Supercritical carbon dioxide was used as an antisolvent for producing recombinant human immunoglobulin G (rIgG) particulate powders. Liquid carbon dioxide (CO2) was premixed with ethanol to create a single‐phase, modified supercritical fluid (SCF). The modified SCF was then vigorously mixed with a pharmaceutically acceptable, aqueous formulation of rIgG, and the mixture was immediately atomized into a pressurized vessel where rapid expansion of the modified SCF extracted the aqueous phase, resulting in precipitation of the protein powder. The process was reproducible, and resulting powder products were characterized by their aqueous solubilities, and the spectroscopic profile, molecular integrity, and antigen binding activity of the individual soluble fractions. Molecular integrity was assessed via size‐exclusion high‐performance liquid chromatography (SEC), whereas antigen binding activity was determined using an enzyme‐linked immunosorbent assay (ELISA). Attempts to characterize particle size and morphology were confounded due to the extremely deliquescent nature of the powders, causing them to absorb moisture rapidly and become gummy. Operational conditions were optimized to a point which yielded powders that were completely soluble, and had ultraviolet (UV) spectroscopic and SEC profiles indistinguishable from those of the reference standard starting solution from which the powders were derived. Antigen binding activities of the powders, however, were ≤50% of expected levels, revealing the need for improvement of this SCF processing approach for rIgGs. From this study, two important conclusions were drawn: (1) Biopharmaceutical product quality assessments must include definitive activity analyses, because profound activity losses, such as those suffered here by the rIgGs, may not be revealed when using standard physicochemical analyses alone; and, (2) A more comprehensive understanding and control of the range of SCF processing effects, particularly on aqueous‐based macromolecular protein therapeutics, is necessary for SCF precipitation to gain full acceptance as a viable option for secondary manufacturing of biopharmaceutical powders. © 2000 John Wiley & Sons, Inc. Biotechnol Bioeng 67: 457–464, 2000.

Supercritical carbon dioxide was used as an antisolvent for producing recombinant human immunoglobulin G (rIgG) particulate powders. Liquid carbon dioxide (CO(2)) was premixed with ethanol to create a single-phase, modified supercritical fluid (SCF). The modified SCF was then vigorously mixed with a pharmaceutically acceptable, aqueous formulation of rIgG, and the mixture was immediately atomized into a pressurized vessel where rapid expansion of the modified SCF extracted the aqueous phase, resulting in precipitation of the protein powder. The process was reproducible, and resulting powder products were characterized by their aqueous solubilities, and the spectroscopic profile, molecular integrity, and antigen binding activity of the individual soluble fractions. Molecular integrity was assessed via size-exclusion high-performance liquid chromatography (SEC), whereas antigen binding activity was determined using an enzyme-linked immunosorbent assay (ELISA). Attempts to characterize particle size and morphology were confounded due to the extremely deliquescent nature of the powders, causing them to absorb moisture rapidly and become gummy. Operational conditions were optimized to a point which yielded powders that were completely soluble, and had ultraviolet (UV) spectroscopic and SEC profiles indistinguishable from those of the reference standard starting solution from which the powders were derived. Antigen binding activities of the powders, however, were

It is frequently necessary to ensure that pharmaceutically active materials of low aqueous solubility meet minimum requirements for specific surface in order to achieve their desired therapeutic activity. The paper describes work on a compound with a very slight aqueous solubility where the need for control over specific surface is indicated. It is customary and often convenient to establish the relationship between surface area and particle size, and to determine the latter, as modern instrumentation allows the rapid measurement of size distribution. For this compound this approach was not feasible as particle size determination proved difficult. The determinations were therefore made directly as surface area measurements using the dynamic flow technique based on the BET nitrogen adsorption technique. Laboratory scale preparations of the compound achieved a sufficiently high specific surface consistent with small particle size, and crystal shape. As these materials were used in early bioavailability studies, it was necessary to achieve a comparable specific surface on scale-up. This initially proved difficult to achieve. A programme of work has been conducted to establish the causes of the lower specific surface observed in these preparations, and to examine the conditions of the precipitation isolation process. The conclusions from the study allowed successful further scale-up of the precipitation to pilot plant scale, achieving acceptable specific surface values.