Comparison of analytical methods for antibody conjugates with application in nuclear imaging - Report from the trenches.

Introduction: Monoclonal antibodies (mAbs) are widely used in nuclear imaging. Radiolabelling with positron emitting radionuclides, typically radiometals, requires the incorporation of a bifunctional chelator for the formation of the radiometal-mAb complex. Additionally, mAbs can be conjugated with small molecules capable to undergo bioorthogonal click reactions in vivo, enabling pre-targeting strategies. The determination of the number of functionalities attached to the mAb is critically important to ensure a good labelling yield or to guarantee pre-targeting efficacy. In this work, we compare three different analytical methods for the assessment of average functionalisation and heterogeneity of the conjugated mAbs.

Methods: Two selected mAbs (Trastuzumab and Bevacizumab) were randomly conjugated through lysine residues with 3-10 equivalents p-isothiocyanatobenzyl-desferrioxamine (p-NCS-Bz-DFO) or 20-200 equivalents trans-cyclooctene-N-hydroxysuccinimide ester (TCO-NHS). The DFO- or TCO-to-mAb ratio were determined using three different methods: direct titration (radiometric for DFO-conjugated mAbs, photometric for TCO-conjugated mAbs), MALDI/TOF MS mass analysis (Matrix-Assisted Laser Desorption-Ionization/Time of Flight Mass Spectrometry), and UPLC/ESI-TOF MS mass analysis (Ultra High Performance Liquid Chromatography/Electrospray Ionization-Time of Flight Mass Spectrometry).

Results: Radiometric and photometric titrations provided information on the average number of DFO and TCO functionalities per mAb respectively. MALDI/TOF MS provided equivalent results to those obtained by titration, although investigation of the heterogeneity of the resulting mixture was challenging and inaccurate. UPLC/ESI-TOF MS resulted in good peak resolution in the case of DFO-conjugated mAbs, where an accurate discrimination of the contribution of mono-, di- and tri-substituted mAbs could be achieved by mathematical fitting of the spectra. However, UPLC/ESI-TOF MS was unable to discriminate between different conjugates when the smaller TCO moiety was attached to the mAbs.

Conclusions: The three techniques offered comparable results in terms of determining the average number of conjugates per mAb. Additionally, UPLC/ESI-TOF MS was able to shed a light on the heterogeneity of the resulting functionalised mAbs, especially in the case of DFO-conjugated mAbs. Finally, while using a single analytical method might not be a reliable way to determine the average functionalisation and assess the heterogeneity of the sample, a combination of these methods could substantially improve the characterization of mAb conjugates.