(W1030-07-45) A Novel Basic Charge Variant Observed in Late-Stage mAb-1 Identified as High Molecular Weight Species

Purpose: A basic charge variant (referred to as 13.5 min basic peak) was observed by cation exchange-ultra-performance liquid chromatography (CEX-UPLC) in thermal stressed mAb-1 manufactured using the late-stage process (P2) that was not observed in mAb-1 manufactured from the early-stage process (P1). Additionally, there was an increase in %HMW species (referred to as 6.5 min HMW peak) observed at high temperature by size exclusion-UPLC (SE-UPLC). This is unusual since historically, analytical and stability comparability has been demonstrated for molecules manufactured using the different processes. An investigation into the nature of this degradation species was initiated to determine if process changes or changes to the formulation, or a combination of both, are the root cause of the observed basic species. In addition to the process change, the mAb-1 formulation was optimized to minimize the rate of a deglycation reaction that was observed in P1¹, by changing the buffer and pH from histidine pH 6.0 to acetate pH 5.0.  The protein concentration was also increased to meet the clinical needs for a wide dosing range.  An investigation into the nature of this degradation species was initiated to determine if process changes or changes to the formulation, or both, are the root cause of the observed basic species.

Methods: Initial studies on the P2 samples looked at mAb-1 at storage, accelerated and stress conditions at protein concentrations ranging from 5 mg/mL to 182 mg/mL, assessing a variety of attributes including CEX-UPLC and SE-UPLC. After observation of new peaks at stressed temperatures (≥40C), a comparability study examining the early-and late-stage formulations was initiated to determine if the new peaks are novel or in small amounts and thus undetected in the early-stage samples. P1 DS was formulated at 50 mg/mL at pH 5.0 (the P2 formulation) and P2 DS was formulated at 50 mg/mL at pH 6.0 (the P1 formulation). As experimental controls, the P1 material was also formulated at 50 mg/mL at pH 6.0 (P1 formulation) and the P2 material was formulated at 50 mg/mL at pH 5.0 (P2 formulation). All formulations were then incubated at a stress temperature of 45°C for 1 month. In addition to assessing the stability by SE-UPLC and CEX-UPLC, the material was also fractionated and enriched. The enriched samples were assessed by native liquid chromatography-mass spectrometry (LC-MS) and later also were screened by size exclusion chromatography-mass spectrometry (SEC-MS) analysis. Statistical analysis includes Pearson correlation test and analysis of covariance (ANCOVA).

Results: The 13.5min basic and 6.5min HMW peaks were both shown to be temperature- and protein concentration-dependent and the relative amounts and rates of formation were correlated. We concluded that these peaks represented the same entity. We further demonstrated that when P1 mAb-1 was formulated in the P2 formulation, the same basic and HMW species seen in the P2 mAb-1 samples were observed, albeit in lesser quantities. We determined that the P2 samples in their native pH 5.0 formulation generated the largest increases in HMW and basic species formation, while the P1 samples in their native pH 6.0 formulation generated the least. In order, from most generation to least, the four conditions were as follows: P2 pH 5 > P1 pH 5 > P2 pH 6 > P1 pH 6. This demonstrates that pH plays a role, with the two pH 5.0 formulations generating more basic and HMW species than the samples in the pH 6.0 formulation. Importantly, it also shows that process plays a role as well, with the late-stage P2 process samples generating more basic and HMW species than the early-stage P1 process samples when at the same protein concentration and in the same buffer. Based on analysis of the fractionated material by peptide mapping and SEC-MS analysis, it was determined that there is not a new impurity in the mAb-1 late-stage process, but that process and formulation both play roles in the relative amount of that species observed.  Additional investigations into the nature of the HMW species suggest that there are greater levels of non-glycosylated species that are present in the P2 samples and those species preferentially aggregate, specifically in the 6.5min HMW peak.

Conclusion: We determined that the 13.5min and 6.5min peaks, while formulation-, temperature- and process-dependent, are the same species, and are present in both P1 and P2, and are not indicative of a new species. The results also indicate that non-glycosylated variants, only observed in the P2 samples, are the driver of the difference between the mAb derived from P1 and P2 processes, leading to the increased propensity for aggregation under stress conditions observed in the mAb-1 P2 samples. Additionally, as the 6.5min and 13.5min peaks are the same species and only appeared in highly thermal stressed samples they therefore would not be observed under normal storage and handling conditions for mAb-1.  As a result, the risk of meaningfully impacting the quality, potency or safety is low.

References: 1. Xu X, O’Callaghan JA, Guarnero Z, Qiu H, Li N, Potocky T, Kamen DE, Graham KS, Shameem M, Yang T-C. Low pKa of Lys promotes glycation at one complementarity-determining region of a bispecific antibody. Biophysical J. 2022; 121:1-13. doi: 10.1016/j.bpj.2022.02.002

Acknowledgements: The manuscript of the research might be submitted for journal publication before the AAPS-2025 PHARMSCI 360 meeting in November.


Basic and HMW peaks observed in late-stage samples but not in early stage process samples. A) CEX-UPLC and SE-UPLC unstressed (black dotted) and stressed (blue solid) traces overlaid for early- and late-stage samples. A new basic peak at 13.5min and a new HMW peak at 6.5min were observed for the stressed late-stage samples. B) CEX-UPLC and SE-UPLC overlaid traces of late-stage samples at varying protein concentrations, with their corresponding graphs displaying the peak % areas below, demonstrating these two peaks are protein concentration-dependent. C) Pearson correlation test demonstrates statistically significant correlation between the 13.5min and 6.5min peaks at all four protein concentrations.


Comparability study results. P1 in P1 buffer (P1 pH 6), P1 in P2 buffer (P1 pH 5), P2 in P1 buffer (P2 pH 6) and P2 in P2 buffer (P2 pH 5) all formulated at 50 mg/mL and then incubated under stress conditions in order to determine if the 13.5min basic peak and the 6.5min HMW peak are observable in samples from both P1 and P2 and in both formulations. A) CEX-UPLC overlaid traces of the 13.5min peak and SE-UPLC overlaid traces of the HMW species, with their corresponding graphs showing the peak % areas below, after incubation for 1 month at 45°C. Both peaks are observable in all four samples, with the greatest intensity as follows: P2 pH 5 > P1 pH 5 > P2 pH 6 > P1 pH 6. B) Analysis of covariance (ANCOVA) between the 13.5min basic and 6.5min HMW peaks of the four formulations following incubation for 1 month at 45°C. C) Pearson correlation test demonstrates statistically significant correlation between the 13.5min and 6.5min peaks.


Late-stage samples are enriched with non-glycosylated species that preferentially aggregate. LC-MS results indicate non-glycosylated species are present at a higher amount in P2 samples compared to P1 samples. SEC-MS results indicate that non-glycosylated species were preferentially depleted from the mAb-1 P2 monomer population and enriched in the 6.5min HMW peak following thermal stress. NG, non-glycosylated; FP, fully glycosylated; PG, partially glycosylated.