Novel theranostic agents containing carborane groups for the treatment of neurotoxic Amyloid aggregates by Neutron Capture Therapy guided by Magnetic Resonance Imaging (MRI)

In this thesis, boron-containing curcuminoids were developed to target amyloid-β (Aβ) aggregates to treat Alzheimer’s Disease (AD), with the goal to integrate molecular imaging and Neutron Capture Therapy (NCT) within a single chemical platform. The study focused on the design and characterization of curcumin-based ligands to perform 19F-MRI preserving affinity for Aβ species, and high-affinity boronated probes to translate NCT into a radiotherapy treatment for Aβ aggregates. Fluorinated curcuminoids bearing nine or eighteen equivalent 19F nuclei linked with different spacers from the curcumin scaffold were synthesized. Among them, Curc-Glu-F9 combined measurable 19F-NMR signal, favorable aqueous solubility, and retained affinity for Aβ fibrils. Notably, its 19F signal was detectable in the presence of monomers and oligomers but disappeared upon binding to mature fibrils, likely due to restricted fluorine mobility and signal broadening. This behavior suggests the potential to detect soluble Aβ species, implicated in early neurotoxicity. Although in vivo validation remains necessary, these results support the feasibility of rational probe design for early amyloid detection. The intrinsic sensitivity limitations of 19F-MRI, however, highlight the need for optimized pharmacokinetics and delivery strategies. To explore therapeutic modulation, a carborane-containing curcuminoid (Phoenix) was synthesized for boron neutron capture therapy (BNCT). Phoenix displayed nanomolar affinity for Aβ fibrils. Upon neutron irradiation, a significant reduction of fibrillar material (~34% at the highest fluence) was observed, whereas irradiation alone produced no structural changes. Spectroscopic and mass spectrometric analyses revealed selective oxidation of histidine residues, suggesting a mechanistic link between localized high-LET particle deposition and destabilization of β-sheet packing. These findings demonstrate that neutron capture can be redirected toward defined protein targets through high-affinity ligands, enabling controlled modification of pathological aggregates. Given the hydrophobicity of Phoenix, formulation strategies were investigated. Inclusion complexes with HP-β-cyclodextrin and TMA-poly-β-cyclodextrin markedly improved aqueous solubility. However, only the HP-β-cyclodextrin formulation preserved selective boron accumulation in amyloid-rich brain regions in vivo following intranasal administration, due to its lower complex stability compared to the cyclodextrin polymer. Structural studies further showed that boron clusters actively contribute to amyloid binding: metallacarboranes exhibited strong affinities, whereas highly hydrated clusters such as dodecaborate, or conjugation with Gd-DOTA, reduced interaction. Overall, this work establishes mechanistic proof-of-concept for amyloid-directed neutron capture and demonstrates that curcuminoids can be functionalized for imaging and therapeutic purposes. While substantial challenges remain, including in vivo validation, blood–brain barrier penetration, and clinical implementation of neutron-based strategies, the results provide a foundation for the development of image-guided molecular interventions targeting amyloidosis