Formulation Excipients and Their Role in Insulin Stability and Association State in Formulation

Abstract

While excipients are often overlooked as the "inactive" ingredients in pharmaceutical formulations, they often play a critical role in protein stability and absorption kinetics. Recent work has identified an ultrafast absorbing insulin formulation that is the result of excipient modifications. Specifically, the insulin monomer can be isolated by replacing zinc and the phenolic preservative metacresol with phenoxyethanol as an antimicrobial agent and an amphiphilic acrylamide copolymer excipient for stability. A greater understanding is needed of the interplay between excipients, insulin association state, and stability in order to optimize this formulation. Here, we formulated insulin with different preservatives and stabilizing excipient concentrations using both insulin lispro and regular human insulin and assessed the insulin association states using analytical ultracentrifugation as well as formulation stability. We determined that phenoxyethanol is required to eliminate hexamers and promote a high monomer content even in a zinc-free lispro formulation. There is also a concentration dependent relationship between the concentration of polyacrylamide-based copolymer excipient and insulin stability, where a concentration greater than 0.1 g/mL copolymer is required for a mostly monomeric zinc-free lispro formulation to achieve stability exceeding that of Humalog in a stressed aging assay. Further, we determined that under the formulation conditions tested zinc-free regular human insulin remains primarily hexameric and is not at this time a promising candidate for rapid-acting formulations.

Keywords: aggregation; diabetes; excipient; insulin; polymer.

Fig. 1

Schematic of insulin association states, preservatives and pharmacokinetics. (a) Schematic showing how preservatives can promote different association states. The insulin hexamer is stable, while the insulin monomer is susceptible to rapid aggregation and amyloid fibril formation. Formulations containing high monomer counts require additional stabilizing agents. (b) Schematic showing how insulin pharmacokinetics can be shifted by altering insulin association state. The insulin monomer is more rapidly absorbed into the blood. In contrast, subcutaneous administration of the insulin hexamer results in a delayed peak and longer duration of action.

Fig. 2

Insulin lispro association states. Using analytical ultracentrifugation, sedimentation coefficients were used to estimate the ratios of insulin association states when formulated with different antimicrobial preservatives in the absence of zinc. Commercial Humalog is shown as a control. It should be noted that higher ordered structures as reported here is thought to be reversible aggregates suchas octamers, rather than high molecular weight polymer or amyloid fibril species.

Fig. 3

RHI association states. Using analytical ultracentrifugation, sedimentation coefficients were used to estimate the ratios of insulin association states when formulated with different antimicrobial preservatives in the absence of zinc. Commercial Humalog is shown asa control. It should be noted that higher ordered structures as reported here is thought to be reversible aggregates suchas octamers, rather than high molecular weight polymer or amyloid fibril species.

Fig. 4

Insulin lispro formulation stability. (a) Change in transmittance traces and (b) Time to aggregation of formulations with different antimicrobial preservatives: Lispro-1 (metacresol), Lispro-2 (metacresol + phenoxyethanol), Lispro-3 (phenoxyethanol), Lispro-4 (methylparaben + propylparaben). (c) Change in transmittance traces and (d) Time to aggregation of phenoxyethanol formulation (Lispro-3) with different concentrations of MoNi polymer excipient. Comparison of stability by aggregation times (t_A), defined as the time to a change in transmittance (λ = 540 nm) of 10% or greater following stressed aging (i.e., continuous agitation at 37°C). Data shown are average transmittance traces for n = 3 samples per group and error bars are standard error mean. Each ☆ represents a sample that did not aggregate before the assay ended at 150 h. A one-way ANOVA with a Tukey–Kramer correction for multiple comparisons was performed in GraphPad Prism 9, and each formulation was compared to all other formulations. Formulations connected by the same letter label are not significantly different. Adjusted p values are listed in the supplemental information Table S5–6. α < 0.05.

Fig. 5

RHI formulation stability. A) Change in transmittance traces and B) Time to aggregation of formulations with different antimicrobial preservatives: RHI - 1 (metacresol), RHI - 2 (metacresol + phenoxyethanol), RHI - 3 (phenoxyethanol), RHI - 4 (methylparaben + propylparaben). C) Change in transmittance traces and D) Time to aggregation of phenoxyethanol formulation (RHI - 3) with different concentrations of MoNi polymer excipient. Comparison of stability by aggregation times (t_A), defined as the time to a change in transmittance (λ = 540 nm) of 10% or greater following stressed aging (i.e., continuous agitation at 37°C). Data shown are average transmittance traces for n = 3 samples per group and error bars are standard error mean. Each ☆ represents a sample that did not aggregate before the assay ended at 150 hours. A one-way ANOVA with a Tukey-Kramer correction for multiple comparisons was performed in GraphPad Prism 9, and each formulation was compared to all other formulations. Formulations connected by the same letter label are not significantly different. Adjusted p values are listed in the supplemental information Table S7–8. α < 0.05.