Synthetic and computational studies towards novel medical imaging agents for PET and MRI
File(s)
Author(s)
Angoh, Shane
Type
Thesis
Abstract
Molecular imaging has a vital role in advancing our understanding of biological processes and treatment monitoring. This thesis explores the synthesis of manganese-based contrast agents (MnBCAs) for dual- modal PET-MRI; the development of predictive tools to aid in the design of both manganese and gallium complexes used in molecular imaging; as well as the modelling of the formation of 89Zr‐doxorubicin complexes inside liposomal formulations such as DOXIL via density functional theory (DFT).
Initial efforts were directed towards the synthesis of novel ligand families for PET-MRI MnBCAs, their subsequent complexation, and testing. Macrocyclic and acyclic ligands containing sulfonamides and amides were explored, with successful Mn2+ complexation achieved with acyclic picolyl amide ligands. DFT calculations were carried out to predict the thermodynamic stability constant of MnBCAs with Gaussian 16; UPW6B95D3/TZVP/IEFPCM(SMD) with two explicit solvent water molecules best predicted the stability constants of the chosen testing dataset (MAE: 1.84 log units, RMSE: 3.72 log units). Furthermore, the possible correlation between MRI-relevant properties and the DFT optimised geometry of 46 MnBCAs was investigated, with no significant correlation found. Machine learning models were assessed to consider features of the geometry, DFT calculated Gibbs energies, and other properties of the MnBCAs for stability constant prediction, and provided predictions which correlated well with experimental values (R2 0.851, MAE: 0.50 log units, RMSE: 0.52 log units).
The identity of potential 89Zr-doxorubicin complexes was modelled via DFT under simulated basic, neutral, and acidic conditions at the M06-2X/Sapporo-TZP-2012/6-31G(d,p)/IEFPCM level. For the acidic conditions inside DOXIL, modelling indicated that Zr(oxinate)4 would almost entirely convert to [Zr(oxinate)(doxorubicinate)2]+ (97.2%) in the presence of doxorubicin.
Finally, a predictive model for the partition coefficient (log P) of gallium complexes with applications in myocardial perfusion imaging was developed. DFT calculations with PW6B95D3/TZVP/SMD predicted log P with excellent correlation to experimental values (R2 0.829).
Initial efforts were directed towards the synthesis of novel ligand families for PET-MRI MnBCAs, their subsequent complexation, and testing. Macrocyclic and acyclic ligands containing sulfonamides and amides were explored, with successful Mn2+ complexation achieved with acyclic picolyl amide ligands. DFT calculations were carried out to predict the thermodynamic stability constant of MnBCAs with Gaussian 16; UPW6B95D3/TZVP/IEFPCM(SMD) with two explicit solvent water molecules best predicted the stability constants of the chosen testing dataset (MAE: 1.84 log units, RMSE: 3.72 log units). Furthermore, the possible correlation between MRI-relevant properties and the DFT optimised geometry of 46 MnBCAs was investigated, with no significant correlation found. Machine learning models were assessed to consider features of the geometry, DFT calculated Gibbs energies, and other properties of the MnBCAs for stability constant prediction, and provided predictions which correlated well with experimental values (R2 0.851, MAE: 0.50 log units, RMSE: 0.52 log units).
The identity of potential 89Zr-doxorubicin complexes was modelled via DFT under simulated basic, neutral, and acidic conditions at the M06-2X/Sapporo-TZP-2012/6-31G(d,p)/IEFPCM level. For the acidic conditions inside DOXIL, modelling indicated that Zr(oxinate)4 would almost entirely convert to [Zr(oxinate)(doxorubicinate)2]+ (97.2%) in the presence of doxorubicin.
Finally, a predictive model for the partition coefficient (log P) of gallium complexes with applications in myocardial perfusion imaging was developed. DFT calculations with PW6B95D3/TZVP/SMD predicted log P with excellent correlation to experimental values (R2 0.829).
Version
Open Access
Date Issued
2024-01
Date Awarded
2024-04
Copyright Statement
Creative Commons Attribution NonCommercial NoDerivatives Licence
Advisor
Long, Nicholas
de Rosales, Rafael T. M.
Sponsor
GlaxoSmithKline
National Institute for Health Research (Great Britain)
Engineering and Physical Sciences Research Council
Publisher Department
Chemistry
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)