Characterisation of a Novel Anti-CD52 Antibody with Improved Efficacy and Reduced Immunogenicity

Abstract

Anti-CD52 therapy has been shown to be effective in the treatment of a number of B cell malignancies, hematopoietic disorders and autoimmune diseases (including rheumatoid arthritis and multiple sclerosis); however the current standard of treatment, the humanized monoclonal antibody alemtuzumab, is associated with the development of anti-drug antibodies in a high proportion of patients. In order to address this problem, we have identified a novel murine anti-CD52 antibody which has been humanized using a process that avoids the inclusion within the variable domains of non-human germline MHC class II binding peptides and known CD4+ T cell epitopes, thus reducing its potential for immunogenicity in the clinic. The resultant humanized antibody, ANT1034, was shown to have superior binding to CD52 expressing cells than alemtuzumab and was more effective at directing both antibody dependent and complement dependent cell cytotoxicity. Furthermore, when in the presence of a cross-linking antibody, ANT1034 was more effective at directly inducing apoptosis than alemtuzumab. ANT1034 also showed superior activity in a SCID mouse/human CD52 tumour xenograft model where a single 1 mg/Kg dose of ANT1034 led to increased mouse survival compared to a 10 mg/Kg dose of alemtuzumab. Finally, ANT1034 was compared to alemtuzumab in in vitro T cell assays in order to evaluate its potential to stimulate proliferation of T cells in peripheral blood mononuclear cells derived from a panel of human donors: whereas alemtuzumab stimulated proliferation in a high proportion of the donor cohort, ANT1034 did not stimulate proliferation in any of the donors. Therefore we have developed a candidate therapeutic humanized antibody, ANT1034, that may have the potential to be more efficacious and less immunogenic than the current standard anti-CD52 therapy.

Fig 1. Selection of high CD52-expressing cell lines.

Raji, REH and CD52-NS0 cells were assessed by flow cytometry for CD52 expression and high CD52 expressing (⁺⁺) clones were identified for the three cell lines. Cells (either parent or ⁺⁺ lines) were stained with irrelevant human IgG (a) or with alemtuzumab (b) and (c).

Fig 2. Flow cytometry analysis of hybridomas or irrelevant mouse IgG.

Selected hybridoma supernatants or media spiked with irrelevant mouse IgG were assessed for binding to either untransfected or NS0-CD52⁺⁺ cells by flow cytometry to identify CD52 binding antibodies.

Fig 3. Amino acid sequence of murine anti-CD52 antibody 2E8 and lead humanized variant.

Amino acid sequences of the murine antibody 2E8 and lead humanized variant (ANT1034 –composed of VH3 and VK4 variable domains). Sourced human sequence segments together with their GenBank accession numbers are indicated below the amino acid sequences. (a) Heavy chain variable domain; (b) Light chain variable domain. Differences between the murine and humanized sequence are shaded. CDRs (shown in orange text) and numbering are as described by Kabat [41]. Sequences are available from GenBank: 2E8_VH—KP877884; 2E8_VK—KP877885; ANT1034_VH—KP877886; ANT1034_VK—KP877887.

Fig 4. Characterisation of anti-CD52 antibodies alemtuzumab and chimeric 2E8 IgG.

Antibodies were characterised by: (a) Direct binding of antibodies to HUT78 cells using flow cytometry; (b) Competition against alemtuzumab (with murine constant regions) for binding to CD52 on the surface of HUT78 cells using flow cytometry; (c) ADCC using REH⁺⁺ target cells (average of 4 PBMC donors for effector cells).

Fig 5. Peptide competition ELISA of selected humanized variants.

Titrations of humanized variants were competed against a fixed concentration (35 ng/ml) of biotinylated murine 2E8 for binding to CD52 peptide that was coated directly on an ELISA plate. Binding was detected with streptavidin-HRP and TMB substrate. Antibodies are named in the format ANT10X₁X₂ where X₁ refers to the VH variant (VH1 to VH5) and X₂ refers to the Vκ variant (Vκ1 to Vκ4).

Fig 6. Characterisation of anti-CD52 humanized antibodies, alemtuzumab and 2E8 chimeric IgG.

Selected humanized antibodies were characterised by: (a) Direct binding of antibodies to REH⁺⁺ cells using flow cytometry; (b) Competition binding against alemtuzumab (with murine constant regions) for binding to CD52 on the surface of REH⁺⁺ cells using flow cytometry; (c) ADCC using REH⁺⁺ target cells (average of 4 PBMC donors) with PBMC effector cells at a target:effector ratio of 50:1, and; (d) CDC using REH⁺⁺ target cells and normal human serum as a source of complement. For clarity, error bars are included in (c) and (d) for alemtuzumab and the lead humanized antibody, ANT1034, only.

Fig 7. Flow cytometry analysis of REH⁺⁺ cells incubated with selected humanized antibodies and alemtuzumab.

Cells were stained with Annexin V as a marker of apoptosis (gate = A) and Propidium Iodide as a marker of necrosis (gate = N). (a) Analysis in the absence of a cross-linking antibody; (b) Analysis with a cross-linking antibody. The percentage of cells in each gate is indicated.

Fig 8. Kaplan-Meier survival curves for the in vivo assessment of ANT1034 and alemtuzumab.

Antibodies were assessed in the Raji human Burkitt lymphoma SCID mouse model. Antibodies were administered on alternate days for seven doses starting 3 days post-injection of tumour cells. Efficacy was determined by comparing the median times to endpoint (either death, or euthanasia for advanced tumour progression).

Fig 9. CD4+ Episcreen™ T cell proliferation responses to ANT1034, alemtuzumab and chimeric 2E8.

CD4⁺ T cells were incubated with autologous mature DC loaded with the samples and assessed for proliferation after 7 days incubation. T cell responses with an SI ≥1.90 (indicated by red dotted line) that were significant (p <0.05) using an unpaired, two sample Student’s t-test were considered positive.