altered activation states of T cells in anti-CD79 treated MRL/lpr mice suggests that the beneficial effects with anti-CD79 may be entirely due to suppression of B cell function and/or removal of autoreactive B cells.
Because of the key role of CD79 molecules in signaling, anti-CD79 immunotherapy may have immunosuppressive effects on B cells that do not require their depletion and may have the potential for multiple immuno modulatory effects. Its effects on T cells, on autoantibody levels, and on host-generated neutralizing Abs are unlike those of anti-CD20, also suggesting that it may be worthy of further investigation as a more potent, yet still quite selective immunosuppressant. Abs against human CD79, as well as potent cytotoxic agents coupled to anti-human CD79 Abs, are effective anti-human B cell lymphoma cells both in vitro and in mouse chimeras (27). This approach may lead to clinical trials in individuals with lupus and other autoimmune disorders.

Disclosures
The authors have no financial conflicts of interest.