Abstract
Bioorthogonal chemistry refers to any chemical reaction that can occur inside a living system without interfering or altering other cellular processes. Over the last two decades, this concept has been researched extensively with numerous applications, such as live cellular labelling and imaging, protein spatial and temporal control, materials science, prodrug activation, and drug delivery. This thesis focuses on the bioorthogonal reaction between aryl azides and trans-cyclooctenes (TCO), driven by a strain-promoted 1,3-dipolar cycloaddition. Methods to further modify reaction and subsequent release rates were investigated (Chapter 2). Then, two biological applications were evaluated; as a technique for cancer specific prodrug activation (Chapters 3 and 4), and as a tool to aid the design and development of new antimicrobials (Chapter 5).
Previously, the Gamble group reported a prodrug activation utilising this bioorthogonal reaction. However, it was hypothesised that the second order rate of reaction between unsubstituted benzylic azide and hydroxyl-functionalised TCO 15 (0.017 to 0.027 M-1 s-1) was too slow for in vivo applications. In a subsequent publication, rates were improved by almost an order-of-magnitude with the addition of fluorine substituents on the benzylic self-eliminating linker (0.110 M-1 s-1). However, the rate of drug (doxorubicin 28) release decreased, leading to reduced in vitro cytotoxicity.
Chapter 2 focused on further improving the rate of reaction between aryl azides and TCOs, without subsequently sacrificing the rate of drug release. A series of novel aryl azide self-eliminating linkers, including halogenated, pyridine, cinnamyl, and benzylic substituted analogues were examined. Additionally, cis-dioxolane fused TCO (d-TCO) 17, developed by the group of Joseph Fox, was also investigated as a tactic for increasing the cycloaddition rate. While all modifications led to faster rates of reaction, the most remarkable result was uncovered from the combination of d-TCO 17 and 2,3,5,6-tetrafluorobenzylic ether analogue 77b, which revealed a second order rate of almost two orders-of-magnitude faster than the initial rate (1.041 M-1 s-1). The addition of alpha-methyl benzylic substitution on the tetrafluorobenzylic linker (77c) improved rates of drug release 7-fold compared to corresponding probe without benzylic substitution 77b. This investigation resulted in reaction and release rates that ranged from minutes to days, thus further expanding the potential applications of this bioorthogonal reaction to a variety of biological settings.
In order to investigate cancer specific prodrug activation, TCO was conjugated to epidermal growth factor receptor (EGFR) antibody cetuximab. EGFR is overexpressed in many cancers including breast, colorectal, and non–small-cell lung cancer. Cetuximab conjugated d-TCO (AB-d-TCO 102) showed significant binding to an EGFR expressing melanoma cell line and selective activation of the doxorubicin prodrug 28a (Chapter 3), displaying comparable cytotoxicity to free doxorubicin (28). No significant prodrug activation was observed in a non-EGFR expressing melanoma cell line. In vivo experiments utilising this prodrug activation technique in a murine melanoma model were conducted. Various dosing and injection methods were investigated, however no significant anti-tumour or increased survival was observed, suggesting there was insufficient drug released at the tumour site. This was supported by in vitro release studies that demonstrated only 20% of doxorubicin (28) release occurs following reaction between prodrug 28a and AB-d-TCO 102 (at pH 7.4).
The stability of this product at pH 7.4 suggested that there was potential for drug delivery as an antibody drug conjugate (ADC). The product formed, AB-T-DOX 111, (via triazoline or imine) was stabilised due to hydrophobic interactions with the antibody, but under acidic conditions (pH 4) the release of doxorubicin (28) increased to 35%. Therefore, it was hypothesised that further doxorubicin (28) release may occur following antibody-receptor internalisation and lysosomal degradation. This alternative drug delivery strategy was investigated in Chapter 4. In vitro cytotoxicity and analytical stability assays with the ADC 111 showed promising results, with improved cytotoxicity and drug release compared to control ADC, 112. The ADC 111 did not demonstrate significant anti-tumour efficacy in mice and further work to examine the biodistribution of the ADC is required.
Finally, in Chapter 5, the bioorthogonal reaction between aryl azides and TCO was investigated as a tool for determining compound uptake and accumulation in Gram-negative bacteria. This assay was designed to enable time, and cost effective fluorescence read-outs in place of a previously reported mass spectrometry (MS) assay. This application was found unsuitable for this bioorthogonal reaction due to TCO instability and slow reaction rates. Instead, an assay was developed using a combination of azide-alkyne and tetrazine-alkyne cycloaddition chemistries which were able to improve detection limits and assay reliability. Although sensitivity limits the accuracy of the fluorescence assay compared to the MS assay, the fluorescence assay may be a valuable tool for large library compound screening. Compounds of interest could then be further screened using the MS assay.