Impact Story: Transitioning a Powerful Analytical Tool into Manufacturing to Improve the Quality of Complex Therapeutics
CDER researchers are evaluating an approach that would harness the potential of mass spectrometry to ensure the quality of complex protein therapeutics to treat cancer and other diseases.
Mass spectrometry (MS) is a powerful analytical tool used in nearly all protein-based biologics license applications as a characterization method to assess a wide variety of protein attributes, including molecular mass, amino acid sequence, and critical chemical modifications. But despite the vast amount of information MS can provide, several barriers have stood in the way of its use for quality control (QC) of protein therapeutics (e.g., antibodies to treat cancer), including a perception that the tool is too complex for non-specialists to use in manufacturing settings.
The multi-attribute method
Recent advances in high-resolution mass spectrometry instrumentation have led to reconsideration of the use of MS for QC of protein therapeutics. Specifically, a method known as the multi-attribute method (MAM) that has allowed MS to be introduced to the quality control environment of manufacturing facilities has been developed.
In MAM, a protein product is digested into smaller units called peptides, which are then separated using a method called ultra-high-performance liquid chromatography. These peptides are then analyzed using MS, which begins by vaporizing a molecule and adding one or more electrical charges to it in a process called ionization. Once the charged peptides are inside the mass spectrometer, they are further separated according to the ratio of their mass and charge, which is measured by a detector to create a mass spectrum. Because each position on the spectrum represents a precise mass-to-charge ratio, peptides can be identified based on the known molecular masses of individual atoms. The relative amount of each peptide corresponds to its peak intensity. To obtain additional information about the protein sequence, the charged peptides can also be fragmented and further separated in an approach known as tandem MS.
A semi-automated data analysis process for MAM has two major components: targeted peak quantification and untargeted peak detection. For targeted peak quantification, the intensities of specified peaks in the mass spectrum, corresponding to specific peptides, are measured. Similar peak sizes of a sample relative to a control sample provide assurance that key attributes are preserved. For example, if levels of a critical sugar molecule to a manufactured protein therapeutic were lower than they should be, certain peaks in the spectrum would be smaller than in a reference sample. Untargeted peak detection uses software algorithms to determine whether any unanticipated peaks appear, indicating the presence of impurities or undesirable changes in the protein.
Evaluating MAM for quality control of protein therapeutics
CDER researchers developed an in-house MAM platform in order to evaluate this method and determine its suitability for quality control in the manufacture of protein-based products. They studied U.S.-approved or foreign-sourced rituximab, a therapeutic antibody (and a relatively complex protein) used to treat certain immune system disorders, including some forms of lymphoma.
The researchers found that MAM successfully distinguished between the domestic and foreign products and could also distinguish lot-to-lot differences in the relative abundance of important product quality attributes. As determined by MAM, the distribution of diverse polysaccharides attached to the protein, often essential to the effectiveness of protein therapeutics, generally agreed with profiling results from the more traditional methods. Other studies have indicated that MAM is generally reproducible and precise for the most abundant attributes, although more work is needed to precisely quantify quality attributes of low-abundance products. CDER scientists also found that several attributes, including amino acid modifications (e.g., by oxidation and deamidation), increased linearly over time when the molecule was subject to forced degradation. These findings suggest that MAM might be used to test the stability of drug products. Further studies are planned to evaluate MAM for assessing long- and short-term stability relative to traditional methods.
Mass spectrometry has the potential to give manufacturers and regulators more detailed information about the critical quality attributes of complex products. CDER’s evaluation of a promising mass spectrometry–based approach is intended to help manufacturers apply this tool toward improved quality control of protein products, including therapeutics for cancer.
Fig. 1. Why is mass spectrometry (MS) so useful to manufacturers for quality control?
(Left panel) In a mass spectrometer, mixtures of molecules (e.g., from a protein digest) are vaporized and ionized (charged). These charged molecules are separated in an electromagnetic field based on their mass and charge. Each molecule is recorded as a separate electrical signal and represented as a peak in what is called a mass spectrum. Each peak corresponds to a precise ratio of mass to charge and, based on this ratio, can be identified as a particular molecule. The size of a peak in a spectrum is proportional to the amount of the molecular fragment.
(Right panel) In their pilot studies evaluating a general MS-based approach that can be applied to quality control in manufacturing, CDER researchers digested rituximab, a protein therapeutic with many types of modifications (as indicated by the different shapes attached to it). The resulting peptides were resolved using a common approach to separating molecules called liquid chromatography. These fragments were then subjected to tandem mass spectrometry, wherein individual fragments separated in a first round of MS are broken up for further analysis in a second round of MS, in order to characterize all of the modifications. In the quality control setting, only the first round of MS is used to measure how much of each of those modifications is present. Computer software is used to identify and align the critical spectrum peaks that need to be quantified as quality attributes from lot to lot — as well as to detect any new peaks that would not be expected. Unexpected peaks could signal impurities or unanticipated changes in the protein. (LC-MS/MS = liquid chromatography-tandem mass spectrometry)