Preparation of experimental varicella (Oka/Merck) viral bulks
Culture flasks with VZV (Oka/Merck)-infected MRC5 cells were obtained from Merck Manufacturing Division (MMD, West Point, PA). The VZV (Oka/Merck) containing MRC5 cells were harvested into two formulations prepared with either a hydrolyzed porcine gelatin (SOL-U-PRO; Dynagel Inc., IL), or 8.5 kD recombinant human gelatin (FG-5001; Lot # 04AE001, FibroGen, Inc., CA), and further harvested in a small-scale process closely mimicking current manufacturing procedure for VZV (Oka/Merck) bulk preparation. Both processed bulks were aliquoted, placed in a liquid nitrogen batch freezer (Kwik-Freeze Freezing System, AIRCO, NJ, USA), frozen and transferred to -70°C. A set of small frozen liquid sample aliquots (1.0 mL) was submitted for VZV plaque assay analysis to determine VZV (Oka/Merck) potency changes during bulk processing for both, SOL-U-PRO- and FG-5001-containing, varicella bulks.
Preparation of experimental, refrigerator-stable varicella vaccine samples
Varicella virus vaccine live (Oka/Merck) is a lyophilized preparation. When this refrigerator-stable vaccine is reconstituted as directed, each 0.5 mL dose contains the following: a minimum of 1350 plaque forming units (PFU) of Oka/Merck varicella virus, approximately 18 mg of sucrose, 8.9 mg of hydrolyzed gelatin, 3.6 mg of urea, 2.3 mg of sodium chloride, 0.36 mg of monosodium L-glutamate, 0.33 mg of sodium phosphate basic, 57 mcg of potassium phosphate monobasic, 57 mcg of potassium chloride. The product also contains residual components of MRC-5 cells including DNA, protein and trace quantities of neomycin and bovine calf serum from MRC-5 culture media. The product contains no preservative .
Experimental, refrigerator-stable varicella vaccine samples containing either the porcine hydrolyzed gelatin, or recombinant human gelatin, were prepared in a small-scale procedure closely mimicking current manufacturing process for varicella vaccine. Briefly, aliquots (40 mL) of SOL-U-PRO- and FG-5001-stabilized VZV (Oka/Merck) bulks were quickly thawed in water bath (30°C), and then diluted into their respective gelatin-containing formulations to a target potency of ≈ 4.4 log10 pfu/mL. Final formulated bulk (FFB) aliquots (0.7 mL) of both experimental vaccines were filled in glass vials, partially stoppered, placed in a liquid nitrogen batch freezer and frozen. These frozen FFB samples were divided into two groups. The first group was transferred to a -70°C freezer and was later used as a control to determine the VZV (Oka/Merck) potency change after lyophilization. The second group of samples was transported to the lyophilization chamber (Usifroid Lyophilizer Model SMH 101, Usifroid SA, France), and lyophilized. After lyophilization, the vaccine vials were inspected, sealed and placed in stability stations.
Short term and long-term vaccine stability study under accelerated and real-time conditions
In addition to storage at -15°C and 2–8°C to examine real-time conditions, the stability stations used for storage were tempered at 15°C and 37°C to examine the vaccine potency stability under accelerated conditions. At pre-determined time points (37°C for 7 days; 15°C for 3, 6, 9, 12 months; 2–8°C and 15°C for 3, 6, 9, 12, 24 months) vaccine vials were removed from stability stations and stored at -70°C until submission for VZV (Oka/Merck) potency analysis. Sample vaccine vials from individual time points were analyzed together with their respective control samples which had been stored at -70°C. In addition, three sample vials from each time point were also submitted for moisture analysis from the long-term stability study executed at 2–8°C for 24 months.
VZV (Oka/Merck) plaque assay analysis
VZV (Oka/Merck) potency in both viral bulk preparations and experimental vaccine samples were determined by VZV plaque assay with liquid overlay medium . Analyzed samples (thawed liquid bulk samples and FFB liquid vaccine samples, as well as reconstituted lyophilized vaccine samples) were diluted with the stabilizer and submitted for analysis in 1 × 12 assay format (one sample in each of 12 independent assay runs). VZV (Oka/Merck) potency was defined as a log10 of VZV plaque forming units (PFU) per mL.
Moisture content analysis of lyophilized vaccine samples
The amount of moisture in the lyophilized vaccine samples was determined by the Karl Fischer coulometric titration method  using an Aquatest™ coulometric moisture titration system (Photovolt Instruments, Inc., Minneapolis, MN) according to the manufacturer's instructions. For each analysis, average moisture content (%) was calculated based on valid results from three tested vaccine samples.
The potency losses associated with lyophilization were calculated as the average of the differences observed between the liquid samples and the -70°C (lyophilized) samples tested in the same 12 assay runs. The standard error of the loss estimate was simply the standard deviation of the observed differences divided by the square root of the number of runs in which a difference was calculated. The same calculations were performed with the stability data for 37°C for one week. Within run differences between the -70°C (lyophilized) samples and the 37°C samples were determined and averaged across 12 independent runs. The data generated for long term stability estimation consisted of concurrent testing of "incubated" samples (those stored at -15°C, 2–8°C, and 15°C for pre-specified interval) along with control samples from the same lot which were stored only at -70°C. For each stability interval, 12 incubated vials and 12 control vials were tested, one vial each, in 12 independent assay runs. The potency loss at that interval was calculated as the average difference between the control and incubated sample within each run. This format helps to minimize the potential for run-to-run differences in the assay affecting the stability estimation. For each of the two formulations, linear regression analysis was performed using the individual loss estimates at each long term storage temperature.