Close relationships between the storage stability and molecular mobility of lyophilized pharmaceutical formulations, as demonstrated by recent studies, [1][2][3][4][5][6] suggest that molecular mobility may represent a useful measure of storage stability.Molecular motions that induce degradation of lyophilized glassy formulations containing polymer excipients are expected not to be localized motions of polymer side chains on small time scales, but rather molecular motions on larger scales involving global segmental motions of polymer backbones. Large-time-scale molecular motions have been determined by thermomechanical analysis and dynamic mechanical analysis at temperatures above glass transition temperature (T g ), 7) but such analyses are unsuitable for evaluation of motions at temperatures below T g . In contrast, NMR relaxation measurements, which have been used as a parameter of molecular mobility in food systems [8][9][10][11] and pharmaceutical formulations, 5,[12][13][14] can provide information regarding largetime-scale molecular motions that occur even in glassy polymer materials at temperatures below T g . Although NMR relaxation times reflect molecular motions on smaller time scales at temperatures substantially lower than T g , large-timescale molecular motions that begin to emerge with increasing temperature can be sensitively detected by NMR relaxation times. For lyophilized dextran, spin-lattice relaxation times of methine proton and carbon in the glucose ring were found to be indicative of the increased large-time-scale molecular motion associated with the beginning of glass transition. 13,14) The purpose of the present study is to examine whether NMR relaxation times can be used as a general indicator of large-time-scale molecular motions related to the storage stability of lyophilized formulations. Laboratory-and rotatingframe spin-lattice relaxation times (T 1 and T 1r ) of 1 H and 13 C in lyophilized poly(vinylpyrrolidone) (PVP) and methylcellulose (MC) are determined to discuss the feasibility of using T 1 and T 1r as a measure of large-time-scale molecular motions.
ExperimentalPreparation of Lyophilized PVP and MC PVP (K-30, 161-03105) was purchased from Wako Pure Chemical Ind. Ltd. (Osaka). MC was synthesized by methylation of cellulose (Merck) with 13 C-labeled methyl iodide (Aldrich) and purified with dialysis tube. Lyophilized PVP and MC were prepared from a D 2 O (Wako Pure Chemical Ind. Ltd.) solution of 2.5% w/w polymer. Freeze drying was carried out at a vacuum level below 5 Pa for 23.5 h in a lyophilizer (Freezevac C-1, Tozai Tsusho Co., Tokyo). The shelf temperature was between Ϫ35 and Ϫ30°C for the first 1 h, 20°C for the subsequent 19 h, and 30°C for the last 3.5 h.The water content of the lyophilized PVP and MC was adjusted by storing at 15°C for 24 h in a desiccator with a saturated D 2 O solution of LiCl D 2 O (12% relative humidity (RH)), potassium acetate (23% RH), K 2 CO 3 2D 2 O (43% RH), NaBr 2D 2 O (60% RH) or NaCl (75% RH).Determination of T 1 and T 1r r by 1 H Pulse...