Optimizing a therapeutic humanized follicle-stimulating hormone-blocking antibody formulation by protein thermal shift assay

Abstract

Biopharmaceutical products are formulated using several Food and Drug Administration (FDA) approved excipients within the inactive ingredient limit to maintain their storage stability and shelf life. Here, we have screened and optimized different sets of excipient combinations to yield a thermally stable formulation for the humanized follicle-stimulating hormone (FSH)-blocking antibody, MS-Hu6. We used a protein thermal shift assay in which rising temperatures resulted in the maximal unfolding of the protein at the melting temperature (T_m ). To determine the buffer and pH for a stable solution, four different buffers with a pH range from 3 to 8 were screened. This resulted in maximal T_m s at pH 5.62 for Fab in phosphate buffer and at pH 6.85 for Fc in histidine buffer. Upon testing a range of salt concentrations, MS-Hu6 was found to be more stable at lower concentrations, likely due to reduced hydrophobic effects. Molecular dynamics simulations revealed a higher root-mean-square deviation with 1 mM than with 100 mM salt, indicating enhanced stability, as noted experimentally. Among the stabilizers tested, Tween 20 was found to yield the highest T_m and reversed the salt effect. Among several polyols/sugars, trehalose and sucrose were found to produce higher thermal stabilities. Finally, binding of recombinant human FSH to MS-Hu6 in a final formulation (20 mM phosphate buffer, 1 mM NaCl, 0.001% w/v Tween 20, and 260 mM trehalose) resulted in a thermal shift (increase in T_m ) for the Fab, but expectedly not in the Fc domain. Given that we used a low dose of MS-Hu6 (1 μM), the next challenge would be to determine whether 100-fold higher, industry-standard concentrations are equally stable.

Keywords: FSH; Good Laboratory Practice (GLP); antibody development; biotherapeutics; colloidal stability.

Figure 1:

Screening of the different buffer and pHs to obtain stable solution of the MS-Hu6. (A) 20 mM citrate buffer (pH 2.9 to 6.1), 20 mM acetate buffer (pH 3.6 to 5.6), 20 mM phosphate buffer (pH 5.6 to 8.0), and 25 mM histidine buffer (pH 5.0 to 7.0). (B) Representation of the cumulative data on change in T_m with increasing pH in different buffers. (C) Maximum stability of the Fab and Fc domain; at pH 5.62 for Fab in phosphate buffer, and at pH 6.85 for Fc in histidine buffer. (D) Effect of phosphate buffer concentrations (10, 20 and 40 mM) on the stability of the Fab and Fc domain of the MS-Hu6. (E) Effect of 20 mM phosphate buffer on the stability of the Fab and Fc domain of the MS-Hu6.

Figure 2:

(A) Screening of the effect of the stabilizers (tween 20, 80 and poloxamer 188) on the stability of the Fab and Fc domain of the MS-Hu6 in the 20 mM phosphate buffer at a pH range (5.5 to 7). (B) Effect of the stabilizers on the stability of the Fab and Fc domain of the MS-Hu6 at pH 6.2.

Figure 3:

(A) Screening of the effect of the NaCl concentrations (150, 100, 50 and 1 mM) on the stability of the Fab and Fc domains of the MS-Hu6 in the 20 mM phosphate buffer at varied pHs (5.5 to 7). (B) Effect of different NaCl concentrations on the stability of the Fab and Fc domains of MS-Hu6 at pH 6.2. (C) Analysis of MS-Hu6 at 1 mM and 100 mM salt concentration using molecular dynamics simulations. The Cα-RMSDs were comparable and minor differences were observed between the two systems at both ionic concentrations. The Cα-RMSD of the two chains of the MS-Hu6 are colored red and black. (D) Ion distribution around the protein surface assessed using the radial distribution function (RDF). The minimal distance between the ions and the protein were calculated as a variation of time. RDF analysis for 100 mM salt simulation indicates that the ions are located closer to the complex surface within the cutoff distance, in accordance with the higher concentration of ions in the solution. (E) The charged ions localized around the protein surface have the ability to influence side chain interactions and alter electrostatic surface charge pattern.

Figure 4:

(A) Screening of the effect of the sugar type and concentrations (260 mM, each, sucrose, trehalose, mannitol, dextran 40, and sorbitol) on the stability of the Fab and Fc domains of the MS-Hu6 in the 20 mM phosphate buffer at varied pHs (5.5 to 7) with 1 mM NaCl, and 0.001 % w/v Tween 20. (B) Effect of the 260 mM concentrations on the stability of the Fab and Fc domains of the MS-Hu6 at pH 6.2, (C) Effect of the different concentration of trehalose and NaCl on the stability of the Fab and Fc domain of the MS-Hu6 at different pH values (5.5 to 7). (D) Effect of the different concentrations of trehalose and NaCl on the stability of the Fab and Fc domains of the MS-Hu6 at pH 6.2. (E) Effect of the different concentrations of the sucrose and NaCl on the stability of the Fab and Fc domain of the MS-Hu6 within a pH range (5.5 to 7). (F) Effect of the different concentrations of the sucrose and NaCl on the stability of the Fab and Fc domains of the MS-Hu6 at pH 6.2.

Figure 5:

Effect of binding of the FSH to the MS-Hu6 on the stability of the Fab and Fc domaina of the MS-Hu6 in (A) 20 mM phosphate buffer (pH 5.76 to 6.42), (B) 20 mM phosphate buffer, 1 mM NaCl, 0.001% w/v tween 20 and 260 mM trehalose (pH 5.76 to 6.42), and (C) 20 mM phosphate buffer, 1 mM NaCl, 0.001 % w/v tween 20 and 260 mM sucrose with (pH 5.76 to 6.42).

Figure 6:

(A) Representative Size Exclusion Chromatography (SEC) chromatographs of MS-Hu6 (1 mg/ml in PBS) and formulation (1 mg/ml) (B) Representative particle size volume distribution graphs of MS-Hu6 (1 mg/ml in PBS) and formulation (1 mg/ml).