Selective ablation of dsDNA-specific plasma cells improves lupus nephritis in NZB/W mice

Significance

Antibody-mediated diseases comprise many different clinical disease entities, and affect more than 10% of the human population. Plasma cells have emerged as a promising therapeutic target for these diseases. However, currently available therapies target both pathogenic and protective plasma cells and carry the risk of humoral immunodeficiency. In a mouse model of autoimmunity leading to lupus nephritis, we show that a selective ablation of plasma cells secreting pathogenic anti-dsDNA autoantibodies using an “antibody-mediated cytotoxicity engager” by labeling them with their antigen, here dsDNA oligonucleotides, conjugated to an antibody that recognizes plasma cells can halt the disease. This concept of antigen-specific ablation of pathogenic plasma cells offers a promising therapeutic perspective for (auto)antibody-mediated diseases.

Abstract

Antibodies specific for double-stranded DNA (dsDNA) are thought to play a crucial role in the pathogenesis of lupus nephritis. Plasma cells have emerged as a promising therapeutic target in antibody-mediated diseases. Here, we show that dsDNA oligonucleotides longer than 80 bp and irrespective of their sequence, react with more than 80% of murine and human dsDNA antibodies. With a conjugate of a modified antibody to mouse CD138 and 100 bp dsDNA oligonucleotides, we demonstrate the selective depletion of dsDNA-specific antibody secreting plasma cells in NZB/W mice, a mouse model of lupus nephritis. Two days after injection of the conjugate, up to 80% of the dsDNA-specific plasma cells in spleen, bone marrow, and kidneys of the mice were ablated, presumably by complement or cellular cytotoxicity, targeted to them by their own secreted antibodies. Plasma cells secreting antibodies of different specificities were not affected. Treated mice showed a sustained reduction of dsDNA-specific antibodies in the serum and a significant improvement in proteinuria, i.e., kidney pathology. This demonstration of the therapeutic efficacy of antigen-specific ablation of pathogenic plasma cells offers a unique option for the precise treatment of antibody-mediated diseases.

Systemic lupus erythematosus (SLE) is a chronic, multifactorial autoimmune disease characterized by the production of autoantibodies against nuclear components, including double-stranded DNA (dsDNA). Anti-dsDNA autoantibodies, a hallmark of SLE, are closely associated with severe organ manifestations such as lupus nephritis, in which they play a critical role in pathogenesis (1, 2). We have shown before in the NZB/W mouse model of lupus nephritis that both short-lived plasmablasts and long-lived plasma cells can contribute to the production of anti-dsDNA autoantibodies (3, 4). In contrast to short-lived plasmablasts, long-lived plasma cells are resistant to conventional immunosuppression, including cyclophosphamide and mycophenolate, and to targeting of precursor B cells with drugs such as anti-CD20 and anti-BLys/BAFF (4–6). Long-lived plasma cells that secrete pathogenic antibodies thus represent a promising, as yet unmet therapeutic target (5, 7). This rationale has been confirmed by several lines of evidence: Generic ablation of plasma cells with the proteasome inhibitor bortezomib protects NZB/W mice from nephritis (8). In SLE patients who do not respond to standard immunosuppression and in whom dsDNA-specific autoantibodies persist, generic ablation of immune cells with antithymocyte globulin, including ablation of plasma cells, followed by autologous stem cell transplantation resulted in a clinical response and disappearance of the autoantibodies (9). The generic ablation of plasma cells with bortezomib (10), daratumumab (11, 12) teclistamab (13), and CD19 CAR T cells, which deplete B cells and CD19+ plasma cells (14, 15), underpin the relevance of (long-lived) plasma cells as a therapeutic target. However, as generic therapies, they target all plasma cells (and B cells) and impair humoral immunity as such, with increased risk of infectious diseases. We have developed an alternative therapeutic strategy that aims at a precise and selective depletion of plasma cells that secrete antibodies of a defined specificity (16).

Here we show that a 100 bp dsDNA oligonucleotide is recognized by almost all autoantibodies against dsDNA in SLE patients and NZB/W mice, a mouse model of SLE. This oligonucleotide conjugated to CD138-specific antibodies results in a significant, selective, and sustained ablation of up to 80% of dsDNA-specific plasma cells and a significant reduction in proteinuria in NZB/W mice.

Results

dsDNA Oligonucleotides React with dsDNA-Specific Autoantibodies.

As early as 1980, it has been described that the length of dsDNA fragments has an influence on their binding to dsDNA-specific autoantibodies (17). Here, we systematically compared dsDNA oligonucleotides with lengths of 9, 21, 30, 40, 61, 80, 100, and 140 bp from a designed random, non-coding sequence for their ability to inhibit the binding of serum antibodies of three SLE patients to calf thymus dsDNA. Oligonucleotides shorter than 61 bp only slightly inhibited binding, while oligonucleotides longer than 61 bp significantly blocked binding (Fig. 1A). Binding inhibition increased with increasing oligonucleotide length and saturated for oligonucleotides longer than 80 bp. The sequence of the oligonucleotides was irrelevant since four 100 bp long dsDNA oligonucleotides with different sequences were all recognized equally by dsDNA-specific antibodies of NZB/W mice (Fig. 1B) and SLE patients (SI Appendix, Fig. S2).

Fig. 1.

dsDNA oligonucleotides can detect anti-dsDNA autoantibodies in SLE patients and NZB/W mice by analyzing their inhibition of the anti-dsDNA binding using an anti-dsDNA ELISA. Blocking rates were calculated by using concentration 0 as 0%. (A) The serum of an anti-dsDNA-positive SLE patient #19 (see also C) was incubated at a 1:200 dilution with different concentrations of a dsDNA oligonucleotide of different lengths. (B) A serum pool of 28-wk-old NZB/W mice was incubated at a 1:1,000 dilution with different concentrations of a 100 bp dsDNA oligonucleotide with four different sequences (SI Appendix, Table S1). (C) Serum samples from 13 anti-dsDNA-positive SLE patients were incubated with different concentrations of a 100 bp dsDNA oligonucleotide. Patient #16 (green curve) showed the highest blocking effect. All serum samples were diluted to an anti-dsDNA antibody concentration of 0.1 I U/mL. (D) Different dilutions between 1:200 and 12,800 of a serum pool from 28-wk-old NZB/W mice were incubated with different concentrations of a 100 bp dsDNA oligonucleotide. The data represent the mean of two values, in B and D with SD.

One of the 100 bp dsDNA-oligonucleotides was selected for further use and examined in more detail for its recognition by dsDNA-specific antibodies. Serum samples from 13 SLE patients with elevated anti-dsDNA antibody levels (range: 27.2 to 150 IU/mL) and pooled serum samples from 28-wk-old NZB/W mice were preincubated with increasing concentrations of the oligonucleotide (0 to 500 nM) and then analyzed for binding to dsDNA. The oligonucleotide inhibited binding in a dose-dependent manner. At a concentration of 500 nM, all samples from SLE patients had blocking rates greater than 52%, with 36% of samples (5 out of 14) having blocking rates greater than 75% (Fig. 1C). Binding to dsDNA from pooled sera of 28-wk-old NZB/W mice was blocked approximately 85% by the 100 bp dsDNA oligonucleotide at a concentration of 500 nM at serum dilutions of 1:200 to 1:800. The inhibitory effect decreased at serum dilutions of 1:1,600 or higher (Fig. 1D). Serum samples from three individual 28-wk-old NZB/W mice exhibited similar blocking behavior (SI Appendix, Fig. S1).

These results show that a 100 bp dsDNA oligonucleotide is recognized by the vast majority of anti-dsDNA autoantibodies in both human SLE and lupus-prone NZB/W mice, dependent on length but not on sequence.

A 100 bp dsDNA Oligonucleotide Identifies dsDNA-Specific Plasma Cells.

dsDNA-specific plasma cells are detectable by ELISPOT technology, using calf thymus DNA (CTDNA) as antigen. Compared to CTDNA, the 100 bp dsDNA oligonucleotide used as an antigen in the dsDNA-specific hybridoma cell line 1E5 produced just as many specific spots (Fig. 2A), although the sharpness of the spots appeared slightly reduced when coated with 100 bp dsDNA oligonucleotide. Comparison of CTDNA and the 100 bp dsDNA oligonucleotide as antigen also showed no significant difference when tested in splenocytes from NZB/W mice (Fig. 2B).

Fig. 2.

100 bp dsDNA oligonucleotide can detect dsDNA-specific plasma cells. (A) Elispot was performed using microtiter plates coated with a 100 bp dsDNA oligonucleotide (oligo), calf thymus DNA (CTDNA), or anti-mouse IgG (anti-muIgG) using a dsDNA- or an ovalbumin (OVA)-specific hybridoma cell line or a mixture. The horizontal lines represent the mean and SD of eight tests. The statistical analysis using one-way ANOVA revealed no significant differences between the three groups. The images show spots of a representative analysis. (B) Elispot was performed with microtiter plates coated with a 100 bp dsDNA oligonucleotide or calf thymus DNA (CTDNA) using splenocytes from 26-wk-old NZB/W mice or Balb/c mice. Six tests were performed. The images represent spots of a representative analysis. The two-tailed Mann–Whitney U-test was used for statistical analysis (ns = not significant). (C) Flow cytometric analysis of dsDNA- and OVA-specific hybridoma cells using 100 bp dsDNA oligonucleotide-Fam and OVA-Alexa 405. A representative example of the three experiments conducted is shown. (D) Flow cytometric analysis of dsDNA-specific plasma cells with 100 bp dsDNA oligonucleotide-Fam and Cy5 in the spleen and bone marrow (BM) of 28-wk-old NZB/W and in PBMC of an active SLE patient. Splenic and BM cells from NZB/W mice were pregated with surface CD138+ and intracellular kappa+. PBMC from the SLE patient were pregated with surface CD19+CD20−CD27++ and intracellular kappa/lambda+. 100 bp dsDNA oligonucleotide-FAM and 100 bp dsDNA oligonucleotide-Cy5 were used exclusively for intracellular double-positive staining. The specificity of the staining was confirmed by blocking with unconjugated 100 bp dsDNA oligonucleotide. A representative example of the three experiments conducted is shown.

While high molecular weight dsDNA cannot be used for immunofluorescence detection of dsDNA-specific B and plasma cells due to its “sticky” nature, the 100 bp dsDNA oligonucleotide conjugated to fluorochrome Fam selectively stained intracellular antibodies of dsDNA-specific 1E5 hybridoma cells, but not those of ovalbumin-specific 4C9 hybridoma cells (Fig. 2C).

For the cytometric detection of dsDNA-specific plasma cells in NZB/W mice and SLE patients, we used the 100 bp dsDNA oligonucleotide conjugated with two different fluorochromes for double staining. Among the gated antibody-secreting cells from spleen and bone marrow of NZB/W mice and of peripheral blood of SLE patients, CD138+ intracellular kappa+ in mice, CD19+ CD27++ intracellular kappa/lambda++ in SLE patients, less than 1% was labeled by intracellular staining with both dsDNA oligonucleotide conjugates. This staining was specific for the 100 bp dsDNA oligonucleotide, as it could be inhibited by unconjugated dsDNA oligonucleotide (Fig. 2D). The 100 bp dsDNA–oligonucleotide–fluorochrome conjugates also detected dsDNA-specific B cells, both in NZB/W mice and in SLE patients, when surface staining was performed (SI Appendix, Fig. S3).

These results show that a 100 bp dsDNA oligonucleotide is as efficient as CTDNA in detecting dsDNA-specific plasma cells by ELISPOT and can be used for immunofluorescent identification and isolation of dsDNA-specific plasma cells and B cells by fluorescence-activated cell sorting.

A Conjugate of Anti-CD138 and dsDNA Oligonucleotide Labels Plasma Cells In Vivo.

Based on our previous proof-of-concept study demonstrating the selective depletion of plasma cells secreting ovalbumin-specific antibodies in mice (16), in the present experiments, we targeted dsDNA-specific plasma cells with a conjugate consisting of a 100 bp dsDNA oligonucleotide and a CD138-specific antibody.

The 100 bp dsDNA oligonucleotide was biotinylated and conjugated to a streptavidin-labeled anti-mouse CD138 antibody with a human IgG1 Fc region. This Fc region is non-functional in mice with respect to triggering effector functions of binding to murine complement or ADCC receptors.

After intraperitoneal injection of 200 μg of the conjugate per mouse, staining of plasma cells in vivo was examined by counterstaining them with anti-human IgG (Fig. 3A). Staining was rapid, reaching a maximum at 2 h and then ceasing within the next 2 d, with staining occurring more rapidly for splenic plasma cells than for plasma cells of the bone marrow (Fig. 3 B and C), reflecting the kinetics previously observed with an anti-CD138/ovalbumin conjugate (16). The staining was specific for plasma cells, as compared to B and T lymphocytes (Fig. 3A).

Fig. 3.

Analysis of plasma cell labeling with the anti-CD138/dsDNA oligonucleotide conjugate in vivo. (A) Representative dot blot and histogram of plasma cell labeling in the spleen and bone marrow of an NZB/W mouse 3 h after i.p. injection of the anti-CD138/dsDNA oligonucleotide conjugate, analyzed by flow cytometry. The conjugate was labeled with anti-human IgG1-APC (anti-huIgG1), plasma cells (PCs) were identified as CD267 (TACI) and CD138 double positive. (B) The kinetics of conjugate labeling on the surface of plasma cells in spleen and bone marrow of NZB/W mice was analyzed by flow cytometry with staining of anti-human IgG1-FITC (ahIgG1) using mean fluorescence intensity (MFI) between 2 and 96 h after injection of the conjugate. The data represent mean values of MFI with SD of 3 mice. (C) Representative histograms of plasma cell labeling kinetics in one of three mice at each time point. The pool of two independent experiment series is shown.

The Anti-CD138/100 bp dsDNA Oligonucleotide Conjugate Selectively Ablates dsDNA-Specific Plasma Cells and Stops Nephritis in NZB/W Mice.

Since the conjugate is no longer detectable on the cell surface of plasma cells 48 h after administration, we determined depletion of plasma cells 2 d after its injection. Plasma cells were enumerated by flow cytometry and ELISpot in the spleen, bone marrow, and kidneys of NZB/W mice treated or not with 200 µg per mouse of the CD138/100 bp dsDNA oligonucleotide conjugate (Fig. 4).

Fig. 4.

Ablation of plasma cells in spleen, bone marrow, and kidneys of 16- to 18-wk-old NZB/W mice 48 h after a single intraperitoneal injection of the conjugate (200 μg/mouse) compared to untreated age-matched mice (n = 10 mice per group), analyzed by flow cytometry using intracellular staining with 100 bp dsDNA oligonucleotide. The data represent mean values with SD. The Mann–Whitney U-test (two-tailed) was used for statistical analysis. Two independent but identical series of experiments were combined.

The treatment did not result in a significant change of the total number of plasma cells in the organs examined. However, dsDNA-specific plasma cells were reduced by 82% in the bone marrow, 73% in the spleen and 69% in the kidney, as determined by flow cytometry (Fig. 4). Similar results were obtained by ELISpot assay, with depletion rates of 68%, 68%, and 55% in the bone marrow, spleen, and kidney, respectively (SI Appendix, Fig. S4). These results demonstrate the selective and efficient ablation of dsDNA-specific plasma cells in the mice by the anti-CD138/100 bp dsDNA oligonucleotide conjugate.

Next, we examined the effect of this selective ablation of dsDNA-specific plasma cells in NZB/W mice on the development of lupus nephritis. To prevent the regeneration of dsDNA-specific plasma cells from their precursors in this murine model (18), we treated the NZB/W mice with the conjugate every 2 wk, starting at 20 wk of age, before the onset of proteinuria. The level of proteinuria was then monitored as a clinical marker of nephritis. Treatment was discontinued at week 30, since the treated mice had developed antibodies against the conjugate (Fig. 5).

Fig. 5.

Behavior of (A) proteinuria, (B) IgG anti-dsDNA, (C) IgG antiovalbumin, and (D) IgG anti-conjugate antibody levels during therapy with anti-CD138/dsDNA oligonucleotide conjugate injected every 2 wk (200 μg/mouse, n = 7) vs. untreated age-matched control group (n = 8). Serum dilutions of 1:1,000 were used in the ELISA. The arrows represent the times of injection. The data represent mean values with SEM. For statistics, two-way ANOVA was used. Two independent but identical series of experiments were combined.

All mice had been immunized with ovalbumin as a reference antigen 8 wk before the start of the treatment with the anti-CD138/100 bp dsDNA oligonucleotide. While the ovalbumin-specific IgG antibody levels remained stable in both treated and untreated mice, IgG anti-dsDNA autoantibodies decreased to approximately 50% in the treated mice as compared to the untreated mice, by week 30. Proteinuria, which was low at week 20 (mean 0.05 g/L), remained low in the treated mice and reached 1.3 g/L at week 28. In comparison, in untreated mice proteinuria increased to a mean value of 9.4 g/L at week 28 (Fig. 5). As expected, also anti-dsDNA antibodies of the IgM class decreased significantly while anti-red blood cell autoantibodies remained stable in the treated group (SI Appendix, Fig. S5).

Discussion

A price to pay for our incredibly powerful adaptive immune system is the accidental development of autoimmunity, in particular in the process of somatic hypermutation of antibody genes of activated B lymphocytes, as exemplified for the somatic evolution of dsDNA-specific autoantibodies (19). If these manage to become long-lived plasma cells, the production of autoantibodies becomes independent of continuous B cell activation and refractory to conventional immunosuppressive therapies, targeting activated B cells. This first has been demonstrated for dsDNA-specific plasma cells in the NZB/W model of lupus nephritis (4). Plasma cells, in particular long-lived ones, thus qualify as a distinct target of innovative immunotherapies aiming at prevention and cure of (auto)antibody-mediated diseases. Indeed, generic ablation of plasma cells by the proteasome inhibitor bortezomib prevents lupus nephritis in NZB/W mice (8) and ameliorates it in patients with refractory SLE (10). Among the autoantibodies of SLE patients and NZB/W mice, those recognizing dsDNA are considered to play a key role in forming immune complexes in the kidneys and causing nephritis (1, 2). Here, we describe an approach to target those plasma cells of mice and humans, which secrete dsDNA-specific autoantibodies. We demonstrate that a 100 base pair (bp) long dsDNA oligonucleotide, independent of its sequence, is recognized by the vast majority of dsDNA-specific autoantibodies of NZB/W mice and patients with SLE. It can be labeled with fluorochromes or other tags to identify and characterize dsDNA-specific plasma cells and B cells. We conjugated the 100 bp dsDNA oligonucleotide to an antibody binding to plasma cells, to deplete dsDNA-specific plasma cells in vivo, by complement-mediated (CDC) or cellular cytotoxicity (ADCC) induced by their own, dsDNA-specific antibodies, relocated to their cell surface. The conjugate, when injected into NZB/W mice, indeed selectively eliminates dsDNA-specific plasma cells in the kidney, spleen, and bone marrow of the mice, stably reduces serum titers of dsDNA-specific autoantibodies and significantly ameliorates nephritis.

For construction of the bispecific conjugate, we used an antibody specific for CD138 (syndecan-1) a protein abundantly expressed on the cell surface of most plasma cells (20). The antibody was labeled with streptavidin and conjugated to a 100 bp long biotinylated dsDNA oligonucleotide. It has previously been described that dsDNA-specific autoantibodies are directed against the phosphodiester backbone or double helix confirmation itself (21, 22). Here, we confirmed that the recognition of dsDNA by the autoantibodies of NZB/W mice and SLE patients is sequence independent. Beyond that we show that a 100 bp dsDNA oligonucleotide is sufficient to be recognized by most of those dsDNA-specific autoantibodies of SLE patients and of NZB/W mice which bind to calf thymus DNA. At a dose of 200 μg per mouse, plasma cells in spleen and bone marrow were efficiently labeled with the conjugate, with a maximum at 2 h. Up to 80% of the dsDNA-specific plasma cells were eliminated from the bone marrow 2 d later, demonstrating the efficacy of the approach. Previous in vitro and ex vivo experiments have shown that plasma cells can be ablated by complement-dependent cytotoxicity (CDC) in an antigen-specific manner (23). Whether antibody-dependent cytotoxicity (ADCC) is also involved in this remains to be investigated. The total number of plasma cells and the number of ovalbumin- and erythrocyte-specific plasma cells did not change significantly, demonstrating the selectivity of the treatment. Obviously, cross-feeding of dsDNA-specific antibodies from dsDNA-specific plasma cells to other plasma cells did not pose a major problem. However, other variables await further optimization. First, we did not block dsDNA-specific autoantibodies already present in the serum, which may have affected the efficiency of the conjugate in labeling the plasma cells. For the treatment of SLE patients with high titers of anti-dsDNA antibodies, we consider it necessary to reduce the concentration of circulating autoantibodies in the bloodstream prior to administration of the conjugate, e.g., through plasma exchange, immunoadsorption, or inhibition of the neonatal Fc receptor (FcRn), in order to achieve a molar excess of the injected conjugate over the circulating anti-dsDNA antibodies. When the conjugate was injected into 20-wk-old NZB/W mice, as described in the present study, there was evidently a significant molar excess of the conjugate relative to the circulating anti-dsDNA antibodies, thereby preventing the binding of all plasma cells by serum antibodies. This is obvious from the results shown, namely that only dsDNA-specific plasma cells were ablated. Second, we did not combine selective ablation of dsDNA-specific plasma cells with a regimen that prevents their regeneration, a systemic feature of the NZB/W mouse model. Third, we did not prevent the generation of anti-drug antibodies (ADA) against the conjugate. When we treated the mice repeatedly, after approximately 10 wk the treatment became inefficient, probably due to ADA. Similarly, we did not observe any induction of dsDNA-specific plasma cells by the conjugate, although this still needs to be formally ruled out. Conventional immunosuppression and/or therapies targeting B cells would likely be sufficient to prevent the regeneration of dsDNA-specific plasma cells and the formation of ADA-secreting plasma cells.

While the ablation of dsDNA-specific plasma cells by labeling the plasma cells with the autoantigen, as described here, is to our knowledge the first approach to target these cells selectively, alternative approaches have been described that either target secreted dsDNA-specific antibodies, or target plasma cells generically. For targeting of dsDNA-specific antibodies, the substance LJP-394 has been developed, which consists of four double-stranded oligonucleotides, each 20 bp in length, bound to a tetrakis platform, which is thought to mediate formation and clearance of drug–antibody complexes and induces B cell tolerance (24). Two randomized double-blind, placebo-controlled trials in patients with lupus nephritis failed to reach the study endpoints, but positive trends in renal endpoints were observed (25, 26). However, the plasma cells that secrete dsDNA-specific antibodies are not affected by this strategy, which could explain the inadequate therapeutic response in the trials. The generic ablation of plasma cells (and other immune cells) by daratumumab, an antibody to CD38, expressed by plasma cells and activated B and T cells, or by teclistamab, a bispecific BCMA-directed CD3 T cell engager targeting all plasma cells, in clinical studies has demonstrated a significant improvement of disease activity (11–13), as has the generic ablation of CD19 expressing B and plasma cells, either by CD19 CAR-T cells (14), or by BCMA-CD19 CAR-T cells (15). The situation is similar with regard to the generic ablation of plasma cells using Atacicept, a recombinant of the extracellular domain of TACI with an IgG1 Fc (27), and the proteasome inhibitor bortezomib (10). In contrast to the generic targeting of most or all plasma cells, B cells and T cells of a patient, the selective therapeutic strategy described here aims at the precise and exclusive depletion of those plasma cells only that secrete pathogenic (auto)antibodies of a given specificity. Protective plasma cells and B cells are spared, no immunodeficiency induced, as compared to therapies addressing all B and plasma cells. Since the selective approach is based on complement-mediated cytotoxicity (CDC) (23), it furthermore will preferentially target those plasma cells secreting pathogenic anti-dsDNA antibodies, that are complement fixing and have a high affinity (1). Hypocomplementemia, which has been described in NZB/W mice (28), does not appear to significantly impair the effectiveness of dsDNA-specific plasma cell ablation. We do not expect hypocomplementemia in human SLE to pose a major problem for plasma cell ablation, as we have previously treated patients with refractory SLE using antithymocyte globulin (ATG) followed by autologous stem cell transplantation and observed complete plasma cell ablation in their bone marrow (9). Treatment with the monoclonal anti-CD38 antibody daratumumab also led to a significant elimination of plasma cells (11). Whether this occurs via CDC or ADCC remains an open question. Finally, while generic ablation of all plasma cells over extended time periods in the patients expectedly bears the risk of significant susceptibility for infections, generic drugs most efficiently will be used at suboptimal doses, ablating not all but only some cells. The new, selective approach does not have this limitation and can aim at maximal depletion of the pathogenic plasma cells.

In summary, here we show that a 100 bp dsDNA oligonucleotide, irrespective of its sequence, is a valuable tool for identification, isolation, and analysis of dsDNA-specific B cells and plasma cells, in the NZB/W mouse model and in SLE patients. Conjugated to a plasma cell-specific antibody, the 100 bp dsDNA oligonucleotide efficiently and selectively eliminates dsDNA-specific plasma cells in NZB/W mice, induces sustained reduction of serum anti-dsDNA antibody titers and prevents nephritis, which also underscores the pathogenic role of anti-dsDNA autoantibodies in lupus nephritis. For SLE patients, this therapeutic strategy aiming at precise ablation of plasma cells according to their expression of pathogenic antibodies would be a promising option to target the cause of nephritis. Furthermore, this approach to selective therapeutic ablation of plasma cells that secrete pathogenic antibodies with defined specificity while preserving humoral immunity would be a promising option for the treatment of other (auto)antibody-mediated diseases.

Materials and Methods

dsDNA Oligonucleotides.

dsDNA oligonucleotides with different lengths between 9 and 140 base pairs (bp) were synthesized by Integrated DNA Technologies, IDT (for sequences see SI Appendix, Table S1). The 100 bp dsDNA oligonucleotide used throughout the study is a random sequence with an AT content of 55%, which was selected such that it does not show relevant homology to any annotated gene when analyzed by NCBI BLAST. In addition, three different variants of the 100 bp dsDNA oligonucleotide were synthesized: Variant 100 bp_m1 consists of a repetitive “ATCG” motif (50% AT), whereas variant 100 bp_m2 and variant 100bp_m3 are composed of repetitive partial segments of the original 100 bp sequence, with AT contents of 32% and 69%, respectively. A biotinylated version of a 100 bp dsDNA oligonucleotide was provided for the preparation of the anti-CD138/dsDNA oligonucleotide conjugate.

Preparation of an Anti-CD138/dsDNA Oligonucleotide Conjugate.

Anti-mouse CD138 antibodies containing a human IgG1 heavy chain constant region (clone: REA104; Miltenyi Biotec) were coupled with streptavidin using LYNX Rapid Streptavidin Antibody Conjugation Kit (Bio-Rad) according to the instructions for use. Biotinylated 100 bp dsDNA oligonucleotide in PBS was added to the streptavidin-coupled antibody with gentle stirring and incubated for 2 h at room temperature. The CD138/dsDNA oligonucleotide conjugate was purified using 100 kDa size exclusion filter tubes (Merck Millipore).

Antibody Detection by ELISA.

IgG anti-dsDNA, and IgG anti-ovalbumin antibodies were measured as described previously (3, 16) using HRP-conjugated goat anti-mouse IgG (SouthernBiotech) for mouse serum samples and HRP-conjugated goat antihuman IgG (SouthernBiotech) for human sera. For the detection of anti-drug antibodies in NZB/W mice, the microtiter plates were precoated with streptavidin or goat anti-human IgG (SouthernBiotech). To detect autoantibodies against red blood cells, the erythrocyte membrane was isolated from mouse blood as described previously (29) and then precoated to the microtiter plates. Optical density (OD) was measured at 450 nm using a SpectraMax Plus384 Microplate Reader (Molecular Devices) after stopping the reaction with H₂SO₄.

To investigate the blocking effect of dsDNA oligonucleotides on the binding of anti-dsDNA autoantibody, diluted serum samples from SLE patients or pooled NZB/W mouse sera were mixed with different concentrations of dsDNA oligonucleotides (from 1,000 to 0.001 nM) and then incubated on DNA-coated microtiter plates at 37 °C for 2 h. The blocking rate was calculated using OD 450, and the results of samples without the addition of dsDNA oligonucleotide were considered as 0% blocking references.

Hybridoma Cell Lines.

The anti-dsDNA mouse hybridoma cell line 1E5 (kind gift from Prof. Dr. Thomas Winkler, Erlangen, Germany) and the anti-ovalbumin (OVA) hybridoma mouse cell line 4C9 (DRFZ) were cultured in RPMI 1640 with 10% FCS (fetal calf serum).

ELISPOT for Detecting dsDNA-Specific Plasma Cells.

ELISPOT was performed as described previously using alkaline phosphatase-labeled anti-mouse IgG (SouthernBiotech) for detection (4). Spots were developed using NBT-BCIP (Thermo Scientific). Microtiter plates were read using an automated ELISPOT reader (AID Autoimmun Diagnostika GmbH), and the number of spots was counted using ELISPOT software (AID Autoimmun Diagnostika GmbH).

Analysis of Plasma Cells by Flow Cytometry.

Flow cytometric analysis was performed on human peripheral blood mononuclear cells (PBMC), hybridoma cell lines and single cell suspensions from spleen, bone marrow (femur and tibia) and kidney samples from NZB/W mice as previously described (16, 30). Live/dead staining was performed with Zombie Violet™ Fixable Viability Kit (BioLegend). For intracellular staining, cells were treated with BD Cytofix-Cytoperm fixation and permeabilization solution (BD Biosciences). The following antibodies/reagents were used for surface staining: anti-mouse CD138-PE (clone 281-2, BD Biosciences), anti-mouse CD138-PE-Vio770 (clone REA104, Miltenyi), anti-mouse CD267 (TACI)-PE (clone REA1227, Miltenyi), anti-mouse CD19-APC-Cy7 (clone: 6D5, BioLegend), antihuman IgG1-APC (clone REA293, Miltenyi), anti-human IgG-FITC (clone IS11-3B2.2.3, Miltenyi), anti-mouse CD3-PerCP (clone: 145-2C11, BioLegend), anti-human CD19 PE/Cy7 (clone HIB19, BioLegend), anti-human CD20-Pacific Blue (clone 2H7, BioLegend), and anti-human CD27 APC/Cy7 (clone M-T271, BioLegend).

The following antibodies/reagents were used for intracellular staining: anti-mouse kappa-pacific orange (clone 187.1, DRFZ), 100 bp dsDNA oligonucleotide conjugated with 6-FAM (6-carboxyfluorescein) and Cy5 (IDT), OVA-Alexa 405 (DRFZ). Plasma cells were identified as CD138 +/intracellular κ+ in mouse or CD19+CD20−CD27+/intracellular κ/λ+ in human samples.

The study was performed with a FACSCanto II cytometer (BD Biosciences), and the data were analyzed using FlowJo software (Tree Star, Inc.).

Patients and Control Blood Samples.

Blood for isolation of PBMC and serum was withdrawn from patients with SLE who met the EULAR/ACR classification criteria (31) and were recruited from the Department of Rheumatology and Clinical Immunology at Charité-Universitätsmedizin Berlin. Healthy individuals were included as controls. All patients gave written informed consent prior to the study entry, including sample collection, in accordance with the Declaration of Helsinki in its current version. The study was approved by the Ethics Committee of Charité-Universitätsmedizin Berlin (EA1/002/24).

Mice and Treatment.

Female NZB/W F1 mice were purchased from the Jackson Laboratory and kept at the animal facility of the German Rheumatism Research Center under defined, pathogen-free conditions. All experiments were performed in accordance with the German Animal Welfare Act and approved by the local authorities.

In vivo labeling of the anti-CD138/100 bp dsDNA oligonucleotide conjugate on plasma cells from spleen and bone marrow was investigated 0, 2, 24, 48, 72, and 96 h after its intraperitoneal administration in 16-wk-old NZB/W mice (200 µg/mouse) by flow cytometry.

The efficacy of plasma cell ablation in spleen, bone marrow, and kidney was investigated 2 d after a single intraperitoneal injection of the conjugate (200 µg/mouse) to 16- to 18-wk-old NZB/W mice.

For the long-term treatment experiment, 10-wk-old NZB/W mice were immunized intraperitoneally with 100 µg ovalbumin (Sigma-Aldrich) emulsified in the same amount of alum (Thermo Scientific) and boosted with the same protocol 4 wk after primary immunization. Starting at 20 wk of age, mice were treated intraperitoneally with the anti-CD138/dsDNA oligonucleotide conjugate (200 µg/mouse) every 2 wk. The control group consisted of untreated age-matched mice. Proteinuria was monitored and semiquantitatively analyzed using Albustix (Siemens) as a biomarker of nephritis. In addition, the levels of anti-dsDNA autoantibodies, anti–red blood cell autoantibodies, anti-ovalbumin antibodies, and anti-drug antibodies directed against streptavidin and the human Fc region of IgG1 were analyzed weekly by ELISA.

Statistical Analysis.

Statistical analysis and figures were generated using GraphPad Prism 11 (GraphPad, La Jolla, CA).

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.