Anticalin N- or C-Terminal on a Monoclonal Antibody Affects Both Production and In Vitro Functionality

Abstract

Bispecific antibodies (BsAbs) represent an important advance in innovative therapeutic strategies. Among the countless formats of BsAbs, fusion with molecules such as anticalins linked to a monoclonal antibody (mAb), represents an easy and low-cost way to obtain innovative molecules. We fused an anticalin against human fibronectin to a molecule biosimilar to trastuzumab (H0) or rituximab (R0), in four different positions, two on the N terminal region of heavy or light chains and two on the C terminal region. The eight BsAbs (H family (HF) 1 to 4 and R family (RF) 1 to 4) were produced and their affinity parameters and functional properties evaluated. The presence of anticalin did not change the glycosylation of the BsAb, shape or yield. The antigenic recognition of each BsAb family, Her2 for HF1 to 4 and CD20 for RF1 to 4, was slightly decreased (HF) or absent (RF) for the anticalin N-terminal in the light chain position. The anticalin recognition of FN was slightly decreased for the HF family, but a dramatic decrease was observed for RF members with lowest affinity for RF1. Moreover, functional properties of Abs, such as CD16 activation of NK, CD32-dependent phagocytosis and FcRn transcytosis, confirmed that this anticalin position leads to less efficient BsAbs, more so for RF than HF molecules. Nevertheless, all BsAbs demonstrated affinities for CD16, CD32 and FcRn, which suggests that more than affinity for FcRs is needed for a functioning antibody. Our strategy using anticalin and Abs allows for rapid generation of BsAbs, but as suggested by our results, some positions of anticalins on Abs result in less functionality.

Keywords: ADCC; CD16; CD32; CDC; Fc gamma receptors; FcRn; anticalin; bispecific antibody; monoclonal antibody.

Figure 1

(A) Cartoon representations of the 8 BsAbs. HF and RF families were constructed on trastuzumab (TTZ) and rituximax (RTX) commercial therapeutic mAbs and produced as biosimilars, as H0 and R0, respectively. Anticalin ACFn (yellow) was covalently linked to heavy (blue) or light (orange) chains by using Glycine-Serine linkers at N- and C-terminal positions. (B) SDS-PAGE of the mAbs and BsAbs in non-reducing and reducing conditions; ACFn is the anti-fibronectin anticalin. Green and yellow arrows indicate mAbs and BsAbs, respectively, in the non-reduced condition. In the reduced condition, heavy chains are indicted by blue or purple arrows and light chains by orange or red arrows for wild-type chains or chains linked to anticalin, respectively. (C) Size-exclusion chromatography on a calibrated Superdex 200 Increase 10/300 GL column of each antibody and the anticalin (right trastuzumab and BsAb, left rituximab and BsAb).

Figure 2

Mass spectra of H0 and R0 monoclonal Abs acquired under denaturing conditions. Insets correspond to the glycoforms identified.

Figure 3

Mass spectra of BsAbs of RF format.

Figure 3

Mass spectra of BsAbs of RF format.

Figure 4

Antigen recognition by BsAbs. Competition assays using (A) fluorescent trastuzumab (TTZ-^A488) or (B) rituxumab (RTX^-A488) on SK-Br-3 or Daudi cell lines, respectively. The mean fluorescent intensity (MFI) of TTZ^-A488 and RTX^-A488 were determined by flow cytometry in the presence of mAbs and BsAbs at different concentrations. Data are percentage TTZ^-A488 inhibition binding on SK-Br3 cells or RTX-^A488 inhibition binding on Daudi cells. RTX^-A488 or TTZ^-A488 alone was considered 100% binding.

Figure 5

BsAbs binding to CD16 (FcγRIIIA). (A) Heatmap representation of mAb and BsAb binding to CD16 evaluated by a flow cytometry competition assay of anti-CD16-FITC binding on NK92 hCD16V in the presence of mAbs or BsAbs for 30 min at 4 °C. Data are percentage inhibition binding of anti-CD16-FITC. (B) CD16-dependent BsAb functionality with NK92 hCD16V cells as effector cells and SK-Br-3 or Daudi cells as target cells in the presence of mAbs or BsAbs. NK92 cells were labelled with CD107a, CD16 and CD56 Abs, then intracellularly labelled for INF-γ. Data are percentage NK92 cells expressing CD107a and/or INF-γ. (One way Anova, *** p < 0.005).

Figure 6

BsAbs binding to CD32 (FcγRIIA). (A) Heatmap representation of BsAbs binding to CD32 evaluated by a flow cytometry competition assay of anti-CD32-FITC binding on THP1 cells in the presence of BsAbs. Data are percentage binding inhibition of anti-CD32-FITC. (B) Phagocytosis assays were used to evaluate CD32-dependent BsAb functionality. Data are percentage of fluorescent positive cells. Results are represented as described in Figure 5.

Figure 7

BsAbs binding to FcRn. (A) Heatmap representation of BsAbs binding to FcRn evaluated by flow cytometry in a competition assay of rituxumab^AF488 on a mix of Jurkat and JurkatΔFcRn cells (ratio 1:1) in the presence of mAbs or BsAbs at different concentrations. Data are percentage inhibition binding of rituxumab^AF488. (B) FcRn-dependent BsAb functionality was evaluated by a transcytosis assay with MDCKII cells expressing human FcRn in the presence of mbs or BsAbs. Rituxumab^-AF488 fluorescence in the compartment receiving transcytosed Rituxumab^-AF488 was measured by microplate reader and data are percentage inhibition of rituximab^AF488 transcytosis. Results are represented as described in Figure 5 and Figure 6.

Figure 8

Spider representation of BsAb properties. Molecules H0, R0 and ACFn were considered as 100% functionality represented by the dotted line.