Automated Data Processing and Analysis for Quality Monitoring of Biotherapeutics by Multi-attribute Method (MAM)

June 4, 2018

Presented at the ASMS 2018, San Diego, CA, USA

Production of biopharmaceuticals requires costly process monitoring strategies and quality systems to ensure final product quality. A number of critical quality attributes (CQAs) — typically measured using a variety of analytical techniques — are used to analyze biopharmaceuticals before release. However, CQAs are generally related to physical properties of the biopharmaceutical and do not characterize the product at the molecular level. As a result, biopharmaceutical producers are exploring MS-based multi-attribute methods (MAMs), which enable measurement of multiple quality attributes in a single test using a single technology, and provide detailed insights at the molecular level. Such methods are especially attractive because they offer the potential of reducing development and manufacturing costs while at the same time, increasing product quality and safety.

 

In all analyses, raw LC-MS data from enzymatically digested samples was loaded directly into the Genedata Expressionist® software platform (Genedata AG, Basel, Switzerland). Where applicable, samples and standards were prepared in parallel on the same day using the same reagents. In an initial test for system suitability, custom scripts were used to confirm that the input data conformed to preset mass accuracy, signal-to-background ratio, and RT shift criteria. Subsequent data analyses were tailored to meet the requirements of the relevant CQA.

 

The system suitability test was able to provide confirmation that the performance of the analytical instrumentation — and the quality of the data that it delivered — was adequate for the intended analyses. Dedicated data processing workflows could be tailored to measure the CQAs for a given biomolecule. Optimized data processing was applied to large data sets and execution times scaled linearly with the number of samples. Multiple quality attributes — such as the presence, location, and relative amount of amino acid modifications in the target molecule — could be determined from a single data set. Comparison of test and reference samples enabled the identification of potential contaminants. The setup of a data analysis workflow that was used to monitor the output of a bioreactor in real time demonstrated that an implementation of the MAM approach created using the software platform was amenable to automation.  The availability of a compliance module offering GxP functionalities — such as audit trails, electronic signatures, and data- and user-management — is intended to facilitate the deployment of such MAM implementations in regulated environments.



Back to list