Preclinical B cell depletion and safety profile of a brain-shuttled crystallizable fragment-silenced CD20 antibody

Abstract

Background: The blood-brain barrier (BBB) presents a major challenge for the development of monoclonal antibody (mAb)-based therapies for brain disorders. To improve the likelihood of success of such therapies, Roche Brainshuttle technology utilizes a single anti-transferrin receptor 1 (TfR1)-antigen-binding antibody fragment linked to a therapeutic antibody, allowing engagement with TfR1 to transport the therapeutic antibody into the brain via receptor-mediated transcytosis.

Methods: We compared Fc-silenced and Fc-competent variants of the Brainshuttle and the parental (non-shuttled) type II CD20 mAb, obinutuzumab in in vitro and in vivo (mouse and cynomolgus macaque) models. Endpoints assessed included B cell binding, B cell killing, tolerability, and ability to cross the BBB.

Results: The Fc-silenced Brainshuttle construct showed a superior safety profile compared with the Fc-competent construct while maintaining the ability to cross the BBB and to deplete B cells in head-to-head comparisons in human and mouse in vitro and in mouse and cynomolgus macaque in vivo models.

Conclusion: Together, our data provide a path forward for the future development of safe and efficacious brain-targeted B-cell-depleting therapies.

Key points: The BBB hinders mAb-based brain disorder therapies A brain-targeted B-cell-depleting mAb for MS that efficiently crosses the BBB via hTfR1 was developed using Brainshuttle^™ technology (1a and 1b) The Brainshuttle^™-CD20 mAb was well tolerated (2a and 2b) and displayed B-cell-killing properties (1c), paving the way for future development and clinical translation of TfR1-targetingtherapies for increased brain penetration.

Keywords: CD20; antibodies; antigens; blood‐brain barrier; central nervous system; monoclonal; multiple sclerosis.

FIGURE 1

Binding, B cell killing, and uptake of Fc‐silent and Fc‐competent Brainshuttle CD20 mAbs. (A) Properties of the different Brainshuttle and control mAbs, with overview of experimental design and model systems used. In some cases, certain assays contributed to multiple categories of evaluation, reflecting their broader utility across different experimental criteria. (B) Schematic representations of Brainshuttle and control mAbs. Grey = Fc‐competent component of IgG1; turquoise = Fc‐silent component of IgG1; red = glycan; blue = anti‐human CD20 Fab arms of IgG1; magenta; non‐CD20‐binding Fab arms of IgG1; green = hTfR1 cross‐Fab; orange = murine‐TfR1 cross‐Fab. (C) FACS analysis of Z‐138 cell binding with Brainshuttle‐CD20 and control mAbs. (D) Direct B cell killing induced by binding of Brainshuttle‐CD20 and control mAbs to Z‐138 cells measured by annV assay and PI uptake. Results expressed as uptake of IgG normalized to total cell protein. (E) hTFR1‐mediated intracellular uptake of Brainshuttle‐CD20 and control mAbs into MDCKII‐hTfR1 and MDCKII‐parental cells (non‐hTfR1 expressing), normalized to total cell protein. No detectable uptake of either compound by MDCKII‐parental cells was observed, confirming that mAb uptake in hTfR1‐transfected cells was mediated by hTfR1. (F) Transcytosis efficiency derived from the mean amount of IgG transcytosed versus the mean amount of IgG in the corresponding intracellular compartment at the beginning (T = 0) of the chase (expressed as a percentage). Data represents the mean ± SD of triplicate measurements in (C–F). annV/PI, annexin V/propidium iodide; CSF, cerebrospinal fluid; FACS, fluorescence‐activated cell sorting; hTfR1, human transferrin receptor 1; huCD20, humanized CD20; huCD20xHIGR3, C57BL/6 huCD20xC57BL/6‐Tg(hIg‐γ1,κ,λ)ait mice; IgG, immunoglobulin G; MDCKII, Madin–Darby canine kidney II; MFI, mean fluorescence intensity; PBMCs, peripheral blood mononuclear cells; PGLALA, P329GLALA mutation (Fc‐silent); PK / PD, pharmacokinetic / pharmacodynamic; SD, standard deviation; WT, wild‐type (Fc‐competent).

FIGURE 2

Kinetics of B cell depletion in blood and spleen. Kinetics of B cell depletion in huCD20 murine blood following administration of Brainshuttle(8D3)‐CD20 (A) and parental (non‐shuttled) (B) mAbs. B cell depletion is statistically significant from baseline with both molecules at later time points, and there is a significant difference between PGLALA and WT in both Brainshuttle(8D3) and parental molecules. Black asterisks represent statistically significant differences between molecules; coloured asterisks represent statistically significant differences to baseline. Kinetics of B cell depletion of Brainshuttle(8D3)‐CD20 mAbs at Day 6 in the spleen (C) and lymph node (D) of huCD20 and naïve mice. Only B cell numbers as a percentage of the CD45+ population are shown. Differences in the means of B cell depletion in the spleen between Brainshuttle(8D3)‐CD20‐WT and Brainshuttle(8D3)‐CD20‐PGLALA‐treated versus naïve mice were assessed by ANOVA. Significant differences in the treated groups versus naïve mice were seen at all dose groups. B cell depletion was significantly higher in Brainshuttle(8D3)‐CD20‐WT versus Brainshuttle(8D3)‐CD20‐PGLALA by two‐tailed paired t‐test at all dose groups. Data represents mean ± SD of n = 6 huCD20 mice. Data for the six naïve mice were used for comparison across all dose groups. ANOVA, one‐way analysis of variance; huCD20, humanized CD20; huCD20xHIGR3, C57BL/6 huCD20xC57BL/6‐Tg(hIg‐γ1,κ,λ)ait mice; mAb, monoclonal antibody; PGLALA, P329GLALA mutation (Fc‐silent); WT, wild‐type (Fc‐competent). *p‐values < .05.

FIGURE 3

Depletion of B cell subpopulations in lymph nodes following NP‐OVA immunization and administration of Brainshuttle(8D3)‐CD20 mAbs. (A) Schematic representation of immunization and treatment of huCD20xHIGR3 mice. Depletion of either B220+ total lymph node B cells (B) or NP‐specific GC B cells (C) was assessed on Day 11 by flow cytometry. Depletion efficiency was determined by comparing B cell counts between either Brainshuttle(8D3)‐CD20‐WT or Brainshuttle(8D3)‐CD20‐PGLALA with the vehicle‐only group (immunized/vehicle treatment) and assessed by ANOVA. A t‐test determined statistical differences between Brainshuttle(8D3) mAbs. Data represents mean ± SEM B cell counts of n = 5 huCD20xHIGR3 mice per group. Mean data values are shown in each bar. ANOVA, one‐way analysis of variance; GC, germinal center; huCD20xHIGR3, C57BL/6 huCD20xC57BL/6‐Tg(hIg‐γ1,κ,λ)ait mice; mAb, monoclonal antibody; NP, 4‐hydroxy‐3‐nitrophenylacetyl; OVA, ovalbumin; PGLALA, P329GLALA mutation (Fc‐silent); SEM, standard error of the mean; WT, wild‐type (Fc‐competent). *p‐value < .05.

FIGURE 4

Cytokine release is associated with Fc‐competence. Cytokine concentrations in human WBAs after 24 h incubation of healthy human blood with non‐shuttled and shuttled mAbs (A–E). (F) Flow cytometry of CD19+ B cells in human WBAs. Data represents the mean +SD of n = 6 healthy blood donors. Mean data values for the 1000 nM compound concentrations are displayed. IFN‐γ, interferon‐gamma; IL‐1β, interleukin‐1 beta; IL‐6, interleukin‐6; IL‐8, interleukin‐8; PGLALA, P329GLALA mutation (Fc‐silent); SD, standard deviation; TNF‐α, tumour necrosis factor‐alpha; WBAs, whole blood assays; WT, wild‐type (Fc‐competent).

FIGURE 5

Effect of Fc silencing on body temperature and cytokine levels in huCD20xHIGR3 mice. (A) Effects of Brainshuttle(8D3)‐CD20 mAbs on body temperature in huCD20xHIGR3 transgenic mice following IV administration at 10 mg/kg compared with non‐shuttled CD20‐WT and administration of vehicle only. (B) Cytokine levels measured 2 h after administration of mAbs. Data points represent mean ± SD of n = 5 huCD20xHIGR3 mice per group. Corresponding data for huCD20 mice is shown in Figure S3. G‐CSF, granulocyte colony‐stimulating factor; huCD20, humanized CD20; huCD20xHIGR3, C57BL/6 huCD20xC57BL/6‐Tg(hIg‐γ1,κ,λ)ait mice; IFN‐γ, interferon‐gamma; IL, interleukin; IV, intravenous; KC, keratinocyte‐derived cytokine; mAbs, monoclonal antibodies; MCP, monocyte chemoattractant protein; MIP, macrophage inflammatory protein; PGLALA, P329GLALA mutation (Fc‐silent); SD, standard deviation; TNF‐α, tumour necrosis factor‐alpha; WT, wild‐type (Fc‐competent).

FIGURE 6

B cell depletion in response to treatment with Brainshuttle(8D3)‐CD20 mAbs by FACS analysis. The frequency of remaining B220+ B cells in blood was calculated by setting the percentage of B220+ cells at the baseline measurement to 100%. Data represents mean ± SD (at baseline and Day 3 n = 15, Day 8 n = 10, Day 22 n = 5 huCD20xHIGR3 mice per group [the reduced number of mice at each time point was due to the scheduled sacrifice of n = 5 mice per time point]). FACS, fluorescence‐activated cell sorting; huCD20xHIGR3, C57BL/6 huCD20xC57BL/6‐Tg(hIg‐γ1,κ,λ)ait mice; mAbs, monoclonal antibodies; PGLALA, P329GLALA mutation (Fc‐silent); SD, standard deviation; WT, wild‐type (Fc‐competent). *p‐values < .05. Black asterisks represent statistically significant differences between molecules; coloured asterisks represent statistically significant differences to baseline.

FIGURE 7

Comparable penetration of Brainshuttle(8D3)‐CD20 mAbs in the mouse brain. (A) PK parameters of Brainshuttle(8D3)‐CD20 mAbs in huCD20xHIGR3 mouse serum following IV administration at 13.3 mg/kg, derived from mean concentration versus time data for n = 12 huCD20xHIGR3 mice. (B) Human IgG immunohistochemical (brown) staining of the cerebral cortex in the brain of huCD20xHIGR3 mice treated with Brainshuttle(8D3)‐CD20 mAbs over 24, 48, and 168 h. Arrows indicate endothelial staining. Blue = hematoxylin counterstain. Scale bars = 50 µm. huCD20xHIGR3, C57BL/6 huCD20xC57BL/6‐Tg(hIg‐γ1,κ,λ)ait mice; IgG, immunoglobulin G; IV, intravenous; mAbs, monoclonal antibodies; PGLALA, P329GLALA mutation (Fc‐silent); PK, pharmacokinetic; WT, wild‐type (Fc‐competent).

FIGURE 8

B cell depletion following Brainshuttle‐CD20 administration in cynomolgus macaques. (A) PK parameters of both Brainshuttle‐CD20 mAbs from two independent PK studies. (B) B cell numbers following Brainshuttle‐CD20 mAb administration are presented as a percentage of B cells measured at baseline. Data represents mean ± SD of n = 4 animals. mAbs, monoclonal antibodies; PGLALA, P329GLALA mutation (Fc‐silent); PK, pharmacokinetic; SD, standard deviation; WT, wild‐type (Fc‐competent).