Enabling adoption of 2D-NMR for the higher order structure assessment of monoclonal antibody therapeutics.

The increased interest in using monoclonal antibodies (mAbs) as a platform for biopharmaceuticals has led to the need for new analytical techniques that can precisely assess physicochemical properties of these large and very complex drugs for the purpose of correctly identifying quality attributes (QA). One QA, higher order structure (HOS), is unique to biopharmaceuticals and essential for establishing consistency in biopharmaceutical manufacturing, detecting process-related variations from manufacturing changes and establishing comparability between biologic products. To address this measurement challenge, two-dimensional nuclear magnetic resonance spectroscopy (2D-NMR) methods were introduced that allow for the precise atomic-level comparison of the HOS between two proteins, including mAbs. Here, an inter-laboratory comparison involving 26 industrial, government and academic laboratories worldwide was performed as a benchmark using the NISTmAb, from the National Institute of Standards and Technology (NIST), to facilitate the translation of the 2D-NMR method into routine use for biopharmaceutical product development. Two-dimensional 1H,15N and 1H,13C NMR spectra were acquired with harmonized experimental protocols on the unlabeled Fab domain and a uniformly enriched-15N, 20%-13C-enriched system suitability sample derived from the NISTmAb. Chemometric analyses from over 400 spectral maps acquired on 39 different NMR spectrometers ranging from 500 MHz to 900 MHz demonstrate spectral fingerprints that are fit-for-purpose for the assessment of HOS. The 2D-NMR method is shown to provide the measurement reliability needed to move the technique from an emerging technology to a harmonized, routine measurement that can be generally applied with great confidence to high precision assessments of the HOS of mAb-based biotherapeutics.

Clearance of extractables and leachables from single-use technologies via ultrafiltration/diafiltration operations.

Quantifying the clearance of extractables and leachables (E/L) throughout ultrafiltration/diafiltration (UFDF) operations allows for greater flexibility in the implementation of single-use technologies in steps upstream of the UFDF process. A proof-of-concept study was completed in which the clearance of 7 E/L from single-use technologies (trimethylsilanol, hexanoic acid, butyrolactone, t-butyl alcohol, caprolactam, acetonitrile, and benzyl alcohol) in four representative proteins were measured and monitored during the UFDF process using quantitative NMR. This study demonstrated that the defined E/L spiked into a variety of protein solutions can be cleared to <1 ppm by 9 diavolumes from a maximum initial load concentration of 1,000 ppm. However, in some cases a rebound effect was observed in the recovered pool to >1 ppm, which is explained in detail. The overall clearance trend observed for both buffer control and protein-containing solutions resembled the ideal clearance trend where no apparent interactions were observed between E/L with the protein, UFDF system, or with other defined E/L which may be present in the system. Additionally, the UFDF system is capable of clearing these potential E/L from single-use technologies below 1 ppm irrespective of initial concentrations in the load (1,000 or 100 ppm), independently from the type of protein. In general, mass recoveries were within ±15% of each spiked compound in protein solutions and their respective buffer controls, suggesting spiked E/L do not interact strongly with protein. By demonstrating the product independent clearance trends of the spiked E/L across UFDF, these results will contribute to the simplification of the E/L toxicology assessment and allow modular manufacturing approach for single-use technologies in biopharmaceutical manufacturing. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 32:718-724, 2016.

The effects of membrane filters used in biopharmaceutical processes on the concentration and composition of polysorbate 20.

Polysorbate 20 (PS-20) is often included in the formulation for therapeutic proteins to reduce protein aggregation and surface adsorption. During the production process of therapeutic proteins, various membrane filters are used to filter product pools containing PS-20. The purpose of this study is to quantify the effects of these membrane filtration processes on the concentration and composition of PS-20. A quantitative understanding of this process provides the knowledge base for better controlling the consistency of formulation excipients in drug products. PS-20 solutions (without protein) were filtered through either 0.2 µm sterilizing filters or membrane filters with 30 kDa MWCO. The concentration of PS-20 was measured by a mixed-mode chromatography method and a nuclear magnetic resonance spectroscopy (NMR) assay. The composition of PS-20 was characterized by (1) H-NMR and a reverse-phase chromatography method. Non-specific adsorption of PS-20 on both the sterilizing filter and 30 kDa MWCO membrane filter was quantified. Composition of PS-20 was altered after 30 kDa MWCO membrane filtration, possibly because the different interactions between heterogeneous PS-20 components and the 30 kDa MWCO membrane were not uniform. As a result, the retentate after the 30 kDa MWCO membrane filtration step contains no POE sorbitan and increased amount of POE sorbitan di-esters and tri-esters.

Trisulfide modification impacts the reduction step in antibody-drug conjugation process.

Antibody-drug conjugates (ADCs) utilizing cysteine-directed linker chemistry have cytotoxic drugs covalently bound to native heavy-heavy and heavy-light interchain disulfide bonds. The manufacture of these ADCs involves a reduction step followed by a conjugation step. When tris(2-carboxyethyl)phosphine (TCEP) is used as the reductant, the reaction stoichiometry predicts that for each molecule of TCEP added, one interchain disulfide should be reduced, generating two free thiols for drug linkage. In practice, the amount of TCEP required to achieve the desired drug-to-antibody ratio often exceeds the predicted, and is variable for different lots of monoclonal antibody starting material. We have identified the cause of this variability to be inconsistent levels of interchain trisulfide bonds in the monoclonal antibody. We propose that TCEP reacts with each trisulfide bond to form a thiophosphine and a disulfide bond, yielding no net antibody free thiols for conjugation. Antibodies with higher levels of trisulfide bonds require a greater TCEP:antibody molar ratio to achieve the targeted drug-to-antibody ratio.

Quantitation and characterization of process impurities and extractables in protein-containing solutions using proton NMR as a general tool.

The ability to detect and quantitate a variety of components in solution has become increasingly important in carrying out efficient and rigorous validation studies for biopharmaceutical manufacturing processes. Here, we demonstrate the general applicability of NMR spectroscopy for the identification and quantitation of leachables and other impurities in protein-based drugs, at low levels previously unattainable in protein-containing solutions. With improved NMR technology (i.e., CryoProbes) and the application of a Carr-Purcell-Meiboom-Gill pulse sequence (CPMG) to attenuate protein signals, we have been able to use NMR to quantify impurities in a protein-based biopharmaceutical product at ~1 µg mL(-1) . The data indicate that NMR spectra can be used to quantitate a range of impurities, from small molecule components to higher molecular weight leachables, without removing protein from solution. Furthermore, quantitation of impurities by NMR is reliable and accurate enough for biopharmaceutical process validation, even for high molecular weight extractables whose structures are not precisely known.