A type II anti-CD20 monoclonal antibody eciently depletes CNS B cells in a mouse model of multiple sclerosis

Background: Successful therapy with anti-CD20 monoclonal antibodies (mAbs) has 21 reinforced the key role of B cells in the immunopathology of multiple sclerosis. While 22 treatment with currently available anti-CD20 mAbs results in rapid and robust elimination of 23 vascular B cells, B cells residing within compartments of the central nervous system (CNS) 24 are not well targeted. The aim of this study was to determine the effects of a novel class of 25 anti-CD20 mAbs on vascular and extravascular CNS-infiltrating B cells in experimental 26 autoimmune encephalomyelitis (EAE), an animal model of multiple sclerosis. 27 Methods: Male double transgenic hCD20xhIgR3 mice and male wild-type C57BL/6 (B6) 28 mice were injected with human myelin oligodendrocyte glycoprotein (MOG) 1–125 to induce 29 chronic EAE. On days 19, 22, and 25 after immunization, the hCD20xhIgR3 mice were 30 injected intravenously with an anti-human CD20 mAb (5 mg/kg), either rituximab (a type I 31 anti-CD20 mAb) or obinutuzumab (a type II humanized anti-CD20 mAb). The B6 mice 32 received a dose of the murine anti-mouse CD20 antibody 18B12. Development of EAE was 33 assessed daily. Seven days after the last injection, mice were euthanized, splenic B-cell 34 subsets were analyzed by flow cytometry, and differential gene expression determined by 35 single-cell RNA sequencing. Total serum immunoglobulin (Ig)G and anti-MOG 1–125 IgG titers 36 were measured by enzyme-linked immunosorbent assay. Reduction in CNS-infiltrated CD19 + 37 and CD3 + cells was analyzed by immunohistochemistry, and ultrastructural CNS pathology 38 was studied by transmission electron microscopy. 39 Results: Treatment with either anti-CD20 mAb had no effect on the clinical course of the 40 disease, animal weight, or serum antibody levels. Obinutuzumab Conclusions: Our results demonstrate differential effects of anti-CD20 mAbs on peripheral immune response and CNS pathology, with type II antibodies potentially being superior to type I in the depletion of tissue-infiltrating B cells. Our results suggest that type II mAbs such as obinutuzumab might be superior to type I 432 mAbs with respect to exerting effects on extravascular B cells. Our data also show that the 433 murine anti-CD20 mAb 18B12 is highly effective in reducing the number of CNS-infiltrating 434 B cells, and might thus be a good murine option for studying the depletion of tissue-infiltrating B cells using WT B6 EAE models.

The updated conceptual understanding of the involvement of B cells in the 81 immunopathophysiology of MS has mainly emerged on the basis of the success of anti-CD20 82 therapy in treating patients with RRMS [17]. In initial studies, depletion of circulating B cells 83 by rituximab, a chimeric anti-CD20 monoclonal antibody (mAb), led to a rapid reduction in 98 reorganization of CD20 into lipid rafts, type II mAbs trigger direct cell death upon binding 109 CD20 without cross-linking by secondary antibodies [33][34][35]. Antibody-dependent cell-110 mediated cytotoxicity and antibody-dependent phagocytosis can be induced by both antibody 111 types in the presence of immune effector cells. However, type II mAbs exert more potent 112 natural killer cell-mediated antibody-dependent cell-mediated cytotoxicity and increased 113 monocyte-and macrophage-mediated antibody-dependent phagocytosis compared with 114 type I mAbs, particularly when glycoengineered [36][37][38]. 115 Obinutuzumab is a humanized type II anti-CD20 mAb that is currently approved as a first-line hence FcγRIIb-mediated CD20 internalization are reduced compared with rituximab, 124 resulting in sustained availability and activity of obinutuzumab [38,44]. 125 In the current study, we tested the therapeutic efficacy of the type II mAb obinutuzumab 126 compared with that of the type I mAb rituximab in a B cell-dependent EAE model using a 127 double transgenic (dbtg) mouse line that expresses both murine and human CD20 on B cells 128 and tolerates the administration and presence of human IgGs [45,46]. 157 The anti-CD20 IgG1 mAbs obinutuzumab and rituximab, the isotype-matched control for 158 obinutuzumab (hIgG1), and the murine anti-CD20 IgG2a antibody 18B12 were kindly 159 provided by F. Hoffmann-La Roche (Basel, Switzerland). The chimeric IgG1 control for 160 rituximab (chIgG1) was purchased from Absolute Antibody (Oxford, UK) and the isotype 161 control antibody for 18B12 (muIgG2a) from Bio X Cell (Lebanon, NH, USA). Mice were 162 grouped (n = 5-6 per group) according to their EAE course and score, and treatment was 163 started at day 19 after immunization, when all mice had reached an EAE score of ≥2.5.

164
Antibodies were injected intravenously into the lateral tail vein at 5 mg/kg body weight, 165 diluted in 50 µL sterile PBS. Treatment was repeated on days 22 and 25. Animals were 166 euthanized 7 days after the last treatment by administration of CO2.

167
Serum collection and antibody enzyme-linked immunosorbent assay (ELISA) 168 Blood was taken from the inferior vena cava of each mouse and allowed to clot. Serum was 169 then collected by centrifugation at 1000 g for 15 minutes at 4°C and stored at -80°C until   The same approach was used to find genes differentially expressed between cells of differentially expressed in at least one of the investigated cell types were visualized as heat maps (pheatmap v1.0.12) using the mean log-normalized expression and gene-wise scaling.

238
Clustering was performed on the Euclidean distances with complete linkage.

239
Immunohistochemistry and histopathology of cerebellum 240 To determine the effect of type I and type II anti-CD20 mAbs on the depletion of CNS-241 infiltrating B cells, the cerebellum was studied during the chronic stage of EAE. In our 242 previous studies [49], the cerebellum showed the highest degree of B-cell infiltration  was assessed by measuring the axonal and the total nerve fiber diameter maxima, from 295 which the g-ratio was subsequently calculated by dividing the diameter of the axon by the 296 total nerve fiber diameter. A mean of 229 ± 37 randomly selected axons in 6-7 images per 297 mouse (n = 5-6 mice per group) were measured. Analysis was performed blinded.

298
Statistical analysis 299 Analysis was performed using GraphPad Prism software v9.0 (GraphPad, San Diego, CA, 300 USA) except for analysis of single-cell RNA sequencing data sets (for details see above).

301
Unless otherwise noted, data are presented as mean and minimum/maximum values. A p-302 value of ≤0.05 was considered statistically significant.

305
B6 and hCD20xhIgR3 mice (n = 5-6 mice per group) were immunized with human MOG1-125 306 and clinical EAE was scored daily. In all mice, immunization caused severe disease with a 307 chronic course that was characterized by complete paralysis of the hind limbs (Fig. 1).

308
Details on disease onset and severity in the different groups are provided in Table 1

315
To evaluate whether the depletion of CD20 + cells by obinutuzumab, rituximab, or the murine 316 mAb 18B12 had an effect on serum Ig levels, total IgG and anti-MOG1-125 IgG were 317 measured by ELISA. Blood was taken from n = 5-6 mice per group, 7 days after the end of  (Fig. 3D), and isotype-switched (CD19 + /IgG + ) 329 B cells (Fig. 3G). Rituximab treatment was less effective and mainly depleted isotype-switched B cells (Fig. 3G). 18B12, which was used as murine anti-CD20 mAb in WT B6 331 mice, was highly potent and depleted approximately 90% of all B-cell subsets except for 332 memory B cells (Fig. 3B-D). Interestingly, although memory (CD19 + /CD80 + /CD73 + ) B cells 333 express CD20, they were spared by anti-CD20 mAb-mediated depletion in all three treatment 334 groups (Fig. 3F). As expected, the relative numbers of CD20 − plasma cells 335 (CD45R − /CD138 + ) (Fig. 3E) and CD3 + T cells (Fig. 3H) remained unaffected. The relative 336 depletion compared to the isotype control is provided in Fig. 3I for each group.  (Fig. 5A, B). As above, different groups of mice were treated with obinutuzumab, 350 rituximab, 18B12, or the matched isotype controls and sacrificed 7 days after the end of 351 treatment. Compared with treatment with the isotype controls, the numbers of both diffuse 352 and dense B-cell infiltrates were reduced following treatment with obinutuzumab and 18B12 353 (Fig. 5B, D). In contrast, rituximab only depleted dense perivascular B-cell infiltrates but did 354 not target B cells infiltrating into the parenchyma (Fig. 5B, D). Obinutuzumab treatment also 355 reduced the number of diffuse T-cell infiltrates, while none of the anti-CD20 mAbs had any 356 effect on dense T-cell infiltrates (Fig. 5C, D).   Our results suggest that type II mAbs such as obinutuzumab might be superior to type I 432 mAbs with respect to exerting effects on extravascular B cells. Our data also show that the 433 murine anti-CD20 mAb 18B12 is highly effective in reducing the number of CNS-infiltrating 434 B cells, and might thus be a good murine option for studying the depletion of tissue-infiltrating We would like to thank Stephanie Link, Anita Hecht, and Andrea Hilpert for their excellent 487 support in preparing the histologic specimens. The authors would also like to thank the pRED