Measurement of ADC Stability in Plasma and Serum via Mass Spectrometry Analysis
Mass Spectrometry Manager
Colin McKee
Head of Technical Services
IntroductionListen Now
Biologics represent a growing class of pharmaceuticals being developed to target a variety of disease and illness vectors. Antibody drug conjugates (ADCs) combine a monoclonal antibody (mAb) with a cytotoxic drug via a linker. This allows the mAb to target a specific cell type or antigen, and deliver a potent molecule without affecting surrounding cells. There are currently twelve approved ADCs with a variety of linkage chemistries of varying specificities.
Stability of biologics is a critical quality attribute (CQA) affecting the levels of dosing and efficacy. Measuring stability can be complicated by the inherent heterogeneity present within the protein, in addition to linker and payloads of ADCs. Mass spectrometry (MS) is an effective technique for these measurements, as loss of payload can be measured on the protein level or as a released drug. Intact measurements have the benefit of minimising sample handling after incubation compared to enzymatic release methods, however, these measurements are not as quantitive. When measuring these samples for stability, either in plasma or during storage, robust analytical methods are imperative to monitor specific CQAs.
Read on as we explore a series of MS approaches for measuring ADC stability with varying levels of sample processing, and then compare their results.
Click the audio buttons throughout to hear more.
The biologics market is expected to grow at a 10.3% CAGR through 2030.1
As of 2023, there are 400+ ADCs in the global pipeline.2
Fig. 1 Structure of val-cit PABC MMAE payload: The val-cit domain is designed to be cleaved by cathepsin leading to the immolation of the spacer group leaving the active warhead free to block tubulin polymerisation
Fig. 2 Hydrophobic Interaction Chromatography (HIC) UV Spectrum at 256nm Fig. 3 PLRP UV spectrum at 214 nm DAR was calculated by HIC as 4.4 and PLRP as 4.5
MethodsListen Now
Here, we outline various analytical strategies to determine the stability of a trastuzumab val-cit PABC monomethyl auristatin E (MMAE) conjugate (DAR 4.5). The samples were incubated in whole plasma (Sprague Dawley (BioIVT)) or IgG depleted serum (Sprague Dawley (BioIVT)) for time periods up to seven days. The ADC was then captured on protein A magnetic beads (BioRad) and washed with PBS.
For the released drug experiments, the beads were exposed to papain to cleave the linker region, whereas for the drug to antibody ratio (DAR) experiments, the ADC was released from the beads intact with 20mM glycine and 0.1% acetic acid. Data was collected over two pooled biological replicates and three technical replicates. Errors are reported based upon standard deviation of technical replicates.
All data was collected on a quadrupole time-of-flight high resolution mass spectrometer linked to an HPLC with buffers of water and acetonitrile each with 0.1% formic acid. The released drug method was conducted with a multiple reaction monitoring (MRMHR) approach monitoring the 718 m/z ion with an internal standard of monomethyl auristatin F (MMAF) (925 m/z) with a C18 column (1.7 μm, 50 x 2.1 mm).
The DAR method was conducted on a C8 column (3.6 μm, 50 x 2.1 mm) for the reverse phase measurement and a dSEC-2 column (1.8 μm, 50 x 2.1 mm).
In addition, the native analysis was performed with 400 mM ammonium acetate as the mobile phase.
Reduced drug-antibody ratio (DAR) analysis Listen Now
Reduced DAR analysis does not require an enzymatic incubation, only a chemical reduction post capture which can be performed either before or after release from Protein A. The data is semi-quantitive, allowing relative measurements between conjugated and unconjugated species across time points. It also requires an increase in sample per injection ~100 ng of conjugate on column.
The ADC at t=0 was observed in H1 through to H3 with very little H0; DAR was calculated at just over 4.5, matching hydrophobic interaction chromatography (Figure 2). Figure 4 shows the total ion chromatogram (TIC) of all the subunit peaks after reduction at t=0 (blue) and after seven days incubation in IgG depleted serum (red). Overall, a decrease in all drug bound species and an increase in L0 and H0 was observed. Payload loss is due to reverse Michael reaction and not cleavage of the val-cit linker. The overall ADC signal decreased over time in plasma, so samples have been normalised against L1 on both TIC and MS for comparison.
Figure 5 shows the normalised mass spectrum of the first glycoform of H0 and H3. We can clearly track the decrease and oxidation of H3 and the corresponding increase of H0. The overall DAR of both light and heavy chains including glycoforms and modifications was recorded in BPV Flex and the output is displayed in Figure 6 where we see a decrease from DAR ~4.5 to DAR ~3 over seven days.
Fig. 4 TIC of reduced trastuzumab vcE conjugate before (blue) and after (red) incubation; separation between each conjugated form is observed with changes in extremes after incubation. Fig. 5 Mass Spectrum of trastuzumab vcE H0 and H3 main glycoforms; overlaid spectra show changes in relative intensity and oxidation corresponding with incubation in serum. Intensity is normalised versus L1
Fig. 6 Structure of val-cit PABC MMAE payload: The val-cit domain is designed to be cleaved by cathepsin leading to the immolation of the spacer group leaving the active warhead free to block tubulin polymerisation
Intact native DAR analysis Listen Now
Native DAR analysis requires the least amount of post incubation sample handling, with only a buffer exchange required before introducing the sample to the LC-MS. The data can be simplified with the removal of glycans through Endo-S, but this step is not required. It is also the least quantitive and sensitive, requiring tens of μg’s of conjugate per injection and only providing a DAR. However, the advantages of the intact measurement is that any breakdown in stoichiometry between heavy and light chains can be observed, unlike other methods where this information is lost. By measuring DAR directly, it can be determined if high DAR species are less stable than low DAR species.
Fig. 7 Intact native spectra over incubation time, native MS slightly underestimates DAR due to suppression of hydrophobic species ionisation but a clear trend is visible with loss of DAR 6 species rapidly and a trend towards DAR 2 and 0 over the seven days
Released drug analysis Listen Now
Ionisation of small molecules is much more efficient than for proteins, allowing increased sensitivity and quantitation. In addition, tandem mass spectrometry (MSMS) can be used for increased specificity and reduction in background. This can be especially useful for complex matrices such as plasma and serum. For cleavable linkers, enzymatically releasing the warhead from the ADC allows these advantages to be exploited and the MSMS response can be to a calibration curve via MRMHR. Figure 8 shows the MSMS spectra for released monomethyl auristatin E (MMAE) with two major fragments (686 and 152 m/z), which were used as quantifier and qualifier ions respectively. Figure 9 shows the calibration curve between 100 and 0.1 ng/ml of MMAE (718>686) corrected for an internal standard.
Fig. 8 MSMS fragmentation of MMAE payload ion 686 m/z was used as the quantifier ion with 152 m/z as the qualifier Fig. 9 Calibration curve of MMAE over four orders of magnitude r2 was >0.99 with an internal standard of MMAF at 2.5 ng/ml
Of the plasma incubated samples, 25 μg were captured on Protein A beads, which would give a final amount as ~950 ng of MMAE. Recovery studies estimated capture efficiency to be ~30%. This could be improved by a higher Protein A: protein ratio.
Conclusions Listen Now
Here, we present three different approaches for the measurement of plasma stability of a model ADC, with various levels of sensitivity and sample processing. These methods are applicable to a variety of payloads with differing levels of hydrophobicity and stoichiometry (stochastic and site specific). In addition they will improve with reduced and more standardised DAR.
Intact and native DAR measurements have little sample handling, though may suffer from suppression of high DAR species. In addition, their measurement is carried out in depleted serum rather than plasma, which may lead to reduced metabolic potential. This could be normalised with the use of a specific antigen capture method rather than a generic Protein A. The intact DAR methods allow the nature of drug loss to be identified as either reverse Michael reaction or cleavage of the linker, whereas this information is lost upon enzymatic cleavage.
Released drug analysis has a further incubation at 37°C after capture which may lead to degradation of the sample, though it provides more sensitive and specific data. For complete understanding of the stability process, intact provides data on how the drug can be lost and released, as well as providing the most quantitive data. Further work is planned on total antibody concentration measured by peptide quantitative analysis, free drug analysis and ADC metabolism.
How Sterling can help
At Sterling, we combine extensive expertise in analytical chemistry and mass spectrometry with our Deeside facility’s specialised capabilities in ADC discovery, development and manufacturing. This enables us to select an optimal approach to stability measurement based on our customers’ specific requirements. As with all of our services, our work in mass spectrometry and bioconjugation is backed by close scientific collaboration and true partnership with our customers. If you’re interested in discovering how we can apply our analytical or bioconjugation expertise in your project, speak to an expert.