Sub Topic | Secondary Topic: Analytical Methodology | HPLC (Biopharmaceutical Molecule)
Authors: Paola Bisol, Aptuit (Presenting Author); Marina Galvani, Aptuit (Main Author); Adam Grobin, Neurocrine Biosciences, Inc.; Arthur Shurtleff, Neurocrine Biosciences, Inc.
Presenting Author: Paola Bisol
Purpose: During assay and impurity testing of a New Chemical Entity (NCE) and related products atypical chromatograms were generated where a new peak, never observed before, was detected. The peak was not consistently present in different analytical sequences run with the same method on different days and analysis set-ups. However, when present, the peak was observed in the same amount in chromatograms of different API samples. Therefore an investigation was raised to confirm that the peak was not an impurity present in the API, while it was instead generated during the chromatographic analysis. The scope of the investigation was also identifying the root cause in order to avoid the presence of this peak in the chromatograms.
Methods: Assay and impurity analysis is performed by reverse phase HPLC using water –acetonitrile-TFA mobile phases and water-acetonitrile diluent. Supporting investigations were conducted by direct injection and HPLC hyphenated mass spectrometry (ion trap and TOF detectors) as well as NMR.
Results: The chromatographic peak under investigation was tested by LC-MS and a molecular weight of + 32 amu vs the active ingredient was found. A slightly different DAD spectrum (from the API) was also recorded, thus indicating that the peak was due to an actual chemical entity (not just a chromatographic baseline effect), correlated with the input API. Accurate mass measurements by LC-MS TOF suggest that two Oxygen atoms were added to the molecular structure of the active ingredient. Additional investigations by LC-MSn and LC-MS in D2O were performed in order to assign a tentative structure to this molecule, but a definitive structure could not be assigned (characterization data not shown). Direct injection in MS of the same sample, which resulted in a chromatographic trace containing about 2% a/a of this peak, did not show the presence of this +32 m/z ion in the mass spectrum. Similarly, the NMR analysis of the same active ingredient batch did not show the presence of any additional unknown drug related component. Both direct injection MS and NMR analyses could, on the contrary, detect the presence of a known drug related impurity present in the batch at about 0.5%. This can be considered as an internal control to demonstrate that, if any additional component had been present in the sample at levels greater than 0.5 %, it would have been detected. The latter evidences support the hypothesis that the observed peak is likely due to an oxidative degradation product of API, not present in the analytical sample, but formed during the chromatographic analysis. In fact, the impurity is observed in variable amounts in different runs, but it is consistent throughout the same run, even for samples pertaining to different input API batches or to different stability conditions. The shape of the peak (with pronounced fronting) might also indicate that the compound is not present at injection time, but might be formed upon analysis time. The analysis conditions were therefore reproduced and stressed in order to force API degradation. A sample solution, where API is dissolved at 0.5 mg/mL in 50% water 50% acetonitrile, was spiked with 0.05% TFA (to simulate the chromatographic mobile phase conditions) and with 1% H2O2 (to simulate the presence of any oxidative peroxide residuals in the analytical system) and stored for 5 days at room temperature. A higher amount of the degradation product was observed in this sample with respect to an equivalent control sample stored in the absence the acidic/oxidative environment. A number of non-pear reviewed publications (e.g. peptide mapping trouble shooting guides) caution that ghost peaks may result from oxidation of aged TFA. Oxidizing residuals in TFA were postulated as the root cause for the generation of this peak. An HPLC run was performed in absence of TFA and no peaks related to this degradation product were detected. Additionally, different HPLC runs were made using different TFA batches and the degradation product was generated only when one particular batch was used. This specific TFA batch has to be considered as a necessary but not sufficient condition to generate the degradation product, since the observed variability in different set-ups with the same mobile phase remains not explained.
Conclusion: The formation of an oxidative degradation product of an active ingredient during the chromatographic analysis was investigated. The combination of the acidic condition and one specific TFA batch were identified as potential root cause and corrective actions were set-up to avoid the generation of the peak during routine analysis.
See attached abstract pdf for images.