It is believed that stability predictions based on the Arrhenius relationship are inappropriate for protein preparations which exhibit complex degradation mechanisms. The degradation mechanisms of some protein drugs have been reported to vary as a function of temperature.' Extrapolation of stability data obtained with such protein drugs to lower temperatures should be limited to the temperature range over which the same degradation pathway is operative. Further, it has been suggested that Arrhenius approach should be conducted in the temperature range over which protein drugs are not susceptible to unfolding, usually below 40 "C.2 However, there have been too few papers concerning degradation kinetics and Arrhenius behavior of protein drugs to determine the valid temperature range for stability studies.
Several papers have described the Arrhenius plots for chemical degradation pathways such as deamidation,3p4 hydrolysis,5,6 and racemi~ation~ of peptides. The Arrhenius plots for inactivation of some proteins during storage have also been reported, even though the inactivation mechanisms were unknown. Inactivation during storage of dry horse serum cholinesterase* and lyophilized human interferon-@g exhibited linear Arrhenius plots in the ranges of 38-140 "C and of 25-80 "C, respectively. Linear Arrhenius plots were also observed for the solid-state inactivation of digestive enzymes such as Rhizopus lipase (45-60 "C) and pancreatic lipase and a-amylase (40-55 "C). The activation energies were calculated from the slopes of the lines to be 23-58 kcal/mol."J Inactivation of lyophilized urokinase showed linear Arrhenius plots from 30 to 50 "C and gave an activation energy of 15 kcalfmol.1'
The present study was carried out to obtain further information on the Arrhenius behavior of protein drug degradation. Commercial preparations of enzymes for medicinal use (a- chymotrypsin troche and tablet, bromelain tablet, kallikrein capsule, and P-galactosidase powder) were used as model protein preparations, and the inactivation during storage of these preparations was studied as a function of temperature.
Stability Studies-Commercial a-chymotrypsin troch, a-chymotrypsin tablet, bromelain tablets, kallikrein capsule, and @-galactosidase powder were purchased from the manufacturers and stored under degradation-accelerating conditions [40-70 "C, 50 or 75% relative humidity (RH)]. The outer packages of these preparations were removed, leaving the formulation enclosed in the primary package. Relative humiditywas adjusted with NaBr (50% RH) or NaCl (75% RH) saturated solutions.
Preparations were also stored at 25 "C and 50 or 75% RH for long-term stability studies. Samples were taken from the stored preparations at appropriate intervals for assay of activity.
a-Chymotrypsin was extracted by crushing troche and tablet preparations in adequate volumes of 0.001 N HCl solution. The coating film of the tablets had been removed with phosphate buffer solution (pH 7.0, 50 mM) prior to the extraction. The a-chymotrypsin solution was centrifuged and the supernate was diluted with 0.001 N HCl to yield an activity unit of about 1.5 pmol N-acetyl-L-tyrosin ethyl ester (ATEE, Aldrich, Milwaukee, WI)/min/mL. The activity of the sample solution was determined by the USP method,12 using ATEE.
Bromelain was extracted with a pH 4.5 solution containing 5.27 mg/mL cystein, 2.23 mg/mL EDTA, and 23.4 mg/mL NaC1,