Rational design of viscosity reducing mutants of a monoclonal antibody: hydrophobic versus electrostatic inter-molecular interactions

Abstract

High viscosity of monoclonal antibody formulations at concentrations ≥100 mg/mL can impede their development as products suitable for subcutaneous delivery. The effects of hydrophobic and electrostatic intermolecular interactions on the solution behavior of MAB 1, which becomes unacceptably viscous at high concentrations, was studied by testing 5 single point mutants. The mutations were designed to reduce viscosity by disrupting either an aggregation prone region (APR), which also participates in 2 hydrophobic surface patches, or a negatively charged surface patch in the variable region. The disruption of an APR that lies at the interface of light and heavy chain variable domains, VH and VL, via L45K mutation destabilized MAB 1 and abolished antigen binding. However, mutation at the preceding residue (V44K), which also lies in the same APR, increased apparent solubility and reduced viscosity of MAB 1 without sacrificing antigen binding or thermal stability. Neutralizing the negatively charged surface patch (E59Y) also increased apparent solubility and reduced viscosity of MAB 1, but charge reversal at the same position (E59K/R) caused destabilization, decreased solubility and led to difficulties in sample manipulation that precluded their viscosity measurements at high concentrations. Both V44K and E59Y mutations showed similar increase in apparent solubility. However, the viscosity profile of E59Y was considerably better than that of the V44K, providing evidence that inter-molecular interactions in MAB 1 are electrostatically driven. In conclusion, neutralizing negatively charged surface patches may be more beneficial toward reducing viscosity of highly concentrated antibody solutions than charge reversal or aggregation prone motif disruption.

Keywords: APR, Aggregation Prone Region; ASA, Accessible Surface Area; ASAFv-HPH, hydrophilic accessible surface area of the Fv portion; ASAFv-HYD, hydrophobic accessible surface area of the Fv portion; CE, Capillary Electrophoresis; CH2; CH3, third constant domain in heavy chain; CHO, Chinese Hamster Ovary; D0, diffusion coefficient at infinite dilution; DFv, dipole moment of Fv; DLS, Dynamic Light Scattering; ELISA, Enzyme-Linked Immunosorbent Assay; Fab, fragment antigen binding; Fc, fragment crystallizable; Fv, fragment variable; HC, heavy chain; IgG, immunoglobulin G; LC, light chain; MAB 1 Control, MAB 1 expressed in CHO cells; MD, molecular dynamics; NTU, Nephelometric Turbidity Unit; PEG, polyethylene glycol; Pagg-VH, aggregation propensity of VH domain; Pagg-VL, aggregation propensity of VL domain; RPM, revolutions per minute; SE-HPLC, Size Exclusion High Performance Liquid Chromatography; Tm, thermal transition temperature; VH, variable domain in the heavy chain; VL, variable domain in the light chain; ZDHH, Debye-Huckel Henry Charge; ZFv, net charge of the Fv; ZFv-app, apparent charge of the Fv; aggregation prone regions; cIEF, capillary Isoelectric Focusing; cP, centipoise; high concentration; kD, protein-protein interaction parameter; mAb, monoclonal antibody; molecular modeling; monoclonal antibodies; negatively charged patches; rational design; second constant domain in the heavy chain; solubility; viscosity; ΔGFv, change in Free energy of Fv; η, solution viscosity; η0, solvent viscosity; ηrel, relative viscosity; ξFv, zeta-potential of the Fv.

Figure 1.

(A) A ribbon diagram showing the schematic structure of Fv portion of MAB 1. V_H (top) and V_L domains (bottom) are shown in dark green and cyan colored ribbons, respectively. Heavy chain CDRs 1 and 2 are colored brown while CDR 3 is colored red. All light chain CDRs are shown in magenta. Aggregation prone regions (APRs) predicted by a TANGO / PAGE combination are shown in yellow. Light chain residues at the sites selected for point mutations are shown in ball and stick. Note that V44 and L45 in light chain lie in an APR at V_H: V_L interface. Light chain E59 on the other hand is situated away from the domain interface. This model is oriented such that the antigen binding site is seen while looking down the plane of paper. (B) APRs spectrum for V_H region is shown by plotting TANGO and PAGE predictions simultaneously with respect to the residue number. The procedure followed here is the same as the one described earlier by Wang et al. The Z-score computed from average and standard deviation values of PAGE aggregation propensity (lnP) score is plotted in blue color while TANGO predicted% aggregation is plotted in green. The 2 horizontal red lines indicate the cut-off values for Z-score (1.96, upper red line) and% aggregation (10%, lower red line) used. (C) APR spectrum for the V_L region is shown. This plot is created in the same way as (B). Note that the second APR in the light chain (44-VLVIY-48) is the strongest one and was targeted for disruption at sites V44 and L45. (D) Solvent exposed hydrophobic (green), positively charged (blue) and negatively charged (red) patches on the surface of MAB 1 Fv. This picture is shown in the same orientation as (A).

Figure 2.

Observations on purity, thermal stability and concentrate-ability of MAB 1 and its variants are presented. (A) SE-HPLC profile overlays of MAB 1 Control and M1–M6. The SE-HPLC profile of M5 indicates presence of large amounts of high molecular mass species. (B) DSC thermogram overlays for MAB 1 Control and M1-M6. Again, it can be seen the M5 is significantly destabilized and shows 4 thermal transitions, while the other variants shown only 3. (C) Comparison of DSC thermal transition temperatures, T_m1, T_m2 and T_m3, for MAB 1 Control (denoted as MAB 1 in the plot), M1-M4 and M6. MAB 1 is MAB 1 Control, the parent MAB 1 produced at Pfizer using CHO cells. M1 is the parent MAB 1 produced at Syngene using transient transfection. The variants, M2, M3, M4 and M6 are single point variants of MAB 1, again, produced at Syngene using transient transection in HEK cell lines. (D) Appearance of M1 and its variants at high concentrations at 2–8°C and 25°C. All materials were at ∼130 mg/mL except for M2 and M4.

Figure 3.

DLS profiles the parent mAb, M1, and variants M2, M3, M4 and M6 are shown along with MAB 1 Control. The X-axis shows concentration of antibody solutions, c (mg/mL), and the Y-axis shows measured diffusion co-efficient, D (cm²/s). Interaction parameters computed from the plots indicate considerable decrease in self-associative behavior for M3 (E59Y).

Figure 4.

Experimentally measured concentration dependent viscosity curves for MAB 1 and its variants (M1, M2, M3, and M6) plotted along with MAB 1 Control (indicated as MAB 1). X-axis indicates concentration of antibody solutions, c (mg/mL). Y-axis shows the solution viscosity, η (cP). The horizontal red line indicates a viscosity of 20 cP. As a general rule, viscosity below 20 cP is desired for highly concentrated antibody solutions to be delivered subcutaneously. It can be seen that viscosity behaviors of M3 and M6 are improved compared to M1.

Figure 5.

Five nanoseconds-long molecular dynamics (MD) simulations were performed on the Fv portions of (A) M1, (B) M2, (C) M4 and (D) M3 at elevated temperatures (400 K). In each plot, X-axis shows the simulation time (ns) and Y-axis indicates distances (Å) between specific atoms of interacting residues in the structural context for position 59 in the light chain of MAB 1. The side chain and backbone distances are shown for interactions formed by the residue at position 59 and I57 with the side chain functional group of R53. The red line shows the distance between central atom in side chain of residue at position 59 and the C^ξ atom in the side chain of R53. The green line indicates the distance of the backbone carbonyl atom of the residue at position 59 with the C^ξ atom in the side chain of R53. Similarly, the blue line indicates the distance of the backbone carbonyl atom of I57 from the C^ξ atom in the side chain of R53. When the distance between carbon atoms is below 6 Å, an interaction is considered to be formed (black horizontal line). This figure shows that side chain guanidium group of R53 engages in promiscuous salt bridging interactions with the back bone carbonyl of I57, and the backbone carbonyl as well as side chain carboxylate of E59 in the parent mAb, M1. Upon charge reversal at position 59 (M2 and M4 variants), the electrostatic repulsions between the side chains of the residue at position 59 and R53 weaken this electrostatic network leading to destabilization of the Fab as evidenced by decreased T_m2 values for M2 and M4. However, the charge neutralization variant, M3, restores the network among these 3 residues by forming tyrosine ring π interactions with the guanidium group of R53. Analogous simulations were also performed at 300 K. These are shown in the supplementary material (Fig. S2A–S2D).