Development and Application of Strategies in Prodrug Chemistry

Abstract

A novel strain-promoted 1,3-dipolarcycloaddition between trans-cyclooctene and a masked p-azidobenzyloxycarbonyl (PABC) pro-probe is introduced and its applications in a click-to-release strategy is discussed in Chapter 2. The reaction of trans-cyclooctenol (TCO-OH, 59) with a caged 7-hydroxycoumarin pro-probe 56 resulted in a rapid 1,3-dipolar cycloaddition with second-order rates of approx. 0.027 M-1s-1. 1H NMR studies indicated that activation proceeded via a triazoline and imine intermediate, both of which upon rapid hydrolysis released the uncaged probe. Doxorubicin prodrug 58 was found to be non-toxic to cells, however, the cytotoxicity was restored in vitro upon activation with 59. While the rates of the developed click reaction were comparable to some examples of the strained azide-alkyne cycloaddition reactions, it was still too slow to achieve low dosing in vivo.

To improve the kinetics of bioorthogonal prodrug activation, a series of five fluorine substituted-azido-PABC pro-moieties 79-88 (masking 7-hydroxycoumarin and doxorubicin) are reported (Chapter 3). The rates of cycloaddition (with TCO-OH 59) escalated as the number of fluorine substituents on the PABC linker increase, with a tetra-fluoro-substituted pro-probe 81 exhibiting a 10-fold increase in reaction rate over the non-substituted PABC pro-probe 56. Additionally, the number of fluorine substituents determined how fast the drug released from the triazoline and imine intermediates that are formed in situ. Experimental results and ab initio calculations demonstrated that increasing the fluorine substituents lowered the transition-state energy for the conversion of the triazoline to the imine intermediate and also indicated that the rate-determining step could be the hydrolysis of the imine intermediate, potentially due to the drop in the predicted pKa values from 8.5 (for the non-substituted analogue) to 5.1 (tetrafluoro-substituted analogue).

Modified PABC linkers used in Chapter 2 and Chapter 3 are ineffective at caging phenols as the linker is attached to the probe via a carbonate ester group, which is susceptible to chemical and enzymatic hydrolysis. To overcome these shortcomings, three cinnamyl ether linkers that are stable under physiological conditions and have varying rates of release (seconds-to-hours) are reported in Chapter 4. The -methyl-linker 134 can release phenols 7-hydroxycoumarin and etoposide (pKa 7.8 and 9.8 respectively) with a t1/2 < 2.5 min in a 1:1 aqueous-organic solvent. The slower releasing linker (-methyl) was found to release 7-hydroxycoumarin with a t1/2 = 54 min. Additionally, the activation studies for the -methyl linker 134 suggest the generation of a highly stable aza-cinnamyl-methide, which is significantly more stable than aza-PABC-methides generated.

The (aza)-quinone-methides 131a-131c and 163, are the major by-products of the strategies involving PABC-self-eliminating linkers (Chapter 2 – Chapter 4). They have been widely underestimated in regards to their toxicity and are assumed to interact with the proximal water molecules to form p-aminobenzylalcohol 60. However, the poor nucleophilicity of water (under physiological conditions) can lead to the alkylation of a wide range of advantageous nucleophiles (e.g. DNA) leading to unwanted side effects (e.g. cancer). Chapter 5 reports the design and synthesis of a nucleophile appended PABA-based self-eliminating linker 167. Upon activation, the attached model-probe is released by a 1,6-elimination. The nitrogen nucleophile that is appended to the linker trapped the generated azaquinone-methide, forming a tetrahydroisoquinolone that could be isolated by HPLC and detected using MS. The proof-of-principle studies indicate that the nucleophile trap is capable of mopping-up the (aza)quinone-methide. In summary, this thesis describes the development of a number of novel self-immolative strategies that can be triggered by biorthogonal or other reactions for use in produg science.