Anti-CD20 mAb-Induced B Cell Apoptosis Generates T Cell Regulation of Experimental Myeloperoxidase ANCA-Associated Vasculitis

Significance Statement

Myeloperoxidase ANCA-associated vasculitis (MPO-AAV) is an important cause of renal failure. Although rituximab has been shown to effectively attenuate MPO-AAV, its underlying mechanism of action beyond depletion of CD20⁺ B cells producing ANCA is unknown. Administration of mouse anti-CD20 mAb to a murine anti-MPO GN mouse model reduced not only serum MPO-ANCA but also, T cell responses. Interestingly, anti-CD20 mAb treatment increased the frequency and functional potency of Tregs. Administration of anti-CD20 mAb rendered B cells apoptotic and resulted in the attenuation of anti-MPO autoimmunity and GN. This highlights a novel pathway by which anti-CD20 mAb therapy may attenuate T cell–mediated autoimmunity.

Abstract

Background

Myeloperoxidase ANCA-associated vasculitis is a major cause of ESKD. Efficacy of anti-CD20 mAb treatment was tested in a mouse model of the disease.

Methods

MPO immunization induced anti-MPO autoimmunity, and a subnephritogenic dose of sheep anti-mouse GBM globulin triggered GN.

Results

Anti-CD20 mAb treatment increased the numbers and immunomodulatory capacity of MPO-specific T regulatory cells (Tregs) and attenuated T cell–mediated and humoral anti-MPO autoimmunity and GN. Disabling of Tregs negated the therapeutic benefit of anti-CD20 treatment. The mechanism of enhancement of Treg activity could be attributed to anti-CD20 mAb effects on inducing B cell apoptosis. Administering anti-CD20 mAb-induced apoptotic splenocytes to mice developing anti-MPO GN was as effective as anti-CD20 mAb treatment in inducing Tregs and attenuating both anti-MPO autoimmunity and GN. A nonredundant role for splenic macrophages in mediating the anti-CD20 mAb-induced immunomodulation was demonstrated by showing that administration of anti-CD20 mAb ex vivo–induced apoptotic splenocytes to unmanipulated mice attenuated autoimmunity and GN, whereas deletion of splenic marginal zone macrophages prevented anti-CD20 mAb-induced immunomodulation and treatment efficacy. Six days after administering anti-CD20 mAb to mice with murine anti-MPO GN, cell-mediated anti-MPO responses and GN were attenuated, and Tregs were enhanced, but ANCA levels were unchanged, suggesting humoral autoimmunity was redundant at this time point.

Conclusions

Collectively, these data suggest that, as well as reducing humoral autoimmunity, anti-CD20 mAb more rapidly induces protective anti-MPO Treg-mediated immunomodulation by splenic processing of anti-CD20–induced apoptotic B cells.

Myeloperoxidase ANCA-associated vasculitis (MPO-AAV) is caused by autoreactivity to the neutrophil enzyme, myeloperoxidase (MPO). Rituximab, an anti-CD20 mAb, is an effective therapy for ANCA-associated vasculitis (AAV) and is recommended as a first-line induction treatment.^1–3 The Rituximab versus Cyclophosphamide for AAV (RAVE) trial showed no inferiority of rituximab. Rituximab depletes CD20 B cells, and the current paradigm is that attenuation of humoral autoimmunity to the vasculitis inducing autoantigens, MPO, and proteinase-3 is rituximab’s main mechanism of action. The primary end point in the RAVE trial that compared rituximab with cyclophosphamide for induction therapy in AAV was disease remission induction after 6 months of therapy. Although deletion of B cells was effective and rapid, there was no correlation between anti-CD20 mAb-induced reduction in ANCA titers and attainment of the primary end point. This observation led the investigators to comment that “the efficacy of both rituximab and cyclophosphamide was likely to be due to mechanisms beyond autoantibody suppression.”²

A murine model of anti-MPO GN involving both humoral (ANCA) and cellular (T cell) anti-MPO autoimmunity^4–15 was used to study the effects of anti-CD20 mAb treatment. Autoimmunity to MPO is induced by immunizing mice with MPO in Freund's adjuvant, and glomerular injury is initiated by planting MPO in glomeruli by administration of a subnephritogenic sheep anti-mouse glomerular basement membrane (GBM) globulin to recruit neutrophils to glomeruli. Glomerular neutrophils degranulate and deposit MPO, resulting in the recruitment of anti-MPO T cells and subsequent delayed-type hypersensitivity (DTH) effectors. Subnephritogenic anti-GBM globulin itself does not induce GN without an active anti-MPO immune response (i.e., OVA-immunized mice administered with anti-GBM globulin do not develop GN).¹⁰ This anti-MPO GN model is highly relevant as a manipulable surrogate for understanding human MPO-AAV pathogenesis and treatment, sharing common features with human disease. Furthermore, the availability of a murine anti-CD20 mAb, which has similar target binding characteristics and functional effects to rituximab,¹⁶ enhances the likelihood that derived observations would be relevant to human disease. As observed in the RAVE trial, we found rapid depletion of B cells associated with significant attenuation of GN before there was any effect on circulating MPO-ANCA IgG. We found that murine anti-CD20 mAb reduced cellular autoimmunity by enhancing T regulatory cell (Treg) expansion and functional (immunomodulatory) capacity. Subsequent studies provided evidence that the murine anti-CD20 mAb induced B cell apoptosis, which in the course of homeostatic clearance involving splenic macrophages, resulted in significant induction of Tregs. These data suggest that modifications to anti-CD20 mAb therapies that do not destroy host humoral immunity while inducing therapeutic Tregs could be developed.

Methods

Mice

C57BL/6 (wild-type) male mice were purchased from Monash University Animal Services. Foxp3^GFP mice were obtained from Alexander Rudensky (University of Washington, Seattle, WA). Wild-type and Foxp3^GFP mice were housed and bred at Monash Medical Centre, and experiments were approved by the Monash Medical Centre Animal Ethics Committee.

Experimental Design

Anti-MPO GN Model

To induce anti-MPO GN, 6- to 8-week-old male mice were immunized on day 0 with 20 µg recombinant mouse myeloperoxidase (rMPO^6,15; generated using the baculovirus system as previously described)¹⁷ in Freund's Complete Adjuvant (FCA; Sigma-Aldrich), subcutaneously. On day 7, mice received a booster injection subcutaneously of 10 µg rMPO in Freund's Incomplete Adjuvant (Sigma-Aldrich). GN was induced by administering an intravenous injection of 1.5 mg sheep anti-mouse GBM globulin at two different time points: during early developing or established anti-MPO autoimmunity (days 16 and 17 or days 28 and 29, respectively). Mice were humanely killed 4 days after administration of anti-GBM globulin. A group of OVA-immunized mice was used to demonstrate that the subnephritogenic dose of sheep anti-mouse GBM globulin administered on days 28 and 29 (established anti-MPO GN model) does not induce any autologous-phase glomerular injury (day 32) (Supplemental Figure 1).

In Vivo Depletion of CD20⁺ B Cells

Mouse anti-mouse CD20 mAb (Biogen Idec) was administered to mice with anti-MPO autoimmunity. For early developing anti-MPO GN (20-day model), mice were administered intraperitoneally with 250 µg mouse anti-mouse CD20 mAb or control mouse IgG2a on day 14. In mice with established anti-MPO GN (32-day model), anti-CD20 mAb was administered on days 14 and 23. To demonstrate that anti-CD20 mAb does not modulate sheep anti-mouse GBM globulin in recruiting glomerular neutrophils, anti-CD20 mAb or control mouse IgG2a was administered to naïve wild-type mice (day 0). Seven days later, 3 mg of anti-GBM globulin was administered, and mice were killed 2 hours later (a time point of peak glomerular neutrophil influx) (Supplemental Figure 2).

In Vivo Depletion of CD20⁺ B Cells and CD25⁺ Tregs

In mice with established anti-MPO GN (32-day model), there was concomitant administration of 250 µg anti-CD20 mAb and 1.5 mg anti-CD25 mAb on days 14 and 25. MPO-immunized treatment controls received rat IgG1 and/or mouse IgG2a.

In Vivo Depletion of Marginal Zone Macrophages

In mice with early developing anti-MPO GN (20-day model), 0.1 ml/10 µg clodronate encapsulated liposomes (ClodronateLiposomes, Amsterdam, The Netherlands) were administered intravenously on day 9. Controls received PBS liposomes (ClodronateLiposomes).

To confirm the depletion of splenic marginal zone macrophages, primary antibody, anti-Metallophilic Macrophages (abcam), was used on 6-µm-thick OCT frozen spleens.

To detect splenic dendritic cells, spleens were stained with CD11c-FITC (Biolegend) and nuclei stained using DRAQ5 Fluorescent Probe (ThermoFisher). Spleens were then mounted with Dako antifade mounting media (Sigma-Aldrich). Fluorescent images were acquired using a Nikon C1 confocal laser scanning head attached to a Nikon Ti-E inverted microscope (Coherent Scientific). Nikon lasers 488 and 647 were used.

Anti-CD20 mAb Induced Apoptotic B Cells In Vitro to Treat MPO Autoimmunity

To obtain apoptotic B cells, spleens were harvested from naïve wild-type mice. Splenocytes were stained for B220⁺ (AF488, Clone RA3–6B2; Biolegend) and sorted for B220⁺ B cells on the BD Influx 1 Cell Sorter (BD Biosciences). Isolated B220⁺ B cells were incubated with 250 µg anti-CD20 mAb for 24 hours. B cells were stained with Annexin V (APC; Biolegend) and propidium iodine to assess apoptosis and cell death. B cells were thoroughly washed for unbound anti-CD20 mAb; 5×10⁶ apoptotic B cells were administered intravenously 1 day before mice were immunized with rMPO in FCA. Control mice received PBS. Mice were humanely killed 10 days later, and anti-MPO immune responses were assessed.

Anti-CD20 mAb Induced Apoptotic Splenocytes In Vitro to Treat Anti-MPO Autoimmunity and GN

To obtain splenocytes, spleens were harvested from naïve wild-type mice. Splenocytes were incubated with 5 µg/ml mouse anti-mouse CD20 mAb for 30 minutes (37°C, 5% CO₂). Splenocytes were washed three times to remove unbound anti-CD20 mAb; 5×10⁷ apoptotic splenocytes were administered to mice with anti-MPO autoimmunity on day 10.

Assessment of Renal Disease and Glomerular Immune Cell Infiltration

Albuminuria was assessed by housing mice in individual metabolic cages to collect urine over 24 hours before the end of the experiment. Urinary albumin concentration was measured by ELISA using an albuminuria kit (Bethyl Laboratories, Montgomery, TX).

Histologic assessment of renal damage was performed on 3-µm-thick formalin-fixed, paraffin-embedded, periodic acid–Schiff-stained kidney sections. A minimum of 30 consecutive glomeruli per mouse were examined, and results are expressed as percentage of segmental glomerular necrosis per glomerular cross-section.

Glomerular immune cells (CD4⁺ T cells, macrophages, and neutrophils) were assessed by an immunoperoxidase-staining technique on 6-µm-thick periodate lysine paraformaldehyde–fixed, OCT-frozen kidney sections. The primary antibodies used were GK1.5 for CD4⁺ for T cells (anti-mouse CD4⁺; American Type Culture Collection, Manassas, VA), FA/11 for macrophages (anti-mouse CD68; from Gordon L. Koch, Cambridge, United Kingdom), and RB6–8C5 for neutrophils (anti-GR1; DNAX, Palo Alto, CA). A minimum of 30 glomeruli were assessed, and results are expressed as cells per glomerular cross-section.

Systemic Immune Responses to MPO

Humoral Anti-MPO Responses

ELISA was used to detect circulating serum anti-MPO IgG titers using a 96-well polystyerene microplate (Invitrogen Technologies). Plates were coated with 1 µg/ml rMPO, and serum was either added directly or serially diluted. Detection was by horseradish peroxidase–conjugated sheep anti-mouse IgG (1:2000; Amersham Biosciences, Rydalmere, Australia).

Systemic Anti-MPO Responses

To assess MPO-specific dermal footpad delayed-type hypersensitive (DTH) responses, mice were challenged by intradermal injection of 10 µg/30 µl of rMPO in the right footpad. The contralateral footpad received intradermal injection of 30 µl saline. DTH was assessed 24 hours later by measuring the difference between footpad thickness (Δmillimeters) using a micrometer. MPO-specific cell proliferation was measured by culturing splenocytes or lymph node (LN) cells at 5×10⁵ cells per well in 96-well flat-bottom plates (Sarstedt, Newton, NC), restimulating with 10 µg/ml rMPO, and incubating for 72 hours. During the last 16 hours of culture, 0.5 µCi of [³H]-thymidine (Perkin Elmer, Waltham, MA) was added. [³H]-thymidine incorporation was measured as previously described. For assessment of cytokine production, lymphocytes or splenocytes were cultured for 72 hours (37°C, 5% CO₂) at 4×10⁶ cells per milliliter per well with 10 µg/ml MPO in supplemented RPMI (10% FCS, 2 mM l-glutamine, 50 µM 2-ME, 100 U/ml penicillin, and 0.1 mg/ml streptomycin; Sigma-Aldrich). IFN-γ, IL-17A, IL-2, and IL-10 secretion was detected in the cultured supernatants by cytometric bead array (BD Biosciences) or by ELISPOT (mouse IFN-γ ELISPOT kit and mouse IL-17A ELISPOT kit; BD Biosciences) with LN cells or splenocytes seeded at 5×10⁵ cells per well restimulated with 10 µg/ml of rMPO for 18 hours. IFN-γ– and IL-17A–producing cells were enumerated with an automated ELISPOT reader system.

Flow Cytometry and Intracellular Staining

The efficacy of anti-CD20 mAb in depleting circulating B cells in blood was assessed using B220 (PE-Cy7, clone RA3–6B2; Biolegend).

To assess T and B cell responses in mice with anti-MPO GN treated with anti-CD20 mAb, splenocytes and LN cells were seeded at 5×10⁵ cells per well (96-well clear, round-bottom microplates; Falcon) with rMPO in 10% FCS in RPMI media. CD4⁺ T cells and CD4⁺ Tregs were stained with CD4, Foxp3, IL-10, CD1, CD5, and CD19. To determine cellular proliferation, cells were labeled with cell trace violet (CTV) dye (CellTrace Violet Cell proliferation Kit; ThermoFisher) before culture. Cells were cultured (37°C, 5% CO₂) for 72 hours in 96-well plates. Results are expressed as percentage of CD4⁺Foxp3⁺CTV^lo for MPO-specific Treg proliferation and CD4⁺CTV^lo for MPO-specific T cell proliferation.

Samples were analyzed on the Navios flow cytometer, and data were analyzed using FlowJo software.

Assessment of Anti-CD20 mAb to Inhibit In Vitro MPO CD4⁺ T Effector Cell Recall Response

To obtain MPO-specific CD4⁺ T effector cells (Teffs), Foxp3^GFP mice (n=10) were immunized with MPO and FCA subcutaneously. Ten days later, draining LNs and spleens were harvested, and CD4⁺ T cells were isolated by magnetic separation using mouse CD4 (L3T4) microbeads (Miltenyi Biotec). CD4⁺ Teffs were sorted on the basis of GFP^−ve expression on the FACS Aria Fusion Cell Sorter (BD Biosciences).

Anti-CD20 mAb- or IgG2a-treated MPO-immunized Foxp3^GFP reporter mice were isolated on day 32. CD4⁺ T cells were isolated (as above) and Tregs were sorted on the basis of GFP^+ve expression.

Statistical Methods

Results are expressed as the mean ± SEM. Unpaired t test was used when comparing two groups. Paired t test for comparison in ex vivo coculture experiments with nonconcordant values excluded. A one-way ANOVA test for experiments was used for comparisons between more than two groups, followed by a Tukey post hoc test for parametric data. All data were analyzed with Graph Pad Version 7 (GraphPad Prism; GraphPad Software Inc., San Diego, CA). Differences were considered to be statistically significant if P=0.05.

Results

Anti-CD20 mAb Treatment Induces CD20⁺ B Cell Depletion, Attenuating Established Anti-MPO Autoimmunity and Kidney Injury

Anti-MPO autoimmunity was induced in mice by MPO immunization and allowed to mature over 32 days, resulting in the generation of both humoral MPO-ANCA IgG and anti-MPO T cell responses. Administration of two doses of anti-CD20 mAb compared with control (Ctrl) murine IgG2a significantly reduced circulating B220⁺ B cells as well as serum MPO-ANCA IgG (Figure 1, A and B). Significant reduction in anti-MPO cellular autoimmunity was also demonstrated by reduced MPO-specific dermal DTH recall responses (Figure 1C), reduced anti-MPO T cell recall responses to MPO measured by [³H]-thymidine uptake, and the number of IFN-γ–producing LN cells isolated from draining LNs compared with controls (Figure 1, D and E). No difference was observed in the number of IL-17A–producing lymphocytes from anti-CD20 mAb-treated mice and controls (Figure 1F). Anti-CD20 mAb-treated mice showed expansion in the frequency of proliferating MPO-specific CD4⁺Foxp3⁺ Tregs (Figure 1G). Glomerular injury was significantly attenuated by anti-CD20 mAb administration. Histologic kidney injury, glomerular segmental necrosis, and glomerular leukocyte infiltration (CD4⁺ T cells) were significantly reduced when anti-CD20 mAb was administered to mice with established anti-MPO autoimmunity (Figure 1, H and I). However, although there was a reduction observed in albuminuria in anti-CD20 mAb-treated mice, this did not reach statistical significance (Figure 1J). The representative photomicrographs of kidney sections illustrate the severity of damaged glomerular lesion and the extent of glomerular leukocyte infiltration (CD4⁺ T cells, macrophages, and neutrophils) (Figure 1K, Supplemental Figure 3). Collectively, these results suggest a role for anti-CD20 mAb-induced enhanced Treg immunosuppression in the attenuation of autoimmunity and GN.

Figure 1.

Anti-CD20 mAb treatment attenuates anti-MPO autoimmunity and GN. (A) Treatment of mice developing anti-MPO GN with two doses of anti-CD20 mAb resulted in rapid and sustained depletion of circulation B220⁺ cells. This was associated with significant attenuation of anti-MPO autoimmunity as shown by (B) reduction in serum MPO-ANCA IgG, (C) MPO-specific dermal DTH responses, and (D) MPO-specific proliferation of [³H]-thymidine–treated LN cells as well as (E) production of IFN-γ, but (F) no difference in IL-17A–producing cells was observed. (G) There was a significant increase in the numbers of proliferating MPO-specific CD4⁺Foxp3⁺ Tregs. The reduction in systemic anti-MPO autoimmunity and enhancement of Tregs was associated with significant attenuation of GN demonstrated by (H) reduction of glomerular segmental necrosis and (I) glomerular leukocyte influx. (J) There was reduction in albuminuria, which did not reach statistical significance. (K) Photomicrographs showing the effects of anti-CD20 mAb treatment on glomerular segmental necrosis and CD4⁺ T cell influx. Error bars represent mean ± SEM with statistical analysis by unpaired t test. cpm, counts per minute; Ctrl, control; FIA, Freund incomplete adjuvant; gcs, glomerular cross-section; Mo, macrophages; OD450nm, Optical Density 450 nm; PAS, periodic acid–Schiff; PMN, polymorphonuclear leukocytes; WT, wild type. *P=0.05; **P=0.01; ***P<0.001; ****P<0.001.

Anti-CD20 mAb-Induced Anti-MPO GN Remission Precedes the Attenuation of Circulating MPO-ANCA IgG

To examine the time course of anti-CD20 mAb therapy in attenuating anti-MPO GN, we studied a group of animals treated with a single dose of anti-CD20 mAb 6 days prior to the termination of the experiment (day 20). At this time point, circulating MPO-ANCA IgG was not significantly reduced by anti-CD20 mAb treatment (Figure 2A), whereas T cell anti-MPO autoimmunity was significantly reduced; anti-MPO DTH, as was MPO recall proliferation of LN cells draining MPO immunization sites (Figure 2, B and C). Concurrently, a significant expansion of MPO-specific CD4⁺Foxp3⁺ Tregs was observed (Figure 2D). Additionally, glomerular injury was attenuated at this time point. The frequency of glomerular segmental necrosis was reduced in anti-CD20 mAb-treated mice, as were the numbers of glomerular effector leukocytes (CD4⁺ T cells and macrophages) (Figure 2, E and F). No difference in albuminuria was observed between groups (Figure 2G). The therapeutic benefit of anti-CD20 mAb occurred before any fall in MPO-ANCA IgG, suggesting that, at this time point, attenuation of disease is mediated by anti-CD20 mAb effects on anti-MPO cell-mediated autoimmunity and not humoral autoimmunity.

Figure 2.

Anti-CD20 mAb treatment attenuates cellular anti-MPO autoimmunity and GN before ANCA titers fall. To examine the time course of the effects of anti-CD20 mAb treatment on autoimmunity and glomerular injury, the effects of using a single dose in a shortened 20-day model were studied. (A) Although single dose of anti-CD20 mAb treatment did not affect humoral (MPO-ANCA) anti-MPO autoimmunity, (B and C) treatment significantly attenuated cellular anti-MPO autoimmunity (MPO-specific DTH response and MPO recall-specific proliferation) and (D) induced expansion of MPO-specific CD4⁺Foxp3⁺CTV⁻ Tregs. As seen in mice given two doses of anti-CD20 mAb over a longer period, the effect of single dose of anti-CD20 mAb treatment induced similar levels of attenuation of (E) glomerular segmental necrosis, (F) leukocyte influx, and (G) albuminuria. These results suggest that the effect of anti-CD20 mAb treatment on cell-mediated immunity is rapid, whereas its effects on MPO-ANCA IgG take longer. Error bars represent mean ± SEM with statistical analysis by unpaired t test. cpm, counts per minute; Ctrl, control; FIA, Freund incomplete adjuvant; gcs, glomerular cross-section; Mo, macrophages; OD450nm, optical density 450 nm; PMN, polymorphonuclear leukocytes. *P=0.05; **P=0.01.

Treg Immunomodulatory Capacity Is Enhanced following Anti-CD20 mAb Treatment

Administration of anti-CD20 mAb to mice with established anti-MPO autoimmunity increased the frequency of Tregs. Next, we sought to determine the functional capacity of Tregs from Foxp3^GFP reporter mice with anti-MPO autoimmunity treated with anti-CD20 mAb. We performed an MPO-specific Treg suppression assay ex vivo. Tregs were isolated (day 32) from Foxp3^GFP mice with established anti-MPO autoimmunity treated with either anti-CD20 mAb or mouse IgG2a (administered on days 14 and 23). Foxp3^GFP cells were isolated and cocultured with anti-MPO CD4⁺Foxp3⁻ Teffs at varying ratios and restimulated with MPO in vitro. Foxp3^GFP Tregs from anti-CD20 mAb-treated mice suppressed MPO-specific Teff proliferation more effectively at lower Treg-Teff ratios (one Treg:32 Teff) than Tregs from control-treated mice. At higher Treg-Teff ratios (1:16 and 1:8), a saturation effect of Tregs was observed (Figure 3A). Cytokine analysis of the coculture supernatant from one Teff:32 Tregs showed there was a significant reduction of IL-2 in supernatants of Tregs from mice treated with anti-CD20 mAb compared with control-treated mice (Figure 3B). This is consistent with enhanced utilization of IL-2 by Tregs from anti-CD20 mAb-treated mice. Although not significant, there was a trend to reduction in the concentration of proinflammatory cytokines (IFN-γ and IL-17A) produced by Teff in the presence of anti-CD20 mAb-treated Tregs (Figure 3, C and D). Taken together, these data demonstrate that anti-CD20 mAb treatment results in an increased frequency and functional potency of MPO-specific Tregs.

Figure 3.

Anti-CD20 mAb enhances Treg immunomodulatory capacity. To further assess the enhancement of immunomodulation induced by anti-CD20 mAb treatment, the functional capacity of anti-CD 20 mAb-induced Tregs was assessed ex vivo. Tregs were isolated from Foxp3^GFP reporter mice with established anti-MPO autoimmunity treated with anti-CD20 mAb or control mouse IgG2a and cocultured with CD4⁺Foxp3⁻ Teffs at various Treg-Teff ratios. Foxp3^GFP Tregs from anti-CD20 mAb-treated animals were more effective in inhibiting CD4⁺Foxp3⁻ Teff proliferation in response to MPO recall responses than control IgG2a-treated Tregs (at a ratio of one Treg:32 Teffs). (A) At higher concentrations of Tregs (1:16 and 1:8), a saturation effect was observed between groups. (B) Analysis of IL-2 concentration showed there was greater consumption of IL-2 by anti-CD20 mAb-treated Tregs compared with control-treated Tregs. There was reduction of inflammatory cytokines (C) IL-17A and (D) IFN-γ by anti-CD 20 mAb-treated Tregs, but this was not statistically significant. Error bars represent mean ± SEM with statistical analysis by paired t test. *P=0.05. cpm, counts per minute.

Tregs Are Required for the Therapeutic Benefit of Anti-CD20 mAb Therapy

To examine the therapeutic effects of anti-CD20 mAbs dependancy on Tregs, we concurrently administered both anti-CD20 mAb as well as anti-CD25 mAb (known to inhibit/deplete Tregs), anti-CD20 mAb alone, or mice administered mouse IgG2a and rat IgG1. As observed previously, anti-CD20 mAb alone significantly increased the frequency of proliferating anti-MPO–specific CD4⁺Foxp3⁺ Tregs. This increased frequency of Tregs induced by anti-CD20 mAb was prevented by concomitant administration of anti-CD25 mAb and was similar to the numbers of Tregs observed in anti-MPO GN control-treated mice (Figure 4A). Concominant administration of anti-CD20 and anti-CD25 mAbs abrogated the capacity for anti-CD20 mAb to attenuate circulating MPO-ANCA IgG (Figure 4B) and cellular autoimmunity shown by enhancement of anti-MPO–specific [³H]-thymidine proliferation, the frequency of IL-17A– and IFN-γ–producing LN cells and dermal DTH (Figure 4, C–F). Anti-CD25 mAb coadministration with anti-CD20 mAb blocked the protection from glomerular injury seen in mice receiving anti-CD20 mAb alone as shown by albuminuria, glomerular segmental necrosis, and leukocyte glomerular infiltration (Figure 4, G–K). Taken together, these results are consistent with the therapeutic effects of anti-CD20 mAb being dependent on the presence of Tregs.

Figure 4.

Anti-CD 20 mAb treatment is Treg dependent. To assess the role of Tregs in the attenuation of anti-MPO autoimmunity and GN induced by anti-CD20 mAb treatment, mice with established anti-MPO autoimmunity were administered anti-CD20 mAb treatment, anti-CD20 mAb treatment and anti-CD25 mAb, or control (Ctrl) IgG2a and rat IgG1. The immunomodulation of anti-MPO autoimmunity induced by anti-CD20 mAb treatment did not occur in mice administered this treatment together with anti-CD25 mAb. Enhancement of immunomodulation assessed by (A) significant increase in the frequency of Foxp3⁺CTV⁻ Tregs and (B) significant reduction of injurious anti-MPO autoimmunity assessed by MPO-ANCA serum levels, (C) MPO recall CD4⁺cell proliferation, frequency of (D and E) MPO recall-stimulated LN cells producing IL-17A and IFN-γ, and (F) skin MPO DTH responses. Similarly, glomerular injury was attenuated by treatment with anti-CD20 mAb compared with Ctrl-treated mice, but when anti-CD25 mAb was administered together with anti-CD20 mAb, GN was either not attenuated or made more severe as assessed by (G) albuminuria, (H) glomerular segmental necrosis, and (I–K) glomerular leukocyte infiltration. Error bars represent mean ± SEM with statistical analysis by one-way ANOVA, followed by a Tukey post hoc test. cpm, counts per minute; FIA, Freund incomplete adjuvant; gcs, glomerular cross-section; OD450nm, opotical density 450 nm. *P=0.05; **P=0.01; ***P<0.001; ****P<0.001.

Administration of Anti-CD20 mAb-Induced Apoptotic Splenocytes to Mice with Established Anti-MPO Autoimmunity Attenuates Anti-MPO Autoimmunity and GN

We next sought to assess the mechanisms of anti-CD20 mAb expansion of Tregs. It is known that anti-CD20 mAb powerfully induces B cell apoptosis. Therefore, we performed a proof-of-concept experiment to test the potential of preventing anti-MPO autoimmunity by transferring anti-CD20 mAb in vitro–induced apoptotic B220⁺ cells (washed free of unbound anti-CD20 mAb) 24 hours prior to immunizing with MPO in FCA. Mice receiving B220⁺ apoptotic B cells (induced by anti-CD20 mAb in vitro) did not develop anti-MPO immune responses with significant increase in the proportion of proliferating anti-MPO–stimulated Tregs compared with mice receiving PBS (Supplemental Figure 4). Given that adoptive transfer of anti-CD20 mAb in vitro–induced B cells was able to prevent anti-MPO autoimmunity, we next examined the therapeutic efficacy of administering apoptotic splenocytes (rendered apoptotic by treating splenocytes from naïve mice with anti-CD20 mAb in vitro) to mice with established anti-MPO autoimmunity (day 10). We compared three groups of mice developing anti-MPO autoimmunity: saline (vehicle), untreated splenocytes, and anti-CD20 mAb in vitro–induced apoptotic splenocytes. Transfer of 5×10⁷ anti-CD20 mAb-induced apoptotic splenocytes significantly reduced systemic anti-MPO–specific DTH and MPO-ANCA IgG compared with mice that received 5×10⁷ untreated splenocytes or vehicle (saline) (Figure 5, A and B). The frequency of proliferating anti-MPO splenic Tregs was increased in MPO-immunized mice receiving in vitro anti-CD20 mAb-induced apoptotic splenocytes compared with mice that received untreated splenocytes or vehicle (Figure 5C). Administration of anti-CD20 mAb-treated apoptotic splenocytes significantly attenuated kidney injury compared with mice that received untreated splenocytes or vehicle (Figure 5, D–H). These data demonstrated that transfer of anti-CD20 mAb-induced splenocytes in vitro (washed free from unbound anti-CD20 mAb) enhances Treg immunoregulation of anti-MPO autoimmunity and attenuates kidney injury. Taken together, these data show that it is possible to produce effective anti-CD20 mAb-mediated treatment indirectly without the direct administration of free anti-CD20 mAb and its associated damage to humoral immunity.

Figure 5.

Administration of in vitro anti-CD20 mAb-induced apoptotic B cells (in the absence of anti-CD20 mAb) to mice developing anti-MPO induces Tregs, immunomodulation, and disease attenuation. Administration of splenocytes (Sp) rendered apoptotic by in vitro exposure to anti-CD20 mAb to mice developing anti-MPO GN significantly reduced the generation of anti-MPO autoimmunity demonstrated by reduced (A) MPO recall dermal DTH responses and (B) MPO-ANCA IgG but (C) significantly enhanced the expansion of CD4⁺Foxp3⁺ Tregs proliferating in response to MPO recall. Compared with mice administered control untreated Sp, anti-CD20 mAb-treated B cell administration significantly reduced glomerular injury, including (D) albuminuria, (E) segmental necrosis, (F) glomerular T cell accumulation, and (G and H) glomerular leukocyte infiltration. Error bars represent mean ± SEM with statistical analysis by one-way ANOVA, followed by a Tukey post hoc test. Apop-Sp, apoptotic splenocytes; FIA, Freund incomplete adjuvant; gcs, glomerular cross-section; OD450nm, optical density 450 nm; Vehicle, saline.*P=0.05; **P=0.01.

Splenic Macrophages Play a Nonredundant Role in Processing Anti-CD20 mAb-Induced Apoptotic B Cells to Achieve Therapeutic Benefit

Splenic macrophages play important roles in clearing apoptotic cells while maintaining immunologic homeostasis. To determine the requirement for specialized splenic macrophages in processing anti-CD20 mAb-induced apoptotic cells, two groups of mice both with induced anti-MPO autoimmunity treated with apoptotic splenocytes (induced by anti-CD20 mAb in vitro) was performed. One group was treated with clodronate liposomes to deplete splenic macrophages, whereas a control group received PBS liposomes (day 10). Clodronate liposomes depleted splenic macrophages as indicated by the lack of a macrophage-specific marker anti-Metallophilic Macrophages while leaving splenic dendritic cells intact (Figure 6A). MPO-immunized mice depleted of splenic macrophages developed significantly enhanced systemic anti-MPO–specific DTH responses and increased numbers of splenocytes producing proinflammatory IFN-γ and IL-17A compared with mice receiving PBS liposomes (Figure 6, B–D). MPO-ANCA IgG titers were unaffected (Figure 6E). Splenic macrophage depletion reduced the capacity of administered anti-CD20 mAb-induced apoptotic cells to enhance the frequency of CD4⁺Foxp3⁺ Tregs (Figure 6E) and IL-10–producing CD4⁺Foxp3⁻ Tr1 cells (Figure 6F) compared with controls with intact splenic macrophages. Additionally, the frequency of Bregs (CD1^hiCD5^hi) and IL-10–producing CD19⁺ B cells was significantly reduced in mice depleted of splenic macrophages (Figure 6, G and H). Therefore, splenic macrophages are important in the clearance of anti-CD20–induced apoptotic B cells to enhance immunoregulation to attenuate established anti-MPO autoimmunity.

Figure 6.

Splenic macrophages play nonredundant roles in processing anti-CD20 mAb-induced apoptotic B cells to drive regulation of anti-MPO autoimmunity. (A) Administration of clodronate-encapsulated liposomes deleted splenic macrophages (anti-Metallophilic Macrophages [MOMA-1⁺] cells) while leaving splenic dendritic (CD11c⁺) cells intact. Control mice were administered vehicle: PBS liposomes. Both groups of mice had anti-MPO autoimmunity induced and were then both treated by transfer of in vitro anti-CD20 mAb-induced apoptotic splenocytes (Sp). (B) Mice with depleted splenic macrophages developed robust anti-MPO autoimmunity characterized by strong dermal MPO DTH responses and production of proinflammatory cytokines by Sp cells after recall MPO challenge, whereas (C and D) mice with intact splenic macrophages responded to transferred apoptotic Sp with significant reduction in the development of anti-MPO autoimmunity. (E) No difference in MPO-ANCA IgG was observed between groups. (F–I) Significant expansion of regulatory cells, including CD4⁺Foxp3⁺, Tr1 cells, B regulatory cells (Bregs), and CD1⁺ IL-10⁺ cells, was observed in mice with intact splenic macrophages compared with control macrophage-depleted mice. Error bars represent mean ± SEM with statistical analysis by unpaired t test. FIA, Freund incomplete adjuvant; OD450nm, optical density 450 nm; *P=0.05; **P=0.01; ***P<0.001.

Discussion

We found that treatment of our murine model of anti-MPO GN with a murine anti-CD20 mAb significantly attenuated disease. As in the human disease, murine anti-CD20 mAb treatment induced rapid and sustained depletion of circulating B cells. Assessment after 18 days of treatment demonstrated that both humoral and cellular anti-MPO autoimmunity as well as GN were significantly attenuated. Thus, we thought this system provides the opportunity to define the mechanisms of anti-CD20 mAb in anti-MPO GN. Moreover, we believe that our anti-CD20 mAb treatment of anti-MPO GN is a model highly relevant to human rituximab treatment of AAV: microscopic polyangiitis. The evidence to support this relates to both the animal model and the availability of murine anti-CD20 mAb treatment. The anti-MPO GN model shares the essential features of the human disease.

Autoimmunity

There is a close overlap in the sequences of the dominant nephritogenic MPO peptides in the mouse model and humans.¹⁸ Immunization of mice with the human nephritogenic MPO peptide induces anti-MPO autoimmunity and GN in mice.¹⁸ Murine MPO-ANCA stains neutrophils in a p-ANCA pattern,¹³ and transfer of high-titer murine MPO-ANCA to normal mice causes GN.¹⁹

Pathology

Anti-MPO autoimmunity in mice induces focal glomerular necrosis with pauci-immune antibody staining and leukocytic infiltration with macrophages, neutrophils, T cells, fibrin, and extracellular MPO.^4,10,13,20 All of these features are common to the human disease.²¹

Clinically

In mice with anti-MPO autoimmunity, the major targeted organ is the kidney, which develops proteinuria, hematuria, and reduction in glomerular filtration. The murine anti-CD20 mAb was engineered to have the same subclass, binding characteristics, and mechanisms of action on B cells as rituximab.¹⁶

The first finding relevant to the mechanisms of anti-CD20 mAb treatment was the major expansion of MPO-specific Tregs. We found enhanced MPO-specific proliferation of CD4⁺Foxp3⁺ Tregs from anti-CD20 mAb-treated mice. In an in vitro Treg suppression assay, anti-CD20 mAb-induced Tregs were more effective in suppressing MPO recall responses of Teffs than control-treated Tregs. Anti-CD20 mAb-induced Tregs also consumed more IL-2 while immunomodulating MPO-stimulated Teffs. To determine the role of anti-CD20 mAb-induced Tregs on mediating immunomodulation of this disease, we compared the effect of anti-CD20 mAb when coadministered with a Treg inhibiting/depleting mAb: anti-CD25 mAb. We found that in the absence of effective Treg capacity, anti-CD20 mAb had no protective effect on autoimmunity or GN. Taken together, these data show that Tregs are a nonredundant requirement for successful anti-CD20 mAb treatment. These observations suggest that in addition to anti-CD20 mAb’s capacity to attenuate humoral autoimmunity, this antibody induces T cell–mediated immunoregulation. To explore the relative contributions of these two components of treatment, we looked at an earlier treatment time point (6 days after a single dose of anti-CD20 mAb). At this earlier time point, significant attenuation of GN as well as significant enhancement in Treg expansion and reduction of anti-MPO cellular autoimmunity was observed. However, there was no significant reduction in circulating MPO-ANCA IgG. These findings strongly suggest that the anti-CD20 mAb treatment induced expansion of Tregs and is functionally relevant to the early attenuation of disease by anti-CD20 mAb therapy. The induction of effective Tregs is rapid, with a swift effect on attenuation of anti-MPO T cell autoimmunity and GN. However, attenuation of humoral autoimmunity takes longer to occur due to the long t _1/2 of circulating Ig and continued function of CD20-negative plasma cells.

We next set out to explore the mechanism of anti-CD20 mAb induction of enhanced immunomodulation by Tregs. We cocultured B220⁺ B cells with anti-CD20 mAb in vitro and observed that >95% became apoptotic. The demonstration that anti-CD20 mAb treatment generated large numbers of apoptotic B cells and enhanced Treg capacity suggested that anti-CD20 mAb treatment may co-opt the homeostatic senescent cell splenic clearance pathway to clear the generated apoptotic B cells. This pathway protects the immune system from recognizing proinflammatory residual cytosolic and nuclear proteins of dying cells by enclosing them via cell apoptosis. These apoptotic cells are cleared in the spleen by phagocytosis by specialized macrophages. This generates immunomodulatory cytokines, ensuring that antigen presented by local dendritic cells directs naïve T cells recognizing self-antigens to preferentially differentiate down T regulatory pathways.^22,23

To test this hypothesis, we first assessed the capacity for in vitro anti-CD20 mAb-induced apoptotic B220⁺ cells to mediate the expansion of Tregs in vivo. We incubated splenic B220⁺ cells with anti-CD20 mAb, then washed the cells to remove free and unbound anti-CD20 mAb, and transferred them into naïve mice. We then induced anti-MPO autoimmunity. Mice that received vehicle developed strong anti-MPO autoimmunity. Mice receiving anti-CD20 mAb-induced apoptotic B220⁺ cells did not develop anti-MPO autoimmunity; instead, they developed expansion of CD4⁺Foxp3⁺ Tregs making IL-10 recall responses to MPO.

Next, we assessed the capacity for in vitro–treated anti-CD20 mAb-induced apoptotic splenocytes to substitute for anti-CD20 mAb in treating anti-MPO GN. Mice with established anti-MPO autoimmunity were treated with either anti-CD20 mAb-induced apoptotic splenocyte cells or untreated splenocytes. Controls included a group of anti-MPO GN mice without splenocyte treatment. Control untreated splenocytes had no therapeutic effects, but anti-CD20 mAb-induced apoptotic splenocyte treatment provided significant protection of the same magnitude as anti-CD20 mAb, despite the absence of anti-CD20 mAb treatment.

To confirm that anti-CD20 mAb treatment co-opts the homeostatic senescent cell splenic clearance pathway, we used clodronate-encapsulated liposomes to delete splenic macrophages in mice with established anti-MPO autoimmunity prior to the administration of in vitro anti-CD20 mAb-induced apoptotic splenocytes. Clodronate liposomes significantly reduced numbers of splenic Tregs and IL-10–producing Tr1 cells as well as regulatory B cells and IL-10–producing CD19⁺ B cells compared with mice with intact splenic macrophages. These results confirm that the homeostatic senescent cell clearance pathway is critical for the effective treatment provided by anti-CD20 mAb. Taken together, administration of anti-CD20 mAb-induced apoptotic cells is sufficient to account for the rapid expansion of Tregs and effective attenuation of anti-MPO autoimmunity induced by anti-CD20 mAb alone. However, a confounder in including analysis of the effect of clodronate liposomes on GN is that clodronate liposomes would also decrease the number of glomerular macrophages so that even if tolerance was not induced in mice receiving apoptotic splenocytes, the expected induced kidney injury would not occur due to the confounding depletion of glomerular macrophages (Supplemental Figure 5).

The landmark RAVE trial showed no inferiority of rituximab to standard of care therapy, cyclophosphamide, in a randomized, prospective trial on patients with acute AAV. In this study, careful observation of patients found no relationship between rituximab-induced disease remission induction and fall in ANCA titers. The authors commented that “as yet undefined” pathways, beyond anti-CD20 mAb-induced depletion of humoral autoimmunity could be responsible for disease remission. The results of this murine study closely followed the RAVE study outcomes, showing the lack of relationship between ANCA titers and disease. Taken together, they support the importance of the induction of tolerance in the therapeutic efficacy of rituximab treatment and suggest the potential pathways that may direct this outcome.

There are growing concerns that prolonged use of rituximab contributes to cases of prolonged hypogammaglobulinaemia.²⁴ Rituximab targeting of B cells in addition to its critical role in B cell ablation also plays an additional therapeutic role by the induction of significant B cell apoptosis and MPO tolerance in the face of disease-inducing autoimmunity to MPO. This knowledge may enable us to devise protocols for using anti-CD20 mAb in ways that minimize innocent bystander damage to B cells.

Disclosures

A.R. Kitching reports consultancy agreements with Toleranzia and Visterra; research funding from CSL Limited and Vifor Pharma; honoraria from Visterra; scientific advisor or membership with the Australian and New Zealand Vasculitis Society; and other interests/relationships as cochair of the Australian and New Zealand Society of Nephrology Research Advisory Committee and a member of the Kidney Health Australia Research Advisory Group. All remaining authors have nothing to disclose.

Funding

This work was supported by National Health and Medical Research Council grant APP1147388.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2020060834/-/DCSupplemental.

Supplemental Figure 1. Ovalbumin-immunized mice administered sheep anti-mouse GBM globulin do not develop anti-MPO GN.

Supplemental Figure 2. Anti-CD20 mAb does not affect glomerular neutrophil recruitment induced by a subnephritogenic dose of sheep anti-mouse GBM globulin.

Supplemental Figure 3. Photomicrographs of kidney sections from mice with established anti-MPO autoimmunity treated with anti-CD20 mAb or mouse IgG2a (control): day 32 model.

Supplemental Figure 4. Administration of ex vivo anti-CD20 mAb-induced apoptotic B cells (in the absence of anti-CD20 mAb) to mice developing anti-MPO autoimmunity induces Tregs and immunomodulation.

Supplemental Figure 5. Mice with established anti-MPO autoimmunity were administered either PBS liposomes or clodronate liposomes before the treatment with anti-CD20 mAb in vitro induced apoptotic splenocytes.