Monoclonal antibodies (mAbs) are a major focus for the biopharmaceutical industry and represent one of the most promising and exciting areas of drug development. These proteins can be engineered to recognize almost any antigen and can be mass-produced on a large-scale.
However, to date, it has been largely unfeasible to characterize the high-resolution 3D structure of mAbs with existing techniques due to their large size. For example, nuclear magnetic resonance (NMR) spectroscopy produces too many overlapping resonances that are difficult to interpret. Additionally, 2D NMR of biomolecules normally requires isotopic labeling because the isotopes 13C and 15N are so low in natural abundance.
Characterizing the 3D structure of mAbs is incredibly important for safety and efficacy as their structure is closely related to their function. Incorrect protein folding can lead to side effects, such as unwanted immune responses, so methods to verify the structure of mAbs in a quality control setting are very much needed.
A recent study by Luke Arbogast and colleagues showed how advances in NMR spectroscopy are beginning to lead to solutions to this challenge.
In this study, the researchers chose to look just at methyl residues as a way to “fingerprint” the structure of a candidate mAb called RM8670. They performed their experiments without isotopic labeling using a Bruker Avance III 900 MHz spectrometer. This was equipped with triple-resonance cryoprobes – a technological advance that has helped make natural isotope abundance spectroscopy more plausible. Looking at methyl groups also helped them take advantage of the greater natural abundance of the 13C isotope (1.1%) compared with the 15N isotope (0.37%).
Performing 2D NMR on intact samples of their mAb using a sequence known as gsHSQC (gradient-selected, sensitivity-enhanced heteronuclear single quantum coherence) allowed the researchers high enough quality spectra to assign and analyse the spectrum peaks. However, their optimized experiment took 12 hours to run.
They therefore repeated it on mAb fragments that had been cleaved into their constituent Fab and Fc domains. Not only did the experiment now only take 4.5 hours to run, but the researchers were able to show that the spectra produced from these fragments closely resembled that produced for the intact mAb. This suggests that the cleaved protein fragments can act as a good proxy for the intact protein in NMR experiments, the team write.
The researchers then found that the experimental time could be reduced even further. They achieved this through combining rapid acquisition techniques – which help to increase the speed of natural isotopic abundance spectroscopy – with non-uniform sampling. Of six combinations tested, the fastest time was 34 minutes using the rapid acquisition technique SOFAST-HMQC with 50% non-uniform sampling.
The researchers, who reported their findings in Analytical Chemistry, explain that developments, such as cryoprobes that facilitate natural abundance spectroscopy and the greater availability of high spectrometer field strengths, make NMR a much more practical tool for quality control settings than in the past. Their results support this and indicate the approach should find a range of uses across the biotherapeutic industry for protein structure assessment, they conclude.